The Themes of Quine's Philosophy.
Edward Becker (University of Nebraska, Lincoln)
New York: Cambridge University Press, 2012, 314 pp.
Becker points out that Quine's view treats the statements of logic and mathematics as differing only in degree, not in kind, from other statements. They are similar, in point of cognitive status, to the statements of theoretical physics. Physics is factual, and its posited objects are real. According to Quine's "naturalism," the facts simply are what our current physical theory says they are. The truths of logic and mathematics are testable only insofar as they are included in testable theories. Statements that are definitional abbreviations of truths of logic should qualify as truths of logic. The thesis of definitional conventionalism, that mathematical truths are logical consequences of definitions, amounts to the claim that mathematical truths are logical truths. Logical truth is a statement that is either itself a truth in which only logical expressions occurs essentially or is a definitional abbreviation of such a truth, whereas mathematical truth is a statement that is either itself a truth containing only logical and mathematical expressions essentially or is a definitional abbreviation of such a truth. The formulation of definitions of all mathematical expressions in purely logical terms is necessary for the tenability of the thesis that mathematics reduces to logic. Truths outside of logic and mathematics can be generated by conventions in the same way as truths found within these disciplines. Statements that are "true by definition" are more accurately described as conventional transcriptions of logical truths. If in "Truth by Convention" the logical truths comprise truths in which only logical words occur essentially and definitional abbreviations of such truths, in "Two Dogmas of Empiricism" statements reducible to logical truths via definitions comprise a category of truths whose defining characteristics are less clear than those of the logical truths. The traditional distinction between a priori and a posteriori truths can be explained behavioristically as a distinction between more and less firmly accepted statements. The fact that the adoption of an alternative logic involves a change in the meanings of logical words seems to support the linguistic doctrine of logical truth. The analytic truths include the logical truths plus those true statements which hold by "essential predication." Those statements that are "analytic" and thus "true" by language are true by language plus logic. To say that a statement is analytic is to say that it is true by language, or true in virtue of the meaning of its words. Although we can make sense of the analytic/synthetic distinction, we cannot make sense of it in a way that is epistemically relevant. Analyticity is not relevant to epistemology. The distinction between "analytic" truths, true purely because of language, and "synthetic" truths, true because of how the world is, falls short of empiricistic standards. Intuition rates the logical truths as belonging to the extension of "analytic." An analytic sentence is one whose truth everyone learns by learning its words. A sentence's genesis is irrelevant to its current epistemic status. The notion of analyticity that we can make sense of cannot be used in making sense of a priori language. A sentence is stimulus analytic for a subject if he would assent to it, if to anything, after every stimulation. Stimulus analyticity can be made to coincide more nearly with its intuitive counterpart by socialization. The relevant stimulation precedes and prompts the subject's assent or dissent. The stimulus meaning is the ordered pair consisting of the affirmative stimulus meaning and the negative stimulus meaning. A stimulation is an event in some person's perceptual apparatus. The stimulus analytic sentences are those to which everyone would assent following any stimulation. A "stimulation" is the action of something on one of a speaker's sensory surfaces.
It follows from the preceding analysis that cognitive equivalence is just stimulus synonymy. The notion of the-feature-that-is-more-prominent/-salient is explicable in behavioral terms. Becker remarks that Quine's brand of behaviorism does not reject innate mechanism. The best account we have of the actual facts is the description of the natural world given by contemporary science. Only substantial chunks of scientific theory or commonsense belief can be confirmed or refuted by experience. Quine rejects the distinction between meaning and collateral information as determinants of responses. The attempt to make sense of conventionalism leads to an infinite regress. If the linguistic theory is to serve as a principle of empiricist philosophy, it should itself meet the empiricist standard of significance, i.e., there should be some way of testing it against experience.
