CCR Intro.fm - PULP - University of Pretoria

The annual Constitutional Court Review is devoted to analysing the 

work of the South African Constitutional Court in the year preceding 

publication. Every volume contains two long essays, each with 

replies, subject-specific articles, and case comments. This inaugural 

2008 volume examines the major themes and currents reflected in 

the decisions handed down by the Court in 2007. 

CLoSA 

Constitutional Law of South Africa 

Konrad 

Adenauer 

Stiftung 

South African Institute for Advanced Constitutional, 

Public, Human Rights and International Law 

Strap ad FINAL.indd 1 5/27/09 9:57:58 AM 

Pretoria University Law Press 

PULP 

www.pulp.up.ac.za 

ISSN: 2073-6215 

CCR 

CONSTITUTIONAL COURT REVIEW 

VOLUME 1 • 2008 


CCR 

VOLUME 1 

• 

2008 

CONSTITUTIONAL 

COURT 

REVIEW 

Lead essays/respones 

On the uses of interpretive charity: Some notes on application, avoidance, equality and 

objective unconstitutionality from the 2007 term of the Constitutional Court of South Africa 

Frank I Michelman ................................................................................................................................... 

The Constitutional Court, court watchers and the Commons: A reply to Professor 

Michelman on constitutional dialogue, ‘interpretive charity’ and the citizenry as sangomas 

Tshepo Madlingozi ............................................................................................................................... 

Normative pluralism and anarchy: Reflections on the 2007 term 

AJ van der Walt ........................................................................................................................................ 

Legal subsidiarity and constitutional rights: A reply to AJ van der Walt 

Karl Klare ............................................................................................................................................ 

Articles 

‘Oh, what a tangled web we weave ...’ Hegemony, freedom of contract, 

good faith and transformation – towards a politics of friendship in the politics of contract 

Jaco Barnard-Naudé ................................................................................................................................. 

Clearing the intersection? Administrative law and labour law in the Constitutional Court 

Cora Hoexter ............................................................................................................................................ 

Case comments 

Sustainable development in practice: Fuel Retailers Association of Southern Africa v Director- 

General Environmental Management, Department of Agriculture, Conservation and Environment, 

Mpumalanga Province 

Loretta Feris ............................................................................................................................................. 

Fuel Retailers, sustainable development & integration: A response to Feris 

Dire Tladi ............................................................................................................................................ 

Cultural and religious accommodations to school uniform regulations 

Patrick Lenta ............................................................................................................................................ 

The case for religious inclusivism and the judicial recognition of religious associational 

rights: A response to Lenta 

Iain T Benson ...................................................................................................................................... 

Media freedom and the law of privacy: NM & Others v Smith & Others 

(Freedom of Expression Institute as amicus curiae) 

Glenn Penfold & Dario Milo .................................................................................................................... 

Wanted: A principled approach to the balancing of policy considerations Steenkamp NO v 

Provincial Tender Board, Eastern Cape 

Sanele Sibanda .......................................................................................................................................... 

Severing the umbilical cord: A subtle jurisprudential shift regarding children and their primary 

caregivers 

Ann Skelton ............................................................................................................................................. 

1 

63 

77 

129 

155 

209 

235 

255 

259 

295 

311 

335 

351

Editors 

Stu Woolman, Associate Professor, University of Pretoria 

Theunis Roux, Professor, University of New South Wales 

Danie Brand, Senior Lecturer, University of Pretoria 

Editorial Board 

Laurie Ackermann, Constitutional Court Justice Emeritus 

Mary Arden, Lady Justice, Court of Appeal of England & Wales 

Danwood Chirwa, Associate Professor, University of Cape Town 

Sujit Choudhry, Professor, University of Toronto 

Christian Courtis, Human Rights Officer, United Nations Office of the High Commissioner 

for Human Rights 

Javier Couso, Professor, Universidad Diego Portales, Chile 

Charles Fombad, Professor, University of Botswana 

Nicole Fritz, Director, Southern Africa Litigation Centre 

Karthy Govender, Professor, University of KwaZulu-Natal 

Michelo Hansungule, Professor, University of Pretoria 

Karl Klare, Professor, Northeastern University 

Heinz Klug, Professor, University of Wisconsin 

Sandy Liebenberg, Professor, Stellenbosch University 

Frank Michelman, Professor, Harvard Law School 

John Mubangizi, Professor, University of KwaZulu-Natal 

Christina Murray, Professor, University of Cape Town 

Charles Ngwena, Professor, University of the Free State 

Enyinna Nwauche, Associate Professor, Rivers State University 

Cheryl Saunders, Professor, University of Melbourne 

Dire Tladi, Legal Counsel, South Africa, United Nations 

AJ van der Walt, Professor, Stellenbosch University

CONSTITUTIONAL 

COURT 

REVIEW 

(2008) 1 

2009

Constitutional Court Review (2008) 1 

Published by: 

Pretoria University Law Press (PULP) 

The Pretoria University Law Press (PULP) is a publisher at the Faculty of Law, 

University of Pretoria, South Africa. PULP endeavours to publish and make 

available innovative, high-quality scholarly texts on law in Africa that have 

been peer-reviewed. PULP also publishes a series of collections of legal 

documents related to public law in Africa, as well as text books from African 

countries other than South Africa. 

For more information on PULP, see www.pulp.up.ac.za 

Printed and bound by: 

ABC Press 

Cape Town 

To order, contact: 


Faculty of Law 

University of Pretoria 

South Africa 

0002 

Tel: +27 12 420 4948 

Fax: +27 12 362 5125 

pulp@up.ac.za 

www.pulp.up.ac.za 

Cover: 

Yolanda Booyzen, Centre for Human Rights 

ISSN: 2073-6215 

© 2009 

The Constitutional Court Review forms part of the Rule of Law in Africa 

Project funded by the World Bank.

CONSTITUTIONAL COURT REVIEW 

The Constitutional Court Review is a once-a-year journal dedicated to 

the analysis of the Constitutional Court’s decisions of the previous 

year. Its purpose is to provide a platform for high-level academic 

engagement with the jurisprudence of the South African 

Constitutional Court. 

To this end, each issue of the Review contains two lead essays 

exploring broad themes arising from a given year’s jurisprudence 

(each ± 20 000 words), each with its own response (± 5 000 words); a 

number of shorter subject-specific articles (each ± 10 000 words); and 

several case comments that engage more narrowly with a given 

decision of the Constitutional Court (each ± 5 000 words). 

Lead essays are solicited by the editors, as are some of the subjectspecific 

articles and case notes, but for the remainder unsolicited 

contributions are invited. Such contributions must be sent to the 

editors at danie.brand@up.ac.za in MS Word format on or before 

31 May of the year following that on which a contribution focuses. 

Contributions will only be considered if they follow the house style, 

available at www.pulp.up.ac.za. 

iii

SUBSCRIPTION 

Subscriptions through agents or bookshops: 

R380-00 (including VAT but excluding postage and packaging) 

Subscriptions directly through PULP: 


Subscriptions by registered students directly through PULP: 


Postage and packaging charges are: 

R45-00 per subscription 

Subscriptions should be directed to the publisher: 

PULP - Pretoria University Law Press 

Faculty of Law 

University of Pretoria 

Pretoria 

0002 

E-mail: pulp@up.ac.za 

Website: http://www.pulp.up.ac.za 

iv

Lead essays/responses 

TABLE OF CONTENTS 

On the uses of interpretive charity: Some notes on application, 1 

avoidance, equality and objective unconstitutionality from the 

2007 term of the Constitutional Court of South Africa 

Frank I Michelman 

The Constitutional Court, court watchers and the Commons: 63 

A reply to Professor Michelman on constitutional 

dialogue, ‘interpretive charity’ and the citizenry 

as sangomas 

Tshepo Madlingozi 

Normative pluralism and anarchy: Reflections on the 2007 term 77 

AJ van der Walt 

Articles 

Legal subsidiarity and constitutional rights: A reply to 129 

AJ van der Walt 

Karl Klare 

‘Oh, what a tangled web we weave ...’ Hegemony, freedom of 155 

contract, good faith and transformation — towards a 

politics of friendship in the politics of contract 

Jaco Barnard-Naudé 

Clearing the intersection? Administrative law and labour law 209 

in the Constitutional Court 

Cora Hoexter 

Case comments 

Sustainable development in practice: Fuel Retailers Association 235 

of Southern Africa v Director-General Environmental 

Management, Department of Agriculture, Conservation 

and Environment, Mpumalanga Province 

Loretta Feris 

Fuel Retailers, sustainable development & integration: 255 

A response to Feris 

Dire Tladi 

v

Cultural and religious accommodations to school uniform 259 

regulations 

Patrick Lenta 

The case for religious inclusivism and the judicial 295 

recognition of religious associational rights: 

A response to Lenta 

Iain T Benson 

Media freedom and the law of privacy: NM & Others v Smith & 311 

Others (Freedom of Expression Institute as amicus curiae) 

Glenn Penfold & Dario Milo 

Wanted: A principled approach to the balancing of policy 335 

considerations. Steenkamp NO v Provincial Tender Board, 

Eastern Cape 

Sanele Sibanda 

Severing the umbilical cord: A subtle jurisprudential shift 351 

regarding children and their primary caregivers 

Ann Skelton 

vi

ON THE USES OF INTERPRETIVE ‘CHARITY’: SOME 

NOTES ON APPLICATION, AVOIDANCE, EQUALITY 

AND OBJECTIVE UNCONSTITIONALITY FROM THE 

2007 TERM OF THE CONSTITUTIONAL COURT OF 

SOUTH AFRICA* 

* I am indebted to Stu Woolman, and to two anonymous referees, for exceptionally 

helpful, trenchant comments. I hope that what follows shows the benefit of their 

efforts, even if I remain incorrigible on some points of difference. 

** Robert Walmsley University Professor, Harvard University. 

1 

D Davidson Inquiries into truth and interpretation (1984) 197 (Davidson Truth). 

2 S Woolman ‘The amazing, vanishing Bill of Rights’ (2007) 123 South African Law 

Journal 762 (Woolman ‘Amazing’). 

3 

Barkhuizen v Napier 2007 7 BCLR 691 (CC). 

4 Masiya v Director of Public Prosecutions 2007 5 SA 30 (CC); 2007 8 BCLR 927 (CC) 

(‘Masiya’). 

5 

NM v Smith 2007 5 SA 250 (CC); 2007 7 BCLR 751 (CC) (‘NM’). 

6 Woolman ‘Amazing’ (n 2 above) 762 (‘chattering classes’). 

1 

Frank I Michelman* 

Charity is forced on us; whether we like it or not, if we want to 

understand others, we must count them right in most matters. 1 

1 Introduction 

To hear Stu Woolman tell it, disturbing lapses and weaknesses — an 

apparent ‘lack of analytical rigour’ suggesting what could be a 

‘penchant for outcome-based decision-making’ — have been showing 

up recently in the work of a Constitutional Court whose prior record 

of performance has deservedly garnered widespread applause. 2 

Woolman cites as evidence three decisions from the Court’s work in 

the year 2007: Barkhuizen, 3 Masiya, 4 and NM, 5 and suggests that his 

reactions to these decisions are widely shared among South Africa’s 

well-informed Court-followers. 6 He makes a worthy, illuminating, 

formidable case, one that the Court would do well to consult and 

ponder.

2 Uses of interpretive ‘charity’ 

Formidable is not, however, conclusive. For reasons I shall come 

to shortly, I have chosen to devote this space to seeing what might be 

said on the other side, specifically with regard to NM and Masiya, on 

behalf of a Court that, I quite agree with Woolman, has left itself with 

a lot of explaining to do. The controlling opinions in these cases are 

indeed, as Woolman says, ‘thinly reasoned’, 7 if by that we mean they 

are in some respects insufficiently explained. It is, however, another 

question whether these cases have been wrongly or irresponsibly 

managed, as measured by reasonably discoverable, valid considerations 

of law and legal administration. 

In particular, I shall be questioning Woolman’s diagnosis from 

these cases of ‘a court uncomfortable with the direct application of 

the specific substantive provisions of the Bill of Rights’ and ‘in full 

flight from any meaningful engagement with Chapter 2 of the 

Constitution.’ 8 Whether a wider survey of the jurisprudence would 

warrant an over-all diagnosis of an excessive flight from substance is 

a question on which I hazard no judgment here. All I say here is that 

NM and Masiya do not, to my eye, support the diagnosis, nor is that, 

in my view, the best way for us to regard these cases. One feature 

common to both is the Constitutional Court’s seeming gravitation to 

its inherent power to develop the common law in terms of 

Constitution sections 173 and 39(2) — as opposed to its judicial review 

power in terms of sections 8 and 172(1) — when undertaking 

modification of common law rules under pressure from the Bill of 

Rights. Woolman believes the Court moves too freely to the inherent 

power. He associates that tendency, as symptom or cause (or both), 

with excessive flight from substance. I aim to raise a doubt about any 

such connection. 9 

The main controversy over Masiya appears to me to turn, at 

bottom, more on a point of substantive disagreement between the 

Court and Woolman than on any notable disregard for Bill-of-Rights 

substance on the Court’s part. The controversy over NM is more 

complicated, and more centrally my concern in these pages. It 

certainly is true that Woolman and the Court divide over when, if 

ever, the Constitutional Court ought to resort to an ‘indirect’ instead 

of a ‘direct’ application of the Bill of Rights to a common law doctrine 

or rule — the Court, in Woolman’s view, making far too much use of 

7 Woolman ‘Amazing’ (n 2 above) 762, 790 n 51. 

8 Woolman ‘Amazing’ (n 2 above) 783. 

9 

I have nothing to say about another symptom or cause of flight that some might 

mention, to wit, the Court’s readiness on some occasions to move to the 

‘justification’ phase of a Bill of Rights case either without having decided the 

question of a substantive infringement, see, eg, Christian Education South Africa 

v Minister of Education 2000 4 SA 757 (CC) para 27, or, perhaps, having found an 

infringement on the basis of a merely ‘notional’ reading of the right in question. 

See S Woolman & H Botha ‘Limitations’ in S Woolman et al (eds) Constitutional 

law of South Africa (2nd Edition, OS, 2005) ch 34 16-18.

(2008) 1 Constitutional Court Review 3 

‘indirect’. Such a division need not, however — or so I shall contend 

— reflect any reduced or absent sense on the Court’s part of 

responsibility to engage with the substance of Chapter Two and its 

several, rights-naming clauses. It might rather come down to a 

question of doctrinal good-housekeeping on which nothing of 

substance depends. The Court and Woolman are differing, I shall 

suggest, over how best to understand and sort out the respective 

offices of the Constitution’s two paths to judicial revision of the 

common law under constitutional pressure: revision as a remedy for 

constitutional violation pursuant to sections 172(1) 10 and 8, 11 and 

revision in the exercise of judicial powers to develop the common 

law, with a view to promoting the spirit, purport, and objects of the 

Bill of Rights in terms of section 39(2). That filing-system question, I 

shall maintain, is entirely distinct from the one about when and how 

regularly the Court regards itself as on or off the hook for an 

elucidation of one or another of the specific clauses in the Bill of 

Rights. 

In developing these claims, I shall be quite openly engaged in 

filling in passages of exposition and explanation that are missing from 

the Court’s opinions in NM and Masiya, to a degree that may 

sometimes strike readers as excessively indulgent of the Court, if not 

as entirely fanciful. In construing and re-presenting the work of the 

Constitutional Court in these cases, I take myself to be following 

something akin to what linguists and language-philosophers have 

called a ‘principle of charity’. ‘Something akin,’ not the genuine 

article, for this is not a work of philosophy, but rather an intended 

contribution to a lawyers’ kibitz on the work of the Constitutional 

Court. 

The ‘principle of charity’, Wikipedia tells us, 

10 172.(1) When deciding a constitutional matter within its power, a court — 

(a) must declare that any law or conduct that is inconsistent with the 

Constitution is invalid to the extent of its inconsistency; and 

(b) may make any order that is just and equitable ... 

11 8.(1) The Bill of Rights applies to all law, and binds the legislature, the executive, 

the judiciary and all organs of state. 

(2) A provision of the Bill of Rights binds a natural or a juristic person if, and to 

the extent that, it is applicable, taking into account the nature of the right and 

the nature of any duty imposed by the right. 

(3) When applying a provision of the Bill of Rights to a natural or juristic person in 

terms of subsection (2), a court — 

(a) in order to give effect to a right in the Bill, must apply, or if necessary 

develop, the common law to the extent that legislation does not give effect to 

that right; and 

(b) may develop rules of the common law to limit the right, provided that the 

limitation is in accordance with section 36(1) ...


is an approach to understanding a speaker’s statements by interpreting 

the ... statements to be rational and, in the case of any argument, 

rendering the best, strongest possible interpretation of an argument. 12 

Donald Davidson, surely one of the principle’s chief philosophical 

architects and expositors, also calls it ‘the principle of rational 

accommodation,’ and summarises as follows: ‘We make maximum 

sense of the words and thoughts of others when we interpret in a way 

that optimises agreement.’ 13 Davidson meant ‘optimise’ as between 

thinking that the other must be holding to beliefs (and, relatedly, 

aims) that differ drastically from our own (else he couldn’t have said 

what he did), and thinking that we must not have heard him right the 

first time. 14 

The philosophical underpinnings of the charity principle are far 

afield from our concerns here and need not delay us. All that matters 

from that department is the motivation for adherence to the 

principle, which the moniker ‘charity’ does not very well convey. The 

aim of interpretive charity is not generosity toward others, or 

anything like that. It is not to pay homage, deference, or respect to 

our interlocutors, or to avoid giving offense. It is not to demonstrate 

our own good manners, or to toe some Goody Two-Shoes line against 

critiques that are not ‘constructive’. (I hold Stu Woolman’s pull-nopunches 

style of court-watching to be entirely constructive and 

admirable.) No, the aim of ‘charitable’ interpretation is not any of 

those. The aim is to learn. It is aggressively to learn what there is to 

be learnt from puzzles the interlocutors pose to us, by assuming there 

is method in their madness and doing our best to ferret that out, using 

everything else we know or can guess (in part from their likeness and 

kinship to us) about where they are coming from. ‘To see too much 

unreason on the part of others’, Davidson says, is ‘to undermine our 

ability to understand what it is they are so unreasonable about.’ 15 It 

is to risk missing issues that might merit our consideration. 

There are hazards — normative pitfalls — in the way of such an 

approach to the construction of judicial opinions. For one: these 

writings enjoy among us the status of utterances of law by a socially 

recognised organ of authority to say what the law is. And maybe, 

therefore, it really does not do for us to go about pretending the 

writers have said things they did not mean to say (perhaps meant not 

to say), in our effort to make their writings accord with what we take 

12 

13 

14 

http://en.wikipedia.org/wiki/Principle_of_charity (accessed 21 June 2008). 

As above. 

‘[T]here is no deciding ... between the view that the Other has used words as we 

do but has more or less weird beliefs, and the view that we have translated him 

15 

wrong.’ Davidson Truth (n 1 above) 101. 

Davidson Truth (n 1 above) 153.


to be legal reason — it being they, after all, and not we, who are the 

authorised law-sayers. 

Lying just ahead are deep jurisprudential waters into which I do 

not choose to wade on this occasion. I mean only to grant up front the 

possible dangers from a ‘charitable’ approach, while pointing also to 

some possible gains. A close examination of NM and Masiya, from a 

methodological stance of all-but-last-ditch resistance against a 

conclusion of blatant error, failure of nerve, or flight from 

responsibility on the Court’s part, can demonstrate, I believe, some 

benefits of interpretive charity as a working disposition on our part — 

not necessarily suited, of course, to every moment in history or every 

judicial performance we may encounter. Such an examination can 

show how that posture may sometimes aid us in performing our task 

of getting at the truth of law or at any rate — more modestly — of 

developing culturally credible resources for the possible guidance of 

judicial performance in the future, with no less a view (here following 

Theunis Roux) to the judiciary’s ‘legal’ than to its ‘sociological’ and 

‘institutional’ legitimacy. 16 

2 ‘Direct’ and ‘indirect’ application 

Much of what is to follow turns on debates over when, if ever, the 

Constitutional Court ought to resort to an ‘indirect’ instead of a 

‘direct’ application of the Bill of Rights to a common law doctrine or 

rule. Such debates sometimes risk falling into confusion and 

misunderstanding, because not everyone conceives or defines the 

distinction in exactly the same way, and perhaps some of us do not 

conceive or define it uniformly at every turn or in every context. It 

will, therefore, behoove us to survey briefly some of the variant 

senses of this distinction, and, in particular, to specify what Woolman 

— consistently — means by it in the work of his to which what follows 

is something of a rejoinder. 

So far as I am aware, the distinction in South African legal 

discourse between direct and indirect application of the Bill of Rights 

16 T Roux ‘Principle and pragmatism on the Constitutional Court of South Africa’ 

(2008) 7 I-CON 106 (Roux ‘Pragmatism’). In Roux’s vocabulary, the ‘legal 

legitimacy’ of judicial review rises and falls with perceptions of a court’s capacity 

to decide cases according to forms of reasoning acceptable to the legal 

community of which it is a part. Legal legitimacy refers to ‘the plausibility (rather 

than correctness) of a judicial decision or body of decisions according to 

applicable standards of legal reasoning.’ Thus defined, Roux explains, legal 

legitimacy is to be distinguished from two other sociological and institutional 

dimensions of the judiciary’s public standing, which Roux names ‘public support’ 

and ‘institutional security’ meaning the court’s capacity to resist political attacks 

on its independence. As above 109. Roux posits that a court presiding over a 

system of judicial constitutional review does well to act with an eye to all three 

dimensions of legitimacy. See as above 110-11.


first arose in regard to the question of so-called horizontal application 

of the Bill of Rights under the Interim Constitution (although, as we 

shall see, it has outgrown its context of origin). In terms of that 

debate, to apply the Bill of Rights ‘horizontally’ was to bring it to bear 

decisively, in a case in which (a) there was no state party and (b) it 

appeared that the pivotal legal doctrine or rule was (or might be) a 

rule or doctrine of the common law, as opposed to the terms of a 

statute. 17 In such a case, either or both of the following would have 

to be true: A plaintiff bases a claim for relief on a rule or doctrine of 

the common law, and/or a defendant bases a defense (including the 

defense of ‘absolution’ or ‘no cause of action’) on a rule or doctrine 

of the common law. One or the other (or both) of the parties might 

question the compatibility with the Bill of Rights of the common law 

rule or doctrine on which the other relies. If that occurs, and the court 

rules in that party’s favor on that basis, a horizontal application of the 

Bill of Rights will have taken place, under the definition we are now 

considering. 

Consider, now, a case such as NM, in which the plaintiff charges 

the defendant with unauthorised disclosure of private information 

concerning the plaintiff, and the defendant seeks to respond by 

proving that she did not knowingly or intentionally publish without 

authorisation, but rather acted honestly (if mistakenly) in the belief 

that authorisation had been given. How might the Bill of Rights be 

brought to bear on this case, specifically section 14, which guarantees 

a right of privacy to everyone? 

A court might proceed by drawing a standard of care from section 

14, and then measuring the defendant’s conduct by that 

constitutionally prescribed standard, with a view to holding the 

defendant liable for a delict if her conduct fails the constitutional 

test. Alternatively, a court might proceed by asking whether the 

extant common law rule — which absolves publishers of private 

information from tort liability where they did not intentionally or 

knowingly publish without authorisation — is consistent with section 

14, with a view to possibly revising the common law rule in such a way 

that the defendant will be found liable for tortious conduct under the 

common law as revised to make it ‘constitutional’. In the first 

instance, the court applies the Bill of Rights to conduct; in the second 

instance, it applies it to law. 

17 

See Du Plessis v De Klerk 1996 3 SA 850 (CC); 1996 5 BCLR 658 (CC) para 49 

(Kentridge AJ) (‘Du Plessis’) (in effect, defining a ‘vertical’ case as one that 

either involves a state party or turns on a statute); S Woolman ‘Application’ in S 

Woolman et al (eds) Constitutional law of South Africa (2nd Edition, OS, 2005) ch 

31 (Woolman ‘Application’) 18 (similarly describing ‘the traditional, vertical, 

view’); 4 (reporting general agreement that ‘the heart of the application debate 

was whether the common law, when relied upon by a private party in a private 

dispute, was subject to constitutional review.’).


This distinction — between testing conduct by a Bill of Rights 

standard, and testing law by a Bill of Rights standard — may have been 

what some early deployments of the direct/indirect distinction had at 

least partly in mind: application of the constitutional standard to 

conduct is ‘direct’ application, whereas application of it to law is 

‘indirect’ application. If so, that usage of the distinction cannot 

survive today. As we shall see below, Constitution section 8(3) 

virtually rules out the immediate application of a Bill of Rights 

standard to any private person’s conduct, directing rather that all 

such applications be made, instead, to the legal rules and doctrines 

on which litigating parties rely to found their claims and defenses. 18 

Consider, then, another sort of choice that might confront a court 

in our invasion-of-privacy case, where the court is already committed 

to test only law, not conduct, under the Bill of Rights. A court 

disposed to rule in the plaintiff’s favor, despite the defendant’s 

proven defense of absence of intention to publish without 

authorisation, might proceed by first holding the extant, delictual law 

of privacy invalid by reason of inconsistency with a specific right 

guaranteed by some rights-granting clause in the Bill of Rights — here 

it would be section 14 (‘Everyone has the right to privacy ...’) — and 

then, by way of remedy for the unconstitutionality, directing a 

curative revision or ‘development’ of the common law. As explained 

further below, the court then would be acting under the aegis of 

Constitution section 172(1), or of Constitution section 8, or (perhaps 

it would be best to say) of both of those sections. Alternatively, the 

court might take its cue from Constitution section 39(2). It might then 

by-pass the question of the extant law’s possible invalidity by reason 

of inconsistency with a specific, rights-granting clause, and simply 

find the extant delictual doctrine unacceptably out of sorts with the 

overall ‘spirit, purport, and objects’ of the Bill of Rights. We might 

classify the first sort of judicial action as ‘direct’ application of the 

Bill of Rights, and the second sort as ‘indirect’. 

We then would have severed any special connection of the direct/ 

indirect distinction to cases involving horizontal application of the Bill 

of Rights, because the distinction as now defined is equally applicable 

18 

But the conduct/law distinction does survive in transmogrified form to postsection 

8 debate. See, eg, H Cheadle ‘Application’ in H Cheadle, D Davis & N 

Haysom (eds) South African constitutional law: The Bill of Rights (2002) 19: ‘ ... 

[B]ecause generally stated rights are not appropriate vehicles for the imposition 

of standards of conduct, section 8 seeks to avoid application of the right to 

conduct of private persons by requiring either legislation or a common law rule 

‘to give effect to the right.’ This constitutional motif of an intermediate law 

between the constitutional right and the conduct of the state or of the private 

actor repeats itself throughout the Bill of Rights.’ Woolman takes Cheadle to be 

saying, contentiously, that ‘the Bill of Rights rarely, if ever, applies directly to 

“conduct”. It applies to law.’ Woolman ‘Application’ (n 17 above) 155.


to cases that are incontestably vertical in nature. 19 Take, for 

example, a case, which surely must be classed as vertical, in which 

the state prosecutes a person for alleged commission of the common 

law crime of sodomy. When that case came before the Constitutional 

Court, the Court employed the form of application of the Bill of Rights 

that we are now calling ‘direct’, so as to invalidate the crime of 

sodomy and erase it from the common law of South Africa. 20 Had the 

Court seen fit to do so, it could have achieve the same result 

‘indirectly’, by holding that the common law of crime must be 

‘developed’ so as to omit the crime of sodomy, in order to stay in tune 

with the overall spirit etc of the Bill of Rights. 

To be as precise as possible about where we have come, we now 

have before us the following three, live possibilities: 

(a) ‘direct’ = declaration of invalidity followed by remedial development 

of the common law; ‘indirect’ = development of the common law 

to keep it attuned to the spirit etc. of the Bill of Rights, with no 

attendant declaration of invalidity. 

(b) ‘direct’ = a finding of the extant common law’s inconsistency with 

norms or standards contained in one or another specific rights-granting 

clause; ‘indirect’ = no such finding, but only a finding of disharmony 

with the general spirit etc. of the Bill of Rights taken whole. 

(c) ‘direct’ = judicial action in terms of Constitution sections 172(1), 8; 

‘indirect’ = judicial action in terms of section 39(2). 

When Stu Woolman speaks of ‘direct’ versus ‘indirect’ application of 

the Bill of Rights, he tends to run these oppositions together. No doubt 

opposition ‘b’ must stand as his core, his official definition of the 

distinction. 21 But Woolman sustains no effort to pry or to hold the 

three oppositions apart. Rather, he sometimes treats the three as a 

cluster, in which each opposition in some way entails, insinuates, 

attracts, or bleeds into both of the other two. Such a treatment flows 

naturally from Woolman’s strongly held, carefully argued normative 

view that South African jurists ought to follow a discipline of holding 

the oppositions in fixed alignment. Woolman’s construction is 

captured in the table below. 22 

19 

See Woolman ‘Application’ (n 17 above) 5 n 1 (observing that the ‘application’ 

debate must now be framed in terms of ‘direct’/’indirect’, not ‘vertical’/ 

’horizontal’). 

20 

See National Gay & Lesbian Coalition v Minister of Justice 1996 3 SA 850 (CC); 

1996 5 BCLR 658 (CC) paras 73, 90 (Sodomy Case). 

21 ‘Direct challenges describe instances in which the prescriptive content of at least 

one specific substantive provision of the Bill of Rights applies to the law or to the 

conduct at issue. Indirect challenges describe instances in which the prescriptive 

content of no specific provision of the Bill of Rights applies to the law or to 

conduct at issue.’ Woolman ‘Application’ (n 17 above) 5 n 1. 

22 See generally Woolman ‘Application’ (n 17 above).

form of judicial power 

(a) 

substantive test of 

inconsistency (b) 

controlling section of 

Constitution (c) 


‘direct’ (1) ‘indirect’ (2) 

judicial-review power 

to declare invalidity 

and prescribe remedy 

involving revision of 

common law [1a] 

is common law 

inconsistent with 

requirements of a 

specific right in the 

Bill of Rights? [1b] 

172(1) and/or 8 [1c] 39(2) [2c] 

inherent power to 

develop common law 

in normal course of 

adjudication [2a] 

is common law 

inconsistent with 

general objects of Bill 

of Rights, even though 

not inconsistent with 

requirements of any 

specific rights? [2b] 

According to what Woolman urges as the best, overall, combined 

construction of the Constitution’s relevant texts and purposes, 

whenever a court chooses to proceed — or finds itself proceeding — in 

terms of any of 1a, 1b, or 1c, it ought to require itself concomitantly 

to proceed, on that occasion, in terms of other two members of triplet 

1; and conversely, of course, for any court that finds itself proceeding 

in terms of any of the members of triplet 2. 

Whether Woolman’s normative/textual arguments in favor of his 

preferred construction, and of a correspondingly disciplined juristic 

practice, entirely succeed is not our question here. My question will 

rather be whether the prism of Woolman’s preferred construction 

works to the overall advantage of clarity of analysis in approaching 

the work of the Constitutional Court. If the Constitutional Court 

should happen to be working with a different (and perhaps itself 

defensible) construction, then it may not. To be more specific: 

Suppose that (as I shall suggest the evidence shows) the Constitutional 

Court is quite comfortable combining 2a and 2c with 1b. Woolman 

finds that combination wrongheaded, or, at any rate suboptimal from 

the standpoint of assigning clear, distinct, and jurisprudentially 

cogent meanings to the several, relevant constitutional clauses. Call 

that the Court’s ‘construction error’. The construction error — 

2a+1b+2c — cannot be counted as contributory towards a flight from 

meaningful engagement with the substance of the rights-naming 

provisions of Chapter Two, because one of its components — 1b — is 

entirely congenial to what Woolman means by engagement. The 

construction error and the ‘flight error’ (as we may call the latter) are 

two different things. Woolman sometimes seems to me to be 

detecting the flight error where only the construction error is clearly 

in evidence.


3 NM v SMITH 

3.1 The case 

Charlene Smith wrote a book about the public career of Patricia De 

Lille, an MP and a public figure. The book recounted De Lille’s 

interventions on behalf of persons living with HIV/AIDS. These 

included an episode involving controversial drug trials at the 

University of Pretoria, in which the three women who became the 

plaintiffs in the case of NM — ‘NM’, ‘SM’, and ‘LH’ — were among the 

volunteer subjects. The three plaintiffs figure significantly in the 

book’s account of the controversy, as persons living with HIV. They 

are in no way otherwise persons of the slightest public note nor would 

their names be recognisable by members of the public outside their 

own special circles of acquaintance. 

Smith learned the plaintiffs’ names from a copy of an in-house 

report of a professor’s investigation of the medical-study controversy, 

conducted on behalf of the University and intended for restricted 

circulation. (No one complains of the means by which Smith obtained 

her copy.) The report refers to ‘annexures’ containing forms signed 

by the three plaintiffs, giving consent to the inclusion of their names 

in the report. The copy provided to Smith did not include the 

annexures. Smith’s book, when published, identified the plaintiffs by 

name as persons living with HIV, as the report had done. Contrary to 

Smith’s expectation, it emerged that the consents given by the 

plaintiffs to the report’s author, Professor Strauss, were not general 

releases to disclose their medical information, with names attached, 

to the public at large, but only authorised a limited disclosure for 

purposes related to the University’s investigation. 

The plaintiffs brought an action against Smith for damages, 

claiming that her book’s identification of them by name, without their 

consent, as persons living with HIV amounted to a violation of their 

common law rights to privacy, specifically in terms of the actio 

iniuriarum. 23 Under the common law as it then stood (and still stands, 

in the wake of the Constitutional Court’s decision in NM), the 

plaintiffs could succeed only if Smith were found to have acted 

toward them in a manner that was not only objectively wrongful but 

also intentionally so. 24 Smith contended that any wrong she might 

have committed toward the plaintiffs, by way of publishing their 

23 

In order to maintain clarity, I focus my examination of NM on the claims for 

damages against defendant Smith; I do not discuss the claims for other forms of 

relief, or the claims for damages against defendants Patricia De Lille and New 

Africa Books Ltd. 

24 See NM (n 5 above) para 55 (Madala J); para 151 (O’Regan J).


private information without their consent, was unintentional. The 

crux of her defense was that she acted in the honest (if mistaken) 

belief that the plaintiffs’ consents, of which she read in the Strauss 

Report, covered unrestricted circulation of the information regarding 

them contained in the report. Apparently persuaded by Smith’s 

testimony to this effect, the High Court absolved Smith of liability for 

tortious invasion of privacy. 25 The Supreme Court of Appeal refused 

review, giving no reasons, and the plaintiffs took their case to the 

Constitutional Court. 26 

Three members of the Court — Langa CJ, O’Regan J, and Sachs J 

— concluded that the case against Smith disclosed a need to develop 

the common law so to remove an incompatibility with the objects of 

the Bill of Rights in the matter of protecting privacy. These members 

would have developed the common law so as to impose liability, on 

‘media defendants’ only, for negligent (or unreasonable), wrongful 

disclosure of private information. 27 The Court’s majority took a 

different tack. They joined an opinion by Madala J, holding that 

Smith’s unauthorised disclosure of the plaintiffs’ private information 

should be found intentional in terms of the extant common law 

standard of liability. 28 They accordingly ordered a judgment for 

damages against Smith 29 while declining to address any question of 

developing the common law. 30 Of the three justices who thought the 

common law must be developed, Langa CJ and Sachs J concurred in 

the order awarding damages to the plaintiffs, on the ground that the 

record showed negligence on the part of Smith. 31 O’Regan J would 

have absolved Smith of negligence and, hence, of liability. 32 

3.2 A Critique 

Woolman’s objections to the Constitutional Court’s performance in 

NM are mainly directed to the decision and opinion of the majority. 

While the objections refer to specific, alleged deficiencies and 

missteps in this particular case, Woolman presents them as exemplary 

of broader, worrisome trends in the work of the Constitutional Court. 

These are most compendiously summed up as a ‘lack of analytical 

rigour’ suggestive of ‘a penchant for outcome-based decisionmaking.’ 

33 In Woolman’s view, the decision in NM ‘appears to rest 

25 

26 

27 

See NM (n 5 above) para 46 (Madala J); para 125 (O’Regan J). 

See NM (n 5 above) para 19. 

See NM (n 5 above) paras 92-94 (Langa CJ); paras 170-19 (O’Regan J); paras 203- 

28 

29 

30 

31 

32 

33 

04 (Sachs J). 

See NM (n 5 above) paras 58-65. 



See NM (n 5 above) para 92 (Langa CJ); para 207 (Sachs J). 


Woolman ‘Amazing’ (n 2 above) 762.


upon a deeply-felt offence to the majority’s moral sensibility about 

how vulnerable persons in our society ought to be treated.’ 34 

Woolman’s bill of particulars includes the following: 

(1) The majority ended up acting ‘as a trier of fact in a run-of-the-mill 

actio iniuriarum matter’ 35 — a sort of task, Woolman evidently means to 

say, that the Constitutional Court was not created to do. 

(2) The majority’s crucial factual determination regarding Smith’s 

intention is ‘contrary to the evidentiary record’, which ‘cannot support 

a factual finding that the respondents had acted intentionally to harm 

the privacy and the dignity interests of the applicants.’ 36 

(3) The majority failed to provide an adequate explanation for why, 

once they had determined that the case would be disposed of by 

applying an unmodified, long-established common law standard of 

liability, they did not dismiss the appeal as being no proper concern of 

the Constitutional Court. 37 

(4) The majority diminished the Constitution by declining to take up for 

consideration a possible ‘challenge to the actio inuriarum grounded in a 

specific substantive provision of the Bill of Rights’ or the possible 

‘creation of a self-standing constitutional action grounded in the right to 

privacy or the right to dignity.’ 38 This objection echoes prior pleas from 

Woolman, for a shift of the Court’s emphasis away from ‘indirect’ 

application of the Bill of Rights pursuant to section 39(2) to ‘direct’ 

application pursuant to section 8. 39 The Court’s penchant for section 

39(2) is deplorable, Woolman maintains, because it plays into the false 

seductions of a so-called minimalist approach to constitutional 

adjudication that is unsuited to the present, early state of development 

of South African constitutional jurisprudence and culture. 40 It invites the 

Court to issue highly contextualised decisions that may feel right to the 

justices and many of those looking on, thus avoiding or postponing 

intracurial division and public controversy, but that remain unexplained 

on the level of principled, clause-by-clause exposition of the meanings 

of the rights-guarantees in Chapter 2 and the values underlying them. 41 

Yet the Court’s presumed, special capacity for that more challenging — 

in part because potentially divisive — task of principled explication is its 

raison d’être par excellence. Performance of that task, Woolman 

perceives, is a service to the country that only the Constitutional Court 




37 

See Woolman ‘Amazing’ (n 2 above) 782 & n 40; Constitution of the Republic of 

South Africa, 1996 sec 167(3)(b) (The Constitutional Court ‘may decide only 

constitutional matters, and issues connected with decisions on constitutional 

matters’). 


39 See generally Woolman ‘Application’ (n 17 above). 

40 Woolman ‘Amazing’ (n 2 above) 764-65 n 4, 784-87. 

41 Woolman ‘Amazing’ (n 2 above) 784-86.


can provide, and that it must perform for the sake of the rule of law in 

South Africa. 42 

(5) In sum, the majority’s narrow, non-venturesome treatment of the 

case shirked a sort of service to the country that the Court was created 

to provide, to wit: authoritative delineation and explication of the 

principles animating the several specific substantive rights in Chapter 2, 

conveyed with definition sufficient to ‘determine the actual validity of 

the rule being challenged in the instant matter and of similar rules 

challenged in subsequent matters’; also to enable citizens and 

government officials to ‘ensure that their behaviour conforms to our 

Constitution’, and also to equip lower courts and lawyers to ‘identify the 

law and thereby settle, litigate and adjudicate, with some confidence, 

Bill of Rights cases.’ 43 NM, Woolman concludes, ‘shows a court in full 

flight from any meaningful engagement with Chapter 2 of the 

Constitution.’ 44 

Woolman’s observations are eminently worth pondering, and by no 

means without warrant in the majority opinion in NM, as written. I 

mean to show, however, how the complaints turn on matters of 

interpretation that can be resolved in the Court’s favour — with, no 

doubt, varying degrees of effort from the reader. The opinion in NM 

is indeed thin and chary of explanation, in places where it could and 

should be thicker and more forthcoming. Contrary to Woolman’s view, 

however, I think the NM majority got to the right answer, by a proper 

path (although the path preferred by Langa CJ seems to me superior), 

and for essentially (and detectably) right reasons. I say so with utmost 

respect to the opinion in the case that most ambitiously addresses the 

task of substantive elucidation of constitutional rights that Woolman 

misses from the Court’s work, the dissenting opinion of O’Regan J. 

Admirable as that opinion is in many ways, in the end it decides the 

case wrongly, in my view. Moreover, it perhaps does so for causes not 

unrelated to O’Regan J’s evident commitment to (and talent for) the 

work of principled exposition of the contents of fundamental rights. 

3.3 A pleading choice 

In order to succeed with their case against Smith, the NM plaintiffs 

had to show that Smith’s conduct not only had harmed them in fact, 

but had done so in a way that was liability-engendering, according to 

the terms of some law meant to protect persons in the plaintiffs’ 

position against the sort of harm at issue. Of course, it would be up to 

the plaintiffs to say what law they had in mind. Their answer was: the 

common law of delict, and specifically the actio iniuriarum. 45 

42 See Woolman ‘Amazing’ (n 2 above) 786-87. 



45 See NM (n 5 above) paras 22, 27-29.


A different response was also imaginably available. The plaintiffs 

might have pleaded as their cause of action Constitution section 14 

(‘Everyone has the right to privacy’), as rendered applicable to 

Smith’s conduct by Constitution section 8(1) (‘The Bill of Rights 

applies to all law, and binds the legislature, the executive, the 

judiciary and all organs of state’) and (8(2) (‘A provision of the Bill of 

Rights binds a natural ... person if, and to the extent that, it is 

applicable ...’). 46 Rather than (or in addition) to claiming damages for 

a common-law tort, the plaintiffs might have claimed on the basis of 

a constitutional tort. They did not do so, and it will repay us to ask 

why they did not. The common-law path selected by the plaintiffs was 

not, after all, free and clear of evident obstacles. The plaintiffs (or 

their lawyers) would have known the list of conventionally established 

requirements for success in a claim for invasion of privacy based on 

the actio iniuriarum. 47 They would have known that among these was 

a showing of the defendant’s intention (‘animus iniurandi’) to 

commit, as against the plaintiff, the wrong consisting of a publication 

of private facts concerning the plaintiff, without the plaintiff’s leave 

or against her will — where the category of ‘private’ facts includes any 

and all facts, not yet publicly available, whose disclosure to the public 

would be expected to ‘cause mental distress and injury to anyone 

possessed of ordinary feelings in the same circumstances’. 48 

Consider, then, a defendant who publishes a plaintiff’s entire, 

anguished medical history (naming her), in the honest but false belief 

that the plaintiff has given leave for publication, or that the 

information has been previously (lawfully) published and hence is no 

longer private. Such a defendant will indeed have committed the 

wrong in question, ‘objectively’ described as above; her action will, 

in that sense, have been wrongful under the law. Missing from the 

case, however, will be the usual basis for an inference of the 

defendant’s intention to mistreat (‘wrong’) the plaintiff in that way 

— to wit, the defendant’s actual knowledge of all of the facts that, in 

combination, would render her action wrongful. 

To put the matter simply: Under the common law of delict as it 

stood when the plaintiffs filed their case against Smith, negligent but 

non-intentional commission of the objective wrong of unauthorised 

publication of private facts was not deemed legally actionable, ‘as a 

46 See NM (n 5 above) para 132 (O’Regan J) (confirming the applicability of 

Constitution sec 14 to private actors). 

47 

See NM (n 5 above) para 55 (Madala J); para 151 (O’Regan J). 

48 NM (n 5 above) para 34 (Madala J), citing National Media Ltd v Jooste 1996 3 SA 

262 (A); 1996 2 All SA 510 (A). O’Regan J preferred to leave open whether the 

Constitution might possibly require some expansion of this common-law test for 

the protected (‘private’) status of the facts disclosed, but she was clear that ‘the 

publication of otherwise confidential information about a life-threatening illness 

is likely to cause distress to the person concerned’, and thus falls under the 

common-law test for liability. NM (n 5 above) para 137.


rule’. 49 Plaintiffs might sometimes have succeeded on the basis of a 

rebuttable presumption of intention, once objectively wrongful 

publication was shown, 50 or by establishing a modified form of 

intention known as dolus eventualis (or reckless disregard) on the 

defendant’s part, 51 but a defendant who could satisfy a judicial factfinder 

that her mistake was bona fide, although careless, would 

escape liability. 

In NM, the available evidence was consistent (at least) with 

defendant Smith’s claim of an honest belief that the plaintiffs had 

consented to unrestricted public naming of them as the persons (living 

with HIV) who had played certain, described roles in the drug-trial 

controversy. Whether that evidence could further support a claim of 

absence of negligence on Smith’s part is much more doubtful, as we 

shall see. Thus there was, from the beginning, reason to calculate 

that the plaintiffs’ prospect for success might depend on whether the 

body of law (the ‘cause of action’) they invoked to support their claim 

would impose liability on a merely negligent defendant. 52 At the 

moment when the plaintiffs sued, the common law of delict evidently 

was against them on that potentially crucial point. By contrast, for 

aught anyone could then tell, the law of the putative constitutional 

tort based on section 14 of the Bill of Rights might have been with 

them. Even so, the plaintiffs chose to base their claim on the actio 

iniuriarum, and not on section 14 of the Bill of Rights. That choice 

requires explanation. 

3.4 Two litigation agendas compared 

In the view of Madala J, the plaintiffs were forced to their pleading 

choice by the Constitutional Court’s 1997 decision in Fose. 53 Fose 

certainly does leave a doubt about when, if ever, ‘constitutional 

damages’ might be awarded in respect of a violation of a provision in 

the Bill of Rights. It seems to me, however, for reasons I am about to 

present, that a complete explanation for the NM plaintiffs’ pleading 

choice would have to include the further point that a claim expressly 

and immediately invoking the prescriptive content of section 14 of the 

Bill of Rights, as a test of the legally actionable character of Smith’s 

conduct, would have brought the plaintiffs to very much the same 

49 

50 

NM (n 5 above) para 55 (Madala J). 

See NM (n 5 above) para 65 (Madala J); para 93 (Langa CJ); paras 125, 155, 169 

51 

52 

(O’Regan J). 

See NM (n 5 above) para 155 (O’Regan J). 

The summons issued by the plaintiffs makes clear they understood that this was 

53 

so. See NM (n 5 above) para 124 (O’Regan J) (describing the summons). 

Fose v Minister of Public Safety and Security 1997 3 SA 786 (CC); 1997 7 BCLR 851 

(CC); see NM (n 5 above) para 27.


pass, regarding the question of that conduct’s intentionality, as 

resulted from their common-law claim. 

This consequence follows inexorably from the mandate of 

Constitution section 8(3), as construed by the Constitutional Court. 54 

That section provides: 

When applying a provision of the Bill of Rights to a natural or juristic 

person in terms of subsection (2), a court — 

(a) in order to give effect to a right in the Bill, must apply, or if 

necessary develop, the common law to the extent that legislation does 

not give effect to that right; and 

(b) may develop rules of the common law to limit the right, provided 

that the limitation is in accordance with section 36(1). 

By this seemingly clear command of constitutional law, any court 

disposed to impose civil liability, in terms of standards of conduct 

drawn from one or another of the rights-guarantees in the Bill of 

Rights, is required to proceed in one of three ways, and not otherwise. 

Either (1) the court must decide the case in terms of legislation that 

gives due effect to the constitutional right in question; or (2) in the 

absence of such legislation, the court must decide the case in terms 

of applicable common law that gives due effect to the constitutional 

right in question; or (3) if the court decides that the applicable 

common law does not currently give due effect to the right, the court 

must decide the case in terms of the common law as developed by the 

court so as to cure the deficiency. 55 

Accordingly, in NM, the result of a claim for a ‘direct’ application 

of section 14 to Smith’s conduct would have been the Court’s having 

to decide two principal issues in tandem: whether giving due effect to 

54 

That is to say, it follows inexorably from the Constitutional Court’s authoritative, 

‘black letter’ construction of the three subsecs of sec 8, as announced in Khumalo 

v Holomisa 2002 3 SA 401 (CC); 2002 8 BCLR 771 (CC) (‘Khumalo’) paras 29-32 and 

precisely summarised by Woolman ‘Application’ (n 17 above) 6, 45. It would not 

follow from Woolman’s ‘preferred reading’ of sec 8, summarised at Woolman 

‘Application’ (n 17 above) 11, 45-46. But of course the NM plaintiffs’ lawyers 

would have had to frame their case in the shadow of the Constitutional Court’s 

reading, not that of any commentator. 

55 This directive carries out what logic would seem to require in its absence. We can 

use the NM case to illustrate. Suppose a court finds that Smith’s conduct has 

‘limited’ (ie, violated) NM’s prima facie right to privacy under Constitution sec 

14. Smith will seek refuge under Constitution sec 36(1). She will claim that she 

has acted in terms of law of general application, namely, the extant common law 

of delict, which — as she will claim — insulates her from liability as long as her 

conduct has not been knowingly and intentionally wrongful. The court, then, will 

have to decide whether that rule of the common law, bestowing a legal privilege 

to commit negligent (although objectively wrongful) publication of private 

information, is reasonable and justifiable in terms of Constitution sec 36(1). If the 

answer is ‘no,’ the court will have no option but to develop the common law so as 

to cure the defect.


section 14 requires that Smith be held liable in this case, given the 

evidence and what it shows; and whether (and, if so, how) the 

common law as it currently stands must be ‘developed’ to encompass 

at least some instances of non-intentionally but objectively wrongful 

invasion of privacy. The first question for decision surely would have 

been (a) whether Smith’s conduct was both objectively wrongful and 

animo iniurandi according to the pre-existing common-law standard. 

If yes, the plaintiffs would prevail. 56 If no, the next question would 

have been (b) whether the plaintiffs’ case against Smith discloses a 

need to expand common-law liability to catch at least some instances 

of non-intentionally but objectively wrongful publication of private 

facts, in order to vindicate fully the right of privacy guaranteed by 

Constitution section 14. If no, the defendant would prevail. If yes, the 

next question would have been (c) precisely what modification of the 

common law is required in order to give due effect to the 

constitutional right of privacy, in the light of possibly competing 

constitutional principles and values such as freedom of expression — 

a question that would have been decided by judges who were at the 

same time, inevitably, thinking about (d) whether one or another 

newly expanded zone of liability that the Court might approve would 

catch Smith’s conduct. 

Compare that agenda with the one that plaintiffs basing their 

claim on the actio iniuriarum must anticipate. Taking the common 

law as it stands, they can succeed by establishing that (a) Smith’s 

conduct was both objectively wrongful and animo iniurandi in terms 

of the extant common law standard of liability. 57 Alternatively, given 

the court’s inherent power to develop the common law, 58 combined 

with the court’s duty to develop the common law so as to promote the 

spirit, purport, and objects of the Bill of Rights, imposed by 

Constitution section 39(2), 59 the plaintiffs can succeed by 

establishing, in the light of their case against Smith, all of the 

following: (b) that the common law’s current, blanket prescription of 

non-liability in all instances of non-intentionally but objectively 

wrongful publication of private facts is not acceptable to the Bill of 

Rights; plus (c) and (d) that when the common law is properly 

56 

For completeness, we should note the possibility of the defendant persuading the 

court that the extant common law standard must undergo development in her 

favor, either in terms of Constitution sec 8(3) (in order to give due horizontal 

effect to the right of freedom of expression guaranteed by sec 15), or in terms of 

sec 39(2) (in order to keep the common law in tune with the objects of the bill of 

rights in the matter of freedom of expression). 

57 

And also, if necessary, successfully defending that standard against complaint 

that it unduly constricts constitutionally valued freedom of expression. 

58 ‘The Constitutional Court, Supreme Court of Appeal and High Courts have the 

inherent power to ... develop the common law, taking into account the interests 

of justice.’ Constitution sec 173. 

59 Carmichele v Minister of Safety and Security 2001 4 SA 938 paras 33-39, makes 

clear that a court is required to consider whether such development of the 

common law is required in any case where a party urges the point.


modified in order to give due effect to the constitutional right of 

privacy, in the light of possibly competing constitutional principles 

and values such as freedom of expression, Smith’s conduct in this case 

will render her liable to the plaintiffs. 

The two agendas look virtually identical. 60 Therefore (the bottom 

line): Since Fose suggests the possibility of special obstacles facing 

claims for constitutional damages, the plaintiffs’ course of least 

resistance was to base their demand for damages on a claim of a 

common law violation of privacy under the actio iniurarum, with 

section 39(2) hovering overhead. To my mind, Madala J’s opinion 

adequately (if somewhat elliptically) evinces a clear appreciation of 

the entire line of thought I have just been unpacking. 61 At the very 

least, Madala J’s opinion is in all respects consistent with this line of 

thought. If, as Woolman truly says, the opinion never poses the 

possibility of ‘the [judicial] creation of a self-standing constitutional 

action grounded in the right of privacy’, 62 that is because the 

applicants, it would seem for sufficient reasons that the majority may 

fairly be taken to have understood, chose not to start up that 

particular formulation of their claim. 

3.5 Jurisdiction 

Although the plaintiffs thus chose to rely on a common law and not a 

constitutional cause of action, their appeal to the Constitutional 

Court was quite patently and undeniably one that fell within that 

Court’s subject-matter jurisdiction — limited though that jurisdiction 

surely is to ‘constitutional matters’ and issues ‘connected to’ 

constitutional matters. 63 The overarching legal problem presented by 

the case is whether the common law of delict, in regard to invasion of 

privacy, requires some modification in order to keep it in line with the 

Bill of Rights taken as a whole. Involved, as one immediately sees, is 

not only a question about whether — and if so, how far — the common 

law must be altered so as to attach liability to at least some instances 

of non-intentionally wrongful disclosure of private facts, in order to 

keep faith with the Bill of Rights in regard to privacy and dignity. 

Equally involved is a question about whether — and if so, how far — 

the common law may permissibly impose such liability, having in mind 

the commitment of the Bill of Rights in regard to freedom of 

expression. These are obviously issues of full-scale constitutional 

import. 

60 Or very closely so. Arguably, there is a difference, which I will address at the 

close of my discussion of the NM case (see below part 3.10) and in my discussion, 

to follow, of Masiya. 

61 See NM (n 5 above) paras 24, 27, 55-57. 


63 Constitution sec 167(3)(b).


Woolman sees NM as a case in which the Constitutional Court 

majority found no occasion to consider these constitutionally fraught 

issues of common-law development, because it decided the case, 

instead, by overturning the High Court’s factual conclusion that 

Smith’s objective violation of the plaintiffs’ privacy was nonintentional 

and hence non-actionable under the extant common law 

standard. That might seem to invite a question about why the 

majority did not order the case dismissed for lack of jurisdiction, once 

it perceived that the case could and would be decided on standard, 

common-law grounds (‘run-of-the-mill’ grounds, Woolman calls 

them). 64 And indeed, the majority must be faulted for its own 

complicity in creating an impression of the irrelevance of the Bill of 

Rights to its disposition of the case. 65 

That impression, however, is mistaken. We shall see that it is at 

least an open question whether the NM majority did not, in effect, 

‘develop’ the common law in response to constitutional pressure. (In 

that regard, the majority may be faulted for misapprehending the 

implications and ramifications of its own decision in the case.) But 

suppose we say that it did not. The majority could still be right in 

affirming the Constitutional Court’s jurisdiction in this case, even 

after reaching its conclusion against any constitutionally imposed 

need for adjustment of the common law. Jurisdiction would hold, as 

long as the case presented a fairly debatable question of such a need 

— including of a need to weaken the actio iniuriandum in deference 

to freedom of expression; for surely such a question is itself a 

constitutional matter. (Ask yourself: Does the Constitutional Court 

forfeit jurisdiction by deciding, after argument, to uphold as 

constitutional a plausibly challenged statute?) 66 

The very presence of the dissents in this case would thus seem to 

settle in the affirmative the question of jurisdiction. At all events, the 

very question of how to construe and apply the extant common law to 

64 See Woolman ‘Amazing’ (n 2 above) 782 n 40: ‘[T]he court often accepts cases 

prior to the meaningful application of its collective “mind” to the question of 

whether or not the case genuinely raises a constitutional issue.’ For an example 

of a jurisdictional dismissal, upon its becoming apparent to the Court that no 

constitutionally-pressured development of the common law is in the offing, see 

Phoebus Apollo Aviation CC v Minister of Safety and Security [2002] ZACC 26; 

2003 1 BCLR 14 paras 10-12. 

65 See NM (n 5 above) para 69 (excusing the Court from considering the possible 

impact on freedom of expression from its decision to hold Smith liable, on the 

ground that ‘this judgment is not extending the common law definition of 

intention to include negligence in relation to the publication of private medical 

facts’). 

66 If, as I assert, the Constitutional Court is competent to hear any non-frivolous 

claim of a need (in terms of sec 39(2)) to develop a common-law rule on which a 

party to the case relies, the effect may be to extend the Court’s jurisdiction too 

broadly for your comfort. If so, the fault lies not in the Court but in the drafters 

of the Constitution. See FI Michelman ‘The supremacy of the Constitution and the 

rule of law’ in S Woolman et al (eds) Constitutional law of South Africa (2nd 

Edition, OS, 2005) ch 11 7-9 (Michelman ‘Supremacy’).


facts of the sort presented by the case of NM, being determinative of 

whether the common law must undergo development in terms of 

section 8 or 39, is itself drenched in constitutional-legal 

considerations. All of the common-law-framed issues in the case — 

whether the facts concerning the HIV status of the plaintiffs were of 

‘private’ quality and whether the private character of those facts had 

been fatally compromised by previous disclosure for allegedly limited 

purposes; 67 whether Smith’s conduct was intentionally wrongful in 

the relevant, legal sense; 68 whether that conduct was in any degree 

negligent, and, if so, in what specific respects 69 — stood to be decided 

in the shadow of the Bill of Rights and of the looming possibility that 

the common law might, depending on how they were decided, have 

to be developed under the mandate of section 39(2). 

Madala J wrote for the majority as follows: 

The dispute before us is clearly worthy of constitutional adjudication 

and it is in the interests of justice that the matter be heard by this Court 

since it involves a nuanced and sensitive approach to balancing the 

interests of the media, in advocating freedom of expression, privacy and 

dignity of the applicants irrespective of whether it is based on the 

constitutional law or the common law. This Court is in any event 

mandated to develop and interpret the common law if necessary. 70 

The last sentence is crucial. In the light of what precedes it, it states 

why the case does ‘genuinely raise a constitutional issue’ 71 and it 

shows that the NM majority in this case necessarily undertook more 

(and did more) than merely ‘act as a trier of fact in a run-of-the-mill 

actio iniuriarum matter.’ 72 

I shall make these matters more concrete in discussion to follow 

below. For now, the point simply is that no one conversant with the 

established commitment of South African constitutional 

jurisprudence to a pervasive influence for the Bill of Rights 

throughout the length and breadth of the country’s law 73 could doubt 

that the knot of common-law controversies in NM deserved — if it did 

not demand — the attentions of at least one of South Africa’s two 

pinnacle courts, exercising a responsibility imposed on both of them 

(given how the plaintiffs chose to frame their case) by section 39(2). 

Had an appeal to the Supreme Court of Appeal been an open 

possibility in this case, and the Constitutional Court nevertheless 

67 

See NM (n 5 above) paras 34-45 (Madala J); paras 135-137; paras 142-43 (O’Regan 

68 

69 

70 

71 

72 

73 

J). 

See NM (n 5 above) paras 58-65 (Madala J); paras 155-69 (O’Regan J). 

See NM (n 5 above) paras 100-11 (Langa CJ); paras 183-89 (O’Regan J). 

NM (n 5 above) para 31. 

Woolman ‘Amazing’ (n 2 above) 782 n 40. 

Woolman ‘Amazing’ (n 2 above) 783. 

See Michelman ‘Supremacy’ (n 66 above) 37-41.


accepted the case for hearing, I would have expected from the Court 

some careful explanation for its choice thus to allow circumvention of 

the SCA. That, however, was not the state of affairs confronted by the 

Constitutional Court when NM’s application arrived at its doorstep. 

The SCA had already removed itself from the case by its summary 

refusal of leave to appeal. 74 

3.6 A puzzle 

The Constitutional Court divided sharply over assessment of the 

evidence concerning the intentionality of Smith’s conduct. Writing for 

a majority of seven justices, Madala J concluded that Smith acted 

intentionally, within the terms of the doctrine then and here 

surrounding the actio iniuriarum. 75 In the views of Langa CJ and Sachs 

J, Smith’s conduct, although not intentionally wrongful, was 

negligent or unreasonable (and thus liability-producing under the 

common law as they would have had it developed). 76 In the view of 

O’Regan J, Smith’s conduct was entirely free of fault. 77 There are two 

divisions worth exploring here: first, the division between Madala J’s 

majority and the three justices — Langa CJ , Sachs J, and O’Regan J 

— who rejected the majority’s finding of intentionally wrongful 

conduct on the part of Smith; and, second, the division among those 

three over the question of Smith’s negligence. 

Smith had directed her trial testimony to explaining how she 

acted in the honest belief either that the plaintiffs’ names had 

already been disclosed to the general public or else that the plaintiffs 

had fully and finally waived objection to any such disclosure. The High 

Court had absolved Smith of any intention to act contrary to the 

plaintiffs’ wishes in this regard. As O’Regan J’s extensive and careful 

summary amply demonstrates, such an inference is, at the very least, 

reasonably supportable from essentially unchallenged components of 

Smith’s testimony. 78 But then a nagging question arises: How are we 

to explain the majority’s apparent deviation from the established 

practice of deference by appellate tribunals to trial court findings on 

issues of fact? That looks to me like Sherlock Holmes’s dog that did 

not bark: a puzzle we should probe, if we — I mean we as academics 

— mean to leave no stone unturned in getting to the innards of the 

problem presented to the Constitutional Court by the NM case. The 

question all the more strikingly demands our attention because the 

74 


75 See NM (n 5 above) paras 58-65. 

76 See NM (n 5 above) paras 92-93 (Langa CJ); paras 205-07 (Sachs J). 

77 

See NM (n 5 above) paras 155-69 (intention); paras 183-89 (negligence). 

78 See NM (n 5 above) paras 157, 159-64.


majority itself declined to address it, despite its having been pressed 

by O’Regan J in dissent. 79 

Academics can speculate — sometimes profitably — about 

imaginable, realist-style explanations for such puzzling judicial 

silences. (Was the Court just hell-bent on imposing a result that it 

likes? Did some intracurial disagreement get in the way of full 

transparency?) In keeping with my general undertaking for this essay, 

I choose a different path of speculation — that of assuming (or 

pretending, if you like that better) that the NM majority thought the 

legal reasons for its action (here, overturning a lower court’s factual 

determination) so obvious as not to require explanation. The idea is 

to see what we might learn about the legal problem presented to the 

court by the case at hand by giving the court every benefit of the 

doubt, working to make their judgment, as a follower of Ronald 

Dworkin might say, the best that it can be. 

Approaching the matter in that spirit, we may discover that the 

majority’s finding of intentionally wrongful action on Smith’s part, 

disputable though it surely is, is no more flatly ‘contrary’ to the 

evidentiary record than an opposite finding is. 80 We may find that, 

pace all of Langa CJ, O’Regan J, and Stu Woolman, the record (as 

described by the opinion-writers in NM) can indeed ‘support a factual 

finding that [Smith] had acted intentionally to harm the privacy and 

dignitary and dignitary interests of the applicants.’ 81 Judgment on 

this point may depend partly on what you mean by ‘intentionally’, but 

that feature in the case is, as I have already pointed out and shall 

explain further below, a part of what makes this case indubitably one 

of constitutional law for the Constitutional Court to decide. 

3.7 ‘Incremental development’ and Smith’s ‘intention’ 

Therefore, consider Constitution section 39(2). Remember how it 

goes? 

When interpreting any legislation, and when developing the common law 

or customary law, every court, tribunal or forum must promote the 

spirit, purport and objects of the Bill of Rights. 

All right, then: Where there is fair room for disagreement over the 

proper gloss on statutory terminology, and the choice is outcomedeterminative 

for the case sub iudice, courts are instructed by the 

Constitution to prefer the gloss that they find to be the more truly 

79 See NM (n 5 above) paras 125, 169. 


81 As above (invoking the opinions of Langa CJ and O’Regan J in NM).


promotive of the objects of the Bill of Rights. How goes it, then, when 

a like question arises with respect to common law terminology? 

Please don’t stand there painting me, along with Stu Woolman, 

into ‘some ... benighted Hartian, positivist view’ 82 of the common 

law, or telling me that there is so such thing as ‘common law 

terminology’ to make a match with statutory terminology. I join 

Woolman in knowing — and you know, too — that there very obviously 

is such a thing, which is not to say that Woolman and I do not also fully 

understand the common law to be, at bottom, an assemblage of 

always open-ended principles, not of closed-ended rule-sentences. 

(We can leave aside whether the latter is any the less true of statute 

law.) 83 Where Woolman and I both stand our ground is on the point 

that lawyers and the rule of law cannot make do with a common law 

discourse that is not densely populated by rule statements, ‘common 

law terminology’. We both — and so do you, Reader — ‘work within a 

tradition ... of which South Africa is most avowedly a part’, which 

‘recognises rules as a necessary feature of the legal [including the 

common law] landscape.’ 84 Witness: 

For the common law action for invasion of privacy based on the actio 

iniuriarum to succeed, the following must be proved: 

(a) Impairment of the applicants’ privacy; 

(b) Wrongfulness; and 

(c) Intention (animus iniurandi). 85 

The elements of the actio injuriarum are the intentional and wrongful 

infringement of a person’s dignitas, fama, or corpus. 86 

Those are two somewhat variant, but essentially identical, renditions 

of the same, single instance of a genus that lawyers routinely call 

‘rules’ of the common law. As such, they exemplify perfectly what I 

mean by my jurisprudentially obtuse expression, ‘common law 

terminology’, coined by way of analogy to ‘statutory terminology’. 

Now, cases arise in which statutory terminology is found 

unacceptably out of sorts with the Bill of Rights, on any reading of the 

statute falling within the outer limits imposed by a decent respect for 

grammar and lexicon. 87 In those cases, by constitutional command, a 

judicial declaration of the statute’s ‘invalidity’ — unconstitutionality 

82 

Woolman ‘Amazing’ (n 2 above) 789. 

83 See generally, on these matters, R Dworkin Law’s Empire (1986). 


85 

NM (n 5 above) para 55 (Madala J). 

86 NM (n 5 above) para 151 (O’Regan J) (citing both Voet Commentary on the 

87 

Pandects 47 10 1 and R v Umfaan 1908 TS 62 at 66 (Innes CJ)). 

See, eg, National Coalition for Gay & Lesbian Equality v Minister of Home Affairs 

2000 2 SA 1 (CC), 2000 1 BCLR 39 (CC) paras 25-26 (Home Affairs).


— must ensue. 88 How goes it with cases in which common law 

terminology is similarly found to be irreparably out of sorts with the 

Bill of Rights: say, because it would flatly restrict civil liability for 

certain sorts of wrongful harming to instances in which the wronging 

is intentional, when the Bill of Rights would be better satisfied by 

lifting that flat restriction? You might answer: In such cases, the 

common law is to be declared invalid as it stands and then remedially 

‘developed’ so as to cure the defect. You might offer as an example 

of such a cure (passing, for now, the absence of any declaration of 

invalidity) the bit of terminological surgery supported by Langa CJ, 

O’Regan J, and Sachs J in NM, according to which the term ‘intention 

(animus iniurandi)’, where it appears in item (c) in Madala J’s above 

recital of the elements of the civil cause of action for invasion of 

privacy, would be replaced (in some, if not all, categories of cases) by 

the term ‘intention (animus iniurandi) or negligence’. We have there, 

as it were, a remedy of ‘reading in’. (Compare it with a replacement 

of ‘spouse’, in a statute, by ‘spouse or partner, in a permanent 

same-sex life partnership in which the partners have undertaken 

reciprocal duties of support’.) 89 

Have you begun to glimpse the point I am after? Section 39(2) 

contains an express instruction — call it a constitution-conforming 

instruction — to judges engaged in the interpretation of statutory 

terminology, according to which the judges must interpret with a view 

to promoting the spirit, purport, and objects of the Bill of Rights, 

presumably before deciding whether they must declare the statute 

invalid and resort (perhaps) to remedial reading-in. On its face, 

section 39(2) entirely omits any parallel, express instruction to judges 

engaged in the contextualised application (as opposed to any surgical 

alteration by excision or insertion of words) of common law 

terminology. Since such an instruction does appear in the Interim 

Constitution, one might pause for a moment to wonder whether the 

omission was intentional. 90 It is difficult, however, to conceive what 

the drafters might have intended by it. Are there, after all, no cases 

in which a due regard for the objects of the Bill of Rights would, or 

should, crucially affect a judge’s interpretive shading and application 

88 

See Constitution sec 172(1)(a). 

89 Satchwell v President of the Republic of South Africa & Another 2002 6 SA 1 (CC), 

2002 9 BCLR 986 (CC) para 37. 

90 

Compare Constitution of the Republic of South Africa Act 108 of 1994 (Interim 

Constitution) sec 35(2): 

No law which limits any of the rights entrenched in this Chapter, shall be 

constitutionally invalid solely by reason of the fact that the wording used prima 

facie exceeds the limits imposed in this Chapter, provided such a law is 

reasonably capable of a more restricted interpretation which does not exceed 

such limits, in which event such law shall be construed as having a meaning in 

accordance with the said more restricted interpretation. 

(The Afrikaans version has ‘reg’ where the English has ‘law’; and so, by the 

reasoning used by Kentridge J in Du Plessis, it presumably means to cover 

common law along with statute law. See Du Plessis (n 17 above) para 44.)


of common law terminology as it stands to the facts of a particular 

case? Thus, without any declaration of invalidity and concomitant 

surgery on the terminology as received? If so, are not the courts 

plainly — if implicitly — required by section 39(2) to shade and apply 

in favor of the Bill of Rights? And is it not a part of the Constitutional 

Court’s responsibility — and jurisdiction — to speak the last word on 

whether lower courts have done so properly in a given case? 

In a luminous opinion for the Court in K v Minister of Safety and 

Security (‘NK’), 91 O’Regan J answered all of those questions with a 

resounding ‘yes’. In NK, the plaintiff, having been assaulted and raped 

by police officers, sued the Minister for damages on a theory of 

respondeat superior. The plaintiff stood ready to contend for 

‘development’ of the common law, but only if her case would not 

succeed ‘on a proper application’ of the extant common law rule of 

vicarious liability. 92 You can see why she had to argue thus in the 

alternative. As O’Regan J explained: 

the common law [normally] develops incrementally through the rules of 

precedent ... [It sometimes happens that a court must] determine 

whether a new set of facts falls within or beyond the scope of an existing 

rule. The precise ambit of each rule is therefore clarified in relation to 

each new set of facts. A court faced with a new set of facts, not on all 

fours with any set of facts previously adjudicated, must decide whether 

a common-law rule applies to this new factual situation or not. If it holds 

that the new set of facts falls within the rule, the ambit of the rule is 

extended. If it holds that it does not, the ambit of the rule is restricted, 

not extended. 93 

O’Regan J had no doubt that a court facing such a decision is bound 

by the mandate of section 39(2) — to heed the Bill of Rights when 

‘developing’ the common law — even though no ‘new’ or ‘changed’ 

common law rule is in the offing: 94 

The overall purpose of section 39(2) is to ensure that our common law is 

infused with the values of the Constitution. It is not only in cases where 

existing rules are clearly inconsistent with the Constitution that such an 

infusion is required. The normative influence of the Constitution must be 

felt throughout the common law. Courts making decisions which involve 

the incremental development of the rules of the common law in cases 

where the values of the Constitution are relevant are therefore also 

bound by the terms of section 39(2). The obligation imposed upon courts 

by section 39(2) of the Constitution is thus extensive, requiring courts to 

be alert to the normative framework of the Constitution not only when 

91 [2005] ZACC 8; 2005 6 SA 419 (CC) (NK). 

92 NK (n 91 above) para 14. 

93 

NK (n 91 above) para 16. 

94 As above.


some startling new development of the common law is in issue, but in all 

cases where the incremental development of the rule is in issue. 95 

By ‘incremental development’, O’Regan J plainly meant the process 

she had just previously been describing: ‘A court [facing] a new set of 

facts, not on all fours with any set of facts previously adjudicated, 

[deciding] whether a [settled] common-law rule applies to this new 

factual situation or not.’ In NK, O’Regan J held that the judicial work 

in such cases should count as ‘development’ of the common law for 

the purposes of section 39(2). 96 It seems she felt she had to say so, in 

order to establish firmly the Constitutional Court’s jurisdiction in the 

case before her, which she saw as presenting a question about 

whether certain facts fell properly within the rule of respondeat 

superior, construing and applying that rule with due regard to the Bill 

of Rights. 

Madala J’s majority in NM found that defendant Smith’s conduct 

fell under the head of ‘intentionally’ wrongful privacy-invasion — 

insomuch as (or so the majority found from the evidence) Smith ‘at 

least foresaw the possibility that the consent had not been given to 

the disclosure’. 97 Might we not well understand them as having 

followed in the tracks of O’Regan J’s work in NK — deciding, under 

pressure from the Bill of Rights, ‘whether the common law rule 

applies to this new factual situation or not’? ‘While the claim falls to 

be dealt with under the actio iniuriaum,’ the majority had already 

reminded us, ‘the Constitution must inform the application of the 

common law’. 98 Were they just vamping there, or might we take 

those words as a cue to understanding that the majority’s application 

to Smith’s conduct of the common law category of ‘intention’ (which, 

we must remember, already included dolus eventualis), was shaded 

or tinctured with a view to bringing or keeping the common law 

doctrine in full conformity with the objects of Bill of Rights? 99 

It is not impossible to see why the majority might have chosen 

such a path to judgment in Smith’s case, in preference to a judicially 

declared, surgical replacement of the term ‘intention’ in Madala J’s 

item (c) by ‘intention or negligence’ — even granting that the latter 

95 NK (n 91 above) para 17. 

96 See NK (n 91 above) paras 16, 17. 

97 

NK (n 91above) para 64 (Madala J). 

98 NM (n 5 above) para 28; see also para 50 (‘If human dignity is regarded as 

foundational in our Constitution, a corollary thereto must be that it must be 

jealously guarded and protected.’). 

99 I am not alone in thus understanding Madala J’s opinion; I have Sachs J for 

company. See NM (n 5 above) paras 201-02 (Sachs J) (‘It is in [the] human rights 

context [addressed by the Bill of Rights] that the competing interests at stake in 

the present matter must be dealt with. In a fittingly accessible manner, Madala J 

has indicated how in the particular circumstances of this case competing needs 

with respect to human dignity, on the one hand, and freedom of expression, on 

the other, should be reconciled.’)


technique carries with it a more immediately accessible package of 

instruction to learners of the law. 100 Terminological replacement is 

inevitably — to some irreducible degree — done at wholesale, and it 

is — to that irreducible degree — inevitably a blunter instrument than 

pure case-by-case steering. If — to illustrate — we follow the carefully 

plotted line of thought in O’Regan J’s NM opinion (agreed to by both 

Langa CJ and Sachs J), a due regard for the Constitution’s privacy 

right calls for addition of negligence liability to the actio iniuriarum 

for privacy invasion. 101 But hold the phone. A due regard for the 

Constitution’s freedom-of-expression principle calls for at-least 

initial or provisional restriction of the to-be-added negligence liability 

to a category of ‘media’ defendants, with ‘media’ defined 

expansively enough to include the authors and publishers of books as 

well as the more strictly journalistic, daily and periodical print and 

electronic media. 102 There is nothing in itself wrong with such a 

definitional-categorical approach, conducted with the sort of caution 

and care evinced by O’Regan J’s opinion. But neither can it be thought 

indefensible for a judge to decide that the process of drawing the 

distinctions that contextual variations will require is better pursued, 

at least for the time being, through the inflections of case-by-case 

application of the extant rule and the operations of stare decisis (as 

explained by O’Regan J in NK) than through process of word-andphrase-surgery-and-implant 

otherwise known as ‘development’ of the 

common law, even if the choice for pure, case-by-case adjustment 

entails some sacrifice of transparency to rote learners. 

Transparency is not entirely foregone, after all, as long as stare 

decisis holds in measurable degree. That is why I say that Madala J 

spoke too hastily in saying that there was no need for him to consider 

any possible ‘chilling effect’ of his judgment on freedom on 

expression, given that (as he said) he was merely applying the 

common law of delict as it then stood, not ‘extending’ it to include 

liability for negligence. 103 As O’Regan J’s analysis in NK makes clear, 

the NM majority’s quite striking application of the ‘intention’ 

standard to the facts of the NM case cannot avoid making waves as a 

precedent, and so the question whether those waves will have a 

constitutionally significant chilling effect cannot be avoided. 

As I am about to show, however, the right answer to that question 

is pretty plainly ‘no’ in Smith’s case, because the true and entirely 

100 

‘Clear and accessible norms’ in the words of Sachs J. NM (n 5 above) para 204. 

101 See NM (n 5 above) paras 92, 95 (Langa CJ); paras 171-82 (O’Regan J); paras 203- 

04 (Sachs ). Sachs J speaks of the extended standard of liability as one of 

‘reasonableness’, but he evidently has in mind approximately what Langa CJ and 

O’Regan J mean by ‘negligence’. 

102 See NM (n 5 above) paras 92-93, 95, 98-99 (Langa CJ); paras 203-05 (Sachs J); 

paras 177-82 (O’Regan J). 

103 NM (n 5 above) para 69.


innocuous ‘rule’ of Smith’s case can easily be taken to be this: If you 

are publishing an account of events and transactions including private 

facts about real persons, and you attach the true names of those 

persons to those facts, without having directly in hand and before 

your eyes proof-positive of their express permission to have you do so, 

and you cannot adduce any credible reason to think that a public that 

knows the names is better informed or otherwise more enriched that 

a public that does not, you are acting in culpable disregard (call it 

‘negligent’, call it ‘reckless’) of those persons’ presumptive interest 

in retaining control over dissemination of that private fact about 

themselves. Lest the bar and lower bench should be left in the 

slightest doubt about the lesson to be drawn, Sachs J took pains to 

point ‘the moral of the story’: 

[U]nless overwhelming public interest points the other way, publishers 

should refrain from circulating information identifying the HIV status of 

named individuals, unless they have the clearest possible proof to 

consent to publication having been given. 104 

That does not seem so hard to learn, even for ‘second-year LL.Bs’. 105 

I have little doubt of the message by now having found its way into 

every well-heeled South African publisher’s and writer’s in-box. 

Of course, it is very far from a full treatise on the values served 

by the constitutional rights of privacy and freedom of expression, by 

comparison, say with the entirely welcome offerings we find in the 

opinion of O’Regan J — memorable passages on the values 

underpinning the constitutional rights to privacy and freedom of 

expression, of which I would not wish the South African constituency 

(or the world constituency, for that matter) to be deprived. 106 For 

those passages, or anything remotely comparable, the NM majority 

opinion can claim no credit. And yet it does not necessarily follow that 

the majority opinion fails to give ‘identifiable content’ to provisions 

in the Bill of Rights, or that it leaves the bar and lower bench without 

worthwhile new instruction about ‘how the Bill of Rights might 

operate’ in an important class of relevantly similar, future 

matters. 107 NM might still be treated as one of those occasions, 

mentioned by Woolman, of an ‘opportunity’ thrown up by the stream 

of cases with varying fact-patterns, to ‘expand our understanding of 

how given rights function in given environments.’ 108 

104 NM (n 5 above) para 209. 

105 Woolman ‘Amazing’ (n 2 above) 787 n 47. 

106 See NM (n 5 above) paras 126-34 (on the values underpinning the constitutional 

right to privacy); paras 144-46 (on the values underpinning the constitutional 

right to freedom of expression). 


108 Woolman ‘Amazing’ (n 2 above) 765.


My proposed account of the NM majority’s finding of liabilityproducing 

‘intention’ on the part of defendant Smith has the further 

merit of providing an honorable explanation for the majority’s refusal 

of deference to the High Court’s contrary finding. When extant 

common law appears to require development in terms of section 39(2) 

or section 8(3), the Constitutional Court follows — and for good reason 

— a general practice of giving the High Courts and the Supreme Court 

of Appeal a first crack at the work. 109 No one, however, would suggest 

that the Constitutional Court should defer when it comes to deciding 

— as the Court had to decide, for example, in Khumalo — whether a 

common law solution reached by the common law judiciary (so to call 

them) is sufficiently in synch with the Bill of Rights as not to call for 

further correction. Nor, by the same token, should we expect the 

Constitutional Court to defer when it judges whether a lower court’s 

in-context application of a common law rule has been carried out with 

due regard for the Bill of Rights — see NK — any more than when the 

Constitutional Court judges whether a lower court’s in-context 

application of a statute has complied adequately with the command 

of section 39(2). 110 

3.8 Smith’s negligence (or ‘intention’?) 

I turn now to the division among the three judges who thought the 

question of Smith’s negligence should properly be decisive of the 

case. I do so with the expectation that our examination of this division 

will cast some further light on the majority’s finding of Smith’s 

tortious ‘intention’. 

Having found an absence of intentional wrongdoing by Smith, 

Langa CJ, Sachs J, and O’Regan J all had to confront the question of 

possibly developing the common law so as to impose liability for at 

least some instances of negligently wrongful invasion of privacy. All 

three agreed, at least tentatively, on the introduction of a negligence 

standard for ‘media’ defendants only, roughly on the model of the 

109 See, eg, Masiya (n 4 above) para 17 (‘Ordinarily, constitutional matters involving 

the development of the common law should first be taken to the Supreme Court 

of Appeal before they reach this Court because of the breadth of its jurisdiction 

and its expertise in the common law.’); Carmichele (n 59 above) paras 55-56: 

The proper development of the common law under section 39(2) requires close 

and sensitive interaction between, on the one hand, the High Courts and the 

Supreme Court of Appeal which have particular expertise and experience in this 

area of the law and, on the other hand, this Court. Not only must the common 

law be developed in a way which meets the section 39(2) objectives, but it must 

be done in a way most appropriate for the development of the common law 

within its own paradigm. There are notionally different ways to develop the 

common-law under section 39(2) of the Constitution, all of which might be 

consistent with its provisions. Not all would necessarily be equally beneficial for 

the common law. 

110 See, eg, Bato Star Fishing (Pty) Ltd v Minister of Environmental Affairs & Tourism 

[2005] ZACC 8; 2005 6 SA 419 (CC) paras 72-98.


SCA’s introduction in Bogoshi 111 of a ‘reasonable-to-publish-even-iffalse-in-fact’ 

standard of defamation liability for such defendants, 112 

approved by the Constitutional Court in Khumalo as an acceptable 

reconciliation of the potentially conflicting demands of the Bill of 

Rights regarding the rights to human dignity and to freedom of 

expression. 113 Where the three came apart was over the application 

of a negligence standard to the evidence in the NM case. 

Of the three, O’Regan J was by far the most outspoken on the 

point — which seems surely correct — that all such applications must 

be sensitive to all of the affected objects of the Bill of Rights, which 

in this case very saliently included freedom of expression. 114 In order 

to avoid imposing too great a burden on freedom of expression, said 

O’Regan J, the law must leave journalists free to ‘to publish 

information provided to them by reliable [published] sources without 

rechecking in each case whether the publication was lawful,’ at least 

where ‘there are no grounds for ... suspicion [of] a risk that the 

original publication was not lawful.’ 115 In the view of O’Regan J, 

grounds for suspicion were lacking in Smith’s case. Therefore, to hold 

her liable, 

one would ... have to find that wherever a reputable source has 

published [names of persons about whom private facts are disclosed], 

secondary publication may not take place without the existence of 

informed consent having been independently verified, so that in each 

case, the subsequent publisher would have to re-ascertain the facts. 116 

And such a rule, O’Regan J concluded, would impose an 

‘unacceptable burden on the dissemination of information.’ 117 

There is an apparent flaw in this reasoning, and it lies in the word 

‘wherever’ (as in ‘wherever a reputable source has published ... ’). 

Let us grant O’Regan J her premise (strenuously disputed by all the 

other justices) that grounds for suspicion of a lack of the plaintiffs’ 

consent to unlimited public identification by name were entirely 

absent from the totality of the circumstances that Smith confronted. 

111 National Media Ltd v Bogoshi 1998 4 SA 1196 (SCA); [1998] 4 All SA 347 (SCA). 

112 See NM (n 5 above) paras 172-77 (O’Regan J). 

113 See Khumalo (n 54 above) paras 43-45. It is interesting to note how Khumalo and 

the NM minority approach the ‘reasonableness’/‘negligence’ standard from 

opposite directions. Khumalo approved a shift of the balance in favor of freedom 

of expression (and thus against dignitas) by moving away from a prior rule of 

strict liability for false-in-fact, defamatory publications, whereas as the NM 

minority would have shifted the balance in favor of dignitas (and thus against 

freedom of expression) by introducing negligence liability in respect of what had 

previously been treated as a strictly intentional tort. 

114 See Khumalo (n 54 above) paras 1444-48. 

115 NM (n 5 above) paras 185, 187. 


117 As above.


(More on this below.) It still does not follow that holding Smith liable 

for negligence in this case would signal liability ‘wherever’ a 

journalist or book-writer follows reputable sources in publishing 

private facts about named persons, and consent from those persons 

unexpectedly turns out to have been lacking. It only follows that 

liability will attach to the conduct of the second publisher when she 

has no sufficient reason for taking the slightest risk of invading 

privacy by including the names. That is a far cry from ‘wherever’, and 

moreover it is exactly what is required by a general norm — we call it 

‘negligence’ — of acting with due regard for other peoples’ interests 

at risk, very saliently including (in the South African legal order), their 

constitutionally recognised interests both in freedom of expression 

and in privacy. 

O’Regan J was prepared to find from Smith’s testimony that Smith 

‘simply did not entertain the possibility that either the University of 

Pretoria or Professor Strauss would have sent a report to a Member of 

Parliament in circumstances where the consent given was only of a 

limited variety in a publication that did not draw attention to that 

fact.’ 118 I take that to signify an ultimate finding by O’Regan J that, 

on a preponderance of the evidence, Smith was just sucked in 

unthinkingly by deceptive appearances that would have prevented 

the question of the scope of the plaintiffs’ consent from even arising 

in her mind or anyone’s. 119 

Yet the evidence on that point, as described by O’Regan J, seems 

to me quite open to an opposite construction. I have in mind a passage 

in Smith’s book, flagged by O’Regan J, that appears to say that Smith, 

while writing the book, decided to include the names after reflecting 

consciously on a series of events that led her to infer that the 

‘consent’ referred to in the Strauss Report must have been ‘a general 

consent on which she could rely’. 120 By what seems to me a quite 

permissible reading, the passage in question shows Smith consciously 

deciding that she has a sufficient basis for inferring that a general 

consent had been given, and the names, therefore, can safely be 

included. I notice also Smith’s testimony ‘point[ing] to three facts 


119 ‘On a preponderance of the evidence’ because O’Regan J allows that Smith 

carries the onus of establishing absence of negligence, given proof of the 

objective fact of her having publishing private facts without consent. See NM (n 5 

above) para 179. 

120 ‘In the chapter concerned, Ms Smith recounts the story of the clinical trial. She 

then points to the fact that details of the applicants’ complaints were originally 

published in the New York Review of Books, though under pseudonyms. She then 

points to the fact that pseudonyms were not subsequently used in the Strauss 

Report but that the real names were given and she uses the real names. The 

chapter itself therefore suggests that she thought that after the New York Review 

of Books article had been published, the applicants had consented to the 

publication of their names by Professor Strauss and that that consent was a 

general consent on which she could rely.’ NM (n 5 above) para 166.


which grounded her belief that Professor Strauss had obtained general 

consent from the applicants for the publication of their names.’ 121 If 

we take this testimony to mean that Smith adverted consciously to 

these facts in forming her belief, at a time when she was still writing 

her book, then the testimony amounts to Smith’s admission that she 

knew then that she was choosing on the basis of something less than 

full and direct knowledge of consent to publication by her. To put the 

matter in language lawyers will understand, Smith was (her own 

testimony suggests) consciously choosing on the basis of 

circumstantial and not ‘real’ evidence. 

The testimony as a whole (I mean as described by O’Regan J; I 

have not read the trial transcript) looks to me quite consistent with — 

I do not say it compels, but Smith has the onus — a finding that Smith, 

while writing, adverted consciously to the question of the scope of 

consent, attempted a judgment about that matter on the evidence 

she had, and judged erroneously. I do not say she judged that 

particular question irrationally, or with the slightest bad faith, or 

even without the very heavy odds in her favor. I do say she 

nevertheless acted irresponsibly, in an important sense that the legal 

notion of negligence precisely nails. I say that, under a negligence 

standard properly understood, Smith’s mistake of judgment cannot be 

excusable, no matter how unlikely the error, when there was literally 

nothing worthwhile to be gained by taken any chance at all. 

But was there really nothing to be gained? ‘Freedom of 

expression,’ states O’Regan J, 

is important because it is an indispensable element of a democratic 

society. But it is indispensable not only because it makes democracy 

possible, but also because of its importance to the development of 

individuals, for it enables them to form and share opinions and thus 

enhances human dignity and autonomy. Recognising the role of freedom 

of expression in asserting the moral autonomy of individuals 

demonstrates the close links between freedom of expression and other 

constitutional rights such as human dignity, privacy and freedom. 

Underlying all these constitutional rights is the constitutional 

celebration of the possibility of morally autonomous human beings 

independently able to form opinions and act on them. 122 

One concurs wholeheartedly, or at any rate I do. The question that 

stares one in the face, though, is how any of these interests in 

democracy, autonomy, or opinion-formation is compromised in the 

slightest (or ‘chilled’, if you like) by holding Smith liable for her 

borderline-inexplicable choice to publish the real names of three 

otherwise non-notable women who figured as persons living with HIV 

121 NM (n 5 above) para 163 (O’Regan J). 

122 NM (n 5 above) para 145 (O’Regan J).


in Smith’s no-doubt highly informative account of Patricia De Lille’s 

public career. What is the value-added by inclusion of these women’s 

names? Smith’s own explanation of the choice is that she included the 

names in order to give her book ‘authenticity’. 123 Please! Would the 

book have been one whit the less authentic, the less informative, the 

less democracy- or autonomy-supportive, if Smith had fictionalised 

the names, explaining in a preface or a footnote that she had done 

so? 124 Surely a person acts negligently when she chooses a course of 

action that anyone can see has the capacity to injure others, and a 

reasonable member of society would perceive an obvious way to avoid 

the risk of injury, which is clearly of lesser cost to anyone than the 

injury might be to any who would suffer it. 125 

In Smith’s case, avoidance was available at the negligible cost of 

forbearing to include the names. The cost of avoidance thus would not 

have been that of the onerous detective work posited by O’Regan J, 

or of the delays in publishing publicly valuable content that such work 

might occasion; it would only have been the loss of the smidgin of 

novelistic coloration that Smith added to her book by inserting the 

publicly meaningless real names of three women who were otherwise 

living their lives entirely out of the public eye — a cost bordering on 

zero, it would seem, in a public perspective. 

And now the circle closes. Consider that the effrontery of seeking 

to excuse the decision to name the plaintiffs by the appeal to 

‘authenticity’ may have been instrumental in leading Madala J’s 

majority to its conclusion that Smith’s choice was ‘intentional’ in the 

pertinent, legal sense. 126 Sachs J, supporting a finding of negligence, 

remarked that ‘if the slightest doubt existed, there was no need to 

publish the actual names of the defendants.’ 127 I so not see why one 

123 See NM (n 5 above) paras 24, 61. 

124 If ‘authenticity’ really means scooping the competition, or showing what a 

superior gumshoe you are, it cannot, in the Constitution’s scales, weigh tellingly 

against a risk to other people’s privacy and dignity. If ‘authenticity’ means 

providing your readership with persuasive indication that you really have had 

access to authoritative sources, or to the sources you say you have accessed, it 

does weigh tellingly in principle, but not in application to this case. Smith had 

plenty of other concrete particulars to carry that burden, not to mention Patricia 

De Lille’s endorsement of her book as an ‘authorised’ biography. 

125 ‘The traditional test for negligence is axiomatic ... : negligence is established if a 

reasonable person in the position of the defendant would have foreseen the 

harm, the reasonable person would have taken steps to prevent it and the 

defendant did not take those steps.’ NM (n 5 above) para 100 (Langa CJ), citing 

Krueger v Coetzee 1966 2 SA 428 (A) at 430E. It would seem that Langa CJ meant 

‘would have foreseen some risk [not the certainty or near-certainty] of harm’. 

Later on in his opinion, he wrote that the crucial question in the case was 

‘whether the reasonable journalist [in Smith’s position] would have foreseen the 

possibility of absence of consent.’ NM (n 5 above) para 105. I have suggested that 

the evidence is compatible with a finding that Smith did in fact foresee that 

possibility — granted that she, with good reason, regarded it as very slight. 

126 See NM (n 5 above) paras 54, 61, 65. 



should not go a bit further and say: Since there was nothing of value 

to be gained by publishing the true names (recall the possibility of 

openly fictionalising them), it was negligent to publish them without 

proof positive of consent, expressly directed to the would-be 

publisher in question. 

For this reason, I am led to quibble with the statement of Langa 

CJ that ‘the’ crucial question to be decided in Smith’s case was 

‘whether the reasonable journalist [in Smith’s position] would have 

foreseen the possibility of absence of consent’. 128 That is, of course, 

a highly pertinent question for any judge approaching this case. To my 

mind, though, an equally crucial question is whether a responsible 

author, aware that she lacked proof-positive of the plaintiffs’ consent 

to the inclusion of their names in her book (as distinguished from 

Professor Strauss’s report), would have included the names. One can 

agree with O’Regan J that the reasonable author would not have 

‘foreseen’ the possibility of a restricted consent, if by that we mean 

that the reasonable author would have rated that possibility 

extremely unlikely, and still find plainly negligent — plainly fraught 

with an unjustifiable risk of harm to others — the publication of 

private facts about named persons whom you do not know and with 

whom you have never met, whose full and clear consents you do not 

have in hand, and who might just as well have gone nameless in your 

publication, so far any anyone’s legitimate interests are concerned. 

Are we perhaps verging here on ‘reckless disregard’? Gaining, then, 

some further insight into the majority’s finding of ‘intentionally’ 

wrongful conduct by Smith? 129 

3.9 A ‘minimalist’ moment? 

O’Regan J’s opinion does not address at all this matter of the 

vanishingly low value for Smith’s book of naming the plaintiffs, even 

though other justices raised it quite vividly in their opinions, and even 

though the omission leaves in limbo both O’Regan J’s conclusion that 

Smith was not negligent and her complaint about the costs to freedom 

of expression of imposing liability in cases such as Smith’s. If my 

analysis is sound, those costs are virtually nil. A flat rule of liability 

for publishing the names of persons attached to private facts, where 

the names are publicly meaningless and add nothing to the opinionforming 

and autonomy- or democracy-supporting value of the 

published content, could easily be derived as a contextual 

specification of the negligence principle for cases to which it applies. 

Such a rule would seemingly be something of a gain for dignitas and 


129 NM (n 5 above) para 125. O’Regan J construes the majority’s finding of 

intentionality in terms of reckless disregard.


an utterly negligible loss for freedom of expression. It would thus 

seem to be a rule upon which all members of the Constitutional Court 

could agree: a triumph, one might say, of ‘minimalism’. 130 

I am thinking here of the dimension of ‘width’, not ‘depth’, 131 for 

(as I hope I have sufficiently shown) there is substantial depth in the 

solution I described for the NM case. In the list of possible benefits of 

a minimalist sensibility marshaled by Iain Currie (following Cass 

Sunstein), 132 there is one that does not appear: Minimalism can be 

conducive to getting the answer right! I mean right in principle, 

getting a principle right. Judges whose habit it is to scan for possible 

narrow grounds of decision, in cases where their court or their 

country’s legal culture is evidently divided over broader-gauged terms 

of debate, may sometimes be the likelier to detect some set of 

relatively contextualised terms of decision that perhaps will not only 

induce a voting consensus, but also may help to clarify for everyone 

concerned the true (or at any rate an agreed) meaning for the 

broader terms of contention (here, ‘intention’ and ‘negligence’) that 

have been leading their debate into trouble. 

Thus, a focus on a relatively contextual point in NM — the point 

that the inclusion of the plaintiffs’ names was sheerly gratuitous from 

the standpoint of the public informational and autonomy-supportive 

value of Smith’s account of Patricia De Lille’s career — might have 

helped clarify for everyone what the negligence principle really 

means and is all about. If your lines of reasoning get fixed at too 

wholesale a level — say, something like ‘negligence liability for 

publishing and freedom of expression do not mix’ — you not only may 

overlook a possibly consensus-forming resolution of the case before 

you, you may fail to see how your generalisations are leading you to a 

conclusion that traduces the true point and meaning of negligence as 

a legal concept, and of negligence liability’s value as an instrument 

for the advancement of the objective normative value system for 

which the Constitution of South Africa is supposed to speak. 

3.10 A case of ‘direct’ or ‘indirect’ application? 

A loose end still flaps in the breeze. In part 3.4 above, I compared the 

litigation agendas respectively implied by a constitution-based and a 

common law cause of action in NM. The Constitution-based action 

succeeds (if it does) by way of an exercise of the court’s power of 

130 See I Currie ‘Judicious avoidance’ (1999) 15 South African Journal on Human 

Rights 139 147-50. 

131 A judicial decision is ‘narrow’ (as opposed to ‘wide’) when it is ‘minimal in 

impact on future decisions’. It is ‘shallow’ (as opposed to ‘deep’) when it is 

‘minimally theorised’. See Currie (n 130 above) 147. 

132 See Currie (n 130 above) 146-50.


judicial review: declaration of invalidity and ensuing, remedial 

modification of the common law. 133 The common law action succeeds 

(if it does) by way of a court’s exercise of its inherent power to 

develop the common law in the ordinary course of adjudication, under 

pressure to conform the common law to the norms of the 

Constitution. Granting this notional difference in the respective 

judicial powers to be put into play, I concluded that the resulting 

litigation agendas would be virtually identical, whichever source of 

power would be chosen. I warned, however, of an arguable difference 

to be taken up later. Later is now; the arguer for the difference is Stu 

Woolman, and the difference argued for is this: Arriving at a 

declaration of invalidity, required for the plaintiffs’ success in a 

constitution-based action, involves a court in a distinctly more 

rigorous and exacting mode of constitutional analysis and exposition 

than does arrival at a decision to exercise the inherent power to 

develop the common law. 

Here it is in a nutshell: 

... [T]he drafters intended for there to be two different processes. The 

first process — direct application — takes the [specifically named] rights 

and freedoms, and the general rules derived from them, as our point of 

departure for determining whether [some challenged parcel of common 

law doctrine] is invalid. The second process — indirect application — 

allows for a mode of analysis that neither specifies whether a particular 

right demands vindication nor permits a finding of invalidity. Instead, ... 

the courts operate under a general injunction to bring all law into line 

with the ‘spirit, purport and objects’ of the Bill of Rights and the 

‘objective, normative value system’ made manifest in the text of the 

Constitution as a whole. 134 

The framers, Woolman says, had a particular reason for their 

provision of these two, distinct processes — direct and indirect 

application — for measuring the constitutional acceptability of a 

challenged common law doctrine or rule. As he explains: 

Section 8 does not mean that the prescriptive content of the substantive 

provisions in the Bill of Rights covers each and every legal dispute ... 

[W]hile the specific provisions in the Bill of Rights cover a large domain 

of law and conduct, they do not engage all law and conduct. The 

133 That is, owing to the command of sec 8(3), it does not succeed by way of 

damages imposed on Smith for a constitutional tort. 



independent purpose of s 39(2) is to engage law and conduct not 

engaged by any of the specific provisions set out in Chapter 2. 135 

This reason for providing the two, distinct processes of direct and 

indirect application implies, as Woolman further points out, a definite 

order of march for courts called upon to consider the constitutional 

acceptability of one or another rule or doctrine of the common law: 

... [O]ne must first ascertain what the ambit is of the allegedly 

applicable constitutional provisions [he means the several rightsgranting 

clauses in the Bill of Rights]. Only when one has determined 

that ambit, and found that it does not speak to the issues raised by a 

[questioned] rule of [the common] law, can one turn to the more 

open-ended invitation of s 39(2). Analysis of the specific provisions of 

the Bill of Rights, and the consistency of law or conduct with those 

provisions, logically must be prior to the analysis of the common law in 

terms of the general spirit, purpose and objects of the Bill of Rights. 136 

In other words, the marching order, logically implied by the 

Constitution’s principle of division of labor between direct and 

indirect application is: direct application first, then indirect, but only 

if direct fails to engage the common law rule or doctrine in 

question. 137 

But of course Woolman’s point is not merely that the Constitution 

implies a particular order of priority of resort to direct and indirect 

application of the Bill of Rights. It is that the obedience of courts to 

this order of march — direct before indirect — is of crucial importance 

to the long-run interest in the proper elucidation of South African 

constitutional law. And that is because of a difference in the sorts and 

levels of explanatory rigour respectively exacted from courts by 

direct and indirect application. When applying the Bill of Rights 

directly in terms of section 8, the Court ‘engages the content of the 

substantive provisions of the Bill of Rights’ and ‘articulates 

135 As above. 

136 Woolman ‘Amazing’ (n 2 above) 777. Woolman elaborates: 

If we reverse the spin, and we first use sec 39(2) to bring the law into line with 

the general spirit, purport and objects of the Bill of Rights, there is simply 

nothing left to be done in terms of direct application. The reason is obvious. If 

the general spirit, purport and objects of Chapter 2, which embraces (at a 

minimum) the entire value domain reflected by the specific substantive 

provisions of the Bill of Rights, does not require a change in the law (or a change 

in conduct brought about by a change in the law), then no narrower set of 

purposes reflected in a single substantive provision of the Bill of Rights could be 

expected to do so. 

137 See Woolman ‘Amazing’ (n 2 above) 83 (‘The logic ... is one that should require 

the courts to ... develop the common law in light of the general objects of the 

Bill of Rights only where no specific right can be relied on by a party challenging a 

given rule of common law ... ’) (emphasis in original).


constitutional rules that amplify that content.’ 138 When applying the 

Bill indirectly in term of section 39(2), it does not. 

Appropriately does not? Obediently does not? Well, Woolman 

wants us to read sections 8 and 39 as setting up two cleanly ‘distinct 

modes of analysis’. 139 The suggestion would seem to be that a court, 

by choosing to frame its inquiry in terms of section 39(2), thereby 

limits itself to the detection of inconsistencies between the parcel of 

common law in question and what we might call the normative 

residue of the Bill of Rights, meaning whatever normative intimations 

a court can draw from the Bill of Rights as a whole but cannot trace 

or ascribe to any of the specific rights-naming clauses. If analysis of a 

rule’s consistency with the specific provisions of the Bill of Rights 

‘logically must be prior’ to analysis of its consistency with the 

normative residue, then the latter process — indirect application — 

must be reserved for cases in which the former cuts no ice. And 

conversely, then, if a court, in a private-on-private case turning on 

common law doctrine, is going to find a ‘direct’ inconsistency — 

meaning an inconsistency with something contained in one or another 

the rights-naming clauses — then that court really ought not, on, that 

occasion, speak the name of 39(2). 140 

But how, after all, can it be so important that this particular 

discipline be observed? I do not mean the discipline of searching for 

clause-based inconsistencies before turning to the question of 

residual inconsistencies, about which I raise no question here. 141 

I mean — and mean only — the discipline of not saying ‘39(2)’ 

when a clause-based inconsistency is what you are after. For this 

specific practice rule, I can find no reason in Woolman’s magisterial 

work on application (which is generally rich in jurisprudential and 

even philosophical reflection) beyond doctrinal tidiness and 

symmetry. The point is to avoid ‘redundancy’, ‘surplusage’, ‘elision’ 

in the constitutional text — or, in other words, overlap between the 

sorts of judicial performances respectively authorised by sections 8 


139 Woolman ‘Amazing’ (n 2 above) 776; see Part II above. 

140 According to Woolman’s preferred construction of secs 8 and 39(2) as a package, 

sec 39(2) should be understood to stand for the following proposition: Where no 

specific right can be relied upon by the party challenging a given rule or law or 

the extant construction of a rule of law the courts are obliged to interpret 

legislation or to develop the law in light of the general objects of the Bill of 

Rights. Woolman ‘Application’ (n 17 above) 12. See also at 80 (indicating that, on 

the preferred construction, ‘s 39(2) is not the intended engine for changes in the 

common law that flow from the direct application of the Bill of Rights’). 

141 A question could, however, be raised about it. See FI Michelman ‘Unenumerated 

rights under popular constitutionalism’ (2006) University of Pennsylvania Journal 

of Constitutional Law 121, 135-36, 152-53 (discussing views of Ronald Dworkin 

and John Rawls).


and 172(1), on the one hand, and 39(2), on the other. 142 We would not 

want two constitutional clauses sharing any of the same work. 143 

Some might wonder whether Woolman’s interpretive construction 

of the framers’ design — his strict correlation of clause-bound 

inspection to sections 8 and 172(1), and residual inspection to section 

39(2) — perhaps proceeds from an unrealistically high expectation of 

formal tidiness and tightness in constitutional drafting. On that score, 

NK may give a hint. In NK, O’Regan J’s flexing of the respondeat 

superior standard plainly was driven, not by some normative residue 

detected in the Bill of Rights, but by attention to the prescriptive 

contents of several, specific clauses, including those ensuring rights 

to ‘freedom and security of the person, and in particular, the right to 

be free from all forms of violence from either public or private 

sources as well as [the] right to dignity, right to privacy and right to 

substantive equality.’ 144 O’Regan J plainly was engaged in the first of 

the two ‘instances’ of need for common-law development in terms of 

section 39(2), recalled by her from Thebus: the kind that arises when 

‘a rule of the common law is inconsistent with a constitutional 

provision’. 145 Was she, therefore, wrong to file the decision under the 

head of 39(2) rather than 8? 

Might that depend on whether O’Regan J considered herself, after 

all, to be engaged in ‘development’ of the common law, as 

contemplated by section 8(3)? Does ‘develop’, there, encompass only 

an express, terminological re-writing, or does it also take in a court’s 

choice about how to apply an established standard to an arguably 

‘new’ set of facts, as in NK? Section 39(2) speaks to a court’s 

142 See Woolman ‘Amazing’ (n 2 above) 771, 777; Woolman ‘Application’ (n 17 above) 

10, 11, 12, 55. The ‘two instances’ reading of sec 39(2) in S v Thebus 2003 6 SA 

505 (CC); 2003 10 BCLR 1100 (CC) para 28 (‘Thebus’) (where one of the instances 

is ‘when a rule of the common law is inconsistent with a constitutional provision’) 

is objectionable to Woolman because it ‘offends the “no surplusage” canon of 

constitutional interpretation by making parts of FC sec 8 redundant.’ Woolman 

‘Application’ (n 17 above) 79. 

143 

Woolman adduces an incisive and entirely persuasive reason for not allowing sec 

39(2) to suck up, exclusively or preemptively, the whole field of constitutionally 

propelled re-examination of the common law. That reason has to do with the 

currently imposed, substantial exclusion of the High Courts from such reexamination 

when conducted under the aegis of sec 39(2). See Afrox Health Care 

v Strydom 2002 6 SA 21 (SCA); Ex parte Minister of Safety and Security v Walters 

2002 4 SA 643 (CC), 2002 2 SACR 105 (CC), 2002 7 BCLR 663 (CC). A like restriction 

cannot be, and has not been, imposed on High Court re-examination where ‘a 

party has a colourable claim grounded in the direct application of a substantive 

provision of the Bill of Rights in terms of section 8'. It is, therefore, of crucial 

importance, Woolman argues, to resist an ‘elision’ in which sec 39(2) would, in 

effect, swallow up sec 8. Woolman ‘Application’ 10, 55-56, 64. This argument is, 

to me, entirely compelling. It does not, however, explain why courts should be 

barred from ever resorting to sec 39(2) when the detected inconsistency is with a 

specific provision, which is the issue that my text, above, addresses. 

144 

NK (n 91 above) para 14. 

145 NK (n 91 above) para 16, citing Thebus (n 142 above) para 28.


‘interpreting’ law, whereas section 8(3) speaks to a choice between 

‘applying’ and ‘developing’ law. Does that difference, perhaps, make 

section 39(2) more apt to the sort of work accomplished by the Court 

in NK? 

It might seem so, to some; it certainly wouldn’t have to, to all of 

us. But I do not mean my string of questions to constrain very strongly 

— much less lead ineluctably — toward a conclusion either way. I only 

mean it to confirm a truth that Woolman, more than anyone, has 

recognised and exposed through his meticulous, sterling, searching 

work on the application conundrum: When the question is that of 

parsing the exact relations and connections among this Constitution’s 

several clauses on application and remedies, no fully neat and 

satisfying answer is to be had. ‘Some surplusage seems inevitable.’ 146 

There are not shells to cover all the peas, and so it is a choice among 

imperfections — in that sense, a matter, as Woolman says, of 

‘interpretation’. 147 

3.11 The proof of the pudding 

But, as Woolman also rightly insists in the same place, it does not 

follow that grounds are lacking for preferring one interpretation — 

one construction — over another. One ground of preference might 

well be the differing ways in which the respective constructions cue 

and guide, concentrate and dissipate, the judiciary’s deployments of 

its efforts and energies. Woolman notes that a ‘muscular’ use of 

section 39(2) ‘could’, in theory, produce the sorts of rights-based, 

rights-expounding work he finds the Court retreating from. 148 But he 

fears that, in practice, a Court that is allowed to give up too easily on 

section 8, or to slide too easily to section 39, will be (has been) a 

Court that too often fails to decide in terms of any of the specific, 

rights-naming clauses in the Bill of Rights, and accordingly fails too 

often to engage in the principled elucidation of those clauses that 

decision in such terms entails. That states a legitimate concern, on its 

face by no means implausible. Still, the claim is an empirical one, 

subject to testing as such. NK, I have just suggested, offers a bit of 

evidence against it. (So, for that matter, does Thebus, where the 

Constitutional Court, under the sign of section 39(2), considered in 

great detail whether a prosecutor’s reliance on the common-purpose 

doctrine ‘trenches on’ or ‘denies’ the constitutional rights of a 

criminal accused not to be deprived of freedom without just cause 

and to be presumed innocent.) 149 

146 Woolman ‘Application’ (n 17 above) 71. 

147 See Woolman ‘Application’ (n 17 above) 13 n 1. 

148 Woolman ‘Application’ (n 17 above) 766 n 6. 

149 Thebus (n 142 above) paras 36, 42.


To what extent, we might ask, does the Constitutional Court’s 

work in NM tend to bear out Woolman’s empirical concern? Start with 

O’Regan J. Her NM opinion certainly exemplifies the sort of 

elucidation that Woolman associates with direct application pursuant 

to section 8. 150 If O’Regan J had given clear indication that she was 

proceeding in terms of section 39(2), as distinct from section 8, that 

would have told against Woolman’s empirical claim. However, she did 

not. What might be a bit worrisome from Woolman’s corner, though, 

is the further fact that O’Regan J did not ever bother to say under 

which set of constitutional auspices — 8 or 39(2) — she was 

undertaking to develop the common law. (But Langa CJ, concurring 

with O’Regan J’s proposed development, did make clear that he 

understood himself to be proceeding in terms of section 39(2).) 151 

Neither did O’Regan J ever pause to declare invalid the extant 

common law of privacy invasion. All this perhaps does tell a bit against 

Woolman, insomuch as it might suggest that O’Regan J believes it 

does not matter which auspices she invokes, which would contradict 

Woolman’s insistence that it does matter, not just because an order 

of march has been prescribed, but because marching under section 

39(2) has an effect on the justices of releasing them from the very 

work of clause-by-clause exposition that O’Regan J’s NM opinion so 

clearly exemplifies. 

What about the NM majority? Madala J, denying that he was 

engaged in any sort of modification of the common law, had no 

occasion to claim any set of constitutional auspices for modifying it. 

But suppose he had said directly that his finding of ‘intention’, on 

Smith’s part, to disclose private facts without authorisation to do so 

from the persons concerned, resulted from a shading of the common 

law category of ‘intention’ under pressure of the Bill of Rights, 

specifically in the matter of privacy protection. What more could 

Madala J then have said by way of explaining his action, than that the 

disclosure went to the ‘inviolable core’ of privacy as previously 

defined in Bernstein, 152 and that the privacy-protection principle of 

section 14 requires exposure to civil liability of anyone who publishes 

core-private facts when the publisher ‘at least foresaw the possibility 

that the consent had not been given to the disclosure’? 153 

Madala J did not say those things in that way, nor could he have, 

because he did not purport to be judging this case in terms of the Bill 

of Rights, à la NK. But can we not fairly guess from his opinion that he 

150 See NM (n 5 above) paras 126-35 (constitutional right to privacy); paras 144-46 

(constitutional right to freedom of expression). 

