22.02.2013 Views

Separability of Light Verb Constructions in Persian*

Separability of Light Verb Constructions in Persian*

Separability of Light Verb Constructions in Persian*

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>Separability</strong> <strong>of</strong> <strong>Light</strong> <strong>Verb</strong> <strong>Constructions</strong> <strong>in</strong> <strong>Persian*</strong><br />

Gholamhosse<strong>in</strong> Karimi-Doostan, University <strong>of</strong> Kurdistan, Sanandaj<br />

E-mail: gh5karimi@uok.ac.ir<br />

‘Abstract’<br />

Persian <strong>Light</strong> <strong>Verb</strong> <strong>Constructions</strong> (LVCs), consist<strong>in</strong>g <strong>of</strong> a light verb and a<br />

preverb, fall <strong>in</strong>to two groups: separable and <strong>in</strong>separable. The nom<strong>in</strong>al preverbs <strong>in</strong> the<br />

former, unlike those <strong>in</strong> the latter, can function as direct object DP argument, can be<br />

modified by an adjective, and can be relativized, scrambled and focused on by Wh<strong>in</strong>terrogatives.<br />

In this paper, it is argued that the (<strong>in</strong>)separability <strong>of</strong> LVCs cannot<br />

necessarily be an <strong>in</strong>dication <strong>of</strong> the fact that LVCs are formed at two levels <strong>of</strong><br />

representation. It is shown that (<strong>in</strong>)separability <strong>of</strong> LVCs follows from the semantic<br />

and morphosyntactic properties <strong>of</strong> preverbs and light verbs <strong>in</strong> the language.<br />

Key words: Persian complex predicates, light verbs, case mark<strong>in</strong>g, a-structure<br />

1. Introduction<br />

After the sem<strong>in</strong>al article <strong>of</strong> Grimshaw and Mester (1988) on the Japanese light<br />

verb suru 'to do', many researchers have focused on light verbs and complex forms <strong>of</strong><br />

verbs <strong>in</strong> different languages (Ahn, 1991; Sato, 1993; Kim, 1994; Mohanan, 1994; Rosen,<br />

1989; Butt, 1995; 2003; Dies<strong>in</strong>g, 1998; 2000; Choi and Wechsler 2001; Butt and<br />

Ramchand 2005; among many others). Persian complex verbs have also attracted some<br />

researchers (Mohammad and Karimi, 1992; Vahedi-Langroudi, 1996; Karimi, 1997;<br />

Karimi Doostan, 1997; Megerdoomian, 2001; and Folli, Harley and Karimi, 2005; among<br />

others). In fact, most <strong>of</strong> the Persian verbs are <strong>in</strong> complex forms consist<strong>in</strong>g <strong>of</strong> a light verb<br />

and a preverb (PV) 1 . The preverbs (PVs) are nouns, adjectives, adverbs or prepositions<br />

(Karimi, 1987 and Folli, Harley and Karimi 2005) and the light verbs lack sufficient<br />

*. I would like to thank the unanimous reviewers <strong>of</strong> Studia L<strong>in</strong>guistica for their enlighten<strong>in</strong>g comments. I<br />

am also grateful to Mrs. Lamiah Hashemi, Mr. Thaher Sarhadi and Mr. Salman Ashabi for read<strong>in</strong>g and<br />

correct<strong>in</strong>g some <strong>of</strong> the grammatical and spell<strong>in</strong>g errors <strong>in</strong> the earlier version <strong>of</strong> this paper.<br />

1 . The follow<strong>in</strong>g abbreviations are used <strong>in</strong> this article: Def. = def<strong>in</strong>ite article; DOM = direct object marker;<br />

DP = determ<strong>in</strong>er phrase; Ez. = Ezafe particle; Fu. Aux = future auxiliary; Impf= imperfective; Indef. Art =<br />

<strong>in</strong>def<strong>in</strong>ite article; Neg = negation particle; NegP = negation phrase; PV = preverb; Prog. Aux. = progressive<br />

auxiliary; VN = verbal noun. Where necessary the terms are def<strong>in</strong>ed <strong>in</strong> the text.


semantic force to function as <strong>in</strong>dependent predicates (Vahedi-Langroudi, 1996; Karimi-<br />

Doostan, 1997; 2005). Any Persian complex verb, referred to as <strong>Light</strong> <strong>Verb</strong><br />

<strong>Constructions</strong> (LVCs) henceforth, can be broken by the Negation Particle (Neg), the<br />

imperfective morpheme mi as <strong>in</strong> (1), modals and auxiliaries such as ba:yad ‘should,<br />

must’, future auxiliary xa:stan ‘will’ and progressive auxiliary da:štan ‘to have’, as well<br />

as clitic pronouns like -eš ‘it’ as <strong>in</strong> (2b-c).<br />

(1) a. Ali gerye ne-mi-konad.<br />

Ali cry Neg. Impf-do<br />

‘ Ali does not cry.’<br />

(2) a. Ali gerye da:rad mi- konad.<br />

Ali cry Prog. Aux. Impf-do<br />

‘Ali is cry<strong>in</strong>g.’<br />

b. be ra:dyo guš xa:had kard.<br />

to radio ear Fu. Aux. do<br />

‘(He) will listen to the radio.’<br />

c. guš-eš ba:yad mi-kardi.<br />

ear-it should Impf.-do-Past<br />

‘(You) should have listened to it.’<br />

It should be noted that the auxiliaries are f<strong>in</strong>ite and carry the agreement features.<br />

In addition to the separation <strong>of</strong> the components <strong>of</strong> LVCs by the negation particle,<br />

the imperfective morpheme mi, the clitic pronouns like -eš ‘it’, modals and auxiliaries,<br />

the nom<strong>in</strong>al element <strong>of</strong> some LVCs may have a specific read<strong>in</strong>g and function like a DP<br />

argument which can be separated from the LV <strong>in</strong> some syntactic constructions. For<br />

<strong>in</strong>stance, a PV, as part <strong>of</strong> a LVC (3a), may occasionally appear <strong>in</strong> a DP which can be<br />

modified by an adjective (3b) and function as the direct object (3c). Moreover, it can be<br />

relativized (3d), focused on by Wh-<strong>in</strong>terrogatives (3e) and scrambled (3f).<br />

(3) a. tegarg be ba:q-e man latme zad.<br />

hail to garden-Ez. I damage beat-past<br />

‘The hail damaged my garden.’<br />

b. tegarg latme-ye bad-i be ba:q-e man zad.<br />

hail damage-Ez. bad-Indef.Par. to garden-Ez.I beat-past<br />

‘The hail caused bad damage to my garden. / The hail damaged my garden<br />

badly.’<br />

2


c. tegarg-e diruz ʔ<strong>in</strong> latme-ra: be ba:q-e man zad.<br />

hail-Ez. yesterday this damage-DOM to garden-Ez. I beat-past<br />

‘The yesterday’s hail caused this damage to my garden’<br />

d. latme-ʔi ke tegarg be ba:q-ha: zad jobra:n na:pazir ʔast.<br />

damage-Indef. Art. that hail to gardens beat-past irretrievable is<br />

‘The damage caused by the hail to the gardens is irretrievable.’<br />

e- Ali če latme-ʔi be šoma: zad?<br />

Ali what damage-Indef. Art. to you beat-past<br />

‘What loss did Ali cause to you?<br />

f. ʔ<strong>in</strong> latme-ra: tegarg-e diruz be ba:q-e man zad.<br />

this damage-DOM hail-Ez. yesterday to garden-Ez. I beat-past<br />

‘The yesterday’s hail caused this damage to my garden’<br />

We should note that <strong>in</strong> (3) not only is a PV modified by an adjective (3b), preceded by a<br />

determ<strong>in</strong>er, i.e. ʔ<strong>in</strong> ‘this’ (3c), and assigned direct object case (3c), but also the LVC is<br />

broken by the PPs be ba:q-e man ‘to my garden’ (3b- c) and a relative clause can appear<br />

between the PV and the LV (3d).<br />

As far as separability is concerned, the LVCs <strong>in</strong> (4a) and (5a) behave differently<br />

from the LVC <strong>in</strong> (3a). Unlike the LVC <strong>in</strong> (3b-f), a PV <strong>in</strong> the LVCs <strong>in</strong> (4b-f) and (5b-f),<br />

can never appear <strong>in</strong> a DP which can be modified by an adjective (4b, 5b) and cannot<br />

function as the direct object (4c, 5c). Moreover, it cannot be relativized (4d, 5d), focused<br />

on by Wh-<strong>in</strong>terrogatives (4e, 5e) and scrambled (4f, 5f). Follow<strong>in</strong>g Karimi-Doostan<br />

(1997), we refer to the LVCs such as those <strong>in</strong> (3) as separable LVCs and those like (4-5)<br />

as <strong>in</strong>separable LVCs.<br />

(4) a. Ali ka:r-aš-ra: ʔanja:m da:d.<br />

Ali work-his-DOM perform<strong>in</strong>g give-past<br />

‘Ali did his work.’<br />

*b. Ali ʔanja:m-e xub-i (ʔaz) (be) ka:r-aš da:d.<br />

Ali perform<strong>in</strong>g-Ez. good-Indef. Art. (<strong>of</strong>) (to) work-his give-past<br />

‘Ali had a good performance <strong>in</strong> his work. / Ali did his work well.’<br />

*c. Ali ʔ<strong>in</strong> ?anja:m-ra: (ʔaz) (be) ka:r-aš da:d.<br />

Ali this perform<strong>in</strong>g-DOM (<strong>of</strong> ) (to) work-his give-past<br />

‘Ali did his work.’<br />

*d. ʔanja:m-i ke Ali (ʔaz) (be) ka:r-aš da:d m<strong>of</strong>id bud<br />

perform<strong>in</strong>g-Indef. Art. that Ali (<strong>of</strong>) (to) work-his give-past useful was<br />

‘Ali’s perform<strong>in</strong>g his work was useful.’<br />

*e. Ali če ʔanja:m-i (ʔaz) (be) ka:r-aš da:d?<br />

Ali what performance-Indef. Art. (<strong>of</strong>) (to) work-his give-past<br />

‘What sort <strong>of</strong> performance did Ali have <strong>in</strong> his work?’<br />

3


*f. ʔ<strong>in</strong> ʔanja:m-ra: Ali (ʔaz) (be) ka:r-aš da:d.<br />

this perform<strong>in</strong>g-DOM Ali (<strong>of</strong> ) (to) work-his give-past<br />

‘Ali did his work.’<br />

(5) a. Ali be ra:dyo guš- kard<br />

Ali to radio ear do-past<br />

‘Ali listened to the radio.’<br />

*b. Ali be ra:dyo guš-e xubi kard<br />

Ali to radio ear-Ez. good do-past<br />

‘Ali listened to the radio well.’<br />

*c. Ali ?<strong>in</strong> guš-ra: be ra:dyo kard<br />

Ali this ear-DOM to radio do-Past<br />

‘Ali did this listen<strong>in</strong>g to the radio.’<br />

*d. guš-i ke Ali be ra:dyo kard va:zeh nabud.<br />

ear-Indef. Art. that Ali to radio do-past clear wasn’t<br />

‘The listen<strong>in</strong>g that Ali did to the radio was not clear.’<br />

*e. Ali če guš-i be ra:dyo kard?<br />

Ali what ear- Indef.Art. to radio do-past<br />

‘What listen<strong>in</strong>g did Ali do to the radio?’<br />

*f. ʔ<strong>in</strong> guš-ra: Ali be ra:dyo kard<br />

this ear-DOM Ali to radio do-past<br />

‘Ali did this listen<strong>in</strong>g to the radio.’<br />

In the present work, it is shown that different behaviors <strong>of</strong> PVs and the<br />

