Truth (Princeton Foundations of Contemporary Philosophy) Read online

  The result is that initially the truth predicate applies only to our own sentences, here and now. For where indexicality is present, (17a) holds no further than that.

  To apply the truth predicate to the sentences of other speakers of our language, or even our own at other times and places, we must first transpose, in precisely the manner needed to turn direct into indirect quotation:

  (18a) WRONG: If Pope Benedict says to Tony Blair, “We disagree with you about gay rights,” then what he says to him is true iff we (the authors) disagree with you (the reader) about gay rights.

  (18b) RIGHT:…iff he (the pontiff) disagrees with him (the ex-premier) about gay rights.

  More importantly, to apply the truth predicate to speakers of other languages we must first translate, thus:

  (19) WRONG: If Carmen sings «L'mour est un oiseau rebelle, que nul ne peut apprivoiser», then what she sings is true iff l'amour est un oiseau rebelle, que nul ne peut apprivoiser.

  (19b) RIGHT:…iff love is a rebellious bird that none can tame.

  How to transpose—to change pronouns and so forth in going from direct to indirect quotation—is something we know as part of our knowledge of our native language. How to translate is not; but with a language like French, where there is an established bilingual community, its customs and conventions establish a standard for correct translation if nothing else does.

  Quine notoriously held that with radical translation, as between our language and that of some group with which there has been no previous contact, there is no right and wrong translation, but only better or worse; and that there might be equally good translations of the same foreign sentence, one turning it into a truth, the other into a falsehood; and that there might even be languages that cannot be translated into ours at all. If all human languages on some deep level share some universal but species-specific features, the languages of intelligent extraterrestrials might be of this kind. The application of the truth predicate to foreign sentences we perhaps cannot translate determinately or at all would be a theoretical extrapolation underdetermined by evidence or even by all nonlinguistic facts, according to Quine, and he does not operate with it.

  Quine's account only implies that if X says what we would say by saying “_______,” then what X says is true iff______. If X says “_ _ _ _ _ ” and there is no “________ ” such that what X says by saying the former is just what we would say by saying the latter, then Quine's account does not imply anything about the conditions under which what X says is true. The view that the truth predicate with which ordinary English speakers actually operate is a theoretical extrapolation, to untranslatable utterances of others, of what started out as an immanent disquotation operator would no longer be deflationist in Quine's sense. Such a view, however, would still remain light-years away from any traditional rival views, and might be called quasi-deflationist.

  In addition to this much-discussed aspect of Quine's view, there is a problem of formulation less often noted: How exactly are we to understand the “equivalence” between saying something and saying that it is true? Quine is much less explicit than Ramsey on this point. To claim interchangeability in absolutely all contexts would be to claim too much. For apart from the matter of stylistics, there is a problem about examples such as the following:

  (20a) Fido believes that his master is at the door.

  (20b) Fido believes that it is true that his master is at the door.

  These do not seem equivalent. For surely a dog, though it may have the belief ascribed in (20a), does not have the concepts needed to have the belief ascribed in (20b), any more than (to borrow Wittgenstein's example) it has those needed to believe that his master will come the day after tomorrow.

  Quine may hold that all T-biconditionals are true, and that all T-introductions and T-eliminations are truth-preserving, but claiming as much will not do as an explanation of equivalence. For the procedure of explaining truth in terms of a notion of “equivalence” itself explained in terms of truth or truth-preservation is unacceptably circular: An explanation of what truth amounts to had better not itself use the T-word. (Besides, if all one knows about X is that all biconditionals of the form “‘________ ’ is X iff ________” are X, then for all one knows, X might be truth or might be falsehood.)

  3.4 OTHER MODERATE THEORIES

  The most-discussed version of deflationism in recent years has been that advocated by Paul Horwich under the label minimalism. Horwich's view differs from Quine's by being propositionalist, and by offering a definite account of “equivalence.” According to minimalism, for one to understand the truth predicate is for one to have the disposition to accept any T-biconditional proposition, or more precisely, anything that one recognizes as such. Against a background of classical logic, this is more or less the same as having the disposition to infer the conclusion proposition from the premise proposition in any T-introduction or T-elimination, or more precisely, anything one recognizes as such.

  The requirement of recognition is imposed in an attempt to get around the problem of mistakes that we correct when they are pointed out to us, and especially the problem of instances no one is disposed to accept because they are simply too long and complicated to take in. (There remain further objections, beyond these problems, to any identification of knowledge of meaning with a disposition, but the matter is too deep and elusive to be gone into here.)

  Horwich emphasizes that his theory does not leave understanding the notion of truth dependent on understanding some A-notion, acceptance or assertion, or some I-notion, inference or implication. To be sure, one can only think a biconditional is acceptable if one has the concept of acceptance, but one can accept a biconditional without having the thought that it is acceptable. To suppose otherwise, that before accepting something one must accept that it is acceptable, involves an infinite regress.