151 See NM (n 5 above) para 120. 

152 Bernstein v Bester NO 1996 2 SA 751 (CC); 1996 4 BCLR 449 (CC) para 77. Compare 

NM (n 5 above) paras 33, 41-42. 

153 Compare NM (n 5 above) para 64.


would have said them, if he had undertaken to explain himself in such 

terms? If so, then NM is, to that very modest extent, a sign that 

section 39(2) does not necessarily, and does not always, function as a 

lure to a form of judicial decision-making that ignores and by-passes 

the regulative content of the rights-naming clauses of the Bill of 

Rights. Masiya, as we are about to see, provides some further 

evidence to like effect. 

4 MASIYA 

4.1 The case and a critique 

The state charged Masiya with the common law crime of rape. Under 

the law as it then stood, rape was defined in terms of non-consensual, 

vaginal penetration (and thus could involve female victims only). The 

evidence in Masiya’s case showed forcible, anal penetration of a nineyear 

old girl. 154 Anal penetration of a female or a male could support 

a conviction of indecent assault, but not of rape. Indecent assault is 

classified as the lesser crime, and it carries a lesser penalty than 

rape. 155 

In the course of the state’s prosecution of Masiya, questions arose 

as to whether the extant common law definition of rape fell to be 

declared invalid as inconsistent with the Constitution, and whether 

that law ought to be developed so as to cover Masiya’s proven act of 

anal penetration. 156 The two questions are not unrelated, but neither 

are they necessarily coupled. In the event of an adjudicative 

154 See Masiya (n 4 above) paras 6, 13. 

155 See Masiya (n 4 above) para 15 (indicating that ‘indecent assault attracts more 

lenient sentences than rape’). 

156 There was, of course, a further question about retrospective application to 

Masiya’s case of any extended definition the Constitutional Court might direct. 

The Court answered ‘no’ to that question, see Masiya (n 4 above) para 56, even as 

it also answered ‘yes’ to the question of whether the advent of Masiya’s case 

before the courts offered an apt occasion for developing the definition. See 

Masiya (n 4 above) para 51: 

If the definition of rape were to be developed retrospectively it would offend the 

constitutional principle of legality. However, if it was to be accepted that the 

principle of legality was a bar to the development of the common law, the courts 

could never develop the common law of crimes at all. Such a conclusion would 

undermine the principles of the Constitution which required the courts to ensure 

that the common law was infused with the spirit, purport and objects of the 

Constitution. 

From a US constitutional lawyer’s perspective, major worries remain, not only 

about prospective-only application of a common law ruling, but also, and 

relatedly, about South African constitutionalism’s evident relaxation of what we 

would call a ‘case or controversy’ prerequisite to judicial jurisdiction. See 

generally R Fallon et al Hart & Wechsler’s the federal courts and the federal 

system 5 ed (2003) 73-85. Woolman’s critique of the Constitutional Court’s Masiya 

decision being in no way directed to such worries, I say nothing more about them 

here.


declaration of the current definition’s constitutional invalidity, 

corrective development of the common law might well follow on as a 

part of a just and equitable remedial order, in terms of Constitution 

s 172(1). 157 But even in the absence of such a declaration, a court 

might arguably still find itself bound by Constitution section 39(2) to 

develop the definition so as cover Masiya’s case. 158 

On what grounds, then, might the Constitutional Court, sitting in 

Masiya’s case, have held constitutionally incompatible the extant 

common law of rape insofar as it fails to cover cases of anal 

penetration? Not, in the first place, on any theory of the law’s 

infringement of Masiya’s constitutional rights or those of other 

putative offenders covered by the definition. For, obviously, 

the current definition of rape criminalises unacceptable social conduct 

that is in violation of constitutional rights. It ensures that the 

constitutional right to be free from all forms of violence, whether public 

or private, as well as the right to dignity and equality are protected. 159 

In other words, any limiting effect of the law of rape on the liberty of 

putative rapists is obviously and amply justified in terms of 

Constitution section 36(1). 160 It easily follows that any declaration of 

the current definition’s unconstitutionality would have to issue in 

contemplation of constitutionally protected interests, not of those 

who might find themselves charged with assaultive crime, but of the 

class consisting of all who might suffer sexual assault — the general 

public, more or less. Thus, it was with a view to ‘afford[ing] society 

the full protection of the Constitution’ that the trial court in Masiya’s 

157 

It seems that, in Masiya’s case, Constitution sec 8 can also serve as the ground of 

an order to develop the common law, inasmuch as an antecedent declaration of 

the invalidity of the common law definition of rape might be said to result (in the 

terms of sec 8(2)) from an application of certain provisions in the Bill of Rights to 

the conduct (or to common law regulating the conduct) of Fanuel Sitakeni Masiya, 

a natural person. Since Woolman generally argues in terms of a choice between ss 

8 and 39(2), as the constitutional authorisation to be invoked by the Court for 

some ensuing directive for common law modification, and since nothing seems to 

depend on whether a remedial order would issue in terms of sec 172(1)(b) or in 

terms of sec 8, I shall, in what follows, refer to the one or the other as the 

context most naturally suggests. 

158 See Masiya (n 4 above) para 33 (quoting from Carmichele (n 59 above) para 36, 

and also citing Thebus (n 152 above) paras 28-31): 

... ‘[C]ourts must remain vigilant and should not hesitate to ensure that the 

common law is developed to reflect the spirit, purport and objects of the Bill of 

Rights ... whether or not the parties in any particular case request the Court to 

develop the common law under section 39(2).’ Where there is deviation from the 

spirit, purport and objects of the Bill of Rights, courts are obliged to develop the 

common law by removing the deviation. 

159 Masiya (n 4 above) para 27. 

160 See also Constitution sec 12(1)(a): ‘Everyone has the right to freedom and 

security of the person, which includes the right — (a) not to be deprived of 

freedom arbitrarily or without just cause’.


case took up sua sponte the question of extending the common law of 

rape to acts of anal penetration. 161 

For convenience, we might imagine that it would be the Director 

of Public Prosecutions who presses the claim for some development of 

the common law that would result in Masiya’s conduct being 

punishable as rape. 162 The DPP, pointing to the state’s constitutionally 

imposed duty to protect and fulfil the rights in the Bill of 

Rights, 163 might charge the received definition of rape with a 

constitutional sin of omission, to wit, its omission to make acts of anal 

penetration punishable as rape. Impermissibly under-protected as a 

result of that omission, the DPP might say, are the right to freedom 

from private violence, conferred by section 12(1)(c), 164 and the right 

to have one’s dignity respected and protected, conferred by section 

10. 165 Alternatively or additionally, the DPP might urge a declaration 

of invalidity based on a sin of commission by the current law, to wit, 

commission of an unfair discrimination prohibited by Constitution 

section 9(3). 166 It is unconstitutionally unfair, the DPP might urge, for 

the state, through its common law, to offer lesser levels of 

recognition, and of protection-by-deterrence, to male than to female 

sufferers of forced, sexualised, bodily penetration, or (within the 

class of female victims) to sufferers of anal than to sufferers of 

vaginal penetration. 

A judicial modification of the common law definition of rape, 

presented as a remedy for a constitutional violation declared on any 

of the grounds aforesaid, would be an instance of what Woolman 

classifies as ‘direct’ application of the Bill of Rights. But a court 

sitting in Masiya’s case might also arguably be authorised and required 

to undertake an extension of the common law definition of rape to 

reach cases of anal penetration, even in the absence of any finding of 

the classical definition’s inconsistency with any clause in the Bill of 

Rights. Start by setting aside the Bill of Rights entirely. Perhaps a 

court might simply decide that considerations of policy and justice 

call, under present social conditions, for extension of the common law 

definition of rape so as to take in cases of anal penetration, either of 

161 Masiya (n 4 above) paras 8-9. 

162 In point of fact the DPP did not urge any modification of the common law 

definition of rape but rather sought conviction for indecent assault. See Masiya (n 

4 above) para 7. I make the opposite assumption for the sake of expository 

convenience. 

163 See Constitution sec 7(2): ‘The state must respect, protect, promote and fulfil 

the rights in the Bill of Rights.’ 

164 ‘Everyone has the right to freedom and security of the person, which includes the 

right — ... to be free from all forms of violence from either public or private 

sources’. 

165 ‘Everyone has inherent dignity and the right to have their dignity respected and 

166 

protected.’ 

‘The state may not unfairly discriminate directly or indirectly against anyone on 

one or more grounds, including race, gender, sex ...’.


females only or of males as well. The court might then invoke its 

ordinary, inherent power to develop the common law as and when 

occasioned by the interests of justice. 167 In doing so, the court might 

or might not refer, beyond its general authority to keep the common 

law attuned to the times and to the interests of justice, to its more 

specific obligation, under section 39(2), to see to the conformation of 

the common law with the spirit, purport, and objects of the Bill of 

Rights. 

In Masiya, the Constitutional Court, in an opinion by Nkabinde J, 

concluded by issuing an order directing a revision of the common law 

definition of rape to include acts of anal penetration of females. 168 It 

did so while (a) expressly declining to issue a declaration of the 

invalidity (unconstitutionality) of the non-extended definition, and 

(b) declining to decide whether acts of anal penetration of males 

would be covered in the extended definition of rape resulting from 

the adjudication in Masiya’s case — thus, in effect, retaining for the 

present a definition of rape in which only females can be victims. 

Those two features of the Court’s Masiya decision, (a) and (b), are 

the targets of vigorous complaint from Woolman. Both stand, in his 

view, as further documentation of the Constitutional Court’s 

unaccountable habit — and that habit’s doctrinally debilitating 

consequences — of refusal to engage in direct application of the Bill 

of Rights or to undertake the rule-like expositions of the contents of 

the several rights-protecting clauses that direct application is 

expected to induce. 

As with Woolman’s objections to the Court’s performance in NM, 

these objections to Masiya seem to me in some part answerable. The 

main idea continues to be this: We want to see what happens if we 

strain to turn these very features in the judgment — the ones that we 

may initially find puzzling or disturbing — into keys to discovery of the 

Court’s own understanding of the legal problem presented by the 

case; and perhaps, through that, to notice something true — I do not 

mean necessarily welcome — about the meaning and structure of the 

law. When we train that sort of effort on the Constitutional Court’s 

work in Masiya, we may find that Woolman’s objections lose some – 

though it will not be all — of their sting. 

167 See Constitution sec 173 (‘The Constitutional Court, Supreme Court of Appeal and 

High Courts have the inherent power to protect and regulate their own process, 

and to develop the common law, taking into account the interests of justice.’). 

168 See Masiya (n 4 above) para 74(5).


4.2 The decision’s gravamen: Under-protection 

In Masiya, the Constitutional Court concluded that the extant 

common law of rape, although not prone to a finding of invalidity, 

nevertheless fell to be developed under pressure from the spirit, 

purport, and objects of the Bill of Rights. By my reading — and this 

point is central to my examination of Woolman’s critique — the 

decisive pressure came (in the Court’s view) not from the Bill’s 

prohibition against unfair discrimination, but rather from its positive 

commitments to protection of persons against violence and indignity. 

The Court does mention section 9 — ‘flirts with’ it, if you like 169 — but 

does not finally rest its action on considerations sounding in equality. 

As I have understood the Court’s opinion, the decisive fault it finds in 

the common law lies not in any unfair discrimination committed by 

the law but rather in protection omitted from it: not in the law’s 

perpetration of inequality, but in its under-protection of 

constitutionally guaranteed dignity and freedom from violence. A 

complaint of under-protection is, of course, a matter distinct in 

principle from a complaint of unequal treatment. By what I presume 

to offer as the ‘charitable’ reading of Nkabinde J’s opinion, it is the 

under-protection complaint that carries the day in Masiya. 

Oh, yes, to be sure, the ‘issues raised’ in the case included 

‘equality’ — specifically citing section 9(1)’s guarantee to everyone of 

‘equality before the law’ and ‘equal protection and benefit of the 

law’. 170 And, yes, Nkabinde J made the comparative observation that 

the trauma associated with anal penetration is ‘just as humiliating, 

degrading, and physically hurtful’ as that associated with vaginal 

penetration. 171 She said, as well, that inclusion of anal penetration in 

the definition of rape ‘will increase the extent to which the 

traditionally vulnerable and disadvantaged group will be protected by 

and benefit from the law’. 172 But all of those remarks are perfectly 

consistent with the view that the current law is constitutionally 

deficient in the specific respect that it under-protects the rights to 

dignity and freedom from violence. They could all easily have been 

written by a judge applying a bill of rights that contains no equality 

clause. 

As Nkabinde J’s opinion proceeds, we read the following: It is the 

office of the crime of rape to ‘ensure[] that the constitutional right 

to be free from all forms of violence, whether public or private, as 

well as the right to dignity and equality are protected.’ 173 It is 


170 Masiya (n 4 above) para 18 & n 32. 


172 As above. 

173 Masiya (n 4 above) para 27.


specifically with regard to these objectives that the current definition 

is fatally ‘under-inclusive’. 174 ‘Non-consensual anal penetration of 

women and young girls such as the complainant in this case 

constitutes a form of violence against them equal in intensity and 

impact to that of non-consensual vaginal penetration.’ 175 Extending 

the definition to include anal penetration will ‘protect the dignity of 

survivors’. 176 It will ‘increase the extent to which [a] traditionally 

vulnerable and disadvantaged group will be protected’ against 

‘humiliating, degrading and physically hurtful’ treatment, and 

‘therefore’ — meaning, in that particular respect — it will ‘harmonise 

the common law with the spirit, purport and objects of the Bill of 

Rights.’ 177 

In sum, it is all about a constitutionally required level of 

protection, not about the distinct injury to the self or to dignity that 

flows from being subjected to second-class treatment. The Bill of 

Rights sections in the foreground here are 10 and 12, not 9. The Court 

convicts the current law of a constitutional sin of omission of due 

protection; it does not convict it of a constitutional sin of commission 

of unfair discrimination. 

4.3 A case of judicious avoidance? 

In Woolman’s view, this short-changing of section 9 counts as 

unimitigated vice in the decision. In mine, it can claim some support 

in ordinary judicial prudence. What, after all, is the context — the 

field — of the differential treatment that troubles Woolman? 178 It is 

the grading of crimes within nests of including and lesser-included 

offenses. (It is not as if Mr Masiya escapes conviction of crime — 

serious crime — if he is convicted of indecent assault and not rape.) 

What is involved is the schedule of penalties devised by the lawmaker 

for more and less aggravated variations on the same core offense. 

Plainly, it is not against the grain of the Bill of Rights, or contrary to 

any clause contained therein, for lawmakers to draw such lines. To 

the contrary, we should think it monstrous if they did not. 

From an equality-clause standpoint, then, an important question 

presented to the Constitutional Court by the Masiya appeal was one 

of best judicial practice: whether it would be generally in the 

interests of justice, and conducive to long-term optimal performance 

174 As above. 




178 Note that I am not yet dealing with the issue raised by the Court’s postponement 

of the question of the new rape definition’s application to cases involving male 

victims.


by the judiciary of its role in the over-all governmental order of a 

democratic South Africa, for the Constitutional Court to swing open 

its doors to complaints of unconstitutionally unfair discrimination, 

assertedly wrought by the lawmaker’s choices of where and how to 

draw the lines required by the gradations of penalties (for more and 

less aggravated incidents of the same class of crime) that the Bill of 

Rights and the rule of law almost certainly demand. 179 If you think 

that is an easy question, it must be because you feel sure that answer 

is ‘no’. If you feel the answer might well be ‘yes’, you must 

nonetheless grant that answer is not immediately self-evident. The 

Constitutional Court, I accordingly conclude, had respectable reason 

for hesitating to tackle the question in a case where the need for a 

challenged law’s correction was sufficiently clear on constitutionally 

resonant grounds of substantive under-protection, equality 

considerations being held at bay. 

4.4 A case of direct application? 

In her Masiya judgment, Nkabinde J concludes that the extant 

common law definition of rape ‘is not inconsistent with the 

Constitution’. 180 Accordingly, no declaration of invalidity is to issue, 

and no development of the common law definition of rape is to ensue 

in the guise of a remedy for constitutional invalidity. 181 Rather, ‘the 

definition is to be extended ... so as to promote the spirit, purport 

and objects of the bill of rights.’ 182 Evidently in play, then, is an 

exercise of the Court’s inherent authority, in terms of sections 173 

and 39(2), to develop the common law, as opposed to its judicialreview 

power, conferred by section 172(1), to declare invalidity and 

craft just and equitable remedial orders. 183 Just as Woolman says. 

Let us next notice what stands immediately behind Nkabinde J’s 

refusal to declare the extant rape law unconstitutional and hence 

invalid. Here is the crucial passage: 

The current definition of rape criminalises unacceptable social conduct 

that is in violation of constitutional rights. It ensures that the 

constitutional right to be free from all forms of violence, whether public 

or private, as well as the right to dignity and equality are protected. 

179 I have in mind here a general standard of appraisal of the Constitutional Court’s 

performance closely aligned to the one projected by Roux ‘Pragmatism’ (n 16 

above). 




183 See Masiya (n 4 above) para 45 (‘[T]he extension of the common-law definition of 

rape to include non-consensual anal penetration of females will be in the 

interests of justice and will have, as its aim, the proper realisation by the public 

of the principles, ideals and values underlying the Constitution.’).


Invalidating the definition because it is under-inclusive is to throw the 

baby out with the bath water. What is required then is for the definition 

to be extended instead of being eliminated so as to promote the spirit, 

purport and objects of the Bill of Rights. 184 

We can all see the conceptual problem to which Nkabinde J is calling 

attention there. Woolman sees it and chalks it up to ‘fuzzy logic’, a 

point to which we shall return. 185 

Nkabinde J suggests a comparison of the rape crime at issue in 

Masiya with the erstwhile common law crime of sodomy, at issue in 

the ‘Sodomy’ case. 186 Let us follow her lead. Decidedly not by way of 

‘a process of developing the common law’, but by way of ‘a direct 

application of the Bill of Rights to a common-law criminal offence’, 

the sodomy crime was simply wiped off the books, declared 

‘constitutionally invalid in its entirety’. 187 If there, why not here? 

Because it is not true in Masiya’s case, as it was in the Sodomy 

case, that the parcel of common law at issue ‘subjected people to 

criminal penalties for conduct which could not constitute a crime in 

our constitutional order’, 188 and therefore could and should be 

declared null and void by reason of inconsistency with provisions in 

the Bill of Rights. Rather, the opposite is true. The rape crime, as 

historically defined, ‘criminalises unacceptable social conduct that is 

in violation of constitutional rights’. 189 Nullification of the sodomy 

crime as defined by the common law leaves all as it should be, 

constitutionally speaking. Nullification of the rape crime as defined 

by the common law leaves ... what? 

Of course, legal insiders know how to spin out an acceptable 

answer: It leaves, we would say, the common law in need of remedial 

modification – say, by a remedy of ‘reading in’ — as a cure for its 

constitutional sin of omission, its defect of inconsistency with sections 

10 and 12(1) of the Bill of Rights. That is the gist of Woolman’s 

complaint against Nkabinde J’s logic. 190 The complaint is entirely 

valid. And yet something also feels right about Nkabinde J’s detection 

of the insult to ordinary reason that lies in condemning, as 

inconsistent with the Constitution, a legal prohibition whose presence 

on the books the Constitution positively demands. A remedial 

authority that would subtend on such a strange-sounding declaration, 

184 Masiya (n 4 above) para 27 (citations omitted). 


186 See Masiya (n 4 above) para 27. 

187 ‘Sodomy’ Case (n 20 above) para 90. 


189 As above. 

190 See Woolman ‘Amazing’ (n 2 above) 768 n 10 (‘The court appears to conclude that 

a finding of invalidity of a rule of common law can result only in a simple 

declaration of invalidity.’).


she quite clearly conveys, ought not to be the first resort of the Court 

when some other resort is available, some other path of reasoning to 

a Constitution-satisfying outcome. And indeed another path is 

available, providentially supplied by the framers. The ram in the 

thicket (block that metaphor!) is section 39(2). 

The choice for section 39(2) is understandable in those terms, if 

not fully defensible. My reason for drawing that distinction will soon 

be on the table, but first we should confirm why Woolman is right to 

criticise the choice. To work the question through, we can start by 

imagining that the Constitution had been so written as to foreclose 

recourse in this case to section 39(2); say, it flatly prohibited the 

courts from performing terminological surgery on the common law, 

except as a remedy for constitutional invalidity declared. With such a 

prohibition in place, should the Constitutional Court have been willing 

to declare invalid the extant common law rape crime, that being the 

only path left open to them to a judicial cure for what they see as a 

constitutionally unendurable sin of omission in the law, that is, its 

failure to impose appropriately severe condemnation and penalty on 

forcible acts of anal penetration? If you ask me, a Court sitting in a 

transformative constitutional order would have no choice but to do 

just that, awkward as the declaration of invalidity might feel to the 

Court. 

The Constitutional Court, it seems, took the view that it did not 

have to perpetrate that awkwardness in Masiya’s case. They had 

access to the omission-fixing cure in question via section 39(2), so why 

should they have engaged in the seeming doubletalk of declaring 

constitutionally invalid a constitutionally required parcel of the 

extant law — only in the next breath to restore what they have just 

condemned as unconstitutional, in the guise of an instantaneously 

ensuing ‘remedy’? Why mangle straight-line, ordinary reason in that 

way, when the drafters have provided a fit alternative? 

A conclusive answer to that question is prefigured by my imagined 

case of the constitutional prohibition on judicial rewriting of the 

common law, except as a remedy for declared constitutional 

invalidity. Notice that, with respect to statutes, such a constitutional 

prohibition is not imagined, it is actual! In terms of section 39(2), 

statutes are open to judicial interpretation but not to judicial 

‘development’. By its holding that the old rape definition’s omission 

of anal penetration is ‘not inconsistent with the Constitution’, the 

Court has, in principle, disabled itself from curing an identically 

flawed act of Parliament should one ever happen to issue. And that, 

surely, is an unacceptable implication. 191 

191 I am indebted for this point, which I had failed to see, to an anonymous referee.


The Masiya Court thus erred in withholding a declaration of 

invalidity of the old common law of rape. Its error, however, consists 

only of resort to section 39(2) where section 172(1) seemed to the 

Court not to fit the case while section 39(2) seemed fully serviceable. 

The error is not at all symptomatic of a flight from the burdens of 

clause-by-clause exposition of constitutional content that attend 

upon direct application. 

The focus of the crime of rape, writes Nkabinde J, ‘ is on the 

breach of “a more specific right such as the right to bodily integrity” 

and security of the person and the right to be protected from 

degradation and abuse.’ 192 Presumably, there was something, some 

line of thought, that Nkabinde meant to convey by that direct allusion 

to O’Regan J’s opinion for the Court in Dawood. What, then? Well, 

that is the passage in Dawood where O’Regan J exactly differentiates 

between dignity as ‘a value that informs the interpretation of many, 

possibly all, other rights’ (or, in other words, dignity acting as a 

condiment in the spirit-brew of the Bill of Rights) and dignity as ‘a 

justiciable and enforceable right that must be respected and 

protected’. Dignity, then, as a target of ‘direct’, not ‘indirect’, 

application of the Bill of Rights, as Woolman understands those terms. 

There, and in other passages I have quoted above, Nkabinde J’s 

opinion makes sufficiently clear that the Masiya Court’s modification 

of the common law is designed to respond, not to some spirit 

summoned from the vasty deeps of the Bill of Rights, but rather to 

certain, specifically identified, distinctly articulated demands of 

sections 10 and 12(1)(c) of the Bill — most particularly, the 

Constitution’s mandate to protect people against private violence. As 

used (however mistakenly) in the Masiya case, section 39(2) is not 

serving as an escape-hatch from the Court’s responsibility for clauseby-clause 

elucidation of the contents of the Bill of Rights. It is rather 

serving as a handy authorisation for a moderately drastic bit of 

judicial surgery on the common law, in circumstances where the 

Court finds inapposite section 172(1)(b)’s license to operate qua 

remedy for invalidity. 

In making this use of section 39(2), Nkabinde J followed precedent 

set by the Constitutional Court in Carmichele. Just as the Court there 

chose not to condemn, as constitutionally out of line, whatever 

affirmative, protective duties were already imposed on the police by 

the extant common law of delict, so the Court in Masiya chose not to 

condemn, as constitutionally out of line, whatever liabilities to 

criminal conviction were already imposed on perpetrators of sexual 

192 Masiya (n 4 above) para 25, quoting from Dawood v Minister of Home Affairs 2000 

3 SA 936 (CC); 2000 8 BCLR 837 (CC) para 35.


violence by the extant common law of crimes. In both cases, the 

extant impositions of duty and liability were in order as far as they 

went. In both, the issue was one of broadening those impositions; 

there was no question of contracting, much less deleting them. Just 

as the Constitutional Court in Carmichele found that development of 

the common law under the aegis of section 39(2) was the apt course 

to take in those circumstances, so did the Court in Masiya. In both 

cases, moreover, the Court made sufficiently clear its view that a 

duty- and liability-extending modification of the common law was 

required not just by some immanent, residual, order of values to be 

distilled from the Bill of Rights as a whole, but also by an application 

of the Constitution’s specific guarantees respecting dignity and 

freedom and security of the person. 193 

It seems we might extract from the Carmichele/Masiya pair a 

general guideline of South African constitutional law, as follows: 

When the constitutional fault detected in some duty- and liabilityimposing 

parcel of the common law lies not at all in the imposition 

of duty and liability, but only in the failure to define the type of 

duty- and liability-raising conduct in question with sufficient 

protective effect, then the authorisation for judicial surgery on the 

common law is to be found not in the court’s judicial-review/ 

remedial power granted by section 172(1), but rather in its inherent 

power to develop the common law in terms of section 39(2). The 

reason for this result is not any thought that under-protection in dutydefining 

laws can only offend against some residual ‘spirit’ of the Bill 

of Rights as a whole, but not against any clear, core, animating 

principle of any particular rights-naming clause. The reason for the 

result simply is that the section 172(1)(b) remedial power is expressly 

conditioned on an antecedent declaration of inconsistency of some 

law with the Constitution, and, in this class of cases — so thinks the 

Court — there is nothing to declare inconsistent. 

In sum: At stake here is a notional distribution of the source of 

judicial authority and obligation to develop the common law, as 

between two clauses in the Constitution. Not at stake is a flight from 

direct application of the Bill of Rights, as Woolman defines ‘direct’. 

Rather, the Carmichele/Masiya pair (and let us also recall NK and 

Thebus) shows the Constitutional Court rejecting Woolman’s 

preferred construction of the relations among sections 8, 39(2), and 

172(1), as outlined in Part 2 of this article. 

193 See Carmichele (n 59 above) para 44 (‘[T]here is a duty imposed on the state and 

all of its organs not to perform any act that infringes these rights. In some 

circumstances there would also be a positive component which obliges the state 

and its organs to provide appropriate protection to everyone through laws and 

structures designed to afford such protection.’); Masiya (n 4 above) paras 25, 27, 

37, 38, 39, reviewed above, text accompanying notes 169-177.


Woolman, of course, is fully aware of this gap between his own 

preferred construction and the Court’s ‘black letter’ construction. He 

is nevertheless careful not to suggest that the gap must necessarily 

make a world of difference in practice. 194 To the contrary, he 

explicitly points out that the Constitutional Court, using its own 

construction, could make a ‘muscular’ use of section 39(2) that ‘could 

generate the kinds of decisions and constitutional rules’ he is looking 

for. 195 But he sees this possibility as ‘entirely hypothetical’, asserting 

that the Constitutional Court has not, in fact, ‘deployed s 39(2) in the 

muscular fashion that would be necessary to rebut my primary 

arguments.’ 196 One can assume, though, that were the Court, acting 

under the aegis of section 39(2), to search specifically for 

incompatibility between some challenged law and particular 

provisions of the Bill of Rights, that should count for Woolman as 

direct application. 197 I believe that Masiya — like Carmichele and NK 

— is most naturally read as just such a case, as opposed to a case of a 

non-differentiated appeal by the Court to some vaguely conjured 

‘spirit’ of the Bill of Rights entire. If Woolman does not see the cases 

that way, that might possibly be because his normative commitments 

are getting in the way, by which I mean his insistence that section 8 

and 39(2) must, on pain of redundancy, have mutually ‘independent’ 

purposes, 198 entailing distinct tests for the common law’s liability to 

revision: incompatibility with ‘the specific provisions of Chapter 2’ 

under section 8, incompatibility with the normative residue under 

section 39(2). 199 

4.5 The sex-discrimination issue: Easy or hard? 

For several pages, now, I have been reflecting on the Constitutional 

Court’s designation, in Masiya, of section 39(2) — and not section 8 or 

172(1) — as the source of its authority to review the High Court’s order 

in respect of developing the common law of rape. I turn now to 

another feature of the Constitutional Court’s Masiya decision that has 

excited Woolman’s concern, namely, the Court’s postponement of 

the question whether the definition of rape, once extended to cover 

anal as well as vaginal penetration, must also be extended to cover 

penetration of males as well as females. My first suggestion (but then 

things will grow more complicated) will be that Woolman’s difference 

with the Court in this regard depends on a substantive disagreement 

concerning the merits of a particular constitutional claim. More 

precisely, it has to do with a conflict between Woolman’s and the 

194 See generally Woolman ‘Application’ (n 17 above). 


196 As above. 

197 See as above. 


199 See Woolman ‘Amazing’ (n 2 above) 769, 771.


Court’s respective preliminary judgments regarding the merits of a 

claim against the rape definition’s consistency with Constitution 

sections 9(1) and 9(3), if that definition, once extended to cover anal 

penetration of females, is not immediately extended again to cover 

anal penetration of males. 

Of course, Masiya’s case does not squarely present that question; 

only a prosecution for rape in a case of anal penetration of a male 

could do that. Woolman does not — and I doubt he ever would — raise 

a blanket objection to judicial hesitancy to reach for hard 

constitutional questions whose resolution is not necessary to a 

complete disposition of the case sub iudice. What he does clearly 

think is that the constitutional violation is, in this instance, so 

exceptionally and blatantly clear that the Constitutional Court could 

appropriately have resolved it on the spot. 200 The hangup is that the 

Court apparently does not share Woolman’s perception of the 

obviousness of the answer to the equality clause question thus posed. 

If the question really is debatable in light of the applicable 

precedents, then the Court’s forbearance to decide on the spot will 

seem (to an American constitutional lawyer, at any rate) to be 

routinely predictable for any court in a constitutional democracy, 

disposing over an assigned power of judicial review; the forbearance 

simply cannot then qualify as probative for a general complaint about 

a judicial addiction to excessive minimalism. And the Court, I am 

about to suggest, had detectible reason to assess the equality clause 

question as debatable under its own precedents. 

In the view of the trial court in Masiya’s case — as well as in the 

apparent view of Woolman 201 — restricting the newly extended rape 

law’s protection to women is ‘irrational and totally senseless’, an 

‘arbitrary discrimination’ 202 — and thus, it would seem, a direct 

violation not only of section 9(3) but also section 9(1), according to 

the ‘tabulation’ for a section 9 inquiry laid down by Goldstone J in 

Harksen. 203 In the view of the High Court, the restriction results in 

‘discriminatory sentencing’, an apparent invocation of section 9(3)’s 

prohibition of unfair discrimination. 204 What is more, the 

Constitutional Court agrees that ‘non-consensual anal penetration ... 

200 Having noted his agreement with Langa CJ’s ‘unassailable’ view that ‘it makes no 

sense to distinguish between men and women’ with respect to the gravity of the 

crime of anal rape, Woolman goes on to say that ‘in these circumstances’ the 

Court was wrong to refrain from addressing the sex-specificity issue in a case 

whose facts did not require its resolution. Woolman ‘Amazing’ (n 2 above) 767. 

201 As above. 


203 See Harksen v Lane NO [1997] ZACC 12; 1997 11 BCLR 1489; 1998 1 SA 300 para 53 

(item ‘(a)’ in the tabulation). 

204 Masiya (n 4 above) paras 2, 15.


is [no] less degrading, humiliating and traumatic’ for males than for 

females. 205 But, again, the hangup is that the Constitutional Court 

does not allow that it automatically follow from this equivalence that 

a gender-specific definition of rape must be unconstitutional. 206 

How could it not? The Court does not spell out its answer, but 

those conversant with the Court’s equality jurisprudence can see one 

sitting there. Section 9(3) prohibits unfair discriminations on grounds 

of sex, but not all sex-based differentiations are unfair (and much less 

are all of them arbitrary and thus in violation of section 9(1)). 207 

According to the Court’s carefully developed and precisely articulated 

doctrine, discrimination is ‘unfair’ when and insofar as it has the 

effect of impairing dignity or inflicting some comparably serious 

harm. 208 That may not be uncontroversial doctrine or (for aught I 

know) the doctrine favored by Woolman, but substantive doctrine it 

surely is, quite carefully and elaborately wrought by the 

Constitutional Court. 

So consider: In the view of the Constitutional Court, 

it is ... widely accepted that sexual violence and rape not only offends 

the privacy and dignity of women but also reflects the unequal power 

relations between men and women in our society. 209 

Rape, said the Court, is understood in South Africa as being ‘not 

simply an act of sexual gratification, but one of physical domination’ 

— ‘an extreme and flagrant form of manifesting male supremacy over 

females.’ 210 Historically, said the Court, ‘rape has been and continues 

to be a crime of which females are its systematic target.’ 211 In the 

prevailing societal view as reported by the Court, not only do females 

thus bear the major brunt of the degradation and brutality inflicted 

by sexual violence, but where such violence is consciously or 

unconsciously directed to the subordination of some sex or gender — 

as it somewhat systematically is, according to what the Court takes to 

be the dominant understanding — that sex or gender is the female or 

feminine, not the male or masculine. 

Perhaps that all adds up to a controversial batch of observations 

about prevailing South African understandings. (I do not know and 

shall not try to say.) But let us ask: Accepting the observations as 

correct, do they really provide a grader of penalties for crimes with a 


206 As above. (‘That this is so does not mean that it is unconstitutional to have a 

definition of rape which is gender-specific.’). 

207 See Harksen (n 203 above) para 53. 

208 See Harksen (n 203 above) paras 46, 50, 51(c). 



211 As above.


decisive reason to tilt the balance more favorably to liberty and 

mercy, as against prevention and condemnation, when sexual 

violence is directed against males? On first look, I find it far from 

obvious that they do. I cannot, however, deny that the question is a 

hard one that requires careful thought. The Court’s observations do, 

then, point to a possible reason, not inherently injurious to the dignity 

of men or of women — any more than, in Hugo, President Mandela’s 

reasons for restricting his commutation order to mothers were found 

injurious to the dignity of men or of women 212 — why a lawmaker 

might decide to treat sexual violence toward women as a more 

seriously graded crime than sexual violence toward men. If not unfair 

in Hugo, why unfair here? One can so inquire in good faith, without for 

a moment doubting that that anal rape is a monstrous affront to the 

dignity and security of male as well as female victims — any more than 

any justice sitting in Hugo could have doubted that imprisonment is 

hard for men as it is for women. 213 

Granted, ‘strong arguments’ could be made — and were, to the 

Constitutional Court in Masiya’s case — ‘to the effect that genderspecificity 

in relation to rape’ must be deemed constitutionally unfair 

insofar as it undoubtedly reflects ‘patriarchal stereotypes 

inconsistent with the Constitution.’ 214 Exactly parallel arguments 

were made in Hugo, including by justices who found, contrary to the 

majority of the Constitutional Court, that the President’s genderspecific 

deployment of his pardon power was indeed an act of unfair, 

sex-based discrimination. 215 In Hugo, the claims of constitutionally 

unacceptable stereotyping met up with both respectful consideration 

and formidable rebuttal from the majority side. 216 In Masiya, the 

Court did not reject the parallel arguments, it rather left them for 

another day. 217 Perhaps there are members of the Court engaged in, 

or open to, reconsideration of the position taken in Hugo. The matter 

at issue is obviously difficult, deep, and of major import for South 

African sex-equality jurisprudence. It may well currently divide the 

justices, just as it divided them in Hugo. If they wait, views might 

undergo maturation and moderation. Parliament may speak, and that 

could make a difference. In Hugo, the sex-discrimination question 

could not wait; if the discrimination were to be found unfair and 

unjustified, a remedial question would then arise that conceivably 

could have issued in Mr Hugo obtaining release from prison. In Hugo, 

the sex-discrimination question could not wait; Mr Hugo's claimed 

right against unfair, sex-based discrimination was directly and 

212 See The President of South Africa v Hugo 1997 4 SA 1 (CC); 1997 6 BCLR 708 (CC). 

213 See Woolman ‘Amazing’ (n 2 above) 767. 

214 Woolman ‘Amazing’ (n 2 above) para 29. 

215 See Hugo (n 212 above) para 80 (Kriegler J); paras 93-94 (Mokgoro J). 

216 See Hugo (n 212 above) paras 37-39 (Goldstone J); para 113 (O’Regan J). 

217 See Masiya (n 4 above) paras 29, 36.


immediately at stake. In Masiya, by contrast, the question can wait; 

the prisoner at bar has nothing at stake in the resolution. 

4.6 Objective unconstitutionality 

But Woolman maintains that the question cannot wait, owing to the 

Constitutional Court’s adherence to the doctrine of objective 

unconstitutionality. 218 Woolman puts the matter thus: ‘[I]f one 

cannot imagine the court’s refusing [on some future occasion] to 

extend the ‘new’ rule to embrace coerced anal intercourse of men, 

then the new rule is objectively unconstitutional as of the moment it 

was announced.’ 219 From this, Woolman apparently deduces that the 

Court is obliged to decide now (Woolman thinks in the affirmative) 

whether extension of the new rule to the protection of men is 

constitutionally required. But how, exactly, would that follow? 

Consider this possible response: Under the doctrine of objective 

unconstitutionality, a court is not free to postpone to some future 

case a decision on the constitutionality of any law whose application 

to a pending case would be decisive, even if the putative 

constitutional defect in that law pertains strictly to rights of persons 

other than the parties to the pending case. (‘The right to challenge 

the constitutionality of a statute which affects you directly cannot be 

made dependent on the finding of some other constitutional right on 

which to base the challenge. What if there is no such right?’) 220 Since 

the Court could never justify convicting Masiya in terms of an 

objectively unconstitutional law, it is obliged to face the question of 

the new rule’s complete constitutionality as a part of its disposition 

of Masiya’s case. 

That cannot be Woolman’s argument, for the simple and obvious 

reason that the Constitutional Court is not, as matters turn out, 

proposing to apply the new definition of rape to the pending 

prosecution of Masiya. 221 The only law the Court proposes to apply to 

him is the old, constitutionally unassailable law defining the crime of 

indecent assault, and so there is no way that an on-the-spot 

assessment of the constitutionality of the new rule can affect the 

prosecution’s outcome. Insofar as the doctrine of objective 

unconstitutionality means to allow parties to pending cases to claim 

218 See, eg, Home Affairs (n 87 above) para 29: 

On the objective theory of unconstitutionality adopted by this Court a litigant 

who has standing may properly rely on the objective unconstitutionality of a 

statute for the relief sought, even though the right unconstitutionally infringed is 

not that of the litigant in question but of some other person. 


220 Ferreira v Levin NO 1996 1 SA 984 (CC), 1996 1 BCLR 1 (CC) para 163 (Chaskalson 

P) (‘Ferreira’). 

221 See Masiya (n 4 above) paras 56-57 (deciding against retrospective application).


the benefit of the unconstitutionality (on whatever grounds) of rules 

and doctrines being then and there invoked against them, it can have 

no purchase in the pending prosecution of Masiya (nor in any 

imaginable future prosecution of him, unless in respect of some 

utterly hypothetical, future criminal act). 222 

Thus, on what we might call a precise or strict reading of the 

doctrine of objective unconstitutionality, it cannot have the effect of 

requiring a decision of the putative section 9(3) claim against the new 

rule, as a part of the Constitutional Court’s disposition of the case of 

Masiya v DPP CCT 54/06. But is there, perhaps, some looser version of 

objective unconstitutionality that might have such an effect? I do not 

see what it would be. The doctrine of objective unconstitutionality 

cannot reasonably be taken to commit the Constitutional Court to onthe-spot 

adjudication of every conceivable claim of constitutional 

defect that might arguably afflict any of the rules and doctrines that 

in any way crop up in a case sub iudice. That way surely lies madness. 

We can use a variation on NM to illustrate. Suppose the Court 

decides to take the case up for review, in the expectation that it 

cleanly presents a question about the consistency of the extant 

common law of privacy-invasion with Constitution section 14, 

insomuch as the extant law flatly excludes liability for nonintentional, 

but negligent, unauthorised disclosures of private 

information. Upon study of the record, the Court discovers that it 

contains a binding stipulation from all sides, to the effect that the 

book-writer was not negligent or in any way at fault for her action — 

so all that could remain for the Constitutional Court to do would be to 

affirm the decision below on that ground. It seems Woolman would be 

among the first to insist (and I would agree) that the Court ought now 

to dismiss the appeal for want of jurisdiction. But why should that 

follow, after all? Why should not the Court rather say: ‘There is still a 

question in the air about whether the extant common law of privacyinvasion 

is constitutional, insomuch as it flatly excludes liability for 

negligence, so we really owe it to the country to decide that now, per 

the doctrine of objective unconstitutionality’. Is that how we want 

the Court to be conducting its affairs? 

222 Thus, this is not a case in which refusal to decide a constitutional question on the 

spot could possibly involve the Court in ‘recognis[ing] the validity of a [law] in 

respect of one litigant, only to deny it to another.’ Ferreira (n 220 above) para 26 

(Ackermann J). Neither — owing to the non-retrospectivity ruling in Masiya — is it 

a case in which the postponement compromises ‘legal certainty’ to the 

disadvantage of any individual contemplating some or other course of action (as 

above). Of course, the objective-unconstitutionality argument would have bite in 

Masiya’s case, if the Court had decided in favor of retrospective application of 

the new rule. But I do not take Woolman to be suggesting that the Court’s ruling 

against retrospective application is itself symptomatic of a flight from substantive 

decisionmaking responsibility.


Nothing that I know of in the Constitutional Court’s oeuvre 

suggests that the Court itself has ever thought so. The Court has 

rejected a ‘narrow’ approach to standing in constitutional cases. 223 

Accordingly, the Court will address constitutional challenges by or on 

behalf of any litigant — whether or not his or her own constitutional 

rights are at stake — who can show that ‘he or she is directly affected 

by’ the law in question, and in that sense has an ‘interest in’ the 

challenge. 224 The Court has furthermore indicated that it will be 

generous in deciding whether such an interest has been shown. 225 The 

Court has also, however, expressly recognised the factors militating 

against allowance of constitutional challenges by persons claiming 

only a ‘hypothetical’ or ‘academic’ interest: 

The principal reasons for this objection are that in an adversarial system 

decisions are best made when there is a genuine dispute in which each 

party has an interest to protect. There is moreover the need to conserve 

scarce judicial resources and to apply them to real and not hypothetical 

disputes. The United States courts also have regard to ‘the proper role 

of the Courts in a democratic society’ which is to settle concrete 

disputes, and to the need to prevent courts from being drawn into 

unnecessary conflict with coordinate branches of government. 226 

These would seem to be reasons why ‘few, if any, countries have at 

all times allowed all persons to invoke the jurisdiction of Courts to 

solve all legal problems.’ 227 South Africa does not draw the line nearly 

as strictly as some countries do (mine, for example), 228 but a line 

there surely must be, and drawing it at the point where the litigant 

raising the challenge has something at stake in its resolution — that 

is, qua litigant, not just qua citizen — seems not a bad place to draw 

it. That, at any rate, is what I understand the doctrine of objective 

unconstitutionality to do. Woolman fears the doctrine may be falling 

into some degree of ‘desuetude’. 229 I respond only that Masiya cannot 

stand as evidence for that proposition. 

4.7 The court’s responsibility as lawmaker 

Suppose you are persuaded that the doctrine of objective 

constitutionality, as currently propounded by the Constitutional 

Court, does not oblige the Court to address a section 9(3) claim 

223 Ferreira (n 220 above) para 162 (Chaskalson P). 

224 Ferreira (n 220 above) paras 165-66 (Chaskalson P). 

225 ‘It is for this Court to decide what is a sufficient interest in the circumstances.’ 

Ferreira (n 220 above) para 168 (Chaskalson P, rejecting Ackermann J’s view that 

the Applicants lacked standing to raise a particular constitutional challenge). 

226 Ferreira (n 220 above) para 164 (citation omitted). 

227 Ferreira (n 220 above) para 31 (Ackermann J). 

228 See, eg, n 154 above. 

229 Woolman ‘Application’ (n 17 above) 50.


against the new rule, as part of its disposition of Masiya’s case. That 

is certainly not yet to say that nothing so obliges it. What most likely 

would so oblige it, if anything does, is the simple fact that the Court 

is the author of the new rule, and we can rightly demand of the 

Constitutional Court that it — above all other agents in the South 

African constitutional order — take care, when it gets into the 

business of writing laws, to write only laws that are fully consistent 

with the Bill of Rights. I mean, just imagine that Cabinet, having 

decided to preempt completely the common law of sexually 

assaultive crimes with a new statutory code, tables before Parliament 

a draft that defines rape as the vaginal or anal penetration of 

females. It will most definitely be in order for objectors to demand 

that Parliament must now, before acting, decide whether, in its 

advised view, such a statute could pass constitutional muster under 

section 9(3). If that is so for Parliament, must it not be even more 

clearly so for the Constitutional Court? 

Maybe not. To say so with perfect confidence is to lose sight 

momentarily of some differences between a court and a parliament in 

a constitutional democracy. When Parliament passes a bill, it, by 

necessary implication, clears that bill of constitutional defect so far 

as it can tell (no doubt subject to possible judicial correction). When 

the Constitutional Court declines to decide the constitutionality of a 

parliamentary enactment, in a case where nothing can possibly turn 

on that decision, it does not do that; it rather says ‘I am a court, I am 

moderately self-restrained by habit that you [meaning the country] 

would not really wish me to shake off, and the time for that is not yet 

ripe.’ 

‘But’, you will justly reply, ‘the new rule we are concerned with 

here is not a parliamentary enactment; it is the Court’s own, direct 

creation; that is just the point! There is no inter-branch dialogue in 

progress here, no consideration of deference to democratic 

accountability, no ball in Parliament’s court, to motivate or warrant 

postponement; there is only the Court refusing to confront itself!’ I 

am on shaky ground here, owing to deficient understanding of the 

exact interplay between common law and statute law in the South 

African law of crimes. (The very idea of common law crime poses a 

major challenge to the American-trained constitutional-legal 

temperament!) My impression, though, from the Masiya case itself, is 

that Parliament has indeed, though the Criminal Procedure Act and its 

various schedules, quite concretely bought into the judiciary’s 

historic definitions of the various crimes and, in a quite real sense,


made those definitions its own — including, specifically, the sexdifferentiating 

aspect in the historic definition of the crime of 

rape. 230 If that is correct, then the full of set of normal, inter-branch 

considerations affecting the degree to which the Court reaches out for 

constitutional questions is operative here. 

It is, to be sure, a set of considerations that can be cashed only 

through exercises of judgment in particular cases. The cost of waiting 

always must be among the factors to be considered. That there is such 

a cost involved in the choice that concerns us here cannot be doubted. 

If (as Woolman believes) the Court must eventually decide to extend 

the rape definition to cover cases of male victims, the next rape-like 

assault on a man still won’t be prosecutable as rape. That is for sure, 

given the anti-retrospectivity ruling in Masiya; and that could make 

all the difference. But the odds don’t seem overwhelmingly that way, 

indecent assault is still a serious crime; and the Court’s public fidelity 

to its anti-imperialist habit is not in itself a negligible value. Had the 

Court grasped the nettle in Masiya that would not, in my view, have 

been over the line (given that there are common law crimes that the 

courts are responsible to define). But neither, in my view on present 

information, was the decision to wait an improper one. 

How we judge that question may depend on how obviously and 

inevitably right we judge the section 9(3) claim against the new rule 

to be. Someone who sees the claim as irresistible is much likelier to 

fault the Constitutional Court for declining to decide it on the spot, 

than is someone, like me, who sees the claim as debatable under the 

precedents as they stand. In presenting the claim in that light, I have 

gone well beyond anything written in the Constitutional Court’s 

opinion; no reference to Hugo or to Harksen, for example, appears 

there. But I have offered nothing that is not entirely consistent with 

what the Court did write, and perhaps nothing that is not more or less 

clearly indicated by what the Court did write, at least to those 

conversant with the body of the Court’s other work. In other words, I 

have followed the course of interpretive charity. Whether with 

resulting net profit or loss to the project of the advancement of the 

benign rule of constitutional law in South Africa must be for the 

reader to judge. 

230 See Masiya (n 4 above) paras 2, 7, 12.

THE CONSTITUTIONAL COURT, COURT 

WATCHERS AND THE COMMONS: 

A REPLY TO PROFESSOR MICHELMAN ON 

CONSTITUTIONAL DIALOGUE, ‘INTERPRETIVE 

CHARITY’ AND THE CITIZENRY AS SANGOMAS 

1 Background 

* I would like to thank the editors for inviting me to write this reply. In particular I 

would like to thank Professor Stu Woolman for regular conversation and extensive 

comments on the form and the substance of this reply. The paper also benefited 

from comments made by the participants at the Constitutional Court Review 

Conference (6-7 August 2008 Somerset West). Needless to say all errors are mine. 

63 

Tshepo Madlingozi* 

When I think back on the Constitutional Court’s 2007 term, one event 

stands out for me. On 20 August 2007, the Court was hearing Merafong 

Demarcation Forum v President of the Republic of South Africa. 1 

After a long period of struggle in which they had employed both 

institutional and extra-institutional mechanisms of democracy, the 

people of Khutsong decided to take their case to the Constitutional 

Court. Why? ‘The government does not want to listen.’ 2 On that day 

more than a 1 000 protesters gathered outside the court room toyi 

toying and singing liberation songs. During the course of the day, 

things turned nasty when protesters started burning tyres, 

brandishing dangerous weapons and allegedly pelting the police with 

stones. 3 For me, this episode vividly demonstrates the fragile state of 

South Africa’s constitutional culture. 

1 Merafong Demarcation Forum & Others v President of the Republic of South 

Africa & Others 2008 5 SA 171 (CC). 

2 I Lekota ‘Flames in Khutsong’ Sowetan 25 May 2007 http://www.sowetan.co.za/ 

3 

Columnists/IdoLekota/Article.aspx?id=473030. 

See ‘Khutsong battle reaches Constitutional Court’ Mail & Guardian On-line 20 

August 2007 http://www.mg.co.za/article/2007-09-20-khutsong-battle-reachesconstitutional-court 

(accessed 23 October 2008). For dramatic photos capturing 

these events, visit http://photos.mg.co.za/view_photo.php?pid=2036&gid=124 

(accessed 23 October 2008).

64 Reply to Michelman 

In their opposition to the incorporation of Merafong municipality 

into the North West province, the people of Khutsong had done 

everything conceivably possible within the law: They had participated 

in public consultation forums where they voiced their opposition 

against the proposal; they sent memoranda to state organs at all three 

levels of government; they had lobbied political parties and other 

stakeholders; they had engaged in peaceful marches and protests. 

Indeed before resorting to the litigation, the people of Khutsong had 

exhausted all institutional mechanisms available to them in a direct, 

participatory and representative democracy. In a phrase, they had 

done all that could be expected of an active citizenry in a 

constitutional democracy. Khutsong — ‘the place of peace’ — will long 

be remembered for the violence and mayhem that have gripped the 

township since December 2005. In scenes very much reminiscent of 

the 1980s township revolts against the apartheid regime, television 

cameras captured daily images of young people blockading the streets 

with burning tyres and stones, preparing ‘molotov cocktails’, torching 

shops, libraries, schools, government buildings and local councillors’ 

homes. Between December 2005 and April 2006, violent protests 

caused damage to private and public property estimated at 

R70 million. 4 The community’s deep level of discontent is further 

reflected by the successful boycott of the municipal elections in 

March 2006. Of almost 30 000 registered voters, only 232 cast their 

ballots. 5 From December 2005 until the end of 2007, the township was 

rendered ‘ungovernable’. 

It took a very long time before the people of Khutsong decided to 

take their struggle to Court. It was not an easy decision. The leaders 

of the Khutsong struggle were convinced that their struggle could ‘be 

won in the street’: 6 

4 

Centre for Development and Enterprise ‘Voices of Anger’ 10 CDE Focus (April 

2007) 10. 

5 As above. 

6 

N Kolisile ‘The crying of Khutsong’ Amandla Publishers 1 October 2008 http:// 

www.amandlapublishers.co.za/content/view/71/32/ (accessed 2 November 

2008). Similarly, after the Court delivered its judgment, the Young Communist 

League noted in a press statement that even though it had not yet studied the 

Court’s decision, it was their ‘contention that matters related to forceful and 

illegitimate demarcations cannot be mitigated or resolved through the organs of 

class rule — the courts.’ See ‘YCL Statement on Khutsong’s ruling by the 

ConCourt’ 13 June 2008 http://groups.google.com/group/yclsa-press/browse_ 

thread/thread/67c0616a73a6f791 (accessed 20 June 2008). It is also important to 

note that — with the 2009 election firmly in their sight — the ruling party has 

recently decided that the Merafong Municipality will be incorporated back into 

Gauteng province. See ‘Khutsong to be returned to Gauteng’ The Mercury 22 

October 2008 http://www.iol.co.za index.php?set_id=1&click_id=13&art_id= 

vn20081022054258319C323057 (accessed 2 November 2008).


It would be very dangerous to put our faith in legal processes. Courts are 

seldom neutral and often tend to serve the interests of the ruling class. 

That is the reason we have delayed taking this matter to the 

Constitutional Court. For more than a year we have relied on our own 

mass strength and we still do. In the coming period we will utilise the socalled 

democratic space provided by the new constitution as well as 

working class resistance. Key strategies include intensifying mobilisation, 

building consciousness and continuing with mass struggle in the 

form of marches and rallies. This struggle will be won in the street! 

Khutsong — during this period of struggle — is both a reminder of the 

fragility of South Africa’s new political order and an indictment of its 

ability to respond effectively to those persons and communities in the 

greatest need. As one of the leaders of the Khutsong anti-demarcation 

movement wrote: ‘The struggle of the people of Greater Merafong is 

not just about forceful incorporation into North West, but about 

fighting for truth from government, people’s rights and democracy.’ 7 

Khutsong is not a singular, if continuous, event in a long but 

peaceful transition from a fascist apartheid state to a democratic 

constitutional order. Every week some poor community engages in 

illegal and often violent actions to back up demands for a more 

responsive government. Patrick Bond reports that ‘the pace of social 

protest intensified dramatically during the mid-2000s, soaring from 

5,813 recorded protests in 2004-05 to more than 10,000/year’. 8 Dale 

McKinley, one of the leaders of the Anti-Privatisation Forum, has 

argued that the increased violence employed by these community 

movements reflects the absence of ‘more inclusive and meaningful 

forms of direct and participatory democracy, that have little to do 

with the institutional forms of representation within bourgeois 

‘democratic’ society.’ 9 McKinley claims convincingly that ‘these 

movements have arisen out of the very failures and betrayals of the 

“main currents” and the institutional framework that gives them 

contemporary legitimacy.’ 10 

The biggest threat to the consolidation of South Africa’s 

constitutional culture is not the unprincipled and crass accusations 

7 Kolisile (note 6 above). 

8 P Bond ‘Rejoinder: Collaborations, co-optation & contestation in praxis-based 

knowledge production’ (2008) 35 Review of African Political Economy 271 272. 

9 D McKinley ‘Democracy and social movements in SA’ (2004) 28 Labour Bulletin 39 

40. 

10 McKinley (note 9 above) 40. For more in-depth analyses and case studies of these 

post-apartheid social movements and community organisations, see R Ballard et 

al (eds) Voices of protest: Social Movements in Post-Apartheid South Africa 

(2006); N Gibson (ed) Challenging hegemony: Social movements and the quest for 

a new humanism in South Africa (2006); D McKinley & P Naidoo (eds) 

Development Update 5 (2004); and T Madlingozi ‘Post-Apartheid social movements 

and the quest for the elusive “new” South Africa’ (2007) 34 Journal of Law 

and Society 77.


lodged by political elites within the African National Congress against 

South Africa’s ‘independent’ judiciary. The time bomb ticking away 

inside our polity is the disillusionment of the overwhelming majority 

of South Africans with the politics and the policies of our postapartheid 

democracy. Constitutional culture thrives when mobilised 

groups feel that they can appeal to constitutional norms and 

procedures in order to have their dispute resolved. It withers when 

such appeals go unheard. To be sure, the state’s response to these 

community movements has been at best to marginalise and vilify 

these movements by labelling them as ‘ultra-leftists’, ‘counterrevolutionaries’ 

organised by white ‘racists’, ‘enemies of the new 

democracy’; 11 and at worst to criminalise and to suppress their 

struggles. 12 

If one of the objectives of the judiciary in a constitutional 

democracy is to facilitate ‘constitutional dialogue’, has the Court 

done enough to facilitate a dialogue between itself and the citizenry? 

The answer is no. More pointedly, I would caution against the 

judiciary’s and the academy’s acceptance of Professor Michelman’s 

proposal that ‘the principle of interpretive charity’ will have a 

meaningful effect on closing the gap between the Constitutional 

Court and the people. 

Professor Michelman is indeed a giant. 13 His breadth of scholarship 

and his ongoing commitment to South Africa have made him one 

11 See, as an example, the response by ANC strategist Michael Sachs to the rise of 

these movements: M Sachs ‘“We don’t want the fucking vote”: Social movements 

and demagogues in South Africa’s young democracy’ (2004) 28 Labour Bulletin 23- 

12 

13 

27. 

For extensive reports on the criminalisation of poor people’s struggle as well as 

its violent repression, see D McKinley & A Veriava Arresting dissent: State 

repression and post-Apartheid social movements (2005) and M Memeza ‘A critical 

review of the implementation of the Regulation of Gatherings Act 205 of 1993 — A 

local government and civil society perspective’ (2006) available at 

www.fxi.org.za (accessed 14 June 2007). 

It is a singular honour for me to be afforded this opportunity to engage with 

Professor Frank Michelman’s excellent foreword to this first issue of the 

Constitutional Court Review (FI Michelman ‘On the uses of interpretive ‘charity’: 

Some notes on application, avoidance, equality, and objective unconstitutionality 

from the 2007 term of the Constitutional Court of South Africa’ (2008) 1 

Constitutional Court Review 1 (Michelman ‘Charity’)). Of those non-South African 

scholars who engage with the work of the South African Constitutional Court and 

South African Constitutional law theory in general, Professor Michelman is 

undoubtedly one of the most dedicated and influential. Indeed, as Professors 

Botha, Andre Van der Walt and Johan Van Der Walt have noted in a preface to a 

South African book dedicated to the scholarship of Prof Michelman, ‘Frank’s ideas 

have been — and continue to be — a source of inspiration for many of us who are 

interested in the possibility — and limits — of ‘transformative constitutionalism’. 

His writings on the judicial function and the capacity of rights discourse to 

facilitate and deepen democratic dialogue are among the most original and 

thoughtful on the topic — as are his reflections on constitutionalism as a tool for 

protecting a plurality of cultures, thoughtways and lifestyles; redressing 

inequality; and promoting social justice. It is therefore hardly surprising that 

South African legal academics and judges would have turned to his ideas in an


of our most important allies in re-thinking the meaning of a truly postapartheid 

constitutional democracy. It is therefore most fitting that 

he should write one of the forewords to the first issue of the 

Constitutional Court Review. My mandate here, however, is not to 

pay homage to Professor Michelman. It is, rather, to somehow engage 

with his response to the equally impressive commentary on the 

Court’s jurisprudence by Professor Woolman. 14 The two interlocutors 

need to be read together. 

Professor Michelman’s piece is a rejoinder to Professor Woolman’s 

critical analysis of what he describes as the Constitutional Court’s 

‘penchant for outcome-based decision-making and a concomitant lack 

of analytical rigour’. 15 This absence of analytical rigour — and its 

deleterious consequences — are exemplified by, though not limited 

to, the Court’s tendency, as Woolman sees it, to ‘refuse’ to engage in 

the direct application of the Bill of Rights: ‘The persistent refusal to 

give rights identifiable content, by avoiding direct application results 

in a Bill of Rights increasingly denuded of meaning.’ 16 Woolman 

further contends that this refusal amounts to a ‘paradigmatic 

violation of the rule of law’. This reticence makes it difficult, if not 

impossible 

for lower court judges, lawyers, government officials and citizens to 

discern, with some degree of certainty, how basic law is going to be 

applied, and to know, with some degree of certainty, that the basic law 

is going to be applied equally ... 17 

Michelman disagrees with Woolman’s conclusion that the thinly 

reasoned judgments in NM 18 and Masiya 19 show that the Court is in 

‘full flight from any meaningful engagement with Chapter 2 of the 

Constitution’ 20 and thus an abdication of the Court’s institutional 

13 attempt to break with the formalism and authoritarianism characterising our 

legal past, and to make sense of the democratic and egalitarian aspirations of the 

Constitution.’ H Botha et al (eds) Rights and democracy in a transformative 

constitution (2003) vii. In a paper presented in 2003 at a symposium dedicated to 

the scholarship of Professor Frank Michelman — Professor Levinson had this to say: 

‘There are many reasons to honour Frank Michelman, not least because there are 

so many different topics — central to the enterprise of thinking about law — about 

which he has been remarkably illuminating. He is truly a giant on whose shoulders 

many of us are proud to stand as we try to think our way through the dilemmas 

presented by trying to take the enterprise of American constitutionalism 

seriously’. S Levinson ‘Perpetual union, “free love”, and secession: On the limits 

to the “consent of the governed”’ (2003) 39 Tulsa Law Review 457 457. 

14 S Woolman ‘The amazing, vanishing Bill of Rights’ (2007) 124 South African Law 

Journal 762 (Woolman ‘Amazing’); S Woolman ‘Application’ in S Woolman et al 

(eds) Constitutional law of South Africa (2nd Edition, OS, 2005) ch 31. 



17 As above. 

18 NM v Smith 2007 5 SA 250 (CC) (NM). 

19 Masiya v Director of Public Prosecutions 2007 5 SA 30 (CC) (Masiya). 



responsibility. Michelman agrees entirely with Woolman’s contention 

that the majority opinions in NM and Masiya are ‘thinly reasoned’. He 

demurs from the thesis that these cases ‘have been wrongly or 

irresponsibly managed, as measured by reasonably discoverable, valid 

considerations of law and legal administration.’ 21 

The move Michelman then makes may be one of the most powerful 

rhetorical ploys in analytic philosophy: the use of the Donald 

Davidson’s principle of interpretive charity. That principle in short 

requires that if we are genuinely interested in understanding another 

person or culture, then we must count them ‘right’ in most of what 

they say and do. To refuse to do so would consign them to the 

madhouse. Michelman describes the principle of interpretive charity 

as ‘an approach to understanding a speaker’s statements by 

interpreting the ... statements to be rational and, in the case of any 

argument rendering the best, strongest possible interpretation of an 

argument.’ 22 The aim of this principle, Professor Michelman tells us 23 

is to learn. It is aggressively to learn what there is to learn from puzzles 

the interlocutors pose to us, by assuming there is method in their 

madness and doing our best to ferret that out, using everything else we 

know or can guess ... about where they are coming from. 

Michelman deploys the technique of ‘interpretive charity’ in order to 

fill in the gaps in the majority’s opinions in NM and Masiya. After an 

extensive and careful examination of the majority opinions in NM and 

Masiya, through a charitable lens, Professor Michelman concludes 

that the fact that the Constitutional Court seems to be gravitating 

towards its inherent power to develop common law in terms of section 

173 and 39(2) of the Constitution — as opposed to its judicial review 

power in terms of section 8 and 172(1) when undertaking a 

modification of common law rules under pressure from the Bill of 

Rights — does not mean that the court is uncomfortable with engaging 

with the substance of the Bill of Rights. 

Woolman sees the refusal to undertake the direct application of 

specific substantive provisions of the Bill of Rights as evidence of a 

court that does not respect its own precedents; as evidence of a court 

unaccountable to the community of constitutional interpreters and 

incapable of telling its citizens how to conform their behaviour to the 

dictates of the basic law; and as evidence that the Court — at the end 

of it all — is responsible for ‘the amazing, vanishing of the Bill of 

21 Michelman ‘Charity’ (n 13 above) 2. 

22 Wikipedia http://en.wikipedia.org/wiki/Principle_of_charity as cited by 

Michelman ‘Charity’ (note 13 above) 4. 

23 Michelman ‘Charity’ (n 13 above) 4.


Rights’. 24 Michelman sees the choice to go the route of indirect 

application made possible by section 39(2) — rather than direct 

application under section 8 — as no more than a matter of ‘doctrinal 

housekeeping’; 25 a ‘filing system’-question 26 over which nothing of 

substance turns. Woolman maintains that form is substance: Direct 

application ultimately leads to a proper elucidation of South African 

constitutional law. Michelman remains unmoved. Woolman’s 

construction — magisterial as it is — proceeds, as Michelman notes, 

from an ‘unrealistic high expectation of formal tidiness and tightness 

in constitutional drafting.’ 27 In Michelman’s opinion, the application 

conundrum — as brilliantly set out by Woolman — can not be resolved. 

Both authors agree that some surplusage is inevitable. 

2 Engendering ‘constitutional dialogue’ with the 

Commons 

I am not a constitutional lawyer. The main direction of my arguments 

can, without apology, be said to be political. However, what I will 

focus on still remains a bone of contention between Woolman and 

Michelman: the Court’s accountability and responsibility to ensure 

principled constitutional dialogue and its alleged failure to do so. I am 

interested in seeing who, beyond Woolman’s ‘chattering classes’, 28 

influences the Constitutional Court. I am interested in asking whether 

the Court truly engages in a dialogue with ordinary persons. I would 

argue — consistent with the positions staked out by Woolman — that 

one of the ways that the Court can make itself accountable to the 

public, and thus facilitate constitutional dialogue, is when its 

judgments are sufficiently reasoned, fully justified and widely 

published. ‘Thinly reasoned’ judgments that require the reader to 

enter into a highly sophisticated and technical reconstruction of a 

judgment — some would say a sangoma-like exercise — do not advance 

the cause of ensuring that ordinary people are part of the ‘principled’ 

— or at least transparent — constitutional dialogue to which the Court 

is ostensibly committed. I must therefore conclude that the principle 

of ‘interpretive charity’ has limited value in bridging the gap between 

the Court and the vast majority of its constituents. 

24 Woolman argues that the Court’s failure to get application jurisprudence right 

and that it ‘continues to err in such quite obvious ways reflects, at a minimum 

three possibilities: (1) the court simply ignores academic interventions; (2) the 

court believes the academic criticism to be incorrect; and (3) the court ignores 

obvious doctrinal errors because it holds itself largely unaccountable to the 

existing community of academic and professional interpreters’ (Woolman 

‘Amazing’ (n 14 above) 783). 


26 As above. 




How did I arrive at this conclusion regarding the limits of the 

principle of interpretive charity in contemporary South Africa? Again, 

not on the basis of a nuanced assessment of the highly technical 

debate between Michelman and Woolman. I have read the exchange 

between Woolman and Michelman to ultimately be about how the 

Constitutional Court should conduct itself in relation to various 

stakeholders. Woolman argues, convincingly, that ‘a principled 

judicial dialogue’ that the Court claims to stimulate, 29 is not possible 

if the Court, in his opinion, ‘refuses to say more than is necessary to 

decide a case on its facts.’ 30 Michelman demands that we try harder, 

and that if we take the principle of ‘interpretive charity’ seriously 

then we might just be able also to enter into a ‘principled dialogue’ 

with the Court. 

Woolman does not deny the gravitational pull of Michelman’s 

charitable approach. Instead, he contends ‘that a court that cannot 

be held accountable to its own precedent, to the text of the 

Constitution, or to the community of constitutional interpreters, and 

that grounds its decisions by mere genuflection in the direction of s 

39(2), more closely approximates a “kryptocracy” than a 

“logocracy”.’ 31 Whether we agree with this characterisation or not, 

all of us should be in agreement that trenchant and sustained critique 

of the Court’s jurisprudence from academics should be encouraged. 

What should we read into the Court’s sense of imperviousness to being 

influenced or at least being accountable to a community of 

‘constitutional interpreters’? If we agree that the court fails to create 

the conditions for a ‘principled dialogue’ between itself and what 

Professor Woolman refers to as ‘a cohort of educated Constitutional 

Court watchers’; 32 through its ‘murky, if not tendentious, lines of 

reasoning’ then what hope is there for the Commons, the masses of 

the people that are directly affected by its decisions? 33 

Writing in 1992, Donald Nicolson noted that although the notion of 

judicial independence is important in a democratic society, 

‘democracy is also threatened by an unaccountable judiciary.’ 34 I do 

not wish to enter into debates about the nature and the form that 

29 As called for by the Court in S v Mhlungu 1995 3 SA 867 (CC) para 129. 




33 Woolman is at pains to point out, however, that he is not simply — or even 

primarily — interested in what he rather glibly calls the chattering classes. 

Indeed, he argues in ‘Amazing’ (n 14 above) and elsewhere that a constitutional 

democracy cannot function as a constitutional democracy when no more than a 

few hundred people can read and decipher the meaning of the Court’s decisions. 

See also S Woolman & M Bishop ‘Law’s Autonomy’ in S Woolman & M Bishop (eds) 

34 

Constitutional conversations (2008) 1. 

D Nicolson ‘Ideology and the South African judicial process — Lessons from the 

past’ (1992) 8 South African Journal on Human Rights 50 70.


judicial accountability should take. Nor do I desire to revisit the 

arguments around how various techniques of judicial accountability 

can ensure a way out of the ‘counter-majoritarian’ dilemma. I am 

solely interested in how ‘public accountability’ can contribute to the 

process of constitutional dialogue between the Court and the 

Commons. It is also important to note that I use ‘accountability’ in the 

very narrow sense proposed by Louis Seidman: 35 

Sometimes we say that persons are accountable when they are required 

to give an account of themselves — that is, to give reasons or 

justifications for conduct and to demonstrate that such conduct is not 

the product of mere whim or caprice. 

Seidman argues that opinion writing exposes the judge’s thought 

process to those who read it, permits criticism of the justifications 

that the court offers and thereby allows judges to be held 

accountable. 

To state that the ‘culture of justification’ 36 that flows from the 

doctrine of the rule of law is just as applicable to the judiciary as it is 

to the executive and to the legislature, is to state the obvious. In 

cases where the Court’s decisions run counter to public opinion — 

assuming that such opinion can be accurately determined — it is 

vitally important that the Court fully justifies its decisions if it is not 

to lose public legitimacy. 37 Chief Justice Pius Langa has rightly 

observed that 

[T]he relationship which [the judiciary] has with rest of the community 

is ... important. It should be regarded as an integral part of the 

community it serves, and it can only function properly if it enjoys the 

complete trust and confidence of that community. 38 

Judging from various public statements and op-ed pieces written by 

members of the judiciary, it would seem that recent public criticism 

of the judiciary has caught members of the bench by surprise. The 

judges appear to have assumed that their public legitimacy was 

secure. However, one simply needs to look at the results of various 

empirical surveys to see that, for the majority of this country, the 

35 LM Seidman ‘Ambivalence and accountability’ (1988) 61 Southern California Law 

Review 1571 1574. 

36 

E Mureinik’s term. See E Mureinik ‘A bridge to where? Introducing the interim Bill 

of Rights’ 1994 (10) South African Journal on Human Rights 31. 

37 See T Roux ‘Principle and pragmatism on the Constitutional Court of South Africa’ 

(2009) 7 International Journal of Constitutional Law 106. See also E van 

Huyssteen ‘The South African Constitutional Court and the death penalty: Whose 

values’ (1996) 24 International Journal of Sociology of Law 291. 

38 

See T Mbeki ANC Today 5(23) http://www.anc.org.za/ancdocs/anctoday/2005/ 

text/at23.txt (accessed 22 October 2008).


Constitutional Court does not enjoy high public support. 39 Indeed, 

recent comments by Anthony Butler are spot on: ‘The idea that the 

courts are suddenly under siege is a fantasy. Their legitimacy has long 

been fragile ...’. 40 

Amongst community movements, activists and ordinary people — 

as opposed to educated Constitutional Court watchers and members 

of the chattering class — the judiciary has not been able to shake off 

the impression that it is an instrument of the elites. 41 Take this 

exchange between the Treatment Action Campaign’s Mark Heywood 

and another activist. Mark Heywood — an unabashedly activist 

Constitutional Court watcher — wrote 42 

The Constitution is at the heart of our democracy. It prescribes 

accountable and ethical government, openness and a culture of justification 

by politicians for their actions. But more than this, it should be 

the touchstone for all public policy, an instrument that can be used to 

insist on public policy that places equality and social improvement at the 

heart of all government action. 

Now, to a ‘sober-minded’ and indeed ‘educated’ Constitutional 

Court watcher, this plea might seem quite reasonable. But that was 

not how it was read by one Commons activist: 43 

This near spiritual, nay, pathological worship of the handwork of neoliberals 

is how the masses and their struggle for full independence, selfreliance 

and liberation got substituted for ‘the constitution’! All political 

power and authority somehow got squeezed into the Constitution, and 

what all need do is learn it, and enforce its provisions, and instantly, all 

will be solved. What is concealed is that the real social and economic 

power in this marginal act of giving themselves a liberal constitution 

automatically moved, and conformed, into the capitalist class. And true, 

to the extent reasonably possible, the people’s real problems will be 

attended to! 

What is also never frequently spoken about are ... the massive racial 

fears of the white supremacists, who, upon being born again as capitalist 

class rights fighters through a liberal constitution, abandoned their 

‘minority racial rights’, and fought to have the constitution invested 

with ‘supreme power’ over all else. 

39 On such public opinion surveys, see Roux (n 37 above) 107 n 4. (Roux reports that 

27.9 per cent ‘attentive’ public support for the CCSA in 1997, ie among citizens 

who had heard about the Court, and ‘34 per cent public support for the CCSA in 

40 

2004’). 

A Butler ‘Emotional responses overlook good points Mantashe makes’ Business Day 

41 

14 July 2008. 

See T Madlingozi ‘The limits of transformative constitutionalism in the struggle 

42 

against poverty and marginalisation’ (2009, forthcoming) copy on file with author. 

M Heywood ‘Political climate dangerous for poor most of all’ Business Day 28 July 

43 

2008. 

A Banda Debate Listserve (28 July 2008) debate-bounces@debate.kabissa.org 

(accessed 28 July 2008).


The point here is that criticism of the judiciary does not only come 

from those who engage in scurrilous attacks on the judiciary in order 

to advance their narrow political agenda. Criticism of the judiciary 

also comes from those who are genuinely concerned about the gap 

between constitutional promises and the lived, impoverished, reality 

of the majority of South Africans. 

The common reaction by judges and legal analysts to criticism of 

certain court judgments is to argue that the people do not understand 

the law or that they have not acquainted themselves sufficiently with 

the facts of the case. The former President of the Supreme Court of 

Appeal, Justice CT Howie has argued that not enough trouble has been 

taken ‘to tell our public, and our media, something of what judicial 

independence is all about’ and that public criticism is often made in 

the absence of getting to know the ‘vital facts’. 44 But as a member of 

the Commons, I must ask: Whose responsibility is it to make sure that 

the public not only knows the ‘vital facts’ but also understands the 

working of the courts? Further, and more significantly, what if the 

public tries to acquaint itself with the reasons for a particular 

decision, only to be confronted with a ‘thinly reasoned’ judgment? 

How reasonable is it to expect members of the public to adopt the 

principle of ‘interpretive charity’ and engage in the laborious exercise 

of filling in meanings and passages in the courts’ decisions as 

Professor Michelman urges us to do? 

3 Conclusion 

During the Court’s 2007 term (and well into its 2008 docket), 

incidences referred to as ‘threats to the judiciary’ reached a fever 

pitch. In the weeks and months preceding the appearance of the then 

deputy president of the African National Congress (ANC) in court to 

face charges of corruption, robust criticism, emanating largely from 

members of the tripartite alliance — the ANC, the South African 

Communist Party and the Congress of South African Trade Unions — 

became common place. The slogan ‘We will kill for Zuma’ if the Court 

did not find in his favour became a disturbingly regular part of public 

discourse in some quarters. 

More than ever, the Court needs to court public support in order 

to protect itself against escalating ‘attacks on the judiciary’ 

emanating from various political quarters. Theunis Roux has shown 

that in the past the Court has not needed public support in order to 

safeguard its institutional security because the ‘ANC political elite has 

shielded the Court from the political repercussions of its most 

44 

Justice CT Howie ‘Judicial independence’ (2003) 120 South African Law Journal 

679 679 & 682 (emphasis added).


unpopular decision ...’. 45 In light of recent events this protection can 

no longer be taken for granted. Indeed, these events dictate that the 

Court change its strategy and work to consolidate its public 

legitimacy. That is no small task. The Court has not yet engaged in any 

meaningful outreach programs. 46 

The Constitutional Court might be based near Hillbrow. It might 

be located in a non-intimidating, inviting building and it might be 

geographically accessible. The judges might be very friendly and well 

meaning people. But the members of the Court continue to operate 

and to move within the circles of ‘educated Constitutional Court 

watchers’. I have seen and enjoyed the presence of Constitutional 

Court judges at various academic conferences. But I am yet to see any 

of them at various workshops and rallies hosted by social movements. 

I have seen members of the Court give frank and engaging public 

lectures at various universities. But I am yet to see any of them 

address members of social movements. 47 I have heard members of the 

Constitutional Court give interviews on English radio stations. I have 

never heard any of them offer an interview on an African language 

radio station. 

An ‘educated Constitutional Court watcher’ might tell me that the 

Court does not need to undertake such a public role. After all, their 

communications are a matter of public record. But where are the 

Sesotho or TshiVenda translations of the Court’s judgments? 48 And 

even if they did appear on the Court’s website, I fail to see how that 

would make any difference at all to the majority of South Africans 

who remain quaintly removed from the internet and other forms of 

new media. 

As we launch this Review, we need to pause and think of ways that 

ordinary people, the Commons, can become part of the ‘principled 

dialogue’ to which the Court and academics such as Woolman and 

Michelman have committed themselves. Finally, as South Africa 

enters what has been aptly described as the ‘second transition’ — 

characterised by, amongst other things, robust engagement with 

45 Roux (note 37 above) 138. 

46 In its annual strategic/business plan, under a section entitled ‘Outreach 

programs’ the Director of the Constitutional Court only manages to mention 

initiatives such as the law researchers program and the setting up of the Court’s 

choir and soccer team. See ‘Annual Strategic/Business Plan 2008/2009’ http:// 

www.constitutionalcourt.org.za/site/Admin/ANNUAL%20STRATEGIC%20BUSINESS% 

20PLAN%202008-2009.pdf (accessed 21 October 2008). 

47 Justice Dennis Davis is a notable exception in this regard. Justice Davis regularly 

gives key note addresses to social movement conferences and other events. 

48 During his interview for position of the Deputy Justice, Justice Dikgang Moseneke 

noted the practical and cost implications in handing down judgments in other 

official languages. See http://www.constitutionalcourt.org.za/site/judges/ 

transcripts/dikgangmoseneke1.html (accessed 1 November 2008).


institutions supporting democracy and the settling of major political 

disputes in courts 49 — the Court’s opinions will more than ever before 

play a significant role in public discourse and democratic politics. It 

is, therefore, vitally important that the Court holds itself 

accountable, not only to ‘educated constitutional court watchers’, 

but also to the Commons — some 46 million ordinary South Africans. 

Fully reasoned and sufficiently justified opinions that do not require 

the public to engage in highly complex exercises in ‘interpretive 

charity’ would be a first step in that direction. Maybe, just maybe, 

the events of 20 August 2007 might not be repeated in the future. 

49 See S Ndlangisa ‘Political ball is in our courts’ City Press 24 August 2008 24.

NORMATIVE PLURALISM AND ANARCHY: 

REFLECTIONS ON THE 2007 TERM 

77 

AJ van der Walt * 

1 The decline of foundation and the allure of 

anarchy 

Waren die klaren Bedeutungen aus dem Haus, tanzte die Sprache auf 

dem Tisch. 1 

As a schoolboy, Robin Davies 2 was intrigued to hear that, according to 

the likes of DH Lawrence, if one really wanted something one ‘should 

have it, ought to have it, must do it or take it if you can’. However, 

Robin soon learnt that social norms often prevent you from taking 

what you want and they cause all kinds of trouble when you do. ‘You 

can have [me and the children], or you can have everyone else. Not 

both’, his wife Nancy told him, years later, laying down the law about 

his philandering. 3 ‘You can’t do both’, his father ruled when, aged 15 

or so, Robin wanted to visit a friend on an afternoon when his parents 

had already arranged something for that evening. 

* B Jur et Art Honns (BA) LLB LLD (Potchefstroom) LLM (Witwatersrand), South 

African Research Chair in Property Law, Faculty of Law, Stellenbosch University 

(hosted by Stellenbosch University, funded by the Department of Science and 

Technology and administered by the National Research Foundation). Thanks to 

Gustav Muller, Elmien du Plessis, Mikhalien Kellermann and Janke Strydom for 

research assistance, and to Karl Klare, the participants in the CCR workshop and 

Geo Quinot, Philip Sutherland, Henk Botha, Irma Kroeze and Wessel le Roux for 

discussions and comments, to Philip Sutherland for access to a draft of ‘Insuring 

contractual fairness in consumer contracts’ (Stellenbosch Law Review 

1 

forthcoming), and to Halton Cheadle for an instructive exchange on application in 

the context of labour disputes. Remaining shortcomings are my own. My apologies 

to the reader for indulging a 30-year obsession. 

T Dorn ‘Ultima ratio’ in D Schwanitz et al Amoklauf im Audimax: Stories (1998) 7- 

2 

3 

46 18. 

In Kingsley Amis’s novel You can’t do both (2004). 

Patrick Standish received a similar ultimatum in Difficulties with girls (1988), as 

did Amis in real life.

78 Normative pluralism and anarchy: Reflections on the 2007 term 

Kingsley Amis, ‘the most clubbable of men’ and a womaniser of 

note, 4 was frustrated by restrictions on his freedom to have things his 

way, but he was familiar with the power of bourgeois morality — 

‘taking it if you can’ wrought havoc when he decided that he should 

have what one of his characters called ‘marriage, 1960s style’, openly 

taking his mistress on holiday. 5 Like his characters, Amis found that 

he could have the wife and the children or the girlfriend; not both. 

Several Amis characters given to ‘take it if they can’ are forced to 

abandon some of what they wanted or lose everything, including their 

dignity. 6 It is easy to scoff at silly ‘rules’ about not having two social 

engagements on the same day, but not all social norms are so easily 

dismissed, even if they appear arbitrary. 

The framing discourse for these Amis novels (and my purpose in 

recalling them) is not about accepting or rejecting social norms — 

despite his scorn for conventional morality, Amis did not subscribe to 

moral anarchy 7 (and in the 2007 term the Constitutional Court said 

little about morality). 8 Characters in the novels who made a life out 

of taking it if they could were not very nice persons and, despite 

flashes of defiance, Amis was not justifying their behaviour. 9 In fact, 

Amis was commenting on what could be described — not in his words 

— as the complexity of human relationships in a time when social 

fundamentalism was being replaced by pluralism 10 and, consequently, 

relativism and uncertainty. World War II liberated men and 

women from their traditional sexual and gendered social roles and 

4 On the autobiographical nature of You can’t do both see Z Leader The life of 

Kingsley Amis (2006) 3 7-9. Martin Amis confirmed that finding parallels between 

his father’s life and novels is not ‘making the elementary mistake of conflating 

the man with the work’, since ‘all writers know that the truth is in the fiction’: 

Experience (2000) 28. 

5 Jenny, in Difficulties with girls (1988); the girlfriend was Elizabeth Jane Howard, 

Amis’s second wife. 

6 ‘Better be a bastard than a fool’, Roger Micheldene told himself, without much 

conviction, towards the end of One fat Englishman (1963). 

7 

In the Memoirs (1991) Amis remembered that his father put his, Kingsley’s, 

shortcomings down to his complete lack of religion. ‘And’, he says, ‘I should not 

be truly his son if I had never felt that he had something there.’ Amis concluded 

the Memoirs with a poem, entitled ‘Instead of an Epilogue’ and dedicated ‘To H’. 

In the final verse Amis lists the admirable characteristics of ‘someone’ he met 

when twenty-four, then proceeds: ‘Well, that was much as women were meant to 

be, / I thought, and set about looking further. / How can we tell, with nothing to 

compare?’ 

8 But see Schaik v S 2008 2 SA 208 (CC) par 76. 

9 

Leader (n 4 above) 468: ‘Roger [Micheldene], like Patrick [Standish], knows how 

horrible he is.’ 

10 Like J Dewar ‘The normal chaos of family law’ (1998) 61 Modern Law Review 467- 

485 I use the term to indicate a diversity of conflicting norms. See LL Mofokeng 

‘The right to freedom of religion: An apparently misunderstood aspect of legal 

diversity in South Africa’ (2007) 11 Law, Democracy & Development 121-131; S 

Louw ‘Mysterium inequitatis: Truth, elections, autonomy in the Southern African 

politic’ (2006) 33 Politikon 221–238 (relying on C Lefort’s notion of ‘the political’ 

to argue that democracy relies on diversity and pluralism). See further M Koenig 

& P de Guchteneire (eds) Democracy and human rights in multicultural societies 

(2007).


from the remnants of Victorian and Christian morality, opening up 

room for experimentation and free choice, 11 but liberation coincided 

with uncertainty and disruption. 12 Freedom from tradition comes at a 

price: once we give up the certainty of transcendent truth, choices 

become contingent and dangerous. Abandoning the certainty of 

tradition created a normative vacuum and raised the spectre of moral 

anarchy and turpitude. As Amis framed it much later, reflecting on 

bad choices he had made: ‘how can we tell, with nothing to 

compare?’ 13 The problem with postwar pluralism was that, once the 

certainty of immutable rules is abandoned, there is nothing to 

compare. When tradition is traded in for pluralism, everything is open 

for questioning, which could create the impression that everything is 

up for grabs; that everything goes. In an effort to avoid anarchy, the 

temptation is to fall back on the formalism of empty rules. 

Amis’s depiction of postwar English society illustrates the slippery 

slope in post-traumatic society from normative pluralism to anarchy 

or, more accurately, to a facile choice between anarchy and 

reactionary formalism. The modernist rejection of foundation and 

acknowledgement of normative pluralism was inevitable, but the 

resulting uncertainty proved harder than expected. A similar process 

took place in law: ever since Realist and Critical Legal Studies scholars 

destroyed the theoretical credibility of fundamentalism and pointed 

out the indeterminacy and contingency implied by pluralism, lawyers 

— who usually deny that normative choice plays any role in law — had 

to account for their normative choices. In the face of nihilism and 

anarchy a reversion to formalism seems like a small price to pay for 

clarity and certainty. 

11 Social liberalisation after World War II can be traced to the aftermath of World 

War I, just like the Critical Legal Studies Movement continued a trend initiated by 

the Legal Realists. M Tushnet ‘Critical legal studies: An introduction to its origins 

and underpinnings’ (1986) 36 Journal of Legal Education 505-517 argues that the 

question whether Critical Legal Studies continued Legal Realism is uninteresting; I 

take continuation between the two world wars for granted. During the wars, 

women were forced to take men’s places in factories, liberating them from their 

home-maker role. Similarly, people fighting in the wars realised that the 

subservient social roles they were used to before the war (because of age, social 

status or race) no longer suited them. In both cases, liberation from tradition 

undermined the authority of social hierarchy. 

12 The street protests of 1968 that shocked mainstream society from Prague to Paris 

and from Berkeley to Berlin originated in the same context: the civil rights 

movement, women’s lib, the hippie movement, anti-war protests and student 

uprisings during the 1960s were founded upon rejection of traditional authority 

13 

and mainstream morality. 

See n 7 above. See IJ Kroeze ‘When worlds collide: An essay on morality’ (2007) 

22 SA Public Law 323-335, especially 331-332 333-335.


John Dewar 14 argues that normative pluralism, which he describes 

as a plurality of rules and values that are chaotic, incoherent and even 

antinomic in that the contradictions between them cannot be 

resolved in the long run, is normal in so far as contradictions at the 

level of rules are held at bay by practices not completely determined 

by text. In Dewar’s view, normative pluralism changes into anarchy 

when it combines with uncertainty about purposes; as soon as 

pluralism extends beyond the level of rules, normative anarchy 

ensues. Dewar identifies normative anarchy in situations where rights 

are mixed with discretion in a way that defies balancing and where 

regulatory legislation moves from rules to utility and back without 

suitable systemic adaptations. 

In a way, the early Amis novels described just such a shift from 

rules to utility and back — like Robin Davies, Amis and thousands of 

others rebelled against the inflexible and arbitrary post-war 

dictatorship of surviving pre-war morality and craved the greater 

freedom and flexibility they experienced during the war. Of course, 

with flexibility comes the burden of discretion but, according to 

Dewar and Amis, abandoning foundation does not necessarily end in 

nihilism: normative anarchy can be avoided as long as contradictions 

are restricted by common purpose. In the less utilitarian language of 

Joseph William Singer, the Critical Legal Studies movement ‘brought 

nihilism to centre stage’ by denying the possibility of determinacy, 

objectivity and neutrality in legal reasoning, but the absence of 

foundation ‘does not condemn us to indifference or arbitrariness, nor 

make it ridiculous to ask, or impossible to answer, the question of 

what we should do or how we should live’. 15 In Singer’s view legal 

theory, as a form of political activity, opens everything up for 

questioning without putting everything up for grabs — critical legal 

theory acknowledges moral and political goals that give direction to 

‘what we must do and how we should live’. 16 

Of course, on one level Dewar and Singer are vulnerable to 

accusations of fundamentalism — how does shared purpose or political 

goal that saves pluralism from anarchy differ from Victorian or 

Christian (or legal Formalist) sources of foundation? 17 How is any 

14 n 10 above, 468-473. See further J Dewar ‘Reducing discretion in family law’ 

(1997) 11 Australian Journal of Family Law 309-326; S Parker ‘Rights and utility in 

Anglo-Australian family law’ (1992) 55 Modern Law Review 311-330; M Maclean 

Making law for families (2000). 

15 

‘The player and the cards: Nihilism and legal theory’ (1984) 94 Yale Law Journal 

1-70 8: asking ‘what shall we do and how shall we live?’ opens up dissent rather 

than consensus. See Tushnet (n 11 above) 509; A Hunt ‘The theory of Critical 

Legal Studies’ (1986) 6 Oxford Journal of Legal Studies 1-45. 

16 Singer (n 15 above) 66-70 identifies cruelty, misery, hierarchy and loneliness as 

the greatest problems that legal theory faces. 

17 

On foundation in law see P Fitzpatrick Modernism and the grounds of law (2001) 

1-7.


political purpose less arbitrary than Mr Davies senior’s decree that 

Robin cannot ‘do both’? Once the foundation of traditional rules is 

rejected, how do we justify the new rules we place in their stead to 

avoid anarchy? 

Developing a satisfactory answer to this challenge, if it can be 

done at all, will take my analysis too far off course, but a starting 

point is to accept the argument of Johan van der Walt and Henk Botha 

that ‘the counter-majoritarian problem and the problem of significant 

social dissent [both sources of normative pluralism in constitutional 

discourse] are not technical hiccups’ but aporias that ‘open up the 

possibility of social deliberation that would exceed technical 

procedure’ and technical ethics to ‘give politics a chance’. 18 If there 

were a purpose that could avoid both anarchy and fundamentalism in 

the face of normative pluralism, it could be nothing else than the 

possibility of open social deliberation and dissent that defies 

foundation and closure to give politics a chance. 19 In the ‘community 

without community’ that Van der Walt and Botha have in mind, a 

society of diverse communities, cultures and conflicting normative 

values, without any justificatory or unifying common will, social 

deliberation and politics can be possible without anarchy. 20 

In the ‘community without community’, recognition of the 

sacrificial nature of decision reveals the essence of law as its very 

failure to resolve social conflict with reference to a common will or 

purpose. In other words, politics is not to be found in a foundational, 

unifying common will that overcomes or mediates plurality; instead, 

politics is possible only in the presence of unmediated and unresolved 

diversity, plurality and dissent. Consequently, a purely dialogic or 

republican politics 21 cannot solve conflict or bring closure by 

18 J van der Walt & H Botha ‘Democracy and rights in South Africa: Beyond a 

constitutional culture of justification’ (2000) 7 Constellations 341-362 350. See H 

Botha ‘Democracy and rights: Constitutional interpretation in a postrealist world’ 

(2000) 63 Tydskrif vir Hedendaagse Romeins Hollandse Reg 561-581 572-574 581; 

K van Marle ‘Lives of action, thinking and revolt — A feminist call for politics and 

becoming in post-apartheid South Africa’ (2004) 19 SA Public Law 605-628 621- 

624; F Michelman ‘Property as a constitutional right’ (1981) 38 Washington & Lee 

Law Review 1097-1114 1110: ‘it is the possibility of partial resolutions that allows 

us to experience the contradiction [between democracy and rights] as generative 

tension rather than as dead end.’ 

19 Van der Walt & Botha (n 18 above) 350. See Van Marle (n 18 above) 621, citing B 

Honig (ed) Feminist interpretations of Hannah Arendt (1995) 135. See further W 

le Roux ‘Bridges, clearings and labyrinths: The architectural framing of postapartheid 

constitutionalism’ (2004) 19 SA Public Law 629-665. 

20 

n 18 above, 350-352. 

21 See le Roux (n 19 above) 646-660 on constitutionalism as a dialogic ‘clearing’.


mediating dissent or pluralism; 22 nor can politics be guaranteed by 

constitutionalism, rights or law. 23 Robert Cover pointed out that an 

important function of law is to ‘kill’ or suppress plurality, dissent and 

meaning; not to open it up. 24 However, to the extent that law is 

inevitable, law, rights and constitutionalism are justifiable in so far as 

it gives politics a chance. A major issue in this article is to ask whether 

the interpretive and adjudicative approach that the Constitutional 

Court has apparently elected to follow in deciding cases where the 

common law tradition and legislation have to be measured against 

constitutional directives is justifiable in the sense that it gives politics 

a chance, instead of just closing up gaps and entrenching certainty. 

2 Pluralism in post-apartheid constitutional 

theory 

Our Constitution does not sanction a state of normative anarchy which 

may arise where potentially conflicting principles are juxtaposed. 25 

South Africans in the post-1994 era understand the anxiety caused by 

abandoning normative fundamentalism. Having faced up to the 

wickedness of apartheid and the bankruptcy of its underlying values, 

the point of constitutional transformation was to replace the 

exclusivism of segregation with an inclusive democracy that 

recognises diverse communities, cultures and traditions as well as 

diverse social, cultural and religious values. The 1996 Constitution’s 

commitment to diversity is apparent from its emphasis on human 

dignity, equality, and freedom 26 and its recognition of the values of 

different social, cultural and religious groups, 27 but recognition of 

diversity as a positive value surpasses mere tolerance — social, 

cultural and religious pluralism (‘more-than-one’) creates and 

supports the very possibility of politics and democracy. 

22 

Van der Walt & Botha (n 18 above) 352 argue that judicial decisions do not solve 

the conflicts that arise from social complexity but suppress them at the cost of 

one party for the sake of a social goal. R Cover ‘Violence and the word’ (1986) 95 

Yale Law Journal 1601-1629 1607: ‘the relationship between legal interpretation 

and the infliction of pain remains operative even in the most routine of legal 

acts’. 

23 

Van Marle (n 18 above) 625-628 warns against ‘constitutional optimism’ based on 

the pervasiveness of law, constitution and rights, referring to E Christodoulidis 

‘Constitutional irresolution: Law and the framing of civil society’ (2003) 9 

European Law Journal 401-432. Compare le Roux (n 19 above) 660-665. 

24 R Cover ‘The Supreme Court 1982 term. Foreword: Nomos and narrative’ (1983) 

97 Harvard Law Review 4-68 41-44. 

25 

Ngcobo J, Fuel Retailers Association of Southern Africa v Director-General: 

Environmental Management, Department of Agriculture, Conservation and 

Environment, Mpumalanga Province, 2007 6 SA 4 (CC) par 93. 

26 

Compare s 1 of the 1996 Constitution. 

27 See ss 9(3) and 9(4), 15, 30, 31; compare ss 39(3), 211-212.


Since 1994 the courts had many opportunities to confirm the 

Constitution’s commitment to accommodation of diversity and 

pluralism. 28 Such an opportunity again arose in 2007 in Pillay, 29 the 

first equality case to be considered by the Constitutional Court in 

terms of the Promotion of Equality and Prevention of Unfair 

Discrimination Act 4 of 2000 (PEPUDA). 30 The Court reiterated the 

confirmation, in the Act, of ‘the Constitution’s commitment to affirm 

diversity’ 31 and decided that reasonable accommodation in matters 

of religion means that 

sometimes the community, whether it is the State, an employer or a 

school, must take positive measures and possibly incur additional 

hardship or expense in order to allow all people to participate and enjoy 

all their rights equally. It ensures that we do not relegate people to the 

margins of society because they do not or cannot conform to certain 

social norms. 32 

Accordingly, a school disciplinary code that inflexibly denied an 

exemption from normal dress code for pupils who wish to wear a nose 

stud as part of their religious or cultural practice discriminated 

against them unfairly, in conflict with the Act. 

Pillay demonstrates that conflict about normative pluralism in the 

context of diversity requires more opportunities for politics and 

difference, not greater consensus. There is no reason why we should 

agree about the wearing of nose studs, headscarves, yarmulkes or 

dreadlocks or why one view about any of these cultural and religious 

symbols should be enforced on everybody, but open deliberation and 

dissent about the underlying social customs and values and about 

tolerance is good for the democratic order. However, difference is 

not tolerated when it contradicts central values of the Constitution or 

undermines democracy 33 — normative anarchy is not permitted. What 

prevents pluralism from sliding into anarchy in the context of diversity 

28 Mostly relating to race, gender, religion and sexual orientation: Amod v 

Multilateral Motor Vehicle Accidents Fund 1999 4 1319 (SCA) (Islamic marriage); S 

v Lawrence; S v Negal; S v Solberg 1997 4 SA 1176 (CC) (Christian public holidays); 

National Coalition for Gay and Lesbian Equality v Minister of Justice 1999 1 SA 6 

(CC) (sexual orientation); Du Toit v Minister for Welfare and Population 

Development 2003 2 SA 198 (CC) (married status); Satchwell v President of 

Republic of South Africa 2002 6 SA 1 (CC) (sexual orientation); J v Director- 

General, Department of Home Affairs 2003 5 SA 621 (CC) (sexual orientation); 

Gory v Kolver NO 2007 4 SA 97 (CC) (sexual orientation). 

29 MEC for Education: KwaZulu-Natal v Pillay 2008 1 SA 474 (CC). 

30 n 29 above, par 39. See I Currie & J de Waal The bill of rights handbook 5 ed 

(2005) 267. 

31 Pillay (n 29 above) par 65. 

32 Pillay (n 29 above) par 73. 

33 

Courts refused to enforce diversity in Bhe v Magistrate, Khayelitsha 2005 1 SA 580 

(CC) (customary law rules of succession); Prince v President, Cape Law Society & 

others 2001 2 SA 388 (CC) (smoking of cannabis as part of religious observance); 

Christian Education in South Africa v Minister of Education 1999 2 SA 83 (CC) 

(corporal punishment as part of Christian education).


as a positive value is exactly what gave Amis and formalist opponents 

of critical legal theory headaches, namely recognising that values are 

contingent and that everything is open for dissent and discussion, 

without implying that everything goes. 

The new constitutional order brought about a second kind of 

normative pluralism, where contradiction results not from different 

values but from the project of constitutional transformation itself. In 

this arena, the issue for debate is the direction, scope and pace of 

constitutional reform and, particularly, the effect of the Constitution 

on existing law. The effect of the Constitution on existing law is a 

logical locus for deliberation (politics) about social and legal reform, 

but the notion that a significant part of existing law — the common 

law — can survive or must be insulated against some or all direct or 

indirect constitutional influence contradicts constitutional values and 

threatens constitutional politics. In a word, the preservative impulse 

in the application debate misrepresents the democratic principle that 

everything is open for debate and threatens to push the normative 

pluralism of transformative politics into the anarchy of fractional 

conflict. 

An important premise of my argument is that the Constitution’s 

commitment to diversity as a positive value is premised upon the 

reduced authority of the common law. 34 The Constitution is founded 

on a set of values and normative commitments related to democracy, 

human dignity, equality and freedom and on the replacement of an 

immoral and discredited social and political system by a new, 

constitutional and democratic system. The shift from exclusivity to 

inclusivity that it presupposes is possible only if the existing legal 

order can be both purified of its exclusivist and authoritarian legacy 

and adapted to accommodate the normative diversity of its reformist 

purposes. In a largely uncodified legal system that relies on judicial 

finding, interpretation and application of old authorities and case 

law, remnants of the old legal system (characterised by exclusivity 

and authoritarianism) will survive political change and the 

establishment of the new constitutional order, for the time being, 

34 

See s 39(3), read with ss 8(3), 39(3) and 173 of the 1996 Constitution, recognising 

the continued existence of the common law, to the extent that it is consistent 

with the Bill of Rights and subject to the power of the higher courts to develop it 

to promote the spirit, purport and objects of the Bill of Rights. Johan van der 

Walt has argued that (direct or indirect) horizontal application of the Constitution 

would promote plurality: JWG van der Walt ‘Perspectives on horizontal 

application: Du Plessis v De Klerk revisited’ (1997) 12 SA Public Law 1-31; 

‘Progressive indirect horizontal application of the bill of rights: Towards a cooperative 

relation between common-law and constitutional jurisprudence’ (2001) 

17 South African Journal on Human Rights 341-363; Van der Walt & Botha (n 18 

above) 350-352.


dragging unwanted survivals from the past behind us. 35 Every 

legislative, executive and judicial decision embodies a choice 

between upholding remnants of the old order and enforcing change 

brought about by the new; every decision to uphold the existing law 

implies a sacrifice of constitutional reform, even when it is 

indubitably correct or unavoidable. Even more importantly, every 

decision in favour of stability and certainty or vested and acquired 

rights inevitably comes at the price of suppressing plurality, dissent 

and change. 

To complicate matters the role of the Constitution is ambiguous. 

On the one hand, any Constitution (with a Bill of Rights) is part of the 

legal machinery that ensures stability and the protection of rights, 

particularly against state interference. In this perspective the 

Constitution is an agent for the protection of established privilege and 

power. On the other hand, though, the South African Constitution is 

clearly also an agent for change, including large-scale social, 

economic and political change that can affect established privilege 

and power negatively. It could be said that the danger of anarchy is 

reduced because the Constitution embodies a direction-giving 

purpose: different value systems are recognised, but we are all 

working towards an open society built on democracy, social justice 

and fundamental rights to human dignity, equality and freedom. Even 

in the face of diversity and pluralism, the Constitution directs us 

towards certain choices and away from others. Frank Michelman 36 

argues that the constitutional supremacy principle in s 1 of the 

Constitution, as interpreted in Pharmaceutical Manufacturers, 37 

establishes 

the value of legal-systemic harmony in the service of the vision of the 

good society staked out by the entire list of founding values set forth in 

FC s 1 and instinct in the rest of the Final Constitution. We deal here 

with the value of the unity of the legal system — meaning the system’s 

35 In the lyrics of Roger Waters: ‘they flutter behind you, your possible pasts / some 

bright eyed and crazy, some frightened and lost / a warning for anyone still in 

command / of their possible future to take care’: Pink Floyd ‘Your possible pasts’, 

from The final cut, subtitled A requiem for the post war dream (1983 © Pink 

36 

37 

Floyd Music Publishers Ltd). 

FI Michelman ‘The rule of law, legality and the supremacy of the Constitution’ in 

S Woolman et al (eds) Constitutional law of South Africa (2nd Edition, OS, 2005) 

34-44. 

Ex Parte President of the Republic of South Africa: In re Pharmaceutical 

Manufacturers Association of South Africa 2000 2 SA 674 (CC) par 44.


normative unity or, as one might say more poetically, its visionary 

unity. 38 

The notion of such a normative unity in the goal of constitutional 

transformation is certainly attractive. For administrative justice, 

Michelman’s analysis of the rule of law, legality and constitutional 

supremacy provisions provides a starting point based on the normative 

unity of the Constitution. Similar arguments can be developed for 

equality and non-discrimination, where transformation also amounts 

to wholesale replacing of apartheid law with a fundamentally new set 

of rules, indicating normative unity in reform. However, some 

constitutional provisions look less like wholesale, unidirectional 

transformations and more like contradictory products of difficult and 

inchoate compromises. In the compromise provisions, Michelman’s 

‘visionary unity’ of the Constitution is less obvious or more contested 

than in abolition and replacement provisions like s 9. 

A striking example is the sections relating to the national flag and 

the national anthem. The national flag described in s 5 (and Schedule 

1) is neither based on the apartheid flag, itself the patchwork product 

of South Africa’s colonial and postcolonial past, nor on the ANC flag, 

nor is it a patchwork amalgam of the two. The Constitution simply 

replaced the apartheid flag with a new flag, signalling the clean break 

style of transformation that illustrates Michelman’s direction-giving 

‘visionary unity’ of the Constitution. However, the national anthem 

illustrates a different approach: under apartheid, Die Stem van Suid- 

Afrika (The Call of South Africa) was the official national anthem, 

although Nkosi Sikelel' iAfrika was favoured by political activists and 

refugees. Under the 1993 Constitution the Republic had two national 

anthems: Nkosi Sikelel' iAfrika and Die Stem van Suid-Afrika / The 

Call of South Africa; 39 under the 1996 Constitution the single national 

anthem consists of one verse each from Nkosi Sikelel' iAfrika, Die 

38 Michelman (n 36 above) 37, referring to Port Elizabeth Municipality v Various 

Occupiers 2005 1 SA 217 (CC) par 35 (emphasis added). Michelman (n 36 above) 

38, meets the objection that the promise of a shared transformative purpose is 

another kind of fundamentalism (see nn 18-20 above and accompanying text) by 

describing legal-systemic unity as a contingent, relative value, arguing that 

contingency is brought about by the linkage of constitutional supremacy in s 1 

with rule of law: the value is contingent on ‘the direction of the pull’. When the 

direction of the pull is towards rule of justice rather than rule of law, the two 

values signify the unity of the legal system in the service of transformation by, 

under, and according to law’. This concern with the ‘pull towards the rule of 

justice’ points in the same direction as the ‘giving politics a chance’ argument of 

Van der Walt & Botha (n 18 above and accompanying text). 

39 Die Stem van Suid-Afrika (translated as The Call of South Africa) was written by 

Afrikaans author CJ Langenhoven in 1918; the music was composed by the 

Reverend ML de Villiers in 1921. The first stanza of Nkosi Sikelel' iAfrika was 

written in Xhosa and the music composed by Enoch Sontonga in 1897; seven 

stanzas in Xhoza were added by the poet Samuel Mqhayi. A Sesotho version by 

Moses Mphahlele was published in 1942. Information: http://www.info.gov.za/ 

aboutgovt/symbols/anthem.htm (accessed 8 May 2008).


Stem van Suid-Afrika and The Call of South Africa. 40 Nothing 

illustrates the accommodation of post-apartheid pluralism by way of 

a makeshift, patchwork compromise between opposing value systems 

better. 

However, patchwork compromises are not fatal. Like the national 

anthem, s 25 embodies a patchwork compromise between constitutional 

protection of existing property interests (s 25(1)-(3)) and 

constitutionally sanctioned land reform (s 25(4)-(9)). 41 Although it is 

difficult to imagine how the land-reform goals in s 25(5)-(9) could be 

promoted without infringing upon existing property interests in 

conflict with s 25(1)-(3), Michelman’s argument suggests a solution: 

the ‘legal-systemic unity’ of ‘every site of law’ (including both parts 

of s 25) is to be found in the constitutional purpose of ‘pulling in the 

justice direction’, ‘in the service of transformation by, under, and 

according to law.’ 42 Reading s 25 in this way suggests that protection 

of existing property interests and promotion of equitable access to 

property are guaranteed in so far as those purposes conform to the 

constitutional purpose of transformation by, under, and according to 

law (including legislation, common law and customary law). What 

renders Michelman’s reading compelling is not that it makes the 

conflict between two seemingly contradictory provisions or normative 

purposes disappear, nor that it identifies an overarching purpose in 

terms of which they can be reconciled or balanced, but that it shows 

that what appears like a contradiction is in fact a tension, 43 a locus of 

significant dissent, that makes social deliberation both inevitable and 

possible and that therefore ‘gives politics a chance’. The very 

possibility of politics is founded on the significant dissent embodied in 

the tension between the two parts of s 25. Constitutional application 

gives politics a chance in so far as it views this tension as part of the 

contingent ‘legal-systemic unity’ of ‘pulling in the justice direction of 

transformation by, under, and according to law’; not in order to deny 

or resolve the tension, but to explore it in the context of the equally 

contingent and contradictory justice goals of constitutional 

transformation itself. The constitutional purpose of transformation 

by, under and in accordance with law is not uncontested; it is a site 

for dissent and social deliberation; for politics. Again, in as far as law 

is inevitable in the process of social and political transformation, its 

40 Ss 2, 248(1) of the 1993 Constitution; s 4, Schedule 1 of the 1996 Constitution. 

The 1997 proclamation was published in Government Gazette 18341. 

41 

AJ van der Walt Constitutional property law (2005) 12-18. 

42 Michelman’s reading finds support in Port Elizabeth Municipality (n 38 above), 

specifically regarding s 25. 

43 

Van der Walt & Botha call this an aporia (n 18 above and accompanying text). 

Michelman explored something like the ‘legal-systemic unity’ (n 36 above) earlier 

(see n 18 above, 1112): ‘My suggestion is to seek a rapprochement of property 

and popular sovereignty in the idea that rights under a political constitution, 

including property rights, are first of all to be regarded as political rights.’


legitimacy must depend on the extent to which it ‘gives politics a 

chance’ by acknowledging the sacrifice and upholding the dissent 

inherent in this tension. 

My goal here is to consider the Constitutional Court’s response, 

during the 2007 term, to the compromise-based, patchwork 

relationship between existing law (particularly the common law) and 

the Constitution. In earlier constitutional discourse this has mostly 

been debated on the basis of the so-called application issue, referring 

to the question how the Constitution should find application in 

existing law. On the one hand, based on the constitutional aspiration 

to rectify the injustices and heal the divisions of the past and 

establish a society based on democratic values, social justice and 

fundamental human rights, the intention is clearly that the 

Constitution is the supreme law and that all law should be subject to 

it 44 — in this respect, the Constitution embodies a strong imperative 

for change in the general direction indicated by the transformation 

goals set out in the Constitution, as explained by Michelman. On the 

other hand, the desire for stability and certainty inherent in the rule 

of law principle implies a constitutional obligation not to disturb 

settled bodies of existing law 45 — and the vested rights protected by 

them — unnecessarily. 46 As I have argued above, this desire for 

stability and certainty will have a stronger effect in areas where the 

relevant constitutional provisions do not reveal the ‘visionary unity’ 

of the Constitution — in the compromise cases, we can expect that 

tradition and stability will offer stronger resistance against change. 

What Karl Klare described as ‘professional legal sensibilities’ will also 

tend to favour the stability of tradition and settled law, preferring to 

restrict change and uncertainty to instances where it is clearly 

required or mandated by unambiguous constitutional or statutory 

prescriptions. 47 

That brings me to my central premise: tension between the push 

for constitutional reform and the pull of traditional stability 

constitutes more than just a minor difference about the pace or the 

direction of change. Underlying this tension is a larger conflict 

44 Ss 1, 2, 7, 8(1) of the Constitution; with regard to ss 1, 2 compare Michelman (n 

36 above) 34-44. 

45 

The phraseology is from S Woolman ‘Application’ in S Woolman et al (eds) 

Constitutional law of South Africa (2nd Edition, OS, 2005) 10. 

46 S 1 (rule of law); s 39(3): the Bill of Rights does not deny the existence of other 

rights or freedoms recognised or conferred by common law, customary law or 

legislation, to the extent that they are consistent with the Bill. Compare ss 173, 

8(3), 39(2) on the higher courts’ power to develop the common law to promote 

the spirit, purport and objects of the Bill. Compare T Roux ‘Continuity and change 

in a transforming legal order: The impact of section 26(3) of the Constitution on 

South African law’ (2004) 121 South African Law Journal 466-492 466 467. 

47 

K Klare ‘Legal culture and transformative constitutionalism’ (1998) 14 South 

African Journal on Human Rights 146-188 166-167.


between two opposing normative values or ideologies, one favouring 

the stability of existing law, based on the presence and the presumed 

political neutrality of the status quo; the other relying on the political 

justifiability of fundamental change indicated by the presumed social 

construction and injustice of the status quo. Those who think that the 

existing distribution of wealth, privilege and power and the law that 

helped create and uphold it are politically neutral and therefore not 

in need of political reconsideration favour a regime of rights in which 

the direct effect of the Constitution is preservation of the status quo; 

those who think that the existing regime and its laws are inherently 

politically constructed and hence in need of political reconsideration 

favour the view that neither law nor rights are immune from 

constitutional effect or democratic redefinition. In this perspective, 

the relationship between existing law and Constitution is an instance 

of the tension between two conflicting sets of normative values, 

respectively favouring stability and change, certainty and justice, 

rights and democracy, tradition and constitution. In the absence of a 

‘visionary unity’ of purpose, conflict between them could precipitate 

a slide into normative anarchy. Simply ignoring or subjugating one in 

favour of the other could, however, undermine the diversity and 

pluralism that is a prerequisite for significant social dissent, politics 

and democracy. In the terminology of Van der Walt and Botha, the 

legitimacy of constitutional review depends upon the extent to which 

constitutional law can ‘give politics a chance’ by acknowledging the 

aporia inherent in the tension between stability and change, and by 

recognising the sacrificial inevitability of upholding, in a concrete 

case, the one at the cost of the other, without the comfort or the 

certainty of an appeal to foundation. 