Quine expresses doubts concerning the possibility of explaining synonymy. Making sense of synonymy and of the analytic/synthetic distinction would involve explaining them in behavioral terms. Statements are cognitively synonymous if their biconditional is analytic. Sentences qualify as synonymous with one another even though their corresponding parts differ in meaning. We cannot explain synonymy and analyticity by appealing to the notion of explication. Synonymy cannot be clarified by an appeal to the notion of definition. Interchangeability salva veritate in all contexts is not a necessary condition of synonymy. A satisfactory explication of synonymy would have to be stated in terms pertaining to linguistic behavior. Synonymy of two sentences for a given speaker cannot be equated with stimulus synonymy of those sentences for him. We can't define synonymy as sameness of utterance conditions (this would imply that no two sentences are synonymous). Quine's explications of cognitive equivalence are not to be taken as explications of synonymy. Synonymy as a relation between expressions in different languages is unintelligible. Synonymy cannot be explicated as stimulus synonymy. Synonymy is beyond rehabilitation: its reprobate status is revealed by the indeterminacy of translation. Explication appeals to synonymy at the level of sentences. The notion of semantical equivalence of sentences is more fundamental than that of synonymy of words. Analyticity involves reducibility to elementary logic via definitions in which definiens and definiendum are synonyms. If logic is to be true purely by convention, the conventions upon which it is ultimately based must be other than definitions. Definitions determine the truth or falsity of statements relative to the truth or falsity of other statements. Statements which are loosely described as being "true by definition" are true by definitions plus logic. All truths, including those of logic and mathematics, are subject to revision in the light of experience. The logical truths are those truths in which only logical words occur essentially. In testing non-logical words for essential occurrence one must vary them simultaneously. We learn of logical laws' truth in the process of learning their component logical particles. The attempt to derive the truths of logic from general conventions must appeal to logic in the metatheory.
Quine rejects the thesis that each statement, taken by itself, has a unique empirical meaning. The possibility of holding any statement true, combined with the possibility of rejecting any statement, makes the analytic/synthetic distinction untenable. If a synthetic statement is construed as one which we might be compelled to give up, there are no synthetic statements. We cannot make sense of the distinction between statements whose truth can be known independently of any confirming experience, and statements whose truth cannot be so known. No statement is immune to revision without having to accept that no statement is analytic. Many statements are confirmable only in the context of a theory. A verificationist view of meaning combined with a Duhemian view of evidence implies the indeterminacy of translation. The indeterminacy thesis is not to be equated with the point about alternative grammatical systems. There would be little or no indeterminacy in the translation of observation sentences. In the case of translation by a bilingual, the translation of all occasion sentences, observational or not, would be free of indeterminacy. When questions of translation are seriously raised, the only evidence to which we could appeal in setting them would be behavioral. Certain questions about translation and meaning cannot be settled in any way at all. In cases where translation is indeterminate there is no failure of discernment, there is nothing to be discerned. Truths about the meanings of standing sentences are nowhere to be found. The linguist has some limited success in translating native utterances by querying whole native sentences and determining their stimulus meanings. Different systems of analytical hypotheses generate radically incompatible translations of native sentences. The main argument for the indeterminacy of translation is based on the inscrutability of reference. Only some cases of inscrutability of reference involve indeterminacy of sentence translation. There may be some cases in which inscrutability of terms result in indeterminacy in the translation of sentences. One could argue for the indeterminacy thesis on the basis of the inscrutability of terms. Some cases of the inscrutability of terms entail indeterminacy. The main argument for the claim that there is indeterminacy appeals to problems in the translation of theoretical sentences. The under-determination of physics entails the indeterminacy of translation. Appealing to verificationism and holism is the best way of arguing for the indeterminacy thesis. Ontology is relative to manuals of translation.
