Truth (Princeton Foundations of Contemporary Philosophy) Read online

  But deflationists claim they can provide an explanation of the utility of truth, after all. Consider first a special case. Imagine an agent with a choice among several courses of action that may lead to various possible outcomes among which the agent has preferences. Call beliefs of the form

  (3) Acting thus-and-so will lead to the outcome most preferred.

  directly action-guiding. Counting a belief as useful to an agent just in case it is useful for obtaining the agent's preferred outcome, and supposing that the agent does not exhibit weakness of the will, so that if the agent believes (3) then the agent will act thus-and-so, it follows tautologically that it will be useful for the agent to believe (3) just in case acting thus-and-so will lead to the outcome most preferred—or in other words, just in case (3) is true, as deflationists understand truth.

  Thus it is explained why having true beliefs is useful, and similarly having false ones is harmful, in the special case when the beliefs are directly action-guiding. And now for the general case. Since directly action-guiding beliefs will generally be inferred as conclusions in some manner from other beliefs taken as premises, it will be indirectly useful if the other beliefs involved are true and the manner of inferring conclusions from premises involved is truth-preserving. This explanation leaves room for exceptions, as any explanation should, since the utility of true beliefs is a general tendency, not an exceptionless law.

  True beliefs can be useless if they do not serve as premises for inferring action-guiding conclusions. True beliefs about trivia of antiquarian scholarship or popular culture may be useless except to contestants on quiz shows. False beliefs can be useful if, taken as premises, they yield true action-guiding conclusions. A false belief that there is a lion behind the door may be as useful as a true belief that there is a tiger in persuading one to leave the door closed. Successful applications of scientific theories now superseded provide other examples. And there are other types of exceptions. (True beliefs can be harmful if, taken as premises together with some false beliefs, they lead to false action-guiding conclusions that would not have been inferred from the false premises alone. Weakness of the will can render true beliefs useless and false ones harmless.) What now is there left to explain about the general tendency of true belief to be useful, for which a physicalist or other inflationary theory of truth would be needed?

  Well, if there is nothing more to be said about the general tendency, there still seems to be more to say about particular cases. The issue may be easier to illustrate in the case not of truth for a particular part of language, but of correctness for a particular kind of nonverbal representation. Parallel to the deflationist account of truth, “Things are as they are said to be,” there is a deflationist account of correctness, “Things are as they are represented as being.” For some particular kinds of representation, however, there seems to something more to be said about what correctness is than just this, and this “something more” may play an important role in explaining the utility of the representation.

  Consider, for instance, subway diagrams. When such a diagram is correct, a certain physical relation holds between it and the subway system, which may be roughly described as follows: For every dot on the diagram there is a unique station in the system whose walls bear a label of the same shape as (but of a much larger size than) the label beside the dot in the diagram; and for every station there is a unique dot thus related to it; and two dots are connected by a line on the diagram iff the stations thus related to them are connected by a subway track. This physical relation can be used to explain why a rider who uses the diagram can successfully navigate the subway system.

  To this the deflationist replies that of course if the diagram is correct, in the deflationist sense that things are as it represents them as being, then this physical relation will hold, for that is how the diagram represents things as being, as anyone knows who knows how to read a subway diagram. With road maps, which unlike subway diagrams represent geometrical features, distances and directions, and not just topological features, order and connections, there will also be a physical relation, but it will be a different one. With celestial globes, which represent the relative directions of stars but not their distances from the earth, there will be yet another relation. In each case, one who understands how the representation is representing things as being will understand that the appropriate relation will hold if the representation is correct in the deflationist sense that things are as it represents them.

  What the physicalist seems to want is a physical, or physiological, account of the rider's or motorist's understanding of subway diagrams or roadmaps. And there may be some scientific, or scientistic, sense of “explanation” in which nothing less than this could provide a genuine explanation of the rider's success in navigating the system relying on the diagram, or the motorist's success in navigating the countryside relying on the map. But the deflationist will insist that the account that is being called for is an account of meaning, in a broad sense applicable to nonverbal as well as linguistic representations, and not of correctness, about which the formula “Things are as they are represented as being” says all there is to be said.

  5.6 NORMATIVITY

  Deflationists often describe truth as a “logical” or “quasilogical” notion. Realists take it to be a metaphysical or physical one. A common feature of logical, metaphysical, and physical notions alike is that they are neutral or value-free. To the extent that truth is useful one may expect it will be valued, but for deflationism and realism alike, “true” is not itself an evaluative term. One common criticism of deflationism and realism, going back to Michael Dummett, the founder of so-called antirealism, is precisely that “true” is an evaluative or “normative” term.