Michelman recognises the contingency of accepting diversity, but 

insists that constitutional transformation is possible to the extent that 

the direction of the pull between conflicting values favours justice 

rather than law. The constitutional tension between legality, rule of 

law and constitutional supremacy involves an aporia; so does the 

tension between stabilising the old and promoting the new, in the 

sense that rule of law safeguards stability via the back door of legal 

certainty while constitutional supremacy demands transformation. 

Consequently, Michelman’s notion of visionary unity under the 

Constitution applies very usefully to the problem of existing law vs 

reform, but we should be careful not to read his analysis as superficial 

pragmatism. The unifying vision he identifies does not indicate an 

inflexible preference for change and against stability: the visionary 

unity of the Constitution identifies justice as a normative guideline for 

making hard decisions in concrete cases, but it does not pre-empt 

decisions in favour of change and against stability. In the language of 

Van der Walt and Botha, Michelman’s argument recognises that every 

choice for or against stability or reform is a sacrificial one: the choice 

cannot be justified, but making an unavoidable decision is justified in


so far as it accommodates social dissent and open democratic 

deliberation. In every case, the question is whether the direction of 

the pull favours justice. 

Relying on the theoretical groundwork of Van der Walt and Botha 

and Michelman, my hypothesis is that normative pluralism in the 

contradictions between constitutional reform and stability could 

easily result in either normative anarchy or reactionary formalism, 

unless the normative unity, the general direction of the pull, between 

these opposing forces favours transformative justice, by which I mean 

the justice that is implied by renunciation of everything the apartheid 

legal order represented and espousal of the central transformative 

values of human dignity, equality and freedom in an open and 

democratic society. This normative unity does not imply an inflexible 

preference for change and against stability, but it indicates the 

direction of the normative pull against which a decision for either 

change or stability must be justified. The questions that I will pose are 

whether the Constitutional Court is in the process of developing a 

direction-giving approach in deciding cases involving the contradiction 

between constitutional reform and stability and, if it is; 

whether this approach is justifiable in that it does not simply entrench 

a new formalism but, instead, opens up space for politics. 

3 One legal system 

There is only one system of law. It is shaped by the Constitution which is 

the supreme law, and all law, including the common law, derives its 

force from the Constitution and is subject to constitutional control. 48 

The tension between preservation of the common law tradition and 

constitutional transformation initially emerged in debates about the 

direct or indirect horizontal application of the Bill of Rights. 

Commentators and policy makers concerned that private-law 

enclaves should not be allowed to escape the reach of constitutional 

reform or to continue ‘privatised’ injustice proposed that constitutional 

rights should apply directly; others queried the wisdom and 

the legitimacy of ‘importing’ constitutional values into private law. 

Clothed as concern about the unnecessary abolition of a welldeveloped 

and supposedly politically neutral legal tradition, the 

latter often concealed anxiety about the continued existence of 

established privilege and wealth. Conflicting decisions in which the 

new constitutional order was sometimes welcomed, sometimes 

accepted with a measure of suspicion, sometimes rejected outright, 

demonstrated uncertainty about the normative scheme that informs 

post-1994 legal reasoning. 

48 Chaskalson P, Pharmaceutical Manufacturers (n 37 above) par 44.


The constitutional text makes it clear that the Constitution is the 

supreme law of the land; 49 that the Bill of Rights is the cornerstone of 

democracy and that it enshrines the values of human dignity, equality 

and freedom; 50 that it applies to all law and that it binds all three 

arms of government and all organs of state; 51 that the state must 

respect, protect, promote, and fulfil the rights in the Bill of Rights; 52 

that the Bill of Rights binds natural and juristic persons if and to the 

extent that it is applicable, taking into account the nature of the 

right; 53 that the courts, in order to give effect to a right in the Bill of 

Rights, must apply or, where necessary, develop the common law to 

the extent that legislation does not give effect to that right and may 

develop rules of the common law to limit the right, 54 and that courts, 

tribunals and other forums must, when interpreting legislation and 

when developing the common law and customary law, promote the 

spirit, purport and objects of the Bill of Rights. 55 However, it was not 

clear how this should work out in practice — when should the 

Constitution or new legislation trump the common law; when should 

the common law be developed to promote the spirit, purport and 

objects of the Constitution; how much should be changed in the 

process? Even more importantly, what should the effect of established 

common law norms and values be when it is unclear whether the 

Constitution or new legislation trumps the common law? 

Some authors and judges share Michelman’s view that 

transformation and stability can exist together and that an 

application theory and practice that uphold the tension is obligatory 

— the very idea of transformative constitutionalism 56 is founded upon 

the notion of a constitutional order that guarantees stability while 

49 S 2. 

50 

S 7(1). 

51 S 8(1). 

52 S 7(2). 

53 

S 8(2). 

54 S 8(3). 

55 S 39(2). 

56 

The concept, coined by Klare (n 47 above) 146-188, inspired a cottage industry in 

constitutional theory; see H Botha ‘Metaphoric reasoning and transformative 

constitutionalism’ 2002 Tydskrif vir Suid-Afrikaanse Reg 612-627; 2003 Tydskrif 

vir Suid-Afrikaanse Reg 20-36; D Moseneke ‘The fourth Bram Fischer memorial 

lecture: Transformative adjudication’ (2002) 18 South African Journal on Human 

Rights 309-319; H Botha ‘Freedom and constraint in constitutional adjudication’ 

(2004) 20 South African Journal on Human Rights 249-283; le Roux (n 19 above) 

629-675; Roux (n 46 above); AJ van der Walt ‘Transformative constitutionalism 

and the development of South African property law’ 2005 Tydskrif vir Suid- 

Afrikaanse Reg 655-689; 2006 Tydskrif vir Suid-Afrikaanse Reg 1-31; M Pieterse 

‘What do we mean when we talk about transformative constitutionalism?’ (2005) 

20 SA Public Law 155-166; S Liebenberg ‘Needs, rights and transformation: 

Adjudicating social rights’ (2006) 1 Stellenbosch Law Review 5-36; P Langa 

‘Transformative constitutionalism’ (2006) 17 Stellenbosch Law Review 351-360. 

See further the contributions of H Botha, D Davis, J Froneman, J van der Walt and 

K van Marle in H Botha, AJ van der Walt & JWG van der Walt (eds) Rights and 

democracy in a transformative constitution (2003).


simultaneously requiring (and authorising or even driving) social, 

economic or political transformation. 57 

The problem is that the debate about the place and role of the 

common law in the new dispensation 58 has been dominated by the 

largely undeclared assumption that the common law should be 

insulated, in some measure, against constitutional amendment, 

presupposing that a substantial part of the common law might remain 

unaffected by the Constitution and that development of the rest can 

take place on its own terms, without raising constitutional issues. The 

jurisdictional format that this turf war adopted under the 1993 

Constitution 59 ended in Pharmaceutical Manufacturers when the 

Constitutional Court rejected the Supreme Court of Appeal’s view 

that the common law grounds for review continued to exist as a body 

of norms ‘separate and distinct from the Constitution’: 60 

There are not two systems of law, ... each operating in its own field with 

its own highest court. There is only one system of law. It is shaped by the 

Constitution which is the supreme law, and all law, including the 

common law, derives its force from the Constitution and is subject to 

constitutional control. 61 

57 Michelman and Van der Walt & Botha argue that the tension in the Constitution is 

central to the possibility of politics; see nn 18-20, 36-38 above. See further Roux 

(n 46 above) 487. 

58 The problem of balancing constitutional supremacy and common law is not unique 

to South African law. A similar problem exists in German constitutional law and 

early commentaries on the South African Constitution found inspiration in German 

theory: D Davis, M Chaskalson & J de Waal ‘Democracy and constitutionalism: The 

role of constitutional interpretation’ in D van Wyk, J Dugard, B de Villiers & D 

Davis (eds) Rights and constitutionalism: The new South African legal order 

(1994) 1-130 89-97. In Carmichele v Minister of Safety and Security (Centre for 

Applied Legal Studies Intervening) 2001 4 SA 938 (CC) par 54 Ackermann & 

Goldstone JJ referred to German constitutional law. Michelman (n 36) 40 relies on 

this passage in Carmichele to show the Constitutional Court’s support for the view 

that a ‘comprehensive (if doubtless somewhat inchoate) vision of the good 

society’ is inherent in the Constitution. Since the European Convention was 

incorporated into English law by the Human Rights Act 1998, English courts have 

struggled to reconcile common law with human rights standards, eg eviction vs 

home interest in art 8 European Convention of Human Rights: Harrow London 

Borough Council v Qazi [2004] 1 AC 983; Kay v London Borough of Lambeth; Leeds 

City Council v Price [2006] UKHL 10 (HL); Connors v United Kingdom [2004] ECHR 

223. 

59 Michelman (n 36 above) 22. 

60 

The SCA decision was Commissioner for Customs and Excise v Container Logistics 

(Pty) Ltd; Commissioner for Customs and Excise v Rennies Group Ltd t/a 

Renfreight 1999 3 SA 771 (SCA); the CC decision was Pharmaceutical 

Manufacturers (n 37 above). Michelman (n 36 above) 15-33 explains Amod v 

Multilateral Motor Vehicle Accidents Fund 1999 4 1319 (SCA) as a model for 

judges who want to keep a judgment out of the constitutional arena to reserve 

for the SCA final control over a category of future cases (18), although Amod 

might simply have avoided ‘thorny issues’; its cause of action preceded 

61 

commencement of the 1993 Constitution. 

Pharmaceutical Manufacturers (n 37 above) par 44. See Michelman (n 36 above) 

21-22.


In Pharmaceutical Manufacturers the Constitutional Court 

‘established, once and for all, that the SCA cannot insulate a decision 

on legality from CC review by dressing it as a merely common law (and 

hence not a constitutional) decision.’ 62 As far as review of 

administrative action is concerned, this decision established a unitary 

vision of the relationship between common law and Constitution: the 

common law continues to exist in as far as it does not directly conflict 

with the Constitution. However, it does not exist as an independent 

body of law — there is just one system of law, shaped by the 

Constitution as the supreme law. The common law derives its force 

from the Constitution and it is subject to constitutional control and 

the tension between the Constitution and the common law is to be 

understood in terms of the transformative goal, the unitary normative 

vision, of the Constitution. The application, interpretation and 

development of the common law must take place in line with this 

vision. The essence, in the words of Michelman, is that application of 

the Constitution should promote the shared vision of an open society 

built on democracy, social justice and fundamental human rights and 

on human dignity, equality and freedom. 

Unfortunately, Pharmaceutical Manufacturers did not bring 

clarity about the effect of the Constitution on existing law (including 

the common law) as a whole; 63 outside of administrative law, the idea 

that the common law might escape constitutional influence and that 

development of the common law can develop on its own terms, 

without raising constitutional issues, retains some traction. Private 

law specialists tended simply to establish absence of direct conflict 

between common law principles and constitutional rights, instead of 

seeking for normative or visionary unity. This approach presupposes 

that the common law could coexist alongside the Constitution, more 

or less unchanged, as long as it is free from the pernicious influence 

of apartheid. The strong version of this argument, that the law in 

general is politically neutral and that the abolition of apartheid laws 

62 Michelman (n 36 above) 18. This is what L du Plessis ‘“Subsidiarity”: What’s in the 

name for constitutional interpretation and adjudication?’ (2006) 17 Stellenbosch 

Law Review 207-231 211 refers to as jurisdictional subsidiarity. 

63 Pharmaceutical Manufacturers (n 37 above) was decided prior to commencement 

of the Promotion of Administrative Justice Act 3 of 2000 (PAJA). Like Fedsure Life 

Assurance Ltd v Greater Johannesburg Transitional Metropolitan Council 1999 1 

SA 374 (CC), Pharmaceutical Manufacturers dealt with the broader notion of 

legality, which includes executive and legislative action outside of the range of 

administrative action regulated by PAJA. With regard to the relationship between 

the common law, PAJA and the Constitution see Bato Star Fishing (Pty) Ltd v 

Minister of Environmental Affairs 2004 4 SA 490 (CC) par 25: the provisions of s 6 

PAJA ‘divulge a clear purpose to codify the grounds of judicial review of 

administrative action as defined in PAJA. The cause of action for the judicial 

review of administrative action now ordinarily arises from PAJA, not from the 

common law as in the past. And the authority of PAJA to ground such causes of 

action rests squarely on the Constitution ... As PAJA gives effect to s 33 of the 

Constitution, matters relating to the interpretation and application of PAJA will 

of course be constitutional matters.’


properly eradicated politics from law, thereby restoring the common 

law to its former glory, no longer enjoys theoretical credibility 64 

outside of doctrinally conservative circles. 65 The weak version, that 

the common law is flexible and capable of adaptation that would 

bring it into line with new demands and requirements in the new 

constitutional democracy, is shared by many lawyers, including highranking 

judges in the new legal order. 66 

In the application debate, this argument supported the view that 

development of the common law, although referred to in the 

Constitution, 67 is something the civil courts are familiar with and that 

they should be left to do the work as they have always done it. 68 This 

view enjoyed wide support and strengthened the division of work 

between the Constitutional Court and the Supreme Court of Appeal 

under the 1993 Constitution. 69 The High Courts, the Supreme Court of 

Appeal and the Constitutional Court now have concurrent jurisdiction 

to decide on the constitutionality of legislation, but the notion 

persists that the Supreme Court of Appeal is the specialist court on 

64 Botha (n 18 above) 561-581. 

65 

See J Potgieter ‘The role of the law in a period of political transition: The need 

for objectivity’ (1991) 54 Tydskrif vir Hedendaagse Romeins-Hollandse Reg 800- 

807 802: ‘It must be stressed that the basic assumption that the South African 

legal system as a whole has become illegitimate, is unfounded. The crisis in South 

Africa lies primarily in the socio-political rather than the legal sphere’; J 

Neethling ‘’n Toekomsblik op die Suid-Afrikaanse privaatreg — volwaardige 

naasbestaan of versoenende sintese?’ (‘A future perspective on South African 

private law — full coexistence or reconciliatory synthesis?) in A van Aswegen (ed) 

The future of South African private law (1994) 1-9 at 3: ‘In the first instance, the 

apartheid era cannot be attributed to Roman-Dutch law — the blame must be 

placed squarely on the shoulders of the ruling minority who introduced the 

system of apartheid by way of legislation, and they were primarily enabled to this 

end by the doctrine of parliamentary sovereignty which derives from English 

constitutional law.’ See further TJ Scott ‘The future of our Roman-Dutch law: 

Reflections and a suggestion’ (1993) 26 De Jure 394-400 399. 

66 MM Corbett ‘Trust law in the 90s: Challenges and change’ (1993) 56 Tydskrif vir 

Hedendaagse Romeins-Hollandse Reg 262-270 264. See also I Mahomed ‘The 

future of Roman-Dutch law in Southern Africa, particularly in Lesotho’ 1985 

Lesotho Law Journal 357-365 360; A Sachs ‘The future of Roman-Dutch law’ (chap 

8) in A Sachs Protecting human rights in a new South Africa (1990) 90-103 (‘we 

know it more or less, we have the books and the rules and the procedures 

available, we might as well use it’). See AJ van der Walt ‘Tradition on trial: A 

critical analysis of the civil-law tradition in South African property law’ (1995) 11 

South Africa Journal on Human Rights 169-206 169-171. 

67 Particularly ss 8(3), 39(2) and 173. 

68 See Brisley v Drotsky 2002 4 SA 1 (SCA); Afrox Health Care Bpk v Strydom 2002 6 

SA 21 (SCA). A Cockrell ‘Private law and the bill of rights: A threshold issue of 

“horizontality”’ in Bill of rights compendium (1998) paras 3A8, 3A7 argues that 

‘the direct application of constitutional rights against private agencies must be 

mediated by the operation of the common law’. 

69 In the 1996 Constitution, as in the 1993 Constitution, the Constitutional Court is a 

specialist court and not a court of general jurisdiction. In terms of s 167 it is the 

court of final instance in constitutional matters only. The Constitutional Court 

must confirm any order of a High Court or the Supreme Court of Appeal to the 

effect that an Act of Parliament, a provincial Act or conduct of the President is 

unconstitutional before that order can have effect: s 167(5). See Currie & de Waal 

(n 30 above) 108-109.


common law matters. 70 In Zantsi 71 the Constitutional Court stated 

that, because of the far-reaching implications of constitutional 

decisions, it would be better if lower courts first heard cases where 

constitutional issues are raised, because that would allow the law to 

develop incrementally. 72 Incremental development is said to be in 

line with the inherent logic of the common law and Zantsi therefore 

reinforced the idea that development of the common law under the 

Constitution could continue largely as before. 73 

The notion that the Supreme Court of Appeal is a specialist court 

in common law matters is not the only reason why cases involving 

development of the common law are left to the civil courts. The 

Constitutional Court prefers not to sit as court of first and last 

instance on any matter and it prefers that the High Court or the 

Supreme Court of Appeal decide on common law issues before it 

considers the matter on appeal. 74 The outcome in Carmichele 75 could 

be seen as proof that the division of labour between the 

70 Amod v Multilateral Motor Vehicle Accidents Fund 1998 4 SA 753 (CC) par 33: the 

SCA’s view on whether the common law should be developed is important because 

of ‘the breadth of its jurisdiction and its expertise in the common law’. 

71 Zantsi v Council of State, Ciskei 1995 4 SA 615 (CC) par 5. This echoed the 

decision in S v Mhlungu 1995 3 SA 867 (CC) par 59 that, where it is possible to 

decide any case without reaching a constitutional issue, that is the course that 

should be followed (dissenting, Kentridge J). This principle was subsequently 

confirmed in Zantsi (above) par 3 and followed in Amod (CC) (n 70 above) par 33. 

Du Plessis (n 62 above) 207-231 calls it the subsidiarity principle. 

72 The notion of ‘interstitial’, incremental development that fills up small gaps in 

received doctrine, case by case, originated in HLA Hart The concept of law (1961) 

37-38. The argument is that larger changes in the common law that affect 

acquired rights should be introduced by legislation. However, judges are wellplaced 

and qualified to make smaller, reasoned ‘elaborations’ or incremental 

changes to the common law on doctrinal reasoning. The argument that the courts 

are neither democratically legitimated to bring about major changes nor qualified 

to judge the wisdom of policy changes is based on the deference approach of the 

legal process school that responded to Realism by resorting to either deference or 

craft: A Bickel ‘The Supreme Court 1960 term: The passive virtues’ (1961) 75 Harv 

LR 40-79; H Wechsler ‘Toward neutral principles of constitutional law’ (1959) 73 

Harv LR 1-35. 

73 Khumalo v Holomisa 2002 5 SA 401 (CC) par 45 (not shown that the common law 

‘as currently developed was inconsistent with the provisions of the Constitution’) 

can be traced to this argument. The defence of reasonable publication was 

developed by the SCA in National Media Ltd v Bogoshi 1998 4 SA 1196 (SCA), 

contradicting earlier case law where publishers were held strictly liable for 

defamation. Bogoshi argued that this was inspired by the dynamic logic of the law 

of delict and not by constitutional developments, despite confirmation of the 

established authority (Pakendorf v De Flamingh 1982 3 SA 146 (A)) in Neethling v 

Du Preez; Neethling v The Weekly Mail 1994 1 SA 708 (A), the (then still) 

Appellate Division of the Supreme Court holding that a newspaper could only 

escape a claim for defamation if it could establish truth. 

74 Carmichele (n 58 above) paras 56-59. 

75 In Carmichele (n 58 above) par 45 the Constitutional Court referred the matter 

back to the trial court (paras 81-83), where it was decided that the common law 

would not have placed a duty upon the state; that the Constitution would place a 

duty upon the state; that the common law therefore had to be developed to place 

a duty upon the state; and damages was granted: Carmichele v Minister of Safety 

and Security 2003 2 SA 656 (C).


Constitutional Court and the civil courts functions well. Given its 

declared preference for developing the common law incrementally 

the fact that the Constitutional Court referred the final order back to 

the High Court could be seen as an effort not to intrude too far into 

the sphere of the legislature, who should be given maximum scope to 

amend the common law by legislation. 76 At the same time, referring 

development of the common law to the high courts need not be seen 

as a sign of deference; the Constitutional Court has indicated that it 

will not compromise the supremacy of the Constitution and that it will 

develop the common law if that is what the interests of justice 

require. 77 As Du Plessis points out, the Pharmaceutical Manufacturers 

decision 78 was 

the Constitutional Court’s response to a tendency of the Supreme Court 

of Appeal (at the time) to pretend that constitution-related issues can 

be disposed of in an ‘enlightened’, rights-friendly manner without 

reference to (let alone reliance on) the Constitution. 79 

In some cases the tendency to regard the Supreme Court of Appeal as 

the specialist court in matters concerning the common law did result 

in outcomes that restricted the influence of the Constitution on 

private law. In Brisley 80 and Afrox 81 it was accepted that public policy 

(of which open-ended norms such as good faith form part) is informed 

by fundamental values in the Constitution 82 and that the Constitution 

might ‘spur on the development of new substantive rules of law’, 83 

but neither good faith nor constitutional values ‘were sufficient to 

outweigh the traditional bias in favour of the strict enforcement of 

agreements’ 84 and the Supreme Court of Appeal was not willing to 

undertake development of the common law that would amount to 

judicial interference with agreements willingly entered into by 

private parties. 85 

76 Zantsi (n 71 above) par 5. 

77 Zantsi (n 71 above) par 4; Harksen v Lane NO 1998 1 SA 300 (CC) par 26. 

78 

n 37 above. 

79 Du Plessis (n 62 above) 217, referring to Bogoshi (n 73 above). Woolman (n 45 

above) 10 refers to the same tendency. 

80 

n 68 above. Roux (n 46 above) 485 argues that the majority in Brisley made a 

policy choice in favour of legal continuity. 

81 n 68 above. 

82 

In view of Carmichele (n 58 above); see Brisley (n 68 above) par 91; Afrox (n 68 

above) par 18. 

83 GF Lubbe ‘Taking fundamental rights seriously: The bill of rights and its 

implications for the development of contract law’ (2004) 121 South African Law 

Journal 395-423 401. 

84 Lubbe (n 83 above) 401 414. 

85 

Du Plessis (n 62 above) 218-221, referring to Afrox (n 68 above). See Woolman (n 

45 above) 95; T Naudé & G Lubbe ‘Exemption clauses — A rethink occasioned by 

Afrox Healthcare Bpk v Strydom’ (2005) 122 South African Law Journal 441-463; S 

Woolman & D Brand ‘Is there a constitution in this courtroom? Constitutional 

jurisdiction after Walters and Afrox’ (2003) 18 SA Public Law 37-82.


During the 2007 term the Constitutional Court decided three cases 

in which the effect of the Constitution on the common law was in 

issue: Barkhuizen, 86 NM 87 and Masiya. 88 Stu Woolman concluded that 

these decisions prove that ‘a penchant for outcome-based decisionmaking, 

and a concomitant lack of analytical rigour, has finally caught 

up with the Constitutional Court’; that ‘the court’s current process of 

(public) reasoning — its preferred mode of analysis — has genuinely 

deleterious consequences.’ 89 In coming to this rather damning 

conclusion, Woolman attaches particular importance to ‘the court’s 

persistent refusal to engage in the direct application of the Bill of 

Rights’ and its insistence upon replacing direct application analysis 

with what he describes as ‘flaccid analysis in terms of three vaguely 

defined values’, namely dignity, equality and freedom. 90 This 

strategy of overemphasising the constitutional values and development 

of the common law (or indirect application) in terms of s 39(2) 

instead of direct application of the rights provisions in the Bill of 

Rights, so Woolman argues, ‘has freed the court almost entirely from 

the text, and thereby grants the court the licence to decide each case 

as it pleases’, thereby also unwittingly undermining the Bill of Rights 

and the rule of law. 91 

These are strong words. If the decisions criticised by Woolman 

indeed indicate a tendency to free the Constitutional Court from the 

text of the Constitution so that it can decide each case as it pleases, 

on the basis of ‘value-speak’, and if that tendency does undermine 

the Bill of Rights and the rule of law, then we are perhaps indeed 

seeing the fruits of normative anarchy. What Woolman is suggesting is 

that the decisions in Barkhuizen, NM and Masiya represent decision 

making at more or less the same level as Mr Davies senior’s ruling that 

Robin could not arrange a visit with a friend in the afternoon when he 

already had another social engagement for that evening, with his 

parents — the decision is arbitrary because it is based on unilateral 

pronouncement of vaguely defined values that are neither 

substantiated by foundation nor open to debate. In the absence of 

foundation, normative pluralism calls for deliberation and politics, 

not flaccid but arbitrary and inflexible rules. 

86 

Barkhuizen v Napier 2007 5 SA 323 (CC) (time-limitation clauses in an insurance 

87 

contract). 

NM v Smith (Freedom of Expression Institute as Amicus Curiae) 2007 5 SA 250 (CC) 

88 

(publication that revealed HIV status of the applicants without consent). 

Masiya v Director of Public Prosecutions, Pretoria (Centre for Applied Legal 

Studies and Another, Amici Curiae) 2007 5 SA 30 (CC) (common law definition of 

89 

rape). 

S Woolman ‘The amazing, vanishing bill of rights’ (2007) 124 South African Law 

90 

91 

Journal 762-794 762. 

Woolman (n 89 above) 763. 

As above.


In the early stages of the constitutional debate much emphasis 

was placed on direct horizontal application of the Bill of Rights. In two 

landmark decisions on direct horizontal application, the 

Constitutional Court first held that s 7(1) and 7(2) of the 1993 

Constitution had to be interpreted restrictively so that direct 

horizontal application on the law of defamation would be excluded; 92 

then subsequently decided that direct horizontal application was 

possible under s 8(2) of the 1996 Constitution. 93 In terms of the 

application debate as it has crystallised in the case law and literature, 

this does not take the matter much further and we are still wavering 

between the preservative impulse to uphold the common law 

tradition unless the Constitution or new legislation explicitly requires 

us to depart from it and the transformative desire to change the legal 

infrastructure that inhibits speedy and effective social and economic 

reform. Most commentators who were initially enthusiastic about 

direct application have given up on the debate or shifted their 

focus, 94 but Woolman insists that the Constitutional Court’s 

persistent evasion of direct application ‘has genuinely deleterious 

effects’. 95 In the next two sections I evaluate decisions from the 2007 

term, arguing that they hold the key to a different perspective on the 

application issue. In the following two sections I identify, from the 

case law of the Constitutional Court, indications of what I describe as 

a subsidiarity approach that enables the courts to simultaneously 

promote constitutional reforms and uphold constitutional stability, 

choosing between the two impulses (and justifying the choice in each 

case) with reference to constitutional principles of justice and 

democracy. In the final section of the article I argue that this 

approach can be justified, even in so far as it creates law that 

resembles fixed doctrine, in so far as it is not aimed at entrenching 

shallow formalism but rather at upholding and enabling a 

constitutionally authorised and required balance between 

constitution and law, rights and democracy, or stability and change, 

within the particular historical and constitutional context of the South 

African Constitution. My underlying assumption, based upon the 

theoretical analysis in the first two sections of the article above, is 

that such a balance is justified in so far as it opens up space to give 

politics a chance, in the sense that it recognises diversity and 

pluralism and creates room for real social contestation and dissent. 

92 Du Plessis (n 62 above) 220, referring to Du Plessis v De Klerk 1996 3 SA 850 (CC). 

93 Du Plessis (n 62 above) 221, referring to Khumalo (n 73 above). 

94 

Eg Van der Walt (n 34 above) 341-363; Roux (n 46 above). 

95 n 89 above.

4 Rights and democracy 


Legislation enacted by Parliament to give effect to a constitutional right ought 

not to be ignored. 96 

4.1 Introduction 

Decisions from the 2007 term indicate the emergence of an ‘angle of 

approach’ 97 — an analytic rhythm — according to which courts could 

decide upon the tension between rights and democracy, while 

acknowledging what Van der Walt and Botha 98 called the irresoluble 

dissent embedded in the countermajoritarian dilemma. Lourens du 

Plessis described this kind of thought process as ‘subsidiarity’, 

defining it as a reading strategy whereby a court refrains from taking 

a decision that can be taken by a lower court or avoids a constitutional 

decision if the matter can be decided on a nonconstitutional basis. 99 

Generally speaking, the notion of subsidiarity poses conceptual and 

analytical problems for the politics-enhancing approach to the 

fundamental contradiction set out in the first two sections of the 

article above. This is particularly true of the way in which the 

principle of subsidiarity has been enunciated in earlier South African 

case law. In 1995 the Constitutional Court enunciated a subsidiarity 

principle in Mhlungu, holding that courts should, whenever possible, 

decide cases without reaching a constitutional issue; 100 in 1998 it 

formulated another one in Amod, 101 stating that the development of 

the common law should (in terms of the 1993 Constitution) preferably 

be undertaken by the Supreme Court of Appeal. The latter principle 

was confirmed (under the 1996 Constitution) in Carmichele. 102 Stated 

this blandly, the principle of subsidiarity looks like a formalist escape 

96 Minister of Health NO v New Clicks South Africa (Pty) Ltd (Treatment Action 

Campaign and Another as Amici Curiae) 2006 2 SA 311 (CC) par 437 (Chaskalson 

CJ). 

97 I am indebted to Henk Botha, who first used this phrase at a workshop in 2007, 

explaining that it pointedly avoids the pretence of a technique that produces 

ready or final answers. The phrase originates from NS Ndebele’s novel The cry of 

Winnie Mandela (2003). 

98 See n 19 above and accompanying text. 

99 

I do not set out or enter into the general subsidiarity debate or literature, where 

this notion has attracted various understandings and where it remains contested; 

see Du Plessis (n 62 above) 207-231. Currie & de Waal (n 30 above) 75-78 call this 

the ‘avoidance principle’, although their interpretation is narrower and more 

mechanistic. See further L du Plessis ‘The South African constitution as memory 

and promise’ (2000) 11 Stellenbosch Law Review 385-394 388-389; W le Roux ‘War 

memorials, the architecture of the Constitutional Court building and countermonumental 

constitutionalism’ in W le Roux & K van Marle (eds) Law, memory and 

the legacy of apartheid. Ten years after Azapo v President of South Africa (2007) 

65-90 87. 

100 n 71 above, par 59 (dissenting, Kentridge J); confirmed in Zantsi (n 71 above) par 

3. 

101 n 70 above, par 14. 

102 n 58 above, paras 50-55.


from politics, but in the historical and constitutional context of the 

South African Constitution it could also be seen as a politicsconfirming 

and –enhancing device that ensures interplay between 

constitutional principles and democratic laws, reformist initiatives 

and vested rights, change and stability. In the former sense, 

subsidiarity would serve a formalist evasion of politics; in the latter 

sense it could promote normative pluralism and constitutional change 

without succumbing to either normative anarchy or reactionary 

formalism. 

In the next two sections I discuss decisions from the 2007 term 

that could indicate that the Constitutional Court is developing a more 

general subsidiarity approach. 103 In the final section I consider the 

implications of this approach for the application debate and evaluate 

the role that subsidiarity could play in promoting normative pluralism 

and avoiding anarchy. 

4.2 The SANDU principle 

In SANDU, 104 the Court confirmed 105 that, once legislation has been 

enacted to give effect to a right in the Constitution, litigants must 

rely on the legislation — and may not rely directly on the 

constitutional provision — when bringing action to protect that right 

against infringement. The principle establishes a subsidiarity rule 

because it excludes direct application of the constitutional provision 

once legislation has been enacted to give effect to that right. 

Legislation in this category include the Labour Relations Act 66 of 

1995 (LRA) (s 23), the Promotion of Administrative Justice Act 3 of 

2000 (PAJA) (s 33), the Promotion of Access to Information Act 2 of 

103 Aspects of this approach have been decided earlier. In New Clicks (n 96 above) par 

97 fn 81 Chaskalson CJ cites with approval a paper in which Cora Hoexter set out 

the logic of both SANDU and Bato Star. The paper was published as C Hoexter 

‘“Administrative action” in the courts’ 2006 Acta Juridica 303-324 (also published 

as H Corder (ed) Comparing administrative justice across the Commonwealth 

(2008)). See to the same effect Currie & de Waal (n 30 above) 649. 

104 South African National Defence Union v Minister of Defence 2007 5 SA 400 (CC) 

paras 51-52. See further Pillay (n 29 above) paras 39-40; Chirwa v Transnet Ltd 

2008 2 SA 24 (CC) paras 59 (Skweyiya J), 69 (Ngcobo J). In Sidumo v Rustenburg 

Platinum Mines Ltd 2008 2 SA 24 (CC) par 248 Ngcobo J confirmed the principle; 

the majority (paras 55-56) did not mention it. In MEC: Department of Agriculture, 

Conservation and Environment v HTF Developers (Pty) Ltd 2008 2 SA 319 (CC) the 

majority (par 24) confirmed that the National Environment Management Act 107 

of 1998 (NEMA) is legislation contemplated by s 24(b) without mentioning SANDU 

(but see Ngcobo J par 62). 

105 Formulated in NAPTOSA v Minister of Education, Western Cape 2001 2 SA 112 (C) 

123B, I-J; mooted in New Clicks (n 96 above) paras 95-96, 434-437 (Chaskalson CJ, 

Ngcobo J); left undecided in National Education Health and Allied Workers Union 

v University of Cape Town 2003 3 SA 1 (CC) par 17; Ingledew v Financial Services 

Board: In re Financial Services Board v Van der Merwe 2003 4 SA 584 (CC) paras 

23-24 (Ngcobo J). See IM Rautenbach ‘Overview of Constitutional Court decisions 

on the bill of rights — 2007’ 2008 Tydskrif vir Suid-Afrikaanse Reg 330-348 330.


2000 (PAIA) (s 32) and the Promotion of Equality and Prevention of 

Unfair Discrimination Act 4 of 2000 (PEPUDA) (s 9). 106 Application of 

the SANDU principle to PAJA and PEPUDA indicates that the 

distinction between vertical and horizontal application becomes 

irrelevant in the context of this principle. 

In line with s 39(2), SANDU requires that where ‘the Legislature 

enacts legislation in the effort to meet its constitutional obligations, 

and does so within constitutional limits, courts must give full effect 

to the legislative purpose’. 107 In Goedgelegen Tropical Fruits 108 the 

Constitutional Court applied the implications, emphasising that the 

Restitution of Land Rights Act 22 of 1994 is remedial legislation under 

the Constitution and that courts, to promote the spirit, purport and 

objects of the Bill of Rights, must acknowledge the historical 

background under which labour tenancies were cancelled. The cause 

of a historical dispossession will often not lie in an isolated moment 

in time or a single act but in the context and, as the contextual 

features of this case ‘constituted a grid of integrated repressive laws 

that were aimed at furthering the government’s policy of racial 

discrimination’ which ‘materially affected and favoured the ability of 

the [landowners] to dispossess the applicants of their labour tenancy 

rights’, 109 the former labour tenants could claim restitution. 

The SANDU principle indicates that, once legislation has been 

enacted to give effect to a constitutional right, direct application of 

the constitutional provision is precluded. However, this applies only 

when litigants seek to protect rights; when the constitutional validity 

of legislation is challenged, the constitutional provision applies 

directly. 110 In Islamic Unity 111 the Court expanded the qualification: 

when challenging the constitutional validity of any statute that 

106 In NAPTOSA (n 105 above), NEHAWU (n 105 above) and SANDU (n 104 above) the 

provision was s 23(5); the legislation the LRA. In Sidumo (n 104) and Chirwa (n 

104) two provisions (ss 23, 33) and statutes (LRA and PAJA) were in dispute. In 

Pillay (n 29) the provision was s 9, the legislation PEPUDA. In New Clicks (n 96) 

the provision was s 33, the legislation PAJA; in Ingledew (n 105) the provision was 

s 32, the legislation PAIA. In HTF Developers (n 104) the provision was s 24, the 

legislation NEMA. 

107 Sidumo (n 104 above) par 249, citing NEHAWU (n 105 above) par 14. Although 

Ngcobo J concurred in the outcome in Sidumo, he wrote a separate judgment: 

see n 104 above. 

108 Department of Land Affairs v Goedgelegen Tropical Fruits (Pty) Ltd 2007 6 SA 199 

(CC). 

109 Goedgelegen Tropical Fruits (n 108 above) paras 53, 66, 69, 71. 

110 New Clicks (n 96 above) par 437; SANDU (n 104 above) par 52; Sidumo (n 104 

above) par 249; Engelbrecht v Road Accident Fund 2007 6 SA 96 (CC) par 15 

111 

(constitutional challenge against legislation involves direct application of s 34). 

Islamic Unity Convention v Minister of Telecommunications 2008 3 SA 383 (CC) 

par 59.


conflicts with the constitutional principle of administrative justice, s 

33 is the review measure. 112 

SANDU, with its qualification, was originally justified in New 

Clicks: allowing litigants to rely on s 23 directly instead of on the LRA 

‘would encourage the development of two parallel streams of labour 

law jurisprudence, one under the LRA and the other under s 23(1)’; 

such a practice would be ‘singularly inappropriate’ 113 and would 

contradict the principle of ‘one system of law grounded in the 

Constitution’. 114 

Where, as here, the Constitution requires Parliament to enact legislation 

to give effect to the constitutional rights guaranteed in the Constitution, 

and Parliament enacts such legislation, it will ordinarily be 

impermissible for a litigant to found a cause of action directly on the 

Constitution without alleging that the statute in question is deficient in 

the remedies that it provides. Legislation enacted by Parliament to give 

effect to a constitutional right ought not to be ignored. And where a 

litigant founds a cause of action on such legislation, it is equally 

impermissible for a court to bypass the legislation and to decide the 

matter on the basis of the constitutional provision that is being given 

effect to by the legislation in question. 115 

The principle expounded by the Court is grounded in the norm that ‘to 

permit the litigant to ignore the legislation and rely directly on the 

constitutional provision would be to fail to recognise the important 

task conferred upon the Legislature by the Constitution to respect, 

protect, promote and fulfil the rights in the Bill of Rights.’ 116 The 

qualification, by contrast, is justified by the norm that the majority 

is accountable to and can be tested against constitutional limits. 

SANDU therefore reflects the tension of the countermajoritarian 

dilemma, with the principle emphasising democracy and the 

qualification emphasising rights: when the democratically elected 

legislature meets its constitutional obligations to enact legislation 

that gives effect to the rights in the Constitution, within 

constitutional limits, the courts should respect and give full effect to 

that legislation. To that extent, rights are subject to democracy. 

However, the courts can review the constitutionality of legislation 

with a direct appeal to the Constitution. To that extent, democracy 

is subject to rights. In so far as the SANDU principle promotes this 

112 Zondi v MEC for Traditional and Local Government Affairs 2005 3 SA 589 (CC): If a 

constitutional challenge is brought against legislation for being inconsistent with s 

33, the proper approach is to consider whether it can be consistent with s 33: s 

39(2) (par 102). The next question is whether it contemplates administrative 

action for purposes of s 33 (par 103), whether it limits the rights in s 33 (par 106) 

and whether the limitation is justifiable (s 36(1)). 

113 New Clicks (par 96 above) paras 434, 436; NAPTOSA (n 105 above) par 123B, J. 

114 Pharmaceutical Manufacturers (n 37 above) par 44. 

115 New Clicks (n 96 above) par 437 (Chaskalson CJ). Footnote omitted. 

116 SANDU (n 104 above) par 52; Pillay (n 29 above) par 40.


balance between democracy and rights one could say that it could be 

employed to promote normative pluralism without falling into the 

trap of either normative anarchy (anything goes) or reactionary 

formalism (escaping real social contestation and dissent through 

formalist reasoning). To that extent, the principle could also be said 

to give politics a chance in that it promotes and enables real social 

contestation and dissent between the democratic push of the 

majority and the stabilising pull of constitutional stability. 

4.3 The Bato Star principle 

In Chirwa 117 the Court confirmed a second principle: 118 once 

legislation has been enacted to give effect to a right in the 

Constitution, and in so far as the legislation was intended to codify 

the common law, litigants may not rely directly on the common law 

when seeking to protect that right against infringement. The 

principle also establishes a subsidiarity rule in that it precludes 

application or development of the common law; 119 once legislation 

has been enacted to give effect to a right, that right is protected 

through constitution-conforming interpretation and application of the 

legislation in accordance with s 39(2). 120 

Bato Star does not render the common law irrelevant — the 

common law informs interpretation of the legislation. 121 The 

stabilising role of the common law should be treated with care, 

though — the common law should inform interpretation of legislation 

only in so far as it is consistent with the Constitution, consistent with 

the legislation and able to augment interpretation of the legislation 

as foreseen in s 39(2), namely to promote the spirit, purport and 

objects of the Bill of Rights. Du Plessis 122 and de Ville 123 explain that 

the relationship between the common law (favouring stability by 

upholding the status quo) and legislation (favouring change through 

policy-driven intervention) has to be inverted in view of the 

centripetal interpretive force of the Constitution. 

The areas of existing law identified as affected by this principle 

are labour disputes (the LRA) 124 and just administrative action 

117 n 104 above, par 23; see further Fuel Retailers (n 25 above) par 37. 

118 Stated earlier in Bato Star (n 63 above) par 25 New Clicks (n 96 above) par 96. 

119 In terms of ss 173, 8(3) or 39(2). See to the same effect Currie & de Waal (n 30 

above) 650. 

120 S 39(2). 

121 n 63 above, par 22. 

122 L du Plessis Re-interpretation of statutes (2002) 179-181. 

123 JR de Ville Constitutional and statutory interpretation (2000) 172 66. 

124 Chirwa (n 104 above) par 23.


(PAJA), 125 but the Restitution of Land Rights Act 22 of 1994, 126 the 

National Environment Management Act 107 of 1998 (NEMA), 127 the 

Water Services Act 108 of 1997, the Mineral and Petroleum Resources 

Development Act 28 of 2002 128 and the Prevention of Illegal Eviction 

from and Unlawful Occupation of Land Rights Act 19 of 1998 (PIE) 129 

could also fit into this category. 

Like SANDU, Bato Star is founded on the principle of ‘one system 

of law grounded in the Constitution’ 130 and intended to prevent the 

development of parallel systems of law and jurisprudence, one based 

on legislation and the other on the common law. Again like SANDU, 

Bato Star acknowledges the countermajoritarian dilemma in that it 

prevents the courts from circumventing democratically enacted 

legislation by reverting to a parallel, alternative system of rules, 131 

although it leaves room for constitutional review of democratic 

legislation. Like SANDU, the Bato Star principle enforces normative 

pluralism by preventing either democratic legislation or constitutional 

rights from usurping all adjudicative space — by forcefully holding the 

balance between the two, the principle gives politics a chance in so 

far as it enables decision-making based on consideration of the 

contestation between both impulses. 

SANDU, Bato Star and the constitutional review qualification, 

considered together, are signs of a subsidiarity approach: once 

legislation is enacted to give effect to a constitutional right, litigants 

must rely on the legislation when bringing action to protect that right; 

they may not rely directly on the constitutional provision or on the 

common law. Direct application of the constitutional provision 

remains possible to attack the constitutional validity of the 

legislation. Neither constitutional principle nor legislative inter- 

125 Fuel Retailers (n 25 above) par 37. 

126 Promulgated to give effect to s 121 of the 1993 Constitution, now authorised by s 

25(7). See Goedgelegen Tropical Fruits (n 108 above). 

127 Enacted to give effect to s 24(b): HTF Developers (n 104 above) par 24. Ss 31A 

and 32 of the Environment Conservation Act 73 of 1989 must also be interpreted 

to give effect to s 24: HTF Developers par 19. 

128 Arguably promulgated to give effect to s 24(b)(iii) and s 25(5), read with s 

25(4)(a). 

129 Enacted to give effect to s 26(3): Port Elizabeth Municipality (n 38 above) paras 

11-23; Occupiers of 51 Olivia Road, Berea Township, and 197 Main Street, 

Johannesburg v City of Johannesburg 2008 3 SA 208 (CC) par 16. S 4(1) of PIE 

explicitly overrides the common law. The CC declined in City of Johannesburg to 

decide whether PIE applied and what the relationship between PIE and s 26 is, 

but held that a local authority that evicts residents without meaningfully 

engaging with them acts at odds with the spirit and purpose of s 26(3). 

130 Pharmaceutical Manufacturers (n 37 above) par 44. 

131 JW Singer & JM Beermann ‘The social origins of property’ (1003) 6 Canadian 

Journal of Law & Jurisprudence 217-248 238 point out that the US Supreme Court 

majority in Lucas v South Carolina Coastal Council 505 US 1003 (1992) 

characterised ‘traditional common law [property] rules as somehow more worthy 

than politics advanced through contemporary legislative action’, threatening to 

derail public policies.


vention or common law tradition is allowed to decide cases 

automatically by its own force; each case is decided only when and on 

the strength of contestation between the conflicting impulses of 

stability and change; rights and constitution; tradition and reform. 

In so far as the cases discussed so far reach, these subsidiarity 

principles are attractive and relatively unproblematic. It is also 

reasonably simple to justify their application and outcome in the 

straightforward cases foreseen in these decisions. However, the 

majority of cases will not be nearly as simple, in that they will not fall 

within the rather narrow ambit foreseen in the decisions discussed so 

far. The difficulty is to decide how far the subsidiarity approach 

reaches; particularly, whether the subsidiarity principles apply only 

where legislation was enacted specifically to give effect to a 

constitutional right and intended to codify existing law; whether they 

also to apply to other legislation; and how they affect the common 

law. 

In the sections that follow I embroider upon the logic of the 

subsidiarity principles that I have identified in the SANDU and Bato 

Star decisions, without thereby pretending either that the 

embroidered or expanded subsidiarity principles reflect the view (or 

would meet with the approval) of the Constitutional Court or that 

they constitute doctrinal necessity. The line of argument in what 

follows is simply to trace what the extent and outlines of the SANDU 

and Bato Star subsidiarity principles could be if their logic is 

developed further, into areas not so far foreseen or discussed by the 

Constitutional Court. I do not present this analysis as a proposal of 

some kind of subsidiarity or deference principle either — in the final 

sections of the article, I still want to subject the outcome of this 

logical exploration to critical scrutiny in view of the normative 

framework set out earlier: if the logic of subsidiarity is embroidered 

upon to present a fuller picture than can at present be gleaned from 

case law, and should this logic be adopted and followed by the courts, 

and in so far as it might eventually establish legal or doctrinal 

principles, are or would they be justified? In asking this question, my 

point of departure will once again be that law (including subsidiarity 

principles), in so far as it is inevitable, will only be justified to the 

extent that it promotes normative pluralism but does not descend into 

either anarchy or reactionary formalism. In allowing room for 

normative pluralism but simultaneously avoiding anarchy, it must give 

politics a chance by allowing for real social contention and dissent. 

In this spirit, I rely on the perceived logic of subsidiarity in SANDU 

and Bato Star to determine whether, and how, the subsidiarity 

principles could apply with regard to other legislation (that it, pre- 

1994 legislation and partial or competing legislation) or to cases 

where there is no applicable legislation at all.


4.4 Application to other legislation 

SANDU was formulated for legislation particularly enacted to give 

effect to a right in the Bill of Rights; Bato Star was aimed at legislation 

that ‘covers the field’. 132 This raises the question whether these 

principles also apply to legislation that protects the right in question 

but that was neither enacted specifically with the constitutional 

provision in mind nor intended to codify existing law. 133 

The first indication that these principles could apply more widely 

is the justification for the subsidiarity principles: if the underlying 

norm is that legislation should not be sidelined in protecting a 

constitutional right out of respect for the transformative role of the 

democratic legislature, the principles should apply as widely as 

possible to legislation that gives effect to a right, including legislation 

not specifically promulgated with the constitutional provision in mind 

or not intended to codify existing law on the protection of that 

right. 134 

The second reason why flexible application of Bato Star makes 

sense is that it is often difficult to decide when legislation was 

intended to codify existing law. 135 On the one hand, some gaps are 

left on purpose: constitutional review of legislative, executive and 

judicial action is covered by neither s 33 nor PAJA and was made 

possible by a direct appeal to the constitutional legality principle. 136 

On the other hand, even legislation intended to ‘cover the field’ 

might leave gaps on purpose or by mistake: PAJA, to which SANDU and 

Bato Star clearly apply, covers the grounds for review of 

administrative action only imperfectly and the common law probably 

132 Stated in New Clicks (n 96 above) par 95 with reference to PAJA; see Sidumo (n 

104 above) par 90. See I Currie ‘What difference does the Promotion of 

Administrative Justice Act make to administrative law?’ 2006 Acta Juridica 325- 

351 328-334 (whether PAJA is codifying legislation). 

133 The Rental Housing Act 50 of 1999 gives effect to s 25(6) (insecure tenure), but is 

not intended to codify landlord-tenant law. 

134 Currie (n 132 above) 328 critiques the tendency to treat the notion that PAJA 

codified the law on administrative review as if it meant that PAJA ‘made no 

difference to the existing law’; in other words, we can apply existing law as if it 

was legitimised by PAJA. Extending Bato Star to partial legislation could force 

interpreters to acknowledge that the common law survives only in so far as it is 

consistent with and augment legislation. 

135 It is normally easier to identify legislation specifically intended to give effect to a 

provision, but even there some cases might be difficult. 

136 See C Hoexter Administrative law in South Africa (2007) 169-175. Review of these 

actions is possible on the basis of the legality principle in the Constitution: 

Fedsure (n 63 above) paras 33-45 (legislative action); President of the Republic of 

South Africa v South African Rugby Football Union 2000 1 SA 1 (CC) par 142 

(executive action); Nel v Le Roux NO 1996 3 SA 562 (CC) par 24 (judicial action). 

See C Hoexter ‘The principle of legality in South African administrative law’ 

(2004) 4 Macquarie Law Journal 165-185.


still applies in the gaps. 137 The subsidiarity principles should 

therefore not be restricted to implementing or codifying legislation 

and it should be accepted that gaps will occur in legislation. Judging 

on the justification argument, a gap in the field covered by legislation 

should mean that the subsidiarity principles do not apply in that 

specific area, allowing direct reference, on that specific point, to 

either the constitutional provision (SANDU) or the common law (Bato 

Star). I return to this point later. 

The third argument that supports wider application of the 

subsidiarity principles is the common law presumption that legislation 

does not amend the common law more than is necessary; 138 

legislation does amend the common law in so far as the intention to 

amend is either stated explicitly or implied clearly. 139 Considering 

that pre-constitutional law survives only to the extent that it is 

consistent with the Constitution 140 and that all legislation must be 

interpreted to promote the spirit, purport and objects of the Bill of 

Rights, 141 it must now be the case that all legislation, whether 

specifically intended to codify existing law or not, either gives effect 

to the rights in the Bill of Rights and promotes the spirit, purport and 

objects of the Bill of Rights or is unconstitutional. 142 It thus becomes 

axiomatic that all legislation must give effect to the rights in and 

promote the spirit, purport and objects of the Bill of Rights to a 

certain extent; the intention to amend the common law must 

therefore be implicit in all legislation and SANDU and Bato Star should 

apply to it. Lourens du Plessis 143 and Jacques de Ville 144 argue in the 

same direction when they state that the common law presumption 

either cannot apply or should be inverted in the context of the new 

Constitution. 145 Existing law must be interpreted to fit in with 

constitutional rights and therefore all legislation must be assumed to 

amend the common law in so far as is necessary to give effect to the 

Constitution, which is why the common law no longer exists as a 

137 Prior to PAJA the courts reviewed decisions taken by private bodies: Theron v Ring 

van Wellington van die NG Sendingkerk in Suid-Afrika 1976 2 SA 1 (A); Dawnlaan 

Beleggings (Edms) Bpk v Johannesburg Stock Exchange 1983 3 SA 344 (W); 

Johannesburg Stock Exchange v Witwatersrand Nigel Ltd 1988 3 SA 132 (A). It has 

been argued that the common law still applies in this situation: Hoexter 

Administrative law (n 136 above) 180-18; Currie & de Waal (n 30 above) 645. 

Other gaps are left by excluding grounds for review from s 6 PAJA, e g the 

common law rule against vagueness. 

138 Du Plessis (n 122 above) 177-181; de Ville (n 123 above) 170-176. 

139 Hoexter (n 103 above) 308: the principle that PAJA cannot be sidestepped by 

resorting directly to the common law is logical, ‘since statutes inevitably displace 

the common law’ (emphasis in original). Compare Rautenbach (n 105 above) 330. 

140 S 39(3). 

141 S 39(2). 

142 S 172. 

143 n 122 above, 179-181. 

144 n 123 above, 172, 66. 

145 Compare Pharmaceutical Manufacturers (n 37 above) par 45.


parallel, alternative system of rules and cannot be reverted to 

directly (Bato Star). 

To put it simply, direct application of the Constitution and the 

application and development of the common law should only come up 

in the absence of legislation. Some legislation will give effect to rights 

in the Bill of Rights more directly and some will affect existing law 

more explicitly and extensively, but in line with SANDU and Bato Star 

all legislation either fails constitutional scrutiny or triggers a 

subsidiarity principle according to which the right must primarily be 

protected via the legislation and not via direct application of the 

constitutional provision or the common law. In litigation aimed at 

protection of a constitutional right, the subsidiarity principles in 

SANDU and Bato Star postpone direct application of the constitutional 

provision and application or development of the common law until it 

is clear that legislation does not cover the case. 

Once a flexible reading of the subsidiarity principles is accepted, 

the next question is how these principles should apply to partial 

legislation; in other words, legislation that gives partial effect to a 

constitutional right but does not attempt or manage to ‘cover the 

field’ as far as protection of the right is concerned. This category is 

difficult to define because it is impossible to say how partial 

legislation must be to disqualify it from being comprehensive, as the 

example of PAJA in the previous section proves. If the subsidiarity 

principles must apply to legislation that ‘covers the field’ as 

imperfectly as PAJA does they should arguably apply to all partial 

legislation when and in so far as the legislation gives effect to a 

constitutional right. An example illustrates the point. 

In Magudu Game Co 146 the KwaZulu-Natal High Court argued that 

the common law principles regarding ownership of wild game should 

be developed, in line with s 39(2), to provide for the exigencies of 

commercial game farming. However, the state had already enacted 

the Game Theft Act 105 of 1991 to protect the rights of commercial 

game farmers. The Act does not codify the law on wild game — outside 

of commercial game farming the common law remains unchanged. 

However, the Act does regulate commercial game farming, and the 

game farmer in this case could not rely on it purely because he failed 

to comply with registration requirements. According to the expanded 

version of Bato Star, development of the common law should not be 

an option in litigation to protect the right in so far as the legislature 

has enacted legislation to give effect to it, which in this case it had. 

146 Magudu Game Co (Pty) Ltd v Mathenjwa and others NNO 2008 2 All SA 338 (N).


SANDU and Bato Star must therefore be applied flexibly; even 

supposedly codifying legislation will leave gaps and, if the 

Constitution or the common law provides for such a gap, the next step 

could be to turn to the Constitution or the common law to fill that gap 

rather than challenge the legislation. Shifting the problem up to the 

Constitution or down to the common law before reverting to a 

constitutional challenge makes sense if a gap in the legislation means 

that the specific aspect is not covered by the legislative scheme, 

which means that the subsidiarity principles do not apply and the gap 

can be filled by application of constitutional provisions or the common 

law, as illustrated by constitutional review of legislative, executive or 

judicial acts 147 and by judicial review of administrative action in 

cases where private bodies exercise public power. 148 

It should have become clear by now that application of the 

subsidiarity principles in SANDU and in Bato Star respectively to 

partial legislation would work out differently. Applying SANDU to 

partial legislation involves the question whether litigants should be 

allowed to appeal directly to the constitutional provision to protect a 

right if legislation that does provide for its protection but was not 

made specifically to give effect to the right does not provide the 

required remedy in a particular case. The extension would probably 

work better with pre-constitutional than with new legislation, but 

generally there is no reason why such a relaxation of SANDU should be 

impossible in principle. Two qualifications should probably apply, 

though. Firstly, the actual purpose and scope of partial legislation 

should be decisive in deciding whether a direct appeal to the 

constitutional provision would be justified if the legislation fails to 

provide a remedy. Secondly, to prevent the development of parallel 

systems of law, this relaxation of SANDU should arguably only be 

possible if the gap cannot be filled by applying or developing the 

common law. In other words, as long as a gap in legislation can be 

filled by application of the common law, a direct appeal to the 

Constitution should not be possible in an action to protect the right 

— this is why the Supreme Court of Appeal’s decision in Tswelopele 149 

to leave the common law ‘untouched’ and to create a new 

constitutional remedy was arguably a mistake. The result of 

Tswelopele is two parallel remedies for ante omnia restoration of 

spoliated possession, based on the Constitution and the common law, 

the only difference being that the constitutional remedy permits a 

restoration order that includes the use of replacement materials. 

Developing the common law, for a clearly circumscribed range of 

cases, would have vindicated and celebrated judicial victories over 

147 n 136 above. 

148 n 137 above. 

149 Tswelopele Non-Profit Organisation v City of Tshwane Metropolitan Municipality 

2007 6 SA 511 (SCA).


apartheid injustice in decisions such as Fredericks 150 and could have 

made a contribution to the rehabilitation of the common law of 

property, which did not cover itself in glory during apartheid, by 

demonstrating how it can support social justice. 151 

Since Bato Star purely prevents direct appeals to the common law 

by circumventing legislation, it is easier to argue that the principle 

should be qualified so that the common law still applies (and can be 

developed) in cases where partial legislation does not in fact give 

complete, covering effect to a constitutional right. However, a 

qualification is once again necessary in line with s 39: the common law 

can only fill gaps in legislation if and to the extent that it is not only 

consistent with the Bill of Rights and with the legislative scheme but, 

in so far as the legislation leaves a gap, offers the possibility (where 

necessary through development) of augmenting the legislation, 

keeping in mind the constitution-enhancing interpretive rule in s 

39(2). The governing principle should not be that the common law 

survives where it remains unaffected by constitutional or legislative 

provisions, but rather that the common law survives only if and in so 

far as it is consistent with the Bill of Rights, consistent with existing 

legislation, and capable of complementing the legislation in giving 

effect to constitutional rights, either as it stands or through being 

developed for the purpose. 

The Magudu Game Co 152 example suggests that SANDU and Bato 

Star could also apply to pre-constitutional legislation. According to s 

39, pre-constitutional law only survives in so far as it does not conflict 

with constitutional provisions and can be interpreted in conformity 

with constitutional rights and values. Accordingly, the 1991 Act 

protects ownership of game to give effect to the property clause (s 

25) in a specific area, namely commercial game farming. A similar 

argument applies to all pre-constitutional legislation that can survive 

the new Constitution and therefore SANDU and Bato Star should apply 

to it. However, although it has to be interpreted in line with the 

Constitution, pre-constitutional legislation cannot be assumed to 

reflect the legislature’s effort to comply with constitutional 

obligations. The Expropriation Act 63 of 1975, being the only general 

expropriation legislation currently in force, predates the Constitution 

and must therefore be interpreted in line with s 25(2) and (3). 153 

Legislation is required to give effect to the expropriation and 

compensation provisions in s 25 and since there is nothing else, the 

150 Fredericks v Stellenbosch Divisional Council 1977 3 SA 113 (C). 

151 See AJ van der Walt ‘Developing the law on unlawful spoliation and squatting’ 

(2008) 125 South African Law Journal 24-36. 

152 n 146 above. 

153 A Draft Expropriation Bill 2008 was tabled in the Portfolio Committee on Public 

Works on 26 March 2008: Parliamentary Monitoring Group’s website at 

www.pmg.org.za. This Bill was recalled for further consideration in August 2008.


1975 Act applies to all expropriation, but it was drafted in another era 

and not intended for application in the remedial and transformative 

context of s 25, and therefore the Act must be read through the 

corrective lens of s 39(2). SANDU and Bato Star should apply to the 

1975 Act, but in applying it the courts must keep in mind that it 

embodies the policies of another time and a different government and 

therefore it requires conscious and comprehensive re-interpretation 

to ensure that it gives effect to current constitutional rights and 

values. The majoritarian consideration that justifies SANDU, namely 

that the legislature’s role in the transformation process must be 

respected and that it must be given full effect, applies only indirectly 

to pre-constitutional legislation: it can only be assumed to embody 

the legislature’s effort to give effect to constitutional rights in so far 

as it allows constitution-promoting interpretation according to 

s 39(2), but there is always the possibility that parts of the old 

legislation still reflect the policies of a disgraced pre-constitutional 

legislature and then a constitutional challenge might be required. Up 

to that point, the subsidiarity principles apply in the form of the 

obligation to read the legislation in conformity with the Constitution, 

as s 39(2) requires. 

4.5 Competing and complementary legislation 

A further problem created by flexible application of SANDU and Bato 

Star is that several pieces of legislation, including supposedly 

codifying legislation like PAJA and the LRA and partial, even preconstitutional 

legislation, sometimes apply to the same dispute 

simultaneously. 154 If more than one statute compete with or 

complement each other, the subsidiarity principle should be that 

competing or complementary legislation must be applied to optimally 

give effect to the Bill of Rights and to promote the spirit, purport and 

objects of the Bill. 155 As a starting point this will require purposive 

interpretation that aligns all the applicable laws with the 

154 Eg SANDU (n 104 above); Sidumo (n 104 above); Chirwa (n 104 above); Fuel 

Retailers (n 25 above). 

155 In Nakin v MEC, Department of Education, Eastern Cape Province 2008 2 All SA 

599 (EC) paras 49-51 Froneman J argues that courts should not be forced to 

‘compartmentalise’, by conceptual exclusion, the possibilities for adjudicating 

labour disputes, namely common law of contract, labour relations and 

administrative justice, since a conceptual selection between the alternatives 

could rob the applicant of an effective remedy because the route selected (e g 

PAJA) might not offer the remedy (e g payment of money) the applicant asks for. 

If this suggests that SANDU / Bato Star should not apply and that litigants (or the 

courts) should have a free choice, it conflicts with normative arguments in favour 

of subsidiarity. However, in so far as it points out that a meaningful remedy is 

important the problem could be overcome by applying subsidiarity (e g PAJA 

applies) but then, if PAJA does not offer a remedy, cascade the issue on remedy 

down the subsidiarity hierarchy, ie by either reverting to the common law, if that 

offers a suitable remedy, or to constitutional review, possibly opening up space 

for reading in a remedy.


constitutional scheme. Some remaining problems could be solved by 

focusing on the separate fields of application of each law 156 or by 

identifying their position in a subsidiarity grid. It has, for example, 

been proposed that PAJA embodies general administrative law while 

other statutes grant particular administrative powers, thus indicating 

their relative status: the LRA is particular legislation with regard to 

labour disputes and therefore its provisions with regard to 

administrative action supersede but are ‘suffused’ by the general 

reasonableness principles of PAJA, which is general legislation. 157 A 

similar hierarchical relationship was construed between NEMA and the 

Environment Conservation Act 73 of 1989. 158 

When legislation limits one right in the process of giving effect to 

another, the affected person can protect her right with a direct 

attack on the constitutional validity of the offending legislation; 

SANDU allows a direct appeal to the constitutional provision. When 

constitutional rights provisions come into conflict, the Constitutional 

Court supports an optimising interpretation strategy. 159 In other cases 

s 39(2) requires the courts to select the interpretation that avoids the 

conflict or promotes the spirit, purport and objects of the Bill of 

Rights. In these cases there does not seem to be a reason to depart 

from the subsidiarity principle that protection of a constitutional right 

should primarily be based on legislation giving effect to that right and 

not on the constitutional right itself or on the common law, except in 

so far as the constitutional validity of the legislation or the common 

law is attacked (in which case the constitutional provision applies 

directly) or in so far as the legislation does not cover the field (in 

which case a direct appeal to the Constitution or the common law 

could be possible). 

Several decisions from the 2007 term concerned regulatory 

legislation that limits a constitutional right, without the legislation 

being intended to give effect to that or any other right. Mohunram 160 

is the latest decision 161 in which the proportionality of civil forfeiture 

of instrumentalities of crime in terms of the Prevention of Organised 

Crime Act 121 of 1998 (POCA) was contested, without attacking the 

156 Land reform laws enacted under s 25(6) to improve security of tenure, e g the 

Land Reform (Labour Tenants) Act 3 of 1996, the Extension of Security of Tenure 

Act 62 of 1997 and the Rental Housing Act 50 of 1999, each has its own field of 

application. 

157 Chirwa (n 104 above) par 23 confirmed Bato Star (n 63 above) paras 59, 149; I 

Currie & J Klaaren The Promotion of Administrative Justice Act benchbook (2001) 

par 1.24; Currie & de Waal (n 30 above) 643 fn 4. Sidumo (n 104 above) paras 104, 

110 states that the LRA is suffused by the constitutional reasonableness principle 

but in fact reads it via PAJA. 

158 HTF Developers (n 104) par 24; Fuel Retailers (n 25). 

159 See eg Christian Education (n 33 above) par 42. 

160 Mohunram v National Director of Public Prosecutions (Law Review Project as 

Amicus Curiae) 2007 4 SA 222 (CC). 

161 See Prophet v National Director of Public Prosecutions 2007 6 SA 169 (CC).


constitutional validity of the Act. It has been accepted that civil 

forfeiture (particularly beyond the ambit of organised crime) could 

result in disproportionate and arbitrary deprivation of property; 

hence the Constitutional Court devised a definition of instrumentalities 

of crime 162 and a proportionality test to determine whether 

forfeiture is justified. 163 Both tests are aligned with the nonarbitrariness 

test in FNB 164 to ensure that civil forfeiture orders under 

POCA would not offend the protection of property against arbitrary 

deprivation in s 25(1). The proportionality test employed in these 

cases to interpret the legislation in a constitution-complying way 

confirms SANDU. 

In two other forfeiture cases, Armbruster 165 and Van der 

Merwe, 166 the facts were similar — the affected persons had large 

sums of foreign currency in their possession when boarding an 

international flight; the currency was seized by customs officials and 

declared forfeit; the affected persons attempted to reclaim the 

money. In Armbruster the applicant launched a constitutional attack 

against the regulations with an appeal to s 34; in Van der Merwe he 

appealed to the common law. SANDU allows a direct appeal to a 

constitutional provision in an attack on the validity of legislation, but 

according to Bato Star a direct appeal to a common law remedy 

should only be available if legislation does not ‘cover’ that particular 

aspect and the common law is capable of interpretation that conforms 

with the Constitution and complements the legislation. 

In Armbruster the Constitutional Court held that the connection 

between the purpose of the forfeiture, the property and the person 

deprived ‘could hardly be closer’; that there is sufficient reason for 

the deprivation and that forfeiture would not be arbitrary. 167 

Forfeiture could produce ‘arbitrarily harsh consequences’ but the 

162 Prophet (n 161 above) was applied in Mohunram (n 160 above) par 44: 

instrumentalities had to be given a wide meaning, but the property must play a 

reasonably direct role in commission of the offence. This test is based on the nonarbitrariness 

standard in First National Bank of SA Ltd t/a Wesbank v 

Commissioner, South African Revenue Service; First National Bank of SA Ltd t/a 

Wesbank v Minister of Finance 2002 4 SA 768 (CC) paras 51, 100. 

163 Prophet (n 161 above) par 65; Mohunram (n 160 above) paras 64, 145-147. In 

Mohunram the majority decided that the forfeiture would be disproportionate 

because part of the immovable property was used for legal purposes and because 

the owner had already been prosecuted, convicted and fined. 

164 The non-arbitrariness standard was set out in First National Bank (n 162 above) 

paras 51, 100. 

165 Armbruster v Minister of Finance 2007 6 SA 550 (CC). The Treasury declared the 

foreign currency forfeit in terms of the Exchange Control Regulations under s 9 of 

the Currency and Exchanges Act 9 of 1933. 

166 Van der Merwe v Inspector Taylor 2008 1 SA 1 (CC). Various grounds were raised, 

ranging from the Criminal Procedure Act 51 of 1977 to the Income Tax Act 58 of 

1962; before the CC consensus was that the seizure was authorised by s 20 of the 

Criminal Procedure Act. 

167 Armbruster (n 165 above) par 71. The nexus remarks are inspired by First 

National Bank (n 162 above) paras 51, 100.


regulations grant an opportunity for representations and a discretion 

to return the currency ‘to ameliorate undue hardship or injustice that 

might be perpetrated on the person affected’; therefore the 

regulations do not allow for arbitrary deprivation. 168 The Court also 

rejected the attack based on s 34, pointing out that the reasons why 

legislation was struck down in Zondi and Chief Lesapo, namely a 

particular historical and social context and complete lack of judicial 

control, were absent in this case. 169 

In Van der Merwe, the applicant elected to claim the money on 

the basis of the common law rei vindicatio, arguing that foreign 

currency bought with funds from his banking account was his 

property. The High Court decided that some of the currency did not 

belong to him because he held it for someone else. The Constitutional 

Court was divided; the majority held that he proved ownership of the 

amount bought and held for himself, but O’Regan J would have 

dismissed leave to appeal because the issues have not been properly 

ventilated. 170 One has to agree that the Court had not seen enough 

evidence or heard enough argument on whether the rei vindicatio is 

at all feasible when cash is reclaimed; 171 when ownership vests in 

cash withdrawn from a bank; and whether the applicant’s state of 

mind with regard to ownership of currency could have had any effect. 

In terms of Bato Star the primary question in Van der Merwe should 

have been whether reclamation of the money was regulated by the 

authorising legislation — if it was, a direct appeal to the common law 

remedy should have been precluded. 

When there is no legislation to trigger SANDU or Bato Star, the tug 

of war between the Constitution and the common law is most intense. 

The question is then whether the civil courts should be allowed to 

withdraw into the comfort zone of the common law, where the reach 

of democratic reform is weak and where the influence of the 

Constitution could be reduced to that of an ally in entrenching vested 

rights, privilege and power against democratic interference. The most 

difficult issues of application feature in this area, with the focus 

ultimately on the question whether (and how far) the independence 

of private law is affected by the supremacy of the Constitution. 

168 Armbruster (n 165 above) par 80, relying on Dawood v Minister of Home Affairs; 

Shalabi v Minister of Home Affairs; Thomas v Minister of Home Affairs 2000 3 SA 

936 (CC) para 53. 

169 Armbruster (n 165 above) paras 59-60, referring to Chief Lesapo v North West 

Agricultural Bank 2000 1 SA 409 (CC), Zondi (n 112 above), Metcash Trading Ltd v 

Commissioner, South African Revenue Service 2001 1 SA 1109 (CC). 

170 Van der Merwe (n 166 above) par 103. 

171 Woodhead Plant & Co v Gunn (1894) 11 SC 4: the currency status of money is 

incompatible with the right of anyone to follow stolen money into the hands of an 

honest receiver for value, whether it had been mixed up with other money or not.


5 Constitutional supremacy vs independence of 

private law 172 

Mir scheint es sich [bei der Konzeption einer ‘Grundrechtsfreiheit des 

Privatrechts’] um eine Art Münchhausentheorem zu handeln, das den 

Privatrechtlern die Möglichkeit verschaffen soll, sich am eigenen Schopf 

aus dem Sumpf des Verfassungsrechts herauszuziehen. 173 

5.1 SANDU and Bato Star in the absence of legislation 

When no legislation applies to the protection of a constitutional right, 

the logic of the SANDU subsidiarity principle would imply that a 

litigant who wants to protect that right must rely on the common law 

and may not rely directly on the constitutional right, unless she 

challenges the constitutional validity of the common law. Mhlungu 

and Zantsi were no doubt intended to justify exactly such a 

subsidiarity principle to prevent every private quarrel being blown up 

into a constitutional issue when it could be dealt with in terms of 

common law. In terms of s 39(2), the obligation to have first resort to 

the common law includes development of the common law to align it 

with the spirit, purport and objects of the Bill of Rights. However, 

once it is clear that the constitutional right cannot be given effect to, 

according to the spirit, purport and objects of the Bill of Rights, by 

either application or development of the common law a litigant can 

resort directly to the constitutional provision, 174 either to launch a 

constitutional attack against a rule or institution of the common law 

that limits or is in conflict with the constitutional right 175 or to craft 

a special constitutional remedy. 176 

It can be argued that the legislature, in so far as it leaves the 

common law unamended by legislation, accepts that the common 

law, for the moment being, sufficiently gives effect (or can be 

developed to give effect) to the rights in the Bill of Rights — if the 

legislature is dissatisfied with the common law it can change it. Out 

of respect for the legislature’s democratic role, protection of the 

constitutional rights should, therefore, in the absence of legislation 

begin with the common law — a direct appeal to the Constitution 

should be reserved for cases where it is impossible to interpret the 

common law in a constitution-compliant way. This is possibly what 

172 Based on the title of M Ruffert Vorrang der Verfassung und Eigenständigkeit des 

Privatrechts (2001). I am indebted to Friedrich Schoch for bringing it to my 

attention and to him and Michael Sachs for explaining German theory to me. 

173 CW Canaris Grundrechte und Privatrecht (1999) 18. 

174 Masiya (n 88 above) par 27. 

175 The CC declared the common law crime of sodomy unconstitutional: Gay and 

Lesbian (n 28 above). 

176 In line with s 172.


inspired the unfortunately phrased remark that the common law 

should be allowed to develop incrementally so as not to interfere with 

legislative authority. 177 

The justification for this residuarity principle is similar to that of 

SANDU: if litigants are allowed to choose freely between 

constitutional provisions and common law remedies as if they were 

independent, parallel systems of law, the common law will never be 

‘brought into’ the new constitutional dispensation. This is what 

happened in Tswelopele: 178 crafting a constitutional remedy and 

leaving the common law ‘untouched’ created two parallel remedies 

instead of developing the common law in line with constitutional 

values. The goal of ensuring one system of law under the guidance of 

the Constitution is promoted not only by preventing the common law 

to develop as a parallel system (Bato Star), but also by preventing 

development of a parallel system of unnecessary constitutional 

remedies. Leaving some aspect of the common law intact and crafting 

a new constitutional remedy should be justifiable only on 

constitutional grounds. 

In was argued in the previous section that SANDU and Bato Star 

apply to legislation regardless of horizontality; 179 the same should 

hold in the absence of legislation. Direct application and development 

of the common law are broader than and should not be confused with 

or collapsed into horizontal application; the former two will come up 

in vertical and horizontal relationships. 180 

5.2 Direct vs indirect application 

The most intractable question when no legislation exists is whether 

the constitutional provisions apply directly or whether they should be 

mediated through the common law. This issue came up in three 2007 

decisions. 

177 Zantsi (n 71 above) par 5; Masiya (n 88 above) par 31. Roux (n 46 above) 487 

notes, with reference to RG Teitel Transitional justice (2000) 24, that judges may 

have to do more work in transitional or transforming societies and leave less for 

the legislature. 

178 Tswelopele (n 149 above). 

179 The principle finds support in s 8(3)(a), which applies particularly to horizontal 

relationships: in applying the provisions of the Bill of Rights to natural and juristic 

persons in terms of subs 8(2) a court, ‘in order to give effect to a right in the Bill, 

must apply, or where necessary, develop, the common law to the extent that 

legislation does not give effect to that rights’ (emphasis added). I pay no direct 

attention to development of the customary law because none of the 2007 

decisions dealt with it. 

180 Tswelopele (n 149 above); Masiya (n 88 above).


In Masiya 181 the question was whether the common law definition 

of rape should be developed to include anal penetration. 182 The 

Constitutional Court decided that the common law definition was not 

unconstitutional because it criminalised morally and socially 

unacceptable behaviour. However, the definition fell short of the 

spirit, purport and objects of the Bill of Rights and had to be 

expanded. 183 The Court distinguished between its roles in testing 

whether legislation is consistent with the Constitution and developing 

the common law, adding that the courts should not ‘appropriate the 

Legislature’s role in law reform’ by developing the common law in a 

way ‘closer to codification than incremental, fact-driven 

development’. 184 This supports the subsidiarity principle that direct 

application (validity challenge) should only follow once it is clear that 

the common law cannot be developed appropriately. Accordingly, the 

case was decided as an opportunity to develop the common law rather 

than a constitutional validity challenge. The majority decided to 

develop the common law only as required by the facts of this case, to 

include non-consensual anal penetration of a female, 185 but in a 

dissenting judgment Langa CJ opined that the development should 

have included anal rape of men. 186 

Woolman 187 criticises Masiya because it ‘never truly considers the 

direct application of the substantive provisions of the Bill of Rights to 

the challenged common-law definition of rape’; instead, the ‘entire 

analysis of the common-law rule takes place within the rubric of s 

39(2) and in terms of indirect application.’ Woolman argues that, in 

181 n 88 above. 

182 The regional court decided that the common law definition was underinclusive 

and extended it to non-consensual anal penetration of any person, but it did not 

have the power to develop the common law: Masiya (n 88 above) par 66. 

Magistrates’ courts are prevented from pronouncing on the validity of any law: s 

110 of the Magistrates’ Courts Act 32 of 1944. The High Court endorsed the a quo 

decision and it came before the CC for confirmation under s 172(2)(a). 

183 Masiya (n 88 above) paras 27, 32, 70. Masiya paras 66, 69 underlines the point 

that s 39(2) has a restricted purpose: the power to develop the common law is 

granted to the CC, the SCA and the High Courts by s 173 and not in s 39(2), which 

merely instructs the courts, when developing the common law or customary law, 

to do it so as to promote the spirit, purport and objects of the Bill of Rights — the 

section concerns the how of constitutional development of the common law but 

not the when. For similar reasons, s 8(3) should not be seen as the authorising 

provision for a general obligation or power to develop the common law either; 

the subsection pertains to cases of horizontal application, as is evident from the 

introductory phrase ‘[i]n applying the provisions of the Bill of Rights to natural 

and juristic persons in terms of subsection (2), ...’ (emphasis added). In spelling 

out the direction of development s 39(2) indicates when the common law might 

require development, but it is not the authorising provision. 

184 Masiya (n 88) par 31. 

185 Masiya (n 88) par 45. 

186 Masiya (n 88) par 75. Sachs J concurred in the dissenting judgment. 

187 n 89 above, 768.


view of the language of ss 8(1) 188 and the decision in Khumalo, 189 the 

substantive provisions in the Bill of Rights should have been applied 

directly instead of via the development of the common law. The point 

in which Woolman differs from Masiya is not easily identifiable; his 

statement that direct application ‘takes the rights and freedoms, and 

the general rules derived from them, as our point of departure for 

determining whether law or conduct is invalid’ echoes the SANDU 

principle almost exactly, 190 as does his view that indirect application 

(protection of the right) does not permit a finding of invalidity. 191 The 

problem is that he describes indirect application as ‘a mode of 

analysis that neither specifies whether a particular right demands 

vindication nor permits a finding of invalidity.’ 192 Woolman’s view 

that indirect application does not tell us whether the right requires 

vindication is based on the narrow assumption that vindication of 

constitutional rights primarily takes place via direct application. In 

terms of SANDU logic, constitutional rights can also be vindicated 

indirectly, via constitution-compliant interpretation of the common 

law. According to Masiya, indirect protection should precede direct 

constitutional challenges as a matter of subsidiarity, out of respect 

for democratic processes, while SANDU declares direct appeals to the 

constitutional provisions impermissible when protection of the right is 

in issue. 193 

Woolman therefore differs from the principle laid down in SANDU 

in that he claims the major part of constitutional litigation for direct 

application and assigns indirect application to a vague and apparently 

insignificant area of ‘law and conduct not engaged by any of the 

specific provisions set out in Chapter 2’, ensuring that this part of all 

law is also brought ‘into line with the “spirit, purport and objects of 

188 

Although s 8(1) provides that the Bill of Rights applies to all law and that it binds 

the legislature, the executive, the judiciary and all organs of state, it does not 

say that the Bill of Rights applies directly or horizontally to all law. See to much 

the same effect Khumalo (n 73 above) paras 31-33. Woolman (n 45 above) 44 

disagrees, arguing that s 8(1) alone makes all law subject to the direct 

application of the Bill of Rights. 

189 

In Khumalo (n 73 above) the CC indicated that direct application was possible 

under the 1996 Constitution. 

190 Woolman (n 89 above) 769. As far as legislative, executive and judicial conduct is 

concerned Woolman reflects the approach of the courts: these forms of conduct 

are excluded from PAJA by the definition of ‘administrative action’ but review is 

possible on the basis of legality: n 136 above. However, as far as conduct by 

organs of state generally is concerned, at least some of it is administrative action 

covered by PAJA, in which case Woolman’s formulation does not reflect SANDU in 

that review of this conduct takes place via PAJA and not directly. 

191 

All law, including legislation and the common law, should only be declared 

unconstitutional once it has been established that it is impossible to find a 

constitutionally consistent interpretation: Masiya (n 88 above) par 27. However, 

rules of the common law can be declared unconstitutional when it is impossible to 

interpret or develop it according to values of the Constitution: Gay and Lesbian (n 

28 above). 

192 

Woolman (n 89 above) 769 (emphasis added). 

193 The legality principle case law is explained by this formulation; n 136.


the Bill of Rights” and the “objective, normative value system” made 

manifest in the text of the Constitution as a whole’. 194 By contrast, 

the residuarity principles establish a rational link between direct 

application and constitutional remedies or validity challenges on the 

one hand and indirect application and protective litigation on the 

other. As was argued earlier, this link is justifiable in view of the 

normative roles it assigns to democratic processes and judicial review 

respectively. Moreover, it makes sense if considered from the 

perspective of remedies: direct application is aimed at reviewing the 

constitutional validity of legislation or the common law and will at 

least sometimes result in declarations of invalidity, which might not 

be what a litigant was after when seeking to protect her right. In some 

cases direct application can admittedly produce a special 

constitutional remedy and therefore the subsidiarity principles keep 

the possibility of direct application open for that purpose, but based 

on the respect-for-democracy norm the subsidiarity approach 

restricts this to a last resort solution rather than an open option that 

exists parallel to statutory or common law protection. 

By shifting the focus from constitutional validity challenge to 

protective litigation, where the point is to develop the common law 

in line with s 39(2), Masiya is consistent with SANDU. The 

Constitutional Court decided that the correct approach was to expand 

the common law definition by developing it in line with s 39(2), in this 

case to include at least non-consensual anal penetration of 

females. 195 The minority decision shows that one can disagree with 

this narrow development of the definition and still agree that indirect 

development of the common law was the right approach. The 

cautious, restricted development undertaken by the majority cannot 

be blamed on their choice not to apply the constitutional provisions 

directly but to develop the common law; much rather, it results from 

earlier decisions in which it was said that the avoidance principle 

(subsidiarity) implies deciding just what is necessary for the facts of 

the case and no more. 196 This is an exaggerated interpretation of 

subsidiarity and hardly defensible in view of constitutional provisions 

and decisions to the effect that all law is governed by the 

Constitution. The subsidiarity principle in Mhlungu and Zantsi could 

be read less restrictively (in view of SANDU) as saying that a court 

should not protect a constitutional right by way of a direct validity 

attack or by way of a direct constitutional remedy before considering 

whether the legislation or common law in question could be 

interpreted in a constitution-conforming and –confirming way. 

194 Woolman (n 89 above) 769. 

195 Masiya (n 88 above) par 27. 

196 Especially Zantsi (n 71 above) par 8.


Apart from an unnecessarily narrow view of subsidiarity or the 

avoidance principle, the Constitutional Court should also not have 

allowed itself to be pulled into common law talk about incremental 

development of the common law. Constitutional transformation will 

sometimes require large developments and changes. The warning in 

Mhlungu and Zantsi against undermining the separation of powers is 

important, but the subsidiarity principles in SANDU and Bato Star 

promote respect for the role of the legislature, without confusing that 

with respect for the common law logic of incremental development. 

Respect for democratic processes should not prevent the courts from 

rooting out remnants of tradition such as discrimination or inequality 

in conflict with the new constitutional dispensation; anal rape will be 

unconstitutional regardless of the sex or gender of the victim and 

hence the courts need not shy away from larger development of the 

common law out of respect for the legislature. 

The decision in NM 197 also turned on a choice whether to develop 

the common law; Woolman 198 again criticised it for failure to apply 

the constitutional provisions directly. The applicants claimed that the 

respondents had violated their rights to dignity and privacy by 

publishing their names and HIV status without permission. The case 

could support direct horizontal application of ss 10 and 14, but the 

applicants claimed on the basis of the common law actio iniuriarum 

and accordingly, in terms of residuarity, their claim must be 

considered on the basis of s 39(2) first. Initially, the question is 

therefore whether the common law provides a remedy as it stands 

and, if it doesn’t, whether it can be developed. The applicants 

brought the matter in the form of a constitutional case because they 

argued that the common law would have to be developed; the 

Constitutional Court decided that the case ‘involves a nuanced and 

sensitive approach to balancing the interests of the media, in 

advocating freedom of expression, privacy and dignity of the 

applicants irrespective of whether it is based on the constitutional 

law or the common law.’ 199 In other words, even though the claim is 

brought under the common law and the application and development 

of the common law are the main issues, balancing of the 

constitutional rights raises a constitutional issue. This reflects the 

meaning of s 8(1) according to SANDU subsidiarity logic: the 

Constitution applies to all law, but not necessarily directly. 

The decision eventually turned on the majority’s factual finding 

that the respondents were aware that the applicants had not given 

permission for the publication of their names, followed by the legal 

decision that this was sufficient to establish intent for the actio 

197 n 87 above. 

198 n 89 above, 781-783. 

199 NM (n 87 above) par 31.


iniuriarum and that the requirements for the action were thus met. 

On the factual finding it was unnecessary to develop the common law. 

The minority, who disagreed with the majority’s factual finding, 

would nevertheless have developed the common law so that 

negligence could also satisfy the requirements. 200 In either case, 

direct application of the constitutional provisions remains uncalled 

for under the subsidiarity principles. 

Woolman 201 criticises Barkhuizen 202 for the same reason as 

Masiya and NM, namely that the Constitutional Court should have 

applied the constitutional provisions directly. Although it is difficult 

to tell whether Barkhuizen was decided in line with SANDU, the case 

was clearly brought as a direct constitutional challenge but not 

decided as one. The constitutional challenge was brought against a 

time-limitation clause in a short-term insurance contract; the 

applicant argued that the clause violated his s 34 right of access to 

court. The High Court upheld the claim but the Supreme Court of 

Appeal dismissed it, arguing that the Constitution did not prevent 

time-limitation clauses in contracts entered into freely and 

voluntarily and that there was inadequate evidence in this case to 

indicate that the contract was not entered into freely and voluntarily. 

The Constitutional Court found that the 90 day time limitation was not 

manifestly unreasonable on the face of it, nor was it unfair, because 

there was no evidence that the contract was not concluded freely 

between parties in equal bargaining positions. Since the applicant did 

not furnish reasons for non-compliance it would not be unjust to 

enforce the time bar against him. 203 

Testing a contractual provision directly against a provision in the 

Bill of Rights (s 34) amounts to horizontal application under s 8(2)- 

(3). 204 The Constitutional Court was uncomfortable with horizontal 

application, noting that the contractual time limitation reveals no law 

of general application. The Supreme Court of Appeal identified the 

underlying principle pacta sunt servanda as the law of general 

application, but the Constitutional Court dismissed this option, 

pointing out that the contractual term was challenged, not the 

general principle. 205 Enforcement of an allegedly unfair contractual 

term admittedly is not law of general application, but it is conduct 

that can be declared unconstitutional in terms of s 172(2). The fact 

200 NM (n 87) paras 44-47 57. Langa CJ and O’Regan J agreed that the journalist and 

publisher might have been negligent, but not that intent had been proved: paras 

111, 125. See Woolman (n 89) 781. 

201 n 89 above, 772-781. 

202 n 86 above. 

203 Barkhuizen (n 86 above) paras 63 66-67 84 86. 

204 Barkhuizen (n 86 above) par 23. It is untrue that the Court yet had to consider 

this issue: Khumalo (n 73 above). 

205 Barkhuizen (n 86 above) paras 24 26.


that conduct does not establish law of general application merely 

means that it cannot limit a constitutional right and cannot be saved 

by the general limitation clause, which is not surprising — the 

limitation clause is intended to give the state the opportunity to save 

otherwise unconstitutional law. However, conduct can infringe upon 

a right and can therefore be the focus of a constitutional validity 

attack. 206 It was therefore conceivable that this case could have been 

decided as a direct constitutional attack on allegedly unconstitutional 

private conduct. If the court should find, in such a case, that 

enforcement of the contract would indeed infringe s 34 or any other 

constitutional right, it would have to declare the conduct invalid. At 

the same time, private conduct that infringes a private right should, 

in most cases, be regulated by private law unless there is a 

constitutional reason for deciding the conflict on a constitutional 

basis. 

Barkhuizen could also have been decided purely on the basis of 

private law — time limitation clauses in private contracts are 

generally invalid if they are against public policy. The question is 

therefore whether there was a reason to decide the case on 

constitutional grounds. One such a reason could have been the desire 

to set a constitutional standard for contractual time limitation clauses 

that limit the right of access to courts. In that case the format of a 

direct constitutional attack on validity of the conduct would have 

been suitable. Another possibility could have been the need to 

develop the common law principle that gives credence to time 

limitation clauses, namely that contracts entered into freely between 

equal parties should be enforced — pacta sunt servanda. Such an 

approach would arguably have been justified if the Constitutional 

Court thought that the common law gave too much weight to the 

sanctity of contract principle and too little to justice, and that the 

common law should be developed in terms of s 39(2) to rectify the 

balance. However, this would have required the case to be 

transformed from a direct validity challenge to an indirect 

development of the common law protection case, as happened in NM 

and in Masiya. In the latter instance one would expect the court to 

test the relevant common law principle against central constitutional 

values and explain how it is to be amended in view of those values. 

In the event, the Constitutional Court did neither of the above, 

preferring to decide the case in another way altogether, namely to 

determine whether the contractual clause was contrary to the values 

underlying the constitutional democracy and contrary to public 

policy. 207 The Court pointed out that contractual terms denying 

206 Woolman (n 89 above) 774-775 makes much the same point. 

207 Barkhuizen (n 86 above) par 36.


access to the courts were contrary to public policy and therefore 

invalid in common law 208 — in principle, the time limitation clause 

could have been tested for fairness and validity purely on common law 

grounds, without any input from the Constitution. It then proceeded 

to analyse, in great detail but without much clarity, the relations 

between the substantive constitutional rights provisions such as s 34, 

the values of the Constitution and public policy. In effect, the 

Constitution and its underlying values are described as a source of 

values that inform the public policy that plays a yardstick role in the 

common law. Woolman correctly describes this as deciding a direct 

constitutional challenge indirectly by ‘speaking in values’, 209 that is, 

not with reference to the relevant rights provisions in the Bill of Rights 

but mediated through a veil of values and policy vaguely infused by 

the substantive content of the constitutional provisions. In deciding 

the case this way, the Court did not set a constitutional standard for 

contractual time limitation clauses, but it also failed to either test or 

develop the underlying common law principle — in fact, the Court 

confirmed that the principle pacta sunt servanda ‘gives effect to the 

central constitutional values of freedom and dignity’, of which ‘selfautonomy, 

or the ability to regulate one’s own affairs, even to one’s 

own detriment’ is the very essence. 210 In other words, the Court not 

only implied that value analysis, rights analysis and policy analysis is 

the same thing 211 but also accepted that the central private-law 

principle pacta sunt servanda operates on the same level and enjoys 

the same status as constitutional values, without giving much 

indication how this should work out in the larger constitutional 

project. 

In a society torn apart by historical injustices and inequality, 

transformation under the aegis of a democratic and supreme 

Constitution means that the seemingly neutral but in fact highly 

contested principal tenets of private law, with pacta sunt servanda 

and absolute ownership top of the list, have to be treated with 

contextual sensitivity in the reconstruction of a new democratic legal 

order. By reducing concrete constitutional rights provisions to a veil 

of values that inform public policy, and by allowing private law 

principles like pacta sunt servanda to join those values in the shaping 

of public policy, the Constitutional Court might appear to subscribe to 

the highly problematic view, in decisions like Brisley 212 and Afrox, 213 

that there are no significant differences between constitutional and 

common law values and that the two systems could co-exist in a way 

208 Barkhuizen (n 86 above) par 34. 

209 n 89 above, 763. 

210 Barkhuizen (n 86 above) par 57. 

211 Woolman (n 89 above) 779. 

212 n 68 above. The critical comments of Roux (n 46 above) are highly relevant here. 

213 n 68 above.


that does not demand substantive changes of the common law. 

Instead of adapting the common law, it looks as if the constitutional 

values have been integrated into the common law. 

One could argue that this result is justifiable in view of the 

importance of the principle involved, namely sanctity of contract, but 

even the Supreme Court of Appeal is apparently not that set on pacta 

sunt servanda. In Linvestment 214 the Court overturned long-standing 

precedent and decided that the owner of a servient tenement can 

insist upon relocation of a defined and registered servitude of right of 

way and that the dominant owner would be obliged to accept the 

relocation, against her will, if non-relocation would cause significant 

harm for the applicant and relocation would not cause significant 

harm for the respondent. The reasons for the decision and the result 

are justifiable on policy grounds, but it is puzzling that the Court 

would be so willing to make what is really a dramatic change to the 

common law on policy grounds, while being so unwilling, despite 

equally strong or stronger policy reasons, to change the common law 

very slightly in Tswelopele. Judging from Linvestment the best 

solution in Barkhuizen was the approach of the Supreme Court of 

Appeal, namely to focus on the general common law principle that 

supports enforcement of time limitations, pacta sunt servanda. Even 

when subjected to a healthy dose of constitutional skepticism, this 

principle embodies significant social and moral values and therefore 

it is highly unlikely that the courts would be willing to declare it 

unconstitutional just because it sometimes allows enforcement of 

unfair contracts. Instead, the approach in Masiya indicates that the 

courts would most likely, if this principle were challenged, attempt to 

rectify any shortcomings by developing it in view of s 39(2). 

Deciding Barkhuizen on this basis would have confirmed the 

subsidiarity approach that the Court seems to be developing, and it 

would have given a welcome indication that the Court is working 

towards a truly substantive, critical analysis of the common law in 

view of s 39(2). The point should not be whether the common law 

protects values like freedom and autonomy that are also important in 

constitutional law, but where and how common law principles need 

to be developed to give effect to the spirit, purport and objects of the 

Bill of Rights. More concretely, it was necessary in Barkhuizen to ask 

where and how the principles of contract need to be developed to 

give effect to the right of access to the courts in s 34, given the 

historical and constitutional context within which inequalities in 

privilege and power developed and within which the right of access to 

courts was denied the poor and weak. It is in this context that the 

mere possibility that contractual time limitations could be unfair and 

214 Linvestment CC v Hammersley 2008 3 SA 283 (SCA).


against public policy in current private law doctrine might not pass 

constitutional muster — the test that was ultimately applied by the 

Constitutional Court majority in Barkhuizen is little more than the 

private law test and it lacks the critical sensitivity for context that 

one would expect of a full-on constitutional challenge. Moreover, as 

in NM, the Court’s analysis of the facts in Barkhuizen was 

disappointing; the factual analysis of Sachs J is a much better example 

of the kind of analysis that was required and that is to be expected. 

As it stands, Barkhuizen neither tested the parties’ conduct for 

constitutional validity, nor did it consider either development of or a 

challenge to the supporting common law principle; instead, common 

law principle was elevated to the sphere of constitutional values that 

indicate whether conduct is against public policy. In the broader 

context of the 2007 decisions, Barkhuizen looks like a constitutional 

non sequitur. 

6 Conclusions 

The subsidiarity principles set out in SANDU and Bato Star could, 

according to the analysis above, reasonably be expanded so as to 

include pre-1994 legislation, partial and competing legislation, and 

instances where no legislation is applicable at all. As has been argued 

in the sections above, the subsidiarity logic of these cases could be 

developed in a way that would make it possible simultaneously to 

promote constitutional reform goals and to respect democratic 

legislative interventions that reflect reformist or stabilising policies. 

Obviously the wider implications of subsidiarity have not been worked 

out by the courts and therefore the analysis above is purely 

speculative. Furthermore, the more comprehensively worked out 

subsidiarity logic set out in the previous sections is not presented as 

a proposal to develop or entrench a judicial doctrine of subsidiarity or 

deference — it is nothing more than an exercise in the logic of the 

SANDU and Bato Star decisions, and it is presented here purely in 

order to have an opportunity to test whether such a logic, if it were 

in fact adopted by the courts, would be justifiable in view of the 

broader normative and constitutional principles discussed in the first 

sections of the article. 

With the exception of Barkhuizen (and perhaps Sidumo), the 

Constitutional Court’s decisions in the 2007 term display a growing 

and reasonably consistent tendency to favour a particular subsidiarity 

approach, always present in our constitutional setup but apparently 

now emerging with increasing clarity, according to which direct 

application of the constitutional rights provisions is restricted to 

constitutional challenges against the validity of legislation. In the 

absence of such a challenge, the Court prefers to treat cases for the


protection of constitutional rights as indirect application cases, 

involving either the interpretation of legislation or development of 

the common law in terms of the constitution-compliance and 

—confirming prescriptions of s 39(2). By extension, taking into 

account earlier jurisprudence, direct application could also be 

possible in the absence of a direct validity challenge, if the goal is to 

craft a special constitutional remedy. However, again extrapolating 

from earlier jurisprudence, the subsidiarity approach indicates that 

direct application (to craft a constitutional remedy or to challenge a 

law) is considered only when indirect protection (interpretation of 

legislation or development of the common law) proves impossible for 

some reason. Furthermore, the subsidiarity hierarchy in the category 

of indirect protection cases prescribes that application and 

development of the common law should only be considered once and 

in so far as there is no legislation to apply. 

The subsidiarity approach of the 2007 cases offers a fresh 

perspective on Mhlungu, according to which courts should, whenever 

possible, decide cases without reaching a constitutional issue. 215 

Woolman articulated five objections against the implication that the 

salutary ‘principle of avoidance’ — or subsidiarity — in Mhlungu should 

turn into a ‘full-blown doctrine in which a court must never 

“formulate a rule of constitutional law broader than is required by the 

precise facts to which it is to be applied”’; 216 some of these 

objections are relevant to my analysis. Firstly, a too narrow 

interpretation of the avoidance rule (as in Zantsi) contradicts the 

project that constitutional interpretation should evolve into a 

principled dialogue between the Constitutional court, other courts, 

the legal profession, law schools, Parliament and, indirectly, the 

public at large. 217 In addition, a too narrow interpretation would 

undermine rational political discourse and the possibility of a 

coherent jurisprudence. These are convincing reasons why the 

subsidiarity principle should not be understood too narrowly, as it 

arguably was in at least Zantsi — it should not become a directive to 

avoid deciding constitutional issues per se. Subsidiarity should be 

understood in terms of its constitutional purpose and justification, as 

that has been spelled out in 2007: to preserve the constitutional 

power and obligation of the courts to control the constitutional 

validity of legislation, while at the same time paying due respect to 

the democratic power and legitimacy of policy makers and 

legislatures in giving effect to their reform obligations under the 

Constitution. 

215 n 71 above, par 59 (dissenting, Kentridge J); Zantsi (n 71 above) par 3. 

216 Woolman (n 89 above) 784-785, citing Zantsi (n 71 above) par 8. 

217 Mhlungu (n 71 above) par 129.


The purpose of subsidiarity can therefore not be pure avoidance. 

Much rather, it should be acknowledgement and respect for the fact 

that courts face the aporia of the countermajoritarian dilemma every 

time they hear a constitutional case. On the one hand they must 

respect and give full effect to the legitimate efforts of the 

democratically elected branches to honour their constitutional 

obligations in bringing about the reforms legitimised by the 

Constitution. Doing that requires the courts to consider the interplay 

between Constitution and legislation and to give the fullest possible 

purposive effect to legislation enacted by the legislature. At the same 

time the courts must apply pre-constitutional legislation and the 

common law in a way that acknowledges the undemocratic origin and 

history of those sources of law and the dilemmas caused by their 

continued validity, in both instances by reading them through the 

corrective lens of s 39(2). In addition to its obligations concerning the 

interpretation of legislation and development of the common law, the 

courts have the constitutional duty and power to control the 

legitimacy of new legislation made by democratically elected 

legislatures, testing it against Constitutional requirements and limits. 

The only way to juggle these tasks is to acknowledge, upfront, that 

these duties of the courts imply contradictory and conflicting 

normative values and that every decision sacrifices one for the sake 

of the other. 

In the introduction I identified the push for constitutional 

transformation and the preservative pull of the common law tradition 

as conflicting normative impulses. I pointed out that this instance of 

normative pluralism differs from normative diversity as a positive 

constitutional value in that, in the case of the conflict between 

tradition and change, the complete picture is not enriched by the 

simultaneous presence of the more-than-one; these two impulses are 

locked in a zero-sum game for control over the balance between 

stability and reform. Either one wants and gets the upper hand at the 

direct cost of the other. Stability is attainable only at the cost of 

reform; change is possible only by sacrificing certainty. You can’t do 

both. And yet, if you are a judge in the new South African 

constitutional order, you have to. 

Normative anarchy, always a threat in a situation like this, is only 

to be avoided, I argued in the introduction, by giving politics a 

chance, which starts with realising that all law is political, but politics 

cannot be enclosed in law. It proceeds with acknowledging that law 

does not resolve the aporia of stability and change; it merely 

sacrifices one for the other, for now. It requires making the sacrifice, 

as far as possible, with a view to open up rather than close down room 

for social deliberation and significant dissent. Obviously, any choice 

for an approach that would avoid anarchy could easily end up by 

entrenching reactionary formalism; in the face of uncertainty and


anarchy it is only too easy to settle for the false security of shallow 

and cheap certainty. 

Giving effect to constitutional rights in a fragmented legal system 

consisting of legislation that was purposely enacted to protect and 

support those rights, legislation that pre-dates the political 

transformation and common law that was traditionally applied as if it 

embodied a-contextual neutral principles is a complex task. This task 

becomes more difficult when the courts are empowered both to 

protect constitutional rights and to control the constitutional validity 

and fittingness of all law, finding a justifiable balance between 

democratic will and constitutional legitimacy. Above all, the courts 

exercise this awesome responsibility as just one small part of the 

large project of social and political transformation, within a dynamic 

context that reflects both the ills of the past and the aspirations of 

the future. 

The Constitutional Court appears to have chosen the analytic 

rhythm of a set of subsidiarity principles to direct the choices it must 

make in this arduous process; in 2007 it followed these principles with 

a measure of consistency and some sense of purpose and justification. 

Developing a subsidiarity approach to give effect to its contradictory 

and conflicting powers and responsibilities is dangerous because it 

could easily disintegrate into shallow formalism, but it also holds the 

possibility of evolving into a more reflective, democracy-confirming 

recognition of the tension between the Constitution and existing law 

than has been evident from earlier judicial explorations into 

application issues. In that respect the direction of the pull of 2007 

must be welcomed.


129 

LEGAL SUBSIDIARITY & 

CONSTITUTIONAL RIGHTS: A REPLY TO 

AJ VAN DER WALT 

* George J & Kathleen Waters Matthews Distinguished University Professor, School 

of Law, Northeastern University, Boston (USA). Thanks to Lucy Williams and Sandra 

Liebenberg for comments and suggestions, and to André van der Walt for the 

friendship and the intellectual challenge. 

Karl Klare * 

The Constitutional Court of South Africa is one of the most respected 

legal institutions in the world. Its work excites interest among lawyers 

and others far beyond South Africa’s borders who are inspired by the 

Constitution’s democratic and egalitarian aspirations. The 

Constitution 1 and the Court 2 are precious accomplishments in the 

human striving to discover and realise the institutional foundations of 

freedom and self-determination. South Africans are guardians of 

these achievements for all of the world’s people. 

It was therefore unnerving when inane and thuggish attacks were 

lobbed at the judiciary in the run-up to Nicholson J’s judgment in 

Zuma’s Case. 3 Prominent public figures branded the judiciary 

‘counter-revolutionaries’. Adolescent poseurs threatened to ‘kill’ for 

their heroes and ‘eliminate’ their enemies. The objects of these 

attacks included the Court’s great jurists such as Pius Langa CJ, who 

has devoted his lifetime to the cause of social justice under law, and 

Dikgang Moseneke DCJ, who at age 15 was the youngest political 

prisoner sent to Robben Island. 

1 ‘Constitution’ or ‘FC’ refers to the Constitution of the Republic of South Africa, 

1996. (In the context of discussing US law, ‘Constitution’ is used in reference to 

the US Constitution.) 

2 The ‘Court’ (with an upper case ‘C’) or ‘the CC’ refers to the Constitutional 

Court. In the context of the concluding discussion of US law, ‘the Court’ refers, as 

is customary, to the United States Supreme Court. References to all other courts 

use a lower case ‘c’. 

3 Zuma v National Director of Public Prosecutions [2009] 1 All SA 54 (N) (setting 

aside decision to prosecute, reversed on appeal in National Director of Public 

Prosecutions v Zuma 2009 2 SA 277 (SCA)).

130 Reply to vd Walt 

In this fraught context, friends of the CC must avoid the 

temptation to rally around the Court by showering it with 

congratulatory platitudes. We best respect the Court by continuing as 

before to study its work meticulously and to provide the most rigorous 

and critical analysis of which we are capable. In that spirit, I have the 

honour, as a non-South African, to contribute this reply to Professor 

van der Walt. 

Professor van der Walt gifts us with a densely layered essay in 

which he tries to invent a new grammar for legal analysis. 4 He argues 

that we must become comfortable with gaps rather than seeking 

refuge in supposedly stable and all-inclusive rules because, first, this 

brings us closer to understanding what actually goes on in legal 

reasoning, and, second, because gaps open space for conversation, 

diversity, dissent, contestation, and judgment. That is, legal gaps 

open space for politics and the emergence of moral communities. Van 

der Walt believes, and I agree, that a legal discourse that embraces 

uncertainty and conflict fits better with a transformative constitution 

than arid syllogisms. Surely he is correct that South African jurists 

must work out new methods for addressing legal problems consonant 

with the Constitution’s transformative project. It is highly implausible 

that legal reasoning methods developed in earlier times, including 

apartheid times, are adequate for this task. 

It is important when reading van der Walt to appreciate that 

‘politics’ carries a positive valence. In the mainstream view, 

something has gone badly wrong when ‘politics’ enters legal decision 

making. By contrast, Van der Walt does not regard politics as an 

enemy of the rule of law. A reception of politics into legal discourse 

can be a welcome contribution to democracy, 5 provided that (1) 

adjudicators acknowledge the discretion that permits and compels 

them to make value-laden (‘ideological’) choices in the routine course 

of their work, (2) they become deeply self-conscious, self-critical, 

and candid about the ‘inarticulate premises, culturally and 

historically ingrained’ 6 that inform their thinking and choices, and (3) 

4 AJ van der Walt ‘Normative pluralism and anarchy: Reflections on the 2007 term’ 

(2008) 1 Constitutional Court Review 77. 

5 

From this perspective, the problem with the recent attacks on the judiciary is not 

that they were political. The problem is that they were uninformed, fatuous, and 

threatening, and they were aimed to achieve power for a particular political 

leader then facing serious criminal charges. In short, their purpose and their 

effect were to close down, rather than to open up, space for discussion and 

dissent. 

6 

Du Plessis v De Klerk 1996 3 SA 850 (CC), 1996 5 BCLR 658 (CC) para 119 (Kriegler 

J).


we understand law making processes as a conversation between 

adjudicators, the other branches, and the public. 7 

Van der Walt focuses on the proper relationship between three 

bodies or sources of law — the Constitution, legislation (pre- and post- 

1994), and the common law. He is concerned about the ‘preservative 

pull of the common law tradition’ in light of the common law’s 

‘undemocratic origin and history’. 8 He fears that the continued 

validity and prestige of the common law threaten to short-circuit the 

transformative project, and he wants to minimise that risk: 

An important premise of my argument is that the Constitution’s 

commitment to [political and ethical, as well as social] diversity as a 

positive value is premised upon the reduced authority of the common 

law. 9 

He warns that uncritical obeisance to common law thinking will 

immobilise the transformative project, and I could not agree more. 10 

2 From normative pluralism to legal subsidiarity 

In quick succession, van der Walt courses through a philosophical 

problem, a jurisprudential problem, and a question of constitutional 

interpretation. Subsidiarity comes in later as an instantiation of the 

approach van der Walt crafts to address more abstract considerations. 

The background philosophical problem is that modernism and postmodernism 

have undermined the foundations of epistemological 

fundamentalism: there is no knowledge-platform from which to 

identify trans-historical and trans-cultural (‘universal’) norms. The 

jurisprudential problem is that Legal Realism and Critical Legal 

Studies ‘destroyed the theoretical credibility of fundamentalism’ by 

demonstrating ‘the indeterminacy and contingency’ of legal norms. 11 

Superficially this would appear to leave lawyers with a choice 

between retreating to empty formalism, on the one hand; or, on the 

7 In this regard, see the comment of Sachs J in S v Mhlungu 1995 3 SA 897 (CC), 

1995 7 BCLR 793 (CC) para 129: ‘[T]he question of interpretation [is] one to which 

there can never be an absolute and definitive answer and ... in particular, the 

search of where to locate ourselves on the literal/purposive continuum or how to 

balance out competing provisions, will always take the form of a principled 

judicial dialogue, in the first place between members of this court, then between 

our court and other courts, the legal profession, law schools, Parliament, and, 

indirectly, with the public at large’. 

8 

Van der Walt (n 4 above) 127. 

9 Van der Walt (n 4 above) 84 (footnote omitted). 

10 On this theme, see generally D Davis & K Klare, ‘Transformative constitutionalism 

and common and customary Law’ (forthcoming). 

11 Van der Walt (n 4 above) 79.


other, sliding into normative anarchy (the idea that ‘anything 

goes’). 12 

Van der Walt refuses this binary, insisting that we can ‘recognis[e] 

that values are contingent and that everything is open for dissent and 

discussion, without implying that everything goes’. 13 He holds out for 

‘normative pluralism’ in the belief that rational discussion about 

values is possible and meaningful even in our post-modern condition. 

We must conceive legal practices and discourses as an opening to 

endless political debate and dissent. 14 This means that every legal 

decision involves a matter of judgment with a decisionist element. 15 

One value-set is chosen over and at the expense of others, but only 

‘for the time being’. All legal decisions are ‘provisional’ and may be 

qualified or superseded in the next case or next generation. Van der 

Walt craves ‘fundamentality without fundamentalism’, to use Morton 

Horwitz’s phrase. 16 

The constitutional problem is the built-in ‘tension between the 

push for constitutional reform and the pull of traditional stability’. 17 

Van der Walt’s ‘central premise’ 18 is that this tension reflects ‘a 

larger conflict between two opposing normative values or ideologies’: 

Those who think that the existing distribution of wealth, privilege and 

power and the law that helped create and uphold it are politically 

neutral and therefore not in need of political reconsideration favour a 

regime of rights in which the direct effect of the Constitution is 

preservation of the status quo; those who think that the existing regime 

and its laws are inherently politically constructed and hence in need of 

political reconsideration favour the view that neither law nor rights are 

immune from constitutional effect or democratic redefinition ... 19 

How are jurists to negotiate the push and pull of these competing 

values without rejecting either in favour of the other — that is, 

12 Van der Walt (n 4 above) 103. 


14 

Van der Walt joins a growing number of commentators who think the time has 

come to replace the evocative and influential metaphor of the Constitution as a 

‘bridge’. See Interim Constitution, Postamble, ‘National unity and reconciliation’ 

para 1 (‘[t]his Constitution provides a historic bridge ... ’); see also E Mureinik’s 

memorable ‘A bridge to where? Introducing the interim Bill of Rights’ 1994 (10) 

South African Journal on Human Rights 31. A new metaphor is needed for the 

deliberative process launched by the Constitution, perhaps seeing it as an 

‘endless highway’. 

15 ‘Decisionist’ refers to Van der Walt’s insistence upon ‘the sacrificial nature’ of 

legal decisions and their inability ever ‘to resolve social conflict with reference to 

a common will or purpose’ (Van der Walt (n 4 above) 6). 

16 M Horwitz ‘The Supreme Court 1992 term — Foreword — The constitution of 

change: Legal fundamentality without fundamentalism’ (1993) 107 Harvard Law 

Review 32. 


18 

As above. 



without retreating to a conservative formalism that sweeps the 

question under the rug, or to a moral anarchy that simply disregards 

it? The answer is to find an idiom of Constitutional adjudication that 

‘uphold[s] the tension’ 20 thereby ‘giving politics a chance’. 21 While 

no particular decision can be justified by reference to foundational 

principles, making decisions can be justified, provided that we are 

able to extract a direction-giving purpose from the Constitution. 22 A 

direction-giving purpose is not a ‘foundational principle’. It is a 

rudder to steer us between the Scylla of formalism and the Charybdis 

of anarchy. 