  This claimed advantage of freedom from entanglements comes at a cost. Minimalism is open to a version of Tarski's objection to the procedure of simply taking T-biconditionals as axioms, namely, that such a conception of truth leaves us unable to deduce certain general laws. Consider these successively more general formulations:

  (21a) If it is true that the tomato on your plate is ripe and juicy, then it is true that the tomato on your plate is ripe.

  (21b) For any tomato, if it is true that it is ripe and juicy, then it is true that it is ripe.

  (21c) For any conjunction, if it is true, then its first conjunct is true.

  The natural ways to argue for these would be as follows:

  (22a) Suppose it is true that the tomato on your plate is ripe and juicy. It follows (by T-elimination) that it is ripe and juicy, hence (by the logic of “and”) that it is ripe, hence (by T-introduction) that it is true that it is ripe.

  (22b) Consider any tomato, call it Tom, and suppose it is true that it is ripe and juicy. From its being true that Tom is ripe and juicy, it follows that Tom is ripe and juicy, hence that Tom is ripe, hence that it is true that Tom is ripe, that is, that the given tomato is ripe.

  (22c) Consider any conjunction, say the conjunction that things are thus and things are so, and suppose it is true.

  From its being true that things are thus and things are so, it follows that things are thus and things are so, hence that things are thus, hence that it is true that things are thus, that is, that the first conjunct is true.

  Here (22a) requires only acceptance of T-biconditionals for specific sentences, expressing specific propositions. Not so with (22bc). For “Tom” does not really denote any specific tomato, and “Tom is ripe” does not really express any specific proposition. That goes double for “things are thus,” which is really only a placeholder for a proposition-expressing sentence. It is not plausible to claim (21b) or (21c) is arrived at inductively from examples like (21a); but arriving at either deductively requires “schematic reasoning,” as the type of reasoning in (22bc) is called. What seems needed is a disposition to infer a conclusion from a premise by T-introduction
and T-elimination (or, what is more or less equivalent, to accept a T-biconditional) even in the case of placeholders.

  That and more also would presumably follow if understanding the truth predicate were identified, not with a disposition to accept each T-biconditional, but with a recognition that all T-biconditionals are acceptable or assertable, or that the conclusion is inferable from or implied by the premise in all T-introductions and T-eliminations. Such an identification makes acquisition of the T-notion depend on prior possession of an A-notion or I-notion, and so has the disadvantage of leaving one unable to explain A-notions or I-notions in terms of the T-notion, lest a chicken-and-egg problem arise: One cannot explain assertion as utterance purporting to be true, or explain implication in terms of truth-preservation.

  If one is comfortable with talk of propositions and properties, the view being contemplated can be formulated as saying that truth is the property of propositions for which its ascription to a given proposition and the proposition itself always imply each other, or the property for whose possession by a given proposition the proposition itself gives the necessary and sufficient condition. Such is Colin McGinn's self-effacement theory. It is not always recognized as a variety of deflationism. Thus there is at present not only no consensus among deflationists as to what is the optimal formulation of deflationism, but not even any consensus as to which views count as genuinely deflationist. Nonetheless, the idea that something fitting our initial characterization of deflationism must be right remains widespread.

  3.5 SLOGANEERING

  In recent years this idea has often been accompanied by rhetoric of a kind one does not find in Ramsey or Quine—it may derive ultimately from the pragmatist philosopher or antiphilosopher Richard Rorty—involving such slogans as “Truth is not a property” or “Truth is not a substantive property.” The thought perhaps is something like the following. The equivalence principle suffices to explain the usage and usefulness of “true” (notably, for avoiding repetition and for formulating generalizations). We do not have to posit that the use of “true” is to report the detection of the presence of some interesting property all truths share. This being so, it would be an improbable coincidence if there were any such interesting property.

  Despite the recent popularity of such slogans, however, there is no hope that any slogan of the kind could provide a characterization of deflationism acceptable to all deflationists. The reason why not is as follows. Colloquially we may always expand “so-and-so does such-and-such” to “so-and-so has the property of doing such-and-such.” Such linguistic padding is called pleonasm, and some philosophers consider this colloquial, pleonastic notion of property the only legitimate one. Others, however, think the only legitimate notion of property is some more substantive one, on which perhaps a natural predicate like “is a human being” would pick out a property, while an artificial one like “is either female or else both over six feet tall and also left-handed” would not. Yet others take “property” to be ambiguous between a pleonastic and a substantive sense, both legitimate. Yet others are uncomfortable with any talk of properties. These divisions exist among deflationists as much as among other philosophers, and cut across the distinction between moderates who do and radicals who do not grant that “is true” is a predicate.

