the analytic and the synthetic- a tenable dualism

Philosophical Review The Analytic and the Synthetic: A Tenable "Dualism" Author(s): Arnold S. Kaufman Source: The Philosophical Review, Vol. 62, No. 3 (Jul., 1953), pp. 421-426 Published by: Duke University Press on behalf of Philosophical Review Stable URL: http://www.jstor.org/stable/2182881 Accessed: 20/04/2009 03:47 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=duke. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org. Duke University Press and Philosophical Review are collaborating with JSTOR to digitize, preserve and extend access to The Philosophical Review. http://www.jstor.org THE ANALYTIC AND THE SYNTHETIC A TENABLE "DUALISM" m HE FIRST part of this paper will constitute a brief exposition of the views held by Morton White and W. V. Quine regarding the analytic-synthetic issue. The expository portion will be restricted to one phase of the problem: the attempt to formulate criteria in terms of which analytic may be distinguished from synthetic judgments. White endeavors to show the futility of attempting to establish be- havioral criteria. Quine deals with various definitions of "analytic" and "synthetic" and tries to show that none of them is adequate. Quine explicitly indicates, in a footnote, that he is in complete agree- ment with White's analysis and is merely trying to extend his argu- ments.' White describes a primitive community, the members of which are to be investigated in order to determine whether visiting anthropolo- gists have correctly characterized the statements, "Man is a featherless biped" as synthetic for the natives, and "Man is a rational animal" as analytic for them. After discarding-a few alternatives he remarks: Probably the most helpful interpretation of this mode of distinguishing ana- lytic and synthetic is that according to which we obtain the following: when the natives have applied the word "man" to certain objects and are then persuaded that these objects are not rational animals, they immediately, without hesitation, withdraw the predicate " man." They contemplate no other means of solving their problem. But when they have applied the word " man" and are then persuaded that the things to which they have applied it are not featherless bipeds, they do not withdraw the predicate "man" immediately but rather contemplate another course, that of surrendering the hypothesis that all men are featherless bipeds.2 White objects to this criterion, indicating that although he thinks it may be "workable," it will "result in our finding that many firmly believed 'synthetic' equivalences are analytic." Quine assumes the adequacy of White's analysis, and attempts to determine whether any proposed definition of "analyticity" will do the 1W. V. Quine, "Two Dogmas of Empiricism," Philos. Rev., LX (January, I951), p. 20. 'M. White, "The Analytic and the Synthetic: An Untenable Dualism," in John Dewey: Philosopher of Science and Freedom, ed. Hook, p. 328. 42I THE PHILOSOPHICAL REVIEW job. Both Quine and White distinguish two types of assertions tradi- tionally called analytic. The first constitutes the class of logical truths, and are of the form "if p then p" or "No unmarried man is married." The second is typified by "No bachelor is married." It is characteristic of the latter type that they may be transformed into logical truths by putting synonyms for synonyms. Thus, by substituting "unmarried man" for "bachelor" we obtain a statement of the first type. But, both Quine and White contend, the notion of "synonymy" is no clearer than "analyticity." They concede that they will regard their objection as overruled if we are able to define or better understand "synonymy." Part of Quine's argument is designed to show that no definition of "synonymy" is possible. Preliminary to his discussion of "synonymy," Quine discusses the proposal that "analytic statement" be defined in the following way: ". . . a statement is analytic when it is true by virtue of its meanings and independently of fact." He suggests this will not do because of difficulties of understanding what is meant by "meanings." He dis- tinguishes between a theory of meaning and a theory of reference and concludes: "Once the theory of meaning is sharply separated from the theory of reference, it is a short step to recognizing as the business of the theory of meaning simply the synonymy of linguistic forms and analyticity of statements; meanings themselves, as obscure inter- mediary entities, may well be abandoned." 3 Thus, if we ask, "What does 'creature with a kidney' mean?," we do not request the denotation or enumeration of objects to which the expression can be correctly applied. In order to answer this question, we must provide some sort of synonymous linguistic form. But if this is how we answer such a question, why speak of "meanings" at all? It is sufficient that we speak of synonyms. However, having discarded these "obscure intermediary entities," defining either "synonymy" or "analyticity" in terms of "meaning" is precluded. It is clear that if Quine's objections to defining "analyticity" in terms of "meaning" are rejected, as far as our problem is concerned the attempt to define "synonymy" loses its prime significance. It is here contended that "analyticity" should be defined in terms of "meanings." Therefore, the various attempts to define "synonymy," which Quine alternately enunciates and attacks, is not discussed in this paper. 'Quine, op. cit., pp. 22-23. 