Borrowing in part from Frank Michelman, 23 van der Walt makes 

the crucial, albeit contestable, assumption that the Constitution does 

indeed provide a direction-giving purpose, viz, to ‘favour 

transformation’ — to renounce everything that the apartheid legal 

order stood for and to embrace the central transformative values of 

human dignity, equality, freedom, social justice, democracy, and an 

open society, values which are the reverse of everything apartheid 

was. Constitutional analysis must ‘yield to the Constitution’s 

direction-giving (transformative) purpose’ while ‘holding open’ the 

tension between stability and transformation, between the law as it 

is and the law as it should be in the free society toward which South 

Africans are transitioning. 24 

At this point Van der Walt proposes subsidiarity theory as a 

transformation-oriented and closure-refusing approach to sorting out 


21 Van der Walt (n 4 above) 127. See also 105: ‘[L]aw ... in so far as it is inevitable, 

will only be justified to the extent that it promotes normative pluralism but does 

not descend into either anarchy or reactionary formalism. In allowing room for 

normative pluralism but simultaneously avoiding anarchy, it must give politics a 

chance by allowing for real social contention and dissent’. 


23 Michelman argues that constitutional supremacy means that all sites of legal 

practice and norm-creation must pull in the same direction, and that the 

principle of constitutional supremacy is a value within the meaning of sec 1 of the 

Constitution (not just a trumping rule) when that direction is toward 

transformation of South African society in light of the ‘vision’ (such as it is) 

contained in the founding values (s 1) and elsewhere instinct in the Constitution. 

See ‘The rule of law, legality and the supremacy of the Constitution’ in S 

Woolman et al (eds) Constitutional law of South Africa (2nd Edition, OS, 2005) 11 

24 

- 36 — 11 - 38. 

Van der Walt (n 4 above) 82 - 90. I agree that, on its best reading, the 

Constitution contains the direction-giving purpose to ‘favour transformation’. But 

more work is needed on the content and historical provenance of the directiongiving 

purpose. As formulated by Van der Walt, the direction-giving purpose has 

greatest analytical power with respect to unravelling apartheid and abolishing its 

legal incidents. ‘Favour transformation’ provides less guidance with respect to 

the construction of the new society of democracy, freedom, and social justice. 

For example, what forms and degree of redistribution of social and political 

power and wealth does this ‘direction’ imply? Could a critic accept Van der Walt’s 

‘direction’ for purposes of discussion and yet plausibly claim that it is silent on 

such questions?


the relationships between common law, legislation, and a 

transformative Constitution. Van der Walt’s subsidiarity refers to the 

hierarchy of sources of law — not, as in European law, to the hierarchy 

of decision making levels. 25 The core tenet is that courts should 

‘avoid[ ] a constitutional decision if the matter can be decided on a 

nonconstitutional basis’. 26 Van der Walt assumes that legislation 

embodies a forward-looking, transformative ‘pull’, whereas the 

common law tugs in a preservative direction. 27 Accordingly he seeks 

to prioritise legislation over the common law and to shield 

Parliament’s work from judicial encroachment. In this respect, 

subsidiarity restates a familiar norm of almost all modern, democratic 

legal systems that courts should defer to the legislature (except 

perhaps in special situations, for example, when the constitutional 

rights of politically excluded and subordinated groups are at stake). A 

second goal is to encourage courts to avoid any tendency to 

proliferate separate tracks or sub-systems of law such as, for 

example, parallel systems of remedies respectively grounded on the 

Constitution itself and on legislation or the common law. 

Van der Walt finds traces of subsidiarity thinking in the Court’s 

recent work, which appropriately balances stability and 

transformation. He ‘embroider[s] upon the logic of the subsidiarity 

principles that [he has] identified in [these] decisions’. 28 The 

provisional verdict is that the Court’s emerging subsidiarity perspective 

will permit normative pluralism to flourish by avoiding 

premature resolution of the stability/transformation tension. In van 

der Walt’s argot, the Court’s cases embrace aporia so as to give 

politics a chance on a continuing basis. 

25 

In EU law, ‘subsidiarity’ denotes a mandate for downward devolution of decision 

making to the lowest level at which a particular decision may appropriately be 

taken. The idea is to harvest local knowledge, facilitate public participation and 

empowerment, and respect local concerns. The principle was formally introduced 

into EU law by the Treaty of Maastricht (entered into force 1 November 1993). 

The present formulation is found in Article 5 of the Treaty Establishing the 

European Community (consolidated version following the Treaty of Nice (entered 

into force 1 February 2003)), which provides in relevant part: ‘In areas which do 

not fall within its exclusive competence, the Community shall take action, in 

accordance with the principle of subsidiarity, only if and in so far as the 

objectives of the proposed action cannot be sufficiently achieved by the Member 

States and can therefore, by reason of the scale or effects of the proposed 

action, be better achieved by the Community’. Strictly speaking, this type of 

subsidiarity is not Van der Walt’s concern. At one point he mentions a precept 

that courts should refrain from taking a decision that can be taken by a lower 

court, but very little discussion is devoted to the hierarchy of courts. 

26 Van der Walt (n 4 above) 99 (footnote omitted). See also (95, 99, and 119 - 120) 

discussing S v Mhlungu 1995 3 SA 867 (CC), 7 BCLR 793 (CC) para 59 (giving rise to 

the principle, as somewhat reformulated by Van der Walt, that ‘a court should 

not protect a constitutional right by way of a direct validity attack or by way of a 

direct constitutional remedy before considering whether the legislation or 

common law in question could be interpreted in a constitution-conforming and 

-confirming way’) (n 4 above) 119. 

27 

Van der Walt (n 4 above) 104 - 106. 



The orchestration might differ, but my thinking is largely in tune 

with van der Walt’s philosophy and jurisprudence. My misgivings 

concern his concluding discussion of subsidiarity. I am not persuaded 

that subsidiarity theory can perform the functions van der Walt asks 

of it or that it is a reliable starting point for developing the 

transformative constitutional methodology he seeks. 

3 Subsidiarity through a critical lens 


The raison d’être of subsidiarity principles is to strike an authoritative 

balance between the conflicting values of judicial deference and 

constitutional supremacy, so that courts are not at large weighing the 

conflict on an ad hoc, case-by-case basis. The central weakness of the 

theory is that, when called upon to resolve difficult cases, subsidiarity 

does little more than instruct courts to reopen the question and 

conduct the ad hoc balancing exercise it counselled them to avoid. 

Moreover, at least as presently formulated, subsidiarity imports 

conventional separation-of-powers discourse without rigorous critical 

examination; it relies upon the deceptively simple but underexamined 

and ambiguous notion of a statute ‘giving effect’ to a 

constitutional right; and, as a result, it tends to skew analysis in the 

direction of stability rather than transformation. 

3.2 Imports 

The subsidiarity approach is not a critical theory. Like most 

conventional discourse of separation-of-powers, subsidiarity rests on 

timeworn and reified conceptions of the various branches and organs 

of government and their inter-relationships. This imagery relies as 

much on wishful thinking as on evidence or analysis. Van der Walt 

knows this of course, but when he recruits subsidiarity to the cause of 

transformation, he cannot shed all of its baggage. Because this is a 

general problem in legal theory, I simply register my objection pro 

forma and leave detailed examination of these issues to another day. 

The trite assumption is that the legislature is the democratic and 

representative branch, to which non-elected judges owe deference. 

We twist ourselves into knots trying to explain why and under what 

circumstances counter-majoritarian judicial review might be 

legitimate in a democratic society. Jurists spin out a catechism of 

legal fictions and maxims from the simplistic notion of legislative


primacy, which on occasion van der Walt incongruously deploys. 29 Van 

der Walt no doubt takes a critical view of all this, but the discourse is 

implicit in the subsidiarity theory he imports. He even goes a step 

farther, consistently portraying the legislature as the forwardlooking, 

progressive branch and a repository of transformative 

intentions and energy, whereas he treats the courts as necessarily 

backward-looking, reactionary, and obstructionist. 

These assumptions fall to be interrogated in the perspective of 

transformative constitutionalism. They make a fine starting point for 

discussion but ought not substitute for investigation and analysis. We 

must move outside the standard frame of the ‘counter majoritarian’ 

dilemma, 30 for at least two reasons. 

First, political institutions (parliaments, courts, agencies) do not 

have an intrinsic nature or an inherent leaning toward or against 

transformation. Whether, for example, the legislature is representative 

and/or plays a democratic, let alone a progressive, role are 

matters of contextual judgment on the historical record. ‘Process 

failures’ happen. 31 Legislatures have been known to ally with 

reaction, to be captured by wealthy interests, and/or to be so 

dominated by an entrenched oligarchy as to lose touch with the grass 

roots. I do not say any such thing has occurred in South Africa since 

1994, although Parliament’s lock-step ratification of the arms deal 

provides food for thought. 32 The point is simply that any separationof-powers 

theory worth having in the 21st century cannot assume 

that, but must inquire whether, legislatures as we find them actually 

function in service to the popular will. 

Second, it is much too late in the day to assume that judicial 

review of legislation under a supreme and rights-rich constitution is 

inherently anti-democratic. Yes, judicial review may derogate from 

majority rule. But at the same time, judicial enforcement of rights 

and other constitutional provisions can deepen and enhance 

29 For example, he cites the fiction that the legislature approves that portion of the 

common law which it has not amended by statute (Van der Walt (n 4 above) 115). 

30 

It is worth noting that Van der Walt makes this precise point in an illuminating, 

earlier essay. See AJ van der Walt ‘Transformative constitutionalism and the 

development of South African property law’ (part 1) 2005 (4) Tydskrif vir die 

Suid-Afrikaanse Reg 655 (contrasting the ‘counter majoritarian dilemma’ — the 

familiar problem of the legitimacy of judicial review of legislation — with a new, 

‘transformative dilemma’). The Constitution ‘created a new adjudicative 

dilemma of simultaneously upholding the supremacy of the new constitution and 

the integrity of a well-developed and established system of private law’ (657 

(footnote omitted)). 

31 

By analogy to the terms ‘market failure’ and ‘regulatory failure’, I use ‘process 

failure’ to refer to a systemic breakdown in the idealised relationships between 

branches of government. 

32 

See P Govender Love and Courage: A Story of Insubordination (2007) 197 - 205 

(former ANC parliamentarian’s account of process of voting on arms deal).


democracy by making legislative bodies more responsive, legislative 

processes more transparent, or elections more competitive, or by 

opening communication channels between the populace and 

government officials. The net result in context might be that the 

democracy-enlarging consequences of the court’s decision outweigh 

its counter-majoritarian form. Treatment Action Campaign 33 comes 

to mind as an example. 

Moreover, South Africa’s Constitution embodies a conception of 

democracy that goes well beyond representation, elections, and 

majority rule. South African constitutional democracy contemplates 

not only robust political rights, but also that the resources and social 

circumstances necessary for meaningful political participation and 

authentic experiences of self-determination will gradually be 

extended to all South Africans. As Theunis Roux meticulously 

demonstrates: 

The principle of democracy in South African constitutional law is not 

that collective decisions shall be taken by majority vote, but something 

far deeper than this, including, at the very least, the notion that the 

people’s will may be trumped by individual rights where this serves the 

democratic values of ‘human dignity, equality and freedom’. 34 

Depending on what they do with their power, a more assertive role for 

courts than is contemplated by the conventional wisdom may be 

consistent with or possibly even required by democracy as envisioned 

by this Constitution. Recognising that judicial enforcement of the Bill 

of Rights can be democracy-enhancing disrupts traditional separationof-powers 

assumptions and calls for a re-examination of the old 

shibboleths. 35 

33 Minister of Health v Treatment Action Campaign (No 2) 2002 5 SA 721 (CC), 2002 

10 BCLR 1033 (CC). 

34 T Roux ‘Democracy’ in S Woolman et al eds (n 23 above) 10 - 66 (quoting FC sec 

7(1)). 

35 At one point, Van der Walt discards the conventional discursive framework and 

reverts to his accustomed, critical self. His insight is so powerful that it deserves 

separate mention. The CC and other courts, he says, conflate the common law 

insistence that courts restrict themselves to incremental development of the law 

with the separation-of-powers imperative that courts defer to the legislature. He 

convincingly argues that courts are perfectly capable of respecting the superior 

authority of the legislature while at the same time taking bold steps as necessary 

and appropriate: ‘Respect for democratic processes should not prevent the courts 

from rooting out remnants of tradition such as discrimination or inequality in 

conflict with the new constitutional dispensation ... hence the courts need not 

shy away from larger development of the common law out of respect for the 

legislature’ (Van der Walt (n 4 above) 120). In fact, Van der Walt hopes that 

subsidiarity will provide courts with a decisional framework authorising and 

encouraging them to take big (transformative) steps when Parliament is silent and 

to challenge Parliament when its efforts to give effect to the Bill of Rights are too 

grudging.


3.3 Problems of application 

Subsidiarity performs a ‘gate-keeping function’. It instructs courts to 

answer certain threshold questions before entertaining suits seeking 

to vindicate a constitutional right or to obtain constitutional damages 

or other special relief on the Constitution itself (‘Constitution-based 

remedies’). In principle, subsidiarity analysis tells a court whether to 

proceed with an action styled as a claim on the Constitution itself (or 

for Constitution-based remedies); or whether, in the alternative, to 

limit the claimant to such rights and entitlements as are available 

within the compass of a pertinent statute giving effect to the 

constitutional right, if any, or the common law, if not. At stake in this 

inquiry is whether a court may heed its own judgment on how to give 

effect to the constitutional right or whether it must leave that task, 

and the prudential and fiscal choices involved, to the legislature. Van 

der Walt explains well enough why in a given type of case courts 

should take the approach prescribed by subsidiarity. Unfortunately, 

subsidiarity often cannot tell us which cases have the specified 

characteristics, notwithstanding van der Walt’s assiduous efforts to 

‘embroider’ the theory. As a practical matter, courts can only answer 

the threshold, gate-keeping questions by examining and balancing the 

pertinent constitutional values and principles. Courts cannot apply 

subsidiarity theory without addressing precisely the questions, and 

making precisely the value judgments, that the theory means them to 

avoid. 

Under SANDU, for example, if Parliament enacts legislation to 

give effect to a constitutional right (‘effect-giving statute’), a 

claimant seeking to protect or enforce that right must rely on the 

legislation and is precluded from bringing a claim directly on the 

Constitution. However SANDU preclusion does not apply when a 

claimant challenges the constitutionality of the statute. 36 Therefore, 

in order to apply SANDU, a court must be able to distinguish rightsenforcement 

claims from constitutional challenges. 37 Drawing that 

distinction will be easy in many cases, but not in others. Important 

cases fall into middle-ground, creating difficulties for the theory that 

limit its usefulness. 

A claimant eligible under an effect-giving statute for the relief 

sought will sue on the statute. She will sue on the Constitution or for 

Constitution-based remedies only if she cannot obtain what she wants 

on the statute. It is reasonable to assume, therefore, that after 

36 

See Van der Walt (n 4 above) 100 - 103, discussing South African National Defence 

Union v Minister of Defence (SANDU) 2007 5 SA 400 (CC), 2007 8 BCLR 863 (CC) 

paras 51 - 52. 

37 

Obviously the claimant’s own description of her claim cannot be determinative. If 

it were, the SANDU principle could be nullified by strategic pleading.


SANDU, whenever a claimant is ineligible for relief on an effect-giving 

statute, she will pursue her claim by alleging that the enforcement 

provisions of the statutory scheme are constitutionally inadequate. If 

the statute is constitutionally adequate as applied, and the claimant 

is remitted to her statutory remedies, she is out of luck. If the statute 

is not constitutionally adequate, the claimant may proceed on the 

Constitution, and the court must hear her case. Thus, every nonfrivolous 

case in which the effect-giving statute denies relief, that is, 

virtually every serious case likely to arise post-SANDU, raises a 

preclusion-avoiding, constitutional question. In all cases seeking 

constitutional relief beyond that provided in an effect-giving statute, 

the courts must make a pre-threshold determination as to whether 

the plaintiff has a legitimate claim of constitutional inadequacy 

before it can make the supposedly threshold, subsidiarity-prescribed 

determination whether the case should be decided on the statute 

alone. 

This renders the SANDU principle a nullity, unless courts are able 

to make the pre-threshold determination on the face of the pleadings. 

But how is a court to know whether it is in the presence of a bona fide 

constitutional question, short of airing and taking a view on the 

matter? Inevitably, the court must consider the full panoply of 

relevant constitutional values, including separation-of-powers values. 

In addition to substantive constitutional concerns, the court must 

inquire whether the question of remedies is one peculiarly within the 

competence of the legislature, or whether the principle of 

constitutional supremacy authorises or obliges the court to consult its 

own views of the matter. By itself, the SANDU principle provides little 

guidance in determining whether a valid claim on the Constitution has 

been pleaded; that determination must be made before SANDU kicks 

in. 

In sum, unless a claim is patently frivolous, a court cannot answer 

the pre-threshold question — whether the remedies provided by an 

effect-giving statute are constitutionally adequate — without fully 

considering and reaching at least a tentative conclusion on that point. 

But the purpose of subsidiarity was to prevent claimants from 

precipitating full-dress adjudication of constitutional issues when the 

legislature has given effect to a constitutional right. The process of 

applying subsidiarity principles reinstates the problem it is intended 

to solve. 

This dilemma, if it is one, derives ultimately from an ambiguity in 

the constitutional provisions that authorise or enjoin Parliament to


‘give effect’ to an enumerated constitutional right. 38 The bare words 

of the text can bear at least two polar meanings and perhaps a range 

of others in between. To ‘give effect’ to a constitutional right might 

mean that the right is free-standing, with a content of its own, but 

that Parliament is invited or mandated to implement and give 

concrete, practical significance to the right. In this interpretation, 

the content of the right is not exhausted by the effect-giving statute. 

Alternatively, and in contrast to textual references to the 

‘progressive realisation’ of constitutional rights through legislation, 39 

the phrase ‘give effect’ might mean that the right has the meaning 

and effects that Parliament gives it. The case for the former 

interpretation seems overwhelming to me, but, followed to its logical 

end-point (a course which van der Walt does not advocate), 

subsidiarity pushes the law toward the latter approach. 

I should pause to say that, in my view, constitutional rights are in 

part constituted by the extent and scope of relief awarded for their 

violation. Generations ago, the Legal Realists argued that remedies 

define legal rights and entitlements, rather than the other way 

around. On the traditional view, courts fashion remedies based on 

their understanding of the nature of the legally protected interests 

infringed and the character of the wrong to those interests. According 

to the Legal Realists, we learn the identity of the legally protected 

interests and the nature of wrongs by examining the remedies courts 

dispense. A wrongful injury to a legally protected interest is that for 

which and to the extent that the courts provide a remedy. 40 One need 

not press the point to an extreme to accept that to some significant 

extent, enforcement and remedies determine what a right means in 

practical effect in the lives of the parties. When Parliament ‘gives 

effect’ to a constitutional right, it may task itself with giving the right 

an enforceable floor of protections and implementations. In practice, 

it may also erect a ceiling and walls around the right. At a certain 

point, ‘giving effect’ to a constitutional right slides into defining the 

right by setting out its metes and bounds. The question we deal with 

here is whether and to what extent the courts are confined within the 

houses that Parliament builds. 

The constitutional adequacy of the relief afforded by an effectgiving 

statute is not, strictly speaking, a question subsidiarity theory 

addresses — it is a substantive problem of constitutional law that must 

be decided by courts. However, the issue of constitutional adequacy 

38 See, eg, FC sec 33(3) (just administrative action) (‘[n]ational legislation must be 

enacted to give effect to these rights’). 

39 

See, eg, FC secs 26(2) (housing) and 27 (health care, food, water & social 

security). 

40 See, eg, WH Hamilton & I Till Property (1933) 12 Encyclopedia of Social Science 

528 536 (‘[i]t is incorrect to say that the judiciary protect[s] property; rather they 

call[ ] property that to which they accord[ ] protection’).


does implicate core subsidiarity concerns in the sense that how one 

approaches it may be strongly influenced by one’s attitude regarding 

the appropriate roles of courts and legislatures. Too eager a judicial 

willingness to find constitutional fault with effect-giving statutes 

might deprive the legislature of the discretion to which it is properly 

entitled in a democratic society. But an overly cautious attitude, 

comfortable with directing most claimants to their statutory 

remedies, might abdicate the courts’ obligation to protect and 

promote the Bill of Rights. 

Particularly in the field of socio-economic rights, when Parliament 

enacts an effect-giving statute providing a range of benefits, rights, 

and entitlements to qualified applicants and imposing corresponding 

obligations and prohibitions on the government, it will trade off 

limitations on the coverage, administration, and enforcement of the 

right in return for greater generosity of benefits and ease of access to 

them. Such tradeoffs are entirely legitimate means to make 

economical use of scarce resources needed to fulfil other 

constitutional rights and to provide for orderly conduct of executive 

and judicial business. Legislatures are supposed to make such tradeoffs, 

and the Constitution requires South African courts to respect the 

considered judgment of Parliament. 41 But must courts give it 

dispositive weight on practical questions of enforcement and 

remedies? Leave aside crudely under-protective and otherwise 

patently inadequate remedial schemes. Assume the legislature’s 

scheme falls within a range of reasonable decision making. In a 

democracy, are there any circumstances under which a court may 

properly substitute its own thinking for that of the legislature? What 

if a court concludes that a different choice of remedies would do a 

significantly better job of protecting the constitutional right in 

question, with a net gain to democracy? Is it nevertheless precluded 

by the legislative judgment (except in extreme cases)? 

Assume that in a statute giving effect to a constitutional right, 

Parliament provides relief for denial or infringement of the right by 

lawsuit, and that the statute specifies a four-year time-limitation 

period for instituting such proceedings. X sues on the statute against 

an appropriate defendant. Her claim is meritorious in all respects save 

that she waited five years to bring the claim. She was tardy for a 

combination of reasons: she was not well-schooled in her rights, she 

faced practical and financial difficulties in obtaining representation, 

and she pursued the claim with insufficient vigour and attention. X 

pleads her cause both on the statute and on the Constitution itself. 

She argues that the four-year cut-off is unconstitutional, and that the 

court should either grant Constitution-based remedies or, alter- 

41 See FC sec 41(1)(e).


natively, read a longer limitations-period into the statute and then 

grant relief on the statute so revised. 

If Parliament had specified a 24-hour limitations period, most 

everyone would agree that the statute is unconstitutional to that 

extent—both for failing to give meaningful effect to the substantive 

right and in terms of the section 34 right of access to courts. 42 Assume 

for purposes of discussion, however, that the Constitutional Court had 

previously ruled that Parliament does not offend the Constitution 

merely by imposing some reasonable limitations-periods in effectgiving 

statutes, the constitutionally acceptable minimum length to be 

determined in light of the nature of the right, the nature of potential 

violations, and any other relevant considerations. If the court in X’s 

case were to conclude that a four-year prescription period plainly, 

unjustifiably, and unreasonably limits enjoyment of the constitutional 

right, presumably it would be common cause that the court may 

properly, indeed must, hold the statute invalid to that extent. 

Suppose instead that the court concludes (1) a four-year time-frame 

is within the range that could reasonably be regarded as appropriate 

and constitutional, but nevertheless (2) a seven-year period would 

better safeguard the right in question. The court arrives at this 

conclusion because it deems the substantive right to be very 

fundamental and in need of generous protection; because of its 

concern about the practical barriers to rights-enforcement faced by 

millions of poor South Africans; and because it feels the more 

forgiving limitations period better comports with ‘the values that 

underlie an open and democratic society based on human dignity, 

equality, and freedom’. 43 May the court impose its preference for 

seven years, thereby trumping Parliament’s choice among a range of 

reasonably acceptable approaches to time-limitation? Are courts, or 

at least the Constitutional Court, charged with the authority to 

optimise protection of fundamental rights? Or would such a decision 

violate a proper understanding of separation-of-powers? 

42 In Moise v Transitional Local Council of Greater Germiston 2001 4 SA 491 (CC), 

2001 8 BCLR 765 (CC), the CC confirmed an order of the Witwatersrand High Court 

finding unconstitutional and invalid sec 2(1)(a) of the Limitation of Legal 

Proceedings (Provincial and Local Authorities) Act 94 of 1970, which barred suits 

against local authorities unless the claimant gave written notice of a claim within 

90 days of the date on which it arose. (Plaintiffs had 24 months within which to 

file suit.) The short notice period was held to impair the FC sec 34 right of access 

to courts in a manner that was not reasonably justifiable within FC sec 36(1). In 

Engelbrecht v Road Accident Fund 2007 6 SA 96 (CC), 2007 5 BCLR 457 (CC), the 

Court held unconstitutional regulation 2(1)(c), GN 17939, 25 April 1997, 

promulgated in terms of the Road Accident Fund 56 of 1996. With some 

qualifications, the regulation barred compensation to a victim of a hit-and-run 

accident unless the claimand filed a police affidavit within 14 days of the 

accident. 

43 FC sec 39(1)(a) (concerning interpretation of the Bill of Rights).


Subsidiarity demands that the court dismiss X’s case even though 

it is in part styled as a constitutional challenge. If we rule out ultrashort 

and patently unfair limitations periods, setting the precise timelimit 

on claims is quintessentially a question of ‘policy’ as to which 

courts should refrain from substituting their own judgment for 

Parliament’s. This is so even if the legislature, in reaching its 

decision, predictably and properly took into account such nonconstitutional 

factors as the government’s need for efficient and 

frugal management of its legal and material resources, its desire to 

remove impediments to economic development, and/or its unadorned 

interest in avoiding litigation. It is surely defensible to argue that 

courts have no business critically scrutinising, let alone secondguessing, 

Parliament. It is worth noting that there is nothing 

particularly ‘transformative’ about this position. Most jurists who 

administer democratic constitutions providing for judicial review 

would probably counsel deference in this situation. Is that the best 

approach for South Africa or one that the Constitution obliges South 

African courts to take? 

It is not obvious, at least to me, that the familiar separation-ofpowers 

considerations supporting judicial deference are equally 

compelling when invoked in a case the outcome of which will define 

the operative bounds of a constitutional right. The legislature may 

have superior institutional competence to make what are 

conventionally called ‘policy decisions’, but it is not necessarily 

better situated than the Constitutional Court to strike the appropriate 

balance between majoritarian interest and constitutional principle. Is 

it not paradoxical to entrench judicially enforceable rights in a 

supreme constitution — a constitution meant to constrain the 

legislature — and then leave it to the legislature effectively to 

determine how those rights are to be protected and enforced, subject 

only to highly deferential judicial review? Given the paramount 

importance of constitutional supremacy in South Africa’s evolving 

constitutional understanding, the courts might legitimately be 

entitled to greater latitude in cases involving the effective definition 

and enforcement of constitutional rights than they may rightly claim 

in reviewing social and economic legislation that does not derogate 

from important rights and freedoms. In a transformative context, a 

bit more active role for the courts (or at least the Constitutional 

Court) and a bit more back-and-forth between the branches might be 

in order. 44 

44 

44 

Courts — particularly the peak level courts (CC and SCA) — could evolve practices 

designed to reduce the institutional tension between the branches and establish a 

more ‘dialogic’ relationship between them. A reviewing court could, for example, 

announce its conclusions but withhold judgment, pending further representations 

from the government or Parliament itself. A court could declare the four-year 

time limitation constitutionally deficient, but suspend the effective date of its 

order to give the legislature time to reconsider and, perhaps, devise an


If the reader will grant that these questions are at least worthy of 

debate, subsidiarity provides us no new tools with which to approach 

them. The theory instructs courts to leave to legislation (or to the 

common law, developing under the mandate of section 39(2) 45 ) cases 

in which it is appropriate in light of separation-of-powers theory to do 

so, but not in cases where that course is inappropriate. But the theory 

lacks the analytical traction necessary to perform any real work in 

assigning actual claims to their appropriate category. Any application 

of subsidiarity in a borderline case pushes the decision maker back to 

the general debate about separation-of-powers. Subsidiarity recites 

some of the countervailing considerations of which account ought to 

be taken, but it adds no new insight as to how they should be 

balanced. 

Likely van der Walt would say that this is exactly his point — there 

are no easy answers to such questions. Precisely what he wants is a 

general acknowledgment that resolving such conundrums calls for 

‘politics’ — discussion, debate, dissent, and the taking of judgments 

about how best to serve competing imperatives (‘for the time being’). 

Discuss, debate, and balance; decide provisionally; hold tensions 

open — I could not agree more. But van der Walt told us that before 

he brought subsidiarity into the picture. Subsidiarity cannot help us in 

pursuing the discussion. 

The reader may doubt that anything of constitutional substance 

can turn on the length of a limitations period. If so, she will find 

sobering a US case in which the Supreme Court sanctioned a man’s 

execution without affording a hearing on his constitutional claims 

because his lawyer was three days late in filing a notice of appeal in 

a lower court. 46 Nevertheless, I turn to examples in which the 

substantive constitutional stakes are more apparent. 

intermediate solution or approach the problem from a different angle. A court 

could invite, ask, or order Parliament to reconsider, standing ready to defer to the 

legislature or review its work with less exacting scrutiny if, upon reconsideration, 

it reaffirms the choice of a four-year limitations period. It could invite, ask, or 

order Parliament to propose a new solution within stipulated guidelines. The 

invention of intermediate review practices between the extremes of deference 

and trumping is one of the great challenges to constitutional law in our time. 

45 FC 39(2) (development of the common law and customary law ‘must promote the 

spirit, purport and objects of the Bill of Rights’). 

46 Coleman v Thompson 501 US 722 (1991) (O’Connor J). Coleman was convicted of 

crimes and sentenced to death. He contended that his trial was infected by 

violations of his constitutional rights. The Court held that the attorney’s latefiling 

was a procedural default rendering Coleman’s conviction final under of the 

state law of Virginia, and therefore unreviewable in the US Supreme Court. 

Coleman was condemned by the equivalent of a ‘subsidiarity rule’ in the 

federalism context. The rule was based on two legal foundations — limitations on 

federal jurisdiction imposed by Article III of the US Constitution, and principles of 

federal/state comity. (The ‘procedural default’ doctrine has exceptions, but the 

Court did not find them applicable in Coleman’s case.)


Imagine that in response to Carmichele 47 and its progeny, 

Parliament enacted a comprehensive programme (the ‘domestic 

violence programme’ or DVP) intended to give effect to women’s 

equality rights (FC section 9), children’s rights (FC section 28), the 

right to security of the person (FC section 12), and the obligation of 

government to protect against violence (FC section 12(1)(c) read with 

FC section 7(2)) in the context of violence against women and other 

domestic violence. DVP commits the national government to massive 

expenditures for increased policing, a well-designed scheme of 

initiatives for police and other officials to receive training on violence 

against women and children, a victim advocate programme in the 

courts and in the community, and secure shelters and temporary 

financial assistance to victims and potential victims. DVP also 

provides scheduled compensation, obtainable through simple and 

prompt administrative procedures, for injuries caused by negligent 

acts or omissions of the police, court authorities, and other public 

officials. In response to concerns about costs, the statute declares the 

specified administrative compensation to be the exclusive remedy in 

such cases, and moreover it grants the government and police 

absolute immunity from suit for negligence in any case involving 

violence against women or children. Assume that all interested groups 

in civil society and leading experts from numerous disciplines were 

consulted and had substantial input into the design of DVP. Many 

objections were raised in the consultation process, but in the end a 

wide consensus emerged that DVP is a state-of-the-art programme 

promising great benefits to vulnerable or victimised persons and to 

the community at large. The vote in Parliament in favour of DVP was 

unanimous. 

H attacks and shoots his wife W with an illegally possessed 

firearm, so seriously injuring and disabling her that she will need caretaking 

assistance, social work, and other services for life. Her young 

children witnessed the attack and suffered deep emotional trauma; 

they will require specialised services for the foreseeable future. H’s 

attack occurs under circumstances reminiscent of Carmichele and 

related cases — H had a history of violence known to local authorities; 

the police failed, for no legitimate reason, to impound H’s guns; and 

no warning or protection was given to W after police officers observed 

H intoxicated and threatening to kill her. H is arrested shortly after 

the attack and is eventually sentenced to a long prison term. 

W claims administrative compensation under DVP, to which she is 

unquestionably entitled. However, the DVP benefits do not provide 

redress for the emotional injury to her or the children, and they do 

47 

Carmichele v Minister of Safety and Security (Centre for Applied Legal Studies 

Intervening) 2001 4 SA 938 (CC), 2001 10 BCLR 995 (CC).


not come close to the estimated life-time cost of her and the 

children’s caring needs. W therefore sues the local police and the 

Minister of Safety and Security for unlawful omission to protect her. 

W brings the claim in delict, urging that, if necessary, the common 

law be developed so as to promote the spirit and objects of the Bill of 

Rights. Alternatively, she sues on the Constitution itself seeking 

Constitution-based damages. The defence concedes that the police 

were guilty of wrongful omissions, that these omissions proximately 

caused W’s injuries, and that the police’s omissions violated common 

law and constitutional duties owed to W. Nevertheless the police and 

Minister seek dismissal of both branches of W’s claim based on the 

DVP immunity provision. W in turn concedes that the defendants are 

immune in terms of DVP, but she argues that in application the 

statutory scheme is unconstitutional to that extent. 

May or must a court hear either her delictual and/or her 

constitutional claim? Should the court remit her to her remedies 

under DVP (with the result that a large portion of her injuries will go 

uncompensated and her constitutional challenge will go unheard)? At 

first glance, the Bato Star principle 48 suggests that W’s common law 

claim is precluded, but she escapes that tenet because she has 

challenged the constitutionality of DVP. If bypass is so easy, can 

subsidiarity effectively constrain courts from second-guessing the 

legislature and from undermining its carefully engineered policy 

compromises? If the court hears W’s claim on the Constitution, must 

it defer to the legislature’s policy judgment, and if so, how much 

deference must it show? On the other hand, if the court may not 

independently examine the constitutionality of the immunity 

provision — if, for example, subsidiarity directs the court to abstain 

from hearing W’s claim on the Constitution — have we not vested the 

legislature with effective power to define and limit the constitutional 

rights at stake, in violation of the fundamental principle of 

constitutional supremacy? 

I do not mean to suggest that there are easy or clear answers to 

these questions. My point is simply that subsidiarity brings us to their 

threshold, and then its analytic power runs out. It leaves us without 

moorings in attempting to figure out the appropriate roles and interrelationships 

of legislation and judge-made law. 

48 

See Van der Walt (n 4 above) 103 - 105 discussing Bato Star Fishing (Pty) Ltd v 

Minister of Environmental Affairs 2004 4 SA 490 (CC), 2004 7 BCLR 687 (CC) para 

25 and Chirwa v Transnet Ltd 2008 2 SA 24 (CC), 2008 3 BCLR 251 (CC) para 23. 

According to Van der Walt (103), these cases state the principle that ‘once 

legislation has been enacted to give effect to a right in the Constitution, and in so 

far as the legislation was intended to codify the common law, litigants may not 

rely directly on the common law when seeking to protect that right against 

infringement’ (italics in original).


3.4 A cautionary tale — of Schweiker v Chilicky 49 

In this case, the US Supreme Court denied claimants the right to 

pursue constitutional claims because Congress had enacted an 

administrative scheme to redress their grievances. At first glance, 

Chilicky is SANDU with an American accent. But subsidiarity thinking 

runs wild in the Court’s opinion, which would make Kafka proud. To 

be clear, I do not believe that subsidiarity theory entails the Chilicky 

result, and I am confident that van der Walt would regard the case as 

wrongly decided. I offer it as a cautionary tale for South Africans — to 

show how much damage can be done when jurists who do not possess 

van der Walt’s sophistication and intellectual integrity use 

subsidiarity-talk to resolve a case. 

Comparative discussion of constitutional torts requires caution 

because the US system is shaped by several factors that do not play a 

role South Africa. The US states are conceived as semi-autonomous 

sovereigns with a much more significant and independent legal 

identity than is associated with South African provinces. US federal 

courts are courts of constitutionally limited, not general, jurisdiction. 

US law seems to be more fastidious in distinguishing between rights, 

sources of law, and causes-of-action for redress of violations. 

Notably, the US Constitution does not contain a general ‘standing’ 

provision such as found in FC section 38, 50 and the case-law is 

generally more restrictive than what FC section 38 permits. 51 The key 

point for present purposes is that the US Constitution does not 

expressly create any causes-of-action under which individuals may 

seek redress for constitutional violations. Congress long ago enacted 

statutes expressly creating causes-of-action against persons acting 

under the colour of state law who deprive individuals of federal 

constitutional rights. 52 There is no comparably generic cause-ofaction 

for violation of constitutional rights by the federal government 

or others acting under the colour of federal law. As a historical 

matter, it was generally assumed, but not entirely certain, that an 

individual could sue a federal actor for redress of constitutional rights 

49 

Schweiker v Chilicky 487 US 412 (1988) (‘Chilicky’) (O’Connor J). Brennan J, 

joined by Marshall and Blackmun JJ, dissented. 

50 FC sec 38 (enforcement of rights) both lists the parties who may approach a court 

when a right in the Bill of Rights ‘has been infringed or threatened’ and gives 

them a constitutional right to do so. 

51 The picture is further complicated because South African constitutional law has 

numerous distinctive features missing or largely absent in the US. These include 

affirmative governmental duties, social and economic rights, and direct 

horizontal application. The latter exists in the US — eg, the Thirteenth 

Amendment prohibition of slavery has direct horizontal application — but US 

constitutional jurisprudence is dominated by a powerful public/private 

distinction. 

52 

The most salient of these is the provision of the Civil Rights Act of 1871 now 

codified at 42 USC § 1983.


violations. In the modern Bivens case, 53 the Court held that the 

Constitution impliedly gives rise of its own force to a cause-of-action 

in favour of an individual deprived of federal constitutional rights by 

a person acting under colour of federal law. 

Bivens remains good law, but the Court has narrowed its scope in 

recent years. In my view, the Court moved in this direction partly 

because of hostility to the Bill of Rights and an unarticulated 

antipathy to some of the kinds of people (eg, poor people) that 

attempt to bring Bivens claims. The Court’s stated reasons are based 

on subsidiarity principles. Specifically, the Court has held that a 

claimant may not pursue a Bivens tort where Congress has enacted an 

enforcement scheme providing meaningful remedies for 

unconstitutional action. In a leading case, Bush v Lucas, 54 a federal 

government employee claimed that his employer had violated his 

constitutional right of freedom of speech. 55 He was precluded from 

bringing a Bivens action and was instead confined to the 

administrative relief available under statutes governing the 

employment of federal civil servants, even though he alleged that 

these remedies were inadequate to deter unconstitutional 

wrongdoing and did not allow compensation for certain of his injuries 

(eg, emotional harms). 56 

The Bush Court assumed that the civil service procedure was 

constitutionally adequate. 57 One might think that, had the Court 

treated Bush’s claim as a direct challenge to the constitutional 

sufficiency of the administrative remedies, 58 it would have permitted 

his Bivens claim to go forward. This would bring the US approach into 

53 

Bivens v Six Unknown Named Agents of Federal Bureau of Narcotics 403 US 388 

(1971). Actions pursuant to the rule of this case are called ‘Bivens torts’. 

54 

462 US 367 (1983). 

55 See US Constitution, 1st amend. 

56 Mr Bush is not related to either of the former Presidents Bush. The Supreme Court 

affirmed that federal courts have the power to establish remedies needed and 

appropriate to vindicate constitutional rights, even if not expressly authorised by 

Congress (Bush (n 54 above) 374). However, it declined to exercise that power in 

Bush’s case because, after careful consideration, Congress had struck a balance 

between ‘the conflicting interests involved in providing job security, protecting 

the right to speak freely, and maintaining discipline and efficiency in the federal 

work force’ (385). Congress had developed ‘an elaborate, comprehensive scheme 

... by which improper [personnel] action may be redressed’, (385), and which 

‘provides meaningful remedies for employees who may have been unfairly 

disciplined for making critical comments about their agencies’ (386) (footnote 

omitted). The Court saw as the central question in the case whether the remedial 

system established by Congress ‘should be augmented by the creation of a new 

judicial remedy for the constitutional violation at issue’ (388). That question 

requires a ‘policy judgment’, including a cost-benefit analysis of the value of 

additional remedies and their possible impact on the efficiency of the civil 

service, a policy judgment that Congress is in a far better position than the Court 

to make (388 - 389). 

57 Bush (n 54 above) 378 n 14. 

58 

Bush’s pleadings were unclear on this point, although on a generous reading his 

suit did advance a constitutional challenge to the administrative scheme.


conformity with SANDU. If one thought that, one would be mistaken. 

Chilicky, to which I now turn, disallows a Bivens tort even where the 

claimant raises an explicit constitutional challenge to the statute (as 

applied), and even though the alternative, statutory scheme to which 

she is remitted excludes recovery for unconstitutional conduct. 

The Federal Government (in cooperation with the states) operates 

several programs dispensing monetary benefits to persons who suffer 

specified forms of income interruption (due, for example, to 

unemployment, retirement, spouse’s death). Chilicky concerned a 

benefits programme for certain categories of people unable to 

perform paid work due to disability. Many recipients rely entirely on 

these benefits for their livelihood and that of their families. At the 

time, eligibility for disability (or other welfare) benefits was in many 

cases a requirement of access to Medicaid, the programme of healthcare 

coverage for poor people. 

Former President Ronald Reagan’s Administration was hostile to 

all forms of welfare and in the early 1980s imposed directives and 

informal pressures on programme managers designed to get them 

drastically to reduce the number of social welfare recipients. As a 

result, the disability-benefit programme alone terminated hundreds 

of thousands of benefit-recipients over a period of years using 

questionable, and often sleazy, methods. The Chilicky claimants 

argued that the procedures the Government used to clear the welfare 

rolls violated the right to Due Process guaranteed by the Bill of 

Rights. 59 Congress eventually brought a halt to the wholesale 

terminations and introduced procedural safeguards. By then, 

approximately 200,000 recipients had been wrongfully terminated, 

by the Government’s own reckoning. Many were eventually 

reinstated, but the episode caused heart-breaking and traumatic 

suffering. Many thousands of recipients and their families lost their 

livelihood, became homeless, and/or scraped by on the kindness of 

friends and relatives for long periods. Regrettably, South Africa is also 

familiar with large-scale delivery-failures in social welfare 

programmes. 

The disability-benefits programme was enacted under the Social 

Security Act, 60 which incorporates an administrative appeals 

procedure for correcting errors. Terminated recipients who were in 

fact eligible invoked these procedures. They were entitled to, and 

most eventually received, reinstatement for future benefits and 

59 

See US Constitution, 5th amend. (‘No person shall ... be deprived of life, liberty, 

or property [by the Federal Government], without due process of law’). The 

Fourteenth Amendment guarantees Due Process in deprivations undertaken by 

state governments. 

60 Social Security Act, Title II (1935), principally codified at 42 USC §§ 420 - 425.


retroactive payment of past benefits wrongfully withheld. The 

problem was that these were the only remedies available under the 

administrative system. In other words, a fully successful claimant 

received at the end of the administrative process only what she was 

entitled to in the first place. The procedure made no provision for 

mandatory or injunctive relief, it afforded no compensation for the 

turmoil and dislocation caused by wrongful termination, and it did not 

entertain, let alone remedy, constitutional claims. 

The claimants in Chilicky were eligible but wrongfully terminated 

benefits-recipients who sought monetary damages for ‘emotional 

distress and for loss of food, shelter and other necessities proximately 

caused by [the Government’s] denial of benefits without due 

process’. 61 Consequential damages and the other remedies requested 

were not available through the administrative appeals system. 

Lacking any other vehicle, plaintiffs lodged their claim as a Bivens 

tort. 

Chilicky differs from the cases van der Walt discusses in that the 

Social Security Act was not enacted in order to ‘give effect’ to a 

constitutional right. The US Constitution contains no guarantee of 

social security such as appears in FC section 27(1)(c). This should not 

greatly alter the analysis of Chilicky, however, because the claim was 

grounded on a constitutional right, the right to Due Process. 

Indisputably, the remedial scheme attached to the disability benefits 

programme must provide constitutional Due Process — in that sense, 

it is a statute designed in relevant part to ‘give effect’ to a 

constitutional right. 

In a 1970s case called Arnett, Justice Rehnquist floated a theory 

that recipients of government largesse or public employment must 

‘take the bitter with the sweet’. 62 By this he meant that any nonconstitutional 

benefit created or extended by law is qualified and 

limited by the procedures specified in the legislation for removal or 

discontinuance of the benefit, even if such procedures do not 

measure up to the minimum requirements of constitutional Due 

Process. Rehnquist J reasoned that no deprivation of property occurs 

if a benefit is terminated or removed according to the procedures for 

removal specified in the law creating the benefit. In these 

circumstances, the claimant is deprived of nothing to which she has a 

61 

62 

Chilicky v Schweiker 796 F 2d 1131 1134 (9th Cir. 1986) (quoting plaintiffs’ 

complaint), reversed by Schweiker v Chilicky 487 US 412 (1988). Other remedies 

were sought, notably an interdict against the benefit cut-offs. The additional 

remedial requests dropped out of the case partly because it meandered through 

the courts for about six years, by which time Congress had halted the Reagan 

campaign to slash the rolls and had provided enhanced procedural safeguards. 

See Arnett v Kennedy 416 US 134 151 - 54 (Rehnquist J for the plurality), 

rehearing denied 417 US 977 (1974).


legal claim of entitlement. Therefore, Rehnquist J argued, the 

strictures of Due Process are inapplicable. 63 

A majority of the Court emphatically rejected Justice Rehnquist’s 

thesis, both in the Arnett decision itself and on subsequent 

occasions. 64 Under settled doctrine, even though the government is 

not constitutionally obliged to confer a certain benefit, once it elects 

to do so, it may not remove or terminate the benefit except by 

procedures consistent with constitutional Due Process. The 

Constitution, not the statute, sets the procedural floor. 

Evidently the Court forgot this when it decided Chilicky, which 

shows legislatures how to immunise their work from constitutional 

review. Therein lies a warning for South Africans. The Chilicky Court 

held that ‘[w]hen the design of a Government program suggests that 

Congress has provided what it considers adequate remedial 

mechanisms for constitutional violations that may occur in the course 

of its administration, we have not created additional Bivens 

remedies’. 65 In the result, the Court precluded the plaintiffs from 

proceeding with an action on the Constitution. Under the banner of 

subsidiarity, they were remitted to statutory remedies which, it was 

common cause, did not even consider, let alone remedy, claims of 

constitutional injury. Under American-style subsidiarity, constitutional 

claimants were channelled to and compelled exclusively to 

utilise a forum that did not entertain constitutional claims. This 

meant that the claimants could not obtain a hearing for their 

constitutional claims in any forum. In practical effect, the legislation 

simply erased their constitutional right to Due Process. 

One might have thought that a self-respecting court would be 

embarrassed to decree something so absurd. But no, the US Supreme 

Court made its conclusion official: ‘the relief sought by the 

[claimaints] is unavailable as a matter of law ... ’. 66 In sum, when 

enacting a benefits programme, the legislature may render the Bill of 

Rights inoperative by attaching remedial procedures that claimants 

are exclusively constrained to utilise when challenging the 

63 See Arnett (n 62 above) (under constitutional text, a ‘deprivation’ is required to 

trigger Due Process guarantees). 

64 

See, eg, Cleveland Board of Education v Loudermill 470 US 532 (1985). 

65 Chilicky (n 49 above) 423. The Court could find no legal distinction between 

Chilicky and Bush: ‘[W]e declined in Bush to create a new substantive legal 

liability [on the Constitution] ... because we [were] convinced that Congress is in 

a better position [than the judiciary] to decide whether or not the public interest 

would be served by creating it. That reasoning applies as much, or more, in this 

case ... In neither case ... does the presence of alleged unconstitutional conduct 

that is not separately remedied under the statutory scheme imply that the 

statute has provided ‘no remedy’ for the constitutional wrong at issue (426 – 428) 

(citations and internal quotation marks omitted; italics in original). 

66 Chilicky (n 49 above) 429.


constitutionality of the programme’s administration, even though 

those procedures expressly disallow constitutional challenges of any 

kind. Chilicky allows Congress to stamp ‘void’ on the Bill of Rights. 67 

The Court’s deference to the statutory remedies fully accords 

with subsidiarity thinking. True, the Court’s articulated rationale 

invoked the theory of institutional competence (the legislature is 

better situated than the courts to study complicated policy matters 

and to engineer comprehensive solutions) rather than classical 

separation-of-powers doctrine (the legislature is representative and 

acts democratically). The distinction might be highly significant in 

some contexts, but here the two theories seem closely intertwined. 

Subsidiarity draws on both. Substitute ‘Parliament’ for ‘Congress’, 

and ‘government agencies’ for ‘state agencies’, and one can easily 

imagine a South African court committed to the SANDU principle 

denying an action on the Constitution with words similar to the 

following: 

Congress ... has addressed the problems created by state agencies’ 

wrongful termination of disability benefits. Whether or not we believe 

that its response was the best response, Congress is the body charged 

with making the inevitable compromises required by the design of a 

massive and complex welfare benefits program ... Congress has 

discharged that responsibility to the extent that it affects the case 

before us, and we see no legal basis that would allow us to revise its 

decision. 68 

I am confident that van der Walt would say that Chilicky goes too far. 

His version of subsidiarity is more flexible than its US counterpart and 

does not rule out constitutional challenges to effect-giving statutes. 

But these subtleties can easily be lost in less able hands than his. 

67 Chilicky is in a line of disastrous Nixon-Reagan era decisions that eviscerated the 

Due Process clauses. In Bishop v Wood 426 US 341 (1976), the Court held that a 

public employer may discharge an employee for cause without providing a 

hearing, an opportunity to respond to the charges against him, or any other 

minimum component of constitutional Due Process, if the government structures 

its relationship with its employees on an at-will basis. In Paul v Davis 424 US 693 

(1976), a local police department distributed circulars to merchants containing 

the names and photographs of individuals identified by the police as known to be 

actively engaged in shoplifting. The names were selected ex parte, without 

notice to the individuals involved, hearing, or any other opportunity for the 

accused to respond. The Supreme Court found no constitutional deficiency in this 

procedure, holding that those who were ‘merely’ publicly branded as criminals 

without trial suffered no impairment of any interest protected by the Due Process 

clauses. Justice Brennan remarked in dissent that ‘[t]he potential of today’s 

decision is frightening for a free people’ (721). 

68 Chilicky (n 49 above) 429 (citations omitted). The legal basis the Court had 

difficulty seeing is, of course, the Constitution’s Fifth Amendment and the Court’s 

own case-law.

5 Conclusion 


I conclude with a hypothetical problem aimed to identify interpretive 

questions that await South Africa’s courts. My purpose is to show the 

limitations of subsidiarity analysis, but also to offer a platform for 

discussion of issues that may well arise as the jurisprudence of SANDU 

and Bato Star evolves. In fact, the hypothetical bears more than a 

passing resemblance to matters that have already concerned South 

African courts. 69 

Imagine that Parliament enacts legislation to give effect to the 

right to social security guaranteed by FC section 27(1)(c). In addition 

to setting forth the substantive content of the programme, the 

legislation contains a remedial scheme intended to address and 

remedy all wrongs committed by the government in connection with 

the benefits programme, including any constitutional violations that 

may occur. Successful claimants are entitled to retroactive payment 

of benefits wrongfully withheld, access to the programme if it was 

wrongfully denied, and restoration to the programme in the event of 

wrongful termination. In addition, a claimant may receive monetary 

compensation for consequential damages, including damages 

proximately caused by a constitutional violation, up to a maximum of 

R100. The statutory procedure makes no provision for mandatory 

relief and does not allow class actions. These restrictions were 

included after careful study, consultation with experts and affected 

stakeholders, and plenary discussion in Parliament. The drafters are 

in good faith convinced that it is necessary to limit litigation and 

remedies in connection with delivery problems in order to preserve 

scarce resources for the beneficial purposes of the programme. 

Parliament declares the statutory remedial procedure to be the 

exclusive avenue of redress for any grievance in connection with the 

programme. Neither the Promotion of Administrative Justice Act 

(PAJA) 70 nor any other statute applies, and any claim that might arise 

under the common law is expressly pre-empted. 

69 See, eg, Permanent Secretary, Department of Welfare v Ngxuza 2001 4 SA 1184 

(SCA), 2001 10 BCLR 1039 (SCA); MEC, Department of Welfare v Nontembiso 

Norah Kate 2006 2 All SA 455 (SCA). In Kate, the SCA sustained an award of 

constitutional damages in a case of excessive delay in the processing of a social 

grant application, as against the view that the claimant should be restricted to 

statutory or delictual remedies. See (para 27) stating that the relief permitted by 

FC sec 38 (enforcement of Bill of Rights): ‘is not a remedy of last resort, to be 

looked to only where there is no alternative — and indirect — means of asserting 

and vindicating constitutional rights. While that possibility is a consideration to 

be borne in mind in determining whether to grant or to withhold a direct section 

38 remedy, it is by no means decisive, for there will be cases in which the direct 

assertion and vindication of constitutional rights is required ... [T]he endemic 

breach of the rights that are now in issue justifies — indeed, it calls out for — the 

clear assertion of their independent existence’. 

70 Act 3 of 2000.


An eligible recipient of welfare benefits is arbitrarily terminated 

from the program for no good reason and without notice or an 

opportunity to be heard. She suffers consequential damages in an 

amount greater than R100 and asserts that the cap on remedies 

renders the statutory remedial procedure constitutionally 

inadequate, citing FC section 27(1)(c) itself and perhaps FC section 33 

(just administrative action), FC section 34 (access to courts), and 

other provisions of the Bill of Rights. May she prosecute a claim on the 

Constitution for the full amount of her damages? Suppose an 

additional 100,000 eligible recipients experience similar problems 

due to the incompetence or malfeasance of the authorities. May they 

sue on the Constitution as a class to remedy their own injuries and 

interdict the authorities from future, administrative misbehaviour? 

If we conclude that these claimants may not proceed on the 

Constitution and are limited to working within the statutory 

procedure, we vindicate the pro-legislature, pro-representation spirit 

of subsidiarity at the expense of constitutional supremacy and 

substantive constitutional rights. We put the legislature in a position 

effectively to define rights entrenched by a Constitution intended to 

constrain it. And because an interdict is precluded by the statute, we 

effectively prevent the courts from granting a timeous remedy that 

might induce the authorities to straighten out the mess. But suppose 

we conclude that these suits may go forward. Would we not be saying, 

in practical effect, that we will observe the tenets of subsidiarity 

except in cases where we prefer not to? And if this is what it comes 

down to, what has subsidiarity taught us?

‘OH WHAT A TANGLED WEB WE WEAVE ...’ * 

HEGEMONY, FREEDOM OF CONTRACT, 

GOOD FAITH AND TRANSFORMATION - 

TOWARDS A POLITICS OF FRIENDSHIP IN THE 

POLITICS OF CONTRACT 

* ‘... when first we practice to deceive.’ Sir Walter Scott Marmion (1825) Canto VI 

Stanza XVII, 349. 

�� BCom(Law) LLB LLD(Pretoria). Senior lecturer, Department of Private Law, 

University of Cape Town. The author would like to thank the following people for 

valuable discussions and comments on earlier drafts of this essay: Karin van Marle, 

Danie Brand, Drucilla Cornell, Dennis Davis and Karl Klare. 

155 

AJ Barnard-Naude** 

‘Q6: “ ... and I wondered how you think your concept of friendship, your 

non-canonical concept of friendship, can address economic exclusion in 

its most extreme form which is the exclusion by those who own capital 

of everyone else. You can’t ask those who own capital to be hospitable 

...”. 

JD: “I ask them nevertheless [laughter].” 

Q6: “It’s naïve to ask them, it’s a naïve request.” 

JD: “Perhaps, but I still do.”’ 1 

‘One must speak for a struggle for a new culture, that is, for a new 

moral life that cannot but be intimately connected to a new intuition of 

life, until it becomes a new way of feeling and seeing reality.’ 2 


The pivot on which this contribution hinges is hegemony in the 

context 3 of the South African law of contract. I must point out at the 

1 From ‘Politics and friendship: A discussion with Jacques Derrida’ available at 

http://www.hydra.umn.edu/derrida/pol+fr.html (accessed 29 May 2008). 

2 A Gramsci Selections from cultural writings (1985) 98. 

3 And it will of course be unavoidable when considering the lexicography of 

hegemony, that the concept of conning (in both its senses) — and the conning of/ 

by the text; the context — will raise its head.

156 Hegemony, freedom of contract, good faith and transformation 

outset that the purpose or aim of this ‘hinging’ on hegemony and/in 

the South African law of contract, is not necessarily to take issue with, 

underscore or re-illustrate the claim that contract holds hegemony — 

in the sense that it ‘occupies a privileged position’ — in the South 

African legal system. 4 Nevertheless, it might very well be the case 

that Alfred Cockrell’s admirable illustration of the hegemony of 

contract in the South African legal system can be explained also by 

the particular focus on hegemony in the South African law of contract 

on which I embark in this contribution. In other words, the hegemony 

of contract might well be attributable to the hegemony within 

contract. This is the case because the hegemony within might very 

well sustain the hegemony of and perhaps also vice versa. (A symbiosis 

comes to mind.) This hegemony within contract relates to a particular 

understanding of freedom of contract. Thus, the focus and point of 

departure in this contribution will be the hegemony — or hegemonic 

order(ing) 5 — of/in a particular understanding of (freedom of) 

contract. This understanding of freedom of contract — and indeed of 

contract itself — is the individualist, rule-committed 6 — liberal 7 — 

understanding of freedom of contract. 

The enactment of the Constitution, 8 the transformative hopes it 

disseminates and the view that its normative framework is explicitly 

post-liberal 9 occasions an opportunity for a re-evaluation of and a 

challenge to the individualist (hegemonic) understanding of freedom 

of contract, its concomitant commitments and, as we shall see, the 

tangled web it weaves in order to sustain the false consciousness on 

4 A Cockrell ‘The hegemony of contract’ (1998) 115 South African Law Journal 286 

286. 

5 

I want to insist on this marking according to which both senses of the word ‘order’ 

remain constantly in play, namely order as socio-political concept of hierarchy 

and order as sequential arrangement. 

6 D Kennedy ‘Form and substance in private law adjudication’ (1976) 89 Harvard 

Law Review 1685 1685: ‘individualism seems to harmonise with an insistence on 

rigid rules rigidly applied.’ 

7 See the description of the liberal self — grounded in individualist morality — 

offered by F Michelman ‘The subject of liberalism’ (1994) 46 Stanford Law Review 

1807 1812. Also see van K Marle & D Brand ‘Enkele opmerkings oor formele 

geregtigheid, substantiewe oordeel en horisontaliteit in Jooste v Botha’ (2001) 

12(3) Stellenbosch Law Review 408 415. 

8 

Constitution of the Republic of South Africa, 1996. 

9 See H Botha ‘Democracy and rights: Constitutional interpretation in a postrealist 

world’ 2000 (63) Tydskrif vir die Hedendaagse Romeins-Hollandse Reg 561 574. 

Botha notes three reasons why the Constitution requires more than a classical 

liberal interpretation. First, he points out that the Constitution contains a 

commitment to an open, value-orientated, participatory democracy. This is a 

commitment that cannot be reconciled with the reduced concept of democracy 

which pervades liberal theory. Secondly, Botha argues that the Constitution does 

not support a liberal conception of rights as boundaries between the individual 

and the collective; the rights in the Bill of Rights have a contingent and nonabsolute 

meaning and to that extent they do not operate as a shield against 

government intervention or as trumps over collective interests. Thirdly, the 

Constitution is structured such that it requires a far more activist stance of the 

judiciary than what would be acceptable under a liberal interpretation (574-576).


which its legitimacy turns. However, in order to embark upon such a 

re-evaluation it will be necessary to begin by revisiting the concept of 

hegemony as encountered in the work of Antonio Gramsci as well as 

the expansions and reformulations of Gramsci’s concept as it relates 

to law. Second, I will articulate more fully what is meant with and 

what are the commitments of the hegemonic liberal understanding of 

contract and freedom of contract. Third, I will track the rise of the 

hegemonic order(ing) in the South African law of contract. I will argue 

that the story of its rise can and must be considered as the story of 

the fall (that is, the marginalisation) of values and concepts in the law 

of contract that came to be understood as somehow threatening the 

hegemonic order(ing) (notably in this regard, good faith). 10 An 

ongoing moment in this contribution interrogates the way in which the 

hegemony confronts the Constitution (both as an interruption, 11 as a 

value system and as law) in order continuously to pose the question 

whether the hegemonic order(ing) within the South African law of 

contract (and the adjudication that keeps this hegemony in place) 

should be resisted in the name of the spirit, purport and objects of 

the Constitution. 

A primary contention in this article will be that the hegemony of 

a radically individualist understanding of freedom of contract is as 

firmly in place under the newly constituted legal order as it was under 

the apartheid legal order in South Africa. The law of contract is still 

seen as an institution of separation rather than relation, of apartness, 

rather than togetherness. That ‘nothing changed again’ 12 is evident in 

a series of cases decided after the reconstitution of the South African 

legal order. 13 What will also become evident, however, is the series 

of contradictions and deconstructions that emerge in the course of 

10 See in this regard G du Toit ‘The significance of postmodern theories of 

interpretation for contractual interpretation: A critical analysis’ unpublished LLD 

thesis, University of Stellenbosch, 2006 234-235. 

11 See, for instance, J Gilbert-Walsh ‘Deconstruction as narrative interruption’ 

(2007) 38(4) Interchange 317. 

12 For a variety of reasons I want to draw explicit attention here to Karin van Marle’s 

discussion of the inscription ‘and nothing changed again’ on one of the etchings in 

Dianne Victor’s celebrated series Disasters of Peace. Van Marle remarks as 

follows: ‘“And nothing changed again”, portrays the state of mind of many South 

Africans who do not experience easy living in post-apartheid South Africa. For 

them, nothing has changed again.’ See K van Marle ‘Art democracy and 

resistance: A response to Professor Heyns’ (2005) Pulp Fictions: Disasters of 

13 

Peace: An Exchange 15 26. 

The most important of these cases are De Beer v Keyser 2002 1 SA 827 (SCA) (De 

Beer), Brisley v Drotsky 2002 4 SA 1 (SCA) (Brisley), Afrox Healthcare Bpk v 

Strydom 2002 6 SA 21 (SCA) (Afrox Healthcare), Johannesburg Country Club v 

Scott and Another 2004 5 SA 511 (SCA) (Johannesburg Country Club), Juglal NO v 

Shoprite Checkers t/a OK Franchise Division 2004 5 SA 248 (SCA) (Juglal N0), 

South African Forestry Co Ltd v York Timbers 2005 3 SA 323 (SCA) (South African 

Forestry Co), Napier v Barkhuizen 2006 4 SA 1 (SCA) (Napier) and Barkhuizen v 

Napier 2007 5 SA 323 (CC) (Barkhuizen).


the marginalisation of good faith in the South African law of contract, 

which of course mirrors the rise of the hegemonic order(ing). 

In the last part of the article I will argue why the hegemonic 

order(ing) of contract, mirrored by the marginalisation of good faith 

can and should be questioned in light of the transformative 

commitments inherent in the spirit of the Constitution. I will argue 

that good faith as the ethical element of contract directly involves the 

constitutional ideal of civic friendship in the South African law of 

contract. Civic friendship as an aspirational ideal of the new legal 

order then enjoins us to transform the hegemonic order(ing) in the 

law of contract by way of a direct involvement of good faith. 14 I see 

the starting point of this transformation as a call for the politics of 

friendship in the politics of contract. In this context, I will argue that 

the provisions of the Draft Consumer Protection Bill stand in stark 

contrast to the commitments of the hegemonic understanding of 

freedom of contract and, although inadequate, these provisions point 

to a future law of contract that will have to be far more concerned 

with the ethical element of contract than it has been up to now. 

These provisions require a transformed understanding of freedom of 

contract in which good faith as the ethical element of contract comes 

to again play a constitutive role in the understanding of freedom of 

contract. This has obvious transformative implications for 

adjudication in the South African law of contract. In this context I will 

take issue with the Constitutional Court’s 2007 judgement in 

Barkhuizen in which it followed a hesitant approach to the transformative 

role of good faith in the South African law of contract. 

Ultimately my contention will be that good faith interpreted in line 

with a transformative reading of the law of contract mandates a 

relational interpretation of freedom of contract that is far removed 

from its standard libertarian version. 

2 Gramsci, hegemony and the law 

In order to come to terms with the present focus on the hegemonic 

understanding of (freedom of) contract, it is necessary — as a point of 

departure — to review the development of hegemony as a conceptual 

tool of critique. For the benefit of the argument that will follow, it 

will also be necessary to enquire into contemporary views on the 

relationship between hegemony and law. 

14 I realise and concede that the theoretical and methodological framework 

developed here is contested. Apart from the justifications offered in the text, the 

primary focus of this article (adjudication on the interaction between the 

common law of contract and the Constitution in 2007) does not allow me to fully 

defend this approach. In this regard, I intend to embark more fully on such an 

explanation in future work.


Hegemony as a critical concept was originally developed in the 

work of Antonio Gramsci 15 who employed the word to explain how 

capitalism maintains control ‘not just through violence and political 

and economic coercion, but also ideologically, through a culture in 

which the values of the bourgeoisie became the “common sense” 

values of all.’ 16 In Gramsci ideology operates in such a way that the 

dominant class in fact instills a false sense in the dominated class that 

the good of the dominant class is also the good of the dominated class. 

This condition serves conformity and perpetuates a status quo. 17 John 

and Jean Comaroff aptly describe the cultural effect of hegemony: 

[W]e take hegemony to refer to that order of signs and practices, 

relations and distinctions, images and epistemologies — drawn from a 

historically situated cultural field — that have come to be taken-forgranted 

as the natural and received shape of the world and everything 

that inhabits it. 18 

Gramsci insisted that hegemony is only transformed once the 

dominated class realised the importance of creating its own culture 

which depends on the realisation that it had unwittingly adopted the 

values of the bourgeoisie and that its own values in fact opposed the 

values of the bourgeoisie. 19 Breaking hegemony thus depends 

fundamentally on the contestation of values and ideals and perhaps 

more importantly on the contestation of the relationships between 

and the content of these values and ideals. 20 

When it comes to the relationship between law and hegemony, 

Litowitz points out that law fulfils a dual function in relation to 

hegemony. 21 On the one hand, the law represses (through the State’s 

monopoly on physical force) any disturbance that might challenge the 

hegemony. On the other hand, the law authorises and legitimates the 

status quo and so produces and perpetuates the hegemony without 

the need to revert to physical force. 22 This contribution will focus 

primarily on this second dimension of hegemony, as I shall interrogate 

15 

A Gramsci Selections from the prison notebooks trans Q Hoare & GN Smith (1971). 

16 JM Russel Philosophical classics (2007) 94. 

17 As above. Also see TR Bates ‘Gramsci and the theory of hegemony’ (1975) 36(2) 

Journal of the History of Ideas 351 352: ‘The concept of hegemony is really a very 

simple one. It means political leadership based on the consent of the led, a 

consent which is secured by the diffusion and popularisation of the world view of 

the ruling class.’ 

18 J Comaroff & J Comaroff Of revelation and revolution: Christianity, colonialism 

and consciousness in South Africa (1991) 23. 

19 

Russell (n 16 above) 95. 

20 Bates (n 17 above) makes the point that the theory of hegemony recognises that 

man is ruled not just by force but also by ideas. The author also points to the 

Marxist mantra that ‘the ruling ideas of each age have ever been the ideas of its 

ruling class.’ 

21 D Litowitz ‘Gramsci, hegemony and the law’ (2000) Brigham Young University Law 

Review 515 531. 

22 As above.


specifically how law — a particular version of contract law — endorses 

and thus normalises or legitimates the individualist or libertarian 

worldview. For Gramsci as for Litowitz this second dimension of 

hegemony is the more dangerous one as it carries within it the 

potential to paralyse all resistance and silence all questioning. 

Litowitz argues that the concept of hegemony requires the attention 

of students of law, precisely because it possesses this profound 

‘ability to induce submission to a dominant worldview.’ 23 

Whereas Gramsci originally maintained that law is a hegemonic 

tool of the dominant class, Litowitz argues that when it comes to 

contemporary legal analysis a more nuanced notion of hegemony is 

required. Litowitz contends that under conditions of post-modernity 

we need an understanding of hegemony that avoids ‘Gramsci’s 

reliance on orthodox Marxist categories that are no longer tenable.’ 24 

Litowitz argues that postmodern theory brought about a profound but 

gradual shift in the understanding of hegemony away from essentialist 

Marxist discourse rooted in rigid class distinction. 25 Postmodern 

theorists understood that oppression does not exclusively exist as the 

imposition of a dominant class and recognised that oppression is often 

invisible and can exist merely in the silencing of alternative 

perspectives. 26 Postmodernism thus changed the operative 

terminology in discourse about hegemony from ‘“class/exploitation” 

to “discourse/marginalisation.”’ 27 

Litowitz explains that this development in hegemony discourse 

can help legal scholars to recognise that law is hegemonic at the 

meta-level (in that it always already represents and legitimises 

dominant interests) while contesting hegemony at the micro-level. 

For Litowitz law participates in the establishment of hegemony 

through the dissemination of a ‘dominant code or map’ 28 that relies 

on ‘unchallenged background assumptions that undergird the law.’ 29 

Thus law always already is hegemonic in that it ‘induces people to 

comply with a dominant set of practices and institutions without the 

threat of physical force’ 30 because people are lulled into a false 

consciousness about the legitimacy of law which causes them 

uncritically to accept legal practices and institutions that might in 

fact be harmful and exploitative. In other words the meta-level at 

which law is hegemony sustains and keeps in place law’s hegemonies 

at micro-level — and vice versa. 

23 

Litowitz (n 21 above) 516. 

24 Litowitz (n 21 above) 518. 

25 Litowitz (note 21 above) 533. 

26 

Litowitz (n 21 above) 534. 

27 As above. 

28 n 21 above, 517. 

29 

As above. 

30 As above.


Under the force of hegemony the assertion that ‘law ... is also the 

site of everyday resistance and struggle’ 31 comes to be either 

dismissed and ignored or actively undermined and suppressed. 

However, contesting the hegemonies upon which the law as hegemony 

is founded can be the first step in raising consciousness upon which 

refusal and resistance of things as they are and supposedly always 

have been, can be built. 32 This is the central motivation behind the 

present focus on the hegemonic order(ing) in the South African law of 

contract. I do not want to suggest that the Marxist roots of hegemony 

discourse can or should be simply or finally denied or that escape from 

these roots should be relentlessly pursued. 33 As we shall see, the 

South African law of contract vividly reflects the Gramscian insight 

that bourgeois best interests come to be uncritically accepted as the 

best interests of the working classes and the poor through the 

phenomenon of false consciousness which generates false consent. 

However, what postmodern theory contributes to hegemony discourse 

is the notion of discourse marginalisation. I will argue in what follows 

that the rise of the hegemonic order(ing) in the law of contract occurs 

through an explicit judicial marginalisation of discourses on equity, 

fairness and good faith. 34 In the context of the reconstitution of the 

South African legal order this raises important and difficult questions 

about the potential of the South African law of contract (in its current 

guise) to contribute to the project of transformative 

constitutionalism. 35 

31 K Calavita ‘Book review: M Lazarus-Black et al Contested states: Law, hegemony 

and resistance’ (1996) 25(1) Contemporary Sociology 94 94-96. 

32 Litowitz (note 21 above) 518. Also see D Kennedy ‘Antonio Gramsci and the legal 

system’ (1982) VI(1) ALSA Forum 32 37: ‘What the Gramscian analysis suggests is 

that one of the ways out of the reform/revolution problem in all these legal 

activities is to try to develop, at the level of conscious communication with other 

people, the extent to which they are letting their goals be perverted by the 

hegemonic false consciousness generated by law.’ 

33 See in this regard E Laclau & C Mouffe Hegemony and socialist strategy (2ed, 

34 

35 

2001) viii. 

A detailed discussion of the link between, on the one hand, the ideas that 

predominate in judicial decisions in the law of contract and, on the other, the 

beliefs about social organisation that are dominant in elites and the general 

populace, falls beyond the scope of this article. But see, for instance, E Mensch 

‘Freedom of contract as ideology’ (1981) 33 Stanford Law Review 753 757 who 

points out that the rise of the classical view of contract in the courts can be 

attributed to the popularity of Victorian moralism amongst the elites of the time. 

There are, in addition, several places in this article where I do not use the term 

‘hegemony’ in any purist Gramscian sense of the word but rather in its more 

general sense to denote predominance. Where the word is primarily used in this 

way there is nevertheless a secondary allusion to the Gramscian notion operating 

in the background. 

See K Klare ‘Legal culture and transformative constitutionalism’ (1998) 14 South 

African Journal on Human Rights 146 146.


3 The rise of the hegemonic order(ing) in 

contract 

3.1 The rise of the market economy, the subjectivity of 

value and the will theory 

The present hegemonic order(ing) in contract begins, of course, in the 

late 18 th century. At this point in history, freedom of contract still had 

an explicit and intricate connection with principles of morality such 

as fairness, conscience and justice. 36 The maxim pacta servanda sunt, 

in this understanding of freedom of contract, was still regarded as 

‘the imperative of good faith in contracts’. 37 As Atiyah points out, 

this meant that: ‘the Courts were, at that time, still more interested 

in seeing that parties to a contract made a fair exchange, than they 

were in enforcing bare promises.’ 38 On this approach, contracts were 

not enforced meticulously. Instead, they were meticulously subjected 

to enquiries into the fairness of the exchange. 

The problem with this approach to freedom of contract was that 

by the nineteenth century it became regarded as anachronistic, 

because it did not serve the interests of the rising commercial classes 

of the emerging market economies. 39 In these emerging economies 

value became regarded as entirely subjective. Horwitz indicates that, 

for this very reason, the late 18th century businessman became 

interested in a freedom of contract that would guarantee the express 

value of the agreed performance, regardless of the substantive 

equality of the exchange. 40 This meant that, as the market economies 

and the commercial classes became dominant, the law of contract as 

it existed up to that point came to be considered unsuited for the 

purposes of the new economies. In short, the eighteenth century 

equitable understanding of freedom of contract was far too invested 

in the concept of objective value. The general view of the eighteenth 

36 MJ Horwitz ‘The historical foundations of modern contract law’ (1974) 87(5) 

Harvard Law Review 917 920. The eighteenth century approach is stated 

succinctly in Evans v Llewellyn [1787] 29 ER 1191 (Evans) 1191: ‘if the party is in 

a situation in which he is not a free agent and is not equal to protecting himself, 

this Court will protect him.’ 

37 Van der Merwe et al Contract general principles (3 ed, 2007) 21 n 7 quoting H 

Eichler Die rechtslere vom vertrauen (1950) 8 (emphasis added). C Visser ‘The 

principle pacta servanda sunt in Roman and Roman-Dutch law’ (1984) 101 South 

African Law Journal 641 647 points out that originally pacta servanda sunt is 

integrally connected to good faith: ‘it would therefore appear that both canon 

law and the lex mercatoria from very early on accepted a general principle of 

sanctity of contract, based upon a moral conception of good faith ...’ (emphasis 

added.) 

38 PS Atiyah The rise and fall of freedom of contract (1979) 438. 

39 Horwitz (n 36 above) 945-946. 

40 MJ Horwitz ‘The triumph of contract’ in AC Hutchinson (ed) Critical legal studies 

(1989) 104 106.


century was that substantive justice existed to ensure that legal 

subjects do not abuse the legal system to exploit each other. 41 In the 

nineteenth century the role and function of contract was no longer 

regarded as guaranteeing the justice of the exchange but to enforce 

so-called ‘willed’ transactions inter partes. 42 So arises the so-called 

‘modern’ or ‘classical’ law of contract in which the source of 

contractual obligation shifted from equitable exchange to 

consensus. 43 

The pervasive political view of the nineteenth century — that of 

unlimited liberal freedom — informed and provided the justification 

for a conception of ‘freedom’ of contract as the shibboleth for the 

privileging of the parties’ wills in conceptualising their 

relationships. 44 Once the concept of value became perceived as 

entirely subjective, the only basis for ascribing value could be 

concurrent, individual will. Principles of substantive justice came to 

be regarded as necessarily arbitrary and an uncertain standard of 

value. 45 At the point where intrinsic value could no longer be ascribed 

to anything, so the argument went, no substantive measure could 

exist by reference to which it could be determined whether one party 

was exploiting the other. 46 On this view, the appearance of consensus 

came to represent the ‘evidence’ that the contractual exchange itself 

was a fair one. 

However, the loss incurred as a result of this shift in emphasis was 

particularly acute. Gordley indicates that nineteenth century jurists 

in fact entirely eliminated the concept of virtue from their discussion 

of contract and were left with the will alone. 47 The motto of modern 

contract thus became this: ‘a man is obliged in conscience to perform 

a contract which he has entered into freely, although it be a hard one 

...’. 48 As Hahlo so aptly described it: Darwinian survival of the fittest, 

the law of nature, also became the law of the marketplace. 49 

41 Mensch (n 34 above) 753, 756. 

42 As above. 

43 Horwitz (n 36 above) 946. 

44 Horwitz (n 36 above) 917. Also see J Gordley The philosophical origins of modern 

contract doctrine (1991) 161 where the point is made that consensus was applied 

almost exclusively to explain the basis of contractual obligations. 


46 As above. 

47 Gordley (n 44 above) 162. 

48 IJ Powell ‘Essay upon the law of contracts and agreements’ (1790) as quoted in 

Horwitz (n 36 above) 917. 

49 HR Hahlo ‘Unfair contract terms in civil-law systems’ (1981) 98 South African Law 

Journal 70.


3.2 The role of individualism 

The above formulation evinces the understanding of freedom of 

contract of advancing and advanced liberal capitalism underpinned by 

what is euphemistically called the ‘morality’ of individualism. 50 

Individualism accepts as given a world of independent individuals who 

are encouraged to prefer the pursuit of self-interest rigorously. A 

consideration or sensitivity for the interests of others falls outside of 

the aims of this way of life, although one should be prepared to obey 

the rules that make it possible to co-exist with other self-interested 

individuals. 51 Other than this, the individual is entirely self-reliant. 

His conduct conforms to the belief that other individuals in the 

community are themselves motivated only by pure self-interest. 52 As 

regards individualism’s economic manifestation, Cockrell puts it 

succinctly: ‘[I]ndividualism assumes a world of traders who meet 

briefly on the market floor, where they engage in discrete and furtive 

transactions.’ 53 This view is often described as market-individualism 

and reveals that the capitalist idea of the market and the ‘moral’ idea 

of individualism are heavily invested in each other. 54 

Individualist beliefs regarding the role of law in society follow 

from the above convictions. Individualism holds that the law cannot 

impose upon legal subjects as a group the liability of shared profits or 

loss. 55 The law merely fixes the boundaries of individual freedom by 

defining and enforcing rights. 56 In the context of the law of contract, 

individualism believes that the parties create their own law through 

agreement (consensus) which is itself a manifestation of the 

individual’s autonomy. Mensch refers to this phenomenon as a ‘magic 

moment of formation, when individual wills created a right whose 

enforcement was necessary for the protection of free will itself.’ 57 

Contractual liability is thus only determined by the formal agreement 

(consensus) of the parties. The law of contract in an individualistic 

world, to borrow from Macaulay, provides the glue that binds 

50 

Kennedy (n 6 above) 1774. Kennedy goes so far as to argue that ‘[t]he “freedom” 

of individualism is negative, alienated and arbitrary. It consists in the absence of 

restraint on the individual's choice of ends, and has no moral content whatever.’ 

51 

JM Feinman ‘Critical approaches to contract law’ (1983) 30 UCLA Law Review 829 

839; Kennedy (n 6 above) 1713. 

52 Kennedy points out that the individualist ethic should be distinguished from the 

egotistical ethic in the sense that the individualist ethic has a strong positive 

‘moral’ content whereas the egotist believes that it is entirely impossible and 

undesirable to place any limits on the perusal of self-interest. See Kennedy (n 6 

above) 1714–1715. 

53 A Cockrell ‘Substance and form in the South African law of contract’ (1992) 109 

South African Law Journal 40 41. 

54 

See CJ Pretorius ‘Individualism, collectivism and the limits of good faith’ (2003) 

66 Tydskrif vir die Hedendaagse Romeins-Hollandse Reg 638 639. 

55 Kennedy (n 6 above) 1713. 

56 

Cockrell (n 53 above) 42. 

57 Mensch (n 34 above) 760.


individuals to their agreements. 58 This emphasis on the creation of an 

own law through the agreement causes individualism to take the view 

that courts should enforce agreements rather than intervene in 

relation to their terms on grounds of fairness or equity. 59 

With the rise of the market economy in the nineteenth century, 

the entire conceptual apparatus of the law of contract thus developed 

around this understanding of the relationship between the individual 

and society, the individual and the State and the individual and the 

law. 60 As Feinman indicates: 

Classical legal thought imagined a world of independent individuals, 

each of whom acts within a broad sphere of legal autonomy to pursue 

her own self-interest. The market was the model of social organisation 

and the acquisitive capitalist the paragon of personal behavior. 61 

In the law of contract this thought established freedom of contract as 

the hegemony (predominance) of the will theory. Duncan Kennedy 

famously added that the hegemony of individualism on the level of 

substance, in turn established the hegemony of rules on the level of 

contract law’s formal commitments. 62 

But because of the fact that domination is implied in the concept 

of hegemony it would be inaccurate to contend that hegemony 

manages to liquidate its other entirely. This means that hegemony 

always implies a relation — a relation of domination and privileging, 

yes, but nevertheless, a relation. The trace of the Other thus always 

haunts any hegemony. The reason why this particular privileging or 

domination is hegemonic is because, as Cockrell indicates, ‘this 

privileging invariably proceeds on the basis that the preference for 

individualism and the rule-form is an axiomatic truth rather than a 

controversial premise in an ongoing argument.’ 63 The danger is that 

the hegemonic version of contract can lead us to think that this is the 

only order(ing) available in the law of contract — that freedom of 

contract grounded in individualism and the commitment to the ruleform, 

is the only version of contract that exists. 64 

58 

S Macaulay ‘An empirical view of contract’ (1985) Wisconsin Law Review 465 466. 

59 Pretorius (n 54 above) 640. 

60 As Dalton points out, the will theory of contract flows from the political doctrine 

of contractualism, which holds that ‘all restraint is evil and that the government 

that governs least is best.’ See C Dalton ‘An essay in the deconstruction of 

contract doctrine’ (1985) 94 The Yale Law Journal 997 1013. 

61 

J Feinman ‘The conservatives, the market and the common Law’ (2005) 37 

Review of Radical Political Economics 288 290. 

62 Kennedy (n 6 above) 1685: ‘[t]he substantive and formal dimensions are related 

because the same moral, economic and political arguments appear in each.’ 

63 Cockrell (n 53 above) 46. 

64 Pretorius (n 54 above) 644: ‘There is no difficulty in recognising that individualism 

is the primary ideology underlying the law of contract. The problem lies in 

suggesting that it is the only one.’


The individualistic or liberal worldview in the understanding of 

contract did not and could not manage to altogether absent the law 

of contract from alternative worldviews. Applied to our current 

context this insight would imply that the traces of the pre-nineteenth 

century understanding of freedom of contract could not be erased by 

either the domination of the commercial classes of the nineteenth 

century or by the courts that represented their interests. The old 

underlying moral conceptions on which the law of contract was 

founded prior to the domination of individualism could thus not be 

abandoned outright. 65 For one, the courts still naively wanted to 

believe that parties to the contract were in fact reaching consensus 

on its terms as honest, just, fair and non-exploitative persons (that is, 

persons who are in good faith). To this extent the courts still 

acknowledged external standards of justice. But the critical legal 

issue had shifted from whether the contract was fair to whether there 

objectively appeared to have been a ‘meeting of the minds’ between 

the contracting parties. 66 

Horwitz points out that although nineteenth century courts could 

not succeed in negating the ancient relation between natural law and 

contract law, they did succeed in setting up a system in which the 

courts could effectively ‘pick and choose’ which groups within the 

broad society they wanted to benefit in a given case. 67 The discourse 

of the nineteenth century managed to set-up an intellectual divide 

between the system of formal rules (associated with the ‘rule of law’ 

or law as rules) 68 and the ancient perceptions of morality and equity 

(which was seen as necessarily undermining the rule of law or 

perhaps, then, the law of rules). As already mentioned, the argument 

in Critical Legal Studies has been that this preference for rules was 

part of the classical laissez-faire individualist morality grounded in 

non-interventionism. 69 It is thus no coincidence that most of modern 

general contract law consists of rules by way of which one determines 

whether the will of the parties coincided and if not, rules that 

regulate the undoing of the exchange. 70 These rules are ordered 

around the doctrines of misrepresentation, duress, undue influence 

and so forth — theories that allegedly turn on the negation of the will 

— and do not visit explicitly — or explicitly do not visit — matters such 


66 

Horwitz (n 36 above) 955. 

67 As above. 

68 C Douzinas & A Geary Critical jurisprudence (2005) 7. 

69 

Kennedy (n 6 above) 1741. 

70 See Van der Merwe et al (n 37 above) 19-32.


as the fairness of the bargain or the good faith of the parties. 71 The 

theory of reasonable reliance on the appearance of consensus — as an 

alternative basis for the existence of a contract — is itself simply seen 

as a qualification of the will theory and not as a negation or in any way 

a complete alternative to it. 72 Yet, as Cockrell indicates, the origin of 

the defect in the defences based on the negation of the will does not 

in fact reside in the will of one of the parties, but rather in the 

improper conduct. The same reasoning applies to the theory of 

reasonable reliance. These defences, as Cockrell argues, are all 

concerned with the legitimacy of conduct, and one way to link them 

would be to say that they all amount to instances of bad faith on the 

part of one of the parties. 73 On the hegemonic understanding, 

however, the open-ended norms or standards of the law of contract 

cannot do this work. It cannot be directly employed or appealed to in 

the adjudication of contractual disputes. In this regard, the point that 

has repeatedly been made by our own courts in the recent past is that 

good faith, reasonableness and fairness, although subjacent to, or 

underpinning the law of contract, can only be used in the resolution 

of contractual disputes to the extent that they have become 

embodied in so-called crystallised rules of the law of contract. 74 We 

will return to this problematic assertion below but for now it is 

important to state the general contention that these assertions serve 

only (in very real political ways) to keep the hegemony firmly in 

place. 

4 The hegemonic order(ing) in the South African 

law of contract 

4.1 Public policy as handmaiden and adversary of freedom 

of contract 

Colonialism and imperialism of course carried the individualist 

understanding of freedom of contract grounded in the will to the 

numerous colonies. It is thus not surprising that the South African law 

of contract from its very inception reflects the hegemony of the will 

71 CFC Van der Walt ‘Die huidige posisie in die Suid-Afrikaanse reg met betrekking 

tot onbillike kontraksbedinge’ (1986) 103 South African Law Journal 646 658. Also 

see Dalton (n 60 above) 1001 who makes the point that dealing with unfairness via 

those constructs that affect the will of the parties (duress, misrepresentation, 

undue influence etc) effectively constitutes a privatisation of the public enquiry 

into contract when ‘the undoing of a defective deal [is] presented as depending 

upon the absence of will or intent rather than on mere inequivalence of 

exchange.’ 

72 Van der Merwe et al (n 37 above) 22. 


74 

See Brisley (n 13 above) 16B-C; Afrox Healthcare (n 13 above) 40H-41A; South 

African Forestry (n 13 above) 338J-339B; and Napier (n 13 above) para 7.


theory of contract. One of the most frequently quoted passages — ‘the 

most privileged statement of all’ 75 — justifying the privileging of (the 

‘will theory version’ of) freedom of contract as the basis of 

contractual obligation in the South African law of contract, is to be 

found in a late nineteenth century case from the English law, Printing 

and Numerical Registering Co v Sampson: 76 

If there is one thing more than another which public policy requires, it is 

that men of full age and competent understanding shall have the utmost 

liberty in contracting, and their contracts, when entered into freely and 

voluntarily, shall be held sacred and shall be enforced by Courts of 

Justice. Therefore you have this paramount public policy to consider — 

that you are not lightly to interfere with this freedom of contract. 77 

In Burger v Central South African Railways 78 Innes CJ ‘developed’ this 

basic premise and held in no uncertain terms that the South African 

law of contract does not recognise the right of a court to release a 

party to a contract from his obligations on considerations of 

fairness. 79 And this is precisely the point, (as will hopefully become 

clearer below): when the focus is on the South African law of contract 

particularly, the rise of the hegemonic order(ing) of (freedom of) 

contract should be considered in the context of the fall of numerous 

values, principles and rights that came to be understood as freedom 

of contract’s adversaries and so became regarded as opposed to it. 

A doctrinal aspect that has remained uncontroversial in the grand 

narrative of the South African law of contract has been the 

recognition that contracts concluded contra bonos mores and/or 

contrary to public policy fall foul of the validity requirement of 

legality and are therefore either void or unenforceable. 80 The extent 

of this power to declare void or unenforceable contracts or terms 

75 Cockrell (n 53 above) 46. See, for example, Wells v South African Alumenite 

Company 1927 AD 69 73 (Wells); Roffey v Catterall, Edwards & Goudré (Pty) Ltd 

1977 4 SA 494 (N) (Roffey) 504G-H; Sasfin (Pty) Ltd v Beukes 1989 1 SA 1 (A) 

(Sasfin) 9F; Edouard v Administrator, Natal 1989 2 SA 368 (D) (Edouard) 379A; 

Baart v Malan 1990 2 SA 862 (E) (Baart) 868A; De Klerk v Old Mutual Insurance Co 

Ltd 1990 3 SA 34 (E) (De Klerk) 44; Benlou Properties (Pty) Ltd v Vector Graphics 

(Pty) Ltd 1993 1 SA 179 (A) (Benlou Properties) 187H; Standard Bank of SA Ltd v 

Wilkinson 1993 3 SA 822 (C) (Standard Bank) 830D; and Basson v Chilwan and 

Others 1993 3 SA 742 (A) (Basson) 761G. See also V Terblanche (2002) ‘The 

Constitution and general equitable jurisdiction in South African contract law’ 

unpublished LLD thesis, University of Pretoria, 2002 153 who refers to the 

‘privileged position of 19th century contract law theory in South Africa.’ 

76 1875 LR 19 (Printing and Numerical Registering Co) Eq 462 per Jessel MR. 

77 

Printing and Numerical Registering Co (n 76 above) Eq 465. The mood is also 

expressed in more general terms by Kotze CJ in Brown v Leyds (1897) 4 OR 17 

(Brown) 31 who held that ‘no Court of Justice is competent to inquire into the 

internal value, in the sense of the policy, of the law, but only in the sense of the 

meaning or matter of the law’. 

78 1903 TS 571 (Burger). 

79 

Burger case (n 78 above) 576. 

80 Van der Merwe et al (n 37 above) 192-193, 200-203.


contrary to public policy was expressed as follows in Morrison v 

Angelo Deep Gold Mines Ltd: 81 

[I]t is a general principle that a man contracting without duress, without 

fraud and understanding what he does, may freely waive any of his 

rights. There are certain exceptions to that rule, and certainly the law 

will not recognise any arrangement which is contrary to public policy. 

The dialectic that emerges here is that, on the one hand, public policy 

generally favours the utmost freedom of contract but, on the other, 

contracts concluded in violation of public policy are unenforceable or 

void. Public policy is thus regarded (from the perspective of freedom 

of contract) as both a legitimating and controlling device in our law 

of contract. 82 The resolution/synthesis of this dialectic is, however, 

not impossible when one understands that the exercise of alleged 

freedom of contract is precisely not freedom of contract — that is, it 

is not freedom that can legitimate a contractual arrangement — when 

it is exercised contrary to public policy. This understanding also points 

to a deeper problem with the notion of autonomy as freedom of 

contract, which is said to form the cornerstone or foundational 

principle of the law of contract. The public policy requirement 

communicates that some contracts that are freely entered into (that 

is, in which the wills of legally competent parties overlap) will not be 

enforced for want of legality (in the broad sense). This means that, as 

a matter of a contract’s validity, there is a constitutive limit on 

freedom of contract that inevitably denies absolute party autonomy. 

And since legality is determined externally (not by the parties but 

with reference to ‘the interests and convictions’ of the society in 

which the contract is concluded) 83 it means that the cornerstone of 

the law of contract is both autonomous and heteronymous. 

Our courts, however, do not approach the matter in this way. 

Freedom of contract is consistently opposed to other ideals, values 

and rights, instead of integrally connected to and thus determined by 

those other values and rights. When public policy is, for instance, 

considered in relation to restraint of trade clauses the position is 

taken that ‘two values or freedoms come into conflict ... namely 

freedom of contract and freedom of trade.’ 84 In our law the position 

(since the decision in Magna Alloys) 85 has been that restraints of trade 

are in principle enforceable in accordance with freedom of contract, 

81 

1905 TS 775 (Morrison) 779. 

82 See Van der Merwe et al (n 37 above) 17. 


84 

Van der Merwe et al (n 37 above) 213. Also see Roffey (n 75 above) 505F (per 

Didcott J): ‘I am satisfied that South African law prefers the sanctity of contracts 

... Freedom of trade does not vibrate nearly as strongly through our jurisprudence 

... it is intrinsically the less commanding of the two ideas.’ 

85 Magna Alloys & Research (SA) (Pty) Ltd v Ellis 1984 4 SA 874 (A) (Magna Alloys).


unless the restraint is contrary to public policy, the measure of public 

policy in relation to restraints of trade being reasonableness. 86 On 

this formulation, freedom of contract is said to be the ‘preferred 

value’. 87 

Recently the constitutionality of the Magna Alloys formulation 

was questioned in the case of Reddy v Siemens Telecommunications 

(Pty) Ltd 88 where it was argued that restraints of trade constitute 

limitations on the right to freely choose a trade, occupation or 

profession — a right expressly protected in section 22 of the 

Constitution. 89 This means, so the argument went, that a restraint of 

trade is enforceable only when it can be proved by the employer who 

wants to enforce it to be a reasonable and justifiable limitation (as 

required by section 36 of the Constitution) on the section 22 right. 90 

The Supreme Court of Appeal (per Malan AJA) held that this argument 

effectively constituted an argument regarding onus and did not 

contest the earlier findings of the courts 

that a restraint that is found to be reasonably required for the 

protection of the party who seeks to enforce it, in accordance with the 

test that has been laid down in the cases, is constitutionally 

permitted. 91 

The Court declined to adjudicate the issue of onus, holding that, 

‘whether the onus was reversed or not, the result in the present case 

would be the same.’ 92 It continued to hold that what is called for in 

86 In the context of the rules/standards discourse it should be pointed out that our 

courts do not seem to experience difficulty with the standard of ‘reasonableness’ 

both in the context of the public policy requirement and outside of it. In addition 

to its role in the area of restraints of trade provisions and public policy, 

reasonableness also plays a decisive role in the determination of contractual 

liability on the basis of reliance or the doctrine of quasi-mutual assent. In this 

regard see, for instance, Sonap Petroleum (SA) (Pty) Ltd v Pappadogianis 1992 3 

SA 234 (A) (Sonap Petroleum) and Constantia Insurance v Compusource (Pty) Ltd 

2005 4 SA 345 (SCA) (Constantia Insurance). 


88 

2007 2 SA 486 (SCA) (Reddy). 

89 In Canon KwaZulu-Natal (Pty) Ltd t/a Canon Office Automation v Booth 2005 3 SA 

205 (N) (Canon) 208-209 the court endorsed this reasoning. 

90 

Reddy (n 88 above) 495C-E. 

91 Reddy (n 88 above) 495D. See Waltons Stationery Co (Edms) Bpk v Fourie en 'n 

Ander 1994 4 SA 507 (O) (Waltons Stationary) 510I-511F (SA); Kotze & Genis 

(Edms) Bpk en 'n Ander v Potgieter en Andere 1995 3 SA 783 (C) (Kotze & Genis) 

786E-I; Knox D'Arcy Ltd and Another v Shaw and Another 1996 2 SA 651 (W) (Knox) 

657Hff and 661D-F, CTP Ltd and Others v Independent Newspaper Holdings Ltd 

1999 1 SA 452 (W) (CTP) 468G-H; Fidelity Guards Holdings (Pty) Ltd t/a Fidelity 

Guards v Pearmain 2001 2 SA 853 (SE) (Fidelity Guards) 861F-862G; and Oasis 

Group Holdings (Pty) Ltd and Another v Bray [2006] 4 All SA 183 (C) (Oasis Group) 

paras 23-25. 

92 Reddy (n 88 above) 496A-B.

the case of restraints of trade is a value judgment 


with two principal policy considerations in mind in determining the 

reasonableness of a restraint. The first is that the public interest 

requires that parties should comply with their contractual obligations, a 

notion expressed by the maxim pacta servanda sunt. The second is that 

all persons should in the interests of society be productive and be 

permitted to engage in trade and commerce or the professions. 93 

It concluded that the common law approach to the determination of 

the reasonableness or not of the restraint ‘gives effect to the precepts 

of section 36(1) of the Constitution’. 94 Thus the enquiry into the 

reasonableness of the restraint simultaneously answers the question 

whether the restraint is a reasonable and justifiable limitation of the 

right in question. 

Clearly, the Court here views the right to choose a trade, 

occupation or profession freely as coming into conflict with freedom 

of contract as pacta servanda sunt. However, an alternative approach 

was followed in the later judgment of Davis J in Advtech Resourcing 

t/a Communicate Personnel Group v Kuhn. 95 In this judgment Davis J 

held that if contractual autonomy is part of the constitutional value 

of freedom and informs the ideal of dignity, 96 then under a 

transformative Constitution such as ours, contractual autonomy must 

be read as internally limited by the rights in the Bill of Rights and ‘the 

spirit of the Constitution read as a whole’. 97 In the context of 

restraints of trade this means that section 22 of the Constitution 

places an internal limit on freedom of contract to the extent that 

agreements in restraint of trade that constitute unreasonable and 

unjustifiable limitations on the section 22 right, will not be enforced. 

And since a restraint of trade constitutes a prima facie violation of the 

section 22 right, an employer who wants to enforce the restraint must 

prove that it is a reasonable and justifiable limitation of the right. 98 

When the Constitution states that the Bill of Rights does not deny 

the existence of any other rights or freedoms that are recognised or 

conferred by common law ‘to the extent that they are consistent with 

the Bill’ of Rights 99 it places the common law notion of freedom of 

contract on an equal footing with all the rights in the Bill of Rights. In 

so doing, it contests the a priori elevation of a liberal notion of 

freedom of contract in the common law in the sense that it calls for 

93 Reddy (n 88 above) 496D. 

94 Reddy (n 88 above) 497 F/G. 

95 

2008 2 SA 375 (C) (Advtech). 

96 Advtech (n 95 above) 387I. 

97 Advtech (n 95 above) 388A. 

98 

Advtech (n 95 above) 387A. 

99 Section 39(3) of the Constitution.


a synchronisation of freedom of contract with the ideals, values and 

rights in the Bill of Rights. Drucilla Cornell distinguishes the concept 

of synchronisation from both Dworkinian coherence and liberal 

conceptions of balancing: ‘Synchronisation recognises that there are 

competing rights situations and real conflicts, which may not be able 

to yield a “coherent” whole.’ 100 For Cornell synchronisation refuses 

closure and so calls for an ongoing, never-ended (re-)thinking of the 

whole even as it marks the failure of constructive coherence. For 

Cornell ‘the tension between the promise of synchronisation and the 

failure of its achievement’ opens the law to its transformative 

potential. 101 

In this context and on this approach, freedom of contract as pacta 

servanda sunt, becomes internally defined by the right in section 22 

and, for that matter, any other right in the Bill of Rights as well as the 

founding values and ideals of the Constitution. But such an approach 

is only possible once one accepts that section 22 (and indeed all the 

rights in the Bill of Rights) is on an equal footing with freedom of 

contract (as the Constitution clearly says it is). Simply put, there 

cannot be talk of freedom of contract where the restraint of trade (or 

whatever contractual provision) is found to unreasonably and 

unjustifiably limit a right in the Bill of Rights or can be said to be 

contrary to the spirit, purport and objects of the Constitution. It goes 

without saying that where the limitation of a constitutional right is 

alleged in the context of restraints of trade, such an approach would 

— in this context — require the prioritisation of direct horizontal 

application of the Bill of Rights, even though indirect horizontal 

application might lead to the same result. 102 

The oppositional/dualistic reading of freedom of contract — by 

way of which its hegemonic understanding is maintained — is, in 

addition, also clearly illustrated through the unwillingness to read a 

fairness jurisdiction into the public policy requirement. 103 As Glover 

argues, the defence that the enforcement of a contract is in specific 

circumstances unfair or unconscionable, has only been approved of in 

100 D Cornell Transformations (1993) 36. 

101 n 100 above, 35. 

102 See in this regard S Woolman ‘The amazing, vanishing bill of rights’ (2007) 3 South 

African Law Journal 762 777. To reach directly for the baton of indirect horizontal 

application by way of sec 39(2) when a constitutional right is alleged to have been 

infringed is to render sec 8 of the Constitution redundant. As Woolman indicates, 

such an approach is inconsistent with the Constitutional Court’s decision in 

Khumalo v Holomisa 2002 5 SA 401 (CC) (Khumalo) para 32 where O’Regan J made 

it clear ‘that we should not attribute a meaning to one section of the Constitution 

that renders another section, quite literally, senseless.’ 

103 Van der Merwe et al (n 37 above) 199. P Aronstam Consumer protection, freedom 

of contract and the law (1979) 43.


the restraint of trade context where, as we have seen, the restraint 

of trade is subjected to a reasonableness enquiry. 104 

However, in Sasfin v Beukes 105 the Appellate Division considered 

the extent to which ‘simple justice between man and man’ 106 can 

trump the public interest in the strict enforcement of contracts 

generally. Primarily, the case centred around the validity of a deed 

of cession in which a customer of a bank (a doctor) ceded all his future 

debtors to the bank regardless of whether he owed the bank money 

or not. 107 The cession effectively rendered the doctor the slave of the 

bank. The majority of the court was of the opinion that these aspects 

of the cession could not be severed from the rest of the agreement 

and held that the entire transaction was unenforceable. 108 

In this case the Appellate Division was willing to evaluate the 

substantive fairness of the disputed deed of cession to come to a 

conclusion that the cession was clearly unreasonable and 

irreconcilable with the public interest. Lewis describes the decision 

as ‘the one decision which yields a ray of light in the field of 

contractual policy, where the court was both bold and innovative in 

escaping the shackles of formalism’. 109 Be that as it may, one again 

sees the opposition between freedom of contract and other values in 

Smalberger JA’s judgment where it is held that unlawfulness comes 

into play where the public interest in the strict enforcement of 

contracts in accordance with the principle of freedom of contract, is 

trumped by other relevant factors. 110 These relevant factors are 

expressed by the Court to ensure that ‘public policy ... properly take 

into account the doing of simple justice between man and man.’ 111 

The uneasiness with which the Appellate Division approaches the 

issue of unfair contracts is illustrated in the following famous dictum: 

No court should therefore shrink from the duty of declaring a contract 

contrary to public policy when the occasion so demands. The power to 

declare contracts contrary to public policy should however be exercised 

sparingly and only in the clearest of cases, lest uncertainty as to the 

validity of contracts result from an arbitrary and indiscriminate use of 

the power. One must be careful not to conclude that a contract is 

104 G Glover ‘Lazarus in the Constitutional Court: An exhumation of the exceptio doli 

generalis?’ (2007) 124 South African Law Journal 449 451. 

105 1989 1 SA 1 (A) (Sasfin). 

106 As above. 

107 Sasfin (n 105 above) 6A-D. 

108 Sasfin (n 105 above) 17-19. 

109 C Lewis ‘Towards an equitable theory of contract: The contribution of Mr Justice 

EL Jansen to the South African law of contract’ (1991) 108 South African Law 

Journal 249 264 n76. 

110 Sasfin (n 105 above) 9D-G. 

111 Sasfin (n 105 above) 13J.


contrary to public policy merely because its terms (or some of them) 

offend one's individual sense of propriety and fairness. 112 

Hawthorne has remarked that the dicta above merely emphasises the 

South African judiciary’s narrow interpretation of the relevance of 

equity considerations in the public interest. 113 By holding that it is 

only in the ‘clearest of cases’ that a court may use its power to refuse 

to enforce an unfair term, and that the power to do this must be used 

‘sparingly’, the court suggests that unconscionability in and of itself 

cannot invalidate a contract. 114 The promise that we can correct for 

clear or extreme cases only, simply suggests, to paraphrase Dalton, 

that the worst features of the system can be held in check, without 

tinkering with its regular operation. 115 Van der Merwe et al confirm 

the matter: 

Sasfin’s decision must not be interpreted to mean that the unreasonably 

harsh or unconscionable effect of a term between the parties is by itself 

a sufficient ground for making an agreement illegal. 116 

In Donelly 117 Kriegler J stressed this point and placed a great deal of 

emphasis on the fact that the Sasfin judgment should not be regarded 

‘as a free pardon to recalcitrant and otherwise defenceless 

debtors’ 118 because it is ‘decidedly not that’. 119 Kriegler J also did 

not let the opportunity pass to emphasise that ‘pacta servanda sunt 

is still a cornerstone of our law of contract’ 120 and that ‘nothing said 

or implied’ 121 in the Sasfin principle can be said to derogate from this 

important fact. 122 In Botha (now Griessel) v Finanscredit (Pty) Ltd 123 

the Appellate Division confirmed that 

a Court’s power to declare contracts contrary to public policy should be 

exercised sparingly and only in cases in which the impropriety of the 

transaction and the element of public harm are manifest. 124 

112 Sasfin (n 105 above) 9A-B. (emphasis added) 

113 L Hawthorne ‘Public policy and micro lending — has the unruly horse died?’ (2003) 

66 THRHR 116 118. 

114 These sentiments are also reflected in cases decided by the lower courts in the 

aftermath of Sasfin. See in this regard Standard Bank (n 75 above) 825C-827A and 

Pangbourne Properties Ltd v Nitor Construction (Pty) Ltd 1993 4 SA 206 (W) 

(Pangbourne Properties) 210G-214F. 

115 Dalton (n 60 above) 1037. 


117 Donelly v Barclays National Bank Ltd 1990 1 SA 375 (W) (Donelly). 

118 n 117 above, 381F. 

119 As above. 

120 Donelly (n 117 above) 381H. 

121 As above. 

122 As above. 

123 1989 3 SA 773 (A) (Botha). 

124 Botha (n 123 above) 782I-783C.


In this case the court refused to declare a deed of suretyship void 

on the grounds of public policy, holding that, although ‘somewhat 

rigorous’, the surety was not left ‘helpless in the clutches of the 

plaintiff’. 125 This prompts one to ask whether one party should 

necessarily be left helpless in the clutches of the other for the court 

to exercise its Sasfin power? If so, very little of the work of reform and 

transformation of the law of contract can be expected to be done 

through the utilisation of the Sasfin principle. 

But the point that I would like to stress going forward is that the 

courts’ concern with maintaining the hegemony of the will theory of 

contract causes them to assume that there is an area in which 

freedom of contract can be exercised independent of public policy, 

fairness, reasonableness and good faith. From a political point of 

view, the paradox which inscribes itself in this opposition between 

freedom of contract and the so-called constraints of public policy, 

fairness, reasonableness and good faith exists in that the claim that 

the reliance on freedom of contract and its concomitant rules is 

somehow a-political and thus unproblematic, generates increasing 

and proliferating politics in the law of contract. The more the courts 

attempt to depoliticise the law of contract by way of limiting the 

open-ended norms in favour of liberal legalism grounded in a rule 

jurisprudence, the more it politicises it by way of these oppositions 

and dualisms. 

Below I will argue that the unwillingness to deal with unfair 

contracts in terms of the public policy requirement is symptomatic of 

a deeper illness, namely the courts’ refusal to accept that the ethical 

element of contract (good faith or the bona fides) — as a foundational 

value of the law of contract — forms part and parcel of not only the 

public policy requirement 126 but also of the very maxim pacta 

servanda sunt. 127 In other words, the opposition between the 

foundational norms of freedom of contract and good faith serves — as 

we shall see — to maintain a certain instrumental politics in the law 

of contract weighted in favour of commercial interests. The 

opposition itself thus furthers the privileging or hegemony of a liberal 

version of freedom of contract, which closes off the law of contract 

from the question of the ethical embodied in the good faith 

requirement. 

125 Botha (n 123 above) 783J. 

126 Van der Merwe et al (n 37 above) 200: ‘in the context of legality there is a close 

relationship between what is fair and reasonable and what is in accordance with 

good faith’. 

127 See n 37 above.


4.2 Good faith and the exceptio doli generalis in the South 

African law of contract 

The South African law of contract has long considered all contracts as 

acts of good faith. 128 This is the case because we inherited substantial 

parts of the Roman-Dutch law of contract in which contracts were 

considered iudicia bonae fidei. 129 In Roman law, however, a 

distinction existed between negotia stricti iuris and the negotia 

bonae fidei. 130 Contracts from the ius stricti bound the debtor to 

perform strictly in accordance with what he promised in the formula 

and not in accordance with what the bona fides could expect of him, 

unless the formula itself referred to the bona fides. 131 In the case of 

the negotia bonae fidei the bona fides were conclusive and the 

absence thereof, whether during negotiations, conclusion or 

institution of the action, gave rise to a defence. 132 The bona fides 

thus operated as an evaluative yardstick to determine the 

enforceability of the negotia bonae fidei. 

To curb possible injustices or unconscionable conduct as a result 

of the enforcement of the negotia stricti iuris, the praetor introduced 

the exceptio doli generalis. 133 Here the defendant was allowed to 

submit facts that he would otherwise not have been able to submit 

because of the operation of the strict ius civile. 134 The exceptio 

therefore functioned to curb the abuse of rights in appropriate 

circumstances and became the instrument with which more equitable 

128 Meskin v Anglo-American Corporation of SA Ltd (1968) 4 SA 793 (W) (Meskin) 802; 

Paddock Motors (Pty) Ltd v Ingesund 1976 3 SA 16 (AD) (Paddock Motors) 28; 

Magna Alloys (n 85 above) 893C; Tuckers Land & Development Corporation (Pty) 

Ltd v Hovis 1980 1 SA 645 (A) (Tuckers Land) 652; Mutual & Federal Insurance Co 

Ltd v Oudtshoorn Municipality 1985 1 SA 419 (A) (Mutual & Federal) 433; Savage 

& Lovemore Mining (Pty) Ltd v International Shipping Co (Pty) Ltd 1987 2 SA 149 

(W) (Savage & Lovemore Mining)198A, Sasfin (n 75 above) 8C-D; Botha (n 123 

above) 782J; LTA Construction Bpk v Administrateur Transvaal 1992 1 SA 473 (A) 

(LTA Construction) 480D; Eerste Nasionale Bank van Suidelike Afrika Bpk v 

Saayman NO 1997 4 SA 302 (SCA) (Eerste Nasionale Bank) 321; Janse van Rensburg 

v Grieve Trust 2000 1 SA 315 (C) (Janse van Rensburg) 325-326; Afrox Healthcare 

(n 13 above) 40I-J; Van der Westhuizen v Arnold 2002 6 SA 453 (SCA) (Van der 

Westhuizen) 469E-F. 

129 Van der Merwe et al (n 37 above) 319. Aronstam (n 103 above) 173: ‘the sole 

question facing a court in Roman-Dutch law was whether the conduct of a party 

to a contract, either in any negotiations that preceded the contract, or in any 

action that was based upon it, had complied with the requirements of good 

faith.’ 

130 See CFC van der Walt (n 71 above) 648 and NJ Grové ‘Kontraktuele gebondenheid, 

die vereistes van die goeie trou, redelikheid en billikheid’ (1998) 61 Tydskrif vir 

Hedendaagse Romeins-Hollandse Reg 687 688 who show that the distinction was 

closely related to the Roman procedural law and the specific defences it allowed 

for in different cases. 


132 As above. 

133 As above. 

134 As above.


principles were introduced in the law of contract by the praetorian 

law. 135 

The question regarding the reception of the exceptio doli 

generalis into the South African law of contract has been contentious. 

The question is complicated by the fact that all contracts became 

negotia bonae fidei in Roman-Dutch Law, thus appearing to do away 

with the need for such a defence. However, Aronstam argues that the 

exceptio doli generalis became part of the Roman-Dutch law in that 

it was implied by good faith: 136 

if all contracts were bonae fidei they had to be performed in accordance 

with the dictates of good faith, and any absence of bona fides could be 

taken into account by a judge in any action based on such contracts. 137 

Aronstam’s argument thus is that the exceptio doli generalis formed 

part of the requirement of good faith and as such was received into 

the South African law of contract. 

Six months before the decision in Sasfin (discussed above), the 

Appellate Division had occasion to consider the question regarding the 

reception of the exceptio doli generalis into the South African law of 

contract in the case of Bank of Lisbon and South Africa v De Ornelas 

& Others. 138 In this case the court primarily concerned itself with the 

question whether the exceptio survived the reception of Roman Law 

into Roman-Dutch Law and of Roman-Dutch Law into the South African 

law. The court held that the exceptio doli generalis never formed part 

of the Roman-Dutch law, 139 could therefore not have been received 

into the South African law 140 and that its occasional appearances on 

the scene of the South African law 141 should finally be prohibited by 

burying it ‘as a superfluous, defunct anachronism.’ 142 

This decision is also generally regarded as the decision in which 

the Appellate Division did away with the perception that a contract 

could be adjudicated and declared unenforceable by a court on the 

135 G Lubbe ‘Bona fides, billikheid en die openbare belang in die Suid-Afrikaanse 

kontraktereg’ (1990) 1 Stellenbosch Law Review 9. Also see C Lewis ‘The demise 

of the exceptio doli: Another route to contractual equity?’ (1990) 107 South 

African Law Journal 26 31-32. 