  Now a radical who admitted only the pleonastic notion of property might agree on the slogan “Truth is not a property,” as might a moderate who admitted only the substantive notion. But a radical who admitted only the substantive notion would find the slogan an understatement, since merely to say that “is true” does not indicate a substantive property is not to say that it is not a predicate. A moderate who admitted only the pleonastic notion would find the slogan plain wrong, since if “is true” is a predicate, it ipso facto indicates a pleonastic property. Those with other combinations of views might find the slogan ambiguous or meaningless. Similar problems plague any slogan formulated in terms of properties.

  We think our third thesis, that there is nothing to be said about what it is to be true, once one has said what it means to call something true, does a little better. The thought is that whether or not an account of meaning leaves room for a further question about being depends on the nature of the account of meaning. For example, if the right definition of “hot” is something like “having whatever physical property it is that causes heat sensations,” then even if one fully understands what it is to be a physical property or a cause or a heat sensation, still there would be room for a question about what heat is, namely, the question “Which physical property is it that causes heat sensations? The presence of caloric fluid? Rapid random motion of molecules? Or what?” But if the right definition of “hot” is something like “having a temperature in °F significantly higher than 150,” then if one fully understands what it is for an object to have a certain temperature in °F, and what it is for a number to be significantly higher than 150, then there seems to be no room for any further question about what heat is. For deflationists, the equivalence principle tells the whole story about the meaning of “true,” and that story, though it is not strictly speaking a definition, seems to be much more like the second than the first candidate definition of “hot,” and to be not at all the sort of account of meaning that leaves room for a further question about being.

  Another slogan sometimes met with as a capsule summary of deflationism is “There are no substantive questions about truth.” This may easily be made to seem absurd. For suppose a Frenchman, Monsieur Orgon, emits certain sounds, and a bystander knowing his circumstances and his language tells us he has said something true. Then surely we may ask why, and surely our question is both substantive and about truth. The deflationist has an answer to this objection, but before stating it, a word about quotation marks is called for.

  Inverted commas may be used to designate at least eight different kinds of things: phonological or orthographic tokens or types, or semantic tokens or types from our own or a foreign language. In writing about Tarski we followed his usage, taking quotation to designate an orthographic type. In stating the deflationist answer to the objection about Monsieur Orgon, a different usage will be in order. In discussing this example it will be convenient to designate phonological entities by using italics and approximate phonetic spelling, and to use ordinary English-style quotation marks to designate expressions of English with their ordinary English meanings, and French-style quotations marks to designate expressions of French with their ordinary French meanings.

  The deflationist account of the supposed example of a substantive question about truth is then that the answer to the question comes in three parts: (i) by emitting sounds very roughly representable as mah mehr eh fahshay, Monsieur Orgon is uttering the French sentence «Ma mère est fachée», which translates as “My mother is angry,” which transposes to “His mother is angry”; (ii) “His mother is angry” is true iff his mother is angry; (iii) his mother is angry. Deflationists claim that (i) and (iii) are substantive but about meaning and emotion, respectively, not truth, while (ii) is about truth, but is a triviality. The original question thus can be substantive and about truth without, so to speak, being substantive about truth, much as (to borrow Plato's example) a dog can be yours, and a father, without being your father.

  3.6 REFERENCE

  Most deflationists about truth are likewise deflationists about other alethic notions. The most important of these is probably the one we called denotation in our discussion of Tarski, and that is elsewhere called reference. (We note that, though we and many others use these terms interchangeably, they are sometimes distinguished. Some may use denoting for what certain expressions do, and referring for what people who use those expressions thereby do. Others, holding that the relation of name to named and the relation of description to described are importantly different, may reserve referring for the former, while allowing denoting to cover both.) The deflationist view is that, call the relation what you will, there are no substantive questions about it, any more
than about truth.

  Why does Monsieur Orgon, by emitting the sounds mah mehr, refer to Madame Prunelle? Well, (i) by emitting the sounds he utters «ma mère», which translates as “my mother,” which transposes to “his mother”; (ii) “his mother” refers to an individual iff that individual is his mother; and (iii) Madame Prunelle is his mother. Everything substantive belongs to (i) semantics or (iii) genealogy; only a triviality belongs to (ii) the theory of reference.

  More specifically, prosententialists about truth tend to be pronomialists about reference, minimalists about truth tend to be minimalists about reference, disquotationalists about truth tend to be disquotationalists about reference, and so on. The referential analogue of (17a) reads as follows:

  (23a) “______” refers to an individual iff he/she/it is _______.

  Curiously, a referential analogue of (17b), undoing indirect rather than direct quotation, is harder to formulate. Some write

  (23b) An individual is the one referred to as ______ iff he/she/it is ______.

  but this is not entirely satisfying, since in ordinary language the form of words “referred to as” seems most often used to distance the speaker from a certain way of referring to an item, as if short for “referred to by some other people as.” For instance, a political reporter who says “The person referred to as The Dear Leader is Kim Jong-Il” probably does not herself regularly refer to the ruler of North Korea as The Dear Leader, unless perhaps ironically.