(<strong>in</strong>)separability <strong>of</strong> LVCs is related to the fact that nom<strong>in</strong>al PVs are <strong>of</strong> three different<br />

types: <strong>Verb</strong>al Nouns (VN), predicative nouns and non-predicative nouns. It will be<br />

illustrated that only the predicative noun PVs, as <strong>in</strong> (3), are able to be developed <strong>in</strong>to DPs<br />

which can be separated <strong>in</strong> the syntax, while the other two groups <strong>of</strong> PVs, such as VN<br />

preverbs <strong>in</strong> (4) and non-predicative nouns <strong>in</strong> (5), are unable to be augmented <strong>in</strong>to DPs<br />

due to their semantic and morphosyntactic properties. Thus, they cannot be separated <strong>in</strong><br />

the syntax as <strong>in</strong> (4) and (5). This issue is expla<strong>in</strong>ed <strong>in</strong> detail <strong>in</strong> section {3.2}. It is<br />

demonstrated that the (<strong>in</strong>)separability <strong>of</strong> LVCs can be accounted for <strong>in</strong> the light <strong>of</strong> the<br />

semantic and morphosyntactic properties <strong>of</strong> PVs and the (<strong>in</strong>)separability <strong>of</strong> LVCs <strong>in</strong><br />

cases like (3-5) cannot necessarily mean that LVCs are formed at two different levels <strong>of</strong><br />

representation, i.e. the lexicon and the syntax.<br />

The data <strong>in</strong> (1-5) can raise some questions. Why are all LVCs uniformly separable<br />

by some morphological and syntactic elements (1-2), while they behave differently <strong>in</strong> (3-<br />

5)? Can the contrasts <strong>in</strong> (3-5) <strong>in</strong>dicate that LVCs are formed at different levels, e.g.<br />

4


lexicon or syntax and the <strong>in</strong>ternal structure <strong>of</strong> cases like (4) and (5) is not visible to the<br />

syntax? How do some PVs sometimes function as parts <strong>of</strong> predicates (3a) and sometimes<br />

as DP arguments play<strong>in</strong>g the role <strong>of</strong> direct object? We have <strong>in</strong>troduced the significant<br />

data and issues related to this work present section. The general properties <strong>of</strong> LVCs <strong>in</strong><br />

Persian are presented <strong>in</strong> section {2}. In section {3}, the components <strong>of</strong> LVCs and their<br />

characteristics are expla<strong>in</strong>ed. The (<strong>in</strong>)separability <strong>of</strong> LVCs is particularly focused on <strong>in</strong><br />

section {4}. The conclud<strong>in</strong>g theoretical remarks <strong>of</strong> the paper are summarized <strong>in</strong> section<br />

{5}.<br />

2. <strong>Light</strong> <strong>Verb</strong> <strong>Constructions</strong><br />

Persian LVCs have attracted a number <strong>of</strong> researchers (Mohammad and Karimi,<br />

1992; Vahedi-Langroudi, 1996; Karimi, 1987; Karimi Doostan, 1997; Dabir-Moghadam,<br />

1995). More recently, Megerdoomian (2001) and Folli, Harley and Karimi (2005) have<br />

ma<strong>in</strong>ly focused on event structure and aspectual properties <strong>of</strong> Persian LVCs; while <strong>in</strong><br />

Megerdoomian (2001) the LVs, <strong>in</strong> Folli, Harley and Karimi (2005) the VNs play an<br />

important role <strong>in</strong> determ<strong>in</strong><strong>in</strong>g the aspectual properties <strong>of</strong> LVCs. Both approaches claim<br />

that Persian LVCs can be accounted for syntactically and f<strong>in</strong>d it difficult to consider these<br />

complex verbs as lexical units. They attempt to expla<strong>in</strong> the <strong>in</strong>ternal structure <strong>of</strong> Persian<br />

LVCs <strong>in</strong> the light <strong>of</strong> Hale and Keyser's (1993, 2002) lexical-syntax approach to verb<br />

formation. A considerable number <strong>of</strong> complex predicates <strong>in</strong>clud<strong>in</strong>g those focused on <strong>in</strong><br />

this work are taken <strong>in</strong>to consideration <strong>in</strong> both works and they take the position that the<br />

LVs are responsible for agentive arguments. Goldberg (2003) argues that Persian LVC is<br />

a zero level verb (V0) by default and whether LVC is expressed as a zero level verb or as<br />

a phrasal entity is determ<strong>in</strong>ed by <strong>in</strong>dependently motivated constructions. Default V0<br />

status accounts for the CP’s zero level properties <strong>in</strong>clud<strong>in</strong>g its resistance to separation<br />

5


and its appearance <strong>in</strong> derivational constructions. V0 status is a default case <strong>in</strong> the sense<br />

that it can be overridden if and only if there is another construction <strong>in</strong> the grammar that<br />

specifically overrides it. In Karimi-Doostan (1997), Persian LVCs consist <strong>of</strong> semantically<br />

bleached LVs and nom<strong>in</strong>al PVs. He argues that the LVs are capable <strong>of</strong> case mark<strong>in</strong>g and<br />

host<strong>in</strong>g verbal features and they, <strong>in</strong> comb<strong>in</strong>ation with T, can assign the nom<strong>in</strong>ative and<br />

accusative case or only the nom<strong>in</strong>ative case. It is shown that the arguments <strong>of</strong> PVs can be<br />

structurally case assigned via T and LV whose case assign<strong>in</strong>g roles are not thematically<br />

restricted. In other words, the arguments <strong>of</strong> PVs can be case marked by T and LVs.<br />

The existence <strong>of</strong> separable and <strong>in</strong>separable LVCs and the dual lexical and<br />

syntactic properties <strong>of</strong> LVCs have led researchers to take different positions on the level<br />

<strong>of</strong> LVC formation and consider LVCs as lexical units or syntactic phrases. Some l<strong>in</strong>guists<br />

(Ghomeshi and Massan, 1994; Goldberg, 1995; 2003) believe that LVC components have<br />

zero level status and LVCs are formed <strong>in</strong> the lexicon. Similarly, <strong>in</strong> Karimi-Doostan<br />

(1997) a LV and a PV, two zero level elements, conjo<strong>in</strong> together <strong>in</strong> the lexicon to form a<br />

LVC with a s<strong>in</strong>gle a(rgument)-structure whose <strong>in</strong>ternal structure is visible to syntactic<br />

rules. Other researchers (Heny and Samiian, 1992; Vahedi-Langrudi, 1996;<br />

Megerdoomian, 2001; 2002 and Folli, Harley and Karimi, 2005) suppose that LVCs are<br />

syntactic units formed <strong>in</strong> the syntax. It seems difficult to account for the separable LVCs<br />

such as those <strong>in</strong> (3) <strong>in</strong> the approaches which take LVCs as lexical units formed <strong>in</strong> the<br />

lexicon. Moreover, cases such as (3b-f) <strong>in</strong> which PVs appear as referential and case<br />

marked full DPs can underm<strong>in</strong>e the positions which consider nom<strong>in</strong>al PVs as unspecific<br />

and non-referential nouns (Vahedi-Langroudi, 1996; Dabir-Moghadam, 1995;<br />

Megerdoomian, 2002).<br />

Karimi (1987), attempt<strong>in</strong>g to <strong>in</strong>troduce various types <strong>of</strong> LVCs, provides the<br />

follow<strong>in</strong>g examples as LVCs <strong>in</strong> Persian. In each <strong>of</strong> the follow<strong>in</strong>g cases, a verb (LV)<br />

comb<strong>in</strong>es with a PV (a prepositional phrase, an adjective, an adverb or a noun) to form a<br />

LVC with a s<strong>in</strong>gle mean<strong>in</strong>g as follows. The literal mean<strong>in</strong>g <strong>of</strong> each PV has been enclosed<br />

by parentheses.<br />

6


(6) a. PP+V<br />

az dast + dâdan (<strong>of</strong> hand giv<strong>in</strong>g) 'to lose'<br />

az yâd + bordan (<strong>of</strong> memory tak<strong>in</strong>g) 'to forget'<br />

be yâd + âvardan (to memory br<strong>in</strong>g<strong>in</strong>g) 'to remember'<br />

be nazar + âmadan (to view com<strong>in</strong>g) 'to appear'<br />

be kâr + bastan (to work ty<strong>in</strong>g) 'to use'<br />

az sar + gereftan (<strong>of</strong> head catch<strong>in</strong>g) 'to restart'<br />

be sar + bordan (to head tak<strong>in</strong>g) 'to spend, to live'<br />

be sar + âmadan (to head com<strong>in</strong>g) 'to expire'<br />

b. A+V<br />

tamiz + kardan (clean do<strong>in</strong>g) 'to clean'<br />

bidâr + shodan (awake becom<strong>in</strong>g) 'to wake up'<br />

xarâb + kardan (destroyed do<strong>in</strong>g) 'to destroy'<br />

sabok + kardan (light do<strong>in</strong>g) 'to degrade'<br />

pahn + kardan (wide do<strong>in</strong>g) 'to spread'<br />

derâz + keshidan (long pull<strong>in</strong>g) 'to lie down'<br />

kam + kardan (little do<strong>in</strong>g) 'to subtract'<br />

c. Adv+V<br />

birun +kardan (out do<strong>in</strong>g) 'to fire (someone),<br />

'to make someone leave'<br />

bâlâ + bordan (up carry<strong>in</strong>g) 'to promote'<br />

pas + dâdan (back giv<strong>in</strong>g) 'to return'<br />

pây<strong>in</strong> + âvardan (down br<strong>in</strong>g<strong>in</strong>g) 'to degrade (someone/<br />

the value <strong>of</strong> someth<strong>in</strong>g)<br />

bar + chidan (over arrang<strong>in</strong>g) 'to pick up, to remove'<br />

dar + kardan (out do<strong>in</strong>g) 'to fire (a shot), to deduct'<br />

dar + gozashtan (<strong>of</strong>f pass<strong>in</strong>g) 'to die'<br />

dar + <strong>of</strong>tâdan (<strong>of</strong>f fall<strong>in</strong>g) 'to quarrel, to oppose'<br />

dur + andâxtan (far throw<strong>in</strong>g) 'to throw away'<br />

d. N+V<br />

da'vat + kardan (<strong>in</strong>vitation do<strong>in</strong>g) 'to <strong>in</strong>vite'<br />

kotak + zadan (beat<strong>in</strong>g hitt<strong>in</strong>g) 'to beat'<br />

shekast + dâdan (defeat giv<strong>in</strong>g) 'to defeat'<br />

panâh + bordan (refuge carry<strong>in</strong>g) 'to take refuge'<br />

panje + andâxtan (palm throw<strong>in</strong>g) 'to grip'<br />

râh + raftan (way go<strong>in</strong>g) 'to walk'<br />

âtash + zadan (fire hitt<strong>in</strong>g) 'to put on fire'<br />

All types <strong>of</strong> LVCs <strong>in</strong> (6) can uniformly be broken by the morphosyntactic elements <strong>in</strong> (1)<br />

and (2). However, the LVCs <strong>in</strong> (6a-c) cannot appear <strong>in</strong> constructions such as those <strong>in</strong> (3).<br />

That means their PVs cannot function as specific DPs function<strong>in</strong>g as direct objects and<br />

7


they cannot be relativized and focused on by Wh-<strong>in</strong>terrogatives. Therefore, they fall<br />

under the category <strong>of</strong> <strong>in</strong>separable LVCs (Karimi-Doostan, 1997). In contrast, the LVCs<br />

with nom<strong>in</strong>al PVs (5d) behave differently. Some <strong>of</strong> them can appear <strong>in</strong> sentences such as<br />

those <strong>in</strong> (3) while some cannot and they resist separation <strong>in</strong> the syntax as <strong>in</strong> (4) and (5).<br />

Thus, the LVCs <strong>of</strong> (6d) type fall <strong>in</strong>to two groups: separable and <strong>in</strong>separable LVCs.<br />

3. The Components <strong>of</strong> <strong>Light</strong> <strong>Verb</strong> <strong>Constructions</strong><br />

We believe that, to expla<strong>in</strong> the separbility and <strong>in</strong>separability <strong>of</strong> LVCs (3-5), it is<br />

necessary to uncover the semantic and morphosyntactic characteristics <strong>of</strong> members <strong>of</strong><br />

LVCs. As mentioned earlier, every LVC <strong>of</strong> (6d) type, separable or <strong>in</strong>separable, consists<br />