136 n 103 above, 173. 

137 n 103 above, 172. 

138 1988 3 SA 580 (A) (Bank of Lisbon). 

139 Bank of Lisbon (n 138 above) 605H. 

140 Bank of Lisbon (n 138 above) 607A. 

141 For a discussion of the various interpretations of the exceptio doli generalis see 

SWJ Van der Merwe et al ‘The exceptio doli generalis: Requiescat in pace — vivat 

aequitas’ (1989) 106 South African Law Journal 235. Also see Weinerlein v Goch 

Buildings Ltd 1925 AD 282 (Weinerlein); Zuurbekom Ltd v Union Corporation Ltd 

1947 1 SA 514 (AD) (Zuurbekom); Paddock Motors (n 128 above); and Arprint Ltd v 

Gerber Goldschmidt Group South Africa (Pty) Ltd 1983 1 SA 254 (A) (Arprint). 

142 Bank of Lisbon (n 138 above) 607B. Also see Lubbe (n 135 above) 9.


basis of good faith. 143 The conflation of these two issues (the 

abolition of the exceptio doli generalis on the one hand and the 

abolition of an explicit good faith jurisdiction, on the other) finds its 

origin in the following dictum of Joubert JA: 

Nor can I find any evidence of the existence of a general substantive 

defence based on equity. This is not surprising inasmuch as the Dutch 

Courts, unlike the English Courts until the Judicature Act 1873 became 

operative in 1875, did not administer a system of equity as distinct from 

a system of law. Roman-Dutch law is itself an inherently equitable legal 

system. In administering the law the Dutch Courts paid due regard to 

considerations of equity but only where equity was not inconsistent with 

the principles of law. Equity could not override a clear rule of law. That 

is also the position of our Courts as regards their equitable jurisdiction 

... Moreover, I cannot find any support in Roman-Dutch law for the 

proposition that in the law of contract an equitable exception or 

defence, similar in effect to the exceptio doli generalis, was utilised 

under the aegis of bona fides. 144 

In his discussion of the majority decision, Cockrell shows that the 

decision that the exceptio doli constitutes ‘a superfluous, defunct 

anachronism’, 145 is founded upon an extreme form of individualism 

which denies that the law may ‘legitimately superimpose an 

overriding duty to act in good faith’ 146 upon the voluntary agreements 

of legal subjects with full capacity. 147 As Cockrell points out, the 

Court’s problem with accepting the exceptio was that it enjoins a 

judge to employ a standard that cannot be cast in ‘a clear rule of 

law’. 148 As the Court held, quoting Van der Linden’s account of Voet 

1.1.16: ‘... [j]udges and jurists ought to look to nothing more 

carefully than this, that they do not forsake the written law for some 

headstrong equity ...’. 149 Joubert JA’s rejection of the exceptio doli 

143 See Eerste Nasionale Bank (n 128 above) 323B. 

144 Bank of Lisbon (n 138 above) 605I-606D. 

145 Bank of Lisbon (n 138 above) 607B. 

146 n 53 above, 44. 

147 As above. 

148 Also see the judgment in Tjollo Ateljees (Eins) Bpk v Small 1949 1 SA 856 (A) 

(Tjollo Ateljees). In this decision the Appellate Division recommended the 

abolition by the legislature of the doctrine of laesio enormis. According to laesio 

enormis a party who had either agreed to pay a price of more than double the 

value of the thing or agreed to sell for less than half of the value of the thing 

could either cancel the agreement or claim price adjustment (see Van der Merwe 

et al (n 37 above) 132). Cockrell (n 53 above) 45 illustrates clearly that the basis 

of the decision was, on the one hand, that laesio enormis constituted an altruist 

intrusion on the terrain of individualism and, on the other, that laesio enormis 

could not be cast in the rule form. 

149 Bank of Lisbon (n 138 above) 610D-E.


is thus clearly linked to his disapproval of the discretion which will be 

afforded to judges by its acceptance. 150 

The attempt to rid our law of the exceptio doli generalis should 

then, given these observations, be understood as part and parcel of 

the attempt to keep the hegemony of freedom of contract as the will 

theory and the maxim pacta servanda sunt, firmly in place. With this 

decision, the Appellate Division attempted finally to seal the 

hegemony of the will theory as freedom of contract. This is 

particularly evident when Bank of Lisbon is considered in the light of 

the Sasfin decision six months later in which possible interference 

with liberal freedom of contract by way of public policy was further 

curtailed. 151 The abolition of the exceptio doli generalis confirmed 

that the enquiry in the South African law of contract when it comes 

to enforceability is conducted with reference to so-called clear rules 

of law that are grounded in individualism. Consequently as we shall 

see, the Bank of Lisbon decision forms the focal point when it comes 

to the marginalisation of good faith as the ethical element of contract 

in South Africa. 

This marginalisation, however, was not achieved without some 

strife. In the first place — regarding the exceptio doli generalis as a 

technical remedy — the Appellate Division contradicted itself when it 

held in the later case of Van der Merwe v Meades 152 that the 

replicatio doli still formed part of our law. As Kerr argued at the time, 

the same court had previously held in Bank of Lisbon that the 

conclusions about the non-reception of the exceptio doli generalis 

‘equally hold for the replicatio doli generalis’. 153 Kerr submits that 

the linking was correct. 154 If, as Kerr points out, the replicatio doli, 

on the later decision of the court, survived the reception, and the 

same court linked the replicatio with the exceptio and has not 

departed from that position, then the only conclusion that can follow 

is that the exceptio doli generalis must also have survived the 

reception. 155 

150 The majority decision clearly does not take account of, or acknowledge, the 

subjectivity/discretion which is at play in considering whether ‘a clear rule of 

law’ is in fact that and, even more problematic, the subjectivity that is at play in 

any decision as regards the question whether the ‘clear rule of law’ is applicable 

or not. 

151 The mantra that public policy should properly take into account the doing of 

simple justice between man and man and the way in which that mantra was 

employed in Sasfin does not, without more, lead to a conclusion that Sasfin 

somehow provided an escape route from strict enforcement of contract. This is 

evident from the cases that followed Sasfin, as discussed earlier. 

152 1991 2 SA 1 (A) (Van der Merwe) 2F-3A. 

153 Bank of Lisbon (n 138 above) 608F-G. 

154 AJ Kerr ‘The replicatio doli reaffirmed. The exceptio doli available in our law’ 

(1991) 106 South African Law Journal 583 584. 

155 n 154 above, 585.


Regarding the question whether the Bank of Lisbon judgment — in 

doing away with the exceptio doli — in fact did away with the power 

to declare a contract or a provision thereof as void or unenforceable 

on the basis of good faith, Olivier JA argued strongly in his famous 

(perhaps notorious) minority judgment in Eerste Nasionale Bank van 

Suidelike Afrika Bpk v Saayman NO 156 that Bank of Lisbon did not and 

could not have done away with this power. Olivier JA argued that the 

reception of all contracts as iudicia bonae fidei into the South African 

law of contract made the exceptio doli generalis (as a technical 

remedy that had to be pleaded), superfluous in that it was implied in 

the idea of iudicia bonae fidei that a court would have discretion to 

examine and pronounce on the bona fides. Olivier JA continued that 

this would be the position from the perspective of those who accept 

the broad, dynamic working of the bona fides and the discretion of 

judges to apply the bona fides. 157 Olivier JA argued bravely that the 

bona fides principle forms part of the public policy requirement 158 

and could be directly applied to the facts before him. This was a 

certain attempt to radicalise the public policy principle as enunciated 

in Sasfin. To quote Olivier JA: 

[e]k hou dit as my oortuiging na dat die beginsels van die goeie trou, 

gegrond op openbare beleid, steeds in ons kontraktereg 'n belangrike rol 

speel en moet speel, soos in enige regstelsel wat gevoelig is vir die 

opvattinge van die gemeenskap, wat die uiteindelike skepper en 

gebruiker van die reg is, met betrekking tot die morele en sedelike 

waardes van regverdigheid, billikheid en behoorlikheid. 159 

(I believe that the principles of good faith, based on public policy, still 

play and should continue to play an important role in our law of contract 

as it does in any legal system which is sensitive to the convictions of the 

community — the ultimate creator and user of the law — with respect to 

the moral and ethical values of justice, fairness and propriety.) 

In his discussion of unfair contracts in the fourth edition of The law of 

contract in South Africa, Richard Christie remarked that 

[t]here is every reason to hope that when the opportunity arises the 

Supreme Court of Appeal will apply Olivier JA’s reasoning, harnessed to 

the concept of public policy, in the context of the unfair enforcement of 

a contract. 160 

In NBS Boland Bank Ltd v One Berg River Drive CC and Others; Deeb 

and Another v ABSA Bank Ltd; Friedman v Standard Bank of SA Ltd 161 

156 n 128 above. 

157 Eerste Nasionale Bank (n 128 above) 323. 

158 Eerste Nasionale Bank (n 128 above) 323-324 

159 Eerste Nasionale Bank (n 128 above) 326 F-H. 

160 RH Christie The law of contract in South Africa (4ed, 2001) 19. 

161 1999 4 SA 928 (SCA) (Friedman).


the Supreme Court of Appeal (SCA) held obiter that in the matter 

before it 

[a]n analogous conclusion may well be reached if one applies the 

modern concept of the role of public policy, bona fides and contractual 

equity to the question in issue (see for example, Eerste Nasionale Bank 

van Suidelike Afrika Bpk v Saayman NO [reference omitted] per Olivier 

JA. 162 

This appeared at the time to have constituted at least an implicit 

endorsement of what Olivier JA said in that case. 

In Miller 163 Ntsebeza AJ interpreted the judgment of Olivier JA in 

Saayman to mean that a court can indeed refuse to uphold a so-called 

non-variation or Shifren clause 164 on the basis of considerations of 

good faith. 165 In turn, Van Zyl J argued — relying explicitly on Olivier 

JA’s judgment in Saayman — in Janse van Rensburg v Grieve Trust 166 

that ‘the principles of justice, equity, reasonableness and good faith’ 

required the development of the common law in the case before 

him. 167 Davis J, in obiter remarks in Mort NO v Henry Shields-Chiat 168 

also invoked Olivier JA’s judgment and added that 

it is clear that if the doctrine [of bona fides] is to be taken seriously then 

the primary importance accorded to the private autonomy of contracting 

parties must be reconsidered as must the hegemony of the will theory of 

contract which survives even in the context of dicta which nod in the 

direction of social responsibility. 169 

Davis J implicitly accepted Olivier JA’s reasoning in Eerste Nasionale 

Bank that good faith forms part of the public policy requirement when 

he stated that Olivier JA’s approach builds on the judgment of 

Smalberger JA in Sasfin and particularly on the dictum from that case 

as quoted earlier. 170 Similar to Olivier JA, Davis J attempted to 

radicalise Sasfin by the attempt to infuse it with the bona fides. 

But what sets Davis J’s judgment in Mort apart is the express link 

made in the judgment between, on the one hand, the argument in 

favour of the bona fides and, on the other, the ethical implications 

162 Friedman (n 161 above) 937F. Also see Pretorius (n 54 above) 643. 

163 Miller and Another NNO v Dannecker 2001 1 SA 928 (C) (Miller). 

164 This name for a non-variation clause (which generally stipulates that no 

amendments to a written agreement shall be valid unless reduced to writing) is 

derived from the decision in SA Sentrale Ko-op Graanmaatskappy Beperk v 

Shifren en Andere 1964 4 SA 760 (A) (Shifren). 

165 Miller (n 163 above) 938 para 19 

166 n 128 above. 

167 n 128 above, 325D-F. 

168 2001 1 SA 464 (C) (Mort). 

169 n 168 above, 474I (Emphasis added). 

170 See note 112 above.


for adjudication in the law of contract of the reconstitution of the 

legal order occasioned by the enactment of the Constitution. To 

quote Davis J: 

Like the concept of boni mores in our law of delict, the concept of good 

faith is shaped by the legal convictions of the community. While Roman- 

Dutch law may well supply the conceptual apparatus for our law, the 

content with which concepts are filled depends on an examination of the 

legal conviction of the community — a far more difficult task. This task 

requires that careful account be taken of the existence of our 

constitutional community, based as it is upon principles of freedom, 

equality and dignity. The principle of freedom does, to an extent, 

support the view that the contractual autonomy of the parties should be 

respected and that failure to recognise such autonomy could cause 

contractual litigation to mushroom and the expectations of contractual 

parties to be frustrated. [Reference omitted.] But the principles of 

equality and dignity direct attention in another direction. Parties to a 

contract must adhere to a minimum threshold of mutual respect in 

which ‘the unreasonable and one-sided promotion of one's own interest 

at the expense of the other infringes the principle of good faith to such a 

degree as to outweigh the public interest in the sanctity of contracts.’ 

[Reference omitted.] The task is not to disguise equity or principle but 

to develop contractual principles in the image of the Constitution. 171 

However, in Brisley v Drotsky 172 the majority of the SCA rejected — 

in no uncertain terms — both Olivier JA’s Eerste Nasional Bank 

judgment as well as the judgments that built on it. In the first 

instance, the majority criticised Olivier JA’s judgment for attempting 

indirectly to resurrect the exceptio doli generalis. 173 With this 

statement the majority seems to have uncritically accepted that the 

exceptio doli generalis is not a part of our law. But as we have seen, 

on the Court’s own jurisprudence it is not at all clear whether the 

exceptio doli generalis has in fact been abolished. The terse 

statement in Brisley regarding the attempt to resurrect the exceptio 

doli generalis dismisses this controversy/contradiction by pretending 

that the exceptio is in fact dead and buried. This statement in Brisley 

can of course be seen as implicitly overriding the court’s Van der 

Merwe decision (in which it, ironically, resurrected the exceptio doli 

generalis), but the Court explicitly stated in a footnote that the 

question regarding the reconsideration of the exceptio did not arise 

in Brisley. 174 This left as uncertain as ever the position in our law 

regarding the status of the exceptio. 

171 Mort (n 168 above) 474I-475D 

172 n 13 above. The dispute in this case also concerned the question regarding the 

enforceability of contractual terms (including a Shifren clause) in circumstances 

where oral amendments to the agreement were made but not reduced to writing. 

173 Brisley (n 13 above) 14 para 17. 

174 Brisley (n 13 above) n 10.


As regards Davis J’s view in Mort that the bona fides (like the boni 

mores in the law of delict) are shaped by the legal convictions of the 

constitutional community (as an ideal community), the majority held 

that if this means that the enforcement of contractual terms depends 

on the convictions of the community (gemeenskapsgevoel), then it 

cannot agree because there are material policy differences between 

the approach to the law of delict and the approach to the law of 

contract. 175 According to the majority the unqualified acceptance of 

Davis J’s judgment will create a state of unacceptable chaos and 

disorder in the law of contract. 176 A court cannot seek shelter in the 

shadow of the Constitution, writes the Brisley court, in order to 

attack and overthrow principles. 177 In this regard it should be 

emphasised that Davis J did not appeal to an unqualified, actually 

existing community in his judgment. He explicitly referred to the 

constitutional community — which can only be an ideal community — 

and that is why Davis J explicitly invoked the constitutional ideals of 

dignity, equality and freedom. The point Davis J was making was that 

the very reconstitution of the legal order required that good faith 

should be given content with reference to the normative (and 

transformative) commitments of the Constitution. He was not hiding 

in the shadow of the Constitution in order to attack and overthrow 

principles from that vantage point. 

Be that as it may, the majority in Brisley accordingly rejected the 

idea that good faith can be utilised as an independent, free-floating 

basis for the invalidation or unenforceability of contractual 

provisions 178 on the basis that good faith as ‘the legal convictions of 

the community’ would prove to be too nebulous a concept to 

determine the enforceability of contractual terms. The Court stated, 

however, that good faith is a fundamental principle which in general 

underlies the law of contract and becomes expressed in its particular 

rules and principles. 179 

At this point in the judgment the Brisley court introduced what it 

called another underlying value of the law of contract: the principle 

175 Brisley (n 13 above) para 21. 

176 Brisley (n 13 above) 15 para 21. ‘Om eensklaps aan regters 'n diskresie te verleen 

om kontraktuele beginsels te verontagsaam wanneer hulle dit as onredelik of 

onbillik beskou, is in stryd met hierdie werkswyse. Die gevolg sal immers wees dat 

die beginsel van pacta sunt servanda grotendeels verontagsaam sal word omdat 

die afdwingbaarheid van kontraktuele bepalings sal afhang van wat 'n bepaalde 

regter in die omstandighede as redelik en billik beskou.’ This statement 

prompted L Hawthorne ‘The end of bona fides’ (2003) South African Mercantile 

Law Journal 271 276 to remark as follows: ‘This cri de coeur has two deplorable 

aspects: the tacit belief that the majority of contractual terms are not 

reasonable or equitable, and the belief that judges are not to be trusted with 

what is reasonable and equitable.’ 

177 Brisley (n 13 above) 16D para 24. 

178 Brisley (n 13 above) 16D para 22 

179 Brisley (as above).


that public policy requires that contracts that have been entered into 

freely and in all earnest by parties with full capacity should be 

enforced. 180 The Court continued to hold that these two underlying 

values (good faith and freedom of contract) sometimes come into 

conflict with each other and that it is the task of the Courts and the 

SCA principally to weigh these values against each other and to make 

gradual and slow changes on occasions when they appear to be 

necessary. 181 

Again, this judgment makes it clear that good faith and pacta 

servanda sunt are positioned as opposing values, that is to say an 

instrumental politics is set up between the two underlying values. But 

it is also the case that it is at this point that the Brisley judgment 

deconstructs itself. 182 It does so in the following way: When the court 

discusses the idea of an underlying value in relation to good faith, it 

makes the point that it is in the very nature of an underlying value 

that courts do not use it directly to adjudicate contracts. Rather, they 

wait for the underlying value to become embodied in the particular 

rules of the law of contract. For this reason the Court concludes that 

good faith is not an independent or free-floating basis for the 

invalidation of contractual terms. But when it comes to the discussion 

of the other underlying value — the enforcement of contracts in the 

public interest — the Court does not realise the implications of its own 

opinion regarding underlying values. Certainly, the consequences of 

identifying freedom of contract as an underlying value should be the 

same — courts should wait for the underlying value to become 

embodied in the rules of the law of contract. Yet, the history of the 

South African law of contract is the history of the direct reliance on 

the so-called underlying principle of freedom of contract. If there 

were to be no direct reliance on so-called underlying values, then it 

would be hard to imagine how the rules of the law of contract were 

formulated in the first place. The fact that freedom of contract is an 

underlying value has never prevented a court from directly relying on 

it. Why then should the ‘underlying value’-character of good faith 

prevent direct reliance on it? Yet this is exactly what the Brisley 

majority wants us to believe: good faith cannot be directly relied 

upon because it is an underlying value. This was stated even more 

explicitly in Afrox Healthcare Bpk v Strydom: 183 

180 Brisley (n 13 above) 16D para 23 quoting Magna Alloys (n 85 above) 893I-894A. 

181 Brisley (n 13 above) 16D para 24. 

182 J Derrida & M Ferraris A Taste for the Secret (2001) 80: ‘I ought to have specified 

that what happens deconstructs itself in the process. It is not I who deconstruct; 

rather something I called “deconstruction” happens to the experience of a world, 

a culture, a philosophic tradition: “it” deconstructs, ça ne va pas, there is 

something that budges, that is in the process of being dislocated, disjointed, 

disadjoined, and of which I begin to be aware. Something is “deconstructing” and 

it has to be answered for.’ 

183 Afrox Healthcare (n 13 above).


Concerning the place and role of abstract ideas such as good faith, 

reasonableness, fairness and justice, the majority in the Brisley case 

held that, although these considerations are subjacent to our law of 

contract, they do not constitute an independent or ‘free-floating’ basis 

for the setting aside or the non-enforcement of contractual provisions; 

put differently, although these abstract considerations represent the 

foundation and very right of existence of rules of law and can also lead 

to the shaping and transformation of these rules, they are not selfcontained 

rules of law. When it comes to the enforcement of a 

contractual provision, the Court has no discretion and does not act on 

the basis of abstract ideas, but precisely on the basis of crystallised and 

established rules of law. 184 

One clearly sees in this dictum the consistent political privileging of 

the individualism/rules position by way of which the hegemonic 

understanding of freedom of contract is kept firmly in place (which in 

turn keeps the hegemonic order(ing) of the law of contract in place). 

The Court’s unease with good faith (described by Cockrell as ‘the 

epitome of a legal standard that embodies communitarian values of 

altruism, care and concern’ 185 (the opposite thus of individualism)) is 

that it cannot be regarded as a rule of law. On the other hand, 

freedom of contract is simply, unproblematically understood as being 

both an underlying value and a rule which could be relied upon to hold 

the defendant in Afrox for instance, to an agreement that exempted 

a hospital from liability for everything but its intentional infliction of 

harm. 186 

A number of deductions regarding good faith and freedom of 

contract in the South African law of contract can be made at this 

point. The first is that good faith is currently not considered in the 

South African law of contract to be a part of the public policy 

requirement (which, from a transformative point of view, presents its 

own set of problems, primarily in that public policy becomes closed 

off from the ethical question generally when good faith does not form 

part of it). Second, good faith is not considered to be an independent 

or free-floating basis for the invalidation or non-enforcement of 

contracts or contractual terms. Third, the latter is the case because 

good faith is an underlying value of the South African law of contract 

and is/becomes embodied in the particular rules of the law of 

184 Afrox Healthcare (n 13 above) 40H-41A (author’s translation from the original 

Afrikaans). The point was restated in South African Forestry (note 13 above) 

338J-339B as follows: ‘although abstract values such as good faith, reasonableness 

and fairness are fundamental to our law of contract, they do not 

constitute independent substantive rules that courts can employ to intervene in 

contractual relationships. These abstract values perform creative, informative 

and controlling functions through established rules of the law of contract. They 

cannot be acted upon by the courts directly ... After all, it has been said that 

fairness and justice, like beauty, often lie in the eye of the beholder.’ 


186 Afrox Healthcare (n 13 above).


contract. (How this happens is uncertain given that courts are not 

allowed to touch good faith.) Fourth, freedom of contract understood 

as the maxim pacta servanda sunt, although itself an underlying value 

of the law of contract, is and can be used in a direct way. 187 Fifth, 

there appears to be no reason other than an ideological or political 

one which would explain why one underlying value — freedom of 

contract — can be used to develop rules by way of which a party can 

be held to or released from her agreements, whereas another 

underlying value — good faith — cannot be used in this way. 

In addition, a matter that is repeatedly ignored in the debate 

regarding the role of good faith in the South African law of contract is 

the implications for the law of contract of the newly constituted legal 

order. In the Brisley case, for instance, the Court faulted Davis J’s and 

other judges’ appeal to this order as putting in place an ideal 

community to which can be appealed in the law of contract by way of 

good faith (as part of the public policy requirement). The stance of 

the Brisley court represents a static, closed view of a legal system in 

which good faith has somehow been worked out in the past and 

incorporated into the rules of the law of contract. As such, this view 

questions the dynamic role of the bona fides in relation to the 

transformation of the South African law of contract and, as Olivier JA 

suggested in Eerste Nasionale Bank, it is out of step both with the 

spirit of our law and with the needs of the community. 188 

In 2007 the Constitutional Court delivered two decisions that 

involved the tangled web weaved with freedom of contract, good 

faith, the exceptio doli generalis and public policy in the South 

African law of contract. Below I shall critically evaluate these 

decisions after which I shall argue for a re-emphasis on good faith as 

the ethical element of contract. I believe that this re-emphasis is 

mandated by an understanding of the new legal order as aspiring to 

the ideal of civic friendship. By deploying the word ‘re-emphasis’ I 

want to signal my belief that the normative character of the 

Constitution requires us to consider good faith as part of the public 

policy requirement in two ways — both in that contracts concluded 

contrary to public policy are unenforceable but also in that public 

policy requires that parties who enter into agreements freely 

187 

In this context, D Hutchison ‘Non-variation clauses in contract: any escape from 

the Shifren straitjacket?’ (2001) 118 South African Law Journal 720 745 has 

remarked that ‘to reach directly for the baton of good faith would be to confess 

to a want of technical expertise or creativity. Palm-tree justice no doubt has its 

virtues, but as lawyers we should adhere to the ideal of justice according to law.’ 

Hutchison does not explain why it is acceptable and in accordance with the ‘ideal 

of justice according to law’ to reach directly for the baton of freedom of 

contract. Surely, another form of ‘palm-tree justice’ is in play when one reaches 

directly for the baton of freedom of contract as itself an underlying value of the 

law of contract? 

188 Eerste Nasionale Bank (n 128 above) 319H-320B.


(ethically, in good faith) should be held to them. The meaning of 

freedom of contract in this way becomes constituted by a 

transformative understanding of good faith. 

5 Good faith and the Constitution: Towards 

transforming 189 the hegemony 


In the Preface to the fifth edition of The Law of Contract in South 

Africa, Richard Christie wrote as follows: 

In the preface to the fourth edition I was rash enough to express the 

hope that the judges would further develop the concepts of good faith 

and public policy. This they have done, but not in the way I expected. 

Since Brisley v Drotsky, Afrox Healthcare Bpk v Strydom and South 

African Forestry Co Ltd v York Timbers good faith as an abstract value 

cannot be used to intervene in contractual relationships, but public 

policy still can. To achieve a just result there are advantages in using the 

concept of public policy which our law has developed over the centuries 

and has linked to the Constitution, rather than the less familiar concept 

of good faith. But it will be a mistake to regard the door as forever 

closed ...’ 190 

Prof Christie has also suggested that the courts are likely to find that 

the Sasfin principle (namely that a court will not enforce a contract 

if its enforcement would be contrary to public policy) is the most 

serviceable instrument for developing the common law of contract to 

give effect to a provision of the Bill of Rights. 191 Below we shall see 

that this has indeed become the preferred route when it comes to the 

enforceability (and of course also constitutionality) of contractual 

terms. But we shall also see that the Constitutional Court did indeed 

leave the door wide open in its 2007 decision in Barkhuizen. 192 The 

question thus currently looming large in the South African law of 

contract is whether the ‘objective, normative value system’ 193 of the 

Constitution requires the introduction of a good faith jurisdiction in 

the law of contract as part of the Sasfin principle. The judgments that 

189 The way in which I employ transformation in relation to hegemony closely relates 

to Drucilla Cornell’s description of transformation. For her transformation means 

that ‘a system can so alter itself that it not only no longer confirms its identity, 

but disconfirms it and, indeed, through its very iterability, generates new 

meanings which can be further pursued and enhanced by the sociosymbolic 

practice of the political contestant within its milieu.’ Cornell (n 100 above) 2. 

190 RH Christie The law of contract in South Africa (5ed, 006) v. 

191 n 190 above, 348. 

192 Barkhuizen (n 13 above). 

193 Carmichele v Minister of Safety and Security and Another (Centre for Applied 

Legal Studies Intervening) 2001 4 SA 938 (CC) (Carmichele) para 54.


supported Olivier JA’s opinion in Eerste Nasionale Bank certainly 

argue that it does. It is implicit in the judgments of the SCA that it 

does not. A related question is whether the above question requires 

reconsideration of the exceptio doli generalis as a technical remedy. 

The latter question was also before the Constitutional Court in 2007. 

5.2 Crown Restaurant CC v Gold Reef City Theme Park 

(Pty) Ltd 194 

The facts in this case were substantially similar to the facts in Brisley 

save for the fact that the parties were both commercial entities. 

Crown Restaurant fell into arrears in terms of a written lease 

agreement with Gold Reef City Theme Park, which subsequently 

cancelled the lease and applied for an ejectment order. Crown 

Restaurant, however, contended that Gold Reef City ‘had verbally 

agreed to grant it an indulgence and allow it time within which to 

make proposals for settling the rental arrears.’ 195 This verbal 

indulgence, it was argued on behalf of Crown Restaurant, meant that 

the company implicitly waived its right to cancel the lease. Gold Reef 

City, however, initially denied the verbal agreement, then conceded 

that it did ‘allow’ Crown Restaurant to make proposals for settlement 

within a stipulated time period 196 but ultimately contended that it 

was in any event entitled to rely on the non-variation clause in the 

written agreement. Gold Reef City further relied on a clause that 

stipulated that no indulgence granted precluded either party from 

enforcing any of its rights in terms of the agreement. 197 

Gold Reef City was successful in the Court a quo. An application 

for leave to appeal to the SCA was unsuccessful. Crown Restaurant 

subsequently applied for leave to appeal to the Constitutional 

Court. 198 In the application for leave to appeal to the Constitutional 

Court, the Court was asked — for the first time in the proceedings - to 

consider the reintroduction of the exceptio doli generalis by way of a 

development of the common law under section 39(2) of the 

Constitution. The applicant contended that the remedy as an 

‘equitable remedy is in line with constitutional values’ 199 and that on 

these particular facts it would — given the verbal agreement — be 

unconscionable for Gold Reef City to rely on the written agreement. 

There were thus two distinct issues before the court: first, whether 

the common law should be developed so as to reintroduce the 

exceptio doli generalis as an ‘equitable remedy in line with 

194 2008 4 SA 16 (CC) (Crown Restaurant). 

195 n 194 above, para 1 

196 See n 194 above, Applicant’s Founding Affidavit para 23.2.(d). 

197 n 194 above, para 2. 


199 As above.


constitutional values’ and second, if the answer to the first question 

is in the affirmative, whether on the facts the applicant could succeed 

with such a defence. 

In its judgment the Constitutional Court pointed out that the trial 

court judge ‘was called upon to deal only with the waiver defence and 

did so. He was not invited to develop the common law of contract to 

promote the spirit, purport and objects of the Bill of Rights, nor to 

address any of the other constitutional issues now raised by the 

applicant.’ 200 After this statement the Court elaborated on the 

undesirability for it to sit as a court of first and last instance. It stated 

that the High Court and the SCA have a vital role to play when it comes 

to the development of the common law of contract and concluded 

that there are no compelling reasons for the Court to deal (as a court 

of first and last instance) with the issues raised by the applicant. It 

again warned litigants that ‘care should be taken to identify properly 

at the time of the institution of the proceedings which constitutional 

issue they wish to have addressed.’ 201 Accordingly, the application for 

leave to appeal was dismissed. 

In order to evaluate critically the apparently unproblematic 

nature of the Constitutional Court’s dismissal of the application for 

leave to appeal, it is necessary that we recall here what was said in 

the decision of the Constitutional Court in Carmichele. In this case the 

applicant (plaintiff in the Court a quo) also did not invite either the 

trial court or the SCA to develop the common law in accordance with 

the spirit, purport or objects of the Bill of Rights. 202 She argued the 

issue relating to the development of the common law for the first 

time in the Constitutional Court and contended that had the High 

Court and the SCA applied the relevant provisions of the Constitution 

in determining whether a legal duty (in the context of the law of 

delict) was owed to her, it would have developed the common law and 

would have found that such a legal duty exists. 203 To quote from the 

judgment of Ackermann and Goldstone JJ: 

Despite the failure by the applicant to rely directly upon the provisions 

of either section 35(3) of the [Interim Constitution] or section 39(2) of 

the Constitution in the High Court and SCA, counsel for the respondent 

did not object to this issue being raised in this Court. If covered by the 

pleadings, and in the absence of unfairness, parties are ordinarily not 

200 Crown Restaurant (n 194 above) para 4 (emphasis added). 

201 Crown Restaurant (n 194 above) para 6. 

202 Carmichele (n 193 above) para 41. 

203 Carmichele (n 193 above) para 28.


precluded from raising new legal arguments on appeal. [Reference 

omitted.] 204 

Furthermore, in the course of the Carmichele judgment the Court 

stated the following: 

In South Africa, the [Interim Constitution] brought into operation, in one 

fell swoop, a completely new and different set of legal norms. In these 

circumstances the courts must remain vigilant and should not hesitate to 

ensure that the common law is developed to reflect the spirit, purport 

and objects of the Bill of Rights. We would add, too, that this duty upon 

judges arises in respect both of the civil and criminal law, whether or 

not the parties in any particular case request the court to develop the 

common law under section 39(2). 205 

The Carmichele judgment made it clear that there is a general 

obligation on courts to develop the common law in accordance with 

the spirit, purport and objects of the Bill of Rights. The judgment also 

makes it clear that the duty exists whether or not the parties in any 

particular case explicitly invite the court to develop the common law. 

Furthermore, it is clear that the Carmichele judgment envisages that 

there might be circumstances where a court is obliged to raise the 

matter on its own and require full argument. This would be required 

presumably when the legal issue ‘can hardly be described as an 

insignificant one, lying at an exotic periphery’ 206 of the law in 

question. 

In the light of these findings in Carmichele regarding the general 

obligation to develop the common law, it is not clear why the 

Constitutional Court considered it material (and thus fatal to the 

applicant’s application for leave to appeal) in Crown Restaurant that 

the trial court was not invited to develop the common law. 

Carmichele tells us that judges have a general obligation to consider 

the development of the common law in light of the spirit, purport and 

objects of the Bill of Rights, even when they are not invited by parties 

before them to do so. This is particularly important when a significant 

legal issue calls for constitutional resolution. By the time that this 

case came before the trial court it was clear that much dispute 

existed over, inter alia, the status of the exceptio doli generalis in 

204 Carmichele (n 193 above) para 31. Also see para 38: ‘It does not appear to have 

been suggested that there was any obligation on the High Court or the SCA to 

develop the common law of delict in terms of section 39(2) of the Constitution.’ 

205 Carmichele (n 193 above) para 36. Also see para 39: ‘we do not mean to suggest 

that a court must, in each and every case where the common law is involved, 

embark on an independent exercise as to whether the common law is in need of 

development and, if so, how it is to be developed under section 39(2). At the 

same time there might be circumstances where a court is obliged to raise the 

matter on its own and require full argument from the parties.’ (emphasis 

added.) 



our law, the role of good faith and the enforceability of a Shifren 

clause and/or a no indulgence clause in circumstances of subsequent 

verbal amendments/indulgences (that is, in the context of the waiver 

question). 207 Moreover, it was (and still is) by no means clear how 

these matters are affected by the normative transformation of the 

legal order occasioned by the Constitution. Given the fact that the 

Constitution ‘brought into operation, in one fell swoop, a completely 

new and different set of legal norms’ it could certainly be asked 

whether the question of the development of the common law in this 

area should not have been raised by the trial court judge — as 

mandated by Carmichele — even though he was not explicitly invited 

to consider the development of the common law. It is thus not clear, 

under the circumstances, why further exploration by the trial court of 

the issues that were eventually raised in the Constitutional Court, was 

‘understandably not undertaken’. 208 

In Carmichele the court held that the trial court and the SCA 

should indeed have considered the question regarding the effect of 

section 39(2) even though the parties did not raise the matter 

regarding development of the common law initially. The way in which 

the court dealt with this difficulty in Carmichele is very different from 

the way in which the Crown Restaurant matter was handled. Crown 

Restaurant was dealt with in terms of Constitutional Court rule 

19(6)(b) which states that ‘[a]pplications for leave to appeal may be 

dealt with summarily, without receiving oral or written argument 

other than that contained in the application itself.’ In Carmichele the 

court heard argument on the merits of the dispute in accordance with 

the principle enunciated in what is now Constitutional Court Rule 

19(6)(c). Inter alia, this move allowed the court to deal with the 

question regarding the prospects of success in considering whether to 

grant the application for special leave to appeal. In Crown Restaurant 

the court did not utilise rule 19(6)(c) and did not explicitly deal in its 

judgment with the question of the prospects of success. The refusal 

to make an order in terms of rule 19(6)(c) leaves one to assume that 

in the view of the court the prospects of success were improbable on 

the papers. Yet this is not the reason the court gives for denying the 

application for leave to appeal. The reason it gives for the failure of 

the application is that the applicant did not invite the trial court to 

develop the common law, that is, it did not raise the constitutional 

argument timeously. In Carmichele the remedy decided on in 

circumstances where the constitutional argument was not considered 

by the trial court or the SCA was to refer the matter back to the trial 

207 The trial judge in fact referred to Brisley v Drotsky (n 13 above). See Gold Reef 

City Theme Park (Pty) Ltd v The Crown Restaurant CC unreported, case no: 2006/ 

595 WLD, 7 June 2006 para 23 — where the controversy regarding good faith and 

the exceptio doli generalis arose in the context of waiver. 

208 Crown Restaurant (n 194 above) para 6.


court. Admittedly, a difference between the Crown Restaurant 

matter and the Carmichele matter is that in Carmichele absolution 

was granted whereas in Crown Restaurant a full trial took place. In 

Carmichele, the fact that absolution was granted by the trial court 

provided a reason for referring the matter back to the trial court in 

addition to the referral on the grounds that the trial court had to be 

given the opportunity to consider section 39(2) in relation to the 

matter before it. 209 Even though a full trial had taken place in Crown 

Restaurant, there is in principle no reason why the matter could not 

be referred back to the trial court in that instance for a consideration 

of the relevance of section 39(2). Given the controversy and 

uncertainty in relation to the contract law issues raised in this matter 

and given that many of these contentious issues are raised in the 

context of waiver, a case could well be made that it would have been 

in the interests of justice for the Constitutional Court to have granted 

leave to appeal, upheld the appeal, set aside the order of the trial 

court and to have referred the matter back to the trial court for a 

proper consideration of the section 39(2) obligation in relation to the 

exceptio doli generalis and its philosophical underpinning — good 

faith. 210 This may be a costly exercise — as was the case in Carmichele 

— but it allows opportunity for the constitutional resolution of an issue 

that is crucially in need of it. However, this did not happen because 

the Constitutional Court was clearly of the view that on the papers 

there were not sufficient prospects of success even if the section 

39(2) enquiry required development of the common law in such a way 

that the exceptio doli generalis be reintroduced. 

From the point of view that even if there were prospects of 

success on appeal, a full trial had already occurred in the High Court 

and the cost to have the matter referred back to the trial court would 

have been prohibitive and unfair to the respondent, the 

Constitutional Court could itself have granted leave to appeal and 

could have decided the issue of the development of the common law 

on appeal. However, this was again not considered a viable option 

both presumably on the basis that there were not sufficient prospects 

of success on appeal and on the basis that it would have been 

extraordinary for the Constitutional Court to sit as court of first and 

last instance, because, as was indicated in Carmichele (referring to 

Bequinot) 211 the Court would be put at a ‘grave disadvantage’ were 

it not to have ‘the benefit of a fully considered judgment from either 

the SCA or the High Court’ 212 on the point of the development of the 

common law. In the Crown Restaurant matter it is a pity that a point 

of law in such urgent need of constitutional resolution fell through the 


210 See Glover (n 104 above) 451. 

211 S v Bequinot 1997 2 SA 887 (CC) (Bequinot). 



cracks as it were and led to a judicial impasse which, seemingly, could 

only be normalised/justified by invoking the idea that the common 

law can only be developed when parties invite the courts to develop 

the common law. And it is this idea that sits uncomfortably with the 

‘general obligation’ to develop the common law in line with 

constitutional values enunciated in Carmichele. 213 

However, it is quite possible that there might be another reason 

for the Court’s decision not to grant leave to appeal in the Crown 

Restaurant matter. This reason might relate particularly to the 

controversy regarding the status of the exceptio doli generalis and 

good faith in our law of contract. In other words, the Constitutional 

Court’s refusal to hear argument on the merits and thus also its 

refusal to grant leave to appeal in Crown Restaurant might relate to 

the apprehensive way in which the substantive legal issue is generally 

treated in our law of contract. Simply put, my contention is that the 

Constitutional Court made a policy decision to deal with this case by 

not dealing with it. Implicit in the Court’s reasoning is that it would 

be inappropriate for it to sit as Court of first and last instance (that is 

without the benefit of judgments from the lower courts) on a matter 

as contentious and as vexed as the status of the exceptio doli 

generalis and the good faith approach in our law of contract. In and 

of itself, this refusal might well have amounted to a silent 

endorsement of the hegemonic order(ing) in our law of contract. 

But when considered in the light of the other contract law case 

before the Constitutional Court in 2007, the judgment of the 

Constitutional Court in Crown Restaurant might have been influenced 

by the fact that a more appropriate contract law case for the 

Constitutional Court (that is, a case where the constitutional issue 

had already passed through the trial court and the SCA) was before it 

at the time of Crown Restaurant that would, in any event, provide the 

Court with the opportunity to pronounce on good faith (as the 

philosophical underpinning of the exceptio doli generalis), albeit by 

not pronouncing on it. In other words, the silence we see in Crown 

Restaurant regarding the exceptio (and for that matter, good faith as 

its philosophical underpinning) foreshadows the silence about good 

faith that we shall see in Barkhuizen. The silence in Barkhuizen, 

however, is different in that it is a silence that opens Kafka’s 

proverbial door of the law, whereas Crown Restaurant closed it. 214 


214 See, in general, F Kafka The trial trans B Mitchell (1998).


5.3 Barkhuizen v Napier 

In this case the Constitutional Court had to consider the 

constitutionality of a time limitation clause in a short-term insurance 

policy. 215 The case came before the Constitutional Court on appeal 

from the SCA which held that there was no evidence before it that 

indicated that the contract/policy containing the time-bar clause was 

not entered into freely and voluntarily in the exercise of the 

constitutional rights to dignity, equality and freedom. This led the 

court to the conclusion that constitutional norms and values could not 

operate to invalidate the bargain concluded. 216 

Before proceeding to the judgment of the Constitutional Court, a 

few important points should be made regarding the SCA’s Barkhuizen 

judgment not the least of which is that this judgment constitutes the 

latest in the series of judgments that attempt to provide a 

constitutional fit and justification for the hegemonic order(ing) of the 

law of contract in South Africa. 217 This legitimation of the hegemonic 

order(ing) was most prominent in the Brisley majority judgment and 

separate concurring judgment of Cameron JA. Of his separate 

concurring judgment in Brisley Cameron JA says the following in his 

judgment in Barkhuizen: 

Brisley rejected the notion that the Constitution and its value system 

confer on Judges a general jurisdiction ... or power to decide that 

contractual terms cannot be enforced on the basis of imprecise notions 

of good faith. 218 

Cameron JA then added that the judgments in Brisley and Afrox 

‘affirmed that the common law of contract is subject to the 

Constitution’ 219 and that public policy, as rooted in the founding 

values of the Constitution — human dignity, the achievement of 

equality and the advancement of human rights and freedoms, nonracialism 

and non-sexism — is the proper instrument by way of which 

the courts can invalidate agreements that are offensive to the values 

of the new constitutional order. 220 

Cameron JA’s judgment openly maintains an explicit opposition 

between public policy and good faith. However, Cameron JA offers no 

215 Barkhuizen (n 13 above) para 1. 

216 Napier (n 13 above) para 28. The court held that this lack of evidence was 

attributable to the way in which the matter came before it, namely by way of ‘a 

terse statement of facts’ (para 9). 

217 In a critique of the judgment of the SCA, Deeksha Bhana argues that the SCA 

reaffirmed the classical liberal position. See D Bhana ‘The law of contract and the 

Constitution: Napier v Barkhuizen (SCA)’ (2007) 2 South African Law Journal 269. 

218 Napier (n 13 above) para 7 (emphasis added). 

219 Napier (n 13 above) para 6. 

220 Napier (n 13 above) para 7.


justification as to why good faith should be regarded as more of an 

imprecise and vague notion than the notions in which public policy are 

supposedly rooted: dignity, the achievement of equality, the 

advancement of human rights and freedoms, non-racialism and nonsexism. 

Dignity particularly has in the past been repeatedly criticised 

for being, an imprecise notion. 221 Some have even gone so far as to 

describe equality as an empty concept. 222 Are the concepts of nonracialism 

and non-sexism not also open concepts? And what about ‘the 

advancement of human rights and freedoms’? Moreover, can it not be 

said that in a constitutional democracy founded on these values, good 

faith — as the ethical element of contract — encompasses, in fact 

embraces, all these founding values at once? Hector MacQueen and 

Alfred Cockrell have argued that the South African courts have long 

conceded that public policy is a dynamic concept in the South African 

law of contract. It nevertheless, argues MacQueen and Cockrell, falls 

to be distinguished from good faith. 223 Whereas good faith applies 

community standards ‘much more directly’ to the relationship 

between the contracting parties, public policy enunciates moral and 

social standards that are not necessarily dependent on the 

relationship between the parties. 224 The SCA appears to have, at least 

implicitly, accepted a dynamic conception of public policy. 225 At the 

same time it appears to have adopted the view that public policy 

should be distinguished from good faith, but not along the lines 

suggested by MacQueen and Cockrell. 226 The SCA rejects any direct 

reliance on good faith on the basis that it is too imprecise a notion. 

Yet, it accepts direct reliance on public policy — a concept as dynamic 

and as imprecise as public policy. Such a basis for the opposition 

between good faith and public policy is incoherent and arbitrary. The 

point is that both good faith and public policy are open-ended norms, 

imprecise notions, text that calls for interpretation and justification 

much like the entire Bill of Rights and the founding values and ideals 

of the Constitution. Both good faith and public policy amount to 

devices by way of which freedom of contract (as embodying the value 

of autonomy) is limited (and thus constituted) in the South African 

law of contract. 227 To privilege public policy on the basis that good 

221 See for instance D Davis ‘Equality: The majesty of Legoland jurisprudence’ (1999) 

116 South African Law Journal 398 413 and C McCrudden ‘Human dignity’ 

University of Oxford Faculty of Law Legal Studies Research Paper Series Working 

Paper 10/2006, April 2006. 

222 P Westen ‘The empty idea of equality’ (1982) 95 Harvard Law Review 537. 

223 H MacQueen & A Cockrell ‘Illegal contracts’ in R Zimmerman et al (eds) Mixed 

legal systems in comparative perspective: Property and obligations in Scotland 

and South Africa (2004) 143 173. 

224 As above. 

225 See Magna Alloys (n 85 above) 891 and Sasfin (n 75 above) 7I-J. 

226 See South African Forestry (n 13 above) para 32. 

227 Perhaps the privileging of public policy can be explained by reference to the 

privileging of the mantra that public policy generally favours the strict 

enforcement of contracts, whereas good faith does not, without more, favour 

such a privileging of liberal ideology.


faith is too imprecise a notion (and thereby suggesting that public 

policy is somehow more precise) sits uncomfortably with the context 

of — and openness to — an entire legal order that is founded on more 

or less ‘imprecise’ notions namely dignity, equality and freedom, 

notions to which meaning is given in a long process over time at the 

same time as their meaning always depends — at least to a certain 

extent — on context. 

A further aspect of the SCA’s Barkhuizen judgment to which I want 

to draw attention is the uncritical — and alarming — assumption in the 

judgment that the hegemonic order(ing) of the law of contract in 

South Africa is already constitutionally entrenched. I use the word 

‘alarming’ here given the vast economic disempowerment in South 

Africa and the grossly unequal, undignified and in fact unfree 

bargaining relations and situations in South Africa that the law of 

contract normalises and on a continuing basis sustains. From this 

perspective, it is hard to believe the SCA when it argues that the law 

of contract — in its current permutation — approximates the ideals of 

a transformative legal order in the best possible way, even when it 

concedes that public policy is a dynamic concept the meaning of 

which is to be determined by reference to the values and rights in the 

Constitution. Of course, some argue that the achievement of socioeconomic 

equality is not — and should not be — something which a 

general law of contract can do much — if anything substantial — 

about. Yet, this approach fails to take into account what Woolman 

calls ‘the structured silence of [contractual] disputes that never make 

it to court.’ 228 Moreover, when the Constitution requires in section 

39(2) that when the common law is developed it must be developed 

taking into account the spirit, purport and object of the Bill of Rights, 

the Constitution explicitly involves all law in the project of 

transformative constitutionalism. It thus explicitly requires that one 

takes a view of the common law as something other than the neutral 

backdrop against which interactions between legal subjects occur. 

Judgments such as those delivered in Brisley, Afrox and the SCA’s 

Barkhuizen judgment are attempts to provide the authoritative 

statements that the Constitution did not significantly change/affect 

— that is, it attempts to legitimise — the hegemonic order(ing) in the 

South African law of contract. The positions taken in these judgments 

of course offer a particular ideological reading of the Constitution. 

The ideology that prevails reads the Constitution in a classical liberal 

way against which it is easy to legitimise the hegemonic order(ing) of 

a classical liberal law of contract as in accordance with it. Cameron 

JA’s judgment in Brisley is exemplary of such a classical liberal 

ideology when it comes to the relationship between the Constitution 

228 See S Woolman ‘Category mistakes and the waiver of constitutional rights: A 

response to Deeksha Bhana on Barkhuizen’ (2008) 1 South African Law Journal 10 

11.


and the law of contract. To recall, Cameron JA held that ‘[t]he 

constitutional values of dignity and equality and freedom require that 

courts approach their task of striking down contracts or declining to 

enforce them with perceptive restraint ... contractual autonomy is 

part of freedom’ 229 and ‘shorn of its obscene excesses’ 230 it ‘informs 

also the constitutional value of dignity.’ 231 This judgment (like the 

judgment of the majority) is predicated upon an individualist ethic 

that is not only difficult to square with the altruistic ideals inherent 

in the spirit of a Constitution that aspires to a post-apartheid society; 

it also uncritically assumes that the strict enforcement of contracts 

(that is, a principled — and thus political — commitment to nonintervention 

by the courts in contractual matters) is in service of the 

constitutional values of dignity, equality and freedom. 

The way in which Cameron JA’s judgment in Brisley was later used 

in Afrox was criticised along the lines suggested above. 232 Lubbe for 

instance argued that when the foundational concepts of the 

Constitution are marshalled in the under-nuanced way in which they 

were in Afrox, the founding values of the Constitution serve to do only 

one thing and that is to dissipate pressure on the legitimacy of this 

hegemony. 233 As Wagener argued, the linking of freedom of contract 

with dignity does not, without more, favour the strict enforcement of 

contracts. 234 On the contrary, dignity might well be invoked to justify 

more active judicial control. Bhana and Pieterse also argued in this 

regard that ‘the [constitutional] value of freedom does not equate 

with complete individual liberty and does not found an independent 

right to unlimited contractual liberty.’ 235 

Turning now to the Constitutional Court’s decision in Barkhuizen, 

it must be said that this decision of the Constitutional Court is, in a 

very real sense, not possible without the decisions and positions taken 

229 Brisley (n 13 above) 35E. 

230 Brisley (n 13 above) 35F. (I am also interested in understanding in which 

circumstances the court will regard freedom of contract as sufficiently ‘shorn of 

its obscene excesses’.) 

231 Brisley (as above). 

232 D Bhana & M Pieterse ‘Towards a reconciliation of contract law and constitutional 

values: Brisley and Afrox revisited’(2005) 122 South African Law Journal 865 874; 

G Lubbe ‘Taking fundamental rights seriously: The Bill of Rights and its 

implications for the development of contract law’ (2004) 121 South African Law 

Journal 395. 

233 See Lubbe (n 232 above) 420. 

234 S Wagener ‘A proposal for the reform of the law of contract brought about by the 

judicial development of novel naturalia’ (2005) Responsa Meridiana 19 39-40. 

Also see K Hopkins ‘The influence of the Bill of Rights on the enforcement of 

contract’ (2003) 425 De Rebus 25. 

235 n 232 above, 879. Bhana and Pieterse also argue that the value of freedom ‘is 

reined in significantly by its interaction with the constitutional values of equality 

and dignity’. Again, I do not believe that freedom is somehow reined in by its 

interaction with equality and dignity. Rather, freedom is constituted by its 

interaction with equality and dignity.


in Brisley and Afrox and of course also in the SCA decision in 

Barkhuizen. Here, in the first case before the Constitutional Court on 

the law of contract, it had the benefit of the judgments of the lower 

courts. Things were as they should be. 

In Barkhuizen the appellant argued that the time limitation 

constituted an unjustifiable limitation on the constitutional right to 

approach a court for redress. 236 In the High Court the appellant 

argued on the basis of the direct horizontal application of the 

constitutional right and succeeded in his claim. 237 However, the 

Constitutional Court confirmed in this case the approach followed in 

the decision by the SCA. The proper approach when dealing with the 

constitutionality of contractual provisions is to subject the term to 

the public policy test ‘as evidenced by the constitutional values, in 

particular, those found in the Bill of Rights.’ 238 This means, of course, 

that the route of indirect horizontal application is the preferred route 

when considering the constitutionality of a contractual 

arrangement. 239 In determining whether a particular clause offends 

public policy as animated by the Constitution, the court held that the 

first question is whether the clause itself is reasonable. 240 The second 

question, held the Court, is whether the clause should be enforced in 

light of the circumstances that prevented compliance with it. 241 In 

this leg of the enquiry unequal bargaining power is a factor, but not 

the only factor, that plays a role in the determination of public 

policy. 242 On the first question, the court held that the 

reasonableness of the provision is determined by weighing up the 

public interest in compliance with freely and voluntarily concluded 

terms (or, otherwise put, freedom of contract) against the public 

interest in the right of access to courts. 243 The second question 

effectively involves an enquiry into the reasons why non-compliance 

is manifest. The onus to prove that it was impossible or unreasonable 

in the circumstances to comply with the clause rests on the party who 

alleges it. 244 Thus, where the clause has been found not to be in 

violation of public policy (objectively speaking), it can still be 

escaped where the defendant proves good reason why there was 

failure to comply. 

But there are two aspects regarding the Court’s statements on 

public policy on which I want to dwell for the moment because they 




239 For criticism see Woolman (note 102 above) 774-781. 


241 As above. 

242 As above. 


244 Barkhuizen (n 13 above) para 58.


have implications for good faith. First, I want to emphasise that the 

Court reconfirmed the trite notion that the validity or otherwise of 

contractual terms is determined by asking whether the clause is in 

accordance with public policy. Second, the Court emphatically stated 

that ‘public policy represents the legal convictions of the community; 

it represents those values held most dear by the society.’ 245 It also 

held that ‘[p]ublic policy imports the notions of fairness, justice and 

reasonableness.’ 246 According to the Constitutional Court then, the 

enforceability or not of contractual terms are determined by the legal 

convictions of the community as represented by ‘the values that 

underlie our constitutional democracy’. 247 This statement flies in the 

face of the Brisley judgment in which the Court explicitly and in the 

strongest terms overruled Davis J’s judgment in Mort because it 

attempted to make the enforceability of contractual terms dependent 

on the legal convictions of the community. 

Recall, however, that Davis J referred to the legal convictions of 

the community in relation not to public policy but to good faith. It 

thus appears that, to the extent that Davis J did indicate that the 

enforceability of contractual terms depends on the legal convictions 

of the constitutional community — as that term is understood as the 

aspirational ‘community of the “ought to be”’ 248 — the Constitutional 

Court is in agreement with him and disagrees with the SCA. The only 

disagreement between the Constitutional Court and Davis J is the 

doctrinal name of these legal convictions. The Constitutional Court 

calls it public policy, Davis J calls it good faith. Can it be that the 

Constitutional Court in this case in fact merged public policy with 

good faith? Can it be that good faith after all forms part of public 

policy as Olivier JA argued in his much-maligned judgment in Eerste 

Nasionale Bank? The answer to this question is by no means clear given 

Ngcobo J’s emphatic statement that the facts of the case did not 

require the Court to answer the question whether, under the 

Constitution, the limited role for good faith expounded since Brisley, 

is appropriate. 249 Yet, the Court attributes exactly the same meaning 

to public policy as the meanings that have been attributed to good 

faith. A striking example is the similarity between the definition by 

the Constitutional Court of public policy as ‘importing the notions of 

fairness, justice and reasonableness’ 250 and Olivier JA’s definition of 




248 See D Cornell Moral images of freedom (2007) 22. 

249 Barkhuizen (n 13 above) para 82. This point is emphasised in the judgment of 

O’Regan J (Barkhuizen (n 13 above) para 120): ‘there is in my view no need for 

this court to consider in what circumstances a court may, in terms of the 

principles of contract, decline to enforce a time-limitation clause against a 

particular applicant based on the defences of impossibility or good faith. That 

difficult question can stand over for decision in an appropriate matter.’ 

250 Barkhuizen (n 13 above) para 73.


good faith as realising the community’s legal convictions regarding 

propriety, reasonableness and fairness. 251 

On the other hand, the Court raised the good faith issue as it was 

dealing with the second leg of its test to determine the enforceability 

of the term on the ground of public policy: whether the terms are 

contrary to public policy in the light of the relative situation of the 

contracting parties. The court held that the applicability of the 

principle of good faith would depend on the reason advanced for noncompliance. 

However, the applicant did not disclose his reason for 

non-compliance. This made it impossible to say whether 

‘enforcement of the clause against the applicant would be unfair and 

thus contrary to public policy.’ 252 In other words, the absence of a 

reason for non-compliance blocked, it seems, the consideration of the 

good faith requirement within the public policy test. In the absence 

of pleaded reasons for non-compliance, it would, according to the 

Court, be unfair to the respondent and contrary to pacta servanda 

sunt to excuse Barkhuizen from non-compliance with the time-bar 

clause. 253 For these reasons, the majority accordingly dismissed the 

appeal. 

On my reading it seems then at least probable that the court 

would have been willing to consider good faith (in the context of its 

formulated public policy test) had reasons been advanced for noncompliance 

with the term. In this way the court seems to both accept 

and reject the category distinction that MacQueen and Cockrell make 

between public policy and good faith. The Court clearly views good 

faith as the measure that would curb the enforcement of the term if 

enforcement ‘would be unfair or unjust to the applicant.’ 254 Given 

the Court’s statement regarding the role of unequal bargaining power 

in this leg of its public policy test, the suggestion seems to be that 

part of the good faith test will involve enquiring into the relative 

bargaining position of the parties and the way in which this impacted 

on the contractual arrangement. In this way it appears that the Court 

left the door wide open as regards good faith’s potential role as part 

of the public policy test. Moreover, this means that the exceptio doli 

generalis, albeit in drag, has risen from the grave 255 — a grave in 

which it was in any event never interred. Another implication of the 

judgment is that the Constitutional Court has confirmed that the 

alleged exercise of freedom of contract in violation of both the rights 

in and the spirit, purport and objects of the Constitution, will not be 

enforced on grounds of public policy. 

251 Eerste Nasionale Bank (n 128 above) 319B. 



254 As above. 

255 Glover (n 104 above) 455.


To what extent this confirmation will constitute internal 

limitations on freedom of contract — and thus amount to a 

synchronisation of freedom of contract with the other constitutional 

rights and freedoms in furtherance of a transformative understanding 

— depends on whether there will be a departure from/transformation 

of the hegemonic order(ing) we have interrogated. Moreover, by 

leaving the door of public policy open to good faith, the court has 

created the possibility that good faith may in the future again become 

constitutive of freedom of contract, rather than marginalised and 

seen as opposed to it. Again, everything here depends on the content 

that will be afforded to good faith. It goes without saying that it is 

always possible (and indeed likely) that good faith can be afforded 

weak content that would just further legitimise the hegemonic 

understanding of freedom of contract. To put it perhaps more 

dramatically, it is of course always an imminent danger that good 

faith will simply become a private and privatising device. 256 For this 

reason it becomes necessary to defend a progressive and 

transformative version of good faith. 

6 Negotiating otherwise — good faith, civic 

friendship and freedom of contract 

When considering the justifiable potential content of good faith — as 

the ethical ideal that undergirds the law of contract — I believe it is 

necessary to take as one’s point of departure the nature and the 

aspirations of the reconstituted legal order within which the law of 

contract operates as an institution of the common law subordinated 

to the Constitution. If it is the case that the ‘proper’ approach (post- 

Barkhuizen) to the influence of the Constitution on the common law 

of contract is the public policy enquiry, then good faith — as the 

ethical element of contract, as ‘the law of the responsible 

transaction’ 257 — can and must come to play a role in the 

determination of the public policy requirement in order for the spirit 

of the Constitution as a whole to have a better chance of a 

transformative effect in the law of contract. In other words if there is 

to be talk of and hope for what Johan van der Walt has called 

progressive indirect horizontal application 258 then we cannot afford 

to lose sight of good faith and the transformative potential it holds for 

the South African law of contract. Good faith holds this 

256 See JWG van der Walt ‘Die toekoms van die onderskeid tussen die publiekreg en 

die privaatreg in die lig van die horisontale werking van die grondwet (deel 2)’ 

2000 Tydskrif vir die Suid-Afrikaanse Reg 605 612. 

257 J Derrida ‘On forgiveness’ in J Derrida On cosmopolitanism and forgiveness (2001) 

56. 

258 See in general JWG van der Walt ‘Progressive indirect horizontal application: 

Towards a co-operative relation between common-law and constitutional 

jurisprudence’ (2001) South African Journal on Human Rights 341.


transformative potential because, more than any other normative 

contractual device, it is good faith that invokes, as an ethical notion, 

respect for the dignity of contracting parties, precisely because, as 

MacQueen and Cockrell suggest, good faith focuses the enquiry 

explicitly on the relationship between the particular contractual 

parties. This respect for dignity when it comes to legal relations is a 

core aspect of the newly constituted legal order and its aspirations to 

civic friendship. 

I would like to suggest, in conclusion then, that the Constitution 

aspires to an ideal of civic friendship (a ‘politics’ of friendship if you 

will) which requires, mandates and demands negotiation 

(contracting) otherwise — that is, with respect for and consideration 

of the other contracting party. I believe that the Constitution aspires 

to the post-liberal ideal of civic friendship precisely because of its 

foundational injunction to respect the dignity of all others. 259 This, I 

would argue, requires that we afford content to good faith in the law 

of contract that would maintain it as a post-liberal concept. I want to 

emphasise here that I do not consider the dignity of which the 

Constitution speaks (and on which the instant conception of civic 

friendship and good faith turns) to be inevitably fated as a radically 

individualistic concept that, moreover, forms the basis of exclusively 

negative freedom. This argument is possible because of the links 

drawn in our early dignity jurisprudence between dignity and 

ubuntu. 260 Drucilla Cornell draws the further link between ubuntu, 

(Kantian) dignity and friendship as follows: 

In Kant, I am a friend to myself because of the dignity of my humanity. 

Under ubuntu, I am a friend to myself because others in my community 

have already been friends to me, making me someone who could survive 

at all, and therefore be in the community. It is only because I have 

always been together with others and they with me that I am gathered 

together as a person and sustained in that self-gathering. 261 

It was Hannah Arendt who, in modern political theory, resuscitated 

and defended a ‘politics of friendship’ relying heavily on the work of 

Aristotle. 262 Arendt’s theory of civic friendship offers an account of 

politics that is both post-individualist and post-liberal. In fact, 

Arendt’s theory of civic friendship invokes an understanding of 

citizenship as a radically ‘horizontal allegiance to friends.’ 263 She 

sees in this understanding of political life the possibility of ‘words that 

259 Constitution sec 10. 

260 See, for instance, S v Makwanyane 1995 3 SA 391 (CC) paras 224, 308 & 311. 

261 D Cornell ‘A call for a nuanced constitutional jurisprudence: Ubuntu, dignity and 

reconciliation’ (2004) 19 SA Public Law 666 669. 

262 H Arendt The human condition ed M Canovan (1998) 243. 

263 ME Stortz ‘Beyond justice: friendship in the city’ (1994) XIV(4) Word & World 409 

414 (emphasis added).


are not empty and deeds that are not brutal.’ 264 As Martha Stortz 

comments on Arendt’s concept of civic friendship: 

Civic friendship of all sorts emphasises the interdependence of citizens 

in public life. It articulates a horizontal understanding of citizenship, 

which prizes the relationship to another citizen and places that 

relationship at the centre of civic life. 265 

Most importantly, though, this form of friendship is not at all about 

the association of the self and the same. At its heart, Arendt’s 

understanding of civic friendship turns on plurality — that is, 

difference — because it is plurality in Arendt that is constitutive of the 

political. 266 And plurality conditions us in the sense that our very 

individuality only takes shape through our recognition that we share 

the world with others. 

This immediately implicates the idea of respect, for without 

respect there can be no durability for plurality. And as Arendt puts it 

in The human condition: ‘what love is in its own, narrowly 

circumscribed sphere, respect is in the larger domain of human 

affairs.’ ‘... [R]espect’, so she continues, 

is a kind of ‘friendship’ ... it is a regard for the person from the distance 

which the space of the world puts between us, and this regard is 

independent of qualities which we may admire or of achievements which 

we may highly esteem. 267 

For Arendt then there can be no talk of civic friendship without this 

respect, which, to emphasise, does not depend on conditions of the 

self and the same. However, Arendt conceded that the modern age is 

characterised by the loss of respect and this ‘constitutes a clear 

symptom of the increasing depersonalisation of public and social 

life.’ 268 It is also because of Arendt’s acute awareness of the 

implications of this depersonalisation due to the loss of respect that 

she insists famously at the end of the preface to The origins of 

totalitarianism that, in the aftermath of the disasters of modernity 

‘human dignity needs a new guarantee which can be found only in a 

new law on earth.’ 269 

264 As above. 

265 As above. 

266 Arendt (n 262 above) 7. 

267 n 262 above, 243. 

268 As above. 

269 H Arendt The origins of totalitarianism (1973) ix.


If we recall Derrida’s assertion in Politics of Friendship that there 

is no friendship without respect of the other, that ‘the co-implication 

of responsibility and respect can be felt at the heart of friendship’ 270 

we will begin to see the trembling overlap between the ethical, the 

political and the juridical in the South African Constitution in which 

respect for the dignity of all others comes to play a constitutive role. 

Justice O’Regan has referred explicitly to respect and responsibility 

as ‘the building blocks of the Constitution.’ 271 And this respect is 

owed to the equal human dignity of everyone. 272 On my reading, this 

unconditional horizontality of the South African Constitution appeals 

to a prioritisation of Arendt’s, but also Derrida’s, vision of friendship 

in which the Greek form of love as philia is instituted as a sociopolitical 

concept but here in holding out the promise of a beyond to 

any condition of fraternity or ‘confraternisation’. 273 Shin Chiba 

argues that it is this form of love that permeates Arendt’s search for 

a new and durable vinculum or social bond, which, let us recall, is the 

original name for law. 274 The implication of this institutionalisation is 

that it is in fact respect, as phronesis and praxis for the Other 275 — 

respect that resides in love of the Other — that becomes constitutive 

of the legal order. 

Judge Dennis Davis has been foremost amongst those who have 

argued that the notion of respect for equal dignity (civic friendship) 

is translated into the law of contract by way of good faith. 276 He has 

also opposed this concept to South Africa’s political past. As Davis J 

puts it: 

this concept of good faith is congruent with the underlying vision of our 

Constitution ... to the extent that our Constitution seeks to transform 

our society from its past, it is self-evident that apartheid represented 

the very opposite of good faith ... Our Constitution seeks to develop a 

community where each will have respect for the other ... Whatever the 

uncertainty, the principle of good faith must require that the parties act 

honestly in their commercial dealings. Where one party promotes its 

own interests at the expense of another in so unreasonable a manner as 

to destroy the very basis of consensus between the two parties, the 

270 J Derrida Politics of friendship trans G Collins (1997) 252. Of course, Derrida is 

problematising Kant’s formulations with regard to friendship here, notably Kant’s 

separation of love from friendship. Time and space does not permit a discussion 

of this deconstruction. 

271 C O’Regan ‘The three r's of the Constitution: Responsibility, respect and rights’ 

(2004) Acta Juridica 86. 

272 As above. 

273 See S Chiba ‘Hannah Arendt on love and the political: Love, friendship, and 

citizenship’ (1995) 57 The Review of Politics 505 512-513. Also see Derrida (n 267 

above) viii. 

274 n 273 above, 509. 

275 Douzinas & Geary (n 68 above) 40. 

276 But also see Grové (n 130 above) 689.


principle of good faith can be employed to trump the public interest 

inherent in the principle of the enforcement of a contract. 277 

In his judgment in Mort Davis J quoted the words of Reinhard 

Zimmermann 278 in relation to this aspect: 

the principles of equality and dignity require that the parties to a 

contract do none other but ‘adhere to a minimum threshold of mutual 

respect in which the unreasonable and one-sided promotion of one’s own 

interest at the expense of the other infringes the principle of good faith 

to such a degree as to outweigh the public interest in the sanctity of 

contracts’. 279 

This is the formulation for which Barkhuizen leaves the door open. 

Ultimately in our context, respect for the Other resides, as Davis J has 

also indicated, in the acknowledgement that: 

[a] transformative constitution needs to engage with concepts of power 

and community ... the concept of contractual autonomy within the 

concept of a community of contracting persons must mean something 

distinct from a libertarian connotation, particularly if the concept of 

ubuntu is to play any role in our law. 280 

These statements about the law of contract as law subordinated 

under the new Constitution are explicit interpretations of the 

constitutional mandate as requiring a re-emphasis on the ethical 

element of contract in the furtherance of a post-liberal or positive 

freedom of contract. A freedom of contract that comes to understand 

that conduct cannot be characterised as free when it disrespects/ 

violates dignity, 281 when it pretends that contract is a relation 

between things and not between persons, when it does not proceed 

according to respect for whoever is on the other side of the 

negotiation. To quote Colombo: ‘Good faith implies a developed sense 

277 Shoprite Checkers (Pty) Ltd v Bumpers Schwarmas CC and Others 2002 6 SA 202 

(C) (Shoprite) 215G-216A. 

278 R Zimmermann ‘Good faith and equity’ in R Zimmerman and D Visser (eds) 

Southern cross: civil and common law in South Africa (1996) 259-260. 

279 n 168 above, 475B-C. 

280 Advtech (n 95 above) 388A. 

281 The mantra in Ackermann J’s celebrated minority judgment in Ferreira v Levin 

NO and Others; Vryenhoek and Others v Powell NO and Others 1996 1 SA 984 (CC) 

(Ferreira) para 49 ‘to deny people their freedom is to deny them their dignity’ 

should perhaps then be reconsidered in this light because, on this understanding 

in the law of contract, to deny people their dignity is to deny one’s own freedom 

(of contract).


of community and a high level of awareness of personal 

responsibilities towards society.’ 282 

When Lubbe argues that our understanding of dignity in contract 

should be informed by Kant’s precept that people should not 

act contrary to the equally necessary self-esteem of others, as human 

beings, that is, [they are] under obligation to acknowledge, in a 

practical way, the dignity of humanity in every other human being 283 

he already acknowledges the relational aspect of contract that would 

amount to a Hegelian version of contract as ‘reciprocal 

recognition.’ 284 Rosenfeld explains that 

under this paradigm contract is not only compatible with the 

reconciliation of individual autonomy and communal values, but it is also 

suited to foster the mutual determination of the individual and — or, 

perhaps more precisely, against — the communal. 285 

In turn this understanding of contract in terms of relationality and 

mutual recognition bears marked resemblances with Nancy’s analysis 

of existence as ‘being singular plural.’ 286 Once we understand that 

both good faith and freedom of contract are essentially relational 

concepts it becomes possible to think a law of contract beyond 

continuing and continuous exploitation, subservience and injustice — 

beyond ‘our proud and deadly dualisms.’ 287 

The legislature has made substantial inroads in this regard and, at 

the time of writing the Consumer Protection Bill had been approved 

by Parliament and is awaiting the President’s signature. Amongst 

various other consumer protection mechanisms the Bill explicitly 

prohibits ‘unfair, unreasonable or unjust’ contracting. 288 In addition, 

courts will be required in terms of clause 52(3) to decide whether a 

contract is unconscionable, unjust, unreasonable or unfair. Although 

guidelines for the determination of the unreasonableness, injustice or 

unfairness are set out in clause 48(2) they are explicitly intended not 

282 S Colombo ‘Fascism, community and the paradox of good faith’ (1994) 113 South 

African Law Journal 482 482. 

283 Lubbe (n 232 above) 422, quoting I Kant The Metaphysics of Morals M Gregor ed 

(1996) 209. 

284 M Rosenfeld ‘Hegel and the dialectics of contract’ (1989) 10 Cardozo Law Review 

1199 1200. 

285 As above. 

286 JL Nancy Being singular plural (2000). In this work Nancy denounces the liberal 

idea of an individual atomistic self as the subject of ontology. As Nancy puts it: 

‘That which exists, ... coexists because it exists. The co-implication of existing 

[l’exister] is the sharing of the world.’ 

287 G Rose Mourning becomes the law (1996) 76. 

288 Consumer Protection Bill 2008 clause 48(1) available at http:// 

www.thedti.gov.za/ccrdlawreview/DraftConsumerProtectionBill.htm (accessed 

20 June 2008).


to limit the generality of the section prohibiting unreasonable, unfair 

or unjust contracting. 289 This Bill thus already points to a future law 

of contract that will be required to be far more concerned than it has 

been up to now, with good faith as the ethical element of contract 

and thus with the spirit of the ideal of civic friendship in South Africa. 

7 Conclusion 

At the heart of contract lies the idea that I have an interest in 

something of yours and that you have an interest in something of 

mine. The hegemonic capitalist over-emphasis on the things and the 

utter neglect of the persons who have these things, has provided an 

extremely distorted version of what the word ‘interest’ in the above 

formulation originally entails. Levinas takes his cue in his re-emphasis 

on the original meaning of the word ‘interest’ from Hegel, for whom 

contract exists as reciprocal recognition and freedom of contract 

exists in the free willing of one’s duty on the road to selfactualisation. 

290 In his essay ‘Sociality and money’, 291 Levinas 

reminds us that the word interest is a composite of ‘inter’ that is, of 

‘between’ and of ‘esse’, that is ‘being’. For Levinas, essence is always 

interestedness or differently put being is always inter-estedness. 

Thus, to have a contractual interest always already implies 

289 For incisive criticism and important suggestions see TJ Naudé ‘Unfair contract 

terms legislation: The implications of why we need it for its formulation and 

application’ (2006) 17 Stellenbosch Law Review 361 374. Also see TJ Naudé ‘The 

use of black and grey lists in unfair contract terms legislation in comparative 

perspective’ (2007) 124 South African Law Journal 128. In the latter article (135- 

136) Prof Naudé criticises this author for — in her opinion, exclusively — 

supporting an open-ended version (based on good faith) of the Unfair Contractual 

Terms Bill proposed by the Working Committee of the South African Law 

Commission (as it then was) in SA Law Commission Discussion Paper 65 

Unreasonable Stipulations in Contracts and the Rectification of Contracts (July 

1996). See AJ Barnard ‘A critical legal argument for contractual justice in the 

South African law of contract’ unpublished LLD dissertation, University of 

Pretoria, 2005 241. Apart from the fact that there is no disagreement between 

myself and Prof Naudé that, to quote Prof Naudé, ‘the use of non-exhaustive 

guidelines and a grey list by a court cannot be equated with the “mechanical 

application of precedent” and does not preclude a decision guided by “the 

relational, the collective and the transformative” at all’, I wish to make it clear 

that my doctoral study focused on the rhetorical history and lexicon of 

adjudication in the South African law of contract (which was and still is primarily 

based on the application of precedent). This mode, I argued, was, and still is in 

danger of perpetuating an individualistic interpretation of both future guidelines 

or lists of suspect terms and an individualistic reification of open-ended norms. 

This is an issue Prof Naudé herself raises — but then dismisses — as a potential 

problem with so-called ‘grey lists’. Where Prof Naudé and I part company as I see 

it, is that whereas I support Kennedy’s belief that individualism prefers the rule 

form and the commitment to formal realisability (which in turn fuels the false 

belief in neutrality), Prof Naudé believes that the rule form does not necessarily 

follow individualist political morality. 

290 See GWF Hegel Elements of the philosophy of right ed A Wood (1991) 104. 

291 E Levinas ‘Sociality and money’ (2007) 16 Business Ethics: A European Review 203 

206-207.


togetherness, being between and with. Levinas admits that ‘in the 

concrete of everyday life, human interestedness always already 

conceals itself’ 292 but at the same time this concealment points to a 

beyond to it. This is the astonishing possibility of the human being 

between, or amongst, beings: the possibility of unconcealment, of 

giving his place, of sacrificing himself for the other, of the goodness 

of giving, ‘the positivity of an attachment to being as being for the 

Other.’ 293 

What is needed in the South African law of contract is the best 

possible negotiation between these two dimensions of the human 

condition — inter-est and dis-inter-estedness. The hegemonic 

order(ing) of the South African law of contract at present does not 

allow for the best possible negotiation between human interestedness 

and human dis-interest. This is the case because 

‘individualism is the structure of the status quo.’ 294 In this sense it 

does not allow for real transformation. This is the case because the 

obsession with individualism and the rule form does not allow for that 

which is inherent in transformation: the turn towards the Other. Yet, 

as Cornell reminds us: ‘[t]ransformation is demanded of us precisely 

because there is no self-enclosed subject who can truly cut herself off 

from the Other. We are constantly being challenged by otherness’. 295 

It is only once the (im)possibility of the ethical relation is explicitly 

inscribed and described in the South African law of contract by way of 

at least the explicit incorporation and exploration of good faith in the 

public policy/legality requirement that the best negotiation/the 

negotiation otherwise, becomes possible. The Constitutional Court’s 

decision in Barkhuizen and the Consumer Protection Bill allow for this 

possibility. To that extent the possibility of the impossible ethical 

relation can be part of the law’s work. 

In the second of the films in the famous Matrix trilogy, the lead 

character, Neo, consults a present day (or perhaps futuristic) version 

of the ancient Greek oracle, which in the film is represented as a wise 

older woman. Let me end with the counsel the Oracle gives to Neo. 

Emphatically she states: ‘I am interested in one thing Neo — the 

future. And believe me, the only way to get there is together.’ 296 

292 n 290 above, 204. 

293 Levinas (n 290 above) 205. 

294 Kennedy (n 6 above) 1775. 

295 Cornell (n 100 above) 41. 

296 A & L Wachowski (dirs) The Matrix reloaded (2003).

CLEARING THE INTERSECTION? 

ADMINISTRATIVE LAW AND LABOUR LAW 

IN THE CONSTITUTIONAL COURT 

* Professor of Law, University of the Witwatersrand, Johannesburg. I am grateful to 

those who read and commented on the first draft of this article: Clive Plasket, the 

participants in the CCR Conference of 6-7 August 2008 and the CCR’s anonymous 

reviewers. Their perceptive comments were of great assistance to me, and they 

are not to blame for any deficiencies that remain. 

209 

Cora Hoexter * 

The dismissal of employees in the public sector is as much a matter of 

administrative law as it is one of labour law — or so we all thought, 

encouraged by the judgment of the Appellate Division in 

Administrator, Transvaal v Zenzile. 1 But that decision comes from an 

era in which labour law was notable for its failure to protect 

employees in the public sector and administrative law had to be 

called on to perform this role. 2 Much has changed since then, and in 

recent years there has been debate about the continuing relevance of 

administrative law in cases of dismissal and other employmentrelated 

conduct. 

In 2007 the Constitutional Court was asked to settle the debate in 

Chirwa v Transnet Ltd & Others, 3 a case of public-sector dismissal 

that called attention to Zenzile once more. The Court also dealt with 

two other cases of dismissal that touched on questions pertinent to 

administrative law. Masetlha v President of the Republic of South 

Africa & Another 4 raised the issue of procedural fairness in the 

dismissal of a high-ranking public official. Sidumo & Another v 

Rustenburg Platinum Mines Ltd & Others 5 arose out of a dismissal in 

the private rather than the public sector, but administrative law was 

again relevant by virtue of an arbitration award by the Commission for 

1 

2 

1991 1 SA 21 (A) (Zenzile). 

The legislation of the time, the Labour Relations Act 28 of 1956, excluded state 

3 

4 

5 

employees from its ambit. 

2008 29 ILJ 73 (CC), handed down on 28 November 2007 (Chirwa). 

2008 1 SA 566 (CC), handed down on 3 October 2007 (Masetlha). 

2008 2 SA 24 (CC), handed down on 5 October 2007 (Sidumo).

210 Clearing the intersection? Administrative law and labour law in the CC 

Conciliation, Mediation and Arbitration (CCMA). In all, the Court had 

three opportunities in 2007 to explore the intersection of labour law 

and administrative law and to elucidate the relationship between 

sections 23 and 33 of the Constitution — and between the statutes 

that are supposed to give effect to those rights, the Labour Relations 

Act 66 of 1995 (LRA) and the Promotion of Administrative Justice Act 

3 of 2000 (PAJA). 

As I shall show in this article, the results are interesting; but they 

are also rather disappointing, at least from the perspective of an 

administrative lawyer. In my commentary on these three cases (in the 

order Sidumo, Chirwa and Masetlha) that is the perspective I adopt. 

My aim is to elucidate aspects of these cases as they touch on 

administrative law and, in the process, to make three main points. 

First, it seems to me that from the various judgments of the 

Constitutional Court no coherent picture emerges of this intersection 

or of the relationship between the rights concerned. Indeed, in their 

constitutional approach or ethos the Sidumo and Chirwa majority 

judgments seem quite different, notwithstanding that the two cases 

were decided by almost identical benches and within weeks of each 

other. A theme emerging strongly from Sidumo, and to a lesser extent 

from Masetlha, is that fundamental rights cannot be hermetically 

sealed from one other. In Sidumo labour law and administrative law 

intermingle and converge, and in Masetlha administrative law in the 

broad sense is permitted to solve a problem that labour law 

apparently fails to address. In Chirwa, by contrast, the majority 

insists on the strict compartmentalisation of fundamental rights and 

on the pre-eminence of one right to the exclusion of the other. 

Sections 23 and 33 of the Bill of Rights are placed in separate 

jurisdictional boxes, and the attitude seems to be that labour law and 

administrative law must have nothing to do with one other. 

Secondly, in Chirwa the various approaches to the ‘administrative 

action’ issue are all problematic to some extent, and the majority 

view on the question of jurisdiction is wholly unconvincing. While the 

Court certainly managed to tidy up the busy intersection of labour law 

and administrative law in this case, it did so at the expense of the 

Constitutional Court’s own precedent — and apparently without much 

faith in the success of its efforts. 

Thirdly, whatever labour lawyers may think of the majority 

judgment in Masetlha, administrative lawyers are likely to be 

perturbed by it. For us it is a decision that appears to set the law of 

procedural fairness back twenty years.

1 Sidumo 


The appellant, Mr Sidumo, had been employed by Rustenburg 

Platinum Mines Ltd (the mine) for almost fifteen years when he was 

dismissed for negligent failure to apply established search procedures 

at a redressing station. He contested his dismissal and referred the 

dispute to the CCMA in terms of section 191 of the LRA. Conciliation 

failed and the matter went for compulsory arbitration. The 

commissioner found that while Sidumo was indeed guilty of 

misconduct, dismissal was not a fair or appropriate sanction; and he 

was reinstated with three months’ compensation. 

The mine applied for review of the award under s 145 of the LRA. 

The Labour Court dismissed its application, and the Labour Appeal 

Court dismissed an appeal to it. The mine was more successful on 

subsequent appeal to the Supreme Court of Appeal, which overturned 

these two decisions and found the dismissal to have been fair. In 

Rustenburg Platinum Mines Ltd (Rustenburg Section) v Commission 

for Conciliation, Mediation and Arbitration 6 it held that the 

commissioner’s decision to reinstate Sidumo had not been rationally 

connected to the information before him. 

The reasoning behind this conclusion is significant, for section 145 

of the LRA lists only four grounds for the review of arbitration awards 

and irrationality does not feature explicitly amongst them. 7 However, 

Cameron JA found for the Court that the PAJA had extended the 

grounds of review available to parties to CCMA arbitrations — there 

being no doubt that CCMA arbitrations qualified as administrative 

action. 8 In this regard he drew on Carephone (Pty) Ltd v Marcus NO, 9 

a pre-PAJA case in which the Labour Appeal Court held that section 

145 was suffused with the standard of justifiability in section 24 of the 

interim Constitution 10 (the predecessor to section 33). Cameron JA 

held that as general legislation relating to administrative action the 

PAJA had ‘superseded the LRA’s specialised enactment within the 

[labour] field’, 11 and it thus subsumed and overrode the more 

restrictive grounds of review in section 145 of the LRA. 12 An unrelated 

but equally significant holding of the Supreme Court of Appeal in 

6 2007 1 SA 576 (SCA) (Rustenburg). 

7 Section 145(2)(a) allows review to be obtained for misconduct on the part of the 

commissioner; for a gross irregularity in the conduct of the arbitration 

proceedings; and where the commissioner exceeded his powers. Section 145(2)(b) 

provides for review where ‘an award has been improperly obtained’. 

8 

Rustenburg (n 6 above) para 25, though without detailed analysis of the question. 

9 1999 3 SA 304 (LAC) (Carephone) paras 15 and 30 et seq. 

10 Constitution of the Republic of South Africa, 1993. 

11 

Rustenburg (n 6 above) para 24. 

12 Rustenburg (n 6 above) para 25.


Rustenburg was that CCMA commissioners should give some deference 

to the employer’s choice of sanction. 13 

An appeal to the Constitutional Court was unanimously upheld and 

the award of the commissioner restored. The Court considered two 

main issues. On the first it held that a commissioner is not required to 

defer to the decision of the employer but rather to consider all the 

relevant circumstances. 14 On the second issue, the applicability of 

the PAJA, the Court agreed on the result but was divided in its 

approach. 

Writing for a narrow majority, Navsa AJ held that arbitral 

decisions of CCMA commissioners were indeed administrative action 

in terms of section 33 of the Constitution, 15 but that the PAJA was not 

applicable to such decisions. In his judgment the LRA had to be 

regarded as specialised legislation alongside the PAJA — that is, as 

legislation dealing with ‘administrative action’ in the labour sphere. 16 

As such, the LRA had to be interpreted in accordance with section 

33. 17 Thus section 145 of the LRA was suffused with the content of 

section 33, including the constitutional standard of reasonableness 18 

— just as in Carephone the Labour Appeal Court had found section 145 

to be suffused with the former standard of justifiability. Applying the 

new standard, Navsa AJ found that the commissioner’s decision was a 

reasonable one: it was not a decision ‘that a reasonable decisionmaker 

could not reach’. 19 Four justices concurred in his judgment. 20 

Ngcobo J, who had the support of three colleagues, 21 arrived at a 

similar conclusion by means of different reasoning. In his judgment 

CCMA arbitrations were not administrative but judicial action, which 

meant that section 33 had no role to play. 22 However, CCMA 

commissioners exercised public power and were thus constrained by 

other constitutional requirements, including those found in the 

doctrine of legality — an aspect of the rule of law — and in sections 

23 and 34 of the Bill of Rights. 23 These constraints all informed the 

interpretation of section 145(2)(a) of the LRA and considerably 

13 

14 

Rustenburg (n 6 above) para 42 et seq, and see para 48(d). 

Sidumo (n 5 above) para 79 in the judgment of Navsa JA, and see para 161 in the 

15 

16 

17 

18 

19 

judgment of Ngcobo J. 

Sidumo (n 5 above) para 88. 


Sidumo (n 5 above) paras 89 and 91. 

Sidumo (n 5 above) paras 106 and 110. 

Sidumo (n 5 above) para 119, using the words of O’ Regan J in Bato Star Fishing 

(Pty) Ltd v Minister of Environmental Affairs 2004 4 SA 490 (CC) (Bato Star) para 

20 

21 

22 

23 

44. 

Moseneke DCJ and Madala, O’Regan and Van der Westhuizen JJ. 

Mokgoro, Nkabinde and Skweyiya JJ. 

Sidumo (n 5 above) paras 207-40. 

Sidumo (n 5 above) para 260.


expanded the three grounds of review listed there. 24 Ultimately, 

however, the facts did not establish any of those grounds of review. 25 

O’Regan J, who concurred in the judgment of Navsa AJ, wrote a 

separate judgment in which she responded to Ngcobo J and gave 

additional reasons for supporting the view that CCMA arbitrations are 

administrative action. 26 

Sachs J, too, produced a separate judgment in which he found 

himself ‘in the pleasant but awkward position of agreeing with 

colleagues who disagree with each other’. 27 He saw the judgments of 

Navsa AJ and Ngcobo J as being animated by the same goal — to 

determine what standard of conduct the Constitution expects of a 

CCMA commissioner — and as substantially in agreement on the 

interests and values involved. 28 The key to the case, as he saw it, was 

to make those interests and values explicit. This he proceeded to do, 

identifying values such as fair dealing and rationality in sections 23, 

33 and 34. His application of these standards to the commissioner’s 

conduct then led him to the same conclusion as his colleagues. 

Of the many points of interest raised by Sidumo, I wish to highlight 

three features for the attention of constitutional lawyers in general 

and administrative lawyers in particular. 

The first is the slender majority achieved on the ‘administrative 

action’ issue — an improvement on the bewildering inconclusiveness 

of New Clicks, 29 but surely further evidence of the intrinsic difficulty 

of deciding what is and what is not administrative action. That 

difficulty exists even when, as here, the Court is relying on the 

general conception of administrative action developed judicially 

under section 33 rather than the more nit-picking definition in the 

PAJA. I think the majority gets it right. Navsa AJ reasons convincingly 

on the basis of significant differences between a court of law and a 

tribunal such as the CCMA, including the relative informality of its 

processes, the absence of a system of binding precedent and the less 

secure tenure of its presiding officers, 30 and O’Regan J adds that ‘it 

is entirely consistent with our constitutional order that the 

24 Sidumo (n 5 above) paras 256-77, and for the grounds of review see n 7 above. 

25 Sidumo (n 5 above) para 289. 

26 

Sidumo (n 5 above) paras 122-41. 


28 As above. 

29 

Minister of Health v New Clicks South Africa (Pty) Ltd 2006 2 SA 311 (CC) (New 

Clicks), a case concerning pharmaceutical regulations. Two members of the Court 

found that regulation-making in general amounted to administrative action under 

the PAJA, while three others regarded the Act as applicable but on a narrower 

basis. Five members found it unnecessary to decide the question, and one found 

that subordinate legislation was generally governed by the principle of legality 

rather than the PAJA. 

30 Sidumo (n 5 above) para 85.


procedures and decisions of the CCMA should be lawful, reasonable 

and procedurally fair’. 31 But the merits are not my main concern 

here. I wish merely to note that Sidumo bears out (yet again) the 

Court’s gloomy prognostication in the SARFU case 32 about the difficult 

boundaries that would have to be drawn in the administrative action 

inquiry. 

The second feature is the further development, or complication, 

of administrative law by the official recognition of yet another 

pathway to review. The pre-democratic era knew only two such 

routes: ‘ordinary’ review of the decisions of public bodies at common 

law 33 and the special statutory review jurisdiction created 

occasionally by the legislature (of which section 145 of the LRA was a 

prime example). With this second type of review the grounds 

specified by the legislature, and indeed the remedies available, could 

be wider or narrower than those associated with ordinary review. 34 

The democratic constitution brought with it a new constitutional 

pathway to administrative-law review in the form of section 33 

(previously section 24 of the interim Constitution), duly informed by 

the common law; but this third route was reserved for exercises of 

‘administrative action’. A few years later the Constitutional Court 

identified a fourth pathway, the principle of legality, for the residual 

review of all those exercises of public power that did not qualify as 

administrative action, 35 and proceeded to develop the principle so 

that it mimics the content of section 33 to a considerable extent. 36 

Then came the PAJA, the national legislation enacted to give effect 

to section 33 and supposedly the default pathway to review 37 — but 

one that often cannot be used because the conduct being reviewed 

does not qualify as administrative action in terms of the Act’s very 

demanding definition. That makes five pathways, as we still have 

common-law review (though it is used only for exercises of private 

power these days) 38 and presumably we still have instances of special 

statutory review in the traditional sense. 


32 

President of the Republic of South Africa v South African Rugby Football Union 

2000 1 SA 1 (CC) (SARFU) para 143. 

33 Johannesburg Consolidated Investment Co Ltd v Johannesburg Town Council 1903 

TS 111 115. 

34 See Nel NO v The Master 2005 1 SA 276 (SCA) para 23 and the discussion by 

Froneman J in the court a quo: Nel NO v Master of the High Court ECD case no 

1633/2001 (undated) 8-9, unreported. 

35 Fedsure Life Assurance Ltd v Greater Johannesburg Transitional Metropolitan 

Council 1999 1 SA 374 (CC) (Fedsure) paras 56-9. 

36 

SARFU (n 32 above) and in Pharmaceutical Manufacturers Association of SA: In re 

Ex parte President of the Republic of South Africa 2000 2 SA 674 (CC) 

37 

38 

(Pharmaceutical Manufacturers Association). 

Bato Star (n 19 above) para 25. 

See C Hoexter Administrative law in South Africa (2007) 119ff.


The majority judgment in Sidumo would seem authoritatively to 

have opened up a sixth route: special statutory review under the LRA 

for labour-related action that also qualifies as administrative action 

under section 33. This, then, is not special statutory review in the 

traditional sense — review, wide or narrow, on whatever grounds the 

legislature specifies within a particular statutory regime. In this new 

type of (extra-special?) statutory review the grounds specified in the 

LRA are suffused with the content of the rights to administrative 

justice in section 33, thus as it were achieving full administrative-law 

review via labour law. Presumably this will apply to other statutory 

regimes too, such as section 131 of the Liquor Act 27 of 1989 39 and 

section 151 of the Insolvency Act 24 of 1936: whenever the conduct in 

question qualifies as administrative action under section 33, the 

grounds of review in such statutes will also be suffused with the 

content of section 33. 

If this is correct, an interesting situation will arise in relation to 

the Promotion of Access to Information Act 2 of 2000. Decisions taken 

by information officers are specially reviewable under that statute 40 

and would almost certainly qualify as administrative action for the 

purposes of section 33 — but they are expressly excluded from the 

more detailed definition of administrative action in the PAJA. 41 This 

piquant scenario is no doubt further evidence of the constitutional 

untenability of having two different and sometimes incompatible 

meanings for ‘administrative action’. 42 (It remains one of life’s 

mysteries that the constitutionality of the PAJA definition has not yet 

been challenged in court.) 

That brings me to a related point, which is that the approach of 

the Constitutional Court has a distinct practical advantage over that 

of Cameron JA in the Supreme Court of Appeal: it means not having 

to engage with the detail of the PAJA. In Rustenburg Cameron JA is 

faced with conflicting time limits within which review must be sought 

— six weeks in terms of section 145 of the LRA and six months in terms 

of the PAJA — and his choice of the shorter time limit 43 sits rather 

awkwardly with the proposition that the PAJA overrides section 145 

of the LRA. For Navsa AJ the difficulty does not arise: the PAJA and 

section 145 of the LRA exist alongside one another as separate and 

39 Cf the approach in Bulk Deals Six CC v Chairperson, Western Cape Liquor Board 

2002 2 SA 99 (C), where the court applied the grounds in the PAJA to a review 

under s 131 of the Liquor Act. 

40 Secs 78-82. 

41 

Sec 1(hh). 

42 On the various problems arising from the PAJA definition and the disparity 

between the two meanings, see C Hoexter ‘“Administrative action” in the courts’ 

2006 Acta Juridica 303 and Hoexter (n 38 above) 216ff. 

43 Rustenburg (n 6 above) para 27.


equally valid manifestations of section 33, and there can thus be no 

conflict between them. 

The third feature to note about Sidumo is the ethos or vision of 

the Constitution that underlies the judgments handed down by the 

Constitutional Court. The majority judgment of Navsa AJ, in 

particular, supports a view of constitutional rights as interconnected 

and overlapping. Nor is this view merely implicit in the judgment. 

There is strong and explicit affirmation of it when Navsa AJ is 

rejecting an argument that the rights implicated in CCMA arbitrations 

are those in sections 23 and 34, and not section 33, of the 

Constitution: 

This submission is based on the misconception that the rights in ss 23, 33 

and 34 are necessarily exclusive and have to be dealt with in sealed 

compartments. The right to fair labour practices, in the present context, 

is consonant with the right to administrative action that is lawful, 

reasonable and procedurally fair. Everyone has the right to have these 

rights enforced before the CCMA acting as an impartial tribunal. In the 

present context, these rights in part overlap and are interconnected. 44 

Although Ngcobo J may not share this view of things, 45 it is given 

particularly strong expression in the judgment of Sachs J. He sees the 

function of the commissioner as a hybrid one, ‘composed of an 

amalgam of three separate but intermingling constitutional rights’ 46 

whose underlying values ‘resist compartmentalisation’. 47 The Bill of 

Rights ‘should not always be seen as establishing independent 

normative regimes operating in isolation from each other, each with 

exclusive sway over a defined realm of public and private activity’, 48 

and indeed the relationship of the various rights should be regarded 

as ‘osmotic rather than hermetic’. 49 In this judgment ‘seepage’, 

‘permeability’, ‘interpenetration’ and ‘hybridity’ are all desirable 

qualities that help ensure the full achievement of constitutional 

justice, while an attempt to establish the primacy of one right or 

another could actually defeat the constitutional objectives to be 

realised. 50 

To say the least, the constitutional vision outlined here seems to 

be absent from the two main judgments in Chirwa. Rather, these 

judgments insist on the primacy of section 23 and the LRA in a labour 

matter, and indeed do their utmost to exclude section 33 and the 

PAJA from the picture. It is to Chirwa that I now turn. 


45 There are hints of it, however: see eg Sidumo (n 5 above) para 266. 

46 




49 

As above. 

50 Sidumo (n 5 above) paras 151-7.

2 Chirwa 


Ms Chirwa, the appellant, had been dismissed by the Transnet Pension 

Fund pursuant to a disciplinary hearing. She alleged procedural 

unfairness in the hearing and referred the dispute to the CCMA for 

conciliation. When that failed she sought administrative-law review of 

the dismissal in the Witwatersrand Local Division. There, relying on 

the common law rather than the PAJA, Brassey AJ set aside the 

decision to dismiss Ms Chirwa and ordered her reinstatement. 51 He 

followed the reasoning of the Appellate Division in Zenzile, 52 holding 

that the termination of an employment contract in the public sector 

is an exercise of public power and not (as the respondent argued) a 

matter of pure contract falling beyond the reach of administrative 

law. Zenzile-type reasoning had indeed become the dominant 

approach 53 in such cases, and here I may mention an important and 

more recent example of it: the judgment of Plasket J in POPCRU v 

Minister of Correctional Services. 54 But, as evidenced by cases such 

as South African Police Union v National Commissioner of the South 

African Police Service, 55 a contrary approach had begun to gain 

momentum — one that challenged the post-1994 status of Zenzile and 

denied administrative law a role in employment matters. 

In Transnet’s appeal against the decision of Brassey AJ two main 

questions were raised: whether this was a matter over which the High 

Court had jurisdiction, and whether the dismissal qualified as 

administrative action under the PAJA. 56 Though the Court agreed that 

the appeal had to be upheld, its members adopted three quite 

different approaches. Cameron JA, Mpati DP concurring, endorsed 

both Zenzile and POPCRU. 57 These judges held that this was a case of 

administrative action and that the High Court did have jurisdiction, 

concurrently with the Labour Court, to hear it in terms of section 157 

51 Chirwa v Transnet Ltd WLD case no 03/01052 of 4 January 2005, unreported. 

52 

Zenzile (n 1 above). 

53 Before 1994 the approach in Zenzile was applied and extended beyond dismissal 

in cases including Administrator, Natal v Sibiya 1992 2 SA 532 (A), Bula v Minister 

of Education 1992 4 SA 716 (Tk) and Hlongwa v Minister of Justice, KwaZulu- 

Natal 1993 2 SA 269 (D); and after 1994 it was applied in cases such as Mbayeka v 

MEC for Welfare, Eastern Cape 2001 1 All SA 567 (Tk) and Simela v MEC for 

Education, Eastern Cape 2001 9 BLLR 1085 (LC). 

54 2006 4 BLLR 385 (E) (POPCRU). See also Nxele v Chief Deputy Commissioner, 

Corporate Services, Department of Correctional Services 2006 10 BLLR 960 (LC) 

(Nxele). 

55 2005 26 ILJ 2403 (LC) (SAPU). Other examples of this approach are Public 

Servants’ Association obo Haschke v MEC for Agriculture 2004 25 ILJ 1750 (LC), 

Louw v SA Rail Commuter Corporation Ltd 2005 26 ILJ 1960 (W), Western Cape 

Workers Association v Minister of Labour 2006 1 BLLR 79 (LC) and Hlope v 

Minister of Safety and Security 2006 3 BLLR 297 (LC). 

56 

Transnet Ltd v Chirwa 2006 27 ILJ 2294 (SCA) (Chirwa SCA). 

57 n 54 above.


of the LRA. 58 The relevant parts of the provision read as follows: 

(1) Subject to the Constitution and section 173, and except where this 

Act provides otherwise, the Labour Court has exclusive jurisdiction in 

respect of all matters that elsewhere in terms of this Act or in terms of 

any other law are to be determined by the Labour Court. 

(2) The Labour Court has concurrent jurisdiction with the High Court in 

respect of any alleged or threatened violation of any fundamental right 

entrenched in Chapter 2 of the Constitution of the Republic of South 

Africa, 1996, arising from – 

(a) employment and from labour relations; 

(b) any dispute over the constitutionality of any executive or 

administrative act or conduct, or any threatened executive or 

administrative conduct, by the State in its capacity as an employer; 

(c) the application of any law for the administration of which the 

Minister is responsible. 

Mthiyane and Jafta JJA agreed that the High Court had jurisdiction 

concurrently with the Labour Court but held that this was not an 

instance of administrative action. These judges relied on SAPU 59 and 

on Supreme Court of Appeal authority, Cape Metropolitan Council v 

Metro Inspection Services (Western Cape) CC, 60 in holding that the 

dismissal was based on contract and so did not satisfy one of the most 

important elements of the PAJA definition: it was not an exercise of 

public power. 61 

Conradie JA was prepared to accept that the dismissal amounted 

to administrative action but held that the High Court did not have 

jurisdiction over what was quintessentially a labour matter. The 

legislature, he reasoned, had set its face against labour matters being 

litigated in other courts. 62 Thus, even if Chirwa had a cause of action 

under the PAJA she was required to pursue it under the LRA in the 

Labour Court. 63 

The two issues were aired again on further appeal to the 

Constitutional Court. 64 Skweyiya J gave judgment for the majority 

with the concurrence of seven justices. 65 Essentially he supported the 

line taken by Conradie JA, and held that the High Court did not have 

58 Chirwa SCA (n 56 above) paras 57-65. Similar reasoning was subsequently adopted 

by the Supreme Court of Appeal in Old Mutual Life Assurance Co SA Ltd v Gumbi 

2007 5 SA 552 (SCA) (Gumbi) and Boxer Superstores Mthatha v Mbenya 2007 5 SA 

450 (SCA). 

59 

n 55 above. 

60 2001 3 SA 1013 (SCA) (Cape Metropolitan). 

61 Chirwa SCA (n 56 above) para 15. 

62 

Chirwa SCA (n 56 above) para 32. 


64 Chirwa (n 3 above). 

65 

Moseneke DCJ, Navsa AJ and Madala, Ngcobo, Nkabinde, Sachs and Van der 

Westhuizen JJ.


jurisdiction. The appellant should therefore have followed to its end 

the route laid down by the LRA for resolving her dispute. Ngcobo J 

produced a separate concurring judgment which dealt in addition 

with the ‘administrative action’ issue and was supported by six 

justices. 66 Langa CJ dissented with the support of Mokgoro and 

O’Regan JJ. 

2.1 Jurisdiction 

The majority conclusion that the Labour Court had exclusive 

jurisdiction in the matter was evidently inspired by considerations of 

policy, and by the legislature’s failure to heed various calls for the 

amendment of section 157. Both Skweyiya and Ngcobo JJ reasoned 

that employees ought not to be in a preferential position merely by 

virtue of their employment in the public sector. Public and private 

employees ought rather to be treated equally, and parity was indeed 

one of the main aims of the current labour regime. 67 That regime 

included a ‘one-stop shop’, 68 a specialised and purpose-built system 

for resolving labour disputes, which should be pursued to its end. 69 

Furthermore, to allow the High Court concurrent jurisdiction with the 

Labour Court in employment matters would be to encourage forumshopping 

and the development of a dual 70 and possibly less coherent 71 

system of law. In the light of these considerations, Ncgobo J thought 

the use of the word concurrent in section 157(2) was ‘unfortunate’. 72 

The problem, however, is that policy arguments cannot alter the 

express terms of section 157 of the LRA — however unfortunate those 

terms may seem. Leaving aside for a moment the cogency of the 

arguments, the majority’s interpretation of the provision is in conflict 

not only with the wording of section 157 but also with the Court’s own 

previous interpretation of that wording in Fredericks & Others v MEC 

for Education and Training, Eastern Cape, & Others. 73 

In Fredericks, teachers who had been refused voluntary 

retrenchment sought review of this action in a High Court, alleging a 

violation of their constitutional rights to equality and just 

administrative action. White J found that his Court had no jurisdiction 

66 

Moseneke DCJ, Navsa AJ and Madala, Nkabinde, Sachs and Van der Westhuizen JJ. 

67 Chirwa (n 3 above) para 66 (Skweyiya J) and paras 99-102 and 149 (Ncgobo J). 

68 Chirwa (n 3 above) para 47 (Skweyiya J). 

69 

Chirwa (n 3 above) paras 65-7 (Skweyiya J), and see paras 105, 110-13 and 117-20 

(Ngcobo J). 

70 Chirwa (n 3 above) paras 65 (Skweyiya J) and 121, 124 (Ngcobo J). 

71 

Chirwa (n 3 above) para 118 (Ngcobo J). 

72 Chirwa (n 3 above) para 121. 

73 2002 2 SA 693 (CC) (Fredericks). See also Fedlife Assurance Ltd v Wolfaardt 2002 

1 SA 49 (SCA); United Public Servants Association of South Africa v Digomo NO 

2005 26 ILJ 1957 (SCA).


in the matter. He saw it in essence as a labour case that had to be 

dealt with under section 24 of the LRA, a provision that envisages the 

involvement of the CCMA. On appeal, however, a unanimous 

Constitutional Court corrected this misapprehension. O’Regan J 

pointed out that section 157(2) is the sole source of the Labour 

Court’s jurisdiction to determine disputes arising from the alleged 

infringement of constitutional rights, and that it clearly gives the 

Labour Court and the High Court concurrent jurisdiction over such 

matters. 74 She emphasised that the LRA gives the Labour Court no 

general jurisdiction over labour matters, so that the High Court’s 

jurisdiction is not ousted by the mere fact that something happens to 

be a labour matter. 75 Nor could section 158(2)(h) be read with section 

157(1) of the LRA to oust the High Court’s jurisdiction, since section 

158(1)(h) ‘does not expressly confer on the Labour Court 

constitutional jurisdiction to determine disputes arising out of alleged 

infringements of the Constitution by the state acting as an 

employer’. 76 O’Regan J also made it clear that in terms of section 169 

of the Constitution, such ousting of the High Court’s jurisdiction is 

permissible only where the matter is assigned to a court of similar 

status — which the CCMA clearly is not. 77 

In Chirwa, notwithstanding the way the matter was framed and 

the many detailed references to the PAJA in the papers, Skweyiya J 

chose to fix on the appellant’s brief reference to provisions in the LRA 

(evidently included for the purpose of establishing grounds of review 

in the PAJA). 78 He held that this was a dispute envisaged by section 

191 of the LRA, which provides a procedure for the resolution of 

disputes about unfair dismissals. It was thus a matter that ‘must, 

under the LRA, be determined exclusively by the Labour Court’. 79 

While Fredericks clearly stood in the way of such an interpretation of 

section 157, Skweyiya J distinguished it on the basis that there had 

been no reliance in the earlier case on section 23 or on the LRA. 80 

That, with respect, seems an unconvincing point of distinction. 

After all, Fredericks was so obviously a ‘labour’ matter that the High 

Court believed it lacked jurisdiction to hear it — despite the 

applicants’ own characterisation of the case as one relating to 

sections 9 and 33 of the Bill of Rights. Conversely, what about the 

inconvenient fact that the appellant in Chirwa evidently abandoned 

74 

75 

76 

Fredericks (n 73 above) para 41. 

Fredericks (n 73 above) paras 38 and 40. 

Fredericks (n 73 above) para 43. Section 158(1)(h) allows the Labour Court to 

review ‘any decision taken or any act performed by the State in its capacity as 

77 

78 

79 

80 

employer on such grounds as are permissible in law’. 

Fredericks (n 73 above) paras 29-31. 

Chirwa (n 3 above) para 157 in the judgment of Langa CJ. 

Chirwa (n 3 above) para 63. 

Chirwa (n 3 above) para 58.


her initial reliance on section 23 and the LRA 81 and deliberately 

framed her case as one in administrative law? 82 

As Nugent JA has remarked, it is difficult to find a ‘clear legal — 

as opposed to policy — reason for the outcome in Chirwa’. 83 The 

majority was apparently bent on characterising the case purely as a 

labour matter whatever the cost. It thus depicted the appellant’s High 

Court application as an instance of asking an ordinary court to decide 

a specialist dispute. 84 Langa CJ bluntly described this manoeuvre as a 

mischaracterisation, 85 for in truth Ms Chirwa was ‘asking a High Court 

to consider an administrative law issue’. 86 Langa CJ pointed out that 

in most cases unfair dismissal claims are decided not by the Labour 

Court but by the CCMA, so that in terms of section 169 of the 

Constitution exclusive jurisdiction could not lawfully be conferred on 

the Labour Court in any event. 87 The Chief Justice also indicated the 

irrelevance of the various policy considerations to the issue of 

jurisdiction, which in this instance was a matter of legislative choice. 

Ultimately, as he said, ‘[w]hile we may question that intention and 

may have preferred a legislative scheme that more neatly divided 

responsibilities between the various courts, that is not the path the 

legislature has chosen.’ 88 

Skweyiya J seems indeed to have been aware of this, and not 

entirely convinced by his own strained interpretation of section 157. 

Otherwise it is hard to see why he ends his judgment by underscoring 

the ‘urgent need for the legislature to revisit the provisions of s 157(2) 

of the LRA’ 89 — for his account of the provision suggests no such need. 

At any rate, wishful thinking cannot change the fact that in 

practice labour law and administrative law are not neatly divided, and 

that in many cases they do overlap. I doubt that neatness will ever be 

achieved at this intersection, legislatively or otherwise, for it seems 

to me that some untidiness is both natural and unavoidable. Nor, I 

think, should neatness be regarded as a particularly worthy aim in this 

81 Chirwa (n 3 above); see para 67 in the judgment of Skweyiya J and paras 84-5 in 

82 

83 

the judgment of Ngcobo J. 

Chirwa (n 3 above); see paras 157-9 in the judgment of Langa CJ. 

In a separate concurring judgment in Makambi v MEC for Education, Eastern Cape 

84 

85 

86 

2008 5 SA 449 (SCA) (Makambi) para 21. 

Chirwa (n 3 above) para 61 (Skweyiya J) and paras 124-5 (Ngcobo J). 


Chirwa (n 3 above) para 173. Not only was the claim couched in the language of 

administrative law, but it was ‘based squarely on the PAJA’. Chirwa (n 3 above) 

87 

para 157. 

Chirwa (n 3 above) para 170. Furthermore, s 157(5) of the LRA states: ‘Except as 

provided for in section 158(2), the Labour Court does not have jurisdiction to 

adjudicate an unresolved dispute if this Act requires the dispute to be resolved 

88 

89 

through arbitration.’ 


Chirwa (n 3 above) para 71.


area of the law. Both of these propositions find some support in the 

facts and reasoning applied in a post-Chirwa case, Nakin v MEC, 

Department of Education, Eastern Cape & Another. 90 

The applicant in Nakin was a school principal who had been 

transferred to a lower-level post through no fault of his own. This 

injustice was officially rectified several years later when the 

department agreed to reinstate him and to correct the accumulated 

shortfall in salary and benefits. The department neglected to effect 

the correction, however, and the applicant approached a High Court 

for an order compelling it to do so in terms of the PAJA. In argument 

the respondents challenged the Court’s jurisdiction to deal with what 

they saw as an unfair labour practice within the meaning of section 

186(2) of the LRA. But Froneman J upheld the High Court’s jurisdiction 

to determine the case as one of unlawful administrative action. 91 

Fredericks, which had not been overruled in Chirwa, was still good 

law, all the more so in a matter not involving dismissal. 92 

The facts of Nakin nicely illustrate the natural overlap just 

referred to. More importantly, however, Froneman J challenges the 

policy behind the tidy-minded majority view in Chirwa. In relation to 

the concerns about forum-shopping and the growth of parallel systems 

of law, he points out that the coherence of labour jurisprudence 

depends on its giving proper expression to section 23 of the 

Constitution rather than its development in a single forum. 93 

Froneman J goes on to show that in practice, labour law has gained 

and not lost from administrative-law insights — an example being the 

recent development of the common-law contract of employment so 

as to include a right to a pre-dismissal hearing. 94 His judgment also 

suggests that the policy of equality may be a misdirected one. 

Froneman J points out that in public employment fairness ‘may 

conceivably have a different content to that in the private sector, for 

reasons relating to constitutional demands of responsiveness, public 

accountability, democracy and efficiency in the public service’. 95 

This is a point of fundamental importance. Public and private 

employers are not the same, for under our Constitution they are not 

under the same public duties. To put it simply, employers in the 

private sector have the luxury (to the extent that they are not 

90 

91 

92 

2008 6 SA 320 (Ck) (Nakin). 

Nakin (n 90 above) para 39. 