422 DISCUSSION II What have the arguments offered by White and Quine proved? I think they have successfully demonstrated that if we want a criterion or set of criteria, which, if applied in the case of any assertion having the grammatical form of a declarative sentence, will enable us to de- termine whether that sentence in that particular context is analytic or synthetic, then our search is doomed to failure. Neither behavioral criteria nor any other kind of definition seems adequate to the task. Does this force us to discard the distinction? I will argue, first, that the difficulty in establishing criteria, be- havioral or otherwise, is not sufficient reason for abandoning the analytic-synthetic distinction; just as a driver would not be expected to give up his distinction between the left and right-hand sides of the road because the Bureau of Highways has neglected to authorize white stripes down the middle of a few outlying thoroughfares. The analogy is apt because it conveys the idea that men are able to fix the distinction whenever it becomes important to do so. Second, I shall argue that an analytic statement is true by virtue of the meanings of component ex- pressions. This two-fold analysis involves answering two questions: (i) What function does the distinction serve? and (2) If the distinc- tion is useful, how are we to explicate it? In answering the first question, let us consider a fictitious historical illustration. Many years ago two famous biologists, Sam and Max, engaged in a long dispute. At issue was the status of the whale. Was it mammal or fish? Sam reasoned as follows: i. All mammals are terrestrial. 2. Whales are not terrestrial. 3. Whales are not mammals. Max insisted that whales were mammals, and that Sam's basic assump- tion, "All mammals are terrestrial," was in error. The controversy raged for many months. The papyrus was filled with arguments pro and con. Biologists split into two camps over the issue; the Mammalians and the Nonmammalians. The country in which the dispute occurred was strange indeed. The king, a -wise and noble fellow, had decreed that all scientific con- troversies extending over a twelve-month period be investigated by his court logicians, and an authoritative decision as to the character and most plausible resolution of the dispute handed down. After twelve months had passed, Sam and Max not yet having come to agreement, 423 THE PHILOSOPHICAL REVIEW George, the king's head logician, was assigned to the case. George saw quickly that if the disagreement over the initial assumption, viz., "All mammals are terrestrial," could be resolved, the controversy would evaporate. He decided to determine whether this assumption, let us call it S, was analytic or synthetic for Sam. If analytic, then the problem would become one of more or less adequate linguistic con- ventions. But, if synthetic, then evidence would be the final arbiter. George asked Sam, "Would a creature being in every respect like a dog, except that his respiratory organs permitted him to inhabit the bottom of the sea, be a mammal?" Sam's reply was entirely incon- clusive. The interrogation continued, Sam admitting that he might be wrong about S, but failing to conceive of any circumstance which would cause him to alter his judgment. What is George to conclude as a result of his investigation? Should he, because he is unable to come to any decision regarding S, as related to Sam, jettison the distinction between analytic and synthetic? As a matter of fact, George concluded that Sam's failure to be decisive about S indicated, not the bankruptcy of the distinction, but precisely that Sam had not been clear and specific about the meanings of the component expressions. It was with good reason that the king had chosen George to be head logician. He was aware that the distinction between analytic and syntheticJunctions as a stimulus to clarity; not necessarily as descriptive of actual usage. At any given stage of inquiry it is indeed possible to point to various statements employed with such universal ambiguity as to defy charac- terization in terms of analytic or synthetic. But far from undermining the validity of the distinction, it calls attention to the need for greater specificity and clarity in linguistic usage. Such distinctions exist for the sake of inquiry; it matters little that certain sentences do not exemplify these dualismss." We discriminate between analytic and synthetic, not because most assertions can be characterized in one of two ways. It is rather as a result of our desire to reform usage, for the purposes of inquiry, that we establish the distinction as a functional ideal which can, in given contexts, be fruitfully attained. It might also be remarked that the precise status of a statement (S) may vary. At one time S is analytic; at another, synthetic. There is no inconsistency here; only awareness that the demands of inquiry are ever changing. Whereas defining "mammal" in terms of "terrestrial" may have suited the needs of biologists at one stage of their science, subsequent discoveries have forced change. Choosing between alterna- tive definitions of a concept is often an entirely arbitrary matter. For example, whether the phrase "objects x and y are of equal weight" be 424 DISCUSSION defined in terms of a lever's equilibrium, or in terms of elongating a spring, equal lengths is a matter of indifference. However, once a definition is chosen, that weights, already known to be equal, behave in prescribed fashion in the other situation, becomes an empirical matter. Thus, assuming there are no special reasons for preferring one definition over