164 INTRODUCTION TO PHILOSOPHY OF SCIENCE 



LOGICAL POSITIVISM 



The placing of logical positivism 1 in the context of con- 

 ventionalism and construct ionalism may strike one as some- 

 what unusual. Yet there is an important sense, it seems, 

 in which this position has close affiliations with these mod- 

 ified positivisms. It certainly is positivistic in its high esteem 

 for Mach, in its insistence that every statement of science 

 "should be based on and reducible to statements of empirical 

 observations," 2 and in its definite rejection of the a priori. 

 But it cannot be a strict positivism for it recognizes the 

 existence of certain types of statement which play a part in 

 science and yet are not empirically verifiable. One thinks 

 immediately of Hobson's idealizations in this connection, 

 until he is told that these statements are incapable of truth 

 or falsity in the ordinary sense of the term, and hence cannot 

 be mere idealizations of the data. One then suspects that 

 they are simply Poincare's conventions; but this proves 

 also to be an erroneous interpretation, for they are strictly 

 without content or meaning — purely formal statements 

 which say nothing and therefore could not be even con- 

 venient or inconvenient. Hence they are symbols which 

 have so completely lost their reference to data that the lat- 

 ter do not function at all in the determination of the char- 

 acter of the symbols. Are the symbols, then, purely arbi- 

 trary? No, they are subject to the rules of language — 

 rules according to which one constructs complex symbols 

 out of simple symbols, and translates symbols into other 

 symbols. But if language itself is merely a device which 

 one employs when he knows — an operator, so to speak — 

 then the purely formal symbols of the logical positivists are 

 descriptive of this device. They are symbols which de- 

 scribe the operations of knowing rather than the data 

 known. They are the values which & takes in the formula 

 aS =f(D,0), when D becomes zero and becomes the 



1 See Chapter I, pp. 11-13. 



2 R. Carnap, The Unity of Science (London: Kegan Paul, 1934), p. 27. 