<strong>of</strong> a LV and a nom<strong>in</strong>al PV. We will exam<strong>in</strong>e and elaborate on the properties <strong>of</strong> LVs <strong>in</strong><br />

(3.1) and the semantic and morphosyntactic features <strong>of</strong> nom<strong>in</strong>al PVs will be presented<br />

and focused on <strong>in</strong> (3.2).<br />

3.1. <strong>Light</strong> <strong>Verb</strong>s 2<br />

In Megerdoomian (2002: 77) LVs are <strong>of</strong> two k<strong>in</strong>ds: Inner LVs and Outer LVs.<br />

The former contributes the <strong>in</strong>ner event, <strong>in</strong>ternal argument and aspect but the latter<br />

contribute the outer event, external argument and aspect. Folly, Harley and Karimi (2005:<br />

1379) believe that LVs play agent/cause, eventive and duration roles <strong>in</strong> LVCs. Karimi-<br />

Doostan (1997: 99-131), study<strong>in</strong>g a wide range <strong>of</strong> LVs <strong>in</strong> Persian <strong>in</strong> detail, concludes that<br />

LVs, lack<strong>in</strong>g thematic roles and fully specified a-structure, have preserved their lexical<br />

aspectual properties and they determ<strong>in</strong>e the (a)telicity <strong>of</strong> the VPs. Karimi-Doostan (2005:<br />

1743-4), claim<strong>in</strong>g that LVs are accusative case markers, po<strong>in</strong>ts out that it is unknown<br />

whether <strong>in</strong> LVCs (a)telicity follows from case properties or vice versa. To understand the<br />

nature and morphosyntactic and semantic characteristics <strong>of</strong> LVs <strong>in</strong> Persian we compare<br />

them with heavy (lexical) verbs and auxiliaries <strong>of</strong> this language <strong>in</strong> sections {3.1.1} and<br />

{3.1.2}.<br />

2 . The relationship between light verbs <strong>in</strong> Persian and little v <strong>in</strong> M<strong>in</strong>imalist Programme (Chomsky, 1995;<br />

1999; 2001; Radford, 2004; among others) is dealt with <strong>in</strong> Karimi-Doostan (2005).<br />

8


3.1.1. <strong>Light</strong> <strong>Verb</strong>s vs. Heavy <strong>Verb</strong>s<br />

In this section, it is shown how LVs and heavy verbs behave differently with<br />

respect to nom<strong>in</strong>alization, the formation <strong>of</strong> manner adverbials, and modification by<br />

adverbs, causativisation and argument structure.<br />

3.1.1.1. Agentive (=Subject) Nom<strong>in</strong>als: The agentive morpheme -ande ‘-er’ is added to<br />

the stems <strong>of</strong> transitive and unergative heavy verbs and forms subject nom<strong>in</strong>als, but the<br />

LVs cannot <strong>of</strong>ten be subjected to this operation <strong>in</strong>dependently. Compare (7) with (8).<br />

(7) Heavy <strong>Verb</strong>s:<br />

a. xa:ndan ‘to s<strong>in</strong>g, to read’ xa:n- (stem)+ -ande → xa:nande ‘s<strong>in</strong>ger, reader’<br />

b. ra:ndan ‘to drive’ ra:n- (stem) + -ande → ra:nande ‘driver’<br />

(8) LVs:<br />

a. kardan ‘to do’ kon- (stem) + -ande → * konande ‘doer’<br />

b. da:dan ‘to give’ dah- (stem) + -ande → * dahande ‘giver’<br />

3.1.1.2. Adverbs <strong>of</strong> Manner: The bare stems <strong>of</strong> some Persian heavy verbs are capable <strong>of</strong><br />

comb<strong>in</strong><strong>in</strong>g with the morpheme -a:n and produce a type <strong>of</strong> adverbials <strong>of</strong> manner such as<br />

xanda:n ‘while laugh<strong>in</strong>g’ <strong>in</strong> (9).<br />

(9) Ali xanda:n va:red-e ʔota:q šod.<br />

Ali while-laugh<strong>in</strong>g enter-Ez room became.<br />

‘Ali, happily (while laugh<strong>in</strong>g), entered the room.’<br />

None <strong>of</strong> the LVs can undergo this morphological operation <strong>in</strong>dependently. Consider the<br />

follow<strong>in</strong>g examples which <strong>in</strong>dicate the contrast between these two classes <strong>of</strong> verb with<br />

respect to this process.<br />

(10) Heavy <strong>Verb</strong>s:<br />

a. xandidan ‘to laugh’ xand- (stem) + -a:n → xanda:n ‘while laugh<strong>in</strong>g’<br />

b. geristan ‘to cry’ geri- ( stem) + -a:n→ geria:n ‘ while cry<strong>in</strong>g’<br />

9


(11) LVs:<br />

a. kardan ‘to do’ kon- (stem) + -a:n → * kona:n ‘while do<strong>in</strong>g’<br />

b. zadan ‘to beat’ zan- (stem) + -a:n → * zana:n ‘while beat<strong>in</strong>g’<br />

Only a full predicate can undergo this morphological process. LVs, be<strong>in</strong>g <strong>in</strong>complete<br />

predicates, are not able to be <strong>in</strong>put to this rule. However, when LVs comb<strong>in</strong>e with other<br />

items and become full predicates, they ga<strong>in</strong> the capacity to undergo this operation. For<br />

<strong>in</strong>stance, the LVs <strong>in</strong> (11) as shown <strong>in</strong> (12) can form adverbs <strong>of</strong> manner like heavy verbs,<br />

after comb<strong>in</strong><strong>in</strong>g with appropriate PVs.<br />

(12) LVCs<br />

a. raqs kardan ‘to dance’ raqs kon- (noun + stem) + -a:n → raqs kona:n<br />

‘while danc<strong>in</strong>g’<br />

b. qadam zadan ‘to walk’ qadam zan- (noun + stem) + -a:n → qadam zana:n<br />

‘while walk<strong>in</strong>g’<br />

3.1.1.3. Modification by Adverbs: LVs, <strong>in</strong> contrast to heavy verbs, cannot<br />

<strong>in</strong>dependently be preceded or modified by adverbs. Consider the follow<strong>in</strong>g sentences:<br />

(13) a. Ali Sasan-ra: šaba:ne zad.<br />

Ali Sasan-DOM nightly beat-past<br />

‘Ali punished Sasan nightly.’<br />

b. tegarg be ba:q-e man šaba:ne latme zad.<br />

hail to garden-Ez. I nightly damage beat-past<br />

‘The hail damaged my garden nightly.’<br />

*(14) tegarg be ba:q-e man latme šaba:ne zad.<br />

hail to garden-Ez. I damage nightly beat-past<br />

‘The hail damaged my garden nightly.’<br />

As can be seen, the adverb šaba:ne ‘nightly’ may precede a heavy verb (13a) or an LVC<br />

(13b), but it cannot precede or modify an LV (14). Thus, it is feasible to say that LVs, <strong>in</strong><br />

contrast to full verbs, cannot be <strong>in</strong>dependently preceded or modified by adverbs.<br />

3.1.1.4. Argument Structure: As po<strong>in</strong>ted out above, LVs and heavy verbs are different<br />

from the viewpo<strong>in</strong>t <strong>of</strong> a-structure. Heavy verbs normally tend to predicate a certa<strong>in</strong><br />

number and types <strong>of</strong> arguments, while their LV counterparts do not seem to be consistent<br />

<strong>in</strong> this regard and they predicate clauses with different number and type <strong>of</strong> arguments.<br />

For example, the Persian da:dan ‘to give’ usually appears <strong>in</strong> sentences with three<br />

10


arguments as <strong>in</strong> (15), when it functions as a heavy verb. However, it predicates different<br />

number and type <strong>of</strong> arguments as <strong>in</strong> (16) <strong>in</strong> its light use.<br />

(15) Ali keta:b-ra: be Sasan da:d. (Dative)<br />

Ali book-DOM to Sasan give-past<br />

‘Ali gave the book to Sasan.’<br />

(16) a. Ali ʔesteʔfa: da:d (Intransitive)<br />

Ali resignation give-past<br />

‘Ali resigned’<br />

b. Ali Sasan-ra: neja:t da:d (Transitive)<br />

Ali Sasan-DOM rescue give-past<br />

‘Ali rescued Sasan’<br />

c. Ali ʔota:q-ra: be Sasan ʔextesa:s da:d (Dative)<br />

Ali room-DOM to Sasan allocation give-past<br />

‘Ali allocated the room to Sasan.’<br />

In contrast to (15), <strong>in</strong> (16) the same LV predicates clauses with different number and type<br />

<strong>of</strong> arguments; depend<strong>in</strong>g on the predicational characteristics <strong>of</strong> the PV elements <strong>of</strong> the<br />

LVCs. Thus, it might be plausible to say that an LV has an unspecified, or defective astructure<br />

(Di Sciullo, 1990; Butt, 1995) which turns <strong>in</strong>to a fully-fledged one when the<br />

given LV co-occurs with another element, while a heavy verb bears a fully specified astructure.<br />

3.1.2. <strong>Light</strong> <strong>Verb</strong>s vs. Auxiliaries<br />

Persian auxiliaries (e.g. šodan ‘to become’, budan ‘to be’, da:štan ‘to have’,<br />

xa:stan ‘will’) <strong>in</strong> their usage as auxiliaries are different from LVs. LVs, bear<strong>in</strong>g lexical<br />

aspectual properties, contribute to the formation <strong>of</strong> predicates (= LVCs) by comb<strong>in</strong><strong>in</strong>g<br />

with PVs. In contrast, auxiliaries, only capable <strong>of</strong> carry<strong>in</strong>g functional features (e.g. tense<br />

and agreement), are unable to participate <strong>in</strong> LVC formation and they may only co-occur<br />

with already fully-fledged predicates either simple or complex. Consider the follow<strong>in</strong>g<br />

examples:<br />

(17) a. Ali na:me-ra: xa:had xa:nd. (Future)<br />

Ali letter-DOM will read<br />

‘Ali will read the letter.’<br />

11


. Ali na:me-ra: ---- xa:nd. (Past)<br />

Ali letter-DOM read<br />

‘Ali read the letter.’<br />

*c. Ali na:me-ra: xa:had ------.<br />

Ali letter-DOM will -----<br />

‘Ali will ---- the letter.’<br />

(17') a. Ali na:me-ra: qera:ʔat xa:had kard (Future)<br />

Ali letter-DOM read<strong>in</strong>g will do-past<br />

‘Ali will read the letter.’<br />

b. Ali na:me-ra: qera:ʔat ---- kard (Past)<br />

Ali letter-DOM read<strong>in</strong>g ---- do-past<br />

‘Ali read the letter.’<br />

*c. Ali na:me-ra: qera:ʔat xa:had -----<br />

Ali Letter-DOM read<strong>in</strong>g will<br />

‘Ali will ---- the letter.’<br />

The (c) examples <strong>in</strong> (17-17’) <strong>in</strong>dicate that the heavy verb xa:ndan and the LV<br />

kardan cannot be omitted and their deletion will result <strong>in</strong> ungrammaticality, but the<br />

deletion <strong>of</strong> the future auxiliary xa:stan does not lead to ill-formed clauses (17b, 17’b).<br />

The only function <strong>of</strong> this auxiliary is to change the tense from past (17b, 17’b) to future<br />

(17a, 17’a). The deletion <strong>of</strong> this auxiliary <strong>in</strong> (b) examples, without any change <strong>in</strong> the<br />

mean<strong>in</strong>g <strong>of</strong> the propositions, shows that it is only a tense marker without any semantic<br />

force and cannot predicate a clause without the existence <strong>of</strong> another verb, but LVs can do<br />

so. Thus, opposed to auxiliaries, LVs can contribute to LVC formation or function as<br />

ma<strong>in</strong> verbs <strong>in</strong> comb<strong>in</strong>ation with PVs. In addition, (17’b) shows that LVs, opposed to<br />

auxiliaries, which can only accompany verbal elements, may co-occur with non-verbal<br />

items <strong>in</strong> order to form complex verbs.<br />

Although researchers might disagree on the semantic force and aspectual<br />

properties <strong>of</strong> LVs, they all argee that LVs can carry verbal features such as person,<br />

number and tense markers and they are able to assign case. This position is supported by<br />

the above data too.<br />

3.2. Nom<strong>in</strong>al Preverbs<br />

Nom<strong>in</strong>al preverbs fall <strong>in</strong>to three groups. For expository reasons, we will refer to<br />

them as <strong>Verb</strong>al Nouns (VN), predicative nouns and non-predicative nouns. As it will be<br />

12


illustrated <strong>in</strong> sections {3.2.1} and {3.2.2}, the predicative nouns and VNs, referr<strong>in</strong>g to<br />

actions or events, bear a-structure while non-predicative nouns, referr<strong>in</strong>g to th<strong>in</strong>gs or<br />

objects, do not bear a-structure. In addition, VN preverbs, unlike predicative and nonpredicative<br />

nouns, lack N features. In (18), we show some LVCs conta<strong>in</strong><strong>in</strong>g these three<br />

types <strong>of</strong> preverbs. The PVs <strong>in</strong> (18a) are VNs, those <strong>in</strong> (18b) are predicative nouns, and<br />

those <strong>in</strong> (18c) are non-predicative nouns.<br />

(18) a. Predicative <strong>Verb</strong>al Nouns:<br />

ʔanja:m da:dan (performance + to give) ‘to perform, to do’<br />

baya:n kardan (express<strong>in</strong>g + to do) ‘to express, to say’<br />

ʔezha:r kardan (stat<strong>in</strong>g + to do) ‘to state, to say’<br />

ʔa:xa:z kardan (beg<strong>in</strong>n<strong>in</strong>g + to do) ‘to beg<strong>in</strong>’<br />

tahye kardan (provid<strong>in</strong>g + to do) ‘to provide’<br />

b. Predicative Nouns:<br />

komak kardan (help + to do) ‘to help’<br />

gerye kardan (cry + to do) ‘to cry’<br />

lotf kardan (favor + to do) ‘to do sb a favor’<br />

latme zadan (damage + to beat) ‘to damage’<br />

ra:hnema:ʔi kardan (advice + to do) ‘to advise’<br />

daʔvat Kardan (<strong>in</strong>vitation + to do) 'to <strong>in</strong>vite'<br />

c. Non-predicative Nouns:<br />

gush kardan (ear + to do) ‘to listen’<br />

ʔarre kardan (saw + to do) ‘to saw’<br />

mash<strong>in</strong> kardan (mach<strong>in</strong>e + to do) ‘to mach<strong>in</strong>e’<br />

ʔatash zadan (fire + to beat) ‘to fire, to set to fire’<br />

q<strong>of</strong>l kardan (lock + to do) ‘to lock’<br />

Karimi-Doostan (2008a: 189-192), based on semantic and morphosyntactic grounds,<br />

claims that predicative VNs (18a) are different from nouns and they cannot be classified<br />

as any <strong>of</strong> the Persian lexical categories, i.e. V, N, P or Adj/Adv. In the rest <strong>of</strong> this<br />

section, different types <strong>of</strong> nom<strong>in</strong>al preverbs are focused on.<br />

3.2.1. Predication and Nom<strong>in</strong>al Preverbs<br />

The PVs may either refer to th<strong>in</strong>gs or objects <strong>in</strong> the universe (18c) or to actions<br />

and events (18a-b). Consequently, the nom<strong>in</strong>al PVs are <strong>of</strong> two types, non-predicative and<br />

13


predicative, with respect to a(rgument)-structure, predication and thematic force. The<br />

non-predicative PVs are PV nouns like guš ‘ear’ and ʔa:taš ‘fire’ as <strong>in</strong> (19-20) which do<br />

not carry a-structure, while the predicative PVs are action or derived nouns such as<br />

ʔanja:m ‘perform<strong>in</strong>g’ and ra:hnama:ʔi ‘advice’ as <strong>in</strong> (21-24) which can bear a-structure.<br />

The reason for such a classification is the fact that the non-predicative nouns cannot<br />

appear <strong>in</strong> nom<strong>in</strong>al form units with arguments which are predicated by the correspond<strong>in</strong>g<br />

LVCs to which these PV nouns belong. However, the VNs and predicative nouns, be<strong>in</strong>g<br />

the nom<strong>in</strong>al element <strong>of</strong> LVCs, can co-occur with their arguments <strong>in</strong> such nom<strong>in</strong>al units<br />

though the verbal elements <strong>of</strong> the correspond<strong>in</strong>g LVCs are not needed. Compare the<br />

follow<strong>in</strong>g sentences predicated by LVCs <strong>in</strong> (a) examples with their nom<strong>in</strong>al forms <strong>in</strong> (b)<br />

examples. The well-formed nom<strong>in</strong>al expressions <strong>in</strong> (b) can be used as titles <strong>of</strong> stories or<br />

newspaper headl<strong>in</strong>es. Also, note that there are no simple verbs correspond<strong>in</strong>g to the<br />

predicative PV nouns (23-26) <strong>in</strong> Persian.<br />

(19) a. Ali be ra:dyo guš da:d / kard.<br />

Ali to radio ear do-past / do-past<br />

‘Ali listened to the radio.<br />

*b. guš-e Ali be ra:dyo<br />

ear-Ez. Ali to radio<br />

‘Ali’s listen<strong>in</strong>g to radio.’<br />

(20) a. Ali bute-ha:-ra ʔa:taš zad.<br />

Ali bushes-DOM fire beat-past<br />

‘Ali set fire to the bushes.’<br />

*b. ʔa:taš-e bute-ha: tavasote Ali<br />

fire-Ez. bushes by Ali<br />

‘sett<strong>in</strong>g fire to bushes by Ali’<br />

(21) a. Ali ka:r-aš-ra: ʔanja:m da:d.<br />

Ali work-his-DOM perform<strong>in</strong>g give-past<br />

‘Ali did his work.’<br />

b. ʔanja:m-e ka:r tavasote Ali.<br />

perform<strong>in</strong>g-Ez. work by Ali<br />

‘Ali’s do<strong>in</strong>g the work. / Do<strong>in</strong>g the work by Ali.’<br />

(22) a. Ali ka:r-aš-ra: ʔa:qa:z kard.<br />

Ali work-his-DOM start do-past<br />

‘Ali started his work.<br />

14


. ʔa:qa:z-e ka:r-e Ali<br />

start-Ez. work-Ez Ali<br />

‘Ali’s start<strong>in</strong>g the work’<br />

(23) a. Ali Sasan-ra: ra:hnama:ʔi kard.<br />

Ali Sasan-DOM advice do-Past<br />

‘Ali advised Sasan.’<br />

b. ra:hnama:ʔi-ye Ali be Sasan.<br />

advice-Ez. Ali to Sasan<br />

‘Ali’s advice to Sasan’<br />

(24) a. Ali be Sasan komak kard.<br />

Ali to Sasan help do-past<br />

‘Ali helped Sasan.’<br />

b. komak-e Ali be Sasan<br />

help-Ez. Ali to Sasan<br />

‘Ali’s help to Sasan’<br />

As can be seen <strong>in</strong> the (b) nom<strong>in</strong>al expressions <strong>in</strong> (21-24), the nom<strong>in</strong>al elements <strong>of</strong><br />

LVCs <strong>in</strong> the (a) examples <strong>in</strong> (21-24) can head NPs which conta<strong>in</strong> all the arguments <strong>of</strong> the<br />

correspond<strong>in</strong>g LVCs. Hence, it is justifiable to hold that the PVs <strong>in</strong> (21-24) are<br />

predicative and responsible for the argument structures <strong>of</strong> the LVCs. On the contrary, as<br />

<strong>in</strong>dicated <strong>in</strong> (19b, 20b) nom<strong>in</strong>al forms, the PVs <strong>in</strong> (19-20) sentences cannot head NPs<br />

conta<strong>in</strong><strong>in</strong>g the arguments predicated <strong>in</strong> (19-20) sentences by the LVCs. Therefore, the<br />

PVs such as those <strong>in</strong> (18c) as shown <strong>in</strong> (19-20) are non-predicative nom<strong>in</strong>al PVs and the<br />

PVs such as those <strong>in</strong> (18a-b) as shown <strong>in</strong> (21-24) are predicative PV nouns. Of course, it<br />

is obvious that the predicational difference between the non-predicative PVs (19-20) and<br />

the predicative ones (21-24) is related to the mean<strong>in</strong>g <strong>of</strong> the PVs. While the former are<br />

semantically opaque nouns referr<strong>in</strong>g to semantically referential objects <strong>in</strong> the universe,<br />

the latter are predicative nouns referr<strong>in</strong>g to actions or events.<br />

3.2.2. Morphosyntactic Properties <strong>of</strong> Nom<strong>in</strong>al PVs<br />

Most Persian lexical words can be categorized <strong>in</strong>to word classes such as nouns,<br />

verbs, prepositions and adjectives/adverbs and these categories can be differentiated from<br />

each other by morphosyntactic and distributional tests. Although Cr<strong>of</strong>t (2005: 435) <strong>in</strong><br />

l<strong>in</strong>e with Schachter (1985: 5-6), Cr<strong>of</strong>t (1991: 45-46), Cr<strong>of</strong>t (2001: 75-83) and Eavan and<br />

Osada (2005) criticizes the distributional method <strong>of</strong> lexical word categorization, he states<br />

that the distributional method is the fundamental empirical method and there is no other.<br />

15


Karimi-Doostan (2008a) shows that, accord<strong>in</strong>g to distributional and morphosyntactic<br />

tests, Persian lexical words fall under the categories <strong>of</strong> noun, verb, preposition and<br />

adjective/adverb. Obviously, the nom<strong>in</strong>al PVs are different from verbs, prepositions and<br />

adjectives and do not fall under the categories <strong>of</strong> any <strong>of</strong> these word classes. To uncover<br />

the characteristics <strong>of</strong> nom<strong>in</strong>al PVs from the viewpo<strong>in</strong>t <strong>of</strong> lexical categorization, we<br />

attempt to illustrate how nouns behave when subjected to distributional and<br />

morphosyntactic tests <strong>in</strong> {3.2.2.1} and then compare nom<strong>in</strong>al PVs to nouns <strong>in</strong> {3.2.2.2}.<br />

3.2.2.1.Nouns<br />

Nouns such as daʔvat ‘<strong>in</strong>vitation’, can be pluralized (25a), selected by adjectives<br />

(25b) and can co-occur with demonstratives (25c), prepositions and Ezafe particle 3 (25e).<br />

They can also function as subject (25f) and object (25a).<br />

(25) a. daʔvat-hâ-ye ziyâdi dâšt-im.<br />

Invitation-pl-Ez a lot had-1 st pl.<br />

‘We had a lot <strong>of</strong> <strong>in</strong>vitations.’<br />

b. daʔvat-e xoob bood.<br />

Invitation-Def. good was<br />

‘The <strong>in</strong>vitation was good’<br />

c. ʔ<strong>in</strong> daʔvat barâ-ye če bood<br />

this <strong>in</strong>vitation for what was<br />

‘What was this <strong>in</strong>vitation for?’<br />

d. ʔaz daʔvat porsid.<br />

from <strong>in</strong>vitation asked<br />

‘He/she asked about the <strong>in</strong>vitation.’<br />

e. daʔvat-e ali az man<br />

<strong>in</strong>vitation-Ez. Ali from me<br />

‘Ali’s <strong>in</strong>vitation to me’<br />

f. ʔ<strong>in</strong> da’vat kâr-o xarâb kard.<br />

this <strong>in</strong>vitation work-DOM ru<strong>in</strong><strong>in</strong>g do-past<br />

‘This <strong>in</strong>vitation ru<strong>in</strong>ed the work.’<br />

3.2.2.2. The Lexical Category <strong>of</strong> Nom<strong>in</strong>al Preverbs<br />

3 - The Ezafe particle –(y/ʔ)e appears between a noun and its complements, and between a noun and a<br />

possessor, between a noun and a modifier, between an adjective/adverb and its complements, and between<br />

a preposition and its complements. The constructs l<strong>in</strong>ked by the Ezafe particle are known as ‘Ezafe<br />

constructions’.<br />

16


Among the three groups <strong>of</strong> nom<strong>in</strong>al PVs <strong>in</strong>troduced above, non-predicative PVs<br />

and predicative nom<strong>in</strong>al PVs behave like nouns. That is, they can be pluralized, selected<br />

by determ<strong>in</strong>ers, demonstratives, prepositions and Ezafe particle and they can function as<br />

subject and object as <strong>in</strong> (25). However, predicative VNs do not morphosyntactically<br />

behave like nouns. Unlike non-predicative nouns (18c) and predicative nouns (18b) as <strong>in</strong><br />

(25), the predicative VNs (18a) are different from the ord<strong>in</strong>ary nouns. They cannot be<br />

pluralized (26a), and selected by determ<strong>in</strong>ers (26b), demonstrative pronouns (26c) and<br />

prepositions (26d). They also cannot function as subject (26e) or object (26a).<br />

(26) *a ʔanjâm-hâ-ye ziyâdi dâštim.<br />

perform<strong>in</strong>g-pl-Ez. a lot had-1 st -pl.<br />

‘We had a lot <strong>of</strong> performances.’<br />

*b ʔanjâm –e xoob bood.<br />

perform<strong>in</strong>g-Def. good was<br />

‘The perform<strong>in</strong>g was good’<br />

*c ʔ<strong>in</strong> ʔanjâm barâ-ye če bood<br />

this perform<strong>in</strong>g for what was<br />

‘What was this perform<strong>in</strong>g for?’<br />

*d ʔaz ʔanjâm porsid.<br />

from perform<strong>in</strong>g asked<br />

‘He/she asked about the perform<strong>in</strong>g.’<br />

*e ʔ<strong>in</strong> ʔanjâm kâr-o xarâb kard.<br />

this perform<strong>in</strong>g work-DOM ru<strong>in</strong><strong>in</strong>g do-past<br />

‘This perform<strong>in</strong>g ru<strong>in</strong>ed the work.’<br />

Based on the above observations, it is reasonable to conclude that non-predicate<br />

PVs and predicative PVs are [+N] 4 while the predicative VNs are [-N] from the<br />

viewpo<strong>in</strong>t <strong>of</strong> lexical categorization. Therefore, one can conclude that different nom<strong>in</strong>al<br />

PVs have the follow<strong>in</strong>g thematic and lexical category properties <strong>in</strong> (27). We use the term<br />

a-structure to <strong>in</strong>dicate the predicative characteristics <strong>of</strong> PVs as expla<strong>in</strong>ed <strong>in</strong> {3.2.1} and<br />

[N] feature to show the lexical category <strong>of</strong> nom<strong>in</strong>al PVs as elaborated on <strong>in</strong> (3.2.2.6}.<br />

(27) a. predicative VN: [+ a-structure] [-N]<br />

b. predicative nouns [+ a-structure] [+N]<br />

c. non-predicative nouns [- a-structure] [+N]<br />

4 . Follow<strong>in</strong>g Chomsky (1970), most generative and related theories posit two b<strong>in</strong>ary features [±N, ±V] to<br />

def<strong>in</strong>e noun [+N, -V], verb [-N, +V], adjective [+N, +V] and preposition [-N, -V] (Haegeman, 1994: 36-37)<br />

17


As it was attested <strong>in</strong> (3), LVCs with predicative PVs are separable, but those with<br />

predicative VNs (4) and non-predicative nouns (5) are <strong>in</strong>separable. The comparison<br />

between the separable LVCs <strong>in</strong> (3) and the <strong>in</strong>separable ones <strong>in</strong> (4) and 5) may persuade<br />

us to conclude that only PVs which carry a-structure and [N] feature can be separated <strong>in</strong><br />

the syntax as <strong>in</strong> (3). To f<strong>in</strong>d out more about the l<strong>in</strong>guistic reasons beh<strong>in</strong>d the separability<br />

<strong>of</strong> LVCs, we are led to the next section attempt<strong>in</strong>g to expla<strong>in</strong> the separability <strong>of</strong> these<br />

compound forms.<br />

4. <strong>Light</strong> <strong>Verb</strong> <strong>Constructions</strong> and their <strong>Separability</strong><br />

Before concentrat<strong>in</strong>g on the separable and <strong>in</strong>separable LVCs (3-5), we will focus<br />

on the issue <strong>of</strong> break<strong>in</strong>g LVCs by some morphosyntactic elements illustrated <strong>in</strong> (1-2). In<br />

such cases, the negation particle, the imperfective particle mi-, the pronom<strong>in</strong>al clitics, the<br />

future auxiliary xa:stan 'will' and the progressive auxiliary da:štan 'to have' appear<br />

between the PVs and the LVs. In the follow<strong>in</strong>g sections, we will account for the<br />

separability <strong>of</strong> LVCs by the aforementioned l<strong>in</strong>guistic elements before expla<strong>in</strong><strong>in</strong>g the<br />

(<strong>in</strong>)separablility <strong>of</strong> LVCs <strong>in</strong> cases like (3-5).<br />

In contrast to some other languages, <strong>in</strong> Persian the negation morpheme is<br />

obligatorily prefixed to LVs as well as heavy verbs. Grimshaw (1991: 44-46) argues that<br />

there is a close relationship between verbness properties, especially tense features, and<br />

negation. It follows from our explanation that LVs have sufficient verbness properties,<br />

<strong>in</strong>clud<strong>in</strong>g tense features, to carry the negation morpheme or to be selected by the NegP<br />

<strong>in</strong>dependently and the <strong>in</strong>ternal structures <strong>of</strong> LVCs, be<strong>in</strong>g visible to syntactic and<br />

morphological rules, allow the <strong>in</strong>sertion <strong>of</strong> the negation particle (head) ne ‘not’ as <strong>in</strong> (1).<br />

The imperfective morpheme mi is prefixed to LVs or heavy verbs and forms<br />

imperfective versions <strong>of</strong> LVCs or heavy verbs. We believe that although LVs are<br />

semantically defective, they have enough verbness properties to function exactly like<br />

heavy verbs <strong>in</strong> some cases like carry<strong>in</strong>g the imperfective morpheme mi <strong>in</strong> (1).<br />

Pronom<strong>in</strong>al clitics such as -eš ‘it’ are attested, as <strong>in</strong> (2c), to <strong>in</strong>tervene between<br />

PVs and LVs. We assume Goldberg’s (1995: 18-20) analysis <strong>of</strong> clitic placement <strong>in</strong> LVCs<br />

<strong>in</strong> which she f<strong>in</strong>ds a relation between clitic placement and stress facts. It is generally<br />

18


accepted that only stressed zero-level items can host clitics. In Persian, pronom<strong>in</strong>al clitics<br />

are usually attached to lexical verbs, as these are the only stressed candidates which can<br />

function as hosts for the clitics. In the case <strong>of</strong> LVCs, clitics usually follow PVs because it<br />

is the PV which functions as the first stress carrier component <strong>of</strong> the LVCs (though <strong>in</strong><br />

some rare cases they follow LVCs as <strong>in</strong> (28b)). This behavior <strong>of</strong> clitics can be captured <strong>in</strong><br />

every analysis <strong>in</strong> which both PVs and LVCs have a zero status and speakers can either<br />

stress a PV, then a host for clitics as <strong>in</strong> (28a), or stress LVCs as a s<strong>in</strong>gle unit to carry<br />

clitics as <strong>in</strong> (28b).<br />

(28) a. guš-eš kard-am<br />

ear-it do-past-1st.sg.<br />

‘I listened to it.’<br />

? b. guš kard-am-eš<br />

ear do-past-1st. sg.- it<br />

‘I listened to it.’<br />

The recently grammaticalized auxiliary xa:stan ‘will’, as <strong>in</strong> (2c), was not used as<br />

an auxiliary <strong>in</strong> Old Persian, Middle Persian and early Modern Persian (Khanlari, 1987:<br />

286-295). In standard Persian, this verb <strong>in</strong> addition to its use as a ma<strong>in</strong> verb, mean<strong>in</strong>g ‘to<br />

want’, developed as the future auxiliary as well. In colloquial Persian, this verb is used<br />

only as a ma<strong>in</strong> verb and it has not been grammaticalized as an auxiliary. xa:stan <strong>in</strong> both<br />

cases, as an auxiliary and a ma<strong>in</strong> verb, precedes a ma<strong>in</strong> verb. The fact that the recently<br />

developed auxiliaries such as xa:stan ‘will’, unlike other auxiliaries which follow the<br />

ma<strong>in</strong> verbs, precede the ma<strong>in</strong> verbs makes the place <strong>of</strong> auxiliaries <strong>in</strong> this language appear<br />

complicated. Other auxiliaries, like budan ‘to be’, follow the heavy verbs and LVs as <strong>in</strong><br />

(29a) and (29d) respectively, but the auxiliary xa:stan ‘will’ obligatorily precedes lexical<br />

verbs, as <strong>in</strong> (29b), and LVs, as <strong>in</strong> (29c). It seems that xa:stan as an auxiliary is still<br />

identical to its use as a ma<strong>in</strong> verb with respect to the fact that both precede a ma<strong>in</strong> verb.<br />

(29) a. b.<br />

V Aux Aux V<br />

nevešte bud-am xa:h-am nevešt<br />

‘written’ ‘was-1st S<strong>in</strong>g’ ‘will-1st S<strong>in</strong>g wrote<br />

‘I had written.’ ‘I will write.’<br />

19


c. d.<br />

PV Aux LV PV LV Aux<br />

ʔanja:m xa:h-am da:d ʔanja:m da:de bud-am<br />

performance will-1st S<strong>in</strong>g gave performance given was-1st.Sg<br />

‘ I will do ...’ ‘I had done ....’<br />

The study <strong>of</strong> the sequence <strong>of</strong> auxiliaries and ma<strong>in</strong> verbs as <strong>in</strong> (29) is beyond the<br />

scope <strong>of</strong> this work. However, what is significant to us is the fact that the existence <strong>of</strong><br />

cases like (29c), where an <strong>in</strong>flected auxiliary appears between the PV and the LV,<br />

<strong>in</strong>dicates, as assumed <strong>in</strong> this work, that the <strong>in</strong>ternal structure <strong>of</strong> the LVC is not opaque to<br />

the syntax. In addition, the fact that LVs and heavy verbs behave identically with respect<br />

to placement <strong>of</strong> auxiliaries supports our idea that LVs and heavy verbs, although<br />

semantically different, are identical with regard to the so called ‘functional categories’<br />

(compare (29a) with (29d), and (29b) with (29c)).<br />

F<strong>in</strong>ally, the separability <strong>of</strong> LVCs by the progressive auxiliary da:štan ‘to have’<br />

(2b) and the modals like ba:yad ‘should’ (2d) can be captured <strong>in</strong> our account <strong>in</strong> which<br />

LVCs rema<strong>in</strong> analyzable <strong>in</strong> the syntax. These items can freely appear <strong>in</strong> different<br />

syntactic positions as <strong>in</strong> (30), so they must be able to <strong>in</strong>tervene between PVs and LVs too<br />

as they do <strong>in</strong> (2b) and (2c).<br />

(30) a. Ali (da:št) Sasan-ra: (da:št) ra:hnama:ʔi (da:št) mikard.<br />

Ali Prog. Aux. Sasan-DOM Prog. Aux. advice Prog. Aux. do<br />

‘Ali was advis<strong>in</strong>g Sasan.’<br />

b. Ali (ba:yad) Sasan-ra: (ba:yad) ra:hnama:ʔi (ba:yad) mikard.<br />

Ali should Sasan-DOM should advice should do<br />

‘Ali should have advised Sasan.’<br />

So far we have been deal<strong>in</strong>g with those <strong>in</strong>stances (1-2) <strong>in</strong> which the components <strong>of</strong><br />

LVCs can uniformly be broken by some items or rules. However, it was illustrated <strong>in</strong> the<br />

previous sections that some LVCs like (3), repeated here as (31), are separable by<br />

syntactic operations while others, such (4) and (5) repeated here as (32) and (33)<br />

respectively, resist separation. This motivated us to classify LVCs <strong>in</strong>to two groups:<br />

separable (31) and <strong>in</strong>separable (32-33).<br />

20


(31) a. tegarg be ba:q-e man latme zad.<br />

hail to garden-Ez. I damage beat-past<br />

‘The hail damaged my garden.’<br />

b. tegarg latme-ye bad-i be ba:q-e man zad.<br />

hail damage-Ez. bad-Indef. Art. to garden-Ez. I beat-past<br />

‘The hail caused bad damage to my gardens. / The hail damaged my garden<br />

badly.’<br />

c. tegarg-e diruz ʔ<strong>in</strong> latme-ra: be ba:q-e man zad.<br />

hail-Ez. yesterday this damage-DOM to garden-Ez. I beat-past<br />

‘The yesterday’s hail caused this damage to my garden’<br />

d. latme-ʔi ke tegarg be ba:q-ha: zad jobra:nna:pazir ?ast.<br />

damage-Indef. Art. that hail to gardens beat-past irretrievable is<br />

‘The damage caused by the hail to the gardens is irretrievable.’<br />

e- Ali če latme-ʔi be šoma: zad<br />

Ali what damage-Indef. Art. to you beat-past<br />

‘What loss did Ali cause to you?<br />

f. ʔ<strong>in</strong> latme-ra: tegarg-e diruz be ba:q-e man zad.<br />

this damage-DOM hail-Ez. yesterday to garden-Ez. I beat-past<br />

‘The yesterday’s hail caused this damage to my garden’<br />

(32) a. Ali ka:r-aš-ra: ʔanja:m da:d.<br />

Ali work-his-DOM perform<strong>in</strong>g give-past<br />

‘Ali did his work.’<br />

*b. Ali ʔanja:m-e xub-i (ʔaz) (be) ka:r-aš da:d.<br />

Ali perform<strong>in</strong>g-Ez. good-Indef. Art. (<strong>of</strong>) (to) work-his give-past<br />

‘Ali had a good performance <strong>in</strong> his work. / Ali did his work well.’<br />

*c. Ali ʔ<strong>in</strong> ʔanja:m-ra: (ʔaz) (be) ka:r-aš da:d.<br />

Ali this perform<strong>in</strong>g-DOM (<strong>of</strong> ) (to) work-his give-past<br />

‘Ali did his work.’<br />

*d. ʔanja:m-i ke Ali (ʔaz) (be) ka:r-aš da:d m<strong>of</strong>id bud<br />

perform<strong>in</strong>g-Indef. Art. that Ali (<strong>of</strong>) (to) work-his give-past useful was<br />

‘Ali’s perform<strong>in</strong>g his work was useful.’<br />

*e. Ali če ʔanja:m-i (ʔaz) (be) ka:r-aš da:d?<br />

Ali what performance-Indef. Art. (<strong>of</strong>) (to) work-his give-past<br />

‘What sort <strong>of</strong> performance did Ali have <strong>in</strong> his work?’<br />

*f. ʔ<strong>in</strong> ʔanja:m-ra: Ali (ʔaz) (be) ka:r-aš da:d.<br />

this perform<strong>in</strong>g-DOM Ali (<strong>of</strong> ) (to) work-his give-past<br />

‘Ali did his work.’<br />

(33) a. Ali be ra:dyo guš- kard<br />

Ali to radio ear do-past<br />

‘Ali listened to the radio.’<br />

*b. Ali be ra:dyo guš-e xubi kard<br />

Ali to radio ear-Ez. good do-past<br />

‘Ali listened to the radio well.’<br />

21


*c. Ali ʔ<strong>in</strong> guš-ra: be ra:dyo kard<br />

Ali this ear-DOM to radio do-past<br />

‘Ali did this listen<strong>in</strong>g to the radio.’<br />

*d. guš-i ke Ali be ra:dyo kard va:zeh nabud.<br />

ear-Indef. Art. that Ali to radio do-past clear wasn’t<br />

‘The listen<strong>in</strong>g that Ali did to radio was not clear.’<br />

*e. Ali če guš-i be ra:dyo kard?<br />

Ali what ear- Indef.Art. to radio do-past<br />

‘What listen<strong>in</strong>g did Ali do to the radio?’<br />

*f. ʔ<strong>in</strong> guš-ra: Ali be ra:dyo kard<br />

this ear-DOM Ali to radio do-past<br />

‘Ali did this listen<strong>in</strong>g to the radio well.’<br />

Some l<strong>in</strong>guists consider nom<strong>in</strong>al PVs as zero level elements/nouns or nonreferential<br />

and <strong>in</strong>def<strong>in</strong>ite nouns ( Barjaste, 1983; Ghomeshi and Massam, 1994; Dabir-<br />

Moghdam, 1995: 39; Vahedi-Langroudi, 1996: 20-50; Megerdoomian, 2001; 2002;<br />

Goldberg, 2003; among others). These researchers have usually focused on LVCs such<br />

as those <strong>in</strong> (32-33) and ignored cases like those <strong>in</strong> (31) or have claimed that such<br />

structures are not LVCs or complex verbs (Dabir-Moghdam, 1995: 40). These<br />

approaches face two problems. First, it is not easy for them to account for the fact that<br />

some nom<strong>in</strong>al PVs (31 and 18b) can co-occur with def<strong>in</strong>ite articles and demonstratives<br />

and function as subject and direct object (31). Moreover, they can be modified,<br />

relativized, scrambled, and focused on by Wh-<strong>in</strong>terrogatives (31). Second, they are not<br />

able to expla<strong>in</strong> the a-structure and predicational nature <strong>of</strong> these k<strong>in</strong>d <strong>of</strong> structures. For<br />

<strong>in</strong>stance, let us consider the cases <strong>in</strong> (34) and (35).<br />

(34) a. Ali Sasan-ra: ra:hnama:ʔi kard.<br />

Ali Sasan-DOM advice do-past<br />

‘Ali advised Sasan.’<br />

b. Ali ra:hnama:ʔi-ye xub-i be Sasan kard.<br />

Ali advice-Ez. good-Indef. Art. to Sasan did.<br />

‘Ali gave good advice to Sasan.’<br />

c. Ali ʔ<strong>in</strong> ra:hnama:ʔi-ra: be Sasan kard.<br />

Ali this advice-DOM to Sasan do-past.<br />

‘Ali gave Sasan this advice.’<br />

22


(35) a. tegarg be ba:q-e-man latme zad.<br />

hail to garden-Ez-I damage beat-past<br />

‘The hail damaged my garden.’<br />

b. tegarg latme-ye bad-i be ba:q-e man zad.<br />

hail damage-Ez. bad-Def. Art. to garden-Ez. I beat-past<br />

‘The hail caused bad damage to my gardens. / The hail damaged my garden<br />

badly.’<br />

c. tegarg-e diruz ʔ<strong>in</strong> latme-ra: be ba:q-e man zad.<br />

hail-Ez. yesterday this damage-DOM to garden-Ez.-I beat-past<br />

‘The yesterday’s hail caused this damage to my garden’<br />

We have two arguments <strong>in</strong> (a) examples <strong>in</strong> (34-35) because the PVs are zero level<br />

non-referential items and can be taken as parts <strong>of</strong> the LVCs. But there are three specific<br />

DP arguments <strong>in</strong> (b-c) examples <strong>in</strong> (34-35). In addition to the subject arguments, the<br />

direct objects have been realized as PPs, and the PVs have developed <strong>in</strong>to DPs modified<br />

and function<strong>in</strong>g as the direct objects marked by the direct object marker ra:. The verbs<br />

<strong>in</strong> (34-35) are not lexical verbs, because the lexical verbs zad ‘beat’ and kard ‘did’ are<br />

cross-l<strong>in</strong>guistically two-place arguments verbs and cannot contribute three arguments <strong>in</strong><br />

(34-35). The most feasible solution is to consider cases like (18b) and (31) as LVCs<br />

whose a-structure is determ<strong>in</strong>ed by the PVs (Mohammad and Karimi, 1992; Vhahedi<br />

Langruodi, 1996, Karimi-Doostan, 1997; 2001; 2005; 2008b) or both PVs and LVs<br />

compositionally contribute to the a-structure <strong>of</strong> LVCs (Mergerdoomian, 2000; Folly,<br />

Harley and Karimi, 2005).<br />

Now we are <strong>in</strong> a position to turn back to the issue <strong>of</strong> the separability <strong>of</strong> LVCs<br />

aga<strong>in</strong>. We have already expla<strong>in</strong>ed the separation <strong>of</strong> LVCs by morphosyntactic elements<br />

such as those <strong>in</strong> (1-2). As noted above, although all LVCs can be uniformly broken by<br />

morphological or syntactic elements (1-2), some LVCs are separable <strong>in</strong> the syntax as <strong>in</strong><br />

(31) but some are not as <strong>in</strong> (32-33). Many scholars such as Grimshaw and Mester<br />

(1988), Sells (1989), Park (1989), Miagawa (1989) and Ahn (1990, 1991), among<br />

others, allude to the notion that <strong>in</strong> Japanese and Korean the <strong>in</strong>corporated LVCs are<br />

formed <strong>in</strong> the lexicon, while the case-marked ones are formed at the syntax. Als<strong>in</strong>a<br />

(1997) also argues that some causatives are formed syntactically and some lexically. We<br />

propose that the existence <strong>of</strong> separable and <strong>in</strong>separable LVCs <strong>in</strong> Persian cannot<br />

necessarily be an <strong>in</strong>dication <strong>of</strong> the fact that LVCs are formed <strong>in</strong> two levels <strong>of</strong><br />

23


epresentation; for example separable LVCs <strong>in</strong> the syntax and <strong>in</strong>separable ones <strong>in</strong> the<br />

lexicon.<br />

Focus<strong>in</strong>g on N-movement to D(eterm<strong>in</strong>er) position and the dist<strong>in</strong>ction between<br />

NP and DP <strong>in</strong> Italian and English, Longobardi (1994) argues that a ‘nom<strong>in</strong>al expression’<br />

is an argument only if it is <strong>in</strong>troduced by the category D. He claims that proper names<br />

that occur <strong>in</strong> the argument function without any overt determ<strong>in</strong>er undergo rais<strong>in</strong>g from N 0<br />

to D 0 <strong>in</strong> order to function as syntactic arguments. Longobardi also suggests that DP can<br />

be licensed either as an argument <strong>in</strong> most cases or as a predicate <strong>in</strong> others (e.g. many<br />

copular or small clause constructions). By contrast, NP is not able to assume argument<br />

function unless <strong>in</strong>troduced by an overt or empty determ<strong>in</strong>er. Masullo (1992), follow<strong>in</strong>g<br />

Abney (1987) and Stowell (1989) dist<strong>in</strong>guishes between bare nouns and DPs and argues,<br />

based on case assignment facts, that only DPs can function as case-marked arguments.<br />

Partee (1987, 1992) proposes that NPs are <strong>of</strong> three types: referential, predicative<br />

and quantificational. She believes that while all NPs have quantificational mean<strong>in</strong>g, only<br />

some can have referential and/or predicative <strong>in</strong>terpretations. It is assumed that some type<br />

shift<strong>in</strong>g pr<strong>in</strong>ciples change NPs from one type to another. These pr<strong>in</strong>ciples might be<br />

lexical or syntactic. Determ<strong>in</strong>ers (THE, A, SOME) are considered as functors which shift<br />

NPs from one type to another. (See Partee (1992: 120) for the pr<strong>in</strong>ciples shift<strong>in</strong>g the three<br />

types <strong>of</strong> NPs <strong>in</strong>to each other.)<br />

Accord<strong>in</strong>g to the properties <strong>of</strong> nom<strong>in</strong>al preverbs and LVs expla<strong>in</strong>ed <strong>in</strong> the previous<br />

sections and <strong>in</strong> the spirit <strong>of</strong> Longobardi (1994), Partee (1987, 1992) and Masullo (1992)<br />

we suggest that a generalization like (36) might be at work <strong>in</strong> the syntactic realization <strong>of</strong><br />