  An assertion that is untrue is open to criticism on that account, or in jargon

  (4) Truth is a norm of assertion.

  According to a favorite analogy of Dummett, truth is to making an assertion what winning is to playing a game. The deflationist account of truth is analogous to the following analysis of winning:

  (5) Winning a game consists in whatever the rules defining the game say constitutes winning.

  This may get the extension of “win” right, but without the principle

  (6) Winning is the aim of playing.

  the point of the win/lose distinction is unaccounted for. Analogously, Dummett claims, without the principle (4) the point of the truth/false distinction is unaccounted for. And any theory, deflationist or other, that leaves the point of the distinction out must be considered inadequate.

  One line of response grants the critic that truth is a norm of assertion, but insists that something's being taken as a norm by some practice is an extrinsic fact about it that does not suffice to make it intrinsically normative. Being 94 feet long remains a neutral, value-free notion even though it is taken as the norm for courts in professional basketball. For truth to be intrinsically normative, (4) must not merely hold but do so as part of the very meaning of the truth predicate.

  But perhaps (4) is better regarded as part of the meaning not of “true” but of “assert,” distinguishing assertion from uttering during an elocution competition, a report of another's speech, a storytelling session, or whatever. If someone is uttering declarative sentences without even pretending to be aiming at uttering true ones, must she be making assertions but flouting a norm of that practice, or may she not be doing something other than making assertions, perhaps narrating fiction? Likewise it is not obvious that (6) must be regarded as part of the meaning of “win” rather than of “play.” If someone is moving checkers on a checkerboard otherwise in accordance with the rules of checkers, but making no effort to capture all the opponent's pieces, must he be playing checkers but violating a norm of that practice, or may he not be doing something other than playing checkers, perhaps demonstrating how the pieces move, or reproducing the moves of some famous game, or playing giveaway checkers?

  Such a response should be available to most ve
rsions of deflationism or realism. It will presumably not, however, be available to a version of deflationism that makes the notion of truth depend on that of assertion (by identifying understanding of the truth predicate with recognition that all T-biconditionals are assertable). But perhaps (4) can be regarded as part of the meaning neither of “true” nor of “assert,” even if by the time both words are learned a child will have learned that there is such a rule as

  (7) Don't assert what isn't true.

  and even though (4) is no more than a jargon-ridden way of saying that there is such a rule.

  How could this come about? A hypothetical scenario might run as follows. Soon after its first words a child begins to hear prohibitions of the form “Don't say…,” and not much later the child learns there are exceptions, thus:

  (8) Don't say “The duck is in the muck” if the duck is not in the muck, or “The goose is on the loose” if the goose is not on the loose, or anything like that.

  (9) But saying when reciting a rhyme, or making believe, or telling what someone else said, or something like that, doesn't count.

  It is obvious why there should be such a rule as (8): We want to be able to infer from the child's saying “Things are thus” that things are thus. It is obvious also why there should be such exceptions as (9). (There may be none for the rule “Don't say the word Daddy said when he hit his thumb with the hammer.”)

  The child later on learns to use “true” as a device for making generalizations that otherwise could only be suggested by a few examples plus “or anything like that,” as in (8). Slightly earlier or later, the child learns to use “asserting” for “saying” minus the open-ended list of kinds of exceptions mentioned in (9) (and “uttering” for “saying” without exception). Once both words have been learned, the child will be able state the rule in the grown-up way (7), but the rule need not be considered part of the meaning of the vocabulary needed to state it. Such a response may have at least some initial plausibility, but other objections remain to be considered, arising from the theory of meaning.

  CHAPTER SIX

  Antirealism

  AS MENTIONED AT THE BEGINNING of the preceding chapter, in traditional philosophy there were several debates pitting a group called “realists” against a group called something else—a different something else in each debate. The realists were those who maintained the “real” or “objective” or “independent” existence of something—a different something in each debate. Early in the last century there was a debate between two groups of mathematicians: classical mathematicians (now the immense majority) and mathematical intuitionists (now a tiny minority). The former accepted certain existence proofs that the latter rejected, and so the debate was usually perceived as being about the existence and nature of mathematical objects, and sometimes perceived as a revival or reincarnation of a traditional debate between “realists” and “conceptualists” over the existence and nature of universals.

  Inspection shows that the mathematicians' differences over which existence proofs are acceptable derive from more fundamental differences over which logical inferences are acceptable. Classical mathematicians accepted and mathematical intuitionists rejected the principle of excluded middle, according to which assertion of A or not A is always warranted. Such differences themselves derive from more fundamental differences over the meanings of “not” and “or” and other logical vocabulary.