What matters for the present discussion is that ontological relativity is a question of relativity relative to a manual of translation. On Becker's view, Quine's acceptance of the thesis of the indeterminacy of translation does not commit him to repudiating lexicography or empirical semantics. The needed behavioral account cannot be found, as it would have to invoke a distinction between assent based purely upon our understanding of language and assent based partly upon non-linguistic information about the world. Weak indeterminacy says that translation is under-determined by behavior, whereas strong indeterminacy says that it is under-determined by behavior plus scientific method. There is no fact of the matter with respect to translation. Conflicting schemes of translation are compatible with all the truths about nature established by our physical theory. As long as conflicting manuals conform to all verbal dispositions, there is no fact of the matter about translation. To say that a person is referring to certain objects is to say that his language can be translated into a fragment of our language involving quantification over those objects. The interpretation of truth-functional connectives presents no insurmountable problems. The early stages of radical translation are obstructed by problems about the intrusion of collateral information and about how terms divide their reference.
Quine assigns the empirical content of theories to "observation categoricals." Whenever the circumstances described in a certain observation sentence obtain, so do the circumstances described in another observation sentence. Quine avoids having to eternalize observation sentences or relativize them to places and times. An observation is categorical "analytic" if the affirmative stimulus meaning of its first component is included in that of the second; other observation categoricals are "synthetic." Observation categoricals are "synonymous" if their respective components are stimulus synonymous. The empirical content of a testable sentence or set of sentences for a speaker is the set of all the synthetic observation categoricals that it implies, plus all synonymous observation categoricals. Systems of analytical hypotheses are justified if they equate the foreigner's observation sentences with stimulus synonymous observation sentences of English. Different speakers of English differ in their "neural hardware."
The basic idea here is that the analytical apparatus developed by Quine on the basis of his notion of stimulus meaning can be used to reconstruct synonymy in terms that would meet his standards of clarity. Becker writes that, according to Quine's holism, only substantial chunks of scientific theory, not isolated theoretical sentences, have empirical implications. Quine's thesis of the inscrutability of reference says that various translations of foreign terms, ascribing various extensions to those terms, can be behaviorally equivalent. The inscrutability extends even to the references of the terms of one's home language. The reference of terms is inscrutable. Synonymy of terms cannot be established on the basis of stimulus synonymy of the corresponding one-word sentences. The introduction of synonymy into semantics or lexicography would be a hindrance to the proper work of these disciplines. Intrasubjective stimulus synonymy approximates more closely than its intersubjective counterpart to synonymy as intuitively conceived. The stimulus synonymy of "Indian nickel" and "Buffalo nickel" can be explained as the result of the influence of collateral information. Questions concerning the synonymy of standing sentences do not have a determinate answer. Synonymy as a relation between standing sentences does not make sense. Given the stimulus synonymy of the one-word sentences "Gavagai" and "Rabbit," one could settle the reference of "gavagai" as a term. A stimulation presenting a rabbit-fly but no rabbit will prompt the informant to assent to "Gavagai" and the linguist to dissent from "Rabbit." A linguist engaged in radical translation should begin with observation sentences. There is no way for the linguist to arrive at uniquely correct radical translations of terms. The identification of native locutions as terms involves the identification of other native locutions as auxiliaries to objective reference. The translation of terms presupposes the translation of auxiliaries to objective reference. The stimulus synonymy of the sentences "Gavagai" and "Rabbit" is compatible with various possible translations of "gavagai" as a term. The issue among the alternative translations of "gavagai" cannot be settled by pointing. The thesis of the inscrutability of reference can be applied to our fellow speakers of English. Reference is meaningful relative to a coordinate system. It is "nonsense" when not so relativized. The background language gives our query about the reference of "rabbit" "relative sense."
It is worth emphasizing that the doctrine of the inscrutability of reference says that it is nonsense to ask about the references of our terms. Becker states that Quine's "relational theory" of objects speaks of how one theory is interpretable or reinterpretable in another. "What are numbers?" or "How should number theory be interpreted in set theory?" have no uniquely correct answers. Questions about the entities to which we are committed are ones about the domains of discourse of the theories we accept. Questions about the references of terms make sense relative to background languages. The references of the terms in a background language are inscrutable. A "background language" functions as both a metalanguage and a subject language. Reference is relative to a background language: statements about the denotations of the terms of a foreground language are proposals for translating those terms into the background language. Ontology is relative to both a background theory and a manual of translation. When the background theory is a containing theory translation takes the degenerate form of homophonic translation. Statements about the denotations of terms must be construed as proposals for translating someone else's terms into our own. Relativization to a background theory includes relativity to a choice of background theory and relativity to the choice of how to translate the object theory into the background theory.