the other, the definition of "objects x and y are of equal weight," is bound to shift from context to context. But in any specific context we should, barring lack of clarity, be able to determine which statement is analytic; which synthetic. If this is not the case we can be quite certain that there is some confusion. However, if the distinction between "analytic" and "synthetic" does play an important role in inquiry, it is fair to ask for an adequate explication. But the fact that the distinction has important conse- quences guarantees its retention no matter what difficulties are in- volved in defining the term "analytic" and distinguishing it from "synthetic." I shall suggest one avenue of analysis in the hope that it may prove useful for dealing with the problem raised by Quine and his followers. Both for Quine and for White the crux of the problem is the practi- cal impossibility of finding behavioral criteria by reference to which a statement in a natural language may be determined as analytic or synthetic. More fundamental is the tacit assumption that even under optimum conditions the rules governing linguistic behavior are so ill- defined, so ambiguous, and so vague, as to doom any attempt to make these rules explicit. But if what has been said is well taken, these con- siderations are irrelevant to the important problem that arises in the context of inquiry: that of fixing meanings, thus obtaining clarity of discourse. In a highly sophisticated scientific debate, the use of lan- guage suffices to distinguish analytic from synthetic judgments. More- over, these verbal cues are important, not for the sake of identifying a sentence as analytic or synthetic, but as indicative of which meanings are, and which are not, fixed in the context of a given discussion. It is only when we have fixed the meanings that we can begin to determine the status of various assertions. A disciple of Quine might object, "By what right do you introduce these obscure intermediary entities, meanings? Are you accepting the view that analytic propositions are true by virtue of meanings? But then how do you cope with Quine's devastating critique of such abstract fictions?" In fact, it can be shown that there are no paradoxes, no philosophi- cal difficulties involved in defining "analytic statement" as "statement true by virtue of the meanings of component expressions." 425 THE PHILOSOPHICAL REVIEW What has Quine's critique of "meaning" proved? He claims to have shown that this conception is unfruitful. The purposes of the theory of meaning, he observes, are served through formulation of synonymous expressions. Thus, when we ask, "What does 'x' mean?," the reply comes in the form of a linguistic expression, 'y', 'y' being a synonym for 'x.' But, though this be granted, one wonders whether the use of such synonyms in answering questions about meaning warrants out- lawing the conception "meaning." After all, our original purpose is usually, not the formulation of a synonymous expression, but communication. It is not the only method. Often we point, gesture, draw pictures. It frequently happens that what we take to be a synonymous ex- pression is not understood by the party with whom we are conversing. Nevertheless, we rely on a common fund of meanings which are prior to any specific attempt at communication. How, one may ask, can we be sure that these meanings of expressions are stable, clear, and uni- vocal? The answer is, of course, that we cannot always be certain. We can only assume this to be the case as long as we seem to com- municate. For example, I can be reasonably sure my brother and I attribute the same meaning to "uncle" as long as he calls those things "uncle" which I believe are his uncles, and, when requested, defines "uncle" or agrees to a suggested definition of that term which I would also accept. Perhaps much more fundamental reasons for dispatching "mean- ing" to the graveyard of archaic concepts underly Quine's argument. But the one explicitly offered in his paper is certainly inadequate. The distinction between analytic and synthetic functions as a means of eliciting clearer formulations. We make sentences analytic or syn- thetic by fixing the meanings of component expressions. Thus, it seems quite reasonable to explicate the distinction by maintaining that analytic sentences are true by virtue of meanings; synthetic judg- ments, true or false by virtue of the relevant facts. ARNOLD S. KAUFMAN The City College of New rork 426 Article Contents p.421 p.422 p.423 p.424 p.425 p.426 Issue Table of Contents The Philosophical Review, Vol. 62, No. 3 (Jul., 1953), pp. 327-492 Front Matter The Subjective as a Problem: An Essay in Criticism of Naturalistic and Existential Philosophies [pp.327-354] Ethical Empiricism and Its Critics [pp.355-373] Frege's Theory of Functions and Objects [pp.374-390] The Use of Simplicity in Induction [pp.391-408] Discussion Quine on Classes and Properties [pp.409-412] What is an Individual? [pp.413-420] The Analytic and the Synthetic: A Tenable "Dualism" [pp.421-426] A Note on the Reflexive Paradoxes [pp.427-428] Reviews of Books untitled [pp.429-432] untitled [pp.432-437] untitled [pp.437-443] untitled [pp.443-447] untitled [pp.447-450] untitled [pp.451-453] untitled [pp.454-458] untitled [pp.458-461] untitled [pp.461-464] untitled [pp.464-467] untitled [pp.468-472] untitled [pp.473-475] untitled [pp.476-479] untitled [pp.479-480] untitled [pp.481-482] untitled [pp.482-484] untitled [p.484] untitled [pp.484-485] Books Received [pp.486-492] Note [p.492] Back Matter