(p.228) Appendix: Charity, Interpretation, and Truth
(p.228) Appendix: Charity, Interpretation, and Truth
Various philosophers, including W. V. Quine (1960), Donald Davidson (1983, 1991), and Daniel Dennett (1978, chap. 1; 1987, chap. 10), have endorsed a “principle of charity” implying either that we must interpret beliefs as being largely correct or, more strikingly, that one's beliefs must actually be largely correct. Given a defensible principle of charity, we could block familiar skeptical worries recommending that we withhold judgment on whether one's beliefs are largely correct. We could then use considerations of charity to show that such skeptical worries are misplaced. What exactly would this antiskeptical argument look like, and is it actually sound? This appendix offers an answer in accord with this book's conditional ontological agnosticism.
1. Twofold Charity
Quine proposes a “maxim of translation” entailing that “assertions startingly false on the face of them are likely to turn on hidden differences of language” (1960, p. 59). He seeks, in this connection, a method of translation specifying how we are to interpret a speaker's logical constants, among other utterances or inscriptions. The rough idea driving Quine's charitable method of translation is that it is useful for translators to translate a speaker's language into their own language in such a way that the speaker's affirmations are largely true relative to the translators' own standards for truth. This pragmatic consideration enables translators to select from the range of empirically adequate alternative translations. Physical facts, according to Quine's thesis of the indeterminacy of translation, do not determine the correctness of either of two conflicting empirically adequate manuals of translation (1981, p. 23; 1990a, pp. 47–48); nevertheless, a principle of charity can assist a choice between such manuals.
Davidson denies that his principle of charity is just a useful device for handling translational indeterminacy. He contends that it not only serves as a requirement on all meaningful interpretation bat also entails that our beliefs (p.229) must actually be largely correct. Davidson's position on the principle of charity is, then, twofold:
Claim (i) amounts to the view that we must interpret others to have beliefs largely correct by our own lights, whereas claim (ii) implies that anyone's set of beliefs is actually largely true.
(i) The general policy … is to choose truth conditions that do as well as possible in making speakers hold sentences true when (according to the theory [of truth and meaning] and the theory builder's view of the facts) those sentences are true. (1974c, p. 152)
(ii) … beliefs are by nature generally true (1983, p. 437); [that is,] it must generally be the case that a sentence is true when a speaker holds it to be. (1975, p. 169)
Claim (i) rests on the following considerations offered by Davidson:
Fallibilism about reasonableness implies that contingent reasonable beliefs can be false, and thus prompts the question whether our finding a great deal of reasonableness in others requires our finding a great deal of truth in them. Even if our understanding others requires our finding their beliefs largely reasonable (at least on the basis of consistency requirements), it evidently is still an open question whether we must also find their beliefs largely correct.
Making sense of the utterances and behaviour of others, even their most aberrant behaviour, requires us to find a great deal of reason and truth in them. To see too much unreason on the part of others is simply to undermine our ability to understand what it is they are so unreasonable about. (1974c, p. 153)
Davidson links understanding and interpreted correctness as follows:
Davidson's point is not that interpretation must eliminate disagreement; it is rather that mistake and meaningful disagreement require considerable agreement: for example, widespread agreement in sentences held true, or in beliefs.
Since charity is not an option, but a condition of having a workable theory [of meaning], it is meaningless to suggest that we might fall into massive error by endorsing it. Until we have successfully established a systematic correlation of sentences held true with sentences held true, there are no mistakes to make. Charity is forced on us; whether we like it or not, if we want to understand others, we must count them right in most matters. … We make maximum sense of the words and thoughts of others when we interpret in a way that optimizes agreement. (1974a, p. 197)
Davidson's position on charity rests on his account of the relation between belief and meaning. “What a sentence means,” according to Davidson, “depends partly on the external circumstances that cause it to win some degree of conviction; and partly on the relations … that the sentence has to other sentences held true. … ” (1983, p. 432). Since the latter relations directly involve beliefs, on Davidson's view, meaning depends on belief. In addition, belief depends on meaning, since “the only access to the fine structure and individuation of beliefs is through the sentences speakers and interpreters of speakers (p.230) use to express and describe beliefs” (1983, p. 432). In short, determinate belief depends on meaningful sentences (see Davidson 1975, p. 170; cf. Moser 1983c).
Davidson infers that to account for the nature of meaning or belief, we must rely on something other than meaning or belief. Following Quine (1960), Davidson uses as his basic explainer “prompted assent”: that is, the causal relation between assenting to a sentence and the cause of this assenting (1983, p. 432). Assenting to sentences, on Davidson's view, is just holding them true. Roughly, the principle of charity recommends that when we as interpreters find a speaker assenting regularly to a sentence under conditions we recognize, we regard those conditions as the truth conditions of the speaker's sentence. Davidson proposes that interpretation derives from a Tarski-style rendition of truth for a speaker's language:
where ‘5’ stands for a description of a sentence of English, and ‘P’ stands for a sentence giving the conditions under which S is true. Indexical and demonstrative components of sentences require the relativity to a speaker and a time.
Sentence S is true (as English) for speaker U at time t if and only if P (1969, p. 45; cf. 1973, pp. 68–9),
Our only available method of interpretation, according to Davidson (1983, p. 434), puts a speaker's beliefs in agreement with our standards of logic, our logical truths. “The point of the principle of charity,” Davidson claims (1983, p. 433), “is to make the speaker intelligible, since too great deviations from consistency and correctness leave no common ground on which to judge either conformity or difference.” Consistency and correctness offer stable ground for identifying beliefs; omitting that ground, according to Davidson, leaves no basis for ascribing beliefs. Davidson thus denies that we can discover a speaker to have largely false beliefs about the world, on the ground that we must interpret “sentences held true (which is not to be distinguished from attributing beliefs) according to the events and objects in the outside world that cause the sentences to be held true” (1983, p. 435).
Davidson seeks support for his twofold principle of charity in the importance of an empirical theory of meaning. He holds that “if a semantic theory claims to apply, however schematically, to a natural language, then it must be empirical in character, and open to test” (1973, p. 73). If we want an empirical theory of meaning, Davidson contends, we need a principle of