  For several decades now Michael Dummett has been urging that such differences over the meanings of specific words should be seen as derivative from a more fundamental difference over the general form an account of meaning should take. He has urged that not just the mathematical debate, but a whole range of debates over various forms of realism, should be reconfigured as a debate over the proper form for an account of meaning. The realist side would be advocates of an approach, to be described shortly, called truth-conditional semantics, while the antirealist side, which Dummett himself generally favors, would advocate a rival approach called verification-conditional semantics.

  When we spoke of “realism” in the preceding chapter, we meant a view directly about truth, to the effect that it consists in an appropriate relation to some portion(s) or aspect(s) of reality. By contrast, when Dummett speaks of “realism” he means a view about meaning (and its relation to truth) that he thinks those traditionally called “realists” about various other matters (beginning, but only beginning, with mathematics) are ultimately committed to, whether or not they acknowledge as much (as not all do). Dummettian antirealism is an influential view about truth (and its relation to meaning), but what it is anti, namely, truth-conditional semantics, is not what we called “realism” about truth in the preceding chapter. (Dummett opposes that, too, but it isn't his main concern.)

  In this chapter, therefore, we will generally ignore the kinds of views discussed in the preceding one, to concentrate on the relations among the trio of views, truth-conditional semantics and verification-conditional semantics and deflationism (fairly often, however, remarking that something we are saying about deflationism applies much more widely). It will be a tangled tale. The concerns and considerations that motivate the three views are so different that the trio more often seem to be addressing different questions than taking differing positions on a single issue. At the end of the chapter we will bring in a fourth, compromise view, called pluralism.

  6.1 MEANING AND TRUTH

  What does meaning have to do with truth? Even deflationists want to say (as we saw in §5.1) that in some sense the state of the world “makes” the proposition that snow is white true. Does it also make the sentence “Snow is white” true? Only so long as we think of the sentence as being not just certain sounds or shapes, but certain sounds or shapes with a certain meaning. This thought may be put by saying that the meaning of a sentence together with the state of the world determine whether the sentence is true. In a slogan:

  (1) Meaning and world together determine truth value.

  Natural though it may seem, the naive formulation (1) calls for critical scrutiny. For any complete account of “the state of the world” or “how things are” would have to include an account of how things are as regards the meanings of sentences, and for that matter, their truth values. The sentence meaning what it does is part of the state of the world, and separate mention of it is otiose, unless we restrict the scope of “the state of the world” to pertinent aspects. Should we, then, understand “the state of the world” to mean “the state of the world as regards nonlinguistic matters”? That will not do, because some sentences are about linguistic matters, and the state of the world as regards linguistic matters will be pertinent to their truth. Rather it is the meaning of the sentence that determines which aspects of the state of the world (usually nonlinguistic aspects, sometimes linguistic aspects) are pertinent. The meaning of the sentence determines what conditions the state of the world must fulfill in order for the sentence to count as true. In a slogan:

  (2) Meaning determines truth conditions: A sentence's truth conditions are consequences of its meaning.

  But there are objections to this slogan, too. For one, where indexicality is present, the meaning of the sentence type is not enough to determine under what conditions a sentence token counts as true; context must provide something more. For another, according to the chain of communication picture (sketched in §5.4), “tigers” may refer to tigers only owing to something “outside the head,” the history of the usage of the term, even though “inside the head” there may be no more meaning attached to the term than a broad sortal classification as “a kind of animal.” If so, then the meaning of “Tigers are striped” will in a sense be insufficient to determine its truth conditions (since from a complete description of the world's fauna, but identified only by their scientific names, plus the information that “tiger” is the common name for an animal species, one could not determine whether tigers are striped).

  We mention objections of these kinds, however, only to ask
the reader to set them aside and not be distracted by them. Our interest here will be in dissent from (2) in a different direction, represented by the counterslogan:

  (3) A sentence's truth conditions constitute, or at least are constituents of, its meaning.

  Whereas (2) suggests meaning comes first, and determines truth conditions, (3) suggests that truth conditions are already a part of meaning, or perhaps even the whole of it. We call (3) the principle of truth-conditional semantics, and its apparent corollary

  (4) Knowledge of meaning consists, at least in part, in knowledge of truth conditions.

  the principle of hardcore truth-conditional semantics.

  By suggesting that any question that involves meaning ipso facto involves truth, (3) and (4) threaten the deflationist slogan “There are no substantive questions about truth,” since (as we saw in §3.5) the defense of that slogan involves claiming that various undeniably substantive questions about meaning are not about truth. More directly and seriously threatening to deflationism, and to many other theories of truth besides, are these further apparent corollaries: