25.01.2013 Views

popper-logic-scientific-discovery

popper-logic-scientific-discovery

popper-logic-scientific-discovery

SHOW MORE
SHOW LESS
  • No tags were found...

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

446<br />

new appendices<br />

more abstract. There are layers of higher and higher degrees of universality,<br />

and thus of transcendence. (In section *15 of the Postscript, an<br />

attempt is made to explain the sense in which these are also layers of<br />

what may be called ‘depth’.)<br />

It is of course because of this transcendence that <strong>scientific</strong> laws or<br />

theories are non-verifiable, and that testability or refutability is the only<br />

thing that distinguishes them, in general, from metaphysical theories.<br />

If it is asked why we use these transcendent universal laws instead of<br />

keeping more closely to ‘experience’, two kinds of answer may be<br />

given.<br />

(a) Because we need them: because there is no such thing as<br />

‘pure experience’, but only experience interpreted in the light of<br />

expectations or theories which are ‘transcendent’.<br />

(b) Because a theorist is a man who wishes to explain experiences, and<br />

because explanation involves the use of explanatory hypotheses which<br />

(in order to be independently testable; see section *15 of the Postscript)<br />

must transcend what we hope to explain.<br />

The reason given under (a) is a pragmatic or instrumentalist one,<br />

and although I believe that it is true, I do not think that it is comparable<br />

in importance with the reason given under (b); for even if a programme<br />

of eliminating explanatory theories for practical purposes<br />

(say, for prediction) were to succeed, the aim of the theorist would be<br />

unaffected. 4<br />

4 That it is possible to do without theories is asserted by Carnap, Logical Foundations of<br />

Probability, pp. 574 f. Yet there is no reason whatever for the belief that Carnap’s analysis,<br />

even if it were otherwise defensible, could be legitimately transferred from his model<br />

language to ‘the language of science’; see my Preface, 1958. In two very interesting articles<br />

W. Craig has discussed certain reduction programmes. (See Journal of Symb. Logic 18, 1953,<br />

pp. 30 f., and Philosophical Review 65, 1956, pp. 38 ff.) To his own excellent critical comments<br />

on his method of eliminating ‘auxiliary’ (or ‘transcendent’) ideas, the following<br />

might be added. (i) He achieves the elimination of explanatory theories, essentially, by<br />

promoting infinitely many theorems to the rank of axioms (or by replacing the definition<br />

of ‘theorem’ by a new definition of ‘axiom’ which is co-extensive with it as far as<br />

the ‘purified’ sub-language goes). (ii) In the actual construction of the purified system,<br />

he is of course guided by our knowledge of the theories to be eliminated. (iii) The purified system<br />

is no longer an explanatory system, and no longer testable in the sense in which<br />

explanatory systems may be testable whose testability is, essentially, related to their

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!