Nakin (n 90 above); see paras 27-39. Cf Makambi (n 83 above) para 17, where 

Farlam JA found it was ‘not possible to hold that this case falls on the Fredericks 

93 

94 

side of the line of distinction drawn in the Chirwa case’. 


Nakin (n 90 above) para 30, with reference to Gumbi (n 58 above); and see also 

95 

paras 31 et seq. 

Nakin (n 90 above) para 34.


thwarted by regulation) of acting out of pure self-interest. Organs of 

state do not have the same freedom, for they are generally 

constrained by their duty to act in the public interest — even when 

there is no legislation explicitly saying so. 96 Thus, while the aim of 

achieving equality between all employees certainly appeals to one’s 

sense of justice, it may not be an attainable ideal under our present 

constitutional dispensation; not, at any rate, if one conceives of that 

equality as a matter of paring down the rights of public-sector 

employees to match those of other employees. 

That does not, however, rule out the possibility of equalising up 

rather than down by allowing the insights of administrative law (or 

some of them) 97 to benefit employees more generally. To borrow the 

words of Froneman J, ‘the substantive coherence and development of 

employment law can only gain’ from administrative law. 98 While such 

reform would optimally be achieved by methodical legislative means, 

one must not underestimate the potential of section 39(2) of the 

Constitution. As I have suggested elsewhere, section 39(2) may be 

regarded as constitutional authority for the courts to set public-law 

standards for ‘private administrators’ to follow, especially in 

relationships characterised by inequality (as the employment 

relationship tends to be) or where coercive power is being 

exercised. 99 

2.2 Administrative action 

Though the Court in Chirwa agreed that the dismissal was not an 

instance of administrative action, its members displayed three 

different approaches to the issue. As I suggest in what follows, none 

of these approaches is unproblematic. 

Skweyiya J found that it was unnecessary to decide the 

administrative action question since the case could be resolved on the 

jurisdictional issue (though he indicated his agreement, obiter, with 

the judgment of Ngcobo J on the administrative action diagnosis). 100 

96 

97 

98 

This most basic principle of administrative law was expressed by Schreiner JA in 

his dissenting judgment in Mustapha v Receiver of Revenue, Lichtenburg 1958 3 

SA 343 (A) at 347E-G — a judgment vindicated by the Supreme Court of Appeal in 

Logbro Properties CC v Bedderson NO 2003 1 SA 460 (SCA) (Logbro) para 13. It is 

in fact unusual for enabling legislation to specify that action has to be taken ‘in 

the public interest’, but an example appears in Nxele supra note 54 para 59. 

The requirements implied by ‘reasonable’ and ‘procedurally fair’ administrative 

action could certainly be extended in this way, but imposing the various 

requirements of ‘lawfulness’ on private employers would make less sense. 


99 Hoexter (n 38 above) 124. Unlike the direct horizontal application of rights under 

s 8(2) of the Constitution, indirect application under s 39(2) does not depend on 

the ‘nature’ of the right sought to be applied. 

100 Chirwa (n 3 above) para 73.


This approach is clearly mistaken, however. As the Chief Justice 

pointed out, the administrative action inquiry could not legitimately 

be avoided: it was indeed the ‘primary question’ 101 in the case, for 

the Court’s jurisdiction depended on whether the dismissal was ‘an 

administrative act or conduct ... by the State in its capacity as an 

employer’ for the purposes of section 157 of the LRA. 

Ngcobo J did not seek to avoid the inquiry. Relying on the section 

33 meaning of administrative action, he found (rightly, I believe) that 

the dismissal clearly amounted to an exercise of public power. 102 

However, it lacked one of the hallmarks of administrative action 

listed in the SARFU case: 103 it did not involve the implementation of 

legislation but had rather been effected in terms of the contract 

between the parties. 104 Ngcobo J thus treated a single factor as 

decisive in a manner arguably not contemplated by the Court in 

SARFU. But the real issue for him seems to have been that the 

dismissal was ‘more concerned with labour and employment 

relations’ than with administration, 105 and he went on to emphasise 

the formal division in the Constitution between labour relations and 

administrative conduct. 106 

By contrast, Langa CJ relied on the definition of administrative 

action in the PAJA, which covers decisions made by an organ of state 

when ‘exercising a public power or performing a public function in 

terms of any legislation’. 107 He, too, held that in the absence of any 

particular statutory authority the dismissal had taken place in terms 

of the contract itself; 108 but unlike Ngcobo J, he went on to find that 

the dismissal was not an exercise of public power. In this regard Langa 

CJ noted the absence of features that tend to point to the existence 

of a public power: any imbalance in the employment relationship was 

not attributable to the respondent’s status as a public body; the 

dismissal had little or no impact on the public; and the source of the 

power to dismiss was contractual. 109 Furthermore, he could see none 

of the ‘strengthening factors’ that had been present in the POPCRU 

case. 110 The Transnet Pension Fund did not have the same obviously 

public character as the Department of Correctional Services, the 

respondent in POPCRU; unlike that department it pursued no 

obviously public goals; and, unlike in POPCRU, the public interest in 


102 Chirwa (n 3 above) para 138, quoting with approval from the judgment of 

Cameron JA in the court below. 

103 n 32, para 143. 


105 As above. 

106 Chirwa (n 3 above) paras 143 et seq. 

107 Section 1 of the PAJA, my emphasis. 


109 Chirwa (n 3 above) paras 186-94. 

110 Chirwa (n 3 above) para 192, referring to POPCRU (n 54 above) para 54.


the administration of the Transnet Pension Fund was not preeminent. 

111 

The different approaches of Ngcobo J and Langa CJ illustrate, 

once again, the problem of having two meanings for ‘administrative 

action’. Given that reality, however, Langa CJ is surely correct to rely 

on the PAJA definition. This approach accords with the constitutional 

principle that general norms should be resorted to only when norms 

of greater specificity have run out, 112 or when testing the 

constitutionality of a specific norm (the PAJA or other original 

legislation) against the more general norm (section 33). 113 

Ngcobo J, who has preferred the section 33 meaning in other cases 

too, 114 does so on the basis that ‘PAJA only comes into the picture 

once it is determined that the conduct in question constitutes 

administrative action within the meaning of section 33.’ 115 This 

approach is rather cumbersome as it calls for two separate inquiries: 

Is the conduct administrative action for the purposes of s 33? If so, 

does the PAJA definition nevertheless exclude it? (Stopping at the first 

inquiry would of course make the PAJA definition completely 

redundant.) 116 More importantly, this approach seems to be 

inconsistent with the principle mentioned above and with the 

Constitutional Court’s jurisprudence on the subject of constitutionally 

mandated legislation. As the statute giving effect to the rights in 

section 33, the PAJA should be resorted to directly — a point that has 

indeed been stressed by Ngcobo J himself. 117 Even if it does not have 

the effect of bypassing the PAJA altogether, the mediation through 

section 33 he insists on seems quite unwarranted. It means, too, that 

in this instance we are deprived of his views on the application of the 

various elements of the PAJA definition. 

As to the merits of the ‘administrative action’ diagnosis in Chirwa, 

it depended largely on a curiosity. Like appointment, dismissal in the 

public sector is almost always governed by statute. It certainly was in 

Zenzile, where ‘the element of public service injected by statute’ 118 

played a considerable part in the reasoning of Hoexter JA. A quirk of 

111 As above. 

112 See I Currie & J de Waal The bill of rights handbook (5 ed 2005) 13. 

113 See Hoexter (n 38 above) 115. 

114 See eg New Clicks (n 29 above) para 446. 


116 Construing the PAJA consistently with s 33 is constitutionally appropriate (as 

indicated by Nugent JA in Grey’s Marine Hout Bay (Pty) Ltd v Minister of Public 

Works 2005 6 SA 313 (SCA) para 22), but it is quite another thing to behave as if 

the statute and its detailed definition of administrative action did not exist. See 

further Hoexter (n 42 above) and I Currie ‘What difference does the Promotion of 

Administrative Justice Act make to administrative law?’ 2006 Acta Juridica 325. 

117 See eg New Clicks (n 29 above) para 436, and in relation to s 23 of the 

Constitution see eg Sidumo (n 5 above) paras 248-9. 

118 n 1 above, 34C.


the Chirwa case was the absence of governing legislation after the 

repeal of the South African Transport Services Conditions of Service 

Act 41 of 1988. Technically, then, the dismissal really did seem to be 

a purely contractual matter in this instance — if not in others. 119 

But the judgment of Langa CJ (and that of Mthiyane JA before 

him) is nevertheless open to criticism on the contractual issue. As 

Stacey has pointed out in a perceptive article, 120 it is difficult to 

square with the approach of the Constitutional Court in AAA 

Investments (Pty) Ltd v Micro Finance Regulatory Council. 121 There 

the Court firmly rejected the idea that a decision loses its public 

character simply because the most immediate source of the power 

happens to be a contract. In that case the contractual element led the 

Supreme Court of Appeal to regard the Micro Finance Regulatory 

Council as ‘a private regulator of lenders who choose to submit to its 

authority by agreement’ 122 — and was rebuked by the Constitutional 

Court for putting form above substance. 123 Is the Chief Justice not 

open to the same charge here? 

In Transnet Ltd v Chirwa, it is interesting to note, the absence of 

governing legislation does not trouble Cameron JA at all. As he sees it, 

Transnet is a public entity created by legislation and operating under 

statutory authority. It would not exist without statute. Its every act 

derives from its public, statutory character, including the dismissal at 

issue here. 124 

On this approach public entities can never be on exactly the same 

footing as private ones, contract or no contract; and even when there 

is a contract, the principles of administrative justice (in one form or 

another) still frame the contractual relationship. 125 In the 

Constitutional Court Ngcobo J supports this line of reasoning: 126 

In my view, what makes the power in question a public power is the fact 

that it has been vested in a public functionary, who is required to 

exercise the power in the public interest. When a public official 

performs a function in relation to his or her duties, the public official 

119 Cf for instance SAPU (n 55 above), where the ‘purely contractual’ label is far less 

convincing, and see further C Hoexter ‘Contracts in administrative law: Life after 

formalism?’ (2004) 121 South African Law Journal 595. 

120 R Stacey ‘Administrative law in public-sector employment relationships’ (2008) 

125 South African Law Journal 307 315. 

121 2006 11 BCLR 1255 (CC) (AAA Investments). 

122 Micro Finance Regulatory Council v AAA Investments (Pty) Ltd 2006 2 SA 27 (SCA) 

para 24. 

123 AAA Investments (n 121 above) para 45. 


125 Cameron JA for a unanimous court in Logbro (n 96 above) para 8, writing an 

important gloss on the Cape Metropolitan case (n 60 above) 



exercises public power. I agree with Cameron JA that Transnet is a 

creature of statute. It is a public entity created by statute and it 

operates under statutory authority. As a public authority, its decision to 

dismiss necessarily involves the exercise of public power, and ‘(t)hat 

power is always sourced in statutory provision, whether general or 

specific, and, behind it, in the Constitution’. 

By contrast, Langa CJ finds that the dismissal is not an exercise of 

public power. In arriving at this conclusion he relies on a number of 

factors that have been used by the courts for determining whether a 

power or function is public. However, his paradigm seems to be quite 

different from that of Cameron JA and Ngcobo J, for in his analysis the 

public and intrinsically statutory nature of the entity apparently 

carries no independent weight. The factors seem to be applied just as 

they would be to a completely private entity such as a club or a 

business. Because of this, and notwithstanding the detailed and 

nuanced reasoning of the Chief Justice and the cumulative force of 

the various features listed by him, one is left with questions. 

Here are some of them. Given that Transnet would not exist 

without statute, does it really matter that a contract happened to 

intervene in this instance? (Langa CJ is clear that where it is achieved 

‘in terms of a specific legislative provision’ a dismissal may amount to 

administrative action.) 127 Given the undoubted public status of 

Transnet, does it really matter that the Transnet Pension Fund is not 

as distinctively public as the Department of Correctional Services was 

in POPCRU, or that it does not pursue its public goals in so obvious a 

manner? Given that Transnet is generally supposed to serve the public 

interest, does it really matter that this particular dismissal was of 

scant interest to the public or that it had no real impact on the public? 

(As Plasket J pointed out in POPCRU, very many actions acknowledged 

to be administrative have little or no such impact.) 128 And is it 

realistic to say — is it ever entirely realistic to say — that no power 

imbalance derived from the employer’s public status? 

Similarly, there is something troubling about the Chief Justice’s 

statement that ‘[w]hilst Transnet is certainly subservient to the 

Constitution, so are all business entities in South Africa’. 129 This fails 

to acknowledge that as an organ of state Transnet is under special 

duties (such as the obligation in section 7(2) of the Constitution) that 

simply do not apply to private employers. 

In short, I wonder whether an entity like Transnet is ever entirely 

free to do as it pleases, and whether it is capable of shedding its 





public, statutory nature so conveniently and completely. I doubt that 

public entities are capable of acting ‘simply in [their] capacity as 

employer’, as Mthiyane JA would have it; 130 and with respect, nothing 

said in Chirwa eliminates that doubt. 

2.3 The interconnectedness of rights 

In Sidumo, 131 as we have seen, the majority fully acknowledges the 

interconnectedness of constitutional rights and their natural 

tendency to overlap. This view of things is also clearly apparent in the 

dissenting judgment in Chirwa, in which Langa CJ more than once 

recognises the overlap of the LRA and the PAJA and their parent 

rights. 132 But there is little hint of this understanding in the majority 

judgments in Chirwa. Instead, emphasis is laid on the entrenchment 

of the respective rights in two separate provisions in the Constitution, 

‘each with its own aims and specialised legislation ... that seeks to 

give effect to its own distinct objectives’. 133 While Skweyiya J 

concedes that the dismissals of public sector employees ‘appear to 

implicate’ not only labour rights but also those of administrative 

justice (or that this ‘is at least what Ms Chirwa is asserting’), 134 and 

though he briefly admits a possible overlap of the LRA and the 

PAJA, 135 he is clear about the ‘pre-eminence of the LRA’ in labour 

disputes, and about the absence of any legislative intention that the 

PAJA should detract from it. 136 In short, he refuses to see Chirwa as 

a case involving administrative law at all. Ngcobo J achieves this view 

of the case too, and he does so by relying heavily on the Constitution’s 

‘clear distinction between administrative action on the one hand and 

employment and labour relations on the other’. 137 

This categorical view of things is a pity — not just because the 

vision of the majority is so completely at variance with the ethos of 

Sidumo, but also because it ignores an important constitutional 

principle. In POPCRU Plasket J warns against formalistic attempts to 

pigeonhole administrative law and labour law, which are after all 

mere classifications of convenience. 138 Under our Constitution, he 

observes, 

[t]here is nothing incongruous about individuals having more legal 

protection rather than less, or of more than one fundamental right 



132 Chirwa (n 3 above) paras 167 and 176. 

133 Chirwa (n 3 above) para 46 (Skweyiya J). 




137 Chirwa (n 3 above) paras 143-4. 

138 n 54 above, para 61.


applying to one act, or of more than one branch of law applying to the 

same set of facts. 139 

In Transnet Ltd v Chirwa Cameron JA finds in the Constitution ‘no 

suggestion that, where more than one right may be in issue, its 

beneficiaries should be confined to a single legislatively created 

scheme’. 140 In Chirwa the Chief Justice expresses the same principle 

in the following way: 141 

Both PAJA and the LRA protect important constitutional rights and we 

should not presume that one should be protected before another or 

presume to determine that the ‘essence’ of a claim engages one right 

more than another. A litigant is entitled to the full protection of both 

rights, even when they seem to cover the same ground. 

It is unfortunate that the majority is impervious to this wisdom. 

3 Masetlha 

Masetlha, 142 chronologically the first case in the series, arose out of 

an atypical case of ‘executive’ dismissal featuring the President of 

the Republic and the head of the National Intelligence Agency (NIA). 

The ‘one-stop shop’ of the LRA was not open to the appellant in this 

case, and nor was the PAJA of application; but it is nevertheless a 

matter that touches both labour law and administrative law more 

broadly. For administrative lawyers its main interest lies in what the 

majority judgment says, or fails to say, about procedural fairness. 

The facts are widely known. Briefly, in December 2004 the 

President appointed the appellant, Mr Billy Masetlha, head and 

Director-General of the NIA for a period of three years. In October 

2005 Masetlha was suspended from this position as a result of the 

‘Macozoma affair’, wherein a prominent businessman was placed 

under surveillance and subsequently lodged a complaint about it with 

the Minister. The suspension followed an investigation by the 

Inspector-General of Intelligence, and was challenged by Masetlha in 

an application for setting aside. In March 2006, before that had been 

decided, the President unilaterally amended Masetlha’s term of office 

so that it expired two days later — thus effectively dismissing him 21 

months before his term was due to end. The President’s justification 

was that the relationship of trust between him and the NIA head had 

139 POPCRU (n 54 above) para 60. See further Stacey (n 120 above) 323ff, where he 

demonstrates the potentially absurd consequences of the proposition that the 

remedies offered by one branch of the law oust those of another. 



142 n 4 above.


broken down irreparably. Masetlha was offered his full salary and 

benefits for the unexpired period. He declined the offer and sought 

review of the President’s decisions and reinstatement in his position. 

Du Plessis J held for the Pretoria High Court that the dispute about 

the suspension had been rendered moot by the dismissal, and that the 

dismissal itself amounted to lawful executive action. 143 This decision 

was upheld on appeal to the Constitutional Court (which did not doubt 

that the President’s action amounted to a dismissal). 144 Moseneke 

DCJ gave judgment for the majority with the concurrence of six 

colleagues; 145 Ngcobo J dissented with the support of Madala J; and 

Sachs J produced a separate judgment. 

The first main issue related to lawfulness. Neither the 

Constitution nor the relevant legislation mentioned dismissal of the 

head of the NIA but, notwithstanding some disagreement about where 

the authority resided, all the justices agreed that the President had 

the implied authority to do what he did. 146 More interesting was the 

second issue: what constraints there were on the exercise of the 

power to dismiss, and whether they included a requirement of 

procedural fairness — for the President had acted without giving the 

appellant notice or a hearing on the issue of his dismissal. 147 

The usual constraints were soon ruled out. They clearly did not 

include the LRA, since it does not apply to members of the NIA. 148 The 

majority did not pause to consider whether the appellant still had the 

benefit of the section 23 right to fair labour practices — which was the 

solution favoured by Sachs J in his separate judgment — and so there 

was no hint here of the Chirwa insistence on the primacy of labour law 

in a case of dismissal. In fact administrative law ultimately came to 

the rescue, though not in the form of the PAJA. The latter did not 

apply either: the dismissal was an exercise of executive powers that 

are specifically excluded from the definition of administrative action 

in the PAJA (and would not count as administrative action for the 

purposes of section 33). 149 As for the common-law right to a predismissal 

hearing recognised by the Appellate Division in Zenzile, 150 

the majority reasoned that this case was distinguishable from the 

143 See Masetlha (n 4 above) paras 22-4 in the judgment of Moseneke DCJ. 

144 Masetlha (n 4 above) para 52. 

145 Langa CJ, Navsa AJ and Nkabinde, O’Regan, Skweyiya and Van der Westhuizen JJ. 

146 The majority found that the necessary power lay in s 209(2) of the Constitution 

and s 3(3) of the Intelligence Services Act 65 of 2002 (see Masetlha n 4 above, 

para 73), whereas Ngcobo J found that provisions of the Public Service Act 103 of 

1994 were equally relevant (Masetlha n 4 above, para 157). 

147 Masetlha (n 4 above) para 84, and see para 197 in the judgment of Ngcobo J. 

148 See s 2(b) of the LRA. 

149 Masetlha (n 4 above) para 76n39, though Moseneke DCJ does not deal fully with 

the s 33 meaning. 

150 n 1 above.


earlier one: it concerned a ‘special legal relationship’ between the 

President and the head of the NIA, and the dismissal involved the 

exercise of an executive power deriving from the Constitution and 

national legislation. 151 In the view of the majority the dismissal 

attracted only the grounds of review implied by that residual pathway 

to administrative-law review, the principle of legality. 152 

As already noted, the content of the principle has been developed 

considerably over the years: it has been held to imply that those who 

wield public power must act lawfully 153 and in good faith, 154 must not 

misconstrue their power, 155 and must act rationally in relation to the 

purpose for which the power was given. 156 These requirements 

effectively cover many of the grounds of review in ‘regular’ 

administrative law (the PAJA), and they explain why the rule of law 

in general, and the principle of legality in particular, have been 

described as ‘administrative law applied under another name’. 157 But 

while procedural fairness may be a standard component of the rule of 

law, 158 the narrower principle of legality has not yet been held to 

require such fairness. In Masetlha the majority showed no inclination 

to change the position. Moseneke DCJ observed that powers to 

appoint and dismiss this type of official were conferred specially on 

the President for the effective business of government and the pursuit 

of national security, and he concluded that ‘it would not be 

appropriate to constrain executive powers to requirements of 

procedural fairness’. 159 

When one considers how well established the audi alteram 

partem principle is in our law generally, and how well established it 

is in the context of dismissal (at least in administrative law), 160 the 

majority’s conclusion seems sadly retrogressive. First, in what way 

would national security or effective government be jeopardised by 

requiring the President to hear the appellant? Secondly, as Ngcobo J 

indicates in his dissenting judgment, it seems inconceivable that the 

content of the rule of law should be less today than it was under the 


152 Masetlha (n 4 above) para 81, where it is referred to as ‘the principle of legality 

and rationality’. 

153 Fedsure (n 35 above) paras 56-9. 

154 The SARFU case (n 32 above) para 148. 

155 As above. 

156 Pharmaceutical Manufacturers Association (n 36 above) para 85. 

157 C Plasket ‘The fundamental right to just administrative action: Judicial review of 

administrative action in the democratic South Africa’ unpublished doctoral thesis, 

Rhodes University, 2002 164. 

158 See C Hoexter ‘The principle of legality in South African administrative law’ 

(2004) 4 Macquarie Law Journal 165. 

159 Masetlha (n 4 above) para 77; and see also para 78. 

160 See the cases cited in n 53 above.


old regime. 161 In Zenzile, a judgment handed down during the 

apartheid era, the lowliest employees, temporary hospital workers, 

were held to have the right to be heard before being dismissed. Now, 

in the democratic era — under a constitution committed to 

accountability, responsiveness and openness 162 — a member of the 

public service 163 is held to have no entitlement at all to procedural 

fairness. The majority apparently made nothing of the differences 

between this appointment and ‘purely political appointees’ such as 

Cabinet Ministers; 164 though I remain unconvinced that under our 

Constitution it is justifiable to dismiss anyone, however political their 

appointment, without the benefit of a hearing. 

In his dissenting judgment Ngcobo J held that the rule of law 

required the appellant to be heard before being dismissed. He linked 

this procedural requirement to the existing substantive requirement 

of rationality, pointing out that fairness provides insurance against 

irrationality or arbitrariness — it ensures that the decision-maker has 

all the facts before making a decision. 165 Ngcobo J also reminded us 

of the importance of the audi principle by quoting from the Court’s 

own judgment in Zondi v MEC for Traditional and Local Government 

Affairs: 166 

A hearing can convert a case that was considered to be open and shut to 

be open to some doubt, and a case that was considered to be 

inexplicable to be fully explained. 

It was not necessary for Ngcobo J to point out that if the President had 

acted fairly before dismissing the appellant, misunderstandings 

between them might conceivably have been cleared up. 167 

Judicial disagreement about the content of the rule of law is not 

all there was to the case, however. Moseneke DCJ went on to say that 

if procedural fairness were indeed a requirement, it had been 

satisfied in this instance. The appellant had had ‘ample occasion’ 168 

to respond to the allegations made against him in relation to the 

Macozoma affair: he had had at least two meetings with the Minister 

at which he was called upon to explain the surveillance and his role in 


162 Section 1 of the Constitution. 

163 See Masetlha (n 4 above) para 37. 

164 Masetlha (n 4 above) paras 288-9, where Sachs J records the differences: 

Ministers ‘know they are hired and can be fired at the will of the President; and if 

fired, they can mobilise politically’, whereas the head of the NIA has ‘one foot in 

government and one in the public administration’ and has no equivalent political 

remedies. 

165 Masetlha (n 4 above) paras 184-7. 

166 2005 3 SA 589 (CC) para 112, quoted in para 204 of his dissenting judgment. 

167 Whether they would have been is of course immaterial: see Masetlha (n 4 above) 

para 204. 

168 Masetlha (n 4 above) para 83.


it, and he had made submissions to the Inspector-General’s 

investigation. There had also been a meeting with the President at 

which the appellant had expressed his dissatisfaction with the 

Inspector-General’s findings and the recommendation that 

disciplinary action should be taken against him. So, 

[a]lthough the President did not ask the applicant for his views at the 

point of dismissing him, he had the benefit of the views of the applicant 

on all material issues that led to the dismissal. 169 

With respect, this reasoning gives cause for concern. To say that there 

is no right to procedural fairness at all in this unique executive setting 

is one thing. It is quite another to say that if there was such a right, 

it was satisfied in the circumstances. The potential effect is to dilute 

the content of the audi alteram partem principle in relation to 

dismissal — which is just where it deserves to be applied rigorously. 

As Ngcobo J points out in his dissenting judgment, fairness surely 

demands that any person being dismissed be heard on the proposed 

dismissal itself. 170 It is certainly not enough, at least not in 

administrative law, for one party to have had the benefit of the other 

party’s views in general. The essence of a fair hearing is that the 

person adversely affected has a chance to comment specifically on 

the proposed adverse action. 171 It is a pity, then, that the majority 

did not make it clearer that this was an exception to the general 

principles of procedural fairness, and it is to be hoped that its 

proposition does not infect those general principles. 

The judgment of Sachs J, too, is worrying in this regard. While he 

found that the appellant was entitled to be treated fairly under 

section 23 of the Constitution and took a broad view of what 

substantive fairness demanded, Sachs J seemed to take a strangely 

constrained view of what procedural fairness required in the 

circumstances. In his judgment the President ought to have consulted 

the appellant on the manner in which the termination was to be 

publicly communicated; 172 but, in the absence of a charge of 

misconduct or other form of breach, Sachs J apparently saw no need 

for a hearing on the dismissal itself. 173 This seems a somewhat 

formalistic approach, too, since misconduct was surely implied: the 

Court apparently accepted that the breakdown of trust between the 

parties was a direct result of the appellant’s conduct in the Macozoma 

affair — for which the Inspector-General had recommended 

disciplinary action against him. 



171 As to the common law, see L Baxter Administrative law (1984) 546; and more 

generally, see Hoexter (n 38 above) 332ff. 


173 Masetlha (n 4 above) para 234.


4 Conclusion 

The cases of Masetlha, Sidumo and Chirwa presented the 

Constitutional Court with opportunities to pronounce on the 

intersection of labour law and administrative law in employment 

matters, and to expound on the relationship between the 

constitutional rights concerned and their associated pieces of 

legislation. As I have tried to show in this article, the judgments of the 

Constitutional Court are full of interest. Sidumo in particular develops 

the connection between the relevant rights in a creative manner, 

while Chirwa raises fundamental questions about the concept of 

administrative action and the nature of public power. 

But the judgments of the Constitutional Court in these cases also 

leave one with a sense of wasted opportunity. Masetlha seems a 

retrogressive decision, at least to an administrative lawyer, for 

allowing an unfortunate exception to the established principles of 

procedural fairness. And Sidumo and Chirwa appear to be far apart in 

the constitutional vision or ethos expressed by the majority in each 

case, for the former encourages the interconnectedness of rights 

while the latter disavows it. As I have argued above, Chirwa is 

unsatisfying in other respects as well: none of the Court’s approaches 

to the ‘administrative action’ issue is unproblematic, and the 

majority’s reasoning on the question of jurisdiction is not only 

unconvincing but also contradicts the Court’s own jurisprudence. The 

decision in this case certainly clears up the intersection between 

labour law and administrative law, but the cost of that tidiness may 

be thought unacceptably high.

SUSTAINABLE DEVELOPMENT IN PRACTICE: FUEL 

RETAILERS ASSOCIATION OF SOUTHERN AFRICA V 

DIRECTOR-GENERAL ENVIRONMENTAL 

MANAGEMENT, DEPARTMENT OF AGRICULTURE, 

CONSERVATION AND ENVIRONMENT, 

MPUMALANGA PROVINCE 


* Associate Professor, Department of Public Law, University of Pretoria. 

1 

2007 6 SA 4 (CC) (Fuel Retailers). 

2 Fuel Retailers (n 1 above) para 109. 

3 The so-called filling station or fuel related cases include All the Best Trading CC 

t/a Parkville Motors, and Others v S N Nayagar Property Development and 

Construction CC and Others 2005 3 SA 396 (TPD); BP Southern Africa (Pty) Ltd v 

MEC for Agriculture, Conservation, Environment and Land Affairs 2004 5 SA 124 

(W); Capital Park Motors CC and Fuel Retailers Association of SA (Pty) Ltd v Shell 

SA Marketing (Pty) Ltd unreported, case no 3016/05, 18 March 2005 (TPD); Sasol 

Oil (Pty) Ltd & another v Metcalf 2004 5 SA 161 (W); and MEC for Agriculture, 

Conservation, Environment and Land Affairs, Gauteng v Sasol Oil and Another 

(368/2004) (2005) SCA 76. 

235 

Loretta Feris* 

In the opening salvo of his dissenting judgment in Fuel Retailers 1 

Justice Sachs notes the irony behind the fact that the first 

environmental rights case before the Constitutional Court came from 

‘an organised section of an industry frequently lambasted both for 

establishing world-wide reliance on non-renewable energy sources 

and for spawning pollution’. 2 However, this fact does not come as a 

total surprise. The fuel sector and filling stations in particular have 

been involved in several cases concerning environmental matters. In 

some of these cases the issue of sustainable development, which is 

guaranteed in the South African environmental right, has taken centre 

stage. 3 As a result, even before Fuel Retailers, we have seen the 

emergence of a South African jurisprudence on sustainable 

development. Fuel Retailers distinguishes itself by having made it all 

the way to the Constitutional Court and so having provided an 

opportunity for this court to express itself on this matter.

236 Sustainable development in practice 

However, it has to be made clear that this case is not primarily 

about section 24, the environmental right 4 of the Bill of Rights. It does 

not focus on the substance and normative content of the right. 

Rather, the cause of action lies in administrative law and governance 

and it is through this prism that one of the concepts central to section 

24, ‘sustainable development’, is explored, analysed and evaluated. 

Much of the case is in fact devoted to sustainable development and 

Ngcobo J, writing on behalf of the majority, spent a significant part 

of the judgment exploring the development of the concept in 

international law, its application in South African law and finally its 

import and consequence for the matter at hand. In addition, Sachs J 

in his dissent provides a contrary, and to my mind a more conceptually 

sound analysis of sustainable development. 

As a result of the majority’s expansive overview of sustainable 

development and Sach’s cogent construction thereof, the case 

contributes significantly to the conceptualisation of sustainable 

development in environmental law, a process that is by its very nature 

ongoing. It underscores and buttresses the development of the 

concept through earlier case law and as such stimulates the debate on 

how exactly we should apply sustainable development. 

In this note I view Fuel Retailers through the lens of sustainable 

development and explore the way in which the Constitutional Court 

has approached the concept. I refrain from engaging in a 

comprehensive analysis of the historical development of the concept 

or an in-depth analysis of the normative content thereof. I do, 

however, consider one of the core elements of sustainable 

development — integration, an element that took centre stage in Fuel 

Retailers and is at the heart of sustainable development decisionmaking. 

In particular, I investigate the value choices inherent in 

sustainable development decision-making and consider the failure of 

the Court to interrogate the practical and normative application of 

integration thoroughly. I furthermore assess the dissenting judgement 

of Sachs J and conclude that his interpretation provides some 

guidance on how to interpret instruments such as legislation and 

policy that require sustainable development. 

4 

Section 24 reads: ‘Everyone has the right — 

(a) to an environment that is not harmful to their health or well-being; and 

(b) to have the environment protected, for the benefit of present and future 

generations, through reasonable legislative and other measures that — 

(i) prevent pollution and ecological degradation 

(ii) promote conservation; and 

(iii) secure ecologically sustainable development and use of natural resources 

while promoting justifiable economic and social development.’

2 The facts 


The applicants in this matter objected to an authorisation that was 

granted by the Mpumalanga provincial environmental authorities for 

the establishment of a filling station in White River in Mpumalanga. 

The filling station was to be established in an area that was already 

the site of six filling stations within a five kilometre radius. An 

environmental impact assessment (EIA) was submitted to the 

Mpumalanga Department of Agriculture, Conservation and Environment 

(DACE) in terms of GN R1182 of 1997 5 issued under the 

Environment Conservation Act (ECA). 6 A record of decision (ROD) was 

signed on 9 January 2002 for the installation of three tanks for petrol 

and diesel, the erection of a convenience store, and ablution 

facilities. The ROD was issued on the basis of a scoping report, a 

geotechnical report and a geohydrological report. 

The scoping report dealt with the potential impact of noise, visual 

factors and traffic and the effect on municipal services, safety and 

crime, and cultural and historical sites; as well as the feasibility of the 

proposed station. However, it did not address the potential economic 

impact on other filling stations in the area. With regard to 

environmental impacts it found that no plant or animal species were 

threatened. The geotechnical report suggested, however, that a 

subterranean aquifer would need protection from pollution and 

recommended that — if required by the Department of Water Affairs 

and Forestry (DWAF) — an impermeable base layer should be installed 

to prevent contamination, and that water quality be tested biannually. 

7 

Fuel Retailers lodged an appeal to the MEC of DACE on the basis, 

inter alia, that the need, desirability and sustainability of the 

proposed filling station had not been considered and that the 

geophysical report regarding prevention of fuel leaks and possible 

contamination of the aquifer is inadequate. The appeal was 

considered and dismissed. In response, the Director argued that need, 

desirability and sustainability need not be considered, as these 

factors had been considered by the local authority when it had earlier 

approved the rezoning of the area from ‘special’ to ‘Business 1’, 8 

5 GG 18261 of 5 September 1997. 

6 Act 73 of 1989. In terms of sec 22(1) read with sec 21(1) no activity which may 

have a substantial detrimental impact on the environment may be undertaken 

without written authorisation from the competent authority. In terms of sec 22(2) 

such authorisation may only be granted ‘after consideration of reports concerning 

the impact of the proposed activity and of alternative proposed activities on the 

environment’; and after the furnishing of a report concerning the impact of the 

proposed development on the environment. 

7 

At no point did DWAF require any preventive measures. 

8 As required by the relevant Ordinance 15 of 1986, which governs rezoning.


which allowed for the construction of a filling station at the identified 

site. 

The applicants unsuccessfully applied to the High Court 9 and to 

the Supreme Court of Appeal 10 to set aside the ROD in terms of the 

ECA, the common law and the Promotion of Administrative Justice 

Act. 11 The applicants’ arguments were based on a number of issues, 

but on application for leave to appeal to the Constitutional Court, 

only one issue was pursued, namely that the environmental 

authorities failed to consider the socio-economic impact of 

constructing the proposed filling station. The applicants argued that 

this obligation is wider than the assessment of need and desirability 

considered in the rezoning process and that it required DACE to 

assess, amongst other things, the cumulative impact on the 

environment produced by the proposed filling station and all existing 

filling stations in close proximity to the proposed one. In sum, the 

proper analysis required an assessment of demand or necessity and 

desirability, not the feasibility, of the proposed station and its impact 

on the sustainability of existing filling stations. In support, the 

applicant relied upon the provisions of section 24(b)(iii) 12 of the 

Constitution, as well as sections 2(4)(a), 2(3), 2(4)(g), 2(4)(i), and 23 

of NEMA. 13 

9 2005 JDR 0915 (T). 

10 

2007 2 SA 163 (SCA). 

11 Act 3 of 2000. 

12 It provides: ‘Everyone has the right ... to have the environment protected, for 

the benefit of present and future generations, through reasonable legislative and 

other measures that ... secure ecologically sustainable development and use of 

natural resources while promoting justifiable economic and social development’. 

13 

2(4)(a): Sustainable development requires the consideration of all relevant 

factors including the following: 

(i) That the disturbance of ecosystems and loss of biological diversity are 

avoided, or, where they cannot be altogether avoided, are minimised and 

remedied; 

(ii) that pollution and degradation of the environment are avoided, or, where 

they cannot be altogether avoided, are minimised and remedied; 

(iii) that the disturbance of landscapes and sites that constitute the nation's 

cultural heritage is avoided, or where it cannot be altogether avoided, is 

minimised and remedied; 

(iv) that waste is avoided, or where it cannot be altogether avoided, 

minimised and re-used or recycled where possible and otherwise disposed of in 

a responsible manner; 

(v) that the use and exploitation of non-renewable natural resources is 

responsible and equitable, and takes into account the consequences of the 

depletion of the resource; 

(vi) that the development, use and exploitation of renewable resources and 

the ecosystems of which they are part do not exceed the level beyond which 

their integrity is jeopardised; 

(vii) that a risk-averse and cautious approach is applied, which takes into 

account the limits of current knowledge about the consequences of decisions 

and actions; and 

(viii)that negative impacts on the environment and on people's environmental 

rights be anticipated and prevented, and where they cannot be altogether 

prevented, are minimised and remedied.


The environmental authorities accepted that the socio-economic 

impact of the proposed filling station had to be considered, but 

argued that they were not required to consider need, desirability and 

sustainability of the proposed filling station as it was taken into 

account by the local authority during the rezoning of the property and 

that ‘rezoning forms part and parcel of the process of an application 

for authorisation in terms of section 22 of ECA’. 14 The applicant, in 

turn, argued that the two processes are distinct and separate. The 

local authority considers a rezoning application from a town-planning 

perspective and focuses on what land uses it would allow, whilst the 

environmental authorities are required to consider the impacts of the 

proposed development on the environment and on socio-economic 

conditions. 15 The applicant also noted that the rezoning had taken 

place approximately eight years prior to the approval of the 

application for authorisation in terms of the ECA and that significant 

environmental changes had occurred in the area since then. 16 

13 2(3): Development must be socially, environmentally and economically 

sustainable. 

2(4)(g): Decisions must take into account the interests, needs and values of all 

interested and affected parties, and this includes recognising all forms of 

knowledge, including traditional and ordinary knowledge. 

2(4)(i): The social, economic and environmental impacts of activities, 

including disadvantages and benefits, must be considered, assessed and 

evaluated, and decisions must be appropriate in the light of such 

consideration and assessment. 

23: The purpose of this Chapter is to promote the application of appropriate 

environmental management tools in order to ensure the integrated 

environmental management of activities. 

(2) The general objective of integrated environmental management is to — 

(a) promote the integration of the principles of environmental management 

set out in section 2 into the making of all decisions which may have a 

significant effect on the environment; 

(b) identify, predict and evaluate the actual and potential impact on the 

environment, socio-economic conditions and cultural heritage, the risks and 

consequences and alternatives and options for mitigation of activities, with a 

view to minimising negative impacts, maximising benefits, and promoting 

compliance with the principles of environmental management set out in 

section 2; 

(c) ensure that the effects of activities on the environment receive adequate 

consideration before actions are taken in connection with them; 

(d) ensure adequate and appropriate opportunity for public participation in 

decisions that may affect the environment; 

(e) ensure the consideration of environmental attributes in management and 

decision-making which may have a significant effect on the environment; and 

(f) identify and employ the modes of environmental management best suited 

to ensuring that a particular activity is pursued in accordance with the 

principles of environmental management set out in section 2. 

(3) The Director-General must coordinate the activities of organs of state 

referred to in section 24(1) and assist them in giving effect to the objectives 

of this section and such assistance may include training, the publication of 

manuals and guidelines and the co-ordination of procedures. 


15 

Fuel Retailers (n 1 above) para 32. 

16 Fuel Retailers (n 1 above) para 33.


3 The judgment 

The cause of action for judicial review centred on whether or not the 

environmental authorities had failed to take into consideration 

matters that they were required to consider before granting the 

authorisation under section 22(1) of the ECA. 17 Ngcobo J framed the 

issue as one that concerns the ‘nature and scope of the obligations of 

environmental authorities when they make decisions that may have a 

substantial detrimental impact on the environment’; and, in 

particular, ‘the interaction between social and economic 

development and the protection of the environment’. 18 It is this 

latter aspect — that is, the interaction between social and economic 

development and environmental protection — that provides the 

impetus for the lengthy interrogation of sustainable development by 

the court. 19 

3.1 Sustainable development 

Ngcobo J’s departure point is the inherent challenge that gave rise to 

the concept of sustainable development: namely that the promotion 

of development requires protection of the environment, whilst the 

environment cannot be protected if development does not pay 

attention to the costs of environmental destruction. 

This challenge has long been recognised and the Court details how 

this challenge gave rise internationally to the development of 

sustainable development and it traces this development to important 

events such as the Brundtland Report, 20 the 1972 United Nations 

Conference on the Human Environment or so-called ‘Stockholm 

Conference’ 21 (which stressed the relationship between development 

and the protection of the environment), the 1992 United Nations 

Conference on Environment and Development or so-called ‘Rio 

Conference’ (which generated principles on sustainable development 

and provided a framework for the development of the law of 

sustainable development) 22 and, finally, the 2002 Johannesburg 



19 

In Fuel Retailers (n 1 above) para 40 Ngcobo J confirms that in light of the 

dictates of sec 24 of the Constitution, the case does, indeed, raise a 

constitutional issue. What is disconcerting though is the way in which he frames 

sec 24. He states that sec 24 ‘guarantees to everyone the right to a healthy 

environment’. This is a positive framing of the right, whilst the right is in fact 

negatively framed, thus lessening the impact and extent of the right. One can 

only hope that this careless use of language was a mere oversight on the side of 

the Court. 


21 


22 Fuel Retailers (n 1 above) para 49-50.


World Summit on Sustainable Development (WSSD) which, according 

to the Court ‘reaffirm[ed] that sustainable development is a world 

priority’. 23 

In delineating the concept the Court refers to the Brundtland 

Commission’s definition of sustainable development as ‘development 

that meets the needs of the present without compromising the ability 

of future generations to meet their own needs,’ 24 but noted that 

recent commentators have refrained from defining sustainable 

development and have instead chosen to identify ‘the evolving 

elements of the concept of sustainable development’. These include: 

the integration of environmental protection and economic development 

(the principle of integration); sustainable utilisation of natural resources 

(the principle of sustainable use and exploitation of natural resources); 

the right to development; the pursuit of equity in the use and allocation 

of natural resources (the principle of intra-generational equity); the 

need to preserve natural resources for the benefit of present and future 

generations (the principle of inter-generational and intra-generational 

equity); and the need to interpret and apply rules of international law in 

an integrated systematic manner. 25 

In highlighting the first principle — that of integration as fundamental 

to the concept of sustainable development — the Court notes that the 

formal application thereof ‘requires the collection and dissemination 

of environmental information, and the conduct of environmental 

impact assessments’ and argues that the ‘practical significance of the 

integration of the environmental and developmental considerations is 

that environmental considerations will now increasingly be a feature 

of economic and development policy’. 26 

With regard to South African law, the Court is of the opinion that 

sustainable development offers a principle for the resolution of 

tensions between ‘the need to protect the environment on the one 

hand, and the need for socio-economic development on the other 

hand’. The concept thus, according to the Court, provides a 

framework for reconciling socio-economic development and 

environmental protection. 27 Ngcobo J states: ‘sustainable 

development does not require the cessation of socio-economic 

development but seeks to regulate the manner in which it takes 

place’. 28 According to the Court the National Environmental 

Management Act (NEMA), which was enacted to give effect to section 


24 



26 Fuel Retailers (n 1 above) para 52 (references omitted). 

27 




24 of the Bill of Rights, defines sustainable development as being the 

‘integration of social, economic and environmental factors into 

planning, implementation and decision-making for the benefit of 

present and future generations’. 29 Ngcobo J argues that ‘whenever a 

development which may have a significant impact on the environment 

is planned, it envisages that there will always be a need to weigh 

considerations of development’, with these considerations being 

‘underpinned by the right to socio-economic development’, against 

‘environmental considerations, as underpinned by the right to 

environmental protection. In this sense it contemplates that 

environmental decisions will achieve a balance between environmental 

and socio-economic developmental considerations through 

the concept of sustainable development’. 30 

3.2 Sustainable development and the EIA process 

The Court argues that the 

nature and scope of the obligation to consider the impact of the 

proposed development on socio-economic conditions must be 

determined in the light of the concept of sustainable development and 

the principle of integration of socio-economic development and the 

protection of the environment. 31 

If this relationship is accepted, so Ngcobo J argues, it follows that 

socio-economic conditions have an impact on the environment. He 

illustrates this point by stating that the proposed filling station may 

have implications for the job security of existing filling stations, but 

may equally have environmental consequences, such as that if 

existing stations close down that will require rehabilitation of the 

environment after closure. 32 

He furthermore cautions that the proliferation of filling stations 

close to each other may increase the likelihood of adverse impacts on 

the environment. In this respect the Court reminds us that NEMA 

‘requires that the cumulative impact of a proposed development, 

together with the existing developments on the environment, socioeconomic 

conditions and cultural heritage must be assessed’. As such 

‘a consideration of socio-economic conditions ... includes the 

consideration of the impact of the proposed development not only in 

29 Fuel Retailers (n 1 above) para 59. The reference indicated is sec 1(1)(xxix). It 

should be noted, however, that this section does not exist and this ‘definition’ is 

30 

31 

32 

in fact from the preamble to NEMA. 



As above.


combination with the existing developments, but also its impact on 

existing ones’. 33 

With respect to the EIA process, the Court notes that NEMA 

requires the consideration, assessment and evaluation of the social, 

economic and environmental impact of proposed activities and this 

‘enjoins the environmental authorities to consider and assess the 

impact of a proposed activity on existing socio-economic conditions, 

which must of necessity include existing developments’. 34 The Court 

explains that the objective of this exercise is not to stamp out 

competition or protect existing developments at the expense of 

future developments. Rather it requires a risk-averse and cautious 

approach when decisions are made. 35 It concludes that what was 

demanded of the environmental authorities was to consider ‘the 

impact on the environment of the proliferation of filling stations as 

well as the impact of the proposed filling station on existing ones’, an 

obligation which is ‘wider than the requirement to assess need and 

desirability under the Ordinance’. 36 

It was common cause that this assessment never happened. The 

Court, however, contends that this duty fell on environmental 

authorities because the two organs of state have different roles and 

functions. Local authorities have a town planning role and as such a 

proposed development may satisfy the need and desirability criteria 

from a town-planning perspective, given that the local authority is not 

required to consider the social, economic and environmental impact 

of a proposed development. Environmental authorities, however, are 

required to do so by the provisions of NEMA. 37 The assumption that 

the duty under the Ordinance to consider need and desirability 

imposed the same obligation as is required by the duty under NEMA to 

consider the social, economic and environmental impacts of a 

proposed development, is therefore wrong. 38 

The Court, as a result sets aside the decision of the environmental 

authorities on a number of bases: 

(i) They misconstrued the nature of their obligations under NEMA and 

consequently failed to apply their minds to the socio-economic impacts 

of the proposed filling station in the instant case. 39 

33 Fuel Retailers (n 1 above) para 72. See also para 74. 


35 

Fuel Retailers (n 1 above) paras 78-81. 



38 


39 As above.


(ii) They left the consideration of need and desirability to the local 

authority and consequently failed to properly discharge of their 

statutory duty. 40 

(iii) They failed to take into account socio-economic conditions as 

required by NEMA and as such they did not comply with a mandatory and 

material condition set by NEMA. 41 

(iv) Given the lack of consideration to the social and economic impact of 

the proposed filling station, including its cumulative impact on the 

environment, the environmental authorities failed to apply their minds 

to the matter. 42 

4 Discussion 

4.1 Sustainable development through the cases 

As mentioned above Fuel Retailers focuses on the concept of 

sustainable development and more specifically how it is to be 

interpreted, contextualised and applied in South African law. The 

Court thoroughly interrogated the concept’s history and status in 

international law and concluded that it is now an established concept 

in international law. With regard to South African law the concept is 

still in its infancy. 43 A number of cases have, however, addressed 

sustainable development and over the last four years a South African 

jurisprudence has begun to see the light. 

The first case to explore this concept was one very similar to Fuel 

Retailers: Sasol Oil (Pty) Ltd & another v Metcalf. 44 The applicants 

40 Fuel Retailers (n 1 above) paras 87-88. 

41 


42 Fuel Retailers (n 1 above) para 90–97. The Court emphasises the fact that the 

rezoning decision by the local authorities, on which the environmental authorities 

relied, had probably occurred in 1995, some eight years prior to the decision by 

the environmental authorities. This was done prior to the coming into effect of 

NEMA and it argues that it was necessary for the environmental authorities to 

consider the matter afresh in the light of the provisions of NEMA, given that a 

significant change in the environment could have taken place in the intervening 

period. 

43 

Note, however, that as early as 1971 our neighbouring (then) Rhodesian court 

recognised that the principle of inter-generational equity can have the effect of 

limiting other rights, such as property rights. Thus in King v Dykes 1971 3 SA 540 

(RA) 545G-H MacDonald ACJ stated: ‘The idea which prevailed in the past that 

ownership of land conferred the right on the owner to use his land as he pleased 

is rapidly giving way in the modern world to the more responsible conception that 

an owner must not use his land in a way which may prejudice his neighbours or 

the community in which he lives, and that he holds his land in trust for future 

generations. Legislation dealing with such matters as town and country planning, 

the conservation of natural resources, and the prevention of pollution, and 

regulations designed to ensure that proper farming practices are followed, all 

bear eloquent testimony of the existence of this more civilised and enlightened 

attitude towards the rights conferred by ownership of land.’ 

44 Sasol (n 3 above).


sought the respondent’s, the Gauteng Department of Agriculture, 

Conservation, Environment and Land Affairs’, authorisation in terms 

of section 22 of the ECA for the construction of a filling station and 

convenience store. This authorisation was refused and in its refusal 

GDACE was guided by departmental guidelines that it had developed 

to assist in decision-making. The guidelines stipulated that 

development must be socially, environmentally and economically 

sustainable and required that filling stations in an urban, residential 

or built-up area should not be situated within three kilometres of an 

existing filling station. The applicants thus approached the High Court 

for a declaration that certain sections of these guidelines issued by 

the respondent were ultra vires the ECA and were, therefore, invalid 

and unenforceable. They furthermore argued that whilst GDACE had 

the power to authorise the construction of structures for the storage 

and handling of hazardous substances at filling stations, this power 

did not include the erection of filling stations itself. As such the 

guidelines did not apply. Finally they argued that even if the power 

did include the erection of filing stations, GDACE was not entitled to 

apply the guidelines as they were predominantly based on socioeconomic 

as opposed to environmental considerations. The High Court 

dismissed the ultra vires argument, but upheld the applicants’ 

argument that GDACE did not have the power to require authorisation 

for the construction of filing stations per se. 45 It furthermore held 

that the sustainable development principles contained in section 2 of 

the NEMA did not extend the mandate of GDACE to take socioeconomic 

(and not just environmental) factors into account. 46 

This decision was a set-back for sustainable development 

jurisprudence. It ignored the core element of integration as part of 

sustainable development. 47 The Brundtland Report, referred to by 

Ngcobo J in Fuel Retailers, contextualises the principle of integration 

and explains that sustainable development not only prioritises the 

needs of the poor, but that it also captures the limitations to 

development imposed by the present state of technology and social 

organisation on the environment’s ability to meet present and future 

needs. 48 The Report suggests an inherent link between social and 

environmental needs and the need for technological advancement 

and development. An imbalance amongst these elements, where 

global patterns of development put the environment under pressure, 

places the earth in crisis. This principle of integration between three 

pillars — environmental protection, economic development and social 

45 

46 

47 

Sasol (n 3 above) 170A-E. 

Sasol (n 3 above) 171E-172B. 

For a more in-depth critique of this case see Patterson A ‘Fuelling the sustainable 

48 

development debate in South Africa’ (2006) 1 South African Law Journal 53. 

Report of the World Commission on Environment and Development: Our Common 

Future (1987).


needs — is thus widely recognised as a core element of sustainable 

development. 49 

This decision was eventually overturned by the SCA. 50 The SCA’s 

reversal was, some ways, due to a conflicting judgement by the same 

court in yet another filling station case, BP Southern Africa (Pty) Ltd 

v MEC for Agriculture, Conservation, Environment and Land Affairs. 51 

The applicant in this case sought the review and setting aside of a 

decision by the Gauteng Provincial Department of Agriculture, 

Conservation, Environment and Land Affairs (GDACE) to refuse its 

application in terms of section 22(1) of the ECA for authorisation to 

develop a filling station on one of its properties. The Department 

based its refusal, inter alia, on environmental concerns. The 

applicant contended, however, that its application was refused not 

because the new filling station itself posed a danger to the 

environment, but rather, because of the fact that there were already 

two other filling stations within three kilometres of applicant's site 

and the Department regarded it as unacceptable to allow 

proliferation of filling stations where existing filling stations were 

economically vulnerable to more competition. It argued that under 

the guise of ‘environmental concerns’, the department was instead 

seeking to regulate the economy on the basis of what were essentially 

economic considerations unrelated to the environment. 

The court explored the concept of sustainable development in 

somewhat more detail than in Sasol and in so doing considered the 

wide definition of ‘environment’ employed by ECA which defines it in 

section 1 as: ‘the aggregate of surrounding objects, conditions and 

influences that influence the life and habits of man or any other 

organism or collection of organisms’. According to the court the broad 

definition of ‘environment’ would include all conditions and 

influences affecting the life and habits of man, which would also 

include socio-economic conditions and influences. 52 With reference 

to the state’s obligation under section 24(b) with regard to 

sustainable development, the Court held that ecologically sustainable 

development and the use of natural resources must be promoted 

jointly with justifiable economic and social development. 53 It stated 

that: 

49 

See for example P Sands Principles of international law (2003) 153. See also D 

Tladi Sustainable development in international law: An analysis of key enviro- 

50 

economic instruments (2007) 58. 

MEC for Agriculture, Conservation, Environment and Land Affairs (Pty) Ltd & 

51 

52 

53 

another (n 3 above). 

2004 5 SA 124 (W). 

BP (n 3 above) 145E. 

BP (n 3 above) 143C-D.


... sustainable development constitutes an integral part of modern 

international law and will balance the competing demands of 

development and environmental protection. The concept of ‘sustainable 

development’ is the fundamental building block around which 

environmental legal norms have been fashioned, both internationally 

and in South Africa ... pure economic principles will no longer 

determine, in an unbridled fashion, whether a development is 

acceptable. Development, which may be regarded as economically and 

financially sound, will, in future, be balanced by its environmental 

impact, taking coherent cognisance of the principle of intergenerational 

equity and sustainable use of resources in order to arrive at 

an integrated management of the environment, sustainable 

development and socio-economic concerns. By elevating the environment 

to a fundamental justiciable human right, South Africa has 

irreversibly embarked on a road, which will lead to the goal of attaining 

a protected environment by an integrated approach, which takes into 

consideration, inter alia, socio-economic concerns and principles. 54 

The court thus concluded that the department was obliged to develop 

an integrated environmental management programme, which took 

cognisance of a wide spectrum of considerations, including 

international conventions and approaches as a result of the broad and 

extensive definition of ‘environment’ in ECA and the principles of 

NEMA which, inter alia, include the consideration of socio-economic 

conditions. 55 

With the more detailed attention to sustainable development 

provided by the Constitutional Court in Fuel Retailers we now have 

significant judicial guidance on interpreting and implementing 

sustainable development. In essence we now know that sustainable 

development is central to the environmental right and to environmental 

regulation. We also know that our jurisprudence 

acknowledges the inter-connected nature of environmental, social 

and economic considerations of sustainable development and that the 

principle of sustainable development aims (albeit, as argued further 

on, mostly unsuccessfully) to serve the purpose of facilitating the 

achievement of this balance. One could, therefore argue, that a 

central tenet of a ‘South African jurisprudence’ on sustainable 

development is the principle of integration. 

4.2 Sustainable development applied 

As noted by Ngcobo J in Fuel Retailers, the overarching definition of 

sustainable development is the one as espoused by the Brundtland 

Report: ‘development which meets the needs of the present 

54 

BP (n 3 above) 144A-144D. 

55 BP (n 3 above) 150D-E.


generation without compromising the ability of future generations to 

meet their own needs’. 56 In attempting to find the practical 

application of sustainable development it may be helpful to view this 

definition of the Brundlandt Report as the aim of sustainable 

development — that is, that which we want to achieve. This aim 

operates in acknowledgment of the fact that whilst human beings are 

driven by their developmental needs to use, exploit and even exhaust 

natural resources, this can and may not happen in a limitless way. 

Thus, as noted by Field, sustainable development could be described 

as the ‘conceptual vehicle chosen by a diverse range of actors to 

negotiate the tensions arising from the need for social and economic 

development on a planet with finite resources’. 57 

This earlier definition has been elaborated upon by more recent 

authors through the identification of ‘the evolving elements of the 

concept of sustainable development’. 58 I would suggest that these 

elements can, in turn, be viewed as the ‘means to achieve the aim’. 

These means would therefore include: sustainable utilisation of 

natural resources, the pursuit of equity in the use and allocation of 

natural resources, and the integration of environmental protection 

and economic development. 59 These elements attempt to give 

concrete existence to a concept that may be viewed as elusive and 

impractical, largely because the concept involves competing 

considerations or normative tensions. 

Out of the three elements, the integration principle has been 

identified as the key principle. 60 It is also the element that the Court 

focused on when it suggested that sustainable development requires 

integrating the often contesting demands of economic development, 

social development and environmental protection. This element 

remains, however, open to criticism. It has been suggested, not 

without justification, that its contents remain rather opaque. 61 

What, for example does ‘integration’ really mean and how is it 

practically achieved? If a waste site is situated close to a residential 

area, where that site generates an income not only for the managing 

company, but also for the residents that live nearby, should that site 

56 Report of the World Commission on Environment and Development: Our Common 

57 

Future (1987) 8. 

Field T ‘Sustainable development versus environmentalism: Competing paradigms 

58 

59 

for the South African EIA regime’ (2006) 123 South African Law Journal 409 411. 


See Sands P Principles of international environmental law (2003) 253. Ngcobo 

refers to these elements, but also includes a wider range of elements (para 51). 

These three elements are, however, the most widely recognised elements of 

60 

sustainable development. 

D Tladi Sustainable development in international law: An analysis of key enviro- 

61 

economic instruments (2007) 58. 

Tladi (n 60 above) 75.


be closed down so as accommodate environmental health 

considerations? Or should it be allowed to remain open in order to 

accommodate social and economic considerations? How does one 

integrate, if at all, these three contesting considerations? One could 

argue that integration is the ‘happy medium’ where one tightly 

regulates the operations of the waste site so as to minimise the 

exposure of the nearby residents, while still ensuring that the site 

contribute to the economy and provides a source of income for its 

workers. In reality this happy medium is a value choice. In this 

instance the decision-making is primarily driven by socio-economic 

considerations. Requiring strict operating conditions, however, means 

that the third pillar, the environment, remains part of the overall 

decision-making process. 

The practical and normative application of integration was only 

marginally addressed by the Court in Fuel Retailers as the ‘collection 

and dissemination of environmental information’, the ‘conduct of 

environmental impact assessments’ and the fact that environmental 

considerations ‘will now increasingly be a feature of economic and 

development policy’. 62 While EIAs and policy tools may be useful in 

the practical application of the integration element of sustainable 

development, they clearly do not address other forms of decisionmaking, 

such as decisions regarding activities where no EIAs were 

conducted, nor do they explain how these decisions are to be made 

even when there is an EIA. 

A current burning issue for example is that of land restitution and 

its relationship vis-à-vis sustainable development. 63 A number of 

current land claims include claims to land that have been declared 

protected areas. This includes both private and state-owned land, 

including for example claims in the Kruger National Park. In making a 

decision on whether to award such claims decision-makers would have 

to take into account the possibility that claimants may not utilise the 

land for conservation purposes, but rather to engage in other 

commercial ventures such as farming. This would clearly promote an 

economic and social goal as opposed to an environmental goal. 

However, whilst the environmental aim of preserving our natural 

heritage may weigh very heavy, equally so the would the idea of 

restoring land to people who were unjustly deprived thereof in the 

past. 

The judgment also does not address the notion that the 

integration principle could be used equally effectively by diverse 


63 For a more detailed discussion see A Du Plessis ‘Land restitution through the lens 

of environmental law: Some comments on the South African vista’ (2006) 1 

Potchefstroom Electronic Law Journal 1 1.


groups with conflicting aims, such as environmentalists versus those 

pursuing economic development aims. 64 Tladi argues that sustainable 

development is inherently a flexible concept, which would have the 

effect that for those advocating economic growth, the emphasis 

would fall on the economic growth value of sustainable development. 

As such sustainable development could mean ‘lasting economic 

growth’, with the aim being to sustain economic growth. 65 Such an 

understanding effectively dilutes and detracts from the original aim 

of sustainable development. The other two elements, sustainable use 

of natural resources and equitable utilisation of natural resources, 

attempt to safeguard the environment against unbridled economic 

development. 

Decisions motivated by socio-economic considerations can 

therefore potentially be disguised as decisions prompted by 

environmental concerns. This was, in fact, the argument made by the 

applicants in BP, namely that the decision to refuse authorisation for 

the proposed activity was motivated not by environmental 

considerations, but rather by the socio-economic considerations as 

set out in the province’s guidelines on EIAs. 

True integration is often depicted as three circles that intersect 

with sustainable development situated in the intersection of these 

circles as shown in the diagram below. 

However, whilst the diagram suggests optimal overlap it is always 

possible that reality indicates otherwise and there may be cases 

where there is very little overlap and where the greater part is in one 

64 

As above. 

65 As above.


of the circles. In other words, the three elements, environmental 

sustainability, economic sustainability and social sustainability do not 

always carry equal weight in decision-making. When a decisionmaker, 

whether it is an administrative official or a judicial officer, 

takes into account sustainable development in the decision-making 

process, he or she makes a value-based judgement. This judgment is 

informed by the values of environmental sustainability, and economic 

sustainability as part of the integration process. However, one (or 

sometimes two) of these values may inevitably triumph. 

Tladi therefore suggests a more nuanced approach in the 

application of sustainable development, one that provides three 

variations of integration based on the value that is the preferred one 

in cases of conflict. In the economic growth-centred variation, 

economic growth takes centre stage, whilst in the environmentcentred 

variation, the natural environment triumphs. Finally in the 

human needs-centred (or social needs centred) variation the social 

needs of humans are placed at the forefront. 66 He argues that such a 

variation approach allows decision-makers to decide which variation 

best serves the aims of sustainable development. 67 In essence his 

analysis suggests that sustainable development decisions are 

inevitably value driven. It means that decision-makers decide in 

advance which of the values they prefer to advance, and whilst still 

taking into account the other two values, base the decision on the 

preferred value. 

If we then apply this ‘variation approach’ to integration, it would 

lead us to the conclusion that the decision in Fuel Retailers may fall 

within the economic growth or human centred (or a combination of 

both) variation since the Court has emphasised the failure of the 

environmental authorities to take socio-economic considerations into 

account? I would suggest not. Whilst the integration process is a valuedriven 

process, the preferred value cannot be without a legitimate 

basis. In other words, a decision maker’s decision should be grounded 

in law and there should be some justifiable base in law for the 

preferred value. Such a basis may be found in a legal or policy 

instrument and may provide an indication of the preferred variation 

of sustainable development. 

In this respect Sachs J’s dissent is instructive. In essence he 

provides us with the application of this ‘variation’ approach to the 

integration element of sustainable development and takes NEMA as 

his ‘legitimising base’. With regard to the application of the preamble 

and principles of NEMA, he notes that ‘economic sustainability is not 

66 Tladi (n 60 above) 80. His idea is not that placing one value centre stage would 

obliterate the other, but rather that it would re-enforce the other two. 

67 Tladi (n 60 above) 82.


treated as an independent factor to be evaluated as a discrete 

element in its own terms’, but rather that the focus is on the interrelationship 

between economic sustainability and environmental 

protection. 68 Accordingly, he argues, NEMA does not envisage that 

social, environment and economic sustainability should proceed along 

separate tracks, with each being assessed separately and only 

considered together at the end of the decision-making process. It is 

his contention that economic sustainability takes on significance only 

to the extent that it implicates the environment. As such, it is only 

‘when economic development potentially threatens the environment 

that it becomes relevant to NEMA’ and it is only at this point that it 

should be considered within the context of the sustainable 

development requirements of NEMA. 69 Sachs bases this conclusion on 

the thesis that ‘all environmental controls were in place and that any 

potential deleterious effect of over-trading was speculative and 

remote’. 70 

Sachs situates his position within the dictates of sustainable 

development as required by NEMA. The overall aim of NEMA is to first 

and foremost ensure environmental protection. NEMA thus grants 

priority to the environment-centred variation of sustainable 

development. In situations of conflict between economic, social and 

environmental considerations, the latter must be preferred. Given 

that NEMA operates within this model Sachs’s argument that social 

and economic considerations are only ‘triggered’ once the 

environment is implicated makes sense. 

Sachs could equally have used section 24 of the Constitution as his 

‘legitimising base’. Section 24(b)(iii) of the Constitution refers to the 

need to ‘secure ecologically sustainable development’. [own 

emphasis]. It can be argued that ‘ecologically’ qualifies the type of 

sustainable development that is envisioned by the Constitution. It 

therefore clearly places an emphasis on environmental considerations 

and as such it places the environmental value centre-stage. Section 

24 of the Constitution therefore mandates the environment-centred 

variation of sustainable development. Any decision-making regarding 

sustainable development that is mandated by section 24 should, 

therefore, be situated within this model. 

The Sachs approach could be seen as ‘applied variation’ as it gives 

us some guidance on how to interpret instruments such as legislation 

and policy that requires sustainable development. In applying the 

model not only to the decision in the Fuel Retailers case but also the 

decision in the BP case, it exposes both for their inadequate and 


69 

As above. 



ultimately unsatisfying application of sustainable development. 

Whilst both decisions were at first glance ‘good for the environment’, 

they were really motivated by socio-economic considerations, as such 

applying economic centred variations of integration when the 

Constitution and NEMA really required environment centred 

variations. If the latter were applied, different outcomes would have 

followed in both cases. 

5 Conclusion 

Fuel Retailers may have been hailed as a victory for the environment 

and for environmental rights, but was it the correct decision from a 

normative point of view? Sustainable development is a truly laudable 

concept, but as pointed out it raises a number of challenges in 

negotiating the increasingly contested demands of development and 

environmental protection. It embodies competing values and requires 

other conceptual and normative framework to address these tensions 

and to provide practical solutions. On its own, the concept fails to 

address practical and normative considerations. 

We therefore need a more principled, value-based approach to 

sustainable development that identifies the value that is being 

prioritised in particular circumstances. The variation approach to 

integration takes into account that certain norms and values will most 

often be paramount in sustainable development decision-making. 

Whilst Fuel Retailers has highlighted the ongoing debate on 

sustainable development in the South African context, it ought not 

have been the final word. The majority judgement failed to 

interrogate the normative nature of sustainable development 

comprehensively and in the process provided us with an inherently 

flawed and incomplete application of the concept. Sachs’ dissent 

provides an alternative, and more nuanced perspective on this 

complex norm.


* 

FUEL RETAILERS, SUSTAINABLE 

DEVELOPMENT & INTEGRATION: 

A RESPONSE TO FERIS 

Legal Counsellor to the South African Permanent Mission to the United Nations in 

New York. Extraordinary Professor in the Department of Public Law at the 

University of Stellenbosch. 

255 

Dire Tladi * 

Fuel Retailers is, without question, a most interesting and significant 

decision: (1) Interesting because the decision tackles head on the 

difficult task of balancing apparently conflicting interests — the right 

to a healthy environment and economic imperatives of development; 

(2) Significant because, as the first decision in which the 

Constitutional Court gives content to section 24 of the Constitution, 

one expects the decision to play a major role in the development of 

environmental law and sustainable development. 

Loretta Feris’ comment 1 provides a fairly detailed (and accurate) 

overview of Fuel Retailers. 2 The exploration of the nuances of 

sustainable development in Professor Feris’ contribution, rooted as it 

is in pragmatism, makes the contribution particularly useful to 

practitioners — especially those who evaluate applications for 

approval for activities with a potential impact on the environment. 

As someone concerned more with the theoretical and normative 

aspects of sustainable development, I thought the paper important 

for engaging the complexity of the process of integration in 

sustainable development discourse — a complexity that is often 

1 L Feris ‘Sustainable development in practice: Fuel Retailers Associatioin of 

Southern Africa v Director-General Environmental Management, Department of 

Agriculture, Conservation and Environment, Mpumalanga Province’ (2008) 1 

Constitutional Court Review 235. 

2 Fuel Retailers Association of Southern Africa v Director-General Environmental 

Management, Department of Agriculture, Conservation and Environment, 

Mpumalanga Province, and Others 2007 6 SA 4 (CC).

256 A response to Feris 

missed because of our collective reverence for the concept of 

sustainable development. 3 Feris and I are in accord over a large range 

of issues, in particular the view that Sachs J’s dissent provides the 

most ‘conceptually sound analysis of sustainable development’. 4 Our 

common ground established, I would now like to offer my slightly 

different take on what the Fuel Retailers Court misses. 

2 The Court’s treatment of sustainable 

development 

Feris states upfront that she will avoid a ‘comprehensive analysis of 

the historical development of the concept or the normative content’ 

of sustainable development. 5 For me, the historical development of 

the concept is critical for understanding the normative content, 

which, in turn, influences its practical application. Ngcobo J’s 

majority judgment does, in fact, survey the historical evolution of 

sustainable development — he traces it back to the 1972 Stockholm 

Conference on the Human Environment. 6 The historical exposition of 

sustainable development by the Court suffers from the same ailment 

as many similar expositions. First, the Court erroneously states that 

the term sustainable development was ‘coined’ by the Brundtland 

Commission in 1987. 7 Second, the Court does not explain how this 

evolution influences the conceptualisation that follows in the rest of 

the judgement. The Court offers, so to speak, a recollection of history 

for history’s sake. While time and space do not allow for a discussion 

of how the historical evolution of sustainable development ought to 

influence its conceptualisation, the Court’s history is notably flawed 

in this regard. 8 

The problem with the Court’s analysis, in my view, begins when 

the Court, in connection with section 24 of the Constitution, makes 

reference to the ‘explicit recognition of the obligation to promote 

justifiable “economic and social development”’ and links this notion 

with the ‘well-being of human beings’ and ‘socio-economic rights’. 9 

Precisely what that link is, the Court never explains or explores. But 

the result of this linkage is that, throughout the judgement, the terms 

3 

4 

Someone, I cannot recall who or in what context, once remarked that sustainable 

development, like human rights, is a concept that is not to be questioned in 

polite company. 

Feris (n 1 above) 236. 

5 Feris (n 1 above) 236. 


7 

Fuel Retailers (n 2 above) para 47. Already in 1980, the IUCN World Conservation 

Strategy contained several references to sustainable development. 

8 For my views, see D Tladi Sustainable development and international law: An 

analysis of key enviro-economic instruments (2007) 34 et seq. 



‘socio-economic rights’, ‘development’ and ‘economic development’ 

are used interchangeably as the values that most often oppose the 

right to a clean and a healthy environment. 10 At one point the Court, 

for example, refers to the integration of environmental protection 

and economic development. 11 Elsewhere, the Court states that as a 

result of sustainable development ‘environmental considerations will 

now increasingly be a feature of economic and development policy’. 12 

Further on, the Court states that ‘economic development, social 

development and the protection of the environment’ are considered 

to be the three pillars of sustainable development. 13 Finally, the 

Court asserts that sustainable development ‘provides a framework for 

reconciling socio-economic development and environmental 

protection’. 14 

By lumping these concepts together the Court misses an 

opportunity to develop a sound understanding of sustainable 

development. The result of treating these concepts as 

interchangeable is that the Court never stops to ask whether the 

factors that the Fuel Retailers Association requested that the 

environmental authorities consider are socio-economic or purely 

economic. To use language from the common definition of sustainable 

development, the Court does not ask whether these factors are social 

or economic. The Court’s judgment implies — incorrectly — that 

economic considerations are the same as social considerations. 

A similar conflation of these concepts is evident in Feris’ analysis. 

After applying the model of three variations of sustainable 

development that I have proposed, she asks whether Fuel Retailers 

would ‘fall within the economic growth or human-centred (or a 

combination of both) variation’ of sustainable development because 

of the failure of the authorities to consider ‘socio-economic 

considerations’. 15 Although Feris answers the question in the 

negative, it is the manner in which she treats economic growth 

centred variation and human well-being centred variation as similar 

that is of some concern. 

One can readily understand the intuitive lumping together of 

these two variations of sustainable development (and the values they 

represent). After all, at the centre of both of these variations are 

human-related concerns, namely economic concerns and social 

concerns. However, in my view these variations (and their values) 

tend to pull in different directions. Factors relevant for economic 

10 Fuel Retailers (n 2 above) paras 51, 52, 53 & 55. 

11 




14 


15 Feris (n 1 above) 249.

258 A response to Feris 

growth variation are, for example, trade related concerns. Access to 

clean drinking water and food reflect a human well-being variation of 

sustainable development. They are not the same. 16 The approach of 

the Court in this instance — and reflected in some of Feris’ comments 

— takes us back in time to the old definition of sustainable 

development. On this outré account, sustainable development 

balanced environmental needs and development needs — the latter 

notion represented both social and economic concerns. 

Sustainable development — and this is where a proper historical 

sketch is important — was born out of a realisation that the existing 

paradigm (in which economic concerns trumped all other concerns 

(social and environment) — could not continue. By blurring the 

distinction between social and economic concerns, our jurisprudence 

flirts with the undesirable outcome of preserving the status quo: 

namely, paying lip service to sustainable development and 

integration. The failure to distinguish more carefully between these 

values facilitates the instrumentalisation of sustainable development 

for economic ends. Fuel Retailers is a case in point. 

3 Concluding remarks 

The treatment of sustainable development by the Constitutional 

Court was long overdue. For that reason alone Fuel Retailers is a 

welcome addition to South African jurisprudence. Given the 

shortcomings of the judgment delineated in Professor Feris’ comment 

and my reply, we can only hope that the Court will, over time, 

develop a more nuanced approach to sustainable development that 

does justice to its history, makes subtle but important distinctions 

between economic concerns and social concerns, and does not allow 

this concept to be captured by those parties with purely pecuniary 

motives. 

16 

For an in-depth discussion of how these variations and their factors, see Tladi (n 8 

above) 81-90.


CULTURAL AND RELIGIOUS 

ACCOMMODATIONS TO SCHOOL 

UNIFORM REGULATIONS 

* MA LLB (Natal) PhD (Cantab). Senior Lecturer, School of Philosophy and Ethics, 

University of KwaZulu-Natal. 

1 Church of Lukumi Babalu Aye Inc v City of Hialeah 508 US 520 (1993). 

2 Section 15(1) of the Constitution of the Republic of South Africa, 1996 (‘the 

Constitution’) provides that ‘everyone has the right to freedom of conscience, 

religion, thought, belief and opinion’. 

259 

Patrick Lenta * 

Laws and regulations in modern liberal democracies rarely 

discriminate deliberately against members of religious and cultural 

groups or target religious and cultural practices. In one relatively 

recent example, Church of Lukumi Babalu Aye, Inc v City of Hialeah 

(Church of Lukumi Babalu Aye), the US Supreme Court invalidated 

municipal ordinances adopted by the city of Hialeah for the specific 

purpose of proscribing animal sacrifice practised by the Santeria 

religion. 1 Since these ordinances did not constitute a neutral law of 

general applicability, but deliberately targeted a religious practice, 

the Supreme Court determined that they were invalid unless they 

served a compelling state interest. Since the state could not show 

such an interest, the ordinances were declared, relatively 

uncontroversially, to be in violation of the Free Exercise clause (the 

US equivalent of South Africa’s right to freedom of religion and 

conscience) 2 and so invalid. 

Laws and regulations that are facially neutral in the sense that 

they do not deliberately target members of particular cultural or 

religious groups are more common than those which are overtly 

discriminatory. In some instances, however, these facially neutral 

laws may have non-neutral effects and impose disparate burdens on 

group members. These individuals sometimes claim from the courts

260 Cultural and religious accommodations to school uniform regulations 

exemptions to generally applicable laws and regulations on the 

grounds that particular laws or regulations impose burdens on them 

that are not imposed on others. For example, Muslim female pupils 

may seek exemptions from school uniform regulations that forbid the 

wearing of headscarves on the grounds that their religious beliefs 

obligate them to wear a headscarf or that wearing a headscarf is a 

genuine expression of their faith. 

Should Muslim girls be required to relinquish this cherished 

religious practice as a condition of availing themselves of the 

opportunity of education, or are schools obligated to meet their 

demands for accommodation? The most compelling argument that 

claims for exemptions in such cases should be granted begins with the 

claim that cultural group membership (including membership of 

religious groups) is intrinsically valuable for individual group 

members. Will Kymlicka argues that cultural membership is 

the context within which we choose our ends, and come to see their 

value, and this is a precondition for self-respect, of the sense that one’s 

ends are worth pursuing. 3 

Membership of groups provides individuals with personal and social 

identities, on which their dignity and prosperity may depend. 4 

3 W Kymlicka Liberalism, community and culture (1989) 193. 

4 For other accounts of the significance of group membership, see J Raz 

‘Multiculturalism: a liberal perspective’ in his Ethics in the public domain (1994), 

B Parekh Rethinking multiculturalism (2000), J Tully Strange multiplicity (1995) 

and C Taylor Multiculturalism and the politics of recognition (1992). For a recent 

acknowledgment by the Constitutional Court concerning the value of group 

membership to members, see Minister of Home Affairs & Another v Fourie & 

Another 2006 1 SA 524 para 89. The existence of cultural groups has social 

benefits too. Characterising religious groups in the United States as a source of 

‘social capital’, R Putnam Bowling alone: The collapse and revival of the 

American community (2000) 79, has observed that ‘faith-based organisations 

serve civic life both directly, by providing social support to their members, and 

indirectly, by nurturing civic skills, inculcating moral values [and] encouraging 

altruism’. It is not fanciful to imagine that the moral values to which Putnam 

refers could potentially provide a counterweight to what B Barry Why social 

justice matters (2005) 236, calls ‘the recrudescent ideology of individualism, 

with its concomitant implications that members of society are nothing to one 

another’, in contemporary capitalist society. Also important are the ‘benefits to 

society of having religious groups operating as vital associations intermediate 

between individuals and government and creating a barrier to government 

domination of social life’ (K Greenawalt Religion and the constitution (2006) 

439). The Constitutional Court recognised certain of the benefits of religion to 

society in Fourie (above) para 90. It must nevertheless be conceded that ‘there is 

nothing automatically good about religion, which can after all lead to war, 

insanity, mutilation, suicide, soul-murder’. (JB White ‘Talking about religion in 

the language of the law: Impossible but necessary’ (1999) 81 Marquette Law 

Review 177 193).


The argument that Muslim headscarves should be accommodated 

by school authorities proceeds as follows. 5 According to the liberal 

principle of equality of opportunity, individuals with differing aims 

should be afforded an equal chance to realise their ambitions and the 

costs that people should have to bear to do so should be, as far as 

possible, equal. But since cultural membership is bound up with 

individual autonomy, dignity, prosperity and self-respect (all of which 

liberals have a reason to value), in assessing whether people have 

equal opportunities, their cultural and religious commitments must 

be taken into account. A prospect only represents an opportunity for 

an individual if she can avail herself of it without incurring excessive 

costs. But in the case of a Muslim pupil, the cost of complying with the 

school uniform regulation may be sufficiently high that the 

opportunity is effectively removed. Genuine equality of opportunity, 

on the liberal multiculturalist conception, requires an exemption to 

permit Muslim female pupils to wear headscarves in school to 

eliminate the excessive costs attached to compliance with the 

uniform regulations. 

A common response to the contention that granting an exemption 

in such cases would amount to unfairly preferential treatment of 

members of the particular group — one that recommended itself to 

the Constitutional Court of South Africa in MEC for Education, 

KwaZulu-Natal and Others v Pillay (Pillay), 6 the most recent occasion 

on which it was confronted with a claim for an accommodation 7 — is 

that when a cultural or religious group constitutes a minority, and so 

is less powerful than majority groups, as Muslims are in South Africa, 

then its members are disadvantaged compared to members of 

majority (or more powerful) groups. Laws and regulations are often 

framed in a way that is consistent with the beliefs and values of the 

dominant, mainstream cultural groups, but not with those of 

vulnerable, minority groups. As a result, members of minority groups 

bear costs in pursuing opportunities with which members of 

mainstream and more powerful groups are not confronted. Since on 

this argument membership of a cultural or religious group is like a 

physical handicap, inasmuch as it is usually an unchosen feature of 

individuals which is in certain respects disadvantaging, exemptions 

5 I have presented this argument in greater detail in P Lenta ‘Muslim headscarves in 

schools and in the workplace’ (2007) 124 South African Law Journal 296 298-299. 

6 

2008 1 SA 474 (CC) (Pillay). 

7 See (n 6 above) paras 44 (‘The norm embodied by the Code is not neutral, but 

enforces mainstream and historically privileged forms of adornment, such as ear 

studs, which also involve the piercing of a body part, at the expense of minority 

and historically excluded forms’), and para 83 (‘many individual communities still 

retain historically unequal power relations or historically skewed population 

groups which may make it more likely that local decisions will infringe on the 

rights of disfavoured groups’).


should be granted to create equality of opportunity for all persons 

regardless of power and influence. 8 

In this paper, I shall comment on Pillay, which concerns the duty 

of school authorities to accommodate religiously and culturally 

expressive clothing and accoutrements — in this case, a nose stud, 

which, it was claimed, expressed the pupil’s Hindu faith and South 

Indian culture, but which contravened school uniform regulations. 9 

The issue of whether a pupil should be permitted to wear a tiny nose 

stud might seem trivial. In fact, however, the politics of dress in 

schools is highly emotive, both in South Africa and elsewhere, because 

items of clothing and adornment often serve as visible markers of 

religious and cultural identity. 10 

It might be imagined that this issue could easily be disposed of 

with reference to the question of whether exemptions should be 

granted to accommodate Muslim headscarves. I have argued 

elsewhere that exemptions from school uniforms should be granted to 

permit Muslim pupils to wear headscarves: in some cases, Muslim 

pupils may sincerely believe that they are under a religious obligation 

8 A version of this argument is made in C Eisgruber & L Sager ‘Equal Regard’ in S 

Feldman (ed) Law and religion (2000) 200. For a similar argument in support of 

religious exemptions, according to which, since members of a majority religion 

will usually protect their own religious practices when enacting laws, but ignore, 

often unintentionally, the burden that otherwise valid laws impose on minority 

religions, exemptions are justified in order to remedy this flaw in the political 

process, see D Laycock ‘Formal, substantive, and disaggregated neutrality toward 

religion’ (1990) 39 DePaul Law Review 993 1014; M McConnell ‘Free exercise 

revisionism and the Smith decision’ (1990) 57 University of Chicago Law Review 

1109 1130-36 and M Nussbaum Liberty of conscience: In defense of America’s 

tradition of religious equality (2008) 20-21 116-119. In Pillay, Langa CJ drew an 

analogy between membership of a minority religious or cultural group and 

disability (n 6 above, 74). 

9 For useful discussions of the right to freedom of religion and the ways in which it 

has been applied in South Africa, see P Farlam ‘Freedom of religion, conscience, 

thought and belief’ in S Woolman et al (eds) Constitutional law of South Africa 

(2nd Edition, OS, 2003) ch 41, S Woolman ‘Community rights: Language, culture 

and religion’ in S Woolman et al (eds) Constitutional law of South Africa (2nd 

Edition, OS, 2003) ch 58, D Meyerson ‘Religion and the South African constitution’ 

in P Radan, D Meyerson & R Croucher (eds) Law and religion (2005) ch 5, G J 

Pienaar ‘The effect of equality and human dignity on the right to religious 

freedom’ (2003) 66 Tydskrif vir Hedendaagse Romeins-Hollandse Reg 579, I Currie 

& J de Waal The bill of rights handbook (5th ed 2005) ch 15, N Smith ‘Freedom of 

religion: The right to manifest our beliefs’ (2002) 119 South African Law Journal 

690. Discussions of cultural expression are few, but see, for example, Woolman 

(above), especially at 31-43, S Woolman ‘Defending discrimination: On the 

constitutionality of independent schools that promote a particular, if not 

comprehensive, vision of the good life’ (2007) 22 Stellenbosch Law Review 31, B 

Fleisch & S Woolman ‘On the constitutionality of single-medium public schools’ 23 

South African Journal on Human Rights 34 and I Currie & J de Waal The bill of 

rights handbook (5th ed 2005) ch 28. 

10 As Frantz Fanon observes, ‘immense cultural regions can be grouped together on 

the basis of original, specific techniques of men’s and women’s dress ... [t]he fact 

of belonging to a given cultural group is usually revealed by clothing traditions’ (F 

Fanon, ‘Algeria unveiled’ in Studies in a dying colonialism (1962) 35).


to wear a headscarf and the objectives furthered by having uniforms 

may not be significantly undermined by granting this exemption. 11 

But if headscarves should be exempted, then so too, one might think, 

should Hindu nose studs. 

My purpose in this essay is to show that, although South African 

Hindus of South Indian extraction are, like Muslims, a minority group, 

the facts of Pillay are not as closely analogous as they might first 

appear to the case of a Muslim pupil who seeks an accommodation to 

permit her to wear a headscarf which she sincerely believes to be a 

matter of religious obligation. Furthermore, the ways in which these 

cases differ provide grounds for concern about the Constitutional 

Court’s decision to accommodate nose studs in Pillay. I shall begin by 

providing, as background, an account of the Constitutional Court’s 

jurisprudence in the two cases in which it has previously considered 

claims for accommodations. I shall then focus on the Constitutional 

Court’s reasoning in Pillay. Throughout, I shall contrast the approach 

of the Constitutional Court to claims for accommodation with similar 

claims that have been adjudicated by courts of ultimate jurisdiction 

in three other jurisdictions: the United States, the United Kingdom 

and Canada. 

2 Corporal punishment and cannabis 

In Christian Education South Africa v Minister of Justice (Christian 

Education), 12 the first occasion on which the Constitutional Court was 

confronted with a claim for an accommodation, it unanimously 

affirmed that accommodations are, in principle, justifiable. Sachs J 

aligned the Court with liberal multiculturalist proponents of 

exemptions by stating that the granting of religious exemptions 

‘would not be unfair to anyone else who did not hold those views’ 13 

and that 

the State should, wherever reasonably possible, seek to avoid putting 

believers to extremely painful and intensely burdensome choices of 

either being true to their faith or else respectful of the law 14 

provided that religious practices do not violate the rights included in 

11 This is the argument of Lenta (n 5 above). 

12 2000 4 SA 757 (CC). 

13 

Christian Education (n 12 above) para 42. 

14 Christian Education (n 12 above) para 35.


the Bill of Rights. 15 

The Constitutional Court nevertheless declined to grant an 

exemption to legislation proscribing corporal punishment in schools to 

permit teachers to inflict corporal punishment in Christian schools. 

The exemption was sought on the grounds that verses in chapters 9, 

22 and 23 of the Book of Proverbs command the corporal punishment 

of children at school. Sachs J argued that the right to dignity, coupled 

with the right not to be subjected to cruel and unusual, inhuman or 

degrading punishment, placed a duty on the state to ‘reduce violence 

in public and private life’, 16 particularly in view of the state’s resort 

to violence to counter protests by children during apartheid and the 

high incidence of child abuse in contemporary South Africa. 

The Constitutional Court’s decision in Christian Education is, I 

think, correct. 17 In a liberal democracy, the state is the ultimate 

guardian and arbiter of children’s interests; it has a duty to protect 

15 The granting of religious accommodations is, of course, far from uncontroversial. 

J Locke argues forcefully in favour of state neutrality, inasmuch as the state must 

not prefer particular religions over others, yet he does not endorse the practice 

of granting exemptions, insisting instead that ‘the private judgement of any 

person conceiving a law enacted in political matters, for the public good, does 

not take away the obligation of that law, nor deserve a dispensation’ (J Locke ‘A 

letter concerning toleration’ in I Shapiro (ed) Two treatises of government and a 

letter concerning toleration (2003) [1689] 243). Locke’s position is that all 

people, regardless of their affiliations, must comply with generally applicable 

laws. If individuals’ religious beliefs require them to disobey the law, they should 

accept the punishment prescribed by law. Thus, for Locke, if generally applicable 

laws are neutral in the sense that they do not deliberately target religion, or a 

particular religion, they should be construed as fair. Nowadays, the principal 

objection to granting religious exemptions is that to do so is unfair to those who 

might also wish to be excused from otherwise legitimate laws that burden 

practices motivated by secular moral beliefs. See, for example, FM Gedicks ‘An 

unfirm foundation: The regrettable indefensibility of religious exemptions’ (1998) 

20 University of Arkansas Little Rock Law Journal 555. Some defenders of the 

approach taken in J Rawls Political liberalism (1993) may be tempted to argue 

that the principle of state neutrality, according to which the state should not 

prefer one religious viewpoint over another, nor religious beliefs over secular 

beliefs, renders the granting of religious exemptions unjustifiable. On this view, 

exempting religious believers immunises religion in a way that advances it in 

comparison to equally heartfelt secular reasons for claiming exemptions. 

Accommodations for religious beliefs, some Rawlsians may argue, amount to 

favourable treatment to which religious believers are not entitled. A common 

response to this objection to mandatory accommodations — one that is implicit in 

Sachs J’s judgment — is that, if a generally applicable law puts members of a 

religious or cultural group (especially a minority group) at a disadvantage relative 

to others, refusing the exemption cannot be justified on the basis that the law 

provides formal equality: justice requires that exemptions be granted in 

appropriate cases to eliminate or reduce the burden as far as possible. 

16 Christian Education (n 12 above) para 47. 

17 

For a more detailed commentary, see P Lenta ‘Religious liberty and cultural 

accommodation’ (2005) 122 South African Law Journal 352 363-371.


children against what it deems to be abusive practices. 18 One of the 

most fundamental liberal rights is to be free from physical assault. 

Most people think that children have this right. 19 Since there is 

considerable evidence that corporal punishment is psychologically 

harmful to pupils, then, even though there may be countervailing 

evidence indicating that mild and infrequent corporal punishment has 

an improving effect on children, the state is entitled to proscribe it in 

schools. Once it does so, there is no room for an exemption for 

Christian schools, since corporal punishment inflicted at these schools 

will not be rendered non-abusive because it is a practice underpinned 

by religious beliefs. To grant an exemption would be to frustrate the 

purpose behind the legislation, which is not to reduce the incidence 

of corporal punishment in school, but to prevent every instance of an 

abusive practice. 

The Court’s decision in Christian Education is consistent with that 

of the UK House of Lords in R (Williamson) v Secretary of State for 

Education and Employment (Williamson), 20 the facts of which were 

virtually identical, except that whereas Sachs J emphasised the duty 

on the state to reduce the incidence of physical violence because of 

South Africa’s circumstances, Lord Nicholls of Birkenhead took the 

view that deference to parliament on an issue of ‘broad social policy’ 

was appropriate. 21 Baroness Hale of Richmond agreed in substance 

with Lord Nicholls, adding that to ban only corporal punishment that 

infringed against children’s right not to suffer abusive punishment 

would present ‘difficult problems of definition, demarcation and 

enforcement’. 22 

18 Although there is a presumption in favour of permitting parents to raise their 

children in accordance with their religious beliefs, the ultimate responsibility of 

the state for the welfare of children may require it to overrule the religious 

beliefs of parents. See, for example, Hay v B & Others 2003 3 SA 492 (WLD), in 

which parents of a minor opposed a doctor’s application to conduct an emergency 

blood transfusion to save the life of their child principally on the grounds that 

receipt of a blood transfusion was contrary to their religious beliefs. Jajbhay J 

held that the parents could not insist that their child should be denied a blood 

transfusion if the child’s survival, in the opinion of a doctor, depended on it. 

Jajbhay J is absolutely correct: adults may reject life-saving medical treatment 

for themselves, but they may not condemn their children to death on the basis of 

their religious beliefs. A similar position was adopted in the UK in Re O (A Minor) 

(Medical Treatment) (1993) 2 FLR 149; Re R (A Minor) (Blood Transfusion) (1993) 2 

FLR 757. The Canadian Supreme Court arrived at this conclusion in B(R) v 

Children’s Aid Society of Metropolitan Toronto (1995) 122 DLR (4th) 1 (SCC), in 

which it upheld a lower court order to remove a baby from the custody of its 

Jehovah’s Witness parents temporarily so that it could receive life-saving medical 

treatment, including a blood transfusion. In the US, judges have readily ordered 

treatment over the parents’ wishes. See Matter of Hamilton, 657 SW 425, 429 

(Tenn Ct App 1983); Custody of a Minor, 393 NE2d 836 (Mass 1979). 

19 See C McKinnon Toleration: A critical introduction (2006) 107. 

20 2005 2 AC 246 (HL). 

21 

n 20 above, para 51. 



In Prince v President of the Law Society of the Cape of Good Hope 

(Prince), 23 a narrow majority of the Constitutional Court refused to 

grant to Rastafarians an exemption from legislation forbidding the 

possession and use of cannabis, despite an existing permit system for 

administering medical exemptions. The majority denied the claim 

principally on the grounds of the practical difficulty of policing the 

exemption, and because South Africa’s obligations under 

international law required a ban to be uniformly enforced. 24 

The applicant, a Rastafarian attorney, applied for an exemption 

to statutes which proscribe the possession and use of cannabis, on the 

grounds that these Acts prevent an activity central to the sacramental 

activities of his religion. The majority of the Court decided that 

although medical exemptions from the Act could be effectively 

controlled and administered, the same would not be true of a 

religious exemption for Rastafarians. Since Rastafarian use of 

cannabis was not restricted to small quantities at religious 

ceremonies, but was consumed at home and on social occasions, there 

would be no way for enforcement officials to distinguish between 

authentic religious consumption and consumption for recreational 

reasons. Moreover, the majority decided, South Africa’s blanket ban 

on the use and possession of cannabis was intended to meet its 

obligations under international law. There would, in addition, be 

financial and administrative difficulties in establishing a permit 

system and this coupled with the fact that the Rastafarian religious 

community is not easily delineated, since its structures are informal, 

would render a permit system incapable of preventing cannabis being 

distributed to non-Rastafarians. The majority, despite aligning itself 

with the approach of the minority of the US Supreme Court in 

Employment Division v Smith (Smith), 25 according to whom religious 

practices could only be subordinated to a general governmental 

interest if the state could show a ‘compelling state interest’ and that 

the law was the ‘least restrictive means’ of serving that interest, 26 

held that a uniform ban on a drug such a cannabis, which is widely 

used for recreational purposes and in which there is significant trade, 

was justified. 

The dissenting judgements of Ngcobo and Sachs JJ took the view 

that a partial rather than full exemption (as sought by the applicant) 

should be granted to enable Rastafarians to use small quantities of 

cannabis for sacramental purposes on religious occasions. A limited 

exemption, Sachs J held, would secure for the Rastafari ‘a modest but 

23 

24 

25 

2002 2 SA 794 (CC). 

For a more detailed analysis, see Lenta (n 17 above) 371-375. 

Employment Division, Department of Human Resources of Oregon v Smith 494 US 

26 

872 (1990). 

Prince (n 23 above) para 122.


meaningful measure of dignity and recognition’ and was required 

because ‘the Constitution requires the state to walk the extra mile’. 27 

Against the majority, Ngcobo and Sachs JJ contended that a permit 

system, coupled with administrative guidelines, could be instituted to 

ensure that the terms of a limited exemption would be effectively 

enforced. Ngcobo J argued that an exemption would not violate South 

Africa’s obligations under international law since provisions to the 

effect that such measures were subject to each party’s constitutional 

principles and limitations were included in international protocols and 

conventions binding on South Africa. If the South African Constitution 

is interpreted by the Court as requiring an exemption, it would not fall 

foul of international conventions. Sachs J similarly claimed that South 

Africa’s international obligations do not rule out an exemption for 

religious purposes. 28 

The majority’s failure to take seriously a limited exemption for an 

act of worship central to faith that would use administrative 

procedures similar to medical exemptions is, I think, regrettable. The 

burden imposed by the relevant legislation on Rastafarians is severe: 

sacramental and liturgical practices are at the core of religion and 

forbidding these practices forces Rastafarians either to violate the 

tenets of their religion or become outlaws. This appears 

discriminatory since it is unthinkable that major faiths should have to 

face this choice. The use of wine in the Catholic Mass is unlikely ever 

to be forbidden. 

On the issue of enforcement, screening devices might have been 

employed to separate genuine Rastafarians wishing to use cannabis 

for purposes of worship from others seeking to smoke cannabis for 

recreational purposes. The terms of the exemption proposed by the 

minority were consistent with dispensing cannabis in small quantities 

directly to Rastafarian priests for distribution restricted to individuals 

participating in church services, the whole to be overseen by state 

officials. 29 People masquerading as Rastafarian worshippers would 

have to go to great lengths to pass themselves off as believers before 

other believers and state officials. They would be required, at 

considerable inconvenience to themselves — which would in most 

cases, one imagines, outweigh the pleasure they would otherwise 

derive from smoking — to participate in the ceremony. It is 

implausible to think that a limited exemption such as this would 

promote the illegal trafficking of drugs any more than would a 

medical exemption. I concede, however, that justifying the limited 

27 

Prince (n 23 above) paras 148-9. 

28 Prince (n 23 above) para 164. 

29 The granting of this limited exemption might have had an economic cost. It might 

have been appropriate to ask Rastafarian worshippers to bear at least part of this 

cost.


exemption might depend on the plausibility of the majority’s 

interpretation of South Africa’s obligations under international 

conventions. 

By aligning itself with the ‘compelling interest test’ established by 

the US Supreme Court in Sherbert v Verner 30 and Wisconsin v Yoder 

(Yoder), 31 the Constitutional Court affirmed its willingness to grant 

exemptions in principle. 32 In doing so, it offered greater protection 

for religious liberty than US Supreme Court did in Smith, in which the 

majority refused to grant an exemption from a criminal prohibition to 

permit Native American Indians to ingest as an element of their 

worship, peyote, a psychotropic drug. The majority in Smith, 

abandoning the ‘compelling interest’ standard (except for 

unemployment compensation cases and ‘hybrid’ cases involving two 

constitutional rights), held that if a law is generally valid, the 

government may apply it against religious claimants without having to 

show more than that the law has a rational basis. Provided the state 

can show this, there is no constitutional requirement to grant 

religious groups exemptions from facially neutral laws which do not 

have the purpose (though they may have the incidental effect) of 

burdening the practice of religion. As Scalia J put it, the right to 

freedom of religion ‘does not relieve an individual of the obligation to 

comply with a “valid and neutral law of general applicability”’. 33 

In response to Smith, the US Congress, at the instigation of a large 

number of religious groups, enacted the Religious Freedom 

Restoration Act (‘RFRA’). The RFRA was intended to restore the 

position prior to Smith relating to the granting of religious 

exemptions, according to which, as the Act states 

Government may not substantially burden a person’s exercise of religion 

even if the burden results from a rule of general applicability … [unless] 

it demonstrates that the application of the burden to the person (1) is in 

furtherance of a compelling government interest; and (2) is the least 

restrictive means of furthering that ... interest. 34 

In City of Boerne v Flores, Archbishop of San Antonio et al, 35 the US 

Supreme Court declared the RFRA invalid as it applies to states and 

30 374 US 398 (1963). 

31 

406 US 205 (1972). 

32 Prince (n 23 above) para 128. 

33 Smith (n 25 above) p 879. 

34 

Religious Freedom Restoration Act, 42 USC §§ 2000bb-1-2000bb-4 (2004). 

35 521 US 507 (1997).


localities, but not as it applies to the federal government, for which 

purposes it remains valid. 36 

In a recent US Supreme Court decision, Gonzalez, Attorney 

General et al v O Centro Espírita Beneficente União Do Vegetal et al 

(Gonzalez), 37 members of a church, the Christian Spiritists, wished to 

import hoasca, a tea containing a federally proscribed hallucinogen, 

which, it claimed, facilitated communion. The Supreme Court 

dismissed the Government’s submission that it had a compelling 

interest in the uniform application of the law to the extent that no 

exemption could be granted to accommodate the sect’s sacramental 

use of hoasca. The Court held that the government’s failure to grant 

an exemption to the claimants was inconsistent with their rights 

under the RFRA. 

Although the Supreme Court in Gonzalez adopted an approach 

similar to that employed in Prince, there are two important respects 

in which the approaches of the two courts differ. 38 First, the majority 

in Prince gave greater weight to a need for uniform enforcement if 

the ban on the use and possession was to be effectively policed. 

Second, the US Supreme Court ruled that the fact that hoasca was 

covered by the Convention on Psychotropic Substances, did ‘not 

automatically mean’ that the federal law furthered a compelling 

interest that outweighed the claim for an exemption under the RFRA. 

Despite acknowledging the importance of ‘honouring international 

obligations and of maintaining the leadership position of the United 

States in the war on drugs’, the invocation of these general interests 

alone, the Supreme Court decided, was ‘not enough’ in the absence 

of evidence addressing the international consequences of granting an 

exemption. 39 By contrast, the majority of the Constitutional Court 

attached greater significance to South Africa’s obligations under 

international law, holding that ‘[t]he use made of cannabis by 

Rastafari cannot in the circumstances be sanctioned without 

36 The American Indian Religious Freedom Act Amendments were enacted by the US 

Congress in 1994 to permit peyote use by Native Americans for the purposes of 

religious worship and many states enacted legislation to exempt Native American 

Indians from existing prohibitions against drug use. This legislation does not 

protect the sacramental use of cannabis by Rastafarians, however, raising the 

issue of whether the 1994 Amendments constitute unfair discrimination. This 

issue arose in the Kansas Court of Appeals in State v McBride 955 P2d 133 (Kan Ct 

App 1998), in which the court distinguished between the religious use of the two 

substances on the following grounds: peyote is used in small quantities and is 

restricted to religious ceremonies by Native American Indians; recreational use 

and abuse of peyote is rarer than is the case with cannabis, and the state in the 

US has a special duty to respect the political and cultural integrity of Native 

Americans. See B Taylor ‘Kansas denies religion-based defense to Rastafarians on 

marijuana charges’ (1998) 38 Washburn Law Journal 307. 

37 No 04-1084 (2006). 

38 See P Edge ‘Religious drug use in England, South Africa and the United States of 

America’ (2006) 1 Religion and Human Rights 165 176. 

39 Gonzalez (n 37 above) 17-18.


impairing the state’s ability ... to honour its international obligation 

to do so’. 40 

Do these differences in approach indicate that the US Supreme 

Court in Gonzalez provided greater protection to freedom of religion 

than the Constitutional Court did in Prince? Before concluding that it 

did, it is important to recognise that the facts of the two cases are 

different. Gonzalez deals with a drug that is not widely used: the 

market for hoasca is small and there had not in the past been practical 

difficulties in policing the ban on recreational use of hoasca. By 

contrast, as the majority in Prince noted, ‘cannabis, unlike peyote, is 

a drug in which there is substantial illicit trade’. 41 Unlike hoasca, 

cannabis is widely used beyond the confines of religious worship for 

recreational purposes. For these reasons, an approach such as that 

employed in Gonzalez may well have denied an exemption for 

religious use of cannabis, just as Blackmun J in his dissenting 

judgement in Smith held that although an exemption should be 

granted in the case of peyote, the state could legitimately deny 

claims for religious exemptions involving drugs such as marijuana and 

heroin, the use of which is not restricted to a ceremonial context and 

in which there is ‘significant illegal trade’. 42 Would the majority’s 

approach in Prince commit it to denying an exemption in a case in 

material respects the same as Gonzalez? Few of the difficulties of 

policing and administering an exemption for cannabis are likely to 

occur in the case of an exemption for hoasca. Nevertheless, if the 

Constitutional Court determined that South Africa’s obligations under 

international law required a ban to be uniformly enforced, the view 

of the majority in Prince suggests that the Court might reject a claim 

for an exemption for hoasca too. 

The refusal of the majority of the Constitutional Court in Prince 

to grant an exemption to enable Rastafarians to use cannabis on the 

grounds of international law obligations is consistent with the 

approach of the UK judiciary. In Taylor, 43 later endorsed in 

Andrews, 44 a Rastafarian had been observed by police officers 

approaching a Rastafarian temple. On being searched, he was found 

to be in possession of cannabis, in violation of a criminal prohibition. 

The Rastafarian’s defence consisted in the contention that the 

cannabis was intended to be consumed as part of an act of worship at 

the temple. Accordingly his possession was a manifestation of his 

religion, protected under the right to freedom of religion (Article 9 of 

the European Convention on Human Rights). The Court of Appeal held, 

40 

Prince (n 23 above) para 139. 


42 Gonzalez (n 37 above) 917-8. 

43 

[2001] EWCA Crim 2263; [2002] 1 Cr App R 37. 

44 [2004] EWCA Crim 947.


on the basis of several international conventions whose purpose was 

to restrict the use, possession and trafficking of psychotropic 

narcotics, that whatever the interference with the Rastafarian’s 

freedom of religion, it was justified in order to prevent public health 

and safety dangers arising from cannabis use, possession and supply. 

3 Nose studs, headscarves and daggers 

Pillay concerned a claim by the parent of a Hindu pupil at a state 

school that the failure of the school’s authorities to grant her 

daughter an exemption from the school’s Code of Conduct to permit 

her to wear a nose stud constituted unfair discrimination on the 

grounds of religion and culture. The claim was brought under the 

Promotion of Equality and Unfair Discrimination Act, 45 which prohibits 

unfair discrimination on the grounds of religion and culture. Although 

the claim was not brought under the rights to freedom of religion and 

culture, the Court held that there may be, as in this case, an 

overlap [between the Promotion of Equality and Unfair Discrimination 

Act and the rights to freedom of religion, belief and opinion and to 

participate in cultural life 46 ] … where the discrimination in question 

flows from an interference with a person’s religious or cultural 

practices’. 47 

Pillay clearly represents a development of the Court’s jurisprudence 

on religious liberty and cultural accommodation. 

The Court ruled that the school’s failure to grant the exemption 

was unfairly discriminatory and that the school was required to 

exempt the pupil concerned, Langa CJ, on behalf of the majority, 

deciding that the wearing of a nose stud represents an expression of 

the Hindu religion and of South Indian Tamil culture. The majority 

found that the failure of the school to grant the accommodation was 

unfairly discriminatory against the pupil concerned relative to other 

pupils whose religious beliefs and cultural practices did not bring 

them into conflict with the Code. Langa CJ stated that although the 

school uniform served ‘admirable purposes,’ these objectives would 

not be undermined by granting religious and cultural exemptions. 48 

In a separate judgement in which she agreed with the majority’s 

finding of unfair discrimination, but disagreed in part with its 

45 

46 

Act 4 of 2000 (‘the Equality Act’). 

Section 30 of the Constitution reads: ‘Everyone has the right to use the language 

and participate in the cultural life of their choice, but no one exercising these 

47 

48 

rights may do so in a manner inconsistent with any provision of the Bill of Rights.’ 

Pillay (n 6 above) para 46. 

Pillay (n 6 above) para 101.


reasoning, O’Regan J found that there is no evidence that the wearing 

of a nose stud has religious significance for the claimant, although she 

accepted that it is a form of cultural expression. She agreed with the 

majority that the failure of the school to grant an exemption was 

unfairly discriminatory, but for a different reason: the applicant was 

discriminated against, not relative to those pupils on whom the Code 

imposed no burden of religious or cultural denial, but relative to other 

pupils whom the school had in the past exempted: Hindu pupils had 

been permitted to wear ‘Lakshmi strings’ in honour of the Goddess 

Lakshmi and other pupils had been allowed to wear hide bracelets as 

a mark of respect on the death of a relative. 49 

Both Langa CJ and O’Regan J recognised that exemptions should 

not be made in response to every claim. Claims for exemptions call 

for a proportionality exercise. The strength of the claim must be 

determined. This involves an enquiry into whether, if a religious 

belief is relied upon, it is genuinely a religious belief, or, in the case 

of a cultural practice, that it is genuinely a cultural practice; the 

sincerity of the claimant, and the nature and severity of the burden 

on religious exercise. The strength of the competing interest of the 

state or other relevant institution in the uniform application of the 

rule must equally be considered. In cases where the law in question 

protects fundamental rights of individuals, or serves a compelling 

state objective that would be significantly frustrated by granting an 

exemption, or where the exemption would present insurmountable 

practical difficulties such that it could not be administered 

effectively, it may be that an exemption is ruled out. In the rest of 

this section, I consider the way in which the Constitutional Court 

negotiated each step in this enquiry. 

3.1 Religious belief or cultural practice? 

The claimant was required to show that the practice for which the 

accommodation is claimed is cultural in nature or that it is motivated 

by religious beliefs. The task of determining what beliefs and 

practices are religious — as opposed to secular, moral or personal — is 

a difficult matter, since religion is a highly complex concept. By what 

criteria can religion be defined and the distinction between it and the 

non-religious drawn, given the ineffable nature of religious 

experience and the fact that individuals have shifting relationships 

with religious communities which may themselves be riven with 

profound theological differences? Defining culture, to the extent that 

49 Pillay (n 6 above) para 130.


it differs from religion, is no easier, since ‘cultural forms possess 

inherent indeterminacy’. 50 

On the issue of the cultural significance of the nose stud, Langa 

CJ asserted two claims. First, ‘cultural convictions or practices may 

be as strongly held and as important to those who hold them as 

religious beliefs are [for religious believers]’ since individual identity 

and dignity is bound up with culture. Second, cultural practices ‘will 

differ from person to person within a culture’, since ‘[c]ultures are 

living and contested formations’. 51 Noting the regrettable absence of 

testimony from the pupil herself, Langa CJ found that an admission by 

the school authorities that the nose stud has ‘cultural significance’ 52 

for the pupil was dispositive of the question of whether the practice 

should be considered to be a cultural practice. 

On the issue of whether the nose stud has religious significance, 

Langa CJ referred to the testimony of Dr Rambilass, an expert on the 

Hindu religion, who had given evidence on behalf of the school. Dr 

Rambilass had conceded that the wearing of a nose stud is a cultural 

practice, but denied that the nose stud has independent religious 

significance, being ‘part of the Shringaar which is concerned with 

love, beauty and adornment’. 53 Langa CJ contested this expert 

witness’s conclusion that the wearing of a nose stud is devoid of 

religious significance on three grounds. First, he argued that it is 

difficult to separate Hindu culture and Hindu religion. Secondly, 

‘there are many different sects of Hinduism with different beliefs and 

practices’. 54 Thirdly, Dr Rambillas’s testimony 

self-consciously focused on defining Hindu religion according to the 

specific wording of the Vedic texts rather than on a broader view of 

religion as being informed and even defined by culture, tradition and 

practice. 55 

On the grounds that ‘the borders between culture and religion are 

malleable and that religious belief informs cultural practice and 

cultural practice attains religious significance’, Langa CJ concluded 

that Sunali’s (the pupil in question’s) wearing of a nose stud was ‘an 

expression of both culture and religion’. 56 

Langa CJ’s reasoning in determining that the wearing of a nose 

stud has religious significance may be questioned from several 

50 

L Rosen, quoted in W Fallers Sullivan ‘Judging religion' (1998) 81 Marquette Law 

51 

52 

53 

54 

55 

56 

Review 441 456. 

Pillay (n 6 above) paras 53-54. 



As above. 

As above. 

Pillay (n 6 above) para 60.


perspectives. There was no evidence to suggest that it bore religious 

significance for Sunali beyond the claimant’s assertion that it did. 

Expert testimony indicated otherwise, and Langa CJ’s rejection of Dr 

Rambilass’ testimony concerning the content of the Hindu religion, 

about which he and not Langa CJ is an expert, is illegitimate. As for 

Langa CJ’s contention that it is often difficult to distinguish between 

culture and religion, it is one thing to say that many religious 

practices are in some sense cultural and that many cultural practices 

are religious — both of which propositions are true, since culture and 

religion are inter-imbricated. It is quite another thing to aver that all 

cultural practices are religious, which is false. There is a subset of 

cultural practices that do not carry religious significance, which the 

evidence of Dr Rambilass indicated included the wearing of a nose 

stud. 

That the borders of religion and culture are porous and that 

cultural practices may attain religious significance does not, contra 

Langa CJ, constitute evidence that the nose stud has religious 

significance. It would do so only if all cultural practices had religious 

significance, which is surely not the case. Langa CJ failed to cite 

credible evidence for thinking that nose studs have religious 

significance. Accordingly, O’Regan J’s assessment is more plausible: 

Although the applicant argues that the nose stud was part of religious 

practice, it is clear that its primary significance to her family arises from 

its associative meaning as part of their cultural identity, rather than 

personal religious beliefs. This is consistent with Dr Rambilass’s evidence 

that wearing a nose-stud is not part of Hindu religion’. 57 

O’Regan J’s approach to determining whether a practice is cultural is 

to be preferred to Langa CJ’s. She contended that since cultural 

practices are associative and that the right to cultural life is to be 

practiced as a member of a community and not primarily a question 

of sincere, but personal belief, Langa CJ’s ‘individualised and 

subjective approach to what constitutes culture is faulty’. Rather ‘in 

probing whether a particular practice is a cultural practice, some 

understanding of what the cultural community considers to be a 

cultural practice is important’. 58 O’Regan J elaborated: 

57 Pillay (n 6 above) para 62. O’Regan J’s conclusion — ‘that the applicant has 

established that the wearing of a nose stud is a matter of associative cultural 

significance, which was a matter of personal choice at least for the learner in this 

case, but that it is not part of a religious or personal belief of the applicant that 

it is necessary to wear the stud as part of her religious beliefs’ — is open to the 

interpretation that she accepted that the practice has non-mandatory religious 

significance for Sunali. But this would be inconsistent with O’Regan J’s failure to 

challenge, and apparent acceptance of, Dr Rambilass’s statement that wearing a 

nose stud is not a religious practice. 



It will not ordinarily be sufficient for a person who needs to establish 

that he or she has been discriminated against on the grounds of culture 

to establish that it is his or her sincerely held belief that it is a cultural 

practice, or that his or her family has a tradition of pursuing this 

practice. The person will need to show that the practice that has been 

affected relates to a practice that is shared in a broader community of 

which he or she is a member and from which he or she draws meaning. 59 

O’Regan J’s distinction between religion and culture has merit. 

Although most religions are also cultures, a religion may be 

constituted by the beliefs and practices of a single individual. A 

religion need not be constituted by shared practices and beliefs (even 

though it will usually be). By contrast, culture signifies the range of 

beliefs and assumptions held in common by members of a group, as 

well as the ensemble of customs and practices (fluid and mobile to be 

sure) that reflect the way members do things in common: modes of 

dress and decoration, types of circulation and social activity, ways of 

working and so on. Thus, where it is claimed that a practice is a 

cultural practice, we can examine the culture concerned to 

determine whether members engage in it in common with other 

members. We are not necessarily in a position to do this when it is 

claimed that a particular practice is religious in nature, since the 

beliefs and practices of a single individual may be sufficient to 

constitute a religion. 

O’Regan J noted that there are a number of women of South 

Indian heritage who wear a nose stud because it identifies them as 

members of that community and connects them with their community 

and with former generations. She rightly considered this sufficient 

evidence to conclude that the wearing of a nose stud by women of 

South Indian extraction is a matter of ‘associative cultural 

significance’. 60 She nevertheless noted, again correctly, that the 

practice is optional (that there is no cultural norm or obligation 

putting pressure on Sunali to wear a nose stud is evidenced by the fact 

that her sisters opted not to do so). 

No evidence was adduced by the applicant to show that the 

wearing of the nose stud by her daughter was religious; indeed, it is 

clear that the wearing of a nose stud is, for Hindu women, an 

expression of cultural identity that is entirely optional and is not a 

religious practice. Nevertheless, that the practice has cultural 

significance is sufficient to bring it within the ambit of the Equality 

Act. 

59 




3.2 Sincerity 

In order to be successful, the claimant must be sincere in her 

assertion that her opposition to conforming to the standard 

requirements rests on religious or cultural grounds: she must 

demonstrate that her beliefs are sincerely held (‘neither fictitious, 

nor capricious ... [nor] an artifice’) 61 or that she is sincerely 

expressing her cultural identity. Although the enquiry into sincerity 

precludes determining the truth or legitimacy of religious claims, it is 

important because charlatans may be tempted, for tactical purposes, 

to misrepresent as religiously or culturally motivated their insistence 

on acting contrary to the rules. The sincerity enquiry 

attempts to ensure that exemptions are not granted to individuals who 

frame their opposition ... in religious [or cultural] terms in order to be 

afforded the legal remedy desired when in fact the religious beliefs are 

not sincerely held because the opposition stems from secular 

convictions. 62 

A claimant is not acting on sincerely held religious beliefs or 

expressing her cultural identity simply because she says she is. The 

claimant must demonstrate to the satisfaction of the court that the 

asserted beliefs are sincerely held or that the practice in question is 

an expression of cultural identity. For Langa CJ, the fact that Sunali 

refused to remove the nose stud after initially agreeing to do so — 

equivocation that is most plausibly explained, he stated, ‘as a young 

woman uncertain about the consequences of standing up against the 

imposing authority of the School’s headmistress’ — ‘points to the 

conclusion that Sunali held a sincere belief that the nose stud was part 

of her religion and culture’. 63 In fact, however, Sunali’s conduct was 

also consistent with the actions of a pupil who was initially prepared 

to forgo wearing the nose stud, but whose parent resolved to defy the 

school’s authorities in order that her daughter could continue to 

engage in a practice expressive of cultural identity only. The sincerity 

of the claimant’s averment that the practice has religious significance 

is questionable, particularly in view of Dr Rambilass’s evidence that 

Hindu women do not regard it as a religiously significant practice. 

Even allowing that different believers within the same faith may hold 

different beliefs and attach different significance to the same 

61 Iacobucci J in Syndicat Northcrest v Amselem (2004) 2 SCR 551 para 52 

(Amselem). 

62 C Kraus ‘Religious exemptions — applicability to vegetarian beliefs’ (2001) 30 

Hofstra Law Review 197 215. As the US District Court observed in US v Kuch 288 F 

Supp 439 443 claimants ‘must not be permitted the special freedoms this 

sanctuary may provide merely by adopting religious nomenclature and cynically 

using it as a shield to protect them when participating in antisocial conduct that 

otherwise stands condemned’. 



practice, the applicant provided no convincing evidence that Sunali’s 

wearing of a nose stud was motivated by sincerely held religious 

belief. In the absence of any evidence that the wearing of a nose stud 

is motivated by a religious belief (such as the scriptural evidence 

provided by the claimants in Christian Education or the verses of the 

Quran referred to by Muslim women claiming exemptions to permit 

them to wear headscarves) the impression that remains is that 

Sunali’s wearing of a nose stud only masqueraded as a religious 

practice for the purposes of securing an exemption. 

This is not to insist on the semiotic neutrality of a nose stud for a 

pupil of South Indian descent, nor is it to dispute the sincerity of the 

claimant’s assertion that Sunali wished to wear the nose stud to 

express her South Indian culture. Far from it: as both Langa CJ and 

O’Regan J rightly observed, it is a practice that is clearly invested 

with cultural significance: an expression of the pupil’s cultural 

identity. Although the nose stud is an adornment worn by women out 

of personal choice within South Indian culture, there is no reason to 

doubt that Sunali wore it as a fashion accessory that expressed her 

cultural identity. 

3.3 Claimant’s burden versus state interest 

Langa CJ observed that the constitution places a duty of ‘reasonable 

accommodation’ on the institution from which an exemption is 

claimed. How should reasonableness be construed in this context? 

According to Langa CJ, the institution must make more than mere 

negligible effort ‘to enable those outside the “mainstream” to swim 

freely in its waters’, but determining the extent of the hardship it 

must suffer to do so will involve a ‘contextual’ determination, ‘an 

exercise in proportionality that will depend intimately on the facts’. 64 

In practice, courts must assess the burden imposed on the claimant’s 

religious freedom or cultural practice and whatever interest is served 

by refusing the exemption, and strike an appropriate balance 

between the two. The claimant should only succeed if she suffers a 

substantial burden on the exercise of her religion or culture and if the 

state or institution’s interest in denying her an exemption is not 

strong. Even if the burden on the claimant is substantial, the claim 

should not succeed if the government’s interest in denying the 

exemption is strong: human sacrifice may be central to a claimant’s 

religion, but it should not be accepted as a defence against a charge 



of murder. Likewise the claim should fail if it affects religious or 

cultural activities only marginally. 65 

The claimant was required to show that, as a result of the 

school’s code, Sunali was placed under a substantial burden. A 

showing of this kind is easier where the religious or cultural practice 

for which the claimant is seeking protection is mandatory rather than 

optional: where there is ‘a Hobson’s choice between observance of 

their faith and adherence to the law’ 66 as there was in Christian 

Education. In Pillay, by contrast, the applicant conceded that her 

wearing a nose stud was not required by the tenets of her faith, but 

was instead a voluntary practice. The Constitutional Court held — 

consistently with the Canadian Supreme Court’s ruling in Amselem 67 

(affirmed, though the Constitutional Court does not say so, by the 

House of Lords in Williamson 68 ) — that it is not essential to the success 

of a claim for a religious exemption that the religious practice be 

required by the claimant’s religious or cultural beliefs, but only that 

it be an expression of those beliefs. That is so, stated Langa CJ, 

because the Constitution is committed to ‘affirming diversity’ and 

‘differentiating between mandatory and voluntary practices does not 

celebrate or affirm diversity but merely tolerates it’. 69 

The Constitutional Court is clearly correct that in order to be 

afforded protection under the right to freedom of religion the belief 

or practice need not be mandatory, although this has nothing to do 

with celebrating rather than tolerating diversity. To be successful, 

the claimant must demonstrate not that the practice for which she 

seeks protection is compelled by her religion, but that the regulation 

imposes a substantial burden on the practice of her religion. A trivial 

impact imposed by a law on an individual would not be sufficient to 

qualify for an exemption. Even if religious practices are not 

mandatory, being prevented from engaging in such practices may be 

substantially burdensome. Kent Greenawalt illustrates this point with 

an example: individual A might not consider wearing a cross to be 

mandatory, yet could regard wearing a cross as ‘an important symbol 

of witness and commitment and a great aid to devotion’. 70 Were a 

school to forbid all jewellery, including crosses, A’s convictions should 

65 See Lyng v Northwest Indian Cemetery Protection Association 374 US 398 (1963) 

at 475, in which Brennan J, dissenting, argued that a trivial interference with 

religious practice was insufficient; the claimants were required to show the 

‘centrality’ of the affected practices and the presence of a ‘substantial and 

realistic threat of frustrating their religious practice’. 

66 


67 n 61 above, paras 67-8. 


69 


70 Greenawalt (n 4 above) 210.


be sufficient to vault her over the threshold of substantial burden. 71 

Of course, not all voluntary practices are similarly justifiable. 

Consider individual B, who wears a cross without attaching great 

religious significance to it, though aware that it represents for 

Christians the crucifixion of Christ. A rule forbidding B from wearing 

jewellery would not constitute a substantial burden. 

Even allowing that proscribing voluntary practices may be 

burdensome, interfering with obligatory practices is likely to be more 

burdensome. That is so, because forgoing a practice mandated by the 

tenets of religion (as wearing a headscarf is for some Muslim women) 

will be intensely injurious to them, because what is being demanded 

of them is that they violate transcendental religious commands, on 

which the favour of their deity or sanctions, salvation or damnation, 

may depend. A clash between religious commandments and civil laws 

or regulations situates believers between competing sources of 

authority, a predicament particularly grave given ‘the widespread 

sense that one’s religious obligations are more ultimate than those of 

the social order and should take priority if the two come into 

conflict’. 72 Voluntary practices, precisely because they are not 

commanded, are less damagingly relinquished, even if forbidding such 

obstacles may be substantially burdensome. 

Langa CJ appeared not to appreciate this when he asserted, ‘that 

we choose voluntarily rather than through a feeling of obligation only 

enhances the significance of a practice to our autonomy, our identity 

and our dignity,’ to which he added that ‘it may be even more vital 

to protect non-obligatory cultural practices’. 73 But this obscures the 

fact that the costs to a believer of forgoing an optional practice will 

almost always be lower than the costs that a believer pays in deciding 

not to engage in a practice that the tenets of his faith mandates, 

because in the case of an optional practice, the individual is not 

required to make the agonising choice between competing sources of 

authority. 

71 

See, by way of illustration, Sasnett v Sullivan 908 F Supp 1429 (WD Wis 1995), 

aff’d, 91 F 3d 1018 (7th Cir 1996), vacated and remanded, 521 US 1114 (1997), in 

which a prisoner challenged under the RFRA regulations limiting the wearing of all 

religious jewellery except wedding bands. Prison officials argued that the 

exercise of a prisoner’s religion was substantially burdened only if the practice 

was required by their religion, which the wearing of crucifixes was not. Posner J 

held that the question was whether the burdened practice was motivated by a 

sincere religious belief, not whether it was ecclesiastically mandated. He found 

that the regulation forbidding the wearing of crosses substantially burdened the 

plaintiff’s religious liberty. Winnifred Fallers Sullivan quotes one of the plaintiffs: 

‘There’s nothing saying you have to wear a cross, but it brings a person closer to 

God’ (Sullivan (n 50 above) 451). 

72 

Greenawalt (n 4 above) 439. 

73 Pillay (n 6 above) paras 64 & 66.


How should the magnitude of the burden imposed on the claimant 

be measured? Langa CJ decided that the failure of the school to grant 

an exemption constituted a ‘significant infringement’ of the 

claimant’s religious and cultural identity on the basis of what he 

referred to as a ‘subjective investigation’. 74 The issue, he stated, is 

whether the pupil considered the nose stud central to her religion or 

cultural identity, not whether other members of the South Indian 

Tamil Hindu community do so (since the status of a practice within a 

particular group may be disputed). Langa CJ acknowledged that ‘the 

Court can properly look at a range of evidence including evidence of 

the objective centrality of the practice to the community at large’. 75 

Yet it is striking that he, unlike O’Regan J, did not consider the way 

in which the practice is viewed by other members of the South Indian 

Hindu community. Instead he considered the pupil’s beliefs and 

attitudes concerning the practice, which he inferred from her 

conduct. Where Langa CJ’s reasoning went awry was in assuming that 

behaviour is an accurate indicator of beliefs. In fact, conduct is not 

reliable evidence of beliefs: it is usually the case that multiple beliefs 

and motivations are consistent with any individual’s activity. 

Langa CJ appears to have accepted uncritically the applicant’s 

assertion that she was placed under a substantial burden and refers 

to the pupil’s insistence on wearing the nose stud under the pressure 

of threatened disciplinary action as evidence of this. 76 But in fact, all 

that the pupil’s conduct evidences is a strong commitment to wearing 

the nose stud, which could reflect a substantial burden but could 

equally be motivated by her obedience to her mother, who appears to 

have engaged in a power struggle with the school authorities. The 

difficulty with Langa CJ’s enquiry into burden is that he seemed to 

treat the conclusory allegation on the part of the claimant regarding 

burden as dispositive. He was too ready to interpret the pupil’s 

conduct as indicative of the strength of her religious and cultural 

commitment to the practice, without allowing that a strong 

commitment to persisting with the practice is not necessarily 

generated by a substantial burden. 

The school claimed that the burden, if any, was slight since the 

pupil could wear the nose stud out of school hours. Langa J rejected 

this claim on the grounds that ‘[t]he practice to which Sunali adheres 

is that once she inserts the nose stud, she must never remove it’. 77 

However, he referred to no evidence in support of the claim that the 

practice of wearing a nose stud adopted by certain members of the 

74 Pillay (n 6 above) paras 85 & 88. 

75 Pillay (n 6 above) para 88. 

76 




South Indian community obliges them to wear the nose stud at all 

times. 

The school claimed that the burden placed on the pupil was slight 

for the additional reason that she could have transferred to another 

school that would have allowed her to wear a nose stud. Langa CJ 

rejected this argument on the grounds that ‘permitting it only when 

no other option remains’ 78 was inconsistent with South Africa’s 

‘constitutional project which not only affirms diversity, but promotes 

and celebrates it’. The effect of requiring the pupil to relocate to 

another school to express her cultural identity, Langa CJ asserted, 

‘would be to marginalise religions and cultures’. 79 

If in Pillay the burden on the pupil did not, contra Langa CJ, 

appear to be particularly substantial, neither was the school’s 

interest in enforcing the school uniform without exception strong. It 

is often argued that uniforms promote crucial educational interests: 

minimising external differences between pupils of different social 

classes; discouraging competitive fashions; promoting school spirit 

and encouraging discipline. 80 At issue was the empirical question of 

whether the accommodation sought by the claimant would 

significantly impair the achievement of these educational objectives. 

The school did not offer evidence on this point and it is difficult to see 

how it could have done so. In fact, as Langa CJ noted, it is difficult to 

see how granting an exemption to permit the wearing of a nose stud 

would interfere substantially with the effective running of a school 

and with the purposes the school uniform was designed to further. 81 

This conclusion is consistent with that of the UK House of Lords in 

Mandla v Dowell Lee, 82 decided under the Race Relations Act of 1976, 

in which a headmaster had refused to admit a Sikh pupil to a private 

school on the grounds that the pupil insisted on wearing a turban, 

mandated by his faith, in violation of the school uniform. Lord Fraser 



80 

There is some debate about whether school uniforms are successful in fostering 

these educational objectives, although (contested) studies conducted by 

American researchers have found school uniforms increase discipline, foster 

school morale, reduce unhealthy competition amongst pupils and so forth. For a 

critical review of the literature, see D Brunsma School uniforms: A critical review 

of the literature (2002). The United States Department of Education’s Manual on 

school uniforms, (http://www.ed.gov/updates/uniforms.html (accessed: 6 June 

2007)) begins with the sentence, ‘A safe and disciplined learning environment is 

the first requirement of a good school’ and under the heading ‘Treat school 

uniforms as part of an overall safety program’ the Manual states as follows: 

‘Uniforms by themselves cannot solve all of the problems of school discipline, but 

they can be one positive contributing factor to discipline and safety’. 

81 


82 [1983] 2 AC 548.


likewise dismissed the headmaster’s argument that an exemption 

would frustrate the objectives furthered by the uniform. 

In Pillay, the available evidence suggested that the burden 

imposed on the pupil and the interest of the school in denying her an 

exemption were roughly in equipoise. The political philosophy 

underlying the Court’s holding that the severity of the burden placed 

on the pupil outweighed the burden placed on the school is revealed 

in its commitment not only to tolerating difference but to ‘affirm[ing] 

it as one of the primary treasures of our nation’. 83 The Court appears 

to be committed to a ‘politics of difference’ that in certain respects 

resembles that outlined by Iris Marion Young in Justice and the 

Politics of Difference. 84 

Young’s contrast between what she considers the traditional 

liberal approach and the politics of difference is instructive: 

The vision of liberalism as the transcendence of group difference seeks 

to abolish the public and political significance of group difference, 

while retaining and promoting both individual and group diversity in 

private, or nonpolitical, social contexts. 85 

For Young, ‘[g]roups cannot be socially equal unless their specific 

experience, culture and social contributions are publicly affirmed and 

recognised’. 86 She deprecates the ‘the typical liberal approach ... 

which tolerates any behaviour so long as it is kept in private’, 87 

arguing that ‘the private, as traditionally conceived, is what should be 

83 


84 IM Young Justice and the politics of difference (1990). It may be useful to deal 

briefly with the change in the way in which the idea of respect for difference has 

been viewed in South Africa. In a perversion of the contemporary politics of 

difference, the apartheid government justified its policy of creating ‘homelands’, 

an order of ‘plural nations’ in semi-autonomous polities with no economic 

viability, with reference to something like a multicultural respect for difference. 

As a result, opponents of apartheid became suspicious of the idea of respect for 

difference. As R Nixon Homelands, Harlem and Hollywood: South African culture 

and the world beyond (1994) 206 asserts, ‘Since the late 1950s, difference (as 

opposed to unity) has been perceived — for sound historical reasons — as a 

government term. This is so because the South African brand of racial supremacy 

has been couched as a form of sensitivity to the special needs and cultural 

particularities of diverse “peoples”. Attentiveness to difference is thus perceived 

as “apartheid” business, a way of coating state racism with a democratic gloss’. 

Nevertheless, as John and Jean Comaroff ‘Criminal justice, cultural justice: The 

limits of liberalism and the pragmatics of difference in the new South Africa’ 

(1994) 31 American Ethnologist 188 195 observe, there has been in South Africa ‘a 

growing recognition of the gravitas of difference’. In the jurisprudence of the 

Constitutional Court, attentiveness to difference has emerged, particularly in its 

jurisprudence dealing with sexual orientation and religious and cultural 

accommodation, as an integral part of constitutional justice. 

85 n 84 above, 168 (emphasis added). 

86 

n 84 above, 174 (emphasis added). 

87 Young (n 84 above) 161 (emphasis added).


hidden from view, or what cannot be brought out to view. It is 

connected with shame and incompleteness’. 88 

The refusal to permit members of cultural groups to manifest 

their culture publicly is considered by Young and the Constitutional 

Court to devalue members by relegating cultural practices to the 

private sphere. It is from this perspective that we should understand 

Langa CJ’s assertion that ‘the symbolic effect of denying her the right 

to wear it ... sends a message that Sunali, her religion and her culture 

are not welcome’. 89 From the perspective of the politics of 

difference, the school’s refusal to permit Sunali to express her 

cultural identity at school (that is, in public) constitutes an injury to 

her. Young’s depreciation of liberal toleration chimes with O’Regan 

J’s statement that ‘treating people as worthy of equal respect in 

relation to their cultural practices requires more than mere 

tolerance’. 90 

3.4 Discrimination, neutrality and fairness 

Under the Equality Act, the claimant was required to show that the 

school’s Code unfairly discriminated against her on the basis of 

religion or culture. The school argued that in this case there was no 

comparator, no group relative to which Sunali received inferior 

treatment. The location of the appropriate comparator was the 

subject of disagreement between the judgements of Langa CJ and 

O’Regan J. Langa CJ identified the appropriate comparator as ‘those 

learners whose sincere religious or cultural beliefs or practices are not 

compromised by the Code, as compared to those whose beliefs are 

uncompromised’. He found that ‘the Code has a disparate impact’ 91 

on certain religions and cultures, the members of which are 

disproportionately burdened or incommoded relative to others who 

are not members. He took to view that the school’s Code 

is not neutral, but enforces mainstream and historically privileged forms 

of adornment, such as ear studs which also involve the piercing of a body 

part, at the expense of minority and historically excluded forms. 92 

This again is consistent with Young’s ‘politics of difference’, 

according to which 

88 Young (n 84 above) 119. 

89 

Prince (n 6 above) para 85. Langa CJ here paraphrases the Canadian Supreme 

Court in Multani v Commission Scolaire Marguerite-Bourgeoys [2006] SCC 6 para 

79 (Multani): ‘A total prohibition against wearing a kirpan to school undermines 

the value of this religious symbol and sends students the message that some 

religious practices do not merit the same protection as others’. 

90 n 84 above, 156. 

91 


92 As above.


blindness to difference disadvantages groups whose experience, culture 

and socialised capacities differ from those of privileged groups [who] ... 

implicitly define the standards according to which all will be measured. 

Because their privilege involves not recognising those standards as 

culturally and experientially specific, the ideal of common humanity in 

which all can participate without regard to race, gender, religion or 

sexuality poses as neutral and universal. 93 

For O’Regan J, by contrast, 

the correct comparator is those learners who have been afforded an 

exemption to allow them to pursue their cultural or religious practices, 

as against those learners who are denied exemption, like the learner in 

this case. 94 

As indicated above, the school had in the past granted exemptions to 

permit the wearing of red ‘Lakshmi strings’ and of hide bracelets to 

mark respect after a funeral, 95 both of which were ‘associative 

cultural or religious practices’. According to O’Regan J, the unfairness 

lay in the school’s inconsistency in exempting certain practices but 

not others without adequate justification of the differential 

treatment. Whereas Langa CJ asserted that the Code was nonneutral, 

O’Regan J appears to have accepted that the Code was 

neutral: she stated that ‘the Code is entitled to establish neutral rules 

to govern the school uniform’. 96 

Which group is the correct comparator? The answer, I think, is 

that both groups are. Langa CJ was certainly correct that the 

unfairness that justifies the granting of an exemption is present when 

a rule imposes a burden on members of a minority or vulnerable 

cultural or religious group to which others are not subject. On the 

other hand, O’Regan J is right that the granting of exemptions to 

permit certain religious and cultural practices allows those who have 

been denied exemptions to claim that they have been unfairly 

treated. Since both judges found that there had been discrimination, 

does it matter which comparator is to be considered the appropriate 

one? It does: on O’Regan J’s approach, had the school refused to grant 

any exemptions there would have been no discrimination and the 

claim would have failed. By contrast, Langa CJ would have still found 

the rule to be discriminatory and the claim would still have been 

successful. The correct approach would be to hold that both groups 

are appropriate comparators and both can give rise to claims for 

exemptions. 

93 n 84 above, 164. 


95 




Given the disagreement between Langa CJ and O’Regan J about 

whether the Code is neutral, who is right? Langa CJ’s determination 

that the code is non-neutral is preferable, comporting as it 

substantially does with the approach of the minority of the US 

Supreme Court in Goldman v Weinberger (Goldman). 97 In that case, 

an orthodox Jewish rabbi serving as a psychologist in the military 

claimed an exemption from an air force regulation that proscribed the 

wearing of headgear indoors. The military’s uniform dress regulation 

permitted the wearing of rings and the wearing of religious symbols 

underneath military uniforms. This allowed latitude for Christians to 

wear articles with religious significance. The minority in Goldman 

were prepared to grant the exemption in part because ‘the 

accommodation of items Christians were likely to wear contrasted 

disturbingly with the rigidity of the rule that disfavoured Orthodox 

Jews and other minorities’. 98 

The School argued that if an exemption were to be granted, then 

‘some of the girls might feel that it is unfair’. 99 Langa CJ, with whom 

O’Regan J agreed on this point, determined, correctly and 

consistently with the Canadian Supreme Court in its recent Multani 

decision, 100 that once the rule is found to be discriminatory the 

appropriate response to complaints from those not covered by the 

exemption is to explain to them that the Code does not impose a 

burden of cultural denial that is comparable to the burden imposed 

on them in requiring them to forgo a fashion item. 101 

3.5 Deference 

The school raised the argument that deference should be accorded to 

the professional judgements of the school authorities regarding school 

uniforms. Langa CJ responded by conceding that ‘the Court must give 

due weight to the opinion of experts, including school authorities, 

who are particularly knowledgeable in their area’. He insisted, 

however, that courts ‘are best qualified and constitutionally 

mandated’ to answer the question ‘whether the fundamental right to 

equality has been violated, which in turn requires the Court to 

determine what obligations the school bears to accommodate 

diversity reasonably’. 102 ‘The Court cannot abdicate its duty by 

deferring to the school’s view on the requirements of fairness’ since 

97 475 US 503 (1986). 

98 Greenawalt (n 4 above) 165. 


100 n 89 above, para 104. 

101 As O’Regan J observed, ‘A school is an ideal place to educate other learners about 

the difference between fashion and cultural practices, and should an exemption 

for nose-studs be granted, a school would be obliged to furnish such education to 

its learners’ (Pillay (n 6 above) para 172). 



to do so would be to make the school a judge in its own case. 103 Langa 

CJ referred approvingly to the fact that the Code had been ‘devised 

after extensive consultation with parents, educators, staff and 

learners’, but decided that the fact of consultation ‘does not 

immunise the resultant decisions ... from constitutional scrutiny and 

review’ since 

many individual communities still retain historically unequal power 

relations or historically skewed population groups which may make it 

more likely that local decisions will infringe on the rights of disfavoured 

groups. 104 

Langa CJ’s view of judicial deference is correct. 105 A moderate 

degree of deference should be accorded school authorities because 

they possess specialist expertise. The Court should allow ample scope 

for the development of policy and the promulgation of regulations by 

the school administration, recognising that rights can sometimes 

properly give way to contrary public needs, such as the creation of 

conditions for effective education. Yet the Court, as ultimate 

authority on questions of right, should arrive at its own independent 

evaluation, rather than accepting the view of the school authorities. 

Brennan J took this view in Goldman, criticising the majority for 

deferring unduly to the judgement of the military. 106 

3.6 A slippery slope? The end of school uniforms? 

Langa CJ rejected an objection that to grant an exemption in this case 

would result in a ‘slippery slope scenario’, an opening of the 

floodgates to endless claims for accommodation for, amongst other 

things ‘dreadlocks, body piercings, tattoos and loincloths’. 107 On this 

objection, once nose studs have been accommodated, fairness 

requires that other claims for exemptions be granted, resulting in the 

erosion of the uniformity of school uniforms to the point that there 

would be a uniform in name only. This argument has no merit, Langa 

CJ claimed, since, first, exemptions must only be considered for 

‘bona fide religious and cultural practices’. 108 Second, claims for 

exemptions may be refused on the grounds that ‘a practice may be so 

insignificant to the person concerned that it does not require a 

departure from the ordinary uniform’. 109 Third, schools could refuse 



105 For a discussion of what constitutes appropriate deference in the context of 

rights adjudication, see P Lenta ‘Judicial deference and rights’ (2006) Tydskrif vir 

Suid-Afrikaanse Reg 456. 

106 n 97 above, 515. 


108 As above. 



to accommodate a particular practice ‘if accommodating [it] would 

impose an unreasonable burden on the School’: 110 if the practice will 

‘create a real possibility of disruption ... threaten[ing] academic 

standards or discipline’. 111 Finally, as to the extent that other 

cultural and religious practices are exempted from the Code, ‘that is 

something to be celebrated, not feared’ since ‘the display of religion 

and public is ... a pageant of diversity which will enrich our schools 

and in turn our country’. 112 

Langa CJ’s response accords with that of Brennan J in Goldman. 

Brennan J stated that each claim would have to be evaluated against 

the reasons for refusing it. He added that the court could defer to 

‘dress and grooming rules that have a reasoned basis in, for example, 

functional utility, health and safety considerations and the goal of 

polished, professional appearance’. 113 Claims for accommodation 

could justifiably be refused in the case of departures from the 

uniform that infringed these requirements. 

Despite the US Supreme Court’s recent characterisation of the 

‘slippery slope’ argument as ‘the classic rejoinder of bureaucrats 

throughout history’, 114 the raising of this concern is not always an 

indication of bad faith. It may instead reflect ‘the reasonable belief 

that the proliferating recognition of difference might itself generate 

an accelerating and potentially destabilising realignment of 

institutional practices’. 115 Langa CJ was adamant that the decision in 

Pillay ‘does not abolish school uniforms’. 116 Yet, given the broadness 

of culture, or as O’Regan J put it, the fact that ‘all human beings have 

a culture’, 117 the potential number of claims for accommodation that 

may have to be granted consistently with the jurisprudence in Pillay 

may be high. Langa J mentions turbans, yarmulkes and 

headscarves, 118 all of which schools would have to accommodate. 

What about dreadlocks? Provided they were tied up, they may have to 

be accommodated as an expression of Rastafarian culture or religion. 

Body piercings and tattoos? Perhaps: tattoos, for example, may 

express not only the faith of Coptic Christians but also Maori culture. 




113 Goldman (n 97 above) 519. 

114 Gonzalez (n 37 above) 15. 

115 A Sarat & R Berkowitz ‘Disorderly differences: Recognition, accommodation, and 

American law’ (1994) 6 Yale Journal of Law and the Humanities 285 288. 





Kilts? A pupil of Scottish extraction may insist on wearing a kilt at 

school and it is not clear, on the approach outlined in Pillay, that 

schools could refuse to grant this exemption. 119 It is difficult to 

predict how many claims for religious and cultural accommodation 

are likely to be pressed, 120 but were all the claims that are required 

to be accommodated in accordance with the approach in Pillay to be 

brought, school uniforms, although they might not be abolished, 

would be far less uniform. In that case, the educational objectives 

promoted by uniforms may be to some extent undermined. This is the 

price that may have to be paid for extending the obligation to 

accommodate to include optional cultural practices. 

A claim that could be refused on the grounds of the approach in 

Pillay is a claim for an exemption to permit a pupil to carry a weapon. 

In Multani, the Canadian Supreme Court — adopting an approach to 

religious exemptions strongly resembling the approach of the 

Constitutional Court in Pillay to religious and cultural exemptions — 

decided that the ‘duty to make reasonable accommodation’ 121 

obliged a school to accommodate a Sikh pupil’s demand for an 

exemption to permit him to wear a kirpan, a metal dagger, which Sikh 

men believe themselves to be religiously obligated to wear at all 

times. Despite accepting that a kirpan is a bladed weapon, the 

Supreme Court held that the increased risk of violence introduced by 

granting the exemption to the school’s code de vie was acceptably 

low, provided the pupil wore the kirpan sealed and sewn inside his 

clothing. To refuse an accommodation in this case, the Court 

(consistently with the view of the Constitutional Court in Pillay) 

added, would be to ‘stifle the promotion of values such as 

multiculturalism, diversity and the development of an educational 

culture respectful of the rights of others’. 122 

The Constitutional Court would be unlikely to grant an exemption 

for a kirpan (and mistaken were it to do so), since South Africa is ‘an 

119 

It might be contended in response that South Indian/Hindu culture is a more 

authentically South African culture than Scots Presbyterian culture, but that 

would be mistaken. While Hindu South Indians arrived in South Africa as 

indentured labourers from the 1860s onwards, the Scots began to arrive before 

that. Two aspects of Scots culture are relevant here. First, it is strongly religious. 

During the mid nineteenth century, a group of Scots who were ministers of the 

Presbyterian Church of Scotland helped to revitalise the Dutch Reformed Church. 

Secondly, it is difficult to imagine a more forceful and nationalistic assertion of 

culture than that characteristic of the Scots. 

120 

The granting of an exemption by the US Supreme Court in Yoder did not result in 

an avalanche of claims. See W Galston Liberal pluralism (2002) 121: ‘The limited 

public education accommodation for the Old Order Amish endorsed by the 

Supreme Court in Wisconsin v Yoder a quarter of a century ago has not lead to an 

escalation of faith-based demands. Indeed few other groups have sought similar 

treatment for themselves’. 

121 

Multani (n 89 above) para 53. 

122 Multani (n 89 above) para 78.


exceptionally, possibly uniquely, violent society’, 123 whereas Canada 

is one of the least violent countries in the world. Yet its probable 

refusal to do so would not necessarily reflect a difference in approach 

between the Constitutional Court and the Canadian Supreme Court: 

the social contexts of Canada and South Africa are sufficiently 

different as to result in the same approach to accommodations 

leading in the case of weapons to different outcomes. 

4 Accommodating dress and adornment: A 

comparative analysis 

The jurisprudence of the UK House of Lords indicates that it may be 

prepared, in certain circumstances, to grant an exemption to permit 

the wearing of religious clothing and adornments, although its 

approach suggests that it is less solicitous towards claims for 

exemptions to school uniforms for religious dress than its South 

African and Canadian counterparts. In Begum, 124 the House of Lords 

rejected a claim for a religious exemption from a Muslim pupil, who 

wished to wear not only a headscarf (hijab), but also a jilbab, a long 

garment the purpose of which is to hide the contours of the female 

body. Despite accepting that the application was motivated by a 

sincere belief on the part of the pupil that she was under a religious 

obligation to wear a jilbab, the Court denied the claim, for the 

following reasons: first, the current school uniform’s dispensation for 

Muslim pupils was designed in consultation with Muslim authorities to 

meet their religious requirements; second, the pupil could have 

transferred to another school in the area that permitted the wearing 

of a jilbab; 125 third, deference — a ‘margin of appreciation’ 126 — 

should be accorded to the decision of Parliament to allow schools to 

make their own decisions about uniforms 127 and, fourth, to grant this 

exemption might encourage undesirable religious extremism and 

might place undue pressure on Muslim female pupils who did not wish 

to wear the jilbab to do so. 128 

Faced with a case with similar facts to those in Begum, the 

jurisprudence of the Constitutional Court would seem to commit it to 

123 A Altbeker A country at war with itself: South Africa’s crisis of crime (2007) 12. 

124 R (on the application of Begum) v Headteacher and Governors of Denbigh High 

School [2006] UKHL 15 (Begum). 

125 Begum (n 124 above) paras 25 & 89. 

126 Begum (n 124 above) para 62. 

127 Begum (n 124 above) para 34. 

128 Begum (n 124 above) para 65.


granting an exemption to permit the pupil to wear a jilbab. 129 Langa 

CJ rejected the argument that the Court should substantially defer to 

the judgement of the school authorities about whether an exemption 

would interfere with the effective running of the school. He also 

disputed the contention that the pupil could have relocated to 

another school that permitted the practice, since this would fail to 

celebrate diversity, and ‘would be to marginalise religions and 

cultures’. 130 The Constitutional Court is likely to take the view, as far 

as claims for accommodation are concerned, that there is only one 

sensible stopping point: the point at which religious and cultural 

practices violate the rights of others or interfere with the business of 

the school. So, for example, the Constitutional Court should not only 

refuse to accommodate weapons in schools; it should also agree with 

Silber J in Headteachers and Governors of Y School 131 that it is not 

reasonable to expect a school to accommodate a niqab, a veil which 

covers the entire face and head save for the eyes, which would 

frustrate educational objectives by impeding identification of and 

communication with pupils. 132 

129 Begum was followed in Playfoot (a minor), R (on the application of) v Millais 

School [2007] EWHC 1698 (Admin) (16 July 2007), in which the High Court refused 

to grant an exemption to permit a female pupil to wear a silver ring that 

expressed her commitment, as a Christian, to remaining sexually abstinent prior 

to marriage. The Court rejected the claim primarily on the grounds that the 

wearing of the ring at school was not required by her religion unlike other 

practices that the school had accommodated, such as the wearing of headscarves 

by Muslim pupils and the wearing of Kara bangles by Sikh girls. Since the 

Constitutional Court has emphasised that the duty of reasonable accommodation 

includes optional as well as mandatory practices, it may well be that it would 

grant an exemption in these circumstances. The reasoning in Playfoot is anyway 

inconsistent with that of the House of Lords in Williamson, in which it was stated 

that a perceived obligation is not a ‘prerequisite to manifestation of a belief in 

practice’ (n 20 above, para 33). 


131 R (on the application of X) v Headteachers and Governors of Y School [2007] 

EWCH 298 (Admin); [2007] HRLR 20 (QBD (Admin)). 

132 Two points about this. First, although the Court would probably agree with Silber 

J’s conclusion, it would decline to follow the reasoning of Silber J, who, adopting 

the approach of the House of Lords in Begum, declined to grant the exemption on 

the grounds that deference is due to school authorities and because the pupil 

concerned could have transferred to another school which permits the wearing of 

a niqab. Secondly, if the school is acting reasonably by declining a pupil’s request 

that she be permitted to wear a niqab, a school’s decision to refuse to permit a 

teacher to wear a niqab is equally reasonable, since a teacher’s obscuring of her 

face and mouth makes non-verbal communication more difficult and, as a result, 

may impede effective communication between the teacher and pupils. The UK 

Employment Appeal Tribunal took this view in Azmi v Kirklees Metropolitan 

Borough Council [2007] UKEAT/0009/07 (30 March 2007). Of course, non-verbal 

communication might be less important with older learners. See, for example, 

Nussbaum (n 8 above) 350.


The US Supreme Court has yet to be confronted with a claim for 

an exemption from school uniform regulations. Would it be likely to 

grant a claim for an exemption to permit religious clothing and 

adornments? 133 In Smith, 134 the US Supreme Court held that, other 

than in exceptional circumstances, there is no constitutional 

requirement to grant religious exemptions from facially neutral laws 

which do not have the purpose, but may have the incidental effect, 

of burdening the practice of religion. Post-Smith, a facially neutral 

school uniform policy, if fairly administered, would be 

constitutionally valid. Nevertheless, as the Supreme Court stated in a 

subsequent decision, Church of Lukumi Babalu Aye, 135 the regulation 

would have to be applied in a religiously neutral manner. If 

exemptions are made for some religious groups but not for others, as 

they were in Pillay, they may be held unconstitutional. Moreover, the 

regulation might be challenged, as in Pillay, under a ‘hybrid’ claim 

that it infringes not only religious liberty but also freedom of 

expression. The US Supreme Court in Smith indicated that hybrid 

claims might be treated differently from free exercise claims standing 

alone. 136 The claim that the wearing of clothing and adornments 

constitutes expressive conduct appears strong. In Pillay, Langa J 

found that the ban on nose studs under the school uniform infringed 

the pupil’s right to freedom of expression. 137 And in her concurring 

judgement in Begum, Baroness Hale of Richmond contended that a 

woman wearing a hijab ‘may have chosen the garment as a mark of 

her defiant political identity’ and may be engaging ‘in a highly 

complex autonomous act intended to use the resources of the 

133 See J Mikhail ‘The free exercise of religion: An American perspective’ in M 

Mahlman & H Rottleuthner (eds) Ein Neuer Kampf Der Religionen? Staat, Recht 

und Religïose Toleranz (2006) 271. 

134 n 25 above. 


136 n 25 above, 881 - 882. The category of the ‘hybrid’ claim, though difficult to 

interpret, has resulted in some successful claims, particularly in areas that link 

free speech to religious liberty. See, for example, Chalifoux v New Caney Ind 

School Dist, 976 F Supp 659 (SC Tex 1997). 



tradition both to change and to preserve it’. 138 

5 Conclusion 

The relative levels of protection for religious liberty afforded by the 

South African Constitutional Court, the UK House of Lords, the 

Canadian Supreme Court and the US Supreme Court in response to 

claims for exemptions from school uniforms may now be assessed. The 

highest level of protection is provided by the South African 

Constitutional Court and the Canadian Supreme Court. These courts 

are most accommodating of clothing and adornment practices of 

religious (and in the case of the South African Constitutional Court, 

cultural) significance, emphasising the values of multiculturalism and 

diversity. 

The UK House of Lords offers a lower level of protection for 

religious claimants. It is more deferential to the legislature, on the 

grounds of a Dworkinian distinction between policy and principle (as 

in Williamson), and to school authorities, on the basis of their 

specialist expertise (as in Begum). It takes the view that pupils should 

as far as possible accommodate themselves to the relevant regulation 

by, for example, transferring where reasonably possible to another 

school that would permit the practice for which protection is being 

sought. By contrast, the South African Constitutional Court and the 

Canadian Supreme Court’s emphasis on affirming culturally and 

religiously differentiated identities suggests a stance in accordance 

with which pupils are to a lesser extent required to accommodate 

themselves to the regulation. The approach of the House of Lords in 

Begum reflects an anxiety about religious extremism which is 

pervasive in post 9/11 and 7/7 UK but not in the different 

138 n 124 above, para 94. The US Supreme Court has not yet addressed the issue of 

whether school uniforms violate the right to freedom of expression, but lower 

courts have decided that school uniforms do not violate the right to free speech. 

The US Court of Appeals upheld the constitutionality of a mandatory public school 

uniform policy in a Louisiana school district (Canady v Bossier Parish School 

Board, 240 F 3d 437, US Ct App 6th Circuit (2001)). The court found that 

‘improving the educational process’ was an important and substantial 

government interest. In upholding the imposition of mandatory uniforms, the 

court noted that the school’s policy was ‘viewpoint-neutral’. Most importantly of 

all for the present case, the Court went on to find that 

[t[he School Board's purpose for enacting the uniform policy is to increase 

test scores and reduce disciplinary problems throughout the school 

system. This purpose is in no way related to the suppression of student 

speech. Although students are restricted from wearing clothing of their 

choice at school, students remain free to wear what they want after 

school hours. Students may still express their views through other 

mediums during the school day. The uniform requirement does not bar 

the important ‘personal intercommunication among students’ necessary 

to an effective educational process.


environment of South Africa: 139 the House of Lords’s concern about 

the prospect of undesirable extremism and the potential for moderate 

Muslims to become radicalised is not even engaged with in Pillay, 

despite being raised in the school’s Heads of Argument. The lowest 

level of protection is that provided by the US Supreme Court, which 

will in most cases decline to grant exemptions to facially neutral state 

laws and local regulations, including school uniform regulations. 

139 JM Coetzee Diary of a bad year (2007) 122 describes South Africa as a country in 

which ‘Islamist extremism still takes a lowly place on the list of public concerns’.

THE CASE FOR RELIGIOUS INCLUSIVISM 

AND THE JUDICIAL RECOGNITION OF 

RELIGIOUS ASSOCIATIONAL RIGHTS: 

A RESPONSE TO LENTA 

* Barrister & Solicitor, Executive Director, Centre for Cultural Renewal (Ottawa, 

Canada). Research Associate, SAIFAC, Johannesburg. Research Fellow, Faculty of 

Law, University of the Free State, Bloemfontein. I would like to acknowledge the 

contributions and suggestions made to this article from those who attended the 

Constitutional Court Review Symposium — sponsored by the South African 

Institute for Advanced Constitutional, International and Human Rights Law, 

Constitutional Law of South Africa and the Konrad Adenauer Stiftung — in August 

2008. I would be remiss if I did not single out the Executive Director of SAIFAC, 

Theunis Roux, for his encouragement. The expert, substantive editorial 

interventions of University of Pretoria Professor, and SAIFAC Senior Research 

Fellow, Stuart Woolman, have significantly improved this work. Justice Laurie 

Ackermann, as always, has been an unstinting source of support and helpful 

insights. In addition, I have benefited from discussions with Shaun de Freitas 

(Associate Professor) Department of Constitutional Law and Philosophy of Law, 

Faculty of Law, University of the Free State: he generously shared with me his 

own unpublished paper on Pillay. Errors and omissions remain the author’s alone. 

295 

Iain T Benson * 

The idea of culture derived from anthropology, a discipline which 

studied the encapsulated exotic, is no longer appropriate. There are no 

longer (if there ever were) single cultures in any country, polity or legal 

system, but many. Cultures are complex conversations within any social 

formation. These conversations have many voices. 1 


Patrick Lenta, in his comment on Pillay and in earlier work, 2 has done 

South African jurisprudence a great service by forcing us to attend to 

the complexity of cultures and the problems associated with the 

1 M Chanock ‘Human rights and cultural branding: Who speaks and how’ in A An- 

Na’im (ed) Cultural transformation and human rights in Africa (2002) 41. See also 

S Benhabib The claims of culture: Equality and diversity in the global era (2002) 

2 

3-9. 

See P Lenta ‘Cultural and religious accommodations to school uniform 

regulations’ (‘School uniforms’) (2008) 1 Constitutional Court Review 259 ‘Muslim 

headscarves in schools and in the workplace’ (2007) 124 South African Law 

Journal 296 (‘Headscarves’) and ‘Religious liberty and cultural accommodation’ 

(2005) 122 South African Law Journal 352 363-371 (‘Accommodation’).

296 A response to Lenta 

accommodation of religious belief and practice in constitutional 

democracies. The usefulness of his work extends beyond South 

Africa’s borders. Questions of religious (and cultural) accommodation 

are on the front burner in many other constitutional democracies. 

This reply focuses on a narrow set of questions that fall within the 

framework Lenta and others have constructed for discussions about 

the relationship between religion and the State. First, I want to 

suggest that: (a) commentators and courts alike reflexively invoke a 

rather reductive conception of ‘the secular’; (b) that this reductive 

conception is generally ‘anti-religious’; and (c) the South African 

Constitution ought to be read so as to recognise that religions and 

religious practice occupy an important role in the formation of public 

discourse and the support of public institutions. 3 Religions are not 

merely to be tolerated: In societies such as South Africa, they ought 

to be embraced. Second, I hope to demonstrate that the failure to 

take seriously religions and cultural communities with comprehensive 

understandings of the good, results in a body of jurisprudence — in 

South Africa and Canada — that privileges individual religious 

autonomy over the communal dimension of rights to religious freedom 

and practice. 

2 The problem of single cultures or 

‘convergence liberalism’ 

John Gray and others have of late suggested that the concept of a 

‘single culture’ or a ‘one-size fits all’ conception of the public sphere 

hides illiberal strains of liberalism implicitly committed to an 

endorsement of ‘convergence’. Gray has described the two main 

approaches to liberalism in relation to pluralism as follows: 

Liberalism contains two philosophies. In one, toleration is justified as a 

means to truth. In this view, toleration is an instrument of rational 

consensus, and a diversity of ways of life is endured in the faith that it is 

destined to disappear. In the other, toleration is valued as a condition of 

peace, and divergent ways of living are welcomed as marks of diversity 

in the good life. The first conception supports an ideal of ultimate 

convergence on values, the latter an ideal of modus vivendi. Liberalism's 

future lies in turning its face away from the ideal of rational consensus 

and looking instead to modus vivendi. 

3 These readings of ‘secularism’ may appear to some to be variant readings. 

However, they are consistent with the meaning of the neologism coined by 

George Jacob Holyoake. While others may use the term in a way that suggests 

that secularism is religion-friendly and consistent with the contours of every 

liberal democracy, I think it important to emphasise the often aggressive strategy 

of religious exclusion and marginalisation that occurs in many ‘secular’ liberal 

democracies (eg, France).


The predominant liberal view of toleration sees it as a means to a 

universal civilisation. If we give up this view, and welcome a world that 

contains many ways of life and regimes, we will have to think afresh 

about human rights and democratic government. We will refashion these 

inheritances to serve a different liberal philosophy. We will come to 

think of human rights as convenient articles of peace, whereby 

individuals and communities with conflicting values and interest may 

consent to coexist. 4 

Such a characterisation of liberalism would seem to fit Lenta’s notion 

that modern liberal states often assume or propound a false view of 

the public sphere as a ‘neutral’ space. However, Lenta does not 

discuss variant interpretations of the ‘secular’. And when he tends to 

use the term at all, he employs it in a Rawlsian sense 5 to mean the 

opposite of ‘religious.’ This disjunction is odd (and flawed). For the 

Rawlsian characterisation 6 of the secular and the religious leads us 

toward religious exclusivism — a terminal point largely at odds with 

the religious inclusivism of Lenta’s analysis. 7 

In religiously exclusive liberal discourse, ‘secular’ often serves as 

a synonym for ‘non-religious’. While the two terms are commonly 

conflated in this manner, this terminology does not sit comfortably 

with Lenta’s call for inclusivity of religious believers and what he 

describes as a constitutional ‘presumption in favour of the 

government’s being required to grant an exemption [for religious 

belief].’ 8 Such an approach, according to Lenta, means ‘departing 

from the principle of uniformity.’ 9 On this matter, Lenta and I are in 

accord. Further evidence of a more nuanced position can be gleaned 

from Lenta’s work as a whole. In ‘Headscarf’, Lenta describes the 

differences between the French approach to laïcité and the Canadian 

approach to multi-cultural inclusion — and then notes that the 

Canadian commitment to inclusivism — even with its limitations — 

4 J Gray Two faces of liberalism (2000) 105. See also PD Lauwers ‘Religion and the 

ambiguities of liberal pluralism: A Canadian perspective’ (2007) 37 The Supreme 

Court Law Review 1. 

5 See Lenta ‘School uniforms’ (n 2 above) 264. 

6 

See J Rawls Political liberalism (1993). For a useful criticism of the approach 

taken by John Rawls, see PF Campos ‘Secular fundamentalism’ (1994) 94 

Columbia Law Review 1814 1825. Campos would have been better served had he 

noted that Rawls’ argument is not so much ‘secular’ as ‘secularist’. That is, a 

secularist is someone who advocates secularism in the construction of the 

political domain. See further CL Niles ‘Epistemological nonsense?: The secular/ 

religious distinction’ (2003) 17 Notre Dame Journal of Law, Ethics & Public Policy 

561 575 n 50 (criticising Rawls' A theory of justice on these same grounds.) 

7 Religiously inclusive secularism, which we might better call a ‘religiously inclusive 

public sphere’, has been adopted by the Supreme Court of Canada. The Canadian 

jurisprudence shall be discussed in greater detail below. See Chamberlain v 

Surrey Sch Dist No 36, [2002] 4 SCR 710 749 (Can) (‘Chamberlain’) 

8 

Lenta ‘Headscarves’ (n 2 above) 307. 

9 As above.


more closely approximates both the South African position and his 

preferred reading of South Africa’s basic law. 10 

And yet, Lenta remains committed to the proposition that the 

Western liberal political tradition is committed to freedom of religion 

largely as a response to historical religious strife. Of course, there’s 

more than a grain of truth to that proposition. For one can mark 

liberalism and constitutionalism’s true beginning with Locke’s Letter 

on Toleration. But if one does mark ‘religious freedom’ as beginning 

with this modest, but still revolutionary, tract, then, upon a closer 

reading, it is hard to view — from a Lockean perspective — state 

neutrality as liberal exclusivism regarding religion, as anything but an 

anti-religious ideology. 11 

For many liberals, the term ‘secular’ means only ‘non-religious’. 

But that mistakes the term’s origins — the saecularum — as a 

particular way of marking periods of time. It is important, therefore, 

to stop and to examine what we mean when we use terms such as 

‘secular’ or ‘secularism’. The failure to do so can lead, I contend, to 

fundamentally anti-religious outcomes. 12 A decision from the 

Canadian courts provides an illuminating and instructive account of 

why getting this term right is of such great import. 

Mr Justice McKenzie, in the unanimous decision of the British 

Columbia Court of Appeal in Chamberlain, stated that: 

In my opinion, ‘strictly secular’ in the School Act can only mean pluralist 

in the sense that moral positions are to be accorded standing in the 

public square irrespective of whether the position flows out of a 

10 The South African approach might well be said to accept a ‘co-operation’ of 

Church and State — rather than any American notion of ‘strict separation’. Of 

course, anyone au fait with recent American constitutional law knows that the US 

Constitution is no longer read as endorsing an absolutely strict separation of 

Church and State. The principles of such co-operation and useful distinctions 

about avoidance of both atheistic and religious theocracy are neatly set out in A 

Sachs Protecting human rights in a new South Africa (1990) 43–49. See also S 

Woolman ‘Community rights: Language, culture and religion’ in S Woolman et al 

(eds) Constitutional law of South Africa (2nd Edition, OS, 2008) ch 58. 

11 Something of this sort is occurring now in Canada. A Commission made up of two 

academics — including noted philosopher Charles Taylor — have endorsed ‘open 

secularism’ and defined secularism at variance with its history. For comment, see 

IT Benson & TL Nguyen ‘The need to re-evaluate the language of the secular and 

secularism in the quest for fair treatment of minorities and belief in Quebec and 

Canada today’ 18 December 2007, http://www.culturalrenewal.ca/downloads/ 

sb_culturalrenewal/BriefTaylorBouchardCommissionDecember2007Fin a.pdf 

12 

(accessed 12 September 2008). 

For more on the history of the terms ‘secularism’ and ‘secular’, see IT Benson 

‘Considering secularism’ in D Farrow (ed) Recognising religion in a secular 

society: Essays in pluralism, religion and public policy (2004) 83–93 and ‘Notes 

towards a (re)definition of the secular’ (2000) 33 University of British Columbia 

Law Review 519 520–538.


conscience that is religiously informed or not. That meaning of strictly 

secular is thus pluralist or inclusive in the widest sense ... 

No society can be said to be truly free where only those whose morals 

are uninfluenced by religion are entitled to participate in deliberations 

related to moral issues of education in public schools. In my respectful 

view ‘strictly secular’ so interpreted could not survive scrutiny in light of 

the freedom of conscience and religion guaranteed by section 2 of the 

Charter [conscience and religion] and equality rights guaranteed by 

section 15. 13 

What is said here about moral positions applies equally to religious 

and cultural beliefs in a public school setting such as Pillay. Simply 

put, convictions emanating from religious beliefs ought to be at no 

disadvantage in terms of public respect by comparison to belief sets 

that emanate from non-religious convictions. When the case reached 

the Supreme Court of Canada, all nine judges agreed with the 

reasoning of McKenzie J. ‘Secular’ in Canadian constitutional 

jurisprudence embraces a religiously inclusive understanding of the 

term. 14 

Justice Gonthier provides further support for this proposition 

when he writes: 

In my view, Saunders J [the trial judge] below erred in her assumption 

that ‘secular’ effectively meant ‘non-religious’. This is incorrect since 

nothing in the Charter, political or democratic theory, or a proper 

understanding of pluralism demands that atheistically based moral 

positions trump religiously based moral positions on matters of public 

policy. I note that the preamble to the Charter itself establishes that ‘ 

... Canada is founded upon principles that recognise the supremacy of 

God and the rule of law’. According to the reasoning espoused by 

Saunders J, if one's moral view manifests from a religiously grounded 

faith, it is not to be heard in the public square, but if it does not, then it 

is publicly acceptable. The problem with this approach is that everyone 

has ‘belief’ or ‘faith’ in something, be it atheistic, agnostic or religious. 

To construe the ‘secular’ as the realm of the ‘unbelief’ is therefore 

erroneous. Given this, why, then, should the religiously informed 

conscience be placed at a public disadvantage or disqualification? To do 

so would be to distort liberal principles in an illiberal fashion and would 

provide only a feeble notion of pluralism. The key is that people will 

disagree about important issues, and such disagreement, where it does 

13 Chamberlain v Surrey School Board (2000) 80 BCLR (3d) 181 (CA) (‘Chamberlain’) 

at paras 33 & 34, reversing Chamberlain v Surrey School Board (1998) 60 BCLR 

(3d) 311 (SC) (emphasis added). For a detailed analysis of this decision, see IT 

Benson & B Miller ‘Court corrects erroneous understanding of the secular and 

respects parental rights’ Lex View No. 40.0, available at http:// 

www.culturalrenewal.ca/qry/page.taf?id=64 (accessed 12 December 2008). 

14 Chamberlain (n 7 above) 749. Madam Justice McLachlin, who wrote the decision 

of the Chamberlain majority, endorsed the reasoning of Justice Gonthier on the 

correct interpretation of ‘secular’.


not imperil community living, must be capable of being accommodated 

at the core of a modern pluralism [emphasis added]. 15 

The approach of the Supreme Court of Canada and the implicit 

recognition in Pillay that a public school must accommodate a variety 

of beliefs (religious or cultural) are at stark variance with the 

approaches taken by countries such as France. In France, the public 

sphere has generally been stripped of all religious accommodation 

under the mistaken belief that the removal of all religious signifiers is 

‘neutral.’ In Pillay, the Constitutional Court of South Africa arrived at 

a similar conclusion about the place of religion in the public sphere as 

the Canadian Supreme Court did in Chamberlain. (I would have 

preferred, however, a somewhat different characterisation of the 

relationship between the ‘sacred’ and the ‘secular’.) In Fourie, Sachs 

J offers the kind of careful and nuanced understanding of the public 

realm as a sphere of ‘co-existence’ that is largely on all fours with the 

notion of religious inclusivism defended here and set out as above in 

Chamberlain: 

In the open and democratic society contemplated by the Constitution 

there must be mutually respectful co-existence between the secular and 

the sacred. The function of the Court is to recognise the sphere which 

each inhabits, not to force the one into the sphere of the other ... The 

hallmark of an open and democratic society is its capacity to 

accommodate and manage difference of intensely-held world views and 

lifestyles in a reasonable and fair manner. The objective of the 

Constitution is to allow different concepts about the nature of human 

existence to inhabit the same public realm, and to do so in a manner 

that is not mutually destructive and that at the same time enables 

government to function in a way that shows equal concern and respect 

for all ... It is clear from the above that acknowledgment by the State of 

the right of same-sex couples to enjoy the same status, entitlements and 

responsibilities as marriage law accords to heterosexual couples is in no 

way inconsistent with the rights of religious organisations to continue to 

refuse to celebrate same-sex marriages. The constitutional claims of 

same-sex couples can accordingly not be negated by invoking the rights 

of believers to have their religious freedom respected. The two sets of 

interests involved do not collide; they co-exist in a constitutional realm 

based on accommodation of diversity. 16 

A more coherent and accurate way of describing what is at issue is to 

recognise that religious beliefs and believers are within the public 

sphere shared by all. Utilising the secular/sacred dichotomy confuses 

the principles at issue. What should be emphasised is the co-operative 

relationship — between politics and religion — in the same manner as 

15 Chamberlain (n 7 above) para 137 (emphasis added). 

16 Minister of Home Affairs & Another v Fourie & (Doctors for Life Intenational & 

Others, amici curiae); Lesbian and Gay Equality Project & Others v Minister of 

Home Affairs 2006 1 SA 524 (CC) paras 94–98.


the passage correctly notes the co-operation of the religious and the 

non-religious within ‘the same public realm.’ Unfortunately, the 

distinction between the ‘secular’ and the ‘sacred’ at the outset does 

little to assist what I take to be South Africa’s religiously inclusive 

position. For a religious citizen, the public order of the State, too, has 

its own sacred dimension. Why? Because everything within creation 

flows, in a some sense, from ‘grace’ or ‘the holy’ or ‘the divine’. The 

Fourie Court would have done better to describe the public realm as 

encompassing believers of all sorts — whether atheist, agnostic or 

religious — and identifying the role of the law, when certain types of 

conflicts emerge, as ordering all relationships according to principles 

of justice. When most people use the term ‘secular’ they mean 

‘public’ — and it would clarify matters greatly if the Constitutional 

Court (and courts elsewhere) said so in the future. 

Lenta’s use of the work of Iris Marion Young, 17 which is critical of 

‘the liberal tendency’ (the singular conception of ‘liberal’ here is 

telling) is good as far as it goes in suggesting a tendency towards the 

privatisation and consequent public irrelevance of religion within one 

strand of liberal thought. Any monofocal view of ‘liberal’ fails, 

however, to look at different conceptions of liberalism and the 

conceptions of ‘the secular’ themselves that play into such an implicit 

or explicit privatisation and marginalisation of the religious and the 

sacred. Secularism (as an anti-religious movement) goes unanalysed in 

Lenta’s work. But the same, it must said, is true of Pillay and Fourie 

and the Supreme Court of Canada’s reasoning in Chamberlain. 

In discussing Pillay, Lenta does not address how we will try to 

accommodate (if we should) non-religious or non-cultural claims 

regarding belief sets and practices. Unfair discrimination, by both the 

State and private parties, including on the grounds of both religion 

and culture, is specifically prohibited by sections 9(3) and (4) of the 

Constitution. Section 9(3) reads: 

The state may not unfairly discriminate directly or indirectly against 

anyone on one or more grounds, including race, gender, sex, pregnancy, 

marital status, ethnic or social origin, colour, sexual orientation, age, 

disability, religion, conscience, belief, culture, language and birth. 

Pillay focuses exclusively on ‘religion’ and ‘culture’ and concludes 

that cultural practices are entitled to the same degree of 

constitutional solicitude as religious practices. One would do well to 

heed without necessarily agreeing with Justice O’Regan’s concerns 

about the extent to which culture and religion differ in terms of the 

17 See Lenta ‘School uniforms’ (n 2 above) 282.


community and the individual. 18 Two important questions, raised by 

Pillay, but not engaged by Lenta, are: (1) can there be ‘a culture of 

one?’ and (2) does the accommodation approach apply if the 

objection to wearing of a uniform is based simply on ‘belief’ or 

‘opinion’? 

3 Equating ‘culture’ with religion: Can there be 

a ‘religion or culture of one?’ 

If it is true, as John Donne famously observed, that ‘no man is an 

island’, 19 then it might be fair to ask if one person can be a culture or 

one person’s beliefs a religion. What are the requirements in order for 

something to constitute a ‘culture’ or ‘religion’? The judgment in 

Pillay does not help us here. Chief Justice Langa’s judgment begins 

and ends its analysis with the following ‘brief introduction’: 

Without attempting to provide any form of definition, religion is 

ordinarily concerned with personal faith and belief, while culture 

generally relates to traditions and beliefs developed by a community. 

However, there will often be a great deal of overlap between the two; 

religious practices are frequently informed not only by faith but also by 

custom, while cultural beliefs do not develop in a vacuum and may be 

based on the community’s underlying religious or spiritual beliefs. 

Therefore, while it is possible for a belief or practice to be purely 

religious or purely cultural, it is equally possible for it to be both 

religious and cultural. 20 

A tension always exists between understanding religion in its personal 

dimensions and broader communal dimensions. Too great a focus on 

the individual’s belief runs the risk of trivialising the communal 

foundation to which any individual belief is invariably related. It is not 

too strong a statement to say that Pillay gets things back to front: The 

meaning that an individual draws from a religious belief or a religious 

practice is contingent upon the existence of a pre-existing religious 

community with well-developed tenets of belief and practice. Put 

differently, in a very important sense, the religious community 

creates and nurtures the religious believer. 21 Unfortunately, Pillay 

fails to recognise this relationship. 

18 KwaZulu-Natal MEC for Education v Pillay 2008 1 SA 474 (CC) (Pillay) paras 

141–148. As I expand upon below, Justice O’Regan’s reasons view religion 

individualistically and fail to accord religion an appropriate associational 

dimension. 

19 J Donne ‘Meditation XVII’ in MH Abrams et al (eds) The Norton anthology of 

English literature Vol. 1 (3rd ed, 1974) 1215. 


21 S Woolman ‘Freedom of association’ in Woolman et al (eds) (n 10 above) ch 44. 

See also M Walzer ‘On involuntary association’ in A Gutmann (ed) Freedom of 

association (1998) 64.


Consider this statement from another Constitutional Court 

decision on religious faith and culture, a decision referred to in the 

most recent Supreme Court of Canada decision touching on religious 

rights: 

For many believers, their relationship with God or creation is central to 

all their activities. It concerns their capacity to relate in an intensely 

meaningful fashion to their sense of themselves, their community and 

their universe. For millions in all walks of life, religion provides support 

and nurture and a framework for individual and social stability and 

growth. Religious belief has the capacity to awake concepts of selfworth 

and human dignity which form the cornerstone of human rights. It 

affects the believer’s view of society and founds the distinction between 

right and wrong. 22 

The reference both to ‘community’ and ‘social stability’ in this 

passage points towards a better framework for locating the right to 

religion (and also to culture). For it is precisely in the relationship 

between the person and the community that the importance of the 

freedom of religion is best viewed. Social stability, after all, is not 

something achieved by the individual. Its achievement is a function of 

the co-ordination of action between and among individuals within a 

community. Neither Pillay nor the Canadian Supreme Court in 

Amselem 23 seem comfortable with affirming this proposition. 

But should they be? For example, we recognise many religiously or 

culturally sponsored charitable endeavours as public goods. Religious 

institutions that promote ‘health care’ or ‘education’ are only 

possible because various communities commit themselves to their 

realisation and their ongoing maintenance. 

In another setting, that of whether a privilege should extend to 

religious communications in the criminal law area, two justices of the 

Supreme Court of Canada recognised the wider context and referred 

to the following passage with approval: 

In a very real sense, then, the value of religious confidentiality is the value 

to society of religion and religious organisations generally. Even from a 

purely utilitarian perspective, that value cannot be overstated. Religious 

organisations based on claims to unchanging truths are a stabilising 

influence in an increasingly fast-paced and atomised society where bonds 

of community are scarce and worth preserving. Moreover, many provide 

22 

Christian Education South Africa v Minister of Education 2000 4 SA 757 (CC) para 

36. See, generally, I Currie & J de Waal The Bill of Rights handbook (5th ed 2005) 

336–357; P Farlam ‘Freedom of religion’ in Woolman et al (eds) (n 10 above) ch 

41. See further Bruker v Marcovitz 2007 SCC 54; K Boonstra & IT Benson ‘When 

should the courts enforce religious obligations?’ Lex View No 63 (2008), available 

at http://www.culturalrenewal.ca/qry/page.taf?id=150 (accessed December 12, 

2008). 

23 Syndicat Northcrest v Amselem 2004 SCC 47. See discussion in n 28 below.


needed social services that government is unwilling or unable to provide in 

a cost-efficient and humane manner. 24 

Religious organisations and their special character (including their 

internal rules, employment contracts etc.) would be ignored or 

threatened were religious or cultural accommodation limited to the 

protection of only individual practices or belief sets or if courts adopt 

too ready a willingness to scrutinise internal arrangements. Moreover, 

the undeniable public benefits of these privately funded institutions 

are what make these projects of interest to the ‘public’. 25 Yet what 

is clearly understood in charities law gets overlooked all too 

frequently in constitutional dicta. Consider the errant approach of the 

Supreme Court of Canada in the following passage: 

While some provisions in the Constitution involve groups, such as section 

93 of the Constitution Act 1867 protecting denominational schools and 

section 25 of the Charter referring to existing aboriginal rights, the 

remaining rights and freedoms are individual rights; they are not 

concerned with the group as distinct from its members. The group or 

organisation is simply a device adopted by individuals to achieve a fuller 

realisation of individual rights and aspirations. People by merely 

combining together, cannot create an entity which has greater 

constitutional rights and freedoms than they, as individuals, possess. 26 

This position — a standard liberal individualistic approach to 

communities — is both epistemologically incorrect and politically 

unsound: it gets the ontological priority of meaning back to front. 

Individuals only come to participate in these institutions because a 

community has — in most instances — already produced them and the 

meaning that flows from their ongoing reproduction. It is because 

communities exist that individuals have anything to choose at all. 

Indeed, sections 15, 30 and 31 of the South African Constitution 

expressly recognise the communal nature of these religious and 

cultural institutions. When the Constitutional Court is ultimately 

faced, again, with questions of religious and cultural accommodation 

it would do well to recall that these charitable institutions are not run 

by an individual but built and supported over time by identifiable 

communities. The express legal recognition of the ‘public benefits’ of 

many religious and cultural charities points us in the right direction. 

A real danger — and not a mere oversight — exists when the 

‘individualising tendency’ of one strand of liberalism is ascendant. 

24 R v Gruenke [1991] 3 SCR 263 299-300 per L’Heureux-Dubé J for herself and 

Gonthier J concurring with the majority (emphasis added). 

25 On the ‘public benefit’ of religious charities, see H Picard (ed) The law and 

practice relating to charities (3rd ed 1999) 84-116; J Warburton (ed) Tudor on 

charities (9th ed 2003) 78-81 and DWM Waters (ed) Waters' law of trusts in 

Canada (3d ed 2005) 716-745. 

26 Public Service Employee Relations Act (Alberta) Ref [1987] 1 SCR 313.


The individualising tendency, I would argue, poses a genuine threat to 

the proper recognition of religious liberties. Holding the individual 

and communitarian dimensions together in theory and in practice is 

essential for the very social stability that the Constitutional Court in 

Christian Education refers to above. As William Galston notes: 

In some measure, religion and liberal policies need each other. Religion 

can undergird key liberal values and practices; liberal politics can 

protect — and substantially accommodate — the free exercise of 

religion. But this relationship of mutual support dissolves if the 

respective proponents lose touch with what unites them. Pushed to the 

limit, the juridical principles and practices of a liberal society tend 

inevitably to corrode moralities that rest either on traditional forms of 

social organisation or on the stern requirements of revealed religion ... 

[L]iberal theorists (and activists) who deny the very existence of 

legitimate public involvement in matters such as family stability, moral 

education, and religion are unwittingly undermining the values and 

institutions they seek to support. 27 

Yet once one recognises the importance religious and cultural rights 

have for a society, we still need to define what count as religions and 

cultures and ask what are the rules or requirements of entrance, exit 

and membership (to name but three) for a ‘culture?’ 28 Is something 

worthy of constitutional protection simply because I say or believe it 

is? If, like religion, protection and recognition extend to those things 

that are voluntarily chosen as well as obligatory (and what would be 

the obligatory aspect of ‘a culture of one’ except mere assertion of 

the will?), should we be concerned that the important category of 

culture could be thinned out to nothingness, deflated to the point of 

implosion, by lack of meaningful definition? If this concern is 

accurate, as I think it is, then too broad a conception of culture can 

have the paradoxical result that in trying to give content to the 

27 W Galston Liberal purposes: Goods, virtues, and diversity in the liberal state 

(1991) 279. 

28 On entrance, membership and exit see Woolman ‘Freedom of association’ (note 

21 above). See also A Hischmann Exit, voice and loyalty (1970); E Brody 

‘Entrance, voice and exit: The constitutional bounds of freedom of association’ 

(2002) 35 UC Davis Law Review 821. Though not dealing with associational rights 

and focused more on some of the other tensions common to Pillay and Amselem, 

see a helpful exchange between two Canadian scholars: R Moon ‘Religious 

identity and commitment: Syndicat Northcrest v Amselem’ (2005) 29 Supreme 

Court Law Review (2d) 201 and DM Brown ‘Neutrality or privilege?: A comment on 

religious freedom’ (2005) 29 Supreme Court Law Review (2d) 221.


concept we actually denude it. 29 

I have already contended that turning ‘religion’ into a merely 

individually assessed matter can have the damaging effect of reducing 

and weakening the category. Benjamin Berger has observed that one 

of the most profound implications of the relationship between 

religious commitment and assessing a contemporary liberal order is 

that: 

... there is a fundamental, though eminently explicable, shortfall at the 

core of liberal legal discourse about religious liberties. Religion is not 

only what law imagines it to be. Law is blind to critical aspects of 

religion as culture. That being so, even if successful at accommodating 

or tolerating what it understands to be religion, aspects of religion as 

culture remain entirely unattended to and, therefore, unresolved in 

their tension with the constitutional rule of law. And with this insight we 

come to one important part of the explanation for why the story we tell 

about law and religion has proven so unsatisfactory: law — in whose 

29 Thus Pillay could be said to give either insufficient attention to the relationship 

between religion and culture and communities (Chief Justice Langa’s judgment) 

or too much attention to religion and culture as individual practices (as in Justice 

O’Regan’s reasons). A richer examination of the communal prerequisites for 

religious practice can be found in Taylor v Kurtstag No and Others 2005 1 SA 362 

(W). Here, in Judge Malan’s reasons for his judgment, the rules of entrance, 

membership, exit and excommunication in the context of Orthodox Judaism are 

taken seriously by the Court. The Canadian Supreme Court has, to date, focused 

more on the individual side of the religious rights ledger. See Amselem (n 23 

above). In this case the Supreme Court of Canada’s majority stated that religion is 

about ‘self-definition and spiritual fulfillment’ (para 39). Amselem collapses the 

individual and communitarian dimensions of religious practice. Amselem focused 

on what might be termed a ‘fringe religious belief’ and held expressly that it was 

not necessary to show the practice in question was required by the religion. A 

more nuanced reading of the judgement, however, shows that since the concern 

about delving too far into religious beliefs would involve the courts in an 

assessment of ‘dogma’ and that would constitute ‘unwarranted intrusions into the 

religious affairs of the synagogues, churches, mosques, temples and religious 

facilities of the nation with value-judgment indictments of those beliefs that may 

be unconventional or not mainstream’ (para 55) the court, as in Pillay, gave 

respect to the communitarian dimension but in a round-about and implicit 

manner which should have been addressed expressly.. So while the 

communitarian aspects of Charter protection were insufficiently elucidated, they 

clearly informed the decision. See also K Boonstra & I Benson ‘Religion is in the 

eye of the beholder’ Lex View No 64 [comment on Amselem] available at http:// 

www.culturalrenewal.ca/qry/page.taf?id=166 (accessed February 23, 2009).


capacity to tolerate, accommodate, and ‘make space’ for cultural 

claims we place so much faith — fails to appreciate religion as culture. 30 

Pillay and Amselem underscore the need for courts to make express 

the proposition that religion acquires meaning through communities 

and that religion matters to individuals precisely because they are 

(usually) born into communities that give their (religious) lives 

meaning. Belief, in the context of religion, cannot be rightly 

understood as an individual dispositional state. The communal 

character of religious believers poses strong, but not ultimately 

incompatible, challenges to liberal theories grounded in predominantly 

individualistic notions of autonomy. Liberal judges ought, 

in fact, to recognise that communal religious freedom is what enables 

many individuals in our society — religious and not — to flourish. 

4 Are belief and opinion to be analysed in the 

same way as religion and culture? 

What becomes of the analysis of accommodation if a student’s 

objection to wearing a uniform or part of a uniform is based on 

conscience, belief or opinion simpliciter? Since a term such as ‘belief’ 

is also a protected category, what analytical tools are we to bring to 

bear on assessing the importance and bona fides of a belief claim that 

some aspect of uniform wearing is offensive to a student? Are religious 

or cultural beliefs elevated forms of belief in terms of constitutional 

protection? Should religion and culture be considered more weighty 

rights than idiosyncratic or individualistic personal beliefs? Given 

what I and others have argued about the ‘public benefits’ of religious 

charities and projects generally, the answer should be ‘yes.’ 

Lenta mentions the Scottishness of kilt wearing as a cultural 

practice that might have to be accommodated in light of the 

reasoning in Pillay. One wonders how and whether a belief sheltered 

by, say, iconoclastic or eccentric expressions of individual will, but 

nonetheless a ‘belief’, would be analysed? What weight is placed on 

30 B Berger ‘Law’s religion: Rendering culture’ in R Moon (ed) Law and religious 

pluralism in Canada (2008) 264 288 available in an earlier version of the article at 

http://ohlj.ca/english/documents/45-2_02_Berger_postFR2_July10.pdf (accessed 

20 September 2008). David Brown’s critique of a certain overly expansive view of law, 

see n 28, above, at 230 - 233 would apply to Berger’s concept that law somehow 

constitutes a ‘culture’ rather than, as another view would have it, a set of rules that 

must order relationships between cultures. Law understood as a ‘culture’ or 

‘community’ in its own right would seem to call into question the ability of the law to 

adjudicate in a neutral fashion between competing claims amongst sub-cultures and 

communities. The laws are the equal property of all and owned especially by none; if 

this is so then it is probably wiser to avoid understanding constitutional law as 

constituting a ‘culture’ or a ‘community’ given how important constitutional 

principles are to citizenship shared by all.


the balancing scales where uniforms are weighed not against religious 

or cultural beliefs and membership (which the cases show can marshal 

significant social arguments for their respect) but against what may 

seem merely idiosyncratic whims or expressions of pique, adolescent 

rebellion or ego? Say, a clerk at the Constitutional Court who sports a 

Mohawk and goes barefoot with or without wearing a kilt. If the test 

here is largely subjective (as it would appear to be with both religion 

and now culture after Pillay), why should we treat the non-religious, 

non-cultural individual claim for respect differently than claims that 

are at bottom communal, but which the Court treats as individual 

forms of expressive conduct? Examining the associational dimension 

and the goods consequent upon such associations might provide some 

way around this difficulty. 

Courts — and commentators — have tried to avoid rank-ordering 

rights. 31 Is this agnosticism about which rights are more weighty than 

others sustainable? In the most recent decision from the Canadian 

Supreme Court touching upon the freedom of religion in relation to 

civil laws, the majority judgment began by affirming that multiculturalism 

and pluralism must be protected, but then goes on to 

employ a fact-specific application of Canadian Charter rights that 

denies ‘bright-line application but does employ rank-ordered choosing 

without naming it as such’: 

Canada rightly prides itself on its evolutionary tolerance for diversity 

and pluralism. This journey has included a growing appreciation for 

multiculturalism, including the recognition that ethnic, religious or 

cultural differences will be acknowledged and respected. Endorsed in 

legal instruments ranging from the statutory protections found in human 

rights codes to their constitutional enshrinement in the Canadian 

Charter of Rights and Freedoms, the right to integrate into Canada’s 

mainstream based on and notwithstanding these differences has become 

a defining part of our national character. The right to have differences 

protected, however, does not mean that those differences are always 

hegemonic. Not all differences are compatible with Canada’s 

31 Dagenais v Canadian Broadcasting Corp [1994] 3 SCR 835 para 72. Chief Justice 

Lamer stated that ‘[w]hen the protected rights of two individuals come into 

conflict ... Charter principles require a balance to be achieved that fully respects 

the importance of both sets of rights.’ See also Trinity Western University v 

British Columbia College of Teachers [2001] 1 SCR 772 para 29. Justice Iacobbucci 

wrote that ‘[n]either freedom of religion nor the guarantee against discrimination 

based on sexual orientation is absolute.’ For a useful critique of balancing in hard 

cases and where matters of principle are at stake, see S Woolman & H Botha 

‘Limitations’ in Woolman et al (eds) (n 10 above) ch 34 citing New Jersey v TLO 

(1985) 469 US 325 369 (Brennan J describes balancing as ‘doctrinally destructive 

nihilism’). See also the Supreme Court of Canada’s decision allowing a student to 

wear a kirpan (Sikh ceremonial dagger worn for religious reasons) in a public 

school as long as it was sealed within his clothing; Multani v Commission Scolaire 

Marguerite Bourgois [2006] 1 S.C.R. 256 and case comment P Lauwers and IT 

Benson ‘Allowing Kirpans in Public Schools’ Lex View No. 57 http:// 

www.culturalrenewal.ca/qry/page.taf?id=119 (last accessed February 20, 2009).


fundamental values and, accordingly, not all barriers to their expression 

are arbitrary. Determining when the assertion of a right based on 

difference must yield to a more pressing public interest is a complex, 

nuanced, fact-specific exercise that defies bright-line application. It is, 

at the same time, a delicate necessity for protecting the evolutionary 

integrity of both multi-culturalism and public confidence in its 

importance. 32 

The French administrative court’s decision that the wearing of a niqab 

(concealing the face of an individual except for the eyes) is a ground 

for denial of citizenship is, itself, grounded in the underlying notion 

that such concealment is ‘inconsistent with Republican values’. 

Perhaps that is true for France. 33 But what should religiously inclusive 

jurisdictions such as South Africa and Canada do in the face of such a 

dispute and what are the grounds for deciding matters differently? 

The relatively empty appeal to ‘balancing’ articulated above is of 

little assistance here. 

In his earlier writings, Lenta hints, without giving a definitive 

viewpoint on the matter, that a niqab might be distinguished from a 

headscarf on the basis that in certain contexts facial covering could 

lead to ‘making identification and communication difficult’. He 

thereby avoids the religiously inclusive requirement to accommodate 

such a face-covering in public settings where such identification is 

essential or important. 34 I think Lenta is correct — if I have understood 

him properly — and look forward to his further analysis on this point. 

The debate in Canada regarding whether Muslim women should be 

allowed to vote while wearing face-coverings suggests that the issue 

may find practical means of resolution which allow for 

accommodation. 35 

To what extent headscarves or face-coverings should or should not 

be accommodated in public settings such as public school classrooms 

32 Bruker v Marcovitz 2007 SCC 54 paras 1-2 per Abella J. 

33 

K Bennhold ‘A veil closes France’s door to citizenship’ The New York Times 19 July 

2008 available at http://www.nytimes.com/2008/07/19/world/europe/19france. 

html?_r=1&pagewanted=1&ei=5070&en=2b01d3cac5c8b204&ex=1217304000&emc 

=eta1&oref=slogin (accessed 22 July 2008). 

34 See Lenta ‘Headscarves’ (n 2 above) 319. 

35 A Canadian report has noted that there are very few women who wear full head 

coverings that obscure the face. One estimate puts the number in Quebec at 50 

Muslim women out of a total number of around 200 000. According to this report, 

Muslim women in Canada are used to showing their faces for identification 

purposes at banks, airports and at the United States border. There has been some 

debate around the question of whether full facial coverings should be allowed at 

election booths and how identity can be assured. One commonly suggested 

solution has been for election officials to make provision for Muslim women to 

uncover to women election officers for the purpose of establishing identity. 

Another solution is to bring with them an adult who can swear to their identity. 

See http://www.cbc.ca/canada/story/2007/09/09/harper-veil.html (accessed 19 

November 2008).


or government offices must be based upon nuanced and careful linedrawing. 

I do not believe that in a Canadian or South African context 

a blanket gendered approach or ‘Republican values’ conclusion of the 

sort that the French have embraced is likely to be the path ahead. A 

proper understanding of the principles of accommodation and the 

religiously inclusive public sphere of the kind adopted by the Canadian 

Supreme Court in Chamberlain and implicitly endorsed in Pillay and 

Fourie, suggests that civil function issues (such as voting 

identification) may be creatively accommodated. 

What is certain, however, is that the courts have yet to address 

sufficiently the communal nature of religious belief and the 

importance that this communal nature (legally comprehended as 

associational rights) plays in serving the goods that religious beliefs 

can offer to free and democratic societies. An individualistic rights 

approach will tend to dissolve this good for the public and, ironically, 

for individuals as well. 36 

36 Thus D Schneiderman ‘Associational rights, religion and the Charter’ in R Moon 

(ed) (n 30 above) 65 notes that: ‘ ... pluralists will call upon state actors to take 

care that they do not impair associational rights more than is necessary. Courts 

have often not been so careful. They have assumed a unity of purpose between 

state and society that should not so readily be presumed’ (80). Compare in the 

same volume (239) L Weinrib ‘Ontario’s Sharia law debate: Law and politics under 

the Charter’, who focuses on the individual aspect of rights and asserts: ‘These 

[Charter] provisions make clear that the relationship between the individual and 

the state is primary and direct: that is, undiminished by personal characteristics 

and unmediated by given or chosen social affiliations’ (247) (emphasis added). To 

argue, as does Weinrib, that relationships between the individual and his or her 

association (church, organisation etc) are not or cannot be, in constitutionally 

significant ways, ‘mediated’ by what a person has chosen (ie contractually 

agreeing to Membership Rules or a Code of Conduct etc) overstates the 

importance that individual rights may have in a constitutional order and 

contributes to the sort of threats to associational rights I have criticised 

throughout this article. The volume in which these papers appears shows this 

tension between approaches and the emerging awareness of associational rights 

as against what might be termed the earlier but incomplete focus upon more 

individualistic approaches. It seems clear the law can and should move towards 

greater associational rights recognition.


MEDIA FREEDOM AND THE LAW OF 

PRIVACY: NM AND OTHERS V SMITH AND 

OTHERS (FREEDOM OF EXPRESSION 

INSTITUTE AS AMICUS CURIAE) 

2007 (5) SA 250 (CC) 

* Directors, Webber Wentzel; Lecturers, University of the Witwatersrand School of 

Law. 

1 C Smith Patricia de Lille (2002). 

2 The clinical trial was conducted by the head of the Immunology Clinic of the 

Medical Faculty of the University of Pretoria, and was aimed at determining the 

efficacy of a combination of drugs on patients’ HIV levels. 

311 

Glenn Penfold & Dario Milo* 

Patricia de Lille, a biography of the well-known Member of 

Parliament, was published in March 2002. 1 The book included a 

chapter on Ms De Lille’s work in campaigning for the rights of persons 

living with HIV/AIDS. In a chapter recounting Ms De Lille’s involvement 

in a dispute relating to a controversial HIV drugs trial at Kalafong 

Hospital, 2 the names of three trial participants, and their HIV-positive 

status, were published without their consent. After the reference to 

their names was drawn to their attention, the three women instituted 

action against the author (Charlene Smith), Ms De Lille and the 

publisher for breach of their rights to privacy, dignity and 

psychological integrity. The plaintiffs alleged that the defendants 

intentionally published their HIV status without the plaintiffs’ 

consent or, alternatively, that they acted negligently in doing so. 

While the former did not require an extension of the actio injuriarum 

(the delictual action for infringement of personality rights), the latter 

did. 

The defendants argued that the women’s HIV status was not 

private at the time of the book’s publication, in that their status had 

been placed in the public domain. They also denied that the 

publication was either intentional or negligent.

312 Media freedom and the law of privacy 

After a trial that lasted eight days in the High Court, Schwartzman 

J held that the defendants lacked both intention and negligence in 

respect of the publication of the plaintiffs’ HIV status. 3 With regard 

to the latter, Schwartzman J found that the defendants reasonably 

believed, at the time of publication, that the plaintiffs had consented 

to the disclosure of their HIV status. 4 The High Court, however, held 

the publisher liable for continuing to distribute copies of the book 

after it became aware that the women had not in fact consented to 

the general disclosure of their HIV status. 5 As a result, the publisher 

was ordered to pay R15 000 in damages to each plaintiff. The 

plaintiffs appealed against the High Court’s exoneration of the 

defendants in respect of the initial publication of the book. 

After the Supreme Court of Appeal denied leave to appeal, the 

matter came before the Constitutional Court, which (by nine judges 

to one) held the defendants liable for the initial publication of the 

plaintiffs’ HIV status. 6 The judges were further divided as to how they 

came to this finding. The majority, in a judgment penned by Madala 

J, 7 did not deviate from the conventional approach to the actio 

injuriarum, holding that, contrary to the factual finding of the High 

Court, the defendants intentionally violated the plaintiffs’ privacy. 

Madala J was satisfied that the defendants ‘were certainly aware that 

the [plaintiffs] had not given their consent or at least foresaw the 

possibility that the consent had not been given to the disclosure’. 8 

This rendered it unnecessary for the majority to consider whether the 

actio injuriarum should be extended to cover harm brought about by 

negligence (as opposed to the traditional mental element of 

intention). 9 

The remaining three judges, in separate judgments, took the legal 

road less travelled, effectively holding that liability for breach of 

privacy should be extended to include liability for media defendants 

based on negligence. On the facts, Langa CJ and Sachs J found that 

the ‘media defendants’ (ie the author and the publisher) fell short of 

3 NM and Others v Smith and Others [2005] 3 All SA 457 (W) (NM High Court) paras 

40-41. 

4 The principal reasons for this were that the women had been identified by name 

as being HIV-positive in a report on various allegations relating to the manner in 

which the clinical trials were conducted, prepared by an outside expert, Prof SA 

Strauss, for the University of Pretoria (‘The Strauss Report’) (para 40 of the High 

Court judgment). The Strauss Report did not suggest that the contents of the 

Report were confidential (para 40.2). 

5 Para 44.2 of the High Court judgment. 

6 NM and Others v Smith and Others (Freedom of Expression Institute as Amicus 

Curiae) 2007 5 SA 250 (CC) (NM or NM v Smith). 

7 In which six judges, Moseneke DCJ, Mokgoro, Nkabinde, Skweyiya, Van der 

Westhuizen and Yacoob JJ, concurred. 

8 




the standard of negligence (or reasonableness), 10 while O’Regan J 

held that they did not. 11 

Against the background of this variety of factual findings by the 

Court, and to some extent because of them, the judgment in NM v 

Smith raises various issues of significant interest, some of which are 

discussed in this note. 12 The first issue that interests us is the 

majority’s cursory treatment of the ‘constitutional matter’ 

requirement (the threshold requirement for the Constitutional 

Court’s jurisdiction) in a judgment that proceeds to decide the factual 

question of the existence of intention for purposes of the actio 

injuriarum. This is a ‘garden-variety judicial task’ 13 that one would 

expect ordinary appellate courts to undertake, but which is not 

generally the domain of a specialist court whose jurisdiction is 

confined to ‘constitutional matters’. The second, related issue that 

we briefly deal with is the questionable merits of the majority’s 

factual finding that the defendants acted intentionally. This is 

followed by a discussion of the most significant legal development 

that arises from this case: the approach of the minority judgments to 

the role of negligence and unreasonableness in an action for breach 

of privacy. The issues that arise in this context include the distinction 

between media and non-media defendants, the transferability of the 

negligence standard that applies in defamation law to privacy law, 

and the application of the negligence test endorsed by the minority 

judges to the facts of the case. 

2 The ‘constitutional matter’ requirement 

The Constitutional Court is a specialist court that does not have 

unlimited jurisdiction. While section 167(3)(a) of the Constitution 

proclaims that the Constitutional Court is the highest court on all 

‘constitutional matters’, the next paragraph stipulates that the Court 

‘may decide only constitutional matters, and issues connected with 

10 

NM (n 6 above) paras 111 & 207. Although Sachs J relies on a case dealing with 

media liability (the decision of the Supreme Court of Appeal in National Media 

Limited and Others v Bogoshi 1998 4 SA 1196 (SCA) (Bogoshi)) for this approach, it 

is unclear whether he limits liability based on reasonableness to the media 

defendants. For example, he states that he supports the ‘reasons and 

conclusions’ in Madala J’s judgment and finds that Ms De Lille (who, on the 

approach of Langa CJ and O’Regan J, is not a media defendant) did not meet the 

standard of reasonableness (paras 202 & 207). 


12 

Other issues that the case addresses but which are outside of the scope of this 

note include the Court's approach to the privacy of a person's HIV/AIDS status, the 

relevance of dignity and reputation claims in this context (especially given the 

stigma attached to HIV/AIDS in society) and the proper approach to the issue of 

costs. 

13 This phrase is taken from FI Michelman ‘The rule of law, legality and the 

supremacy of the Constitution’ in S Woolman et al (eds) Constitutional law of 

South Africa (2nd Edition, OS, 2005) (Michelman ‘The rule of law’) 11-10.


decisions on constitutional matters’. 14 The threshold of a 

‘constitutional matter’ (and connected issues) thus determines the 

line beyond which the Constitutional Court’s authority does not 

extend. While the Constitutional Court is the highest court on 

constitutional matters, the Supreme Court of Appeal has the last word 

on non-constitutional matters. 15 

A ‘constitutional matter’ is defined in section 167(7) as including 

‘any issue involving the interpretation, protection or enforcement of 

the Constitution’. The Constitutional Court has clarified this 

definition by identifying certain categories of constitutional 

matters. 16 The category that most commonly applies when one is 

dealing with the common law (like the actio injuriarum) is the 

obligation, under section 39(2) of the Constitution, to develop the 

common law in light of the spirit, purport and objects of the Bill of 

Rights. 17 Nevertheless, much uncertainty remains as to the meaning 

of a ‘constitutional matter’. 18 

Given the fact that classification of an issue as a ‘constitutional 

matter’ determines whether the Constitutional Court is lawfully 

empowered to deal with it, and given the lack of clarity as to the 

meaning of this requirement, it is both surprising and disappointing 

14 

15 

16 

Sec 167(3)(b). 

Sec 168(3). 

Frazer v ABSA Bank Limited (National Director of Public Prosecutions as amicus 

curiae) 2007 3 484 (CC) (Frazer v ABSA) para 38; Boesak v S 2001 1 SA 912 (CC) 

17 

(Boesak) para 14. See also Michelman ‘The rule of law’ (n 13 above) 11-7 - 11-8. 

See, for example, Khumalo and others v Holomisa 2002 5 SA 401 (CC) (Khumalo v 

Holomisa); Carmichele v Minister of Safety and Security and Another (Centre for 

Applied Legal Studies intervening) 2001 4 SA 938 (CC); and K v Minister of Safety 

18 

and Security 2005 6 SA 419 (CC) (K v Minister of Safety and Security). 

Part of the reason for this uncertainty is the breadth of our Constitution. As Van 

der Westhuizen J remarked in Frazer v ABSA (n 16 above) para 36: 

‘[p]hilosophically and conceptually it is difficult to conceive of any legal issue 

that it not a constitutional matter within a system of constitutional supremacy’. 

See also Ngcobo J in Van der Walt v Metcash Trading Limited 2002 4 SA 317 (CC) 

(Metcash) para 32; and C Lewis ‘Reaching the pinnacle: principle, policies and 

people for a single apex Court in South Africa’ (2005) 21 South African Journal on 

Human Rights 509 519. Michelman ‘The rule of law’ (n 13 above) 11.2(b) suggests 

that the Constitutional Court’s acceptance of the legality principle (ie the 

exercise of all public power must be in accordance with law) logically means that 

all appeals are constitutional matters. This is because, as Michelman points out, a 

court’s decision amounts to the exercise of public power and must therefore be 

correct as a matter of law. Michelman notes that this is apparently in conflict 

with the Court’s endorsement of its ‘less-than-plenary subject-matter 

competence’ (11-12). Although engagement with Michelman’s thought-provoking 

analysis falls beyond the scope of this case note, we note that it may be based on 

an overly broad conception of the principle of legality (an argument that is 

apparently acknowledged by Michelman at 11-14). In any event, this is not the 

approach taken by the Constitutional Court, which has remarked that ‘[a] 

contention that a lower Court reached an incorrect decision is not, without more, 

a constitutional matter’ (Frazer v ABSA (n 16 above) para 40). See also the cases 

cited below at ns 24, 26 & 27.


that the majority’s judgment in NM v Smith contains only three short 

paragraphs under the heading ‘Is this a constitutional issue?’: 

The applicants approached this Court with a view to vindicate their 

constitutional rights to privacy, dignity and psychological integrity 

which, they allege, have been violated by the respondents. Their claim 

is, however, based on the actio iniuriarum and, therefore, falls to be 

determined in terms of the actio iniuriarum. 

It is important to recognise that even if a case does raise a constitutional 

matter, the assessment of whether the case should be heard by this 

Court rests instead on the additional requirement that access to this 

Court must be in the interests of justice and not every matter will raise 

a constitutional issue worthy of attention. 

The dispute before us is clearly worthy of constitutional adjudication 

and it is in the interests of justice that the matter be heard by this Court 

since it involves a nuanced and sensitive approach to balancing the 

interest of the media, in advocating freedom of expression, privacy and 

dignity of the applicants irrespective of whether it is based on the 

constitutional law or the common law. The Court is in any event 

mandated to develop and interpret the common law if necessary. 19 

It is apparent that only the first paragraph quoted above deals with 

the mandatory requirement of a constitutional matter. The remaining 

paragraphs deal with whether granting leave to appeal is ‘in the 

interests of justice’. 20 All that one gleans from this first paragraph is 

that the aim of the plaintiffs was to vindicate their constitutional 

rights through the actio injuriarum. Nevertheless, as Woolman points 

out, the desire of litigants cannot be decisive of whether a 

constitutional matter arises. 21 The third paragraph perhaps provides 

an indication of the majority’s thinking on this issue, stating that the 

case ‘involves a nuanced and sensitive approach to balancing the 

interests of the media, in advocating freedom of expression, privacy 

and dignity of the applicants’. While this may be true of other 

portions of the judgment, 22 it is difficult to envisage how this 

balancing comes into play in the most significant aspect of the 

19 NM (n 6 above) paras 29-31. 

20 While the existence of a constitutional matter is a necessary requirement for the 

Constitutional Court to entertain a matter, it is not in itself sufficient. The Court 

has a discretion as to whether or not to hear a particular constitutional matter. 

This discretion is exercised on the basis of the ‘interests of justice’, taking into 

account a range of factors. See, for example, Boesak (n 16 above) para 12. In 

relation to direct appeals to the Constitutional Court, the ‘interests of justice’ 

test is expressly included in sec 167(6) of the Constitution. 

21 

S Woolman ‘The amazing, vanishing Bill of Rights’ (2007) 124 South African Law 

Journal 762 782. See also Frazer v ABSA (n 16 above) para 40. 

22 For example, the majority’s reasoning on the quantum of damages (which 

balances various constitutional rights and values) or, more clearly, the minority’s 

extension of the actio injuriarum to negligence in relation to media defendants. 

The latter is a clear constitutional matter, in that the judges sought to develop 

the common law in light of the values of the Constitution (as contemplated in sec 

39(2) of the Constitution).


majority’s judgment, namely, the assessment of intention. This is a 

subjective assessment of the question whether the defendants knew 

that they were disclosing private information without the plaintiffs’ 

consent or actually foresaw that this may be the case and proceeded 

recklessly to disclose the information. It seems to us that this is purely 

a factual enquiry which leaves no room for balancing of interests. 

The paucity of the majority’s explanation as to why the issues in 

its judgment amount to ‘constitutional matters’ is even more 

surprising given the fact that, irrespective of whether it was right or 

wrong to exercise jurisdiction on this matter, it is difficult to 

reconcile the Court’s consideration of this matter with its previous 

jurisprudence on the meaning of ‘constitutional matters’. 

Earlier decisions of the Constitutional Court emphasise that, while 

the jurisdiction of the Constitutional Court is extensive and the term 

‘constitutional matter’ should be given a wide meaning, 23 it is not allembracing 

and some proper meaning must be given to it. 24 Most 

strikingly for our purposes, some of the Court’s decisions prior to NM 

v Smith indicate that purely factual disputes do not amount to 

constitutional matters. 25 We briefly discuss two of them. 

The first is Boesak, in which the applicant alleged that his 

constitutional rights to be presumed innocent and to freedom and 

security had been infringed by a decision of the Supreme Court of 

Appeal upholding fraud and theft convictions against him. According 

to the applicant, the Court erred in finding him guilty beyond 

reasonable doubt. During the course of his judgment, on behalf of a 

unanimous Constitutional Court, Langa DP identified certain broad 

principles that apply to the classification of criminal cases, including 

that a challenge to a decision of the SCA on the basis only that it is 

wrong on the facts is not a constitutional matter; and ‘[u]nless there 

is some separate constitutional issue ... no constitutional right is 

engaged when an applicant merely disputes the findings of fact made 

23 S v Basson 2005 1 SA 171 (CC) (Basson) para 90-1; and Boesak (n 16 above) para 

14. See also the later decision of the Constitutional Court in Frazer v ABSA (n 16 

above) para 37 & 39. 

24 See Basson (n 23 above) para 91; Metcash (n 18 above) para 32; and Dikoko v 

Mokhatla 2006 6 SA 235 (CC) para 123. Van der Westhuizen J, on behalf of a 

unanimous Court in the more recent decision in Frazer v ABSA (n 16 above) para 

39 stated as follows: ‘While the conception of a constitutional matter is broad, 

the term is of course not completely open. The fact that sec 167(3)(b) of the 

Constitution limits this Court’s jurisdiction to constitutional matters presupposes 

that a meaningful line must be drawn between constitutional and nonconstitutional 

matters and it is the responsibility of this Court to do so’. See also 

Wary Holdings (Pty) Ltd v Stalwo (Pty) Ltd and another, Trustees of the 

Hoogekraal Highlands and another v Minister of Agriculture and Land Affairs 

25 

[2008] ZACC 12 para 39. 

I Currie & J De Waal The Bill of Rights handbook (5th ed, 2005) 104 state that: 

‘[m]atters that turn purely on questions of fact are not constitutional matters’.


by the SCA’. 26 The second decision is Phoebus Apollo Aviation CC v 

Minister of Safety and Security, 27 in which the Court held that it did 

not have jurisdiction in relation to a dispute about whether the State 

should, on the facts of the particular case, be held vicariously liable 

for the dishonest acts of three off-duty police officers. Kriegler J 

stated as follows: 

It is not suggested that in determining the question of vicarious liability 

the SCA applied any principle which is inconsistent with the 

Constitution. Nor is there any suggestion that any such principle needs to 

be adapted or evolved to bring it into harmony with the spirit, purport or 

objects of the Bill of Rights. On the contrary, counsel for the appellant 

expressly conceded that the common-law test for vicarious liability, as it 

stands, is consistent with the Constitution. It has long been accepted 

that the application of this test to the facts of a particular case is not a 

question of law but one of fact, pure and simple. The thrust of the 

argument presented on behalf of the appellant was essentially that 

though the SCA has set the correct test, it had applied that test 

incorrectly — which is of course not ordinarily a constitutional issue. 28 

It may be that these cases are not, on a proper reading, authority for 

a general principle that a dispute of fact can never be a constitutional 

matter, particularly if one has regard to the fact that in both cases 

the Court found that no constitutional matter arose only after stating 

that no fundamental right was implicated by the factual dispute. In 

Boesak, for example, Langa DP remarked that upholding a conviction 

when the SCA ought to have had reasonable doubt as to the 

applicant’s guilt ‘is no violation of the applicant’s right to be 

presumed innocent’ before he stated that the appeal did not involve 

a constitutional matter. 29 Similarly, in Phoebus Apollo the Court 

found that the right to property clearly did not apply, before stating 

that no constitutional matter arose. 30 

These cases may therefore leave room for an argument that a 

dispute of fact can amount to a constitutional matter where a 

plausible argument can be made that the matter is relevant to the 

26 

Boesak (n 16 above) para 15(a). 

27 2003 1 BCLR 14 (CC) (Phoebus Apollo). 

28 Phoebus Apollo (n 27 above) para 9. 

29 

n 16 above, para 16. See also para 35. This approach is consistent with the 

decisions of the Constitutional Court to the effect that an incorrect decision by a 

court does not infringe a constitutional right. While one has the right to a fair 

trial (s 35(3) of the Constitution) or a fair public hearing (s 34), one does not have 

a right to the correct outcome in judicial proceedings. See Lane and Fey NNO v 

Dabelstein and others 2001 2 SA 1187 (CC) para 4; Metcash (n 18 above) paras 14 

& 19; and Basson (n 23 above) paras 91, 94 & 99. 

30 n 27 above, paras 4-6 & 9. See also Metcash (n 18 above) para 14.


vindication of a constitutional right (whether directly or indirectly). 31 

This reading of Boesak and Phoebus Apollo finds some support in the 

following dictum of Chaskalson CJ in Basson: 

Where no constitutional right is engaged, a challenge to a decision of 

the SCA or any other court, on the basis only that it is wrong on the 

facts, is not a constitutional matter. And a dispute that does not impact 

upon or give effect to an entrenched right or other constitutional 

provision, will not ordinarily be a constitutional matter. 32 

There may thus have been scope for distinguishing Boesak and 

Phoebus Apollo, and arguing that because the claim in NM Smith was 

a plausible one that was clearly aimed at indirectly vindicating 

constitutional rights, any factual dispute that is relevant to 

vindicating those rights (including one as to intention on the part of 

the defendants) amounts to a constitutional matter. However, the 

difficulty which such an argument would face is the unanimous 

decision of the Constitutional Court in K v Minister of Safety and 

Security. This case involved a claim based on vicarious liability for 

rape committed by uniformed, on-duty police officers. The claim 

sought to vindicate a number of the victim’s constitutional rights, 

including her rights to freedom and security of the person and dignity. 

The Court, in assessing whether a constitutional matter arose, 

considered the meaning of the ‘development’ of the common law for 

purposes of section 39(2) of the Constitution. O’Regan J, writing for a 

unanimous Court, noted that the common law is most clearly 

developed when a common law rule is changed or a new rule is 

introduced. 33 She pointed out, however, that courts more commonly 

decide cases within the framework of an existing rule. O’Regan J 

identified two types of cases where this occurs. The first is where a 

court merely applies the rule to ‘a set of facts which it is clear fall 

within the terms of the rule or existing authority’, in which case the 

rule is not developed. 34 The second instance is where a court 

determines ‘whether a new set of facts falls within or beyond the 

31 

Michelman ‘The rule of law’ (n 13 above) 11-10 puts it as follows: ‘what the 

[Constitutional Court] is really telling us in Pheobus Apollo Aviation is that it 

sometimes will decline to hear argument on a claim of unconstitutionality 

because of the extreme prima facie implausibility of the claim ...’. An example of 

the indirect vindication of a constitutional right would be an action, such as that 

in NM, for breach of privacy (a constitutional right entrenched in sec14 of the 

Constitution) through the actio iniuriarum. 

32 n 23 above, para 91, emphasis added. See also Rail Commuters Action Group and 

others v Transnet Limited t/a Metrorail and others 2005 2 SA 359 (CC) (Rail 

Commuters) para 52, where O’Regan J, after quoting from Boesak, stated as 

follows: ‘This reasoning does not imply that disputes of fact may not be resolved 

by this Court. It states merely that where the only issue in a criminal appeal is 

dissatisfaction with the factual findings made by the SCA, and no other 

constitutional issue is raised, no constitutional right is engaged by such a 

challenge’. 

33 

n 17 above, para 16. 

34 As above.


scope of an existing rule’ so that the ‘precise ambit of each rule is ... 

clarified in relation to each new set of facts’. 35 This falls within the 

concept of the development of the common law contemplated in 

section 39(2), which is aimed at ensuring ‘that our common law is 

infused with the values of the Constitution’. 36 

Turning to NM v Smith, it seems to us that a finding as to whether 

the defendants, as a matter of fact, knew that the information was 

private or that the plaintiffs had not consented to its disclosure or 

that the defendants actually foresaw that possibility, does not 

amount to the development of the common law in the manner 

contemplated in K v Minister of Safety and Security. It is difficult to 

see in what way the values of the Constitution are relevant to this 

enquiry. 37 Moreover, if this does amount to the development of the 

common law, it is difficult to reconcile this with the approach of the 

Constitutional Court in Minister of Safety and Security v Luiters, 38 

which was argued after, and decided before, NM v Smith. In this case, 

dealing with vicarious liability in respect of the actions of an off-duty 

police officer who allegedly placed himself on duty, the Court 

unanimously held (with reliance on K v Minister of Safety and 

Security) that the question as to whether the officer subjectively 

intended to act within the course and scope of his employment was 

not a constitutional matter. Langa CJ remarked as follows: 

The Minister queried the finding of the Supreme Court of Appeal that 

Constable Siljeur subjectively intended to act as a policeman at the time 

of the shooting. A number of reasons were given in support of the 

Minister’s criticism of this finding by the Supreme Court of Appeal. The 

thrust of the Minister’s submission, however, was to urge this Court to 

reconsider the facts as found by the High Court and the Supreme Court 

of Appeal. This submission does not raise a constitutional issue for, as it 

was made clear in K, the question whether a police officer has 

subjectively acted as a police officer is purely factual. 

35 

As above. 

36 K v Minister of Safety and Security (n 17 above) para 17. O’Regan J referred to 

this as ‘the incremental development of the rule’ in light of the normative 

framework of the Constitution. 

37 As Woolman (n 21 above) 783 comments: ‘the majority in NM acts as a trier of 

fact in a run-of-the-mill actio iniuriarum matter’. It may be suggested that the 

approach of the majority in NM falls within the second instance in K v Minister of 

Safety and Security, ie determining whether a particular set of facts falls within 

or beyond the ambit of an existing rule, in that the majority determined that the 

facts fell within the ambit of dolus eventualis in circumstances which would, in 

the absence of the Constitution, amount only to negligence. For a variation on 

this argument, see FI Michelman ‘On the uses of interpretive ‘charity’: Some 

notes on application, avoidance, equality, and objective unconstitutionality from 

the 2007 term of the Constitutional Court of South Africa’ (2008) 1 Constitutional 

Court Review (Michelman ‘Interpretive charity’) 30 - 32. For the reasons set out 

in this article, we do not find this argument persuasive. 

38 2007 2 SA 106 (CC) (Luiters).


The Constitutional Court’s own case law therefore suggests that a 

good argument can be made that the factual question as to the 

defendants’ state of mind in NM v Smith was not a constitutional 

matter, notwithstanding the fact that the plaintiffs were seeking to 

vindicate their constitutional rights. 

Elsewhere in this volume, Michelman, adopting what he describes 

as a ‘charitable’ reading of the majority judgment in NM v Smith, 

suggests that the majority may have developed the common law ‘in 

response to constitutional pressure’ and that all of the issues in this 

case 

stood to be decided in the shadow of the Bill of Rights and of the 

looming possibility that the common law might, depending on how they 

were to be decided, have to be developed under the mandate of section 

39(2). 39 

Michelman goes on to assert that the majority, in finding that the 

conduct of the defendants amounted to intention, may be said to 

have been considering whether the common law rule applies to the 

new factual situation as envisaged in K v Minister of Safety and 

Security. 40 While Michelman’s thoughtful argument merits a more 

detailed response than this case note permits, it seems to us that 

there are at least three possible responses. The first is that one 

cannot get away from the fact that a finding of intention (dolus) 

involves a finding of subjective state of mind (ie actual knowledge or 

foresight) on the part of the defendant. This is a purely factual finding 

which, it seems to us, leaves no space for a consideration of 

constitutional rights or values. Second, if the majority intended, in 

light of the Bill of Rights, to extend the scope of dolus eventualis to 

conduct which would otherwise not have constituted this form of 

intention, one would expect them to have said so expressly. On the 

contrary, the approach of the majority was simply to apply the dolus 

eventualis test to the facts of the case before them. 41 Third, even if 

one is able to bring the majority’s approach within the wording of 

O’Regan J’s judgment in K v Minister of Safety and Security, it seems 

to us more difficult to do so in relation to Luiters. 

Irrespective of whether or not the majority was correct in holding 

that the defendants’ intention (or lack thereof) properly fell within 

the Court’s jurisdiction, our primary difficulty with the judgment of 

39 Michelman ‘Interpretive charity’ (n 37 above) 20; see also 21 and 28. 

40 Michelman ‘Interpretive charity’ (n 37 above) 28. 

41 

As Michelman ‘Interpretive charity’ (n 37 above) points out at 19, Madala J 

expressly states that, in light of the fact that his judgment ‘is not extending the 

common-law definition of intention to include negligence in relation to the 

publication of private medical facts, there will be no “chilling effect” on freedom 

of expression in South Africa ...’ (para 69).


the majority in NM v Smith is that none of these issues were canvassed 

in the judgment and no attempt was made to reconcile the majority’s 

approach with the previous case law. The Court thus dealt with a 

matter that could conceivably fall beyond its jurisdiction — an 

outcome that would have serious implications for the rule of law — 

without providing a meaningful explanation of its reasons for doing so. 

Moreover, if the majority’s reason for regarding the defendants’ 

intention as a constitutional matter is that it was an issue that needed 

to be decided in order to determine whether the plaintiffs’ 

constitutional rights should be vindicated, the effect of this is that a 

wide range of ordinary factual disputes may well occupy the Court’s 

resources in the future. 42 These matters could, for example, include 

an extensive factual dispute as to whether an allegation in a 

defamation action is true (which implicates the right to dignity) or a 

dispute as to whether a person who caused physical harm to another 

did so with intention or negligence (which implicates the right to 

security of the person and bodily integrity). 

Finally on this issue, we reiterate that the Constitution provides 

that the Constitutional Court’s jurisdiction extends to ‘issues 

connected with decisions on constitutional matters’. The Court has 

held that this phrase should be given a wide interpretation so as to 

extend the Court’s jurisdiction to matters that ‘stand in a logical 

relationship’ to constitutional matters. 43 It includes ‘any anterior 

matter [that], logically or otherwise, is capable of throwing light on 

or affecting the decision by this Court on the primary constitutional 

matter’. 44 It is unfortunate that the Court did not consider whether 

the question as to the defendants’ intention fell within the scope of 

this concept. 45 

42 Subject to the ‘interests of justice’ filter discussed above. 

43 Alexcor Limited and another v Richtersveld Community and others 2003 12 BCLR 

1301 (CC) para 30. 

44 As above. The Court went on to state it more formally: ‘when any factum 

probandum of a disputed issue is a constitutional matter, then any factum 

probans, bearing logically on the existence or otherwise of such factum 

probandum, is itself an issue “connected with a decision on a constitutional 

matter”.’ See also Rail Commuters (n 32 above) para 52; and Basson (n 23 above) 

para 22. 