PVs <strong>in</strong> Persian.<br />

(36) A nom<strong>in</strong>al preverb can be a specific referential DP argument if and only<br />

if the nom<strong>in</strong>al PV carries +N feature and bears a-structure 5 .<br />

It is possible to capture the separabilty and <strong>in</strong>separability <strong>of</strong> LVCs <strong>in</strong> the light <strong>of</strong><br />

(36). The predicative nouns are predicational (a-structure bear<strong>in</strong>g) and at the same time<br />

5 . See Karimi-Doostan (1997: 184-207) and Wechsler (2008) for the discussion <strong>of</strong> predicative role <strong>of</strong> PVs<br />

<strong>in</strong> LVCs. Why only predicative nouns may develop <strong>in</strong>to DPs or LVs select predicative DP PVs needs<br />

further <strong>in</strong>vestigation.<br />

24


they have N features as <strong>in</strong> normal nouns enabl<strong>in</strong>g them to occur with nom<strong>in</strong>al elements<br />

such as those <strong>in</strong> (31) and to function as N head <strong>in</strong> DPs. When nom<strong>in</strong>al predicative<br />

preverbs (18b) can be augmented <strong>in</strong>to DPs, then they are liable to be modified by<br />

adjectives (31b) and function as objects (31c). They can consequently be relativized<br />

(31d), focused on by Wh-<strong>in</strong>terrogatives (31e) and scrambled (31f). That is, as an example<br />

<strong>in</strong> (37a) the nom<strong>in</strong>al PV ra:hnema:ʔi ‘advice’ provid<strong>in</strong>g the two arguments <strong>of</strong> the LVC,<br />

has rema<strong>in</strong>ed unspecified and non-referential as a part <strong>of</strong> the LVC. Both arguments <strong>of</strong> the<br />

clause are case marked. The subject argument, Ali, receives nom<strong>in</strong>ative case and the<br />

object argument, Sasan, receives accusative case 6 and the sentence is well-formed. In<br />

(37b), Ali receives nom<strong>in</strong>ative case but, unlike (37a), the PV has appeared as a DP<br />

argument receiv<strong>in</strong>g accusative case 7 . The PV, capable <strong>of</strong> contribut<strong>in</strong>g two arguments,<br />

discharges its <strong>in</strong>ternal argument, Sasan, as PP. The appearance <strong>of</strong> the <strong>in</strong>ternal argument<br />

as a PP is because <strong>of</strong> the fact that there is no other structural case 8 available for this<br />

argument, so it is licensed as a PP receiv<strong>in</strong>g its case from a P.<br />

(37) a. Ali Sasan-ra: ra:hnama:ʔi kard.<br />

Ali Sasan-DOM advice do-past<br />

‘Ali advised Sasan.’<br />

b. Ali ?<strong>in</strong> ra:hnama:ʔi-ra: be Sasan kard.<br />

Ali this advice-DOM to Sasan do-past.<br />

‘Ali gave Sasan this advice.’<br />

Now let us consider (38) as an example with an LVC conta<strong>in</strong><strong>in</strong>g a VN PV.<br />

Although the VN PVs such as ʔanja:m and ʔa:qa:z (18a), like predicative noun PVs (37)<br />

have a-structure, they lack N features (27) or have an <strong>in</strong>termediate status between a noun<br />

and a verb (Grimshaw, 1991). Therefore, they cannot satisfy the requirement to function<br />

as lexical heads to develop <strong>in</strong>to DPs (38b) and rema<strong>in</strong> as parts <strong>of</strong> LVCs non-separable <strong>in</strong><br />

6 . We assume as argued <strong>in</strong> Mohammad and Karimi (1992); Vahedi-Langroudi (1996) and Karimi-Doostan<br />

(1997; 2002; 2005) that LVs lack a-structure and can assign case while nom<strong>in</strong>al PVs contribute the<br />

semantic force and the a-structure <strong>of</strong> the LVCs.<br />

7 . Note that PV DPs can only be case marked by the accusative case marker ra:, if they co-occur with a<br />

determ<strong>in</strong>er as <strong>in</strong> (37b). However, a regular direct object can be case marked even without a determ<strong>in</strong>er as <strong>in</strong><br />

(i).<br />

(i). Ali keta:b-ra: xa:nd.<br />

Ali book-DOM read<br />

'Ali read the book.'<br />

8 . Persian LVs and their case mark<strong>in</strong>g properties are expla<strong>in</strong>ed <strong>in</strong> Karimi-Doostan (2005) <strong>in</strong> detail.<br />

25


the syntactic structures such as those <strong>in</strong> (32-33). Thus, when VN preverbs cannot be<br />

realized as DPs, we should not expect them to be used <strong>in</strong> the so called separable LVCs,<br />

cf. (32 and 33) with (31).<br />

(38) a. Ali ka:r-aš-ra: ʔanja:m da:d.<br />

Ali work-his-DOM perform<strong>in</strong>g give-past<br />

‘Ali did his work.’<br />

*b. Ali ʔ<strong>in</strong> ʔanja:m-ra: (ʔaz) (be) ka:r-aš da:d.<br />

Ali this perform<strong>in</strong>g-DOM (<strong>of</strong> ) (to) work-his give-past<br />

‘Ali did his work.’<br />

F<strong>in</strong>ally, let us look at (39) as an example <strong>of</strong> <strong>in</strong>separable LVCs (33) with nonpredicative<br />

nouns and expla<strong>in</strong> the <strong>in</strong>separability <strong>of</strong> this k<strong>in</strong>d <strong>of</strong> LVCs (18c) too.<br />

(39) a. Ali be ra:dyo guš- kard<br />

Ali to radio ear do-past<br />

‘Ali listened to the radio.’<br />

*b. Ali ?<strong>in</strong> guš-ra: be ra:dyo kard<br />

Ali this ear-DOM to radio do-past<br />

‘Ali did this listen<strong>in</strong>g to the radio.’<br />

Although this k<strong>in</strong>d <strong>of</strong> PV, like predicative nouns (37), can have N features, they lack astructure<br />

(27c). It was also shown <strong>in</strong> {3.1.1} that LVs are different from heavy verbs,<br />

especially from the viewpo<strong>in</strong>t <strong>of</strong> a-structure. Almost all researchers (Mohammad and<br />

Karimi, 1992; Vhahedi-Langruodi, 1996; Karimi-Doostan , 1997; 2001; 2005; 2008b;<br />

Mergerdoomian, 2000; and Folly, Harley and Karimi, 2005; among others) agree that<br />

LVs have defective a-structure or lack a-structure. Therefore; <strong>in</strong> (39a), unlike (37a) and<br />

(38a), none <strong>of</strong> the members <strong>of</strong> the LVC can <strong>in</strong>dependently be responsible for contribut<strong>in</strong>g<br />

the a-structure <strong>of</strong> the LVC. Both members are compositionally responsible for the astructure<br />

<strong>of</strong> LVCs. S<strong>in</strong>ce accord<strong>in</strong>g to (27) PVs, lack<strong>in</strong>g a-structure, cannot be realized as<br />

DPs. Then, the PV guš ‘ear’, be<strong>in</strong>g non-predicative, cannot be augmented to a DP. The<br />

reason beh<strong>in</strong>d this fact is obvious. The LVs merge with zero level PVs (37a-39a) or<br />

argument bear<strong>in</strong>g DP PVs (37b) to ga<strong>in</strong> thematic force and a-structure to function as<br />

predicates. However, if the non-predicate PVs are realized as DPs, they are concrete DP<br />

arguments which can only be predicated by heavy verbs not by LVs. For example, <strong>in</strong><br />

(39b), three specific and referential DP arguments have appeared <strong>in</strong> a clause with a LV.<br />

26


Neither can the LV itself predicate the clause nor is there any semantically transparent<br />

argument bear<strong>in</strong>g DP or N to complete its defective or bleached a-structure.<br />

5. Conclusion<br />

LVCs as compound units, function<strong>in</strong>g as s<strong>in</strong>gle predicates, are different from<br />

simple lexical verbs because the simple verbs, unlike LVCs (1-5 or 31-33), cannot be<br />

broken and separated. However, the <strong>in</strong>separability <strong>of</strong> cases like (32-33) could mislead<br />

some to believe that the <strong>in</strong>ternal structure <strong>of</strong> some LVCs is not visible to the syntax and<br />

the LVCs are lexical items identical to simple verbs. It has been clarified that the<br />

<strong>in</strong>separability <strong>of</strong> LVCs is because <strong>of</strong> the semantic and morphosyntactic properties <strong>of</strong> PVs<br />

and LVs, so it cannot be because <strong>of</strong> the fact that separable and <strong>in</strong>separable LVCs are<br />

formed at two different levels <strong>of</strong> representation.<br />

The role <strong>of</strong> LVs and PVs <strong>in</strong> LVCs or clauses, especially the dual role <strong>of</strong> PVs as a<br />

part <strong>of</strong> a predicate and at the same time as a case marked specific DP argument is a<br />

complicated issue. Although researchers do not agree on the exact role <strong>of</strong> LVs and PVs,<br />

<strong>in</strong> Karimi-Doostan (2005) PVs contribute the thematic roles or a-structure <strong>of</strong> LVCs while<br />

LVs at least carry verbal features and may assign accusative case. Case mark<strong>in</strong>g ability <strong>of</strong><br />

the LVs is not semantically restricted. This means that the LVs may case mark DPs<br />

which are not the arguments <strong>of</strong> the LVs. In Megerdooma<strong>in</strong> (2001; 2002) and Folly,<br />

Harley and Karimi (2005), the <strong>in</strong>ternal arguments are contributed by the PVs and they<br />

may be case marked as direct objects. Obviously, they receive their case from the LVs.<br />

When the LVs can case mark the arguments <strong>of</strong> PVs we should expect them to be able to<br />

case mark the DP PVs too, provided that the PV is predicative (36). On the contrary, <strong>in</strong><br />

the <strong>in</strong>separable LVCs (18a, 18c, 32-33), the PVs cannot be augmented <strong>in</strong>to DP and<br />

cannot receive accusative case from the LVs because they do not satisfy the requirements<br />

<strong>in</strong> (36).<br />

Reference:<br />

Abney, S. (1987) The English Noun Phrase <strong>in</strong> its Sentential Aspects. Ph.D. Dissertation,<br />

Cambridge: MIT.<br />

Ahn, H. (1990) 'On light verb construction <strong>in</strong> Korean and Japanese,' In Hajime Hoji<br />

27


(ed.) Japanese and Korean L<strong>in</strong>guistics 221-37. Stanford: Centre for Study <strong>of</strong> language<br />

and Information.<br />

Ahn, H. (1991) <strong>Light</strong> <strong>Verb</strong>s, VP Movement, Negation and Clausal Architecture <strong>in</strong><br />

Korean and English . Ph.D. Dissertation, The University <strong>of</strong> Wiscons<strong>in</strong>, Madison.<br />

Als<strong>in</strong>a, A. (1997) 'Causatives <strong>in</strong> Bantu and Romance'. In Complex predicates. (eds.) A.<br />

Als<strong>in</strong>a et al. Standford: CSLI. 203-246.<br />

Barjaste, D. (1983) Morphology, Syntax and Semantics <strong>of</strong> Persian Compound <strong>Verb</strong>s: A<br />

lexicalist Approach Ph.D. Dissertation, University <strong>of</strong> Ill<strong>in</strong>ois at Urbana-Campaign.<br />

Butt, M. (1995) The Structure <strong>of</strong> Complex Predicates <strong>in</strong> Urdu. Stanford, Clifornia: CSLI<br />

Publications.<br />

Butt, M. (2003) 'The <strong>Light</strong> <strong>Verb</strong> Jungle'. In G. Aygen, C. Bowern and C. Qu<strong>in</strong> (eds.),<br />