What these latter observations reveal is that statements about the denotations of terms are to be construed as statements about how we propose to translate those terms into a background language. Becker claims that Quine speaks about interpretation because he wants to address the question of what is involved in specifying a theory's universe of discourse, and about reinterpretation because he wants to extend his relativistic thesis to the case of reduction, whereas to interpret an ontology is to say what its objects are (to reinterpret one is to say that its objects can be modeled in a smaller universe of different objects). Questions and statements about the ontology of a theory or about the references of its terms make sense only "relative to" a background language. Talk about the denotations of terms must be couched in a background language and reconstrued as talk about how we propose to translate those terms into the background language. We must use a back ground language in speaking of reference or ontology. We can meaningfully ask questions about the references of terms only "relative to" a background language. "Relativity to" a network of expressions or to a background language is not relativity in the usual sense. Quine compares the regress of background languages in specifying the references of terms to the regress of coordinate systems in specifying position and velocity. A language can sometimes be used as a background language for itself. Quine compares the use of networks of expressions or of background language with the use of frames of reference or coordinate systems.
As is no doubt obvious, statements about ontology are ones about how one theory of objects can be interpreted or interpreted in another. Becker maintains that Quine's theory of objects construes statements about ontology as statements about the relations between theories of objects. Global questions about the references of a language's terms are ones about how we propose to translate those terms into a different background language. Questions about the position or velocity of a coordinate system are ones about the position of that system in a different coordinate system. Questions about ontology are ones about the relations between theories of objects. Statements about the position or velocity of coordinate systems should be construed as statements about their relations to other coordinate systems. Questions about the references of terms make sense relative to background languages. The references of the terms in a background language are inscrutable. Because of the inscrutability of reference, we should make proposals for translating terms of the foreground language into the background language. Inscrutability can be applied to ourselves and such application leads to absurdities. This "quandary" can be resolved by relativizing reference to coordinate systems. It is meaningless to question the references of all of our terms. There is no fact of the matter as to what our terms refer to. Relative to our "coordinate system " of terms, predicates, and auxiliaries we can and do talk meaningfully and distinctively of rabbits and parts. "Relativity" to background languages is a matter of using a background language rather than mentioning one. Statements about reference have to be "relativized to" background languages, being couched in a background language and construed as proposals for translating terms into the background language. "Inscrutability of reference" describes the thesis that there is no fact of the matter as to the references of a person's terms. Interpreting a person's ontology is a matter of specifying the values of his variables and the references or extensions of his terms. There is no fact of the matter of how we should interpret a person's terms. There is no fact of the matter of the references of the terms of other speakers of our language. There is a fact of the matter of the references of our own terms and there is a fact of the matter of the references of other speakers of our language. Our current theory takes the references of our own terms at face value. There is no such thing as reference, conceived as a relation between words and objects. The relations of terms to one another within a theory and the ones of translation amongst the various reference schemes are real. The inscrutability of reference is an epistemological thesis. Coextensiveness is a clear notion as it applies to our own language. Inscrutability applies to foreign languages but not to the home language. There is a fact of the matter of the references of one's own terms. Quine appeals to innate quality spaces in explaining language learning.
I recommend Becker's addition to the ever expanding literature on Quine's philosophy, a body of scholarship to which Becker has contributed significantly.
Reviewed by George Lazaroiu, PhD
Institute of Interdisciplinary Studies in
Humanities and Social Sciences, New York
|Printer friendly Cite/link Email Feedback|
|Publication:||Review of Contemporary Philosophy|
|Article Type:||Book review|
|Date:||Jan 1, 2013|
|Previous Article:||About Europe: Philosophical Hypotheses.|
|Next Article:||Imagining the University.|