45 The judgment in Luiters (n 38 above) suggests that this may well be the case: ‘[i]t 

was argued on behalf of the Minister that once the Court assumes jurisdiction on 

one basis, it has the power to alter the findings of both the High Court and the 

Supreme Court of Appeal on factual issues even if we do not find it necessary to 

change the test. It may well be that the factual issues are issues that are 

connected to a decision on a constitutional matter in such circumstances’ (para 

29).


3 The majority’s finding of intention 

The most extraordinary aspect of the judgment in NM v Smith is the 

extent to which the Court was divided on the factual question as to 

the defendants’ state of mind. 46 The majority held that the 

defendants had intended wrongfully to infringe the plaintiffs’ privacy, 

finding that they ‘knew well of the wrongfulness of their conduct and 

that the disclosure of private facts was likely to invade the privacy 

rights of the applicants’. 47 The majority’s criticism of the defendants 

is, at times, strident, stating at one point that Ms Smith 

did a half-hearted check [for the consents] but soon became tired of the 

exercise and so decided to go ahead and produce the book without 

having obtained the consents of the applicants. 48 

These are very serious findings for the majority of the Constitutional 

Court to make in relation to persons (the defendants) who had 

testified that they honestly assumed that suitable consents had been 

given and did not consider the possibility that this was not the case, 49 

and in circumstances where the Court states, ‘[t]hat they are good 

activists in the field of HIV/AIDS admits of no doubt’. 50 

Not only are the findings of the majority severe, they are also at 

odds with the decision of Schwartzman J in the trial court who, having 

heard several days of evidence at the trial, held that the defendants 

had lacked both intention and negligence. In relation to the former, 

he stated that Ms Smith and Ms De Lille had 

by their long standing involvement with people infected with HIV, 

demonstrated that they are two of the most unlikely people to 

intentionally invade the privacy of a person infected with HIV. 51 

The findings of the majority are also dramatically at odds with 

O’Regan J’s dissenting judgment. It is remarkable that the majority 

finds that the defendants knew that the plaintiffs had not consented 

46 See also Woolman (n 21 above) 781; and J Steinberg ‘Generous judgment instills 

stigma’ in Notes from a fractured country (2007) 221. 



49 

NM (n 6 above) paras 159-168. 


51 NM High Court (n 3 above) para 40. Michelman ‘Interpretive charity’ (n 37 above) 

states that the majority's refusal to defer to the High Court's finding on the facts 

is explicable if it viewed the High Court's application of the common law rule of 

intention as not being compliant with the Bill of Rights (at current page 37: 

Editors to insert new page number). The difficulty we have with this approach is 

that it assumes that the majority’s discussion of whether intentional conduct was 

present on the facts, constitutes a constitutional matter with which the Court 

ought to have concerned itself. We have argued above that the Court itself does 

not identify this as a constitutional matter raised by the case.


to the disclosure of their HIV status (or had actually foreseen this 

possibility and recklessly disregarded it) where O’Regan J finds, on 

considering the same record, that the reasonable journalist would not 

have been obliged to make further enquiries in this regard. 52 

Given this context, one would expect that the majority would 

have been careful to set out their reasons for differing so dramatically 

from the factual findings of the trial court and O’Regan J. We agree 

with Langa CJ that, in light of the defendants’ heartfelt denials of 

intention, the Court ‘would need a great deal of evidence to find that 

these activists would intentionally infringe the rights of the very 

people whom they are committed to protect’. 53 

The reasons that the majority advances in this regard are, with 

respect, not convincing. While we do not intend to engage in a 

detailed factual analysis for purposes of this note, we mention two 

aspects of concern. The first is that the majority did not engage in any 

detail with the reasons offered by O’Regan J as to why she found that 

the defendants did not act with intention or negligence. The second 

is that a number of the majority’s statements in the portion of its 

judgment dealing with intention indicate that the defendants did not 

have intention and were, at best for the plaintiffs, negligent. For 

example, Madala J states that: ‘[b]oth respondents assumed, without 

any enquiry, that the information contained in the Strauss Report was 

not confidential’ and ‘[Ms Smith] assumed that the applicants had 

consented to the public disclosure of their names and HIV status 

because the source of the publication came from a reputable 

institution’. 54 The Court also notes that ‘[Ms De Lille] failed to take 

sufficient steps to ascertain whether the [plaintiffs] had in fact given 

unlimited consent’. 55 It is difficult to reconcile these statements, 

which suggest that the defendants actually made certain (incorrect) 

assumptions, and ought to have made further investigations, with a 

finding of intention, even in the form of dolus eventualis. 56 

52 

NM (n 6 above) paras 185-187. 


54 NM (n 6 above) paras 60-61, emphasis added. 

55 


56 Michelman ‘Interpretative charity’ (n 37 above) suggests that the fact that the 

inclusion of the names of the three plaintiffs in the book could add no value to 

the publication (and in particular, despite protestations by Ms Smith to the 

contrary, certainly was not required to give authenticity to the publication) ‘may 

have been instrumental in leading Madala J's majority to its conclusion that 

Smith's choice was “intentional” in the pertinent, legal sense’ (at current page 

43: editors to insert new page no). We submit, however, that the utility of 

including the plaintiffs' names in the book can at most be a factor to consider in 

determining whether the state of mind of Ms Smith was such that she acted with 

intention (either in the sense that she knew that she was acting unlawfully or 

foresaw that she may be doing so and acted in reckless disregard of that 

foresight). This factor cannot elevate otherwise negligent but non-intentional 

conduct to the status of dolus.


4 The Court's approach to the development of 

the common law of privacy 

The law of defamation underwent radical reform with regard to media 

defendants in the Bogoshi case. 57 Although the decision has been 

rightly criticised for blurring the doctrinal boundaries between fault 

and unlawfulness in the law of delict, 58 it now appears to be generally 

accepted by our courts that the effect of Bogoshi is that media 

defendants sued for liability in defamation will be entitled to rebut 

the presumption of unlawfulness by proving that their conduct was in 

all the circumstances objectively reasonable (a position that is 

favourable to the media relative to non-media defendants). In 

addition, if the defence of absence of knowledge of unlawfulnness, 

and hence lack of fault, is pursued by a media defendant, the media 

defendant must establish that it was not negligent in this regard (in 

this respect, the media is in a less favourable position relative to nonmedia 

defendants). 59 

Despite these developments in the law of defamation, until the 

NM v Smith decision, the common law principles that governed the 

57 

n 10 above. The reform in fact started at the High Court level with the decision of 

Cameron J (as he then was) in Holomisa v Argus Newspapers Ltd 1996 2 SA 588 

(W) (Holomisa v Argus Newspapers). 

58 

The Court's discussion of lawfulness in the form of reasonable publication 

overlaps significantly with its discussion of lack of fault in the form of a 

reasonable belief that the publication was lawful. For instance, Hefer JA at 1214 

stated that ‘the indicated approach is intended to cater for ignorance and 

mistake at the level of lawfulness; and in a given case negligence on the 

defendant's part may well be determinative of the legality of the publication’. 

For discussion of the blurring between the elements of fault and unlawfulness by 

the Court, see J Burchell Personality rights and freedom of expression: The 

modern actio injuriarum (1998) 226; JR Midgley ‘Media liability for defamation’ 

(1999) 116 South African Law Journal 211; D Milo ‘The cabinet minister, the Mail 

& Guardian, and the report card: The Supreme Court of Appeal's decision in the 

Mthembi-Mahanyele case’ (2005) 122 South African Law Journal 28 38-39; J 

Neethling ‘The protection of false defamatory publications by the mass media: 

Recent developments in South Africa against the background of Australian, New 

Zealand and English law’ (2007) 40 Comparative and International Law Journal of 

South Africa 103 123-124. For further discussion on how courts in the area of 

delict continue to grapple with the distinction between fault and unlawfulness, 

see eg A Fagan ‘Rethinking wrongfulness in the law of delict’ (2005) 122 South 

African Law Journal 90; J Neethling ‘The conflation of wrongfulness and 

negligence: is it always a bad thing for the law of delict?’ (2006) 123 South 

African Law Journal 204; RW Nugent ‘Yes, it is always a bad thing for the law — a 

reply to Professor Neethling’ (2006) 123 South African Law Journal 557; J 

Neethling & JM Potgieter ‘Wrongfulness and negligence in the law of delict: A 

Babylonian confusion?’ (2007) 70 Tydskrif vir Hedendaagse Romeins-Hollandse 

Reg 120; FDJ Brand ‘Reflections on wrongfulness in the law of delict’ (2007) 124 

South African Law Journal 74; A Fagan ‘Blind Faith: A Response to Professors 

Neethling and Potgieter’ (2007) 124 South African Law Journal 285. 

59 The leading judicial recognition of these principles is the Constitutional Court's 

unanimous decision in Khumalo v Holomisa (n 17 above) paras 19-20. See also 

Mthembi-Mahanyele v Mail & Guardian Ltd and Another 2004 6 SA 329 (SCA) 

(Mthembi-Mahanyele) paras 45-46.


law of privacy — more particularly liability based on intention — had 

not been challenged under the Constitution. 60 The facts of NM v Smith 

raised squarely the issue of the fault element of liability in privacy 

cases: assuming that Ms Smith, Ms De Lille and the publisher held the 

view that consent had been provided to reveal the plaintiffs’ 

identities and that this view was mistaken, was it necessary for them 

to prove that such a mistake was reasonable? As stated, the majority 

of the Court did not reach this issue; it found that the defendants 

published with knowledge that the plaintiffs had not provided the 

requisite consent (or at least with reckless disregard as to whether 

consent had been granted). 61 On the other hand, the minority judges 

grappled with the issue at some length. Their judgments raise a 

number of issues of profound significance to media liability in privacy 

cases, and we deal with each in turn: the distinction between media 

and non-media defendants for purposes of privacy (and defamation) 

liability and the desirability of a negligence test in this context; and 

how negligence on the part of media defendants should be assessed 

where journalists rely on official reports. 

5 The distinction between media and non-media 

defendants and the negligence test 

In Bogoshi, the Court expressly perpetuated the distinction between 

media and non-media defendants that was created in Suid-Afrikaanse 

Uitsaaikorporasie v O'Malley, 62 where the Appellate Division opined 

that strict liability applied in defamation cases to members of the 

mass media. 63 Although the Court in Bogoshi rightly jettisoned the 

restrictive principle of strict liability, it held that it would not be just 

to allow media defendants to escape liability on the same basis as 

non-media defendants, that is, by relying on the absence of animus 

injuriandi. 64 Hefer JA relied for this distinction on the Australian High 

Court's judgment in Lange v Australian Broadcasting Corporation: 65 

'the damage that can be done when there are thousands of recipients 

of a communication is obviously greater than when there are a few 

60 

A number of academic commentators had, however, expressed the view (albeit 

without much analysis) that the developments in defamation law had the effect 

that equivalent principles now applied in the law of privacy. See eg D McQuoid- 

Mason ‘Privacy’ in S Woolman et al (eds) Constitutional law of South Africa (2nd 

Edition, OS, 2003) 38-19; Burchell (n 58 above) at 429. 

61 See text accompanying n 8 & 9 above. Madala J, for the majority, stated at para 

57 that this was not an appropriate case to depart from ‘the age-old approach to 

the actio injuriarum. I do not, by any means, wish to be understood to say the 

common law should or could never be developed in this regard’. 

62 

1977 3 SA 394 (A). 

63 O’Malley (n 62 above) 403. This was confirmed by the Appellate Division in 

Pakendorf en Andere v De Flamingh 1982 3 SA 146 (A). 

64 

Bogoshi (n 10 above) 1214. 

65 (1997) 189 CLR 520 (Lange).


recipients'; 66 this made the 'additional burden' of proving lack of 

negligence 'entirely reasonable'. 67 

These principles were developed by the minority in NM v Smith in 

judgments that are now the leading discussions in South African law 

with respect to the question of how to distinguish between media and 

non-media defendants, and why such a distinction is necessary. All 

the minority judges were clear that, as in defamation law, a higher 

level of care is required in privacy law for media defendants as 

opposed to non-media defendants. As O'Regan J stated: 

For purposes of this case, I accept that the legal principles developed in 

Bogoshi should apply not only in the law of defamation but also to the 

infringement of privacy rights by the media ... Modern electronic, print 

and broadcast media are immensely, and indeed, increasingly powerful. 

Publications often reach hundreds of thousands of readers, viewers and 

listeners. It is accordingly appropriate, given the scale of damage to an 

individual that can be caused by such widespread publication, to confer 

special obligations upon the media in respect of publication. In so doing, 

we recognise that the media are not only bearers of rights under our 

constitutional order, but also bearers of obligations. 68 

Given the immense power wielded by the media, 69 we submit that, while 

a negligence test for breaches of privacy by the media clearly limits their 

rights to media freedom, such a limitation is justifiable. We agree with 

Sachs J in this context that ‘the principles developed in [the Bogoshi] case 

are eminently transportable to the law of privacy’, 70 at least in the 

66 

Lange (n 65 above) 572. 

67 Bogoshi (n 10 above) 1214. 

68 

NM (n 6 above) para 177. See also Langa CJ at para 94: 'It is ... constitutionally 

appropriate that the media should be held to a higher standard than the average 

person'. 

69 

For foreign studies in this regard, see generally J Curran & J Seaton Power 

without responsibility: The press, broadcasting, and new media in Britain (6th ed, 2003). For an interesting South African perspective, see R Calland Anatomy of 

South Africa: Who holds the power? (2006) 194-204. 

70 NM (n 6 above) para 203. See also J Neethling ‘The right to privacy, HIV/AIDS and 

media defendants’ (2008) 125 South African Law Journal 36 45. For a different 

view, see H Scott 'Liability for the mass publication of private information in 

South African law: NM v Smith (Freedom of Expression Institute as Amicus 

Curiae)' (2007) 18 Stellenbosch Law Review 387 396-398. Scott argues that the 

analogy between privacy and defamation law in this context is flawed, amongst 

other things, because in Bogoshi the result was to extend much greater 

protection to the media than was previously thought to be the case, whereas the 

impact in relation to breach of privacy is to significantly limit freedom of the 

media (398). However, we submit that this was only the case because of the 

unconstitutional imposition of strict liability on the media; the point is that media 

defendants, post-Bogoshi, are not treated as generously as non-media defendants 

in the law of defamation. We consider that there is no principled objection to 

utilising defences that have developed in defamation law in the privacy context, 

in clearly analogous contexts, as the common law of privacy itself recognises. See 

eg Jansen van Vuuren and another NNO v Kruger 1993 4 SA 842 (A) 850.


context of fault liability. 71 The Constitutional Court in Khumalo v 

Holomisa stated that the defence of reasonable publication — which is in 

many respects akin to a negligence test — strikes the correct balance 

between freedom of expression and reputation and thus limits each right 

as little as possible. O'Regan J stated in that case for a unanimous court: 72 

The defence of reasonable publication avoids ... a winner-takes-all 

result and establishes a proper balance between freedom of expression 

and the value of human dignity. Moreover, the defence of reasonable 

publication will encourage editors and journalists to act with due care 

and respect for the individual interest in human dignity prior to 

publishing defamatory material, without precluding them from 

publishing such material when it is reasonable to do so. 73 

We submit that, in principle, assessing the media’s conduct through 

the prism of a negligence test achieves a reasonable balance between 

the competing considerations of privacy and dignity on the one hand, 

and freedom of the media on the other. On this approach, unless a 

defence negating unlawfulness (such as consent) applies, the media 

would have to establish lack of negligence in order to escape liability, 

such as, on the facts of NM v Smith, that its mistake as to the 

existence of consent was not negligent. 74 Requiring the media to 

establish lack of negligence does not in our view constitute a 

disproportionate restriction on media freedom; indeed, the standards 

expected of reasonable journalists are often reflected in codes of 

conduct, such as the Press Council’s Code, which media houses 

71 Another respect in which the Bogoshi principles seem to us to be capable of 

application in privacy actions is in the context of the defence of truth and public 

interest. While the ‘truth’ aspect of this defence is seldom in issue in privacy 

cases (which typically turn on information that is true but invasive: see eg E 

Barendt ‘Privacy and the press’ 1995 Yearbook of Media and Entertainment Law 

22 26), the truth of an allegation that implicates the privacy of the plaintiff may 

nevertheless need to be established by the media in order to rely upon the 

defence of public interest (see also Tshabalala-Msimang and others v Makhanya 

and others [2008] 1 All SA 509 (W) para 48). Thus a plaintiff is entitled in principle 

to vindicate her right to privacy and dignity where false but non-defamatory facts 

are published, where the matter ‘is by its nature such as to attract the law of 

[privacy]’: Ash v McKennitt and others [2007] 3 WLR 194 (CA) para 80. We submit 

that the application of Bogoshi to the law of privacy in this context has the result 

that facts which may not be true but which have been published reasonably, 

should also benefit from the defence of public interest. See also the approach to 

defences to breach of confidence actions in England: Attorney-General v 

Guardian Newspapers Ltd (No 2) [1990] 1 AC 109 at 283. See also the decision of 

the English High Court in Mosley v News Group Newspapers Ltd [2008] EWHC 1777 

(QB) para 143 and what follows (testing the invasion of privacy in issue against 

the test of whether the journalist acted responsibly based on information then 

available to him). 

72 Khumalo v Holomisa (n 17 above) para 43. 

73 

Compare also the decision of Cameron J in Holomisa v Argus Newspapers (n 57 

above) 617. 

74 We submit that there is also no reason in principle why the qualification that a 

mistake need be reasonable should not extend to mistakes about other elements 

of the delict, such as whether the information constituted a private fact.


themselves espouse, 75 and negligence is a concept with which our law 

is familiar. 76 Moreover, a negligence standard represents a 

reasonable compromise between the extremes of strict liability, on 

the one hand, and a purely subjective test of intention, on the other. 

While the former provides no protection to media freedom, the latter 

may have the deleterious result that privacy concerns are effectively 

eviscerated. 77 A standard based on negligence also finds support in 

the analogous context of defamation law in such jurisdictions as 

England 78 and Australia, 79 and in the jurisprudence of the Privy 

Council 80 and the European Court of Human Rights. 81 

Accepting that negligence should apply to ‘media defendants’, 

the difficulty that remains is a definitional one: how should the line 

be drawn between media and non-media defendants? While 

traditional print and electronic media entities are easily recognisable, 

the digital revolution has the result that ‘[l]ines between different 

forms of media are blurring, making it difficult to apply mediumspecific 

definitions’. 82 O’Regan J’s guidance in NM v Smith in regard 

to the definitional category of media defendants is instructive: 

Professional and commercial purveyors of information are well placed to 

ensure that appropriate systems prevent the unreasonable disclosure of 

private facts and the negligent disclosure of those facts. This is not the 

case for ordinary citizens. Moreover, generally, disclosure by ordinary 

75 As Lord Hoffmann recently stated in an English defamation case, ‘the standard of 

responsible journalism is made more specific by the Code of Practice which has 

been adopted by the newspapers and ratified by the Press Complaints 

Commission’ (Jameel v Wall Street Journal Europe Sprl [2007] 1 AC 359 (HL) 

(Jameel) para 55). Expert evidence as to reasonable conduct on the part of 

journalists could also be led, as it was in the NM, where Professor Anton Harber 

led evidence as to what is expected of a journalist reporting the identity of 

persons living with HIV/AIDS. 

76 See Sachs J in NM (n 6 above) para 204. 

77 We submit that the experience of the US courts with the actual malice standard 

that applies in defamation and false light privacy cases is instructive in this 

regard. This rule — first articulated by the US Supreme Court in New York Times v 

Sullivan 376 US 254 (1964) — has had a disproportionate impact on the rights of 

the plaintiff and has also adversely affected the quality of public discourse. See 

generally D Milo Defamation and freedom of speech (2008) 203-206. This was, 

however, not the outcome on the approach of the majority in NM v Smith, who 

found intention on the part of the defendants. 

78 See Reynolds v Times Newspapers Ltd [2001] 2 AC 127 (HL) at 202, 205. See also 

Jameel (n 75 above) paras 32, 55, 149. 

79 Lange (n 65 above) 573. 

80 See eg Bonnick v Morris [2003] 1 AC 300 (PC) paras 22, 23. 

81 

See eg Bladet Tromso and Stensaas v Norway (2000) 29 EHRR 125 (Bladet) para 

65. For a discussion of the respects in which these jurisdictions articulate a 

negligence test for media liability, see Milo Defamation and freedom of speech (n 

77 above) 188-99. 

82 DA Anderson ‘Freedom of the press’ (2002) 80 Texas Law Review 429 438.


citizens will not be as widespread as disclosure by those involved in 

professional or commercial dissemination of information. 83 

On the basis of this test, the minority judges rightly regarded the 

publisher and Ms Smith as media defendants, but viewed Ms De Lille 

as a non-media defendant. 84 As we have discussed, our courts in 

defamation cases have adopted the proposition that the primary 

rationale for imposing a negligence standard on the media is the 

increased capacity to cause harm inter alia because of the 

widespread manner in which allegations may be published. We note 

that it is not only the traditional media that has the capacity to cause 

severe harm to a person’s reputation and privacy because of the 

extent of the publication; the wide circulation of a leaflet containing 

private information may also have this effect. 85 It remains an open 

question whether the negligence standard should apply beyond the 

traditional media categories, as is the case in Australian law. 86 This 

will require an assessment of various factors, including the capacity 

to cause harm; the resources available to the publisher to guard 

against such harm occurring; 87 whether professional codes of conduct 

— such as the Press Council's Code and the Broadcasting Complaints 

Commission Code — govern the publishers; 88 the credibility attached 

to the medium of communication in issue; 89 and the commercial 

motive of the publication. 90 We caution that the imposition of a 

negligence standard on publishers other than the traditional media 

83 NM (n 6 above) para 181. See also Langa CJ at para 94: ‘It makes sense that media 

defendants, who are experts in the field and who routinely distribute facts to vast 

numbers of people, with a particular air of authority and for commercial gain, 

should be held liable for disclosures which they should reasonably have foreseen 

would cause harm’. 

84 NM (n 6 above) per Langa CJ paras 98-9; O’Regan J para 182. It is more difficult to 

understand Sachs J's decision in this regard. Sachs J concludes that all three 

defendants ‘did not meet the standard of reasonableness’, albeit that he 

accepted that the Bogoshi test applied to media defendants. It is not clear 

whether Sachs J regards Ms De Lille as a media defendant, thus attracting the 

higher standard of care, or whether he means to impose the higher standard of 

care beyond the traditional categories of media defendants. 

85 In one High Court case, Marais v Groenewald en ’n ander 2001 1 SA 634 (T), it was 

precisely such reasoning that led to the Court imposing defamation liability based 

on gross negligence on the defendant, who had widely disseminated a letter to 

various political party structures. Van Dijkhorst J held that ‘[s]kriftelike laster 

kan anders as deur die media — soos hier deur partystrukture — tog ook wyd 

versprei word en ’n goeie naam skaad. Publikasie kan soveel wyer wees as in ’n 

plaaslike koerantjie’ (646). See further our comments in the text in regard to 

imposing negligence liability on non-media defendants. 

86 See also Lange (n 65 above) 572. O’Regan J in NM (n 6 above) para 181 left open 

the issue of whether harmful targeted disclosure of information, e.g. to a small 

community, should also trigger a negligence standard of fault. 

87 See also O'Regan J in NM (n 6 above) para 181. 

88 See the dictum of Lord Hoffmann in Jameel (n 75 above). 

89 

See Bogoshi (n 10 above) 1213. 

90 See NM (n 6 above) para 98 (Langa CJ) & para 181 (O’Regan J).


should be carefully considered on a case-by-case basis, in view of the 

serious consequences for freedom of expression that are involved. 91 

6 The negligence test and reliance on official 

reports 

The final aspect of NM v Smith that we address in this note is the 

application by the minority judges of the negligence standard to the 

facts. It is here that O’Regan J parts ways with Langa CJ and Sachs J; 

while O’Regan J held that the media defendants did not act 

negligently, Langa CJ and Sachs J reached the opposite conclusion. 

The crisp issue for resolution was whether a reasonable journalist and 

publisher would have assumed from the Strauss Report that the 

plaintiffs had consented to their identities and HIV-status being 

revealed. 

Langa CJ held that the Strauss Report was not a public document 

and was not widely distributed; in these circumstances: 

The reasonable media defendant would ... not have relied on the Strauss 

Report as a document that removed their duty to ensure informed 

consent had been obtained ... 

The inescapable conclusion is that a reasonable journalist or a 

reasonable publisher would have foreseen the possibility that there was 

not consent. Because the possible harm was great, the effort necessary 

to avoid the harm minimal and the benefit of publishing the names 

negligible, a reasonable journalist or publisher would have taken steps 

to avoid that harm. 92 

A diametrically opposed factual conclusion was reached by O’Regan 

J, who would have exonerated Ms Smith and the publisher from any 

negligence liability. 93 The interesting aspect of O’Regan J’s judgment 

91 

As Langa CJ stated in NM (n 6 above) para 94: ‘However, to extend that standard 

to ordinary people, and thus to everyday relationships, would be to extend the 

law too far into intensely personal space ... [This] is not a matter that is 

appropriate for the law to regulate’. 

92 NM (n 6 above) paras 110-111. Langa CJ also attached significance to the expert 

evidence of Professor Harber, that a journalist cannot assume that consent has 

been given in relation to HIV/AIDS reporting. Compare Sachs J at para 205, who 

held that Ms Smith ‘should have left no stone unturned in her pursuit of 

verification’. We submit that this dictum, which Langa CJ also adopted (para 111) 

sets the bar too high for media defendants. Even allowing for the nature of the 

private information in issue, it is, we submit, not correct as a matter of principle 

to postulate that reasonableness required that every possible step available to 

verify whether consent had been given, should have been followed. Rather, what 

is required of Ms Smith is that, within the context of a book publication (as 

opposed to, for instance, the reportage of breaking news in a daily newspaper), 

she ought to have taken reasonable steps to verify the issue of consent. 

93 A conclusion also reached by the High Court (NM High Court (n 3 above) para 41).


for media freedom is her treatment of whether a reasonable 

journalist and publisher would be entitled to rely on the contents of 

the Strauss Report, which had disclosed the identities of the 

plaintiffs. O’Regan J reasoned as follows: 

To hold that in the circumstances … [Ms Smith and the publisher] were 

under a further duty to contact either the University [of Pretoria] or the 

[plaintiffs] to ensure that they had in fact consented to publication of 

their names would impose a significant burden on freedom of 

expression. The result of such a finding would be that where personal 

private facts have been published already by a reputable organisation, 

another organisation may not rely on that publication as having been 

done lawfully and without infringement of privacy. 94 

While we agree with much of O’Regan J’s reasoning, it is difficult to 

fault Langa CJ’s analysis on the facts. We consider that it may well be 

that O’Regan J did not attach enough significance to the expert 

evidence of Professor Anton Harber and the nature of the privacy 

interests at stake, which would seem to us to require that a 

reasonable author and publisher of a book — presumably not under 

significant time pressures of the kind experienced by daily 

newspapers and broadcasters where breaking news is reported — 

ought to have taken additional steps to ascertain whether consent had 

been given for widespread disclosure to the public at large. 95 This is 

more particularly so given that the inclusion of the names of the 

plaintiffs could not, in the circumstances, have added much of value 

to the discussion in the biography of Ms De Lille's role in the saga. 96 

What is ground-breaking about O’Regan J’s judgment from the 

perspective of media freedom is her view that, as a general rule, a 

reasonable journalist is entitled to rely on the contents of official 

documents: 

Journalists must be entitled to publish information provided to them by 

reliable sources without rechecking in each case whether the publication 

was lawful, unless there is some material basis upon which to conclude 

that there is a risk that the original publication was not lawful. If there 

is a reasonable basis for suspecting that the publication of private 

information was without consent, a journalist will, of course, bear an 

94 


95 Langa CJ in NM (n 6 above) para 111 suggests that the steps that a reasonable 

journalist would have taken on the facts included finding the annexures to the 

Strauss Report containing the terms of the consent by the plaintiffs or contacting 

the plaintiffs directly. 

96 Michelman ‘Interpretive charity’ (n 37 above) [current page 45 of article; editors 

to insert] of this volume.] attaches great significance to this point in rejecting 

O'Regan J's analysis on the facts.


obligation to check. If there are no grounds for such suspicion, it cannot 

be said that a journalist acts negligently in not checking. 97 

This observation takes our reasonable publication law a significant 

step further than the Bogoshi progeny to date. In Bogoshi, Hefer JA 

set out a number of relevant factors to determine reasonableness of 

media conduct, including the ‘character and known provenance’ of 

the information, ‘the nature, extent and tone of the allegations’, ‘the 

reliability of their source as well as the steps taken to verify the 

information’, and ‘the opportunity given to the person concerned to 

respond, and the need to publish before establishing the truth in a 

positive manner’. 98 However, it has been disappointing that, in the 

ten years since Bogoshi was handed down, very little guidance has 

emerged from the Courts as to how the reasonable publication test 

should be applied in a concrete set of circumstances. 99 The leading 

decision on the application of the Bogoshi case, Mthembi-Mahanyele 

v Mail & Guardian Ltd, 100 muddied the waters, with the Supreme 

Court of Appeal split evenly on the issue of whether the newspaper 

could benefit from a reasonable publication defence. Whereas Lewis 

JA — with whom Howie P concurred — regarded the journalists' 

conduct as reasonable, inter alia on the basis that the tone of the 

article was irreverent and that the newspapers had relied on 

information in the public domain, 101 Mthiyane JA — with whom Mpati 

DP concurred — ruled that none of the guidelines suggested in Bogoshi 

was followed by the newspaper. 102 Mthembi-Mahanyele did little to 

illustrate how to concretise some of factors that Hefer JA had 

enunciated in the Bogoshi case. 

O’Regan J’s approach in NM v Smith that a journalist is, all things 

being equal, entitled to publish information derived from a reliable 

source, and particularly to repeat allegations made by a reputable 

organisation, therefore represents a desirable development in our 

jurisprudence in favour of media freedom. We submit that O’Regan J 

97 NM (n 6 above) para 187. Langa CJ endorsed this suggestion; he held that 

‘journalists should not be forced to verify disclosures made by reputable 

organisations’, but this would depend on ‘the nature of the document, the nature 

of the institution that produced the document, the importance of the interests 

involved and the relevant circumstances of the case’ (para 102). 

98 Bogoshi (n 10 above) 1212-1213. 

99 There have been only a handful of reported decisions applying Bogoshi, and not 

much clarity as to its application has emerged. For instance, in Sayed v Editor, 

Cape Times 2004 1 SA 58 (C), Davis J held that the defence was available for a 

story that had been produced with ‘considerable care’, whereas the defence 

failed in Lady Agasim-Pereira v Johnnic Publishing Eastern Cape (Pty) Ltd [2003] 

2 All SA 416 (SE), inter alia on the basis that obvious sources had not been 

contacted to verify the allegations. 

100 n 59 above. 

101 Mthembi-Mahanyele (n 59 above) 74. 

102 Mthembi-Mahanyele (n 59 above) para 110. The fifth judge, Ponnan AJA, held 

that the publication was not defamatory to begin with; he therefore did not need 

to address defences available to the media (para 85).


is correct to observe that, in general, it would set the standard of 

reasonableness too high to require journalists to verify the contents 

of official reports or documents, and that to require this would have 

a chilling effect on freedom of the media. This position is in harmony 

with the developing reasonable publication jurisprudence in England 

and in the European Court of Human Rights. 

In England, the responsible publication defence has been held to 

entitle publishers to argue that their reporting of an important issue 

of public interest qualifies for protection under the ‘neutral 

reportage’ doctrine. In one leading case, Charman v Orion Group 

Publishing Group Ltd and others, 103 the Court of Appeal confirmed 

that a responsible publication defence will be available where 

'judging the thrust of the report as a whole, the effect of the report 

is not to adopt the truth of what is being said, but to record the fact 

that the statements which were defamatory were made'. 104 In such 

cases, journalists will not be required to verify the truth of the 

underlying allegations. 105 

The European Court of Human Rights has expressly adopted the 

proposition contended for by O’Regan J. In Bladet Troms and Stensaas 

v Norway, 106 the facts were that defamation proceedings were 

brought against a newspaper and its editor in relation to an article 

alleging breaches of seal hunting regulations. The Court held that an 

interference with the article 10 rights 107 of the press was not 

justifiable. The newspaper had relied on the official report prepared 

by a seal hunting inspector for the Ministry of Fisheries. The Court 

held as follows: 

In the view of the Court, the press should normally be entitled, when 

contributing to public debate on matters of legitimate concern, to rely 

on the contents of official reports without having to undertake 

independent research. Otherwise, the vital public-watchdog role of the 

press may be undermined ... 

[T]he Court considers that the paper could reasonably rely on the official 

... report, without being required to carry out its own research into the 

accuracy of the facts reported. It sees no reason to doubt that the 

newspaper acted in good faith in this respect. 108 

The sanctioning in this jurisprudence of reliance on official reports 

and other documents of solid providence has the potential to grant 

103 [2008] 1 All ER 750 (CA) (Charman). 

104 Charman (n 103 above) para 48. 

105 See also Roberts v Gable [2008] 2 WLR 129 (CA) para 53. 

106 n 81 above. 

107 Article 10 of the European Convention on Human Rights protects freedom of 

expression. 

108 Bladet (n 81 above) paras 68, 72.


the media significant breathing space to pursue important stories on 

matters of public interest. For instance, where a media publisher 

repeats allegations made in an official report, it should not be 

required of the media publisher to take extensive steps to verify the 

contents of the report (although the subject of the allegation should 

generally be provided with a right of reply). Stated differently, the 

media defendant's conduct in not verifying the truth of the allegations 

emanating from such a source should not count against it when 

assessing its conduct, both for purposes of the reasonableness 

defence (which negates unlawfulness) and for the defence of lack of 

negligence (which negates fault). 

7 Conclusion 

NM v Smith is a fascinating — and disappointing — decision of the 

Constitutional Court on many levels. We have discussed some aspects 

of the decision in this note. From the perspective of constitutional 

jurisprudence, it is unfortunate that the majority of the Court treated 

the factual enquiry as to whether the defendants acted with intention 

in publishing the plaintiffs’ identities as a constitutional matter, 

without undertaking any substantive analysis of this difficult issue. 

This is particularly the case where the majority’s approach is difficult 

to reconcile with previous decisions of the Constitutional Court on 

what is and what is not a ‘constitutional matter’. The majority of the 

Court in our view also appeared to reach the incorrect conclusion on 

the facts, finding that the defendants had intentionally violated the 

plaintiffs’ privacy. We prefer the judgments of the minority judges, 

who reached a conclusion on the major constitutional issue involved 

in the case — that the media need to establish absence of negligence 

in publishing private information. The decision of Langa CJ on the 

facts — visiting negligence liability on the media defendants — 

appears to us to be defensible. Further, O’Regan J’s endorsement of 

a general rule that journalists are entitled to rely on reliable sources 

and information from reputable organisations, represents a 

significant advance in our media freedom jurisprudence.

WANTED: A PRINCIPLED APPROACH TO 

THE BALANCING OF PUBLIC POLICY 

CONSIDERATIONS. STEENKAMP NO V 

PROVINCIAL TENDER BOARD, 

EASTERN CAPE 

It matters how judges decide cases 1 


* 

Senior Lecturer, University of Witwatersrand. BA, LLB (Cape) LLM (Wits). I wish to 

thank the editors of the Constitutional Court Review for inviting me to be a 

contributor to the inaugural edition and for affording me the opportunity to 

present an earlier version of this note at the Constitutional Court Review 

Conference in Somerset West. I am thankful for the comments and questions 

raised by the participants at the conference, particularly those of Cora Hoexter. I 

would especially like to thank Stu Woolman for his editorial input, as well as for 

his continuing support, encouragement and patience. In the spirit of Harambee, I 

must express my heartfelt thanks to Mtendeweka Mhango for his insightful 

comments and valuable engagement with an earlier version of this paper. Finally, 

I must convey my thanks to Jonathan Klaaren for his critical and insightful 

comments. Any mistakes or omissions remain my own. 

1 

See R Dworkin Law’s empire (1986) 1. 

2 2007 3 BCLR 300 (CC). 

335 

Sanele Sibanda* 

Steenkamp NO v Provincial Tender Board, Eastern Cape 2 is unlikely to 

generate too much controversy in academic circles or in the 

profession. The majority decision entrenches the status quo in so far 

as the delictual liability of tender boards is concerned. And apart from 

the majority judgment’s rather confounding comments that appeared 

to conceive the subject of remedies in rather rigid and dichotomous 

terms, nothing in the written text appears outwardly contentious. 

Even the disagreement between the majority and the minority 

concerns a point of law upon which reasonable people could disagree. 

However, that is not to say that the judgment does not raise any 

issues of interest and concern. My aim in this note is to reflect upon, 

and where appropriate critique, the manner in which the majority and 

minority arrived at their respective positions with respect to the

336 Wanted: A principled approach to the balancing of public policy considerations 

determination of wrongfulness. In particular, I will consider the 

manner in which the both judgments dealt with the balancing of the 

public policy considerations raised. I critique the rather thinly 

substantiated, individualised and subjectivised approach of the 

majority to the balancing of the policy considerations. In particular, I 

take issue with the fact that the majority’s approach is focused on 

justifying the outcome of the matter as and between the parties, thus 

missing an important opportunity to guide us on how matters involving 

similar considerations should be approached in future. I then contrast 

this approach with the minority’s clearer identification of the values 

and interests represented by the policy considerations raised. 

The minority’s approach to the balancing of policy considerations 

— grounded in the principle of proportionality — is to be preferred. 

Proportionality doe not simply require that courts clearly identify the 

competing values or interests at stake in a dispute. Properly 

conceived, proportionality demands that courts elaborate in some 

detail the reasons that support their conclusions. 

2 Background and facts 

Although judgment in Steenkamp was handed down in 2006 and 

reported in 2007, the facts giving rise to this matter preceded the 

Constitutional Court’s decision by over a decade. The matter has its 

origins in disputed tender proceedings in the Eastern Cape. In 1996, 

Balraz Technologies (Pty) Ltd (‘Balraz’) was awarded a tender by the 

provincial tender board. However, that award was subsequently set 

aside in review proceedings instituted by one of the unsuccessful 

tenderers, Cash Paymaster Services (Pty) Ltd. 3 In Cash Paymaster 

Services (Pty) Ltd v Eastern Cape Province, the court held that given 

the irregularities in the tender adjudication process, the tender board 

had acted in an administratively unfair manner. 

Unfortunately, Balraz did not survive to tender another day. It was 

placed under final liquidation before it could retender. As a result, all 

future claims relating to the original tender were instituted by 

Balraz’s liquidator. At the trial court, the applicant, Balraz’s 

liquidator, was unsuccessful in its claim for the recovery of out-ofpocket 

expenses incurred as a result of its having commenced to 

deliver services as per the contract arising out of the original tender 

award. The court held that the applicant’s status as an unregistered 

entity at the time of submitting its tender documentation rendered 

the tender award void. No relationship giving rise to a duty of care 

could have arisen as and between itself and the tender board. Put 

3 

For the review decision, see Cash Paymaster Services (Pty) Ltd v Eastern Cape 

Province 1999 1 SA 324 (Ck).


differently, the conduct of the tender board vis-à-vis the applicant 

could not be properly conceived as being wrongful. 

Undeterred by this outcome, the applicant appealed to the 

Supreme Court of Appeal. That appeal was dismissed. The Supreme 

Court of Appeal’s decision focused on the delictual liability of the 

tender board as a public body. It held, on grounds of public policy, 

that neither statute nor common law ought to be extended to cover 

alleged delictual damages caused by the tender board (to the 

applicant) that were of a purely economic nature. The Supreme Court 

of Appeal held so despite the fact that it had found that the tender 

board had acted negligently (albeit in good faith). 

3 The majority decision 4 

In the Constitutional Court, the issue was crystallised as follows: Can 

a once successful tenderer recover its out-of-pocket expenses 

incurred after the contract was awarded, but before the said award 

was set aside on review? 5 The majority judgment, delivered by 

Moseneke DCJ, turned on whether or not the respondent’s decision to 

award the tender to the applicant, whilst in breach of its 

constitutional and statutory mandate to exercise its powers in an 

administratively fair manner, was wrongful in a delictual sense. 6 The 

respondent had conceded that it had acted in an administratively 

unfair manner, albeit in good faith, when it assessed and awarded the 

tender to Balraz. 7 

In considering the import of the tender board’s failure to uphold 

the principles of administrative justice, the majority rejected the 

idea that the breach of a constitutional or a statutory duty should as 

a matter of course be regarded as being delictually wrongful. 

According to the majority, the determination of wrongfulness giving 

rise to the recognition of a duty of care and liability on the part of a 

defendant could not be made without having regard to public policy 

considerations of what was fair and reasonable in the particular 

circumstances. 8 

4 I do not consider the concurring decision delivered by Sachs J because the issues 

raised therein are not germane to those presently under consideration. 

5 

Two other issues were raised in the Constitutional Court: firstly, whether or not 

there was a constitutional issue raised; and secondly, whether or not it was in the 

interests of justice to consider the matter. These issues will not be discussed in 

this note. For present purposes it suffices to say that in both instances the 

decision of the court was in the affirmative. 

6 See sec 187 of the Constitution of the Republic of South Africa, Act 200 of 1993 

(Interim Constitution) and sec 217(1) of the Constitution of the Republic of South 

Africa, 1996 (‘Final’ Constitution); and the State Tender Board Act 2 of 1994 

(Repealed by Provincial Tender Board Repeal Act (Eastern Cape) 6 of 2004). 

7 

Steenkamp (n 2 above) para 36. 

8 Steenkamp (n 2 above) para 37.


In holding that the applicant’s claim must fail, the majority held 

that the applicable constitutional and statutory provisions in terms of 

which the tender board acted did not envision a tenderer in the 

position of Balraz having a claim to damages. 9 Instead the majority 

took the view that had Balraz been prudent in the circumstances, it 

would have utilised other avenues of redress available to it to 

mitigate its risk of loss. 10 For example, according to the majority, 

Balraz could have tendered again after the adverse outcome of the 

review proceedings. 11 Alternatively it could have negotiated for a 

right to reimbursement for out-of-pocket expenses at the time of 

contracting in order to safeguard it from loss in the event of the 

tender award being overturned on review. 12 

In concluding that Balraz was not owed a duty of care by the 

tender board, the majority held that no rights or values in the 

Constitution would justify an extension of the ambit of the law of 

delict in the manner sought. In reaching this conclusion, the majority 

found that compelling public policy considerations required that it 

provide some level of immunity to tender boards that made innocent 

but negligent decisions. 13 One such policy consideration that weighed 

heavily on the majority was a deep-seated concern that an extension 

of liability in the manner sought would open the floodgates of 

litigation by unsuccessful tenderers. This prospect, the majority 

feared, would have the undesirable consequence of eroding the 

efficacy of the tendering process and tarnishing the image of tender 

boards in the public’s eyes. 14 Given that the majority had disposed of 

the matter on the basis of wrongfulness, the majority declined to 

answer questions relating to the validity of Balraz’s tender or the 

tender board’s negligence. 15 

4 The minority decision 

Langa CJ and O’Regan J, in delivering the dissenting minority 

judgment, identified one especially critical ground upon which they 

9 Steenkamp (n 2 above) para 47. 

10 

Moseneke DCJ, clearly unmoved by the applicant’s contentions, makes this point 

when he writes that ‘[o]n the facts, Balraz wasted no moment to accept the 

tender award. But once the order to supply goods and services was made by the 

department, Balraz should have curbed its commercial enthusiasm as it was well 

within its right to require that its initial expense not lead to its financial ruin 

should the award be nullified. Balraz unnecessarily chose the more hazardous 

course which is to incur mainly salary expenses of directors without fashioning 

out an appropriate safeguard. Its loss could have been easily curbed by prudent 

conduct and precaution.’ Steenkamp (n 2 above) para 52. 

11 



13 Steenkamp (n 2 above) para 55(a). 

14 

Steenkamp (n 2 above) para 55(b). 



differed with the majority, namely the determination of wrongfulness 

on the part of the tender board. Langa CJ and O’Regan J deemed it 

important to characterise the applicant’s claim as one of pure 

economic loss. Such a characterisation had historically played a part 

in the courts’ disinclination to recognise a duty of care. 16 

According to the minority, the nature of the wrongfulness inquiry 

in the post 1994-era must, in essence, accord with the norms and the 

values of Constitution. In a matter such as the one before the Court, 

the minority held that the following norms or principles had to be 

taken into account when determining wrongfulness: 

(1) The need to guarantee that successful tenderers have an enduring 

faith in the tender process and the resultant contracts. Such a guarantee 

ensures that tenderers commence to deliver on their contracts timeously 

without having to labour under the fear of facing financial ruin in the 

event that a tender is subsequently set aside through no fault of their 

own. Further, a failure to provide such a guarantee posed a greater risk 

of financial ruin to smaller or new business enterprises that government 

procurement policies sought to promote. 17 

(2) The need to properly differentiate between claims for damages 

arising out of loss of profit instituted by an unsuccessful tenderer on the 

one hand; and damages resulting from out-of-pocket expenditure 

incurred by an initially successful tenderer who had commenced to 

deliver services in accordance with its contractual obligations. The 

former would constitute a windfall claim and the latter a claim for 

reimbursement for services rendered. 18 

(3) The need to promote accountability in the entire tender process in 

order to ensure that that once successful tenderers who incur out-ofpocket 

expenditure spent in the good faith fulfilment of a tender 

contract were not unduly burdened with unrecoverable expenditure. 

Such a burden would inevitably have a detrimental impact on the future 

delivery of tender obligations. 

Having considered the factors set out above, the minority concluded 

that in the circumstances of the case, the tender board acted 

wrongfully, owed the applicant a duty of care, and thus created the 

basis for a claim for delictual damages. 


17 




5 Comparing the approach of the majority and 

the minority to the balancing of public policy 

considerations 

My focus in this section will be on what I take to be the main point of 

disagreement between the two opinions: namely the approach to the 

balancing of public policy considerations. I will not, therefore, 

concern myself with the actual decision reached on the nonconstitutional 

point of delictual wrongfulness. For purposes of this 

note I am satisfied that the Court could conceivably have gone either 

way. 

5.1 Wrongfulness as a requirement for delictual liability 

Before I turn squarely to the central concern of this note, namely the 

balancing of public policy considerations I want to consider briefly the 

requirement of wrongfulness as an element of delictual liability. The 

requirement of wrongfulness is a sine qua non for the attachment of 

delictual liability. Wrongfulness, in a delictual sense, is said to exist 

where particular conduct either infringes upon a person’s legally 

recognised right or where it results in the breach of a legal duty owed 

to a person by another. 19 According to Neethling et al, whether 

particular conduct is in fact wrongful is determined with reference to 

the legal convictions of the community or the community’s boni 

mores. This inquiry, according to the authors, entails an objective 

test based on the criterion of reasonableness. The pertinent question 

to ask is whether or not the conduct in question was justifiable or 

unjustifiable. 20 

In the context of the wrongfulness inquiry, the determination of 

reasonableness requires (a) that a court must take into consideration 

all the circumstances of the matter pertaining to the parties and (b) 

that a court must consider issues of public and legal policy in making 

a value judgment as to how the alleged wrongdoer’s conduct should 

be viewed by society. 21 The second proviso requires that the 

19 JR Midgley & JC Van der Walt ‘Delict’ LAWSA (2nd ed, 2005) Volume 8(1) para 60. 

See also J Neethling et al Law of delict transl & ed J Knobel (5th ed, 2005) 31. 

20 Neethling et al (n 19 above) 34. 

21 

See Olitzki Property Holdings v State Tender Board and Another 2001 8 BCLR 779 

(SCA) para 12 (Cameron JA commenting on how reasonableness is determined 

puts it as follows: ‘The determination of reasonableness here in turn depends on 

whether affording the plaintiff a remedy is congruent with the court’s 

appreciation of the sense of justice of the community. This appreciation must 

unavoidably include the application of broad considerations of public policy 

determined also in the light of the Constitution and the impact upon them that 

the grant or refusal of the remedy the plaintiff seeks will entail.’ (footnotes 

omitted).) See also Minister of Safety and Security v Van Duivenboden 2002 6 SA


reasonableness of particular conduct be determined in terms of the 

legal convictions of the community. 22 Historically, the test for 

reasonableness in this context has been described as being openended 

and flexible. It necessarily takes cognisance of the everevolving 

nature of the legal convictions of the community. 23 

5.2 The important role of public policy considerations in 

determining wrongfulness in novel situations 

The role played by public policy considerations in the determination 

of reasonableness in areas where liability has not previously been 

established cannot be overstated. 24 Van Aswegen, writing in the preconstitutional 

era, pointed out that the most important role played by 

policy considerations has been in ‘the new developments, expansions 

and adaptations to the settled body of delict law.’ 25 The overriding 

utility of public policy considerations is that they afford the court an 

opportunity to consider novel problems for which there is no 

precedent or authority. In other words, despite the clearly openended 

nature of the concept of ‘public policy’, the concept, at the 

very least, requires judges to provide some level of principled 

reasoning grounded in social policy for the decisions that they make. 

The importance of public policy has been entrenched further in 

the current constitutional dispensation. Significant developments 

have occurred in extending the limits of delictual liability through the 

recognition of new duties of care. These new duties have often been 

21 431 (SCA) para 16; Premier of the Province of the Western Cape v Fair Cape 

Property Developers (Pty) Ltd [2003] 2 All SA 465 (SCA) para 41. 

22 

Neethling et al (n 19 above) 34. See also A Fagan ‘Rethinking wrongfulness in the 

law of delict’ (2005) 122 South African Law Journal 90 93-94. Although, strictly 

speaking, it is outside the scope of this paper, it is worth mentioning that the 

Fagan’s robust analysis of wrongfulness as an element of delict in the 

aforementioned article has ignited a debate with respect to the manner in which 

our courts engage in the wrongfulness inquiry (and its relationship to negligence). 

See J Neethling ‘The conflation of wrongflness and negligence: Is it always such a 

bad thing for the law of delict’ (2006) 123 South African Law Journal 204; RW 

Nugent ‘Yes, it is always a bad thing for the law: A reply to Professor Neethling’ 

(2006) 123 South African Law Journal 557; J Neethling & JM Potgieter 

‘Wrongfulness and negligence in the law of delict: A Babylonian confusion?’ 

(2007) 70 Tydskrif vir Hedendaagse Romeins-Hollandse Reg (2007) 120; J 

Neethling & JM Potgieter ‘In (self)defence of the distinction between 

wrongfulness and negligence’ (2007) 124 South African Law Journal 280; and A 

Fagan ‘Blind Faith: A response to Professors Neethling and Potgieter’ (2007) 124 

South African Law Journal 285. 

23 See Midgley et al (n 19) para 60. See also Amod v Multilateral Motor Vehicle Fund 

(Commission for Gender Equality Intervening) 1999 2 All SA (SCA) para 23. 

24 

MM Corbett ‘Aspects of the role of policy in the evolution of our common law’ 

(1987) 104 South African Law Journal 67. 

25 A van Aswegen ‘Policy considerations in the law of delict’ 1993 (56) Tydskrif vir 

Hedendaagse Romeins-Hollandse Reg 171; See also Midgley et al (n 19 above) 

para 60.


rooted in the ‘constitutionalisation’ of public policy. In Carmichele, 26 

the Constitutional Court recognised that in making the type of value 

judgments required in order to determine wrongfulness in novel 

situations, the constitutional injunction contained in s 39(2) requires 

all courts to promote the spirit, object and purport of the bill of 

rights. 27 In deciding to extend the delictual liability of police officers 

and prosecutors to instances wherein the wrongful conduct 

constituted an actionable omission, the Carmichele Court accepted 

that the determination of wrongfulness demands that some regard be 

paid to public policy considerations. Ackermann J and Goldstone J 

wrote: 

Before the advent of the [Interim Constitution], the refashioning of the 

common law in this area entailed ‘policy decisions and value judgments’ 

which had to ‘reflect the wishes often unspoken, and the perceptions, 

often dimly discerned, of the people’. A balance had to be struck 

between the interests of the parties and the conflicting interests of the 

community according to what ‘the (c)ourt conceives to be society’s 

notions of what justice demands’. Under s 39(2) of the Constitution 

concepts such as policy decisions and value judgments reflecting ‘the 

wishes ... and the perceptions ... of the people’ and ‘society’s notions of 

what justice demands’ might well have to be replaced, or supplemented 

and enriched by the appropriate norms of the objective value system 

embodied in the Constitution. 28 (emphasis added) 

The Court’s acknowledgment of the fact that the advent of the 

Constitution has established ‘an objective normative value system’ 

has influenced the development of notions of public policy and, 

concomitantly, the understanding of what constitutes the legal 

convictions of the community. 29 The influence of this ‘phrase’ has led 

26 See Carmichele v Minister of Safety and Security 2001 4 SA 938 (CC) (hailed as a 

groundbreaking judgment that saw the extension of the delictual liability for 

omissions to instances in which the police and public prosecutors failed to act in 

circumstances where they could have prevented the plaintiff from being raped). 

Section 39 (2) of the Constitution reads as follows: ‘When interpreting any 

legislation, and when developing the common law or customary law, every court, 

tribunal or forum must promote the spirit, purport and objects of the Bill of 

Rights’. 

27 Woolman has made the point that the effect of sec 39(2) is to place a mandatory 

injunction on all courts to interpret all laws in a way that promotes the spirit, 

purport and objects of the Bill of Rights. Put differently, sec 39(2) establishes an 

interpretive canon that dispels any notions of judicial discretion. In accordance 

with this view, a failure by the lower courts to fulfil the obligation imposed by sec 

39(2) renders their decisions susceptible to constitutional review by the 

Constitutional Court. See S Woolman ‘Application’ in S Woolman et al (eds) 

Constitutional law of South Africa (2nd Edition, OS, 2005) ch 31. See also F 

Michelman ‘The rule of law, legality and the supremacy of the Constitution’ in 

Woolman et al (eds) (above) ch 11. 

28 Carmichele (n 26 above) para 56 (footnotes omitted). 

29 See Woolman ‘Application’ in Woolman et al (eds) (n 27 above) 31-93, for a 

discussion on the origins, application and problems with use of the phrase ‘an 

objective normative value system’ in South African constitutional jurisprudence.


to positive developments in the lower courts. Olitzki, 30 Premier, 

Western Cape v Fair Cape Property Developers, 31 and the Supreme 

Court of Appeal decision of Steenkamp 32 all turn on the grounding of 

fairness and reasonableness in terms of constitutional rights and 

constitutional norms. 33 

However, in spite of these positive developments, a couple of 

important questions are left unanswered. How does a court determine 

what public policy considerations to take into account when engaging 

with the wrongfulness inquiry in a novel situation? And further, is 

there a principled basis upon which courts go about balancing public 

policy considerations; or is it a purely a matter of judicial discretion? 

In fact, what becomes evident from a review of the South African 

academic literature and case law is that although the determination 

of wrongfulness is ultimately a legal question, no explicit legal rules 

or principles guide the selection or the deployment of public policy 

considerations. 34 

So despite the apparent mandatory injunction of s 39(2), the 

courts enjoy and exercise a wide discretion with regard to the type of 

public policy considerations that they deem appropriate and relevant 

in any particular situation. At a minimum, the litigants’ pleadings will 

influence the choice of policy considerations. However the courts, as 

was demonstrated in Carmichele, must consider, mero motu, other 

relevant policy considerations including those that flow from our basic 

law. While sound reasons exist for this type of judicial discretion — 

the inevitability of novel disputes — the same considerations do not 

30 Olitzki (n 21 above) para 31. 

31 Premier of the Province of the Western Cape v Fair Cape Property Developers 

(Pty) Ltd [2003] 2 All SA 465 (SCA) para 33. 

32 

Steenkamp NO v Provincial Tender Board, Eastern Cape [2006] 1 All SA 478 (SCA) 

para 25 (Steenkamp (SCA)). 

33 See Pharmaceutical Manufacturers Association of South Africa: In re Ex parte 

President of the Republic of South Africa 2000 2 SA 674 (CC) para 44. In this 

matter, Chaskalson CJ emphatically pronounced that ‘there is only one system of 

law. It is shaped by the Constitution which is the supreme law, and all law, 

including the common law, derives its force from the Constitution and is subject 

to constitutional control.’ Although the context and issues under consideration 

when Chaskalson CJ made this pronouncement were quite different from the 

issues under consideration presently, the potency of this holding can equally be 

extended to the present context. This is to say that the courts have, in my 

opinion, correctly recognised the overarching influence of the Constitution such 

that even in the determination of matters of public policy, such determination 

must be made within the rubric of constitutional norms and principles. See also B 

Leinius & JR Midgley ‘The impact of the Constitution on the law of delict’ (2002) 

119 South African Law Journal 20; Neethling et al (n 19 above) 36. 

34 See van Aswegen (n 25 above) 174. She considers the philosophical underpinnings 

of the nature of policy considerations and whether or not they can appropriately 

be considered to constitute legal rules when applied in judicial decision-making. 

Although Van Aswegen argues convincingly that policy considerations when 

invoked by judges are equally part of the law as are other legal sources, she does 

not deal with the question as to how judges should decide on what these 

considerations are to be, nor how they should be applied in a particular instance.


come into play when it comes to the determination of how competing 

public policy considerations should be assessed and reconciled. 

5.3 A comparative evaluation of the approach taken by the 

majority and the minority to the balancing of policy 

considerations 

In Steenkamp, both the majority and minority judgments took into 

account similar policy considerations with respect to the impact that 

an extension of delictual liability would have on the functioning of 

tender boards. 35 However, the two judgments came to vastly 

different conclusions as far as the weight to be attached to these 

considerations is concerned and whether the tender board acted 

wrongfully. The majority was clearly moved by arguments advanced 

against the extension of liability, whilst the minority was less 

persuaded. The arrival at different conclusions is not something that 

in itself is notable. Different judges will disagree regarding the 

construction of the facts and proper application of the law. I would 

even go so far as to contend that such differences of opinion amongst 

judges are a desirable feature of any legal system that is attuned to 

the contested views of the role of law from different sectors of a 

heterogeneous polity. However, for present purposes my concern is 

not with the different conclusions reached by judges of the majority 

and the minority. It remains, rather simply, focused on the manner in 

which they engage competing policy considerations. 

One major problem with the majority’s approach to wrongfulness 

is that it reveals a decidedly individualised inquiry that places undue 

emphasis on Balraz’s particular circumstances and fails to focus on 

the public policy considerations underlying its claim. 36 It would 

appear from the majority decision that although the majority does 

take cognisance of the public policy grounds raised by the plaintiff, it 

did not seriously engage them. 37 By way of contrast, the 

considerations of public policy in support of not extending liability to 

35 See Steenkamp (n 2 above) paras 49-50 & 55 (majority judgment) and paras 87-92 

(minority judgment). 

36 See Steenkamp (n 2 above) paras 43-46 (majority judgment). 

37 In fact Moseneke DCJ, Steenkamp (n 2 above) para 47, states: ‘I must at the 

outset say that the submissions of the applicant are attractive but not 

sustainable’.


tender boards, although also not engaged with at great length, 38 are 

considered in a decidedly more objective fashion that clearly 

identifies the public interests that they seek to promote without 

subjectivising the inquiry to focus on the conduct of this particular 

tender board. Ultimately, the competing public policy considerations 

pit the individualised considerations of the applicant against the 

potentially negative social consequences and public interest concerns 

that militate against the extension of delictually liability to tender 

boards. The entire exercise ineluctably results in the emasculation of 

the applicant’s claim. This diminishment of the applicant’s (public 

policy) interests not only creates an extra hurdle for the applicant, 

but unjustifiably shifts the focus away from the open and candid 

consideration of competing interests that a proportionality exercise 

demands when determining the question of wrongfulness. 

Put somewhat differently, the primary problem with the approach 

taken by the majority is its failure to identify and elaborate upon the 

values and interests underlying the applicant’s claim for a delictual 

remedy and properly weighing them up against the competing values 

and interests of the tender board as a public body. The applicant had 

an underlying right to administrative justice 39 and a right to 

appropriate relief. 40 And yet neither the right nor the desired remedy 

is engaged by the majority’s opinion. Instead the majority’s reasoning 

is predicated on the belief that the applicant should have shown that 

it had done everything within its power to mitigate its loss before it 

could allege wrongfulness. 

The minority opinion takes a markedly different approach to the 

identification, elaboration and assessment of competing policy 

considerations when determining the question of wrongfulness. The 

minority’s approach, although also clearly grounded in the factual 

matrix of the matter before the court, is more deliberate in the 

manner in which it engages with these competing considerations. By 

this I simply mean that the minority’s approach displays a more 

deliberate effort to identify the societal and public interests that 

38 

See Steenkamp (n 2 above) para 55: Moseneke DCJ holds that the majority concur 

with the ‘significant findings and conclusions’ of the Supreme Court of Appeal. It 

is interesting to note that Moseneke DCJ finds it adequate to concur in this 

fashion and simply leave the matter there. What makes this fact all the more 

peculiar is the fact that earlier on in the decision the Deputy Chief Justice goes to 

some length to justify the constitutional importance of the matter at hand; and 

its potential to influence the future development of the law in respect of appeals 

to private law remedies in matters that are clearly founded in public law. See 

Steenkamp (n 2 above) paras 17-23. 

39 

See sec 33(1) of the Constitution: ‘Everyone has the right to for the right to 

administrative action that is lawful, reasonable and procedurally fair.’ 

40 Section 38 of the Constitution affords a person a right to approach a court seeking 

‘appropriate relief’ where that person alleges that their rights have been 

infringed.


underpinned the applicant’s claim. 41 Although not expressly 

articulated in these terms, the minority approach to the assessment 

of public policy considerations would appear to be grounded in the 

principle of proportionality. 42 

41 The best way to illustrate this point is by way of example. The approach of the 

majority and that of the minority differ on the point of whether or not the law 

should be developed to accommodate a claim for out-of-pocket expenses 

instituted by a once successful tenderer. The majority position on this point is set 

out as follows: 

The residual question is whether there is justification to develop the 

common law to embrace this narrow claim for damages based on out-ofpocket 

expenses in favour of an initially successful tenderer where the 

award is subsequently set aside by the court and the tenderer retains the 

right to participate in the subsequent tender process. I think not. First, 

there is no magic in characterising financial loss as out-of-pocket. If 

public policy is slow to recompense financial loss of disappointed 

tenderers it should not change simply because of the name the financial 

loss bears. Second, even if there may not be a public law remedy such as 

an interdict, review or appeal this is no reason for resorting to damages 

as a remedy for out-of-pocket loss. This is so because first, as I found 

earlier, the loss may be avoided and second it is not justified to 

discriminate between tenderers only on the basis that they are either 

disappointed tenderers or initially successful tenderers. To do so is to 

allot different legal rights to parties to the same tender process. There is 

no justification for this distinction particularly because ordinarily both 

classes of tenderers are free to tender again should the initial tender be 

set aside. Steenkamp (n 2 above) para 54. 

The minority position on this point is set out as follows: 

A claim for out-of-pocket expenses, however, is a far more modest claim. 

Moreover, the inability to recover out-of-pocket expenses may well 

render smaller and less financially viable tenderers at risk of liquidation. 

Indeed, such was the case here. In our country, government procurement 

is one of the key mechanisms for ensuring that those previously locked 

out of economic opportunity by the policies of apartheid, are given an 

opportunity to participate. By definition such companies and individuals 

are often new, small and not financially robust. In our view, this is an 

additional and important factor which supports our conclusion that both 

Moseneke DCJ and the Supreme Court of Appeal are incorrect to conclude 

that aquilian liability cannot arise on the circumstances of this case. 

Steenkamp (note 2 above) para 80. 

42 See Steenkamp (n 2 above) para 69. Although Langa CJ and O’Regan make no 

direct reference to the principle of proportionality, it is, in my view, implicit in 

their approach to the balancing of policy considerations: ‘Determining whether 

conduct is wrongful or unlawful for the purposes of aquilian liability, as has been 

stated on many occasions, essentially involves a question of legal policy which 

must be answered in the light of the norms and values of our society and, since 

1994, in the light of the norms of our Constitution. Even though the question is a 

normative one, it is not to be answered on the basis of 'an intuitive reaction to a 

collection of arbitrary factors', but instead requires an identification of the 

relevant norms followed by an analysis and, if necessary, balancing of those 

norms to determine the outcome.’ (parqa 70) (emphasis mine and footnotes 

omitted.)


6 The prinicple of proportionalty: A possible 

basis upon which the balancing of public 

policy considerations should be performed? 

The principle of proportionality is firmly established in South African 

constitutional law. The invocation of the principle of proportionality 

has been experienced mainly within the context of limitations 

analysis. 43 The most fundamental idea underlying the principle of 

proportionality is that courts will candidly assess the competing 

rights, values and interests reflected in both the asserted 

constitutional norm and in the existing law that allegedly impairs that 

norm. 44 In other words, the principle of proportionality requires more 

than a mechanical exercise of summing up on some imagined scale the 

respective virtues of the two sides to the disputes. Chaskalson P, in S 

v Makwanyane, offers up a lucid exposition of the nature of the 

principle of proportionality and what it entails: 

The limitation of constitutional rights for a purpose that is reasonable 

and necessary in democratic society involves the weighing up of 

competing values, and ultimately an assessment based on 

proportionality. This is implicit in the provisions of section 33 (1). The 

fact that different rights have different implications for democracy, and 

in the case of our Constitution, for an ‘open and democratic society 

based on freedom and equality’, means there is no absolute standard 

which can be laid down for determining reasonableness and necessity. 

Principles can be established, but the application of those principles to 

particular circumstances can only be done on a case by case basis. This 

is inherent in the requirement of proportionality, which calls for the 

balancing of different interests. 45 

The value of the principle of proportionality would appear to be most 

apparent in those areas or instances where the courts must address 

rather novel issues, where no clear legal rule or precedent exists and 

where important matters of principle are at stake. 46 In such instances 

the principle of proportionality will require a court deciding upon such 

43 

Section 36 of the Constitution provides the test for the limitation of 

constitutional rights. 

44 See S Woolman & H Botha ‘Limitations’ in Woolman et al (eds) (n 27 above) ch 34; 

I Currie & J de Waal The Bill of Rights handbook (5th ed, 2005) 172-8; G Erasmus 

‘Limitation and suspension’ in D van Wyk et al (eds) Rights and constitutionalism: 

The new South African legal order (1994) 629. 

45 

1995 3 SA 391 (CC) para 104. This dictum has been cited with approval in many 

subsequent cases. See, eg, National Coalition for Gay and Lesbian Equality v 

Minister of Justice and Others 1999 1 SA 6 (CC) para 33. 

46 

An obvious example of such an issue is where a law is being challenged for 

violating the rights protected in the Bill of Rights. In most instances there will be 

some right demand — fundamental and therefore worthy of constitutional 

protection — being pitted against some law that will often seek to attain some 

important societal purpose.


an issue to make explicit, at least, the following: the exact nature of 

the conflict; what the competing or conflicting values or interests are 

and the importance of these; and the substantive reasons for choosing 

one set of values or interests over the other. In other words, the 

principle of proportionality demands that where the courts are 

required to engage in the exercise of balancing they should do so in a 

manner that clearly identifies all the pertinent competing values or 

interests, before weighing these up and providing fully substantiated 

conclusions. This type of application of the principle of 

proportionality is one that we have become accustomed to seeing at 

work in our bill of rights jurisprudence. 

The application of the principle of proportionality or at the very 

least an approach that is closely akin to it is nothing novel in delict. 

Courts have often had to consider whether or not to extend liability. 

In Carmichele, Judges Ackermann and Goldstone allude to the 

importance of [the principle of] proportionality: 47 

... in determining whether there was a legal duty on the police officer to 

act, Hefer JA in Minister of Law and Order v Kadir referred to weighing 

and striking a balance between the interests of the parties and the 

conflicting interests of the community. This is a proportionality exercise 

with liability depending upon the interplay of various factors. 

Proportionality is consistent with the Bill of Rights, but that exercise 

must now be carried out in accordance with the ‘spirit, purport and 

objects of the Bill of Rights’ and the relevant factors must be weighed in 

the context of [a] constitutional State founded on dignity, equality and 

freedom and in which government has positive duties to promote and 

uphold such values. (emphasis added) 

The core of the idea underlying the principle of proportionality and 

how it can add an important dimension to public policy considerations 

in the determination of wrongfulness was powerfully captured by 

Nugent JA in Van Duivenboden: 

When determining whether the law should recognise the existence of a 

legal duty in any particular circumstances what is called for is not an 

intuitive reaction to a collection of arbitrary factors, but rather a 

balancing against one another of identifiable norms. 48 (emphasis added) 

A similar approach to rethinking public policy considerations grounded 

in an application of the principle of proportionality was also followed 

in Steenkamp (SCA). 49 The Supreme Court of Appeal’s judgment — 

like the minority’s opinion in the Constitutional Court — illustrates 

how the principle of proportionality can be applied in the 


48 

Van Duivenboden (n 26 above) para 21. 

49 Steenkamp (SCA) (n 32 above).


determination of wrongfulness. The SCA’s decision accomplishes this 

task by meticulously identifying and engaging with the public policy 

considerations it deemed relevant to the determination of 

wrongfulness. 50 The SCA focused on the public and societal interests 

that underlay both the appellant’s claim for a remedy and the 

respondent’s denial of liability. Although the final decision with 

respect to wrongfulness in the Supreme Court of Appeal accords with 

that of the majority of the Constitutional Court, it is certainly more 

palatable. It gives due consideration to the competing rights and 

interests of both the applicant and the respondent in an objective and 

judicious manner. In short, the conclusion reached as to the question 

of wrongfulness in the Supreme Court of Appeal decision is more 

transparently reasoned and does not leave one with an uneasy sense 

that the exercise of balancing masks the subjective whims of 

judges. 51 

7 Conclusion 

Again. My discomfort with the Constitutional Court’s majority 

decision in Steenkamp is not with its conclusion that the conduct of 

the tender board was not wrongful. The matter really could have gone 

either way. However, my main concern and the reason for writing this 

note is to take issue with the unsatisfactory and somewhat opaque 

manner in which the majority reached its conclusion. The most 

profound shortcoming of this decision is that it leaves the 

Constitutional Court vulnerable to the critique that the ultimate 

conclusion as to wrongfulness reflects a mere subjective preference 

rather than a decision grounded in principle. That some may consider 

this charge to be unfair criticism is undoubted. However, as long as 

courts continue to engage in an amorphous balancing exercise, 

critiques such as mine will be difficult to rebut. A failure to develop 

a principled framework will always leave those who are sitting on the 

wrong side of a decision in which public policy considerations were 

considered with reason to be sceptical about what actually animated 

50 

See Steenkamp (SCA) (n 32 above) paras 29-44 

51 See Woolman & Botha (n 44 above) 34-93 to 34-104 9: The authors provide a 

sustained critique of the concept of balancing and its relationship with 

proportionality; in particular, the inappropriateness of the balancing metaphor in 

constitutional adjudication.


a court’s decision. 52 

Steenkamp has given us an opportunity to consider how the 

exercise of balancing policy considerations might be profitably 

supplanted by a more principled framework. The principle of 

proportionality is a useful departure point in the creation of such a 

framework. The appeal of this principle, to my mind, is that it is 

already established in our law, and more importantly, that it accords 

with the unarticulated practices that have already taken hold in our 

courts under the new constitutional dispensation. 

52 In a recent article Stuart Woolman has raised similar concerns about the 

Constitutional Court’s thin reasoning. Woolman objects to this minimalist mode of 

adjudicative reasoning by the Constitutional Court for several reasons, not least 

of which is that he considers it to be tantamount to abdication by the 

Constitutional Court of its responsibility to develop clearly articulated rules and 

to provide clearly substantiated reasons for its decisions. These clearly 

articulated rules and substantiated reasons are, according to Woolman, essential 

for purposes of developing a coherent constitutional jurisprudence that informs 

political discourse, as well as serving to guide the lower courts and the other 

coordinate branches of government. S Woolman ‘The amazing, vanishing Bill of 

Rights’ (2007) 124 South African Law Journal 762 784-787. See also F Michelman 

‘On the uses of “interpretive charity”’: Some notes on application, avoidance, 

equality and objective unconstitutionality from the 2007 term of the 

Constitutional Court of South Africa’ (2008) 1 Constitutional Court Review 1 

(Contends that while the judgments in question are thinly reasoned, they can, if 

patiently (and charitably) assessed, and if given sufficient expiation by 

commentators, be reconstructed so as to avoid some of the alleged problems with 

both their form and substance.)


SEVERING THE UMBILICAL CORD: A 

SUBTLE JURISPRUDENTIAL SHIFT 

REGARDING CHILDREN AND THEIR 

PRIMARY CAREGIVERS 

351 

Ann Skelton 

In S v M (Centre for Child Law as Amicus Curiae) 1 the Constitutional 

Court overturned a High Court judgment in which a primary caregiver 

of children had been sentenced to correctional supervision, including 

a brief period of imprisonment. The appellant had been convicted on 

multiple counts of fraud in the Regional Court. The sentence handed 

down by the High Court on appeal required her to undergo 

imprisonment prior to possible release on correctional supervision. 

Her further appeal to the Constitutional Court focused on the duties 

of a court when sentencing a primary caregiver, given the 

constitutional injunction that a child’s best interests are of 

paramount importance in all matters concerning the child. The 

Constitutional Court’s judgment subtly adjusts the balancing exercise 

at sentencing, as governed by the so-called ‘triad of Zinn’ principles. 

The judgment also provides an endorsement of the usefulness of 

correctional supervision, one of the benefits of which is the scope that 

it allows for the application of restorative justice. 

The more profound aspect of the judgment in S v M, however, lies 

in its development of child law, and in particular in what the 

Constitutional Court adds to its previous pronouncements on section 

28 of the Constitution 2 — that the section is a statement of realisable 

and enforceable rights. In S v M, the Court pays more detailed 

attention than previously to the ‘best interests of the child’ principle. 

The Court also expands for the first time on what the Constitution 

means when it says that the best interests of the child are of 

paramount importance in all matters concerning the child. This case 

note examines the focus in the judgment on the child’s rights as 

1 

2008 3 SA 232; 2007 12 BCLR 1312 (CC); 2007 2 SACR 539 (CC) (S v M). 

2 Constitution of the Republic of South Africa, 1996 (the Constitution).

352 Children and their primary caregivers 

distinct from the rights of their caregivers. It is argued that this 

distinction presents a new development in the Court’s child rights 

jurisprudence, and that some of the Constitutional Court’s earlier 

pronouncements on children in relation to their caregivers should be 

re-examined in the light of S v M. 

What are the duties of a sentencing court in the light of section 

28(2) of the Constitution when the person being sentenced is the 

primary caregiver of minor children? This question was the central 

pivot of the arguments and judgment in S v M. The case wound its way 

up to the Constitutional Court by means of a long journey that began 

in the Wynberg Regional Court. The applicant in the case was M, 3 the 

mother of three minor children. She was the sole caregiver of the 

children, and was also the main provider of financial support for their 

care. She had raised a bond on a modest home in which the family 

lived on the income she derived from two small businesses. She was 

convicted on various counts of fraud and theft and was sentenced 

initially by the Regional Court to four years imprisonment. 4 On 

appeal, the High Court confirmed her conviction on 38 counts of fraud 

and three of theft amounting to a total value of just over R19 000. The 

High Court set aside her sentence and replaced it with a sentence of 

four years in terms of section 276(1)(i) of the Criminal Procedure Act 

51 of 1977, a sentence which incorporates correctional supervision, 

subject to an initial portion of the sentence being served in prison. 

The High Court refused the applicant leave to appeal to the Supreme 

Court of Appeal, and that Court itself turned down a petition for leave 

to appeal. 5 

The effect of the High Court sentence was that the applicant 

would have had to serve no less than a total of eight months 

imprisonment before the Commissioner of Correctional Services could 

have considered releasing her on correctional supervision. When the 

applicant noted an appeal in the Regional Court against her sentence, 

that court denied her bail pending the determination of her appeal. 

Consequently, the applicant began serving her sentence on 29 May 

2003. The High Court granted the applicant appeal bail on 25 July 

2003, after she had served just under two months of her sentence. She 

remained on bail at the time when the matter was heard in the 

Constitutional Court. The fact that she had served two months in 2003 

meant that she was in effect facing a period of no less than six months 

of further imprisonment in terms of the sentence imposed by the High 

Court. 

3 

The Constitutional Court issued an order on the day of hearing that the citation of 

the case name should include only the initial of the applicant’s surname, in order 

to protect the identity of her children. 

4 

S v M (n 1 above) para 3. 

5 S v M (n 1 above) para 4.


The Constitutional Court decided to enrol her application for 

leave to appeal. The Chief Justice issued directions in which he posed 

three questions, 6 and required that the parties should confine the 

scope of their submissions to answering these. In addition to the 

central question of defining the role of a sentencing court when 

sentencing a primary caregiver of minor children, the Chief Justice 

also directed that once the duties had been identified, there should 

be consideration of whether the duties were observed in this case. 

The final question posed was that if the Court were to find that the 

duties were not observed, what order (if any) should the Court make? 

The Court had asked the Cape Bar to appoint a curator ad litem 

for the children, and the curator compiled a comprehensive report for 

the Court, in addition to making oral submissions. The Centre for Child 

Law, which had been granted leave to enter the case as amicus 

curiae, also presented ‘wide ranging written and oral submissions on 

the constitutional, statutory and social context in which the matter 

fell to be decided’. 7 

The applicant, the curator and the amicus all contended that the 

sentencing court had a duty to give specific and independent 

consideration to the effect that sentencing a caregiver to 

imprisonment would have on minor children. 8 It was conceded that 

there would be cases where the best interests of the child would be 

outweighed by the imperative to sentence a caregiver to 

imprisonment where considerations of proportionality required that. 

On the facts of M’s case, however, it was argued that due 

consideration of the best interests of the children led to the 

conclusion that the appropriate sentence would be one that did not 

require M to spend any further time in prison. It was proposed that 

conditions could be added to the sentence of correctional supervision 

to redress the impact of the crime on the victims and to recognise the 

interests of society. 

6 S v M (n 1 above) para 5. 


8 

S v M was not the first case to find that a sentencing court had such a duty. In S v 

Kika 1998 2 SACR 428 (W) the High Court set aside a sentence of a fine which 

resulted in imprisonment of a primary caregiver, finding that the magistrate had 

not considered the impact that the sentence might have on the children. The 

judge caused the circumstances to be investigated, and sent the matter back to 

the court a quo for the setting of a fresh sentence, pointing out a court must, 

when sentencing a primary caregiver, be satisfied that the welfare needs of the 

children will be met. The direct link with section 28(1)(b) and section 28(2) of the 

Constitution was first articulated in S v Howells 1999 1 SACR 675 (C); 1999 2 All SA 

233 (C) affirmed on appeal by the Supreme Court of Appeal in Howells v S 2000 

JOL 6577 (SCA). In that case the court considered the effect that imprisonment of 

a primary caregiver convicted of fraud would have on the children, but 

nevertheless decided to set a sentence of imprisonment, whilst making ancillary 

orders to ensure the children would be adequately cared for.


The response of the National Director of Public Prosecutions, 

supported in the main by the Department of Social Development and 

the Department of Justice and Constitutional Development, was that 

the current operation of the general principles of sentencing provided 

sufficient room for the consideration of the best interests of the 

primary caregiver’s children. Their submissions were that the courts 

a quo in this matter had given adequate consideration to the situation 

of the children, and that the sentence should not be interfered with. 

The majority judgment of the Court was written by Sachs J. 

Characterised by ringing language and memorable passages, the 

judgment is the Court’s clearest and most detailed explanation to 

date of the content and scope of children’s rights as set out in section 

28 of the Constitution. The judgment also makes some interesting 

observations about recent developments in the principles of 

sentencing, such as the growing recognition of restorative justice, 9 

and re-affirms the importance and usefulness of correctional 

supervision as a form of sentence. 10 For the sake of completeness, 

this case note will provide an overview of the judgment, including a 

brief consideration of the Court’s findings and observations on 

sentencing. The major part of the note will expand on the Court’s 

pronouncements on children’s rights, and in particular the Court’s 

interpretation of the paramount importance of the best interests of 

the child as stated in section 28(2) of the Constitution. 

2 Overview of the judgment 

The majority judgment proceeds by delivering responses to the key 

questions that had been posed by the Court. Expounding on the duties 

of the sentencing court in the light of section 28(2) of the Constitution 

when the person being sentenced is the primary caregiver of minor 

children, the Court begins by considering the current approach to 

sentencing. Sachs J finds that the Zinn 11 triad still retains the status 

of the ‘sentencing north star’. 12 However, this departure point must 

now be considered in the light of the Constitution, which has 

transformed the traditional aims of sentencing. 13 In the view of the 


10 

S v M (n 1 above) para 58-59. 

11 In S v Zinn 1969 2 SA 537 (A) at 540G-H the appellate division described what has 

to be considered at sentencing as ‘the triad consisting of the crime, the offender 

and the interests of society’. 

12 S v M (n 1 above) para 10 fn 4. 

13 S v M (n 1 above) para 10. The court makes reference to the case of Director of 

Public Prosecutions, KwaZulu-Natal v P 2006 3 SA 515 (SCA); 2006 1 All SA 446 

(SCA); 2006 1 SACR 243 (SCA); and Brandt v S 2005 2 All SA 1 (SCA); S v B 2006 1 

SACR 311 (SCA). These cases deal with the way in which the Constitution has 

changed the approach to sentencing of offenders who were below the age of 18 

years at the time of the commission of the offence.


Court, the question to be asked was whether the rights of children as 

set out in section 28 of the Constitution ‘add an extra element’ 14 to 

the responsibilities of a sentencing court over and above those 

imposed by the Zinn triad, and if so, how are those responsibilities to 

be fulfilled? In order to examine this question properly, the Court 

turned its attention to the significance of section 28(2) of the 

Constitution, namely the ‘best interests of the child’ provision. 15 The 

detail of this aspect of the judgment will be discussed below. For the 

moment it is sufficient to note that the Court reiterates that section 

28(2) is a self-standing right, as well as being a guideline for the 

balancing of other rights. The question is not whether section 28 

creates enforceable rights — which clearly it does — but rather ‘what 

reasonable limits can be placed on their application’? 16 In exploring 

this, the court tackles the issue of what is meant by ‘paramount’. The 

judgment strives to establish how the paramountcy principle operates 

in practice, particularly when the child’s rights are being weighed 

against other important competing rights. The weighing exercise in 

M’s case was that between the rights of the children not to be 

separated from parental or family care and to have their best 

interests considered, and the rights of society to be protected by the 

punishment of criminal misconduct. 

Having established the scope and significance of section 28, the 

inquiry shifted to what the proper approach of a sentencing court 

should be where the convicted person is the primary caregiver of 

minor children. Is it enough to consider the plight of the children as 

part of the Zinn triad — that part of the triangle dealing with the 

criminal and his or her circumstances? Counsel for the Director of 

Public Prosecutions had argued that the principles of sentencing 

provided sufficient scope to consider the situation of the children, 

and that the law needed no development. The Court took a different 

approach, favouring the view of the amicus that a child of a primary 

caregiver is not a ‘circumstance’ but an individual whose interests 

needed independent consideration. 17 The weight that would be 

accorded to those interests would depend on the facts of each case, 

but the Court determined that there was a duty on any court 

sentencing a primary caregiver to give separate and specific 

consideration to the impact that the intended sentence would have 

on the children. The Court took cognisance of information that had 

been placed before it about the effects on children of the 

imprisonment of their primary caregivers. Also considered was the 

international legal framework, the UN Convention on the Rights of the 

14 

15 


Section 28(2) reads as follows: ‘A child’s best interests are of paramount 

16 

17 

importance in every matter concerning the child’. 


S v M (n 1 above) para 30.


Child, and in particular the African Charter on the Rights and Welfare 

of the Child. Article 30(1) of that Charter contains a unique provision 

dealing expressly with ‘Children of Imprisoned Mothers’. Of particular 

relevance were clauses calling on states parties to ‘ensure that a noncustodial 

sentence will always be first considered when sentencing 

such mothers’; and to ‘establish and promote measures alternative to 

institutional confinement of such mothers’. Article 30(1) concludes by 

reminding the states parties that ‘the essential aim of the 

penitentiary system will be the reformation, the integration of the 

mother to the family and social rehabilitation’. 

The Court held as follows: 

[F]ocused and informed attention needs to be given to the interests of 

children at appropriate moments in the sentencing process. The 

objective is to ensure that the sentencing court is in a position 

adequately to balance all the varied interests involved, including those 

of the children placed at risk. This should become a standard 

preoccupation of all sentencing courts. 18 

The result of the judgment in S v M, therefore, is that in each case 

the sentencing court must give specific attention to the impact the 

sentence will have on the child or children of a primary caregiver. This 

does not mean that a primary caregiver will never, henceforth, be 

given a custodial sentence. The judgment explains quite clearly that 

the choice of the sentencing option least damaging to the interests of 

the children is made ‘within the legitimate range of choices in the 

circumstances available to the court’. 19 In other words, if a noncustodial 

option is clearly appropriate, the court must impose such a 

sentence, bearing in mind the interests of the children. If there is a 

range of appropriate sentences under consideration, the likely 

negative impact of imprisonment on the children of the primary 

caregiver will generally tip the balance in favour of a communitybased 

sentence. 

If the court cannot reasonably consider a non-custodial sentence 

in the circumstances, the court’s responsibilities do not end there. 

The judgment gives step-by-step guidelines indicating that 

if, on the Zinn triad approach the appropriate sentence is clearly 

custodial and the convicted person is a primary caregiver, the court must 

apply its mind to whether it is necessary to take steps to ensure that the 

children will be adequately cared for while the caregiver is 

incarcerated. 20 


19 

As above. 



Having enumerated the duties, the majority then applied them to 

the facts of M’s case and decided to overturn the High Court sentence. 

It was replaced with a sentence that allowed her to serve no further 

time in prison. The remainder of the sentence was suspended, on 

various conditions, including correctional supervision for a period of 

three years, and repayment to the victims of the crime of the monies 

of which they had been defrauded. In addition to the need to protect 

the children’s best interests, the fact that M had not committed 

offences for a long time, and had made demonstrable efforts to be a 

contributing member of society were strong factors influencing the 

choice of sentence. The Court pondered whether it should, having 

decided to set aside the order of the High Court, replace the sentence 

or refer the matter back for the court a quo to reconsider. It decided 

to set a new sentence itself, due to the fact that a significant amount 

of up-to-date and good quality information had been placed before it. 

The Court made several interesting observations regarding 

sentence. The judgment weighs up the advantages of correctional 

supervision versus a custodial sentence. 21 Correctional supervision is 

described as ‘a multifaceted approach to sentencing comprising 

elements of rehabilitation, reparation and restorative justice.’ 22 It is 

also a sentence that, if applied to those who are likely to respond 

positively, can protect society without the destructive impact that 

imprisonment can have on innocent family members. The Court 

remarked on prison overcrowding, pointing out that rehabilitation is 

difficult to achieve under such circumstances, and that therefore 

correctional supervision is a sentence which is more likely to result in 

rehabilitation. 23 The Court went on to say that ‘[a]nother advantage 

of correctional supervision is that it keeps open the option of 

restorative justice in a way that imprisonment cannot do’. The court 

21 

The judgment recalls the Court’s earlier judgment in S v Williams and Others 

1995 3 SA (CC), per Langa J as he then was: ‘The development of this process [of 

correctional supervision] must not be seen as a weakness, as the justice system 

having “gone soft.”’ 

22 S v M (n 1 above) para 59. The Court refers to the following judgments regarding 

correctional supervision: S v R 1993 1 SA 476 (A); S v Siebert 1998 1 SACR 554 

(SCA); S v Ingram 1995 1 SACR 1 (SCA). 

23 S v M (n 1 above) para 61. The Court cites S v Lebuku 2006 JOL 17622 (T) para 13 

– 15 with approval, where Webster J refers to the 2003/2004 Annual Report of the 

Inspecting Judge of Prisons in which Judge Fagan recommended the use of noncustodial 

sentences in appropriate cases, in order to avoid further over-crowding.


explained that central to the notion of restorative justice24 is the 

recognition that the community, and not the criminal justice system, 

that is the site of crime control. The sentence handed down by the 

Court demonstrates a restorative justice lens, as in addition to being 

confined to her home when not working, and being required to do 10 

hours of community service per week, M was required to pay back her 

victims, which she had offered to do. With regard to this, a particular 

point is made: 

It would have special significance if she is required to make the 

repayments on a face-to-face basis. This could be hard for her, but 

restorative justice ideally requires looking the victim in the eye and 

acknowledging wrongdoing. 25 

Writing for the minority, 26 Madala J confirmed the principles decided 

upon by the majority. He fully agreed that the sentencing court does 

have a duty to consider the rights of the children when sentencing a 

primary caregiver. However, applying that approach, Madala J came 

to a different conclusion. The fact that M was a repeat offender, and 

that she had committed further offences whilst on bail were factors 

that, according to the minority, outweighed the children’s interests. 

The overall judgment is not weaker for the minority judgment. The 

principles are common to both, although the judges differed on the 

outcome after all the relevant factors had been considered. 

3 Development of Child Law 

S v M has developed child law significantly, through its detailed 

enunciation of the scope and application of section 28. The concept 

of best interests is put under the spotlight, and the Court spells out 

the way in which ‘paramountcy’ referred to in section 28(2) should be 

24 

The court cites the following sources on restorative justice: D Pinnock ‘What kind 

of justice?’ University of Cape Town, Institute of Criminology Occasional Paper 

Series 4-95 (1995) available at http://web.uct.ac.zadepts/sjrp/publicat/ 

25 

whatkind.htm, accessed on 16 August 2007; and M Batley in Maepa (ed) Beyond 

retribution: Prospects for restorative justice in South Africa Institute for Security 

Studies, Monograph 111 ‘Restorative justice in the South African context’ 21 

http://www.iss.co.za/pubs/Monographs/No111/Chap2.htm (S v M (n 1 above) n 

67). See also A M Skelton ‘The influence of the theory and practice of restorative 

justice in South Africa with special reference to child justice’ unpublished LLD 

thesis, University of Pretoria, 2005; A Skelton & M Batley ‘Charting progress, 

mapping the future: Restorative justice in South Africa’ Institute for Security 

Studies and Restorative Justice Centre (2006). 

In Dikoko v Mokhatla 2006 6 SA 235 (CC), in the dissenting judgment of Sachs J, at 

para 114, the key elements of restorative justice are listed as encounter, 

reparation, reintegration and participation. See also the dissenting judgment of 

Mokgoro J, particularly at para 68. Since S v M was heard there have been three 

other judgments reported which expound upon restorative justice, see S v 

Maluleke 2008 1 SACR 49 (T); S v Shilubane 2008 1 SACR 295 (T); S v Saayman 

26 

2008 1 SACR 393 (E). 

Nkabinde J and Navsa AJ concurred with the dissenting judgment.


understood and applied. The Court acknowledges that the concept of 

best interests is not new to South African courts, as it has long been 

used as the applicable standard in matters such as care, contact or 

maintenance. In the family law field it is everyday currency, its most 

common application being where adult parties are in dispute about 

the care of or contact with a child. In such cases, the decisions will 

always be guided fundamentally by what is in the child’s best 

interests. 

However, in the new constitutional order, ‘the scope of best 

interests has been greatly enlarged’. 27 Section 28(2) declares the best 

interests of the child to be of paramount importance, not just in 

family matters, but in all matters concerning the child. At first blush 

this formulation seems to have such an all-encompassing reach that 

questions should be asked about how it can be effectively used in 

practice. To illustrate this, Sachs J quoted from the judgment of 

Jooste v Botha 28 in which Van Dijkhorst J was so concerned that 

section 28(2) would over-ride all other legitimate interests, that he 

felt constrained to interpret the provision only as a general guideline 

and not as a rule of law of horizontal application. Sachs J points out 

that Van Dijkhorst J was wrong about this. He explains that while 

section 28 undoubtedly serves as a general guideline to the courts, its 

‘normative force’ does not stop there. 29 In fact, the Constitutional 

Court has previously described section 28(2) as an independent, selfstanding 

right, 30 and as ‘an expansive guarantee’. 31 Sachs J sums up 

the unequivocal position of the Court: section 28 establishes a set of 

children’s rights that courts are obliged to enforce. 32 

This leads to another question: If the ambit of children’s rights is 

so wide, in what circumstances can such rights be limited? It is a 

question that has been pondered by academics for a number of years. 


28 

2000 2 SA 199 (T). 


30 In Minister of Welfare and Population Development v Fitzpatrick and Others 2000 

3 SA 422 (CC); 2000 3 BCLR 713 (CC) per Goldstone J: ‘The plain meaning of the 

words clearly indicates that the reach of section 28(2) cannot be limited to the 

rights enumerated in section 28(1) and section 28(2) must be interpreted to 

extend beyond those provisions. It creates a right that is independent of those 

specified in section 28(1).’ 

31 Sonderup v Tondelli 2001 1 SA 2001 (CC); 2001 2 BCLR 152 (CC) (Sonderup) para 

29. 



Clark, 33 writing in 1998, took the view that children’s rights would 

trump opposing rights. However, her comments related to family law, 

where the best interests principle is considered the determining 

factor. Bekink and Bekink considered the matter more broadly, 

arguing that ‘paramount’ means that in weighing up competing 

interests, the scales must tip in favour of the child. 34 

In the case of De Reuck v Director of Public Prosecutions, 

Witwatersrand Local Division and Others 35 the Constitutional Court 

provided some clarity on the issue. In the High Court decision in the 

same matter, the Court had found that children’s best interests can 

never be trumped by the rights of others. 36 The Constitutional Court 

disagreed, reiterating its approach that constitutional rights are 

mutually interrelated and interdependent, forming a single 

constitutional value system. Citing its earlier judgment in Sonderup, 37 

the Constitutional Court held that section 28(2), like other rights 

enshrined in the Bill of Rights, is subject to limitations that are 

reasonable and justifiable in accordance with section 36. 

4 What is the meaning of ‘paramount 

importance’ of children’s best interests? 

The question that remained hovering in the air even after the 

Constitutional Court had ruled in De Reuck was: What is to be 

understood by the phrase ‘paramount importance’? Friedman and 

Pantazis 38 posed the question thus: ‘Indeed, if a child’s best interests 

are not always supreme, what is the point of section 28(2)?’ 

The authors went on to observe that section 28(2) is highly 

unusual, because it is the only section in the Constitution that applies 

to a group of people in relation to all aspects of their lives. In their 

view the section appears to be aimed at creating a right for children 

33 

B Clark ‘Competing custody rights: New concepts of family and the best interests 

of the child’ (1998) 35 Comparative and International Law Journal of South 

Africa 288 290. See further J Heaton ‘Some general remarks on the concept “best 

interests of the child”’ (1990) 53 Tydskrif vir Hedendaagse Romeins Hollandse Reg 

96; B Clark ‘A golden thread? Some aspects of the application of the standards of 

the best interests of the child in South African family law’ (2000) Stellenbosch 

Law Review 3; F Mahlobogwane ‘South African courts and the “best interests of 

the child” in custody disputes’ (2005) 38 Comparative and International Law 

Journal of South Africa 246. 

34 

B Bekink & M Bekink ‘Defining the standard of the best interest of the child: 

Modern South African perspectives’ (2004) 37 De Jure 21 26. 

35 2004 1 SA 406 (CC); 2003 12 BCLR 1333 (CC) para 55. 

36 

De Reuck v Director of Public Prosecutions, Witwatersrand Local Division 2003 3 

SA 389 (W) (De Reuck) para 45. 

37 2001 1 SA 1171 (CC); 2001 2 BCLR 152 para 27-30. 

38 

A Friedman & A Pantazis ‘Children’s rights’ in S Woolman et al (eds) 

Constitutional law of South Africa (2nd Edition, OS, 2006) 47-34 - 47-35.


as children — addressing the vulnerability of children, and ensuring 

that their rights do not, as in the pre-constitutional era, frequently 

have to give way to the rights of others. Friedman and Pantazis 

concluded that section 28(2) implies that in every matter where a 

child’s rights are (substantially) involved, those interests must be 

taken into account. In addition, they said that a child’s interests have 

a leg up vis-à-vis other rights and values, though this does not amount 

to these rights necessarily acting as trumps. 39 The judgment in S v M 

seems to have proved them right. 

Sachs J states that what lies at the heart of section 28 is the right 

of the child to be a child and enjoy special care. In a poetic passage 

of the judgment, he explains what he means by this: 

Every child has his or her own dignity. If a child is to be constitutionally 

imagined as an individual with a distinctive personality, and not merely 

as a miniature adult waiting to reach full size, he or she cannot be 

treated as a mere extension of his or her parents, umbilically destined to 

sink or swim with them. The unusually comprehensive and emancipatory 

character of section 28 presupposes that in our new dispensation the sins 

and traumas of fathers and mothers should not be visited on their 

children. Individually and collectively all children have the right to 

express themselves as independent social beings, to have their own 

laughter as well as sorrow, to play, imagine and explore in their own 

way, to themselves get to understand their bodies, minds and emotions, 

and above all to learn as they grow how they should conduct themselves 

and make choices in the wide social and moral world of adulthood. And 

foundational to the enjoyment of the right to childhood is the promotion 

of the right as far as possible to live in a secure and nurturing 

environment free from violence, fear, want and avoidable trauma. 40 

This forms the crux of the Court’s decision that sentencing courts do 

have a duty to consider the rights of the children when sentencing the 

primary caregiver. The state’s action — imprisonment of an adult — 

causes the removal of the caregiver from the family home. This 

directly impacts on the child’s right to family and parental care. The 

Court went on to say that although no constitutional injunction can 

completely protect children from life’s hard knocks, the law can 

create conditions to protect them from abuse or neglect. In S v M, the 

Court defined the legal duty as ensuring that the best interests of the 

child are considered by the sentencing court wherever there is a 

likelihood that the primary caregiver will be removed from the family 

through imprisonment. The likely effect, where the range of available 

sentencing options permits it, is to give children’s rights ‘a leg up’ by 

keeping the caregiver in the family. Where a custodial option is 

considered by the sentencing court to be the only legitimate course, 

39 

Friedman & Pantazis (n 38 above) 47-35. 



then the court must nevertheless be satisfied that the children will be 

properly cared for and may make orders to minimise the negative 

effects on the children. 

If children’s rights do not always win out over other rights, what 

is the meaning of ‘paramount importance?’ The judgment embarks on 

an exploration of this thorny question. Sachs J comments that the 

very expansiveness of the paramountcy principle appears to promise 

everything but deliver little in particular. 41 The best interests 

concept is indeterminate, resulting in differing views about its 

meaning by professionals and judges. These problems accepted, the 

Court has recognised that it is precisely the contextual nature and 

inherent flexibility of section 28 that constitutes the source of is 

strength. The determination will depend on the circumstances of each 

case, and this is not a weakness, but a strength. A truly child-centred 

approach requires an in-depth consideration of the needs and rights 

of the particular child in the ‘precise real-life situation’ he or she is 

in. To apply a pre-determined formula for the sake of certainty, 

irrespective of the circumstances, would in fact be contrary to the 

best interests of the child. 42 

The exercise of weighing up the best interests of the child has 

been clearly articulated by the Court. S v M has built on the Court’s 

previous jurisprudence, 43 creating a nuanced model that can be 

applied in a range of different legal and factual contexts. A more 

difficult question, concedes Sachs J, is to ‘establish an operational 

thrust for the paramountcy principle’. S v M goes further than any 

previous judgment, though it still defines the principle more by 

stating what it is not, than by saying what it is. It is not an 

‘overbearing and unrealistic trump’, 44 it cannot be interpreted ‘to 

mean that the direct or indirect impact of a measure or action on 

children must in all cases oust or override all other considerations.’ 

Sachs J concludes that ‘the fact that the best interests of the child 

41 

S v M (n 1 above) para 23. See also H Reece ‘The paramountcy principle: 

Consensus or construct?’ (1996) 49 Current Legal Problems 267; D Matlala ‘The 

law reports: Parent and child — paramountcy of a child’s best interests’ (2006) 

June De Rebus 41. 

42 The Court has re-stated this position in the subsequent judgment of AD and 

Another v DW and Others (Centre for Child Law as Amicus Curiae; Department 

for Social Development as Intervening Party) 2008 3 SA 183 (CC), per Sachs J para 

55: ‘Child law is an area that abhors maximalist legal propositions that preclude 

or diminish the possibilities of looking at and evaluating the specific 

circumstances of the case ... This means that each child must be looked at as an 

individual, not as an abstraction. It also means that unduly rigid adherence to 

technical matters, such as who bears the burden for proof, should play a 

relatively diminished role; the courts are essentially guarding the best interests 

of a child, not simply settling a dispute between litigants.’ 

43 Minister of Welfare and Population Development v Fitzpatrick and Others 2000 3 

SA 422 (CC); De Reuck (n 35 above); Sonderup (n 30 above). 



are paramount does not mean that they are absolute.’ 45 To 

acknowledge all of these realities is important because if the best 

interests principle is spread ‘too thin’ it risks becoming devoid of 

meaning, instead of promoting the rights of children as it was 

intended to do. 46 

Later in the judgment the operation of the principle is explained 

in a more positive manner. The judgment states that sentencing 

officers should pay appropriate attention to the children of a primary 

caregiver and take reasonable steps to minimise damage. It continues 

thus: 

The paramountcy principle, read with the right to family care, requires 

that the interests of children who stand to be affected receive due 

consideration. It does not necessitate overriding all other 

considerations. Rather, it calls for appropriate weight to be given in each 

case to a consideration to which the law attaches the highest value, 

namely the interests of children who may be concerned. 47 

5 The child’s rights separate from adult’s rights: 

a subtle shift 

A further important feature of the majority judgment in S v M is the 

fact that it carefully focuses on the best interests of the child, and 

does not get entangled in a debate about the issues relating to the 

rights of the primary caregiver. The discourse centres on children’s 

rights to family and parental care, and their right to have their best 

interests given appropriate weight. Counsel for the National Director 

of Public Prosecutions attempted to argue that if the sentencing court 

considers children’s rights, then primary caregivers will gain an unfair 

advantage in the sentencing process. The majority of the Court 

rejected this out of hand, saying that ‘the issue is not whether parents 

should be allowed to use their children as a pretext for escaping the 

otherwise just consequences of their own misconduct.’ 48 That, said 

the Court, would be a mischaracterisation of the interests at stake. 

The purpose of emphasising the duty of the sentencing court to 

consider the best interests of the child is not to allow errant parents 

to escape punishment, but rather to protect innocent children from 

avoidable harm. 



47 


48 S v M (n 1 above) para 35. See, however, Madala J’s dissenting judgment at para 

117, where he states that the Court must guard against creating a perception that 

encourages offenders to use the interests of children as ‘a tool in the judicial 

process’.


This focus on the child’s rights separate from the rights of their 

caregivers heralds a subtle but significant shift in the Constitutional 

Court’s child rights jurisprudence. Read together with the strong 

statements earlier in the judgment about children being possessed of 

their own dignity, and being separate beings who are not merely 

destined to sink or swim with their parents, the Court’s ‘new’ 

enunciation of children’s rights calls for a re-examination of some of 

its earlier pronouncements. 

The High Court decision in Grootboom v Oostenberg Municipality 

& Others 49 concerned an application brought by a group of 390 adults 

and 510 children (all of whom were living with their parents) who had 

not been provided with adequate housing and who, at the time of the 

initial application, were living on a sports field, following their 

eviction from land destined for low-cost housing. The applicants 

succeeded in their argument that because children have an 

unqualified right to shelter in terms of section 28(1)(c), coupled with 

the fact that they have a right to be cared for by their families or 

parents, children had the right to be provided with shelter (together 

with their parents) until such time as their parents could provide them 

with shelter. Davis J was careful to point out that the children were 

the bearers of the right. The fact that they should not be separated 

from their parents in order to access their rights did not amount to a 

derivative right for the parents (as had been argued by the 

applicants), but rather that ‘an order that enforces a right to shelter 

should take account of the need of the child to be accompanied by his 

or parent’. 50 

The approach of the majority in S v M is reminiscent of this way 

of thinking. The similarities are that in both cases the judgments 

make it clear that the rights accrue to the child, but because of the 

imperative to keep children in their families wherever possible, the 

adult caregiver might also benefit from the required order. 

However, the Constitutional Court did not uphold the judgment of 

Davis J. In Government of the Republic of South Africa & Others v 

Grootboom & Others, 51 the Court found that section 28(1)(c) did not 

create any primary obligations on the State to provide shelter on 

demand to children and their parents, but rather that the primary 

obligation to care for children lay with their parents, and that the 

state only has the obligation in the alternative to provide shelter 

49 2000 3 BCLR 277 (C) (Grootboom 1). 

50 Grootboom 1 (n 48 above) 289D. 

51 2001 1 SA 46 (CC); 2000 11 BCLR 1169 (CC) (Grootboom 2).


when, for example, children are removed from their families. 52 The 

Court found that responsibilities of the state to children in the care of 

their families extended to providing the legal and administrative 

infrastructure to ensure that children are protected in terms of 

section 28(1). Section 28(1)(c), although it contains no internal 

qualifier suggesting that it is subject to progressive realisation, does 

not, according to Grootboom, provide children with a direct and 

immediately enforceable right to housing. 53 The Court was concerned 

that ‘[c]hildren could become stepping stones to housing for their 

parents instead of being valued for who they are’. This statement jars 

with the fact that children in South Africa receive other poverty or 

care-related benefits, such as child support grants, 54 and there is no 

serious suggestion that this might lead to them being valued for the 

money that their presence brings into the household rather than for 

who they are. 55 

In S v M the majority of the Court does not get distracted by any 

ideas that primary caregivers might use their children as ‘stepping 

stones’ to avoid imprisonment, although this kind of argument was 

advanced by counsel for the National Director of Public Prosecutions. 

The minority judgment in S v M does, however, get mired in this way 

of thinking. Although Madala J agrees with the majority on the 

principle that the children of primary caregivers should have their 

best interests considered and protected, he is concerned about 

children being ‘used’ by adults seeking to avoid imprisonment. He 

cautions that ‘[t]his Court should be wary of setting a precedent that 

creates a perception that courts will give primary caregivers a 

sentence that is disproportionate to what they deserve and which 

encourages them to use the interests of children as a tool in the 

judicial process.’ 56 These remarks, in contrast to the explanation in 

the majority judgment that the children are the bearers of the rights 

52 Grootboom 2 (n 51 above) para 77. The judgment did not state clearly that 

children living with their parents will not have a claim for the enforcement of 

their rights in terms of section 28(1)(c), as it uses the term ‘for example’ when 

referring to children removed from their primary caregivers. 

53 The Court came to the puzzling conclusion that there was no difference between 

the right to ‘adequate housing’ provided for in section 26 and ‘shelter’ in section 

28. Friedman & Pantazis (‘Children’s Rights’ in Woolman et al (n 37 above) 46-8) 

have criticised this interpretation by the court, firstly because it ignores the plain 

meaning in the two different phrases, and also because the Constitutional 

drafters’ choice of different terminology suggests different conceptual 

extensions. 

54 In the case of Khoza and Others v Minister of Social Development and Others; 

Mahlaule and Another v Minister of Social Development and Others 2004 6 BCLR 

569 (CC) the Court recognised that denial of support to South African born 

children of foreign permanent residents in South Africa was not only 

discrimination on the grounds of the parents’ nationality, but also trenches on the 

rights that children have under section 28(1)(c). 

55 Friedman & Pantazis (n 38 above) 47-6 n 2: ‘In any case, the state is committed to 

providing all needy children with child care grants.’ 



and to suggest otherwise is a mischaracterisation, strike a chord that 

is an echo of children being used as ‘stepping stones’. 

Much disappointment has been expressed with regard to the 

Constitutional Court’s decision in Grootboom, viewed from a 

children’s rights perspective. 57 However, it has also been 

acknowledged by several writers 58 that the effect of the decision in 

Grootboom — vis-à-vis children’s rights — underwent some 

adjustment in the case of Minister of Health & Others v Treatment 

Action Campaign & Others. 59 In that case, which dealt with access to 

treatment to avoid mother-to-child transmission of HIV/AIDS, the 

Court held as follows: 

The state is obliged to ensure that children are accorded the protection 

contemplated by s 28 that arises when the implementation of the right 

to parental or family care is lacking. Here we are concerned with 

children born in public hospitals and clinics to mothers who are for the 

most part indigent and unable to gain access to private medical 

treatment which is beyond their means. 60 

So the approach adopted in Grootboom to the effect that children 

living with their parents will have to look to their parents, rather than 

the state, for the fulfilment of their section 28(1)(c) rights was 

ameliorated. 61 Although the two cases are clearly distinguishable by 

the fact that shelter is arguably easier for parents to provide than a 

specialised form of health care, it was significant that the TAC Court 

57 

J Sloth-Nielsen ‘The child’s right to social services, the right to social security, 

and primary prevention of child abuse: Some conclusions in the aftermath of 

Grootboom’ (2001) 17 South African Journal on Human Rights 210; E Bonthuys 

‘The South African bill of rights and the development of family law’ (2002) 119 

South African Law Journal 748; M Pieterse ‘Reconstructing the private/public 

dichotomy? The enforcement of children’s constitutional social rights and care 

entitlements’ (2003) Tydskrif vir Suid-Afrikaanse Reg 1; S Liebenberg ‘Taking 

stock: The jurisprudence of children’s socio-economic rights and its implications 

for government policy’ (2004) 5:4 ESR Review 2; and K McLean ‘Housing’ in S 

Woolman et al (eds) Constitutional law of South Africa (2nd Edition, OS, 2006) 

55.51 – 55.54. 

58 J Sloth-Nielsen ‘Children’ in D Davis & H Cheadle (eds) The South African 

constitution: The bill of rights (2nd ed [date]) 421; Friedman & Pantazis (n 38 

above) 47-13; E Bonthuys ‘Children’ in I Currie & J de Waal (eds) The bill of rights 

handbook (5th edition [date]) 613. 

59 

2002 1 BCLR 1033 (CC) (TAC). 

60 TAC (n 59 above) para 79. 

61 S Liebenberg (‘The judicial enforcement of social security rights in South Africa’ 

in E Riedel Social Security as a Human Right (2007) 69) points out that the Court 

did not conclude that children enjoyed an unqualified, direct claim to the 

provision of basic health care services, but rather based their finding on the fact 

that the government’s policy was unreasonable because it excluded a particularly 

vulnerable group.


gave clarity that section 28(1)(c) rights do not only set an obligation 

for the state where the child is separated from the parent. 62 

6 Conclusion 

S v M appears to herald a new era in which the fear of children being 

used as instruments by their parents to advance their parents’ rights 

have been supplanted by a clear recognition of children as rights 

bearers. Comparisons have been adduced in this article to show that 

this approach may indicate a subtle shift by the Court to see children 

as separate from their parents and therefore not ‘umbillically 

destined to sink or swim with them’. 

There are, of course, significant differences between cases such 

as Grootboom and TAC on the one hand, and S v M on the other. The 

first-mentioned cases deal with socio-economic rights of children. 

This means that the orders given in those matters potentially had 

significant cost implications. S v M does not create major cost 

implications, it requires only an investment of time and care on the 

part of state officials already employed. In such cases it is clearly 

easier for the Court to make the kind of pronouncements that it has 

made in S v M. Grootboom and TAC can themselves be separated out 

from one another. Health care is not something that parents can ever 

fully provide unless assisted either by private health care providers, 

or if the parents are poor, by the state. Although the judgments do 

not fully explain the difference, commentators have mentioned that 

nutrition and shelter are more likely to be within the scope of what 

parents can be expected to deliver, whilst health care and social 

services are more likely to require input by the state. 63 

Children of primary caregivers who are due to be sentenced pose 

an interesting counter-point. They are still living with their parents, 

but are at risk of being separated from them. S v M found that the 

state has a duty to create a positive legal environment, and thus 

sentencing courts now have clear duties that have been articulated by 

the Court. These arise from the rights of the children, and the 

majority judgment had no difficulty in dispelling any concerns about 

62 

The jurisprudence related to the socio-economic rights of children who are 

separated from their parents has gradually developed, with the courts finding 

that such children do have an immediate and directly enforceable claim against 

the state for the rights included in section 28(1)(h): see Centre for Child Law and 

Another v Minister of Home Affairs and Others 2005 6 SA 50, which dealt with the 

rights of unaccompanied foreign children to receive social services, and Centre 

for Child Law and Others v MEC for Education, Gauteng and Others 2007 1 SA 223 

(T), which dealt with children who had been removed from state care and placed 

in a school of industries where they were living in extremely poor conditions, 

without adequate social services. 

63 Friedman & Pantazis (n 38 above) 47-12; J Sloth-Nielsen (n 56 above) 225.


the primary caregivers reaping unfair advantages from the fact that 

they had children. The only distinction that can be made is that S v M 

deals with the right to family or parental care, and does not deal 

directly with socio-economic rights of the children. 

The effect of S v M is ultimately to force the state to create a legal 

environment to facilitate children’s rights being recognised and their 

best interests being met as far as possible through the orders of 

sentencing courts. To that extent, the judgment accords with the role 

of the state as outlined in Grootboom, which was to create the 

regulatory environment for the achievement of children’s rights. 

However, in the way that it ‘constitutionally imagine[s]’ 64 children as 

separate rights bearers with separately identifiable and enforceable 

rights, S v M takes children’s rights a jurisprudential step forward.

CCR Intro.fm - PULP - University of Pretoria

Create successful ePaper yourself

Delete template?

Save as template?