Harvard Work<strong>in</strong>g Papers <strong>in</strong> L<strong>in</strong>guistics, Vol., 9.<br />

Butt, M. and G. Ramchan. (2005). 'Complex Aspectual Structure <strong>in</strong> H<strong>in</strong>di/Urdu.'<br />

In Nomi Ertischik-Shir and Tova Rapoport (eds.) The Syntax <strong>of</strong> Aspect, 117–153.<br />

Oxford: Oxford University Press.<br />

Choi, I. and S. Wechsler, (2001) 'Mixed categories and argument transfer <strong>in</strong> the Korean<br />

LVCs'. In Proceed<strong>in</strong>gs <strong>of</strong> 8 th International HPSG Conference, Norwegian University<br />

<strong>of</strong> Science and technology, 3<br />

Chomsky, N. (1970) ‘Remarks on Nom<strong>in</strong>alization’ In Read<strong>in</strong>g <strong>in</strong> Transformational<br />

Grammar, (ed.) Rodrick Jacobs and Peter S. Rosenbaum. Boston: 184-221.<br />

Chomsky, N. (1995) The M<strong>in</strong>imalist Program. Cambridge Mass: MIT Press.<br />

Chomsky, N. (1999) Derivation by Phase. Manuscript. MIT.<br />

Chomsky, N., 2001. Beyond Explanatory Adequacy. Manuscript. MIT<br />

Cr<strong>of</strong>t, W. (1991) Syntactic Categories and Grammatical Relations: The Cognitive<br />

Organization <strong>of</strong> Information. Chicago: University <strong>of</strong> Chicago Press.<br />

Cr<strong>of</strong>t, W. (2001) Radical Construction Grammar: Syntactic Theory <strong>in</strong> Typological<br />

Perspective. Oxford: Oxford University Press.<br />

Cr<strong>of</strong>t, W. (2005) ‘Word classes, parts <strong>of</strong> speech, and syntactic argumention’. L<strong>in</strong>guistic<br />

Typology 9: 431-441.<br />

Dabir-Moghadam, M. (1995) ‘Compound <strong>Verb</strong> <strong>in</strong> Persian’. Iranian Journal <strong>of</strong> L<strong>in</strong>guistics<br />

Vol.12. No. 1&2, 2-46.<br />

Dies<strong>in</strong>g, M. (1998) '<strong>Light</strong> <strong>Verb</strong>s and the Syntax <strong>of</strong> Aspect <strong>in</strong> Yiddish'. The Journal <strong>of</strong><br />

Comparative L<strong>in</strong>guistics 1, 119-156.<br />

Dies<strong>in</strong>g, M. (2000). 'Aspect <strong>in</strong> Yiddish: The Semantics <strong>of</strong> an Inflectional Head'. Natural<br />

Language Semantics 8, 231-253.<br />

Di Sciullo, A.M . and Rosen, S.T. (1990) '<strong>Light</strong> and Semi-light <strong>Verb</strong> <strong>Constructions</strong>'. In<br />

K. Dziwirek, P. Farrel and E.M. Bikandi (eds.) Grammatical Relations. Stanford:<br />

Center for the Study <strong>of</strong> Language and Information.<br />

Evans, N. and T. Osada (2005) Mundari: The myth <strong>of</strong> a language without word classes.<br />

L<strong>in</strong>guistic Typology, 9: 351-391.<br />

Folli, R., Harley, H. and Karimi S., 2003. Determ<strong>in</strong>ants <strong>of</strong> the event type <strong>in</strong> Persian<br />

Complex predicates. In: Astruc, L. and M. Richards (eds.) Cambridge Occasional<br />

Papers <strong>in</strong> L<strong>in</strong>guistics.<br />

Ghomeshi, J. and D. Massam (1994) “Lexical/Syntactic Relations Without Projection,”<br />

L<strong>in</strong>guistic Analysis, 24: 173-217.<br />

Goldberg, A. E. (1995) 'Persian Complex Predicates and the Lexicon'. Manuscript,<br />

University <strong>of</strong> California, San Diego.<br />

Goldberg, A. E. (2003) ‘Words by Default: The Persian Complex <strong>Constructions</strong>'. In E.<br />

28


Francis and L. Machaelis (eds.), mismatches: Form-Function <strong>in</strong>congruity and the<br />

Architecture <strong>of</strong> Grammar, CSLI Publications, 83-112.<br />

Grimshaw, J. (1991) ‘Extended Projection’. Manuscript, Brandies University.<br />

Grimshaw, J. and A. Mester (1988). '<strong>Light</strong> <strong>Verb</strong>s and Theta-Mark<strong>in</strong>g'. L<strong>in</strong>guistic<br />

Inquiry, 19: 205-232 .<br />

Haegeman, L. (1994). Introduction to Government and B<strong>in</strong>d<strong>in</strong>g Theory (2 nd Edn). Oxford:<br />

Basil Blackwell.<br />

Hale, K. and S. J. Keyser (1993) 'On Argument Structure and Lexical Expression <strong>of</strong><br />

Syntactic relations,' In K. Hale and J. Keyser (eds.), Views from Build<strong>in</strong>g 20, MIT.<br />

Hale, K. and S. J. Keyser (2002) Prolegmenon to a Theory <strong>of</strong> Argument Structure.<br />

Cambridge, MA: MIT Press.<br />

Heny, J. and V. Samiian (1992) 'Three Cases <strong>of</strong> Restructur<strong>in</strong>g <strong>in</strong> Modern Persian,' In K.<br />

Hunt, T. Perry and V. Samiian (eds.), Proceed<strong>in</strong>gs <strong>of</strong> the Western Conference <strong>in</strong><br />

L<strong>in</strong>guistics.<br />

Karimi, S. (1987) 'Compounds <strong>in</strong> Persian,' In L<strong>in</strong>guistics Work<strong>in</strong>g Papers, University<br />

<strong>of</strong> Wash<strong>in</strong>gton, Department <strong>of</strong> L<strong>in</strong>guistics, Seattle.<br />

Karimi-Doostan, Gh. (1997) <strong>Light</strong> verb constructions <strong>in</strong> Persian. Ph.D dissertation. Essex<br />

University. England.<br />

Karimi-Doostan, Gh. (2001) 'N + V Complex Predicates <strong>in</strong> Persian,' In Structural Aspects<br />

<strong>of</strong> Semantically Complex <strong>Verb</strong>s, Dehe & Warner (eds.), 277-292. Frankfort: Peter<br />

Lang.<br />

Karimi-Doostan, Gh. (2002) ‘Compositionality Complex Predicates <strong>in</strong> Persian’. In<br />

Proceed<strong>in</strong>gs <strong>of</strong> the L<strong>in</strong>guistics Society <strong>of</strong> Korea vol. 1, 255-268. Seoul.<br />

Karimi-Doostan, Gh. (2005) '<strong>Light</strong> <strong>Verb</strong>s and Structural Case'. L<strong>in</strong>gua, 115: 1737-1756.<br />

Karimi-Doostan, Gh. (2008a) ‘Predicative nouns and adjectives’. Grammar 3: The Journal<br />

<strong>of</strong> Iranian Academy <strong>of</strong> Persian Language and Literature 3, 187-202.<br />

Karimi-Doostan, Gh. (2008b) ‘Event structure <strong>of</strong> verbal nouns and light verbs’. In S.<br />

Karimi, V. Samiian and D. Stilo (eds.), Aspects <strong>of</strong> Iranian L<strong>in</strong>guistics.<br />

Khanlari, P. N. (1987) Tarix-e Zaban-e Farsi (Persian Language History). Tehran: Nashr-e<br />

Nou Press.<br />

Kim, S. W., 1994. A Study <strong>of</strong> <strong>Light</strong> <strong>Verb</strong> <strong>Constructions</strong>. Language Research, 30; 137-<br />

159.<br />

Longobardi, F. (1994) 'Reference and Proper Names: A Theory <strong>of</strong> N-Movement <strong>in</strong><br />

Syntax and Logical-Form'. L<strong>in</strong>guistic Inquiry, 25: 609-665.<br />

Masullo, P. J. (1992) Incorporation and Case theory <strong>in</strong> Spanish. Ph.D. Dissertation,<br />

University <strong>of</strong> Wash<strong>in</strong>gton.<br />

Megerdoomian, K. (2000) Beyond words and Phrases: A unified Theory <strong>of</strong> Predicate<br />

Composition. Ph. D. Dissertation, University <strong>of</strong> Southern California.<br />

Megerdoomian, K. (2001) Event Structure and Complex Predicates <strong>in</strong> Persian. Canadian<br />

Journal <strong>of</strong> L<strong>in</strong>guistics, 46:97-125.<br />

Miyagawa, S. (1989) '<strong>Light</strong> verbs and the Ergative Hypothesis'. L<strong>in</strong>guistics<br />

Inquiry, 20: 659-668.<br />

Mohammad and Karimi (1992) '<strong>Light</strong> <strong>Verb</strong>s are taken over: Complex <strong>Verb</strong>s <strong>in</strong> Persian'.<br />

Proceed<strong>in</strong>gs <strong>of</strong> the Western Conference on L<strong>in</strong>guistics, Vol. 5: 195-212.<br />

Mohanan, T. (1994) Argument Structure <strong>in</strong> H<strong>in</strong>di. Stanford, Clifornia: CSLI Publications.<br />

Park, K. (1989) 'A Lexical Approach to the syntax <strong>of</strong> Compound verbs,' Harvard Studies<br />

<strong>in</strong> Korean L<strong>in</strong>guistics III : 319-328.<br />

29


Partee, B. (1992) 'Syntactic Categories and Semantic types,' In Rosener, J. and R.<br />

Johnson (eds.), Computational L<strong>in</strong>guistics and Formal Semantics. Cambridge<br />

University Press.<br />

Partee, B. H. (1987) 'Noun Phrase Interpretation and Type-Shift<strong>in</strong>g Pr<strong>in</strong>ciples,' In<br />

Groenendijk, J.and D. de Jongh and M. Stokh<strong>of</strong> (eds) Studies <strong>in</strong> Discourse<br />

Repres<strong>in</strong>tation Theory and the Theory <strong>of</strong> Generalized quantifiers. Dordrecht: Foris.<br />

Radford, A. (2004) M<strong>in</strong>imalist Syntax: Explor<strong>in</strong>g the structure <strong>of</strong> English. Cambridge:<br />

Cambridge University Press.<br />

Rosen, S. (1989) Argument Structure and Complex Predicates. Ph.D. Thesis, Brandeis<br />

University.<br />

Samiian, V (1983) Orig<strong>in</strong> <strong>of</strong> Phrasal Categories <strong>in</strong> Persian: an X-bar analysis. Ph.D.<br />

UCLA<br />

Sato, Y. (993) Complex Predicate Formation with <strong>Verb</strong>al Nouns <strong>in</strong> Japanese and Korean:<br />

Argument Transfer at LV. Ph. D. Dissertation, University <strong>of</strong> Hawaii.<br />

Schachter, P. (1985) Parts <strong>of</strong> Speech Systems. In Timothy Shopen (ed.) Language<br />

typology and Syntactic Description, Vol. I: Clause Structure, 3-61. Cambridge<br />

University Press.<br />

Sells, P. (1989) 'More on <strong>Light</strong> <strong>Verb</strong>s and Theta-mark<strong>in</strong>g'. Manuscript, Stanford<br />

University.<br />

Stowell, T. (1989) 'Rais<strong>in</strong>g <strong>in</strong> Irish and the projection Pr<strong>in</strong>ciple,' Natural Language and<br />

L<strong>in</strong>guistic Theory, 7: 317-360.<br />

Vahedi-Langrudi, M. (1996) The syntax, Semantics and Argument Structure <strong>of</strong> Complex<br />

Predicates <strong>in</strong> Modern Farsi. Ph.D. Dissertation, University <strong>of</strong> Ottawa.<br />

Wechsler, S. (2008) 'Idioms, <strong>Light</strong> <strong>Verb</strong>s and Lexical Decomposition'. NORMS Work<strong>in</strong>g<br />

on Argument Structure. Lund, February, 5-6, 2008.<br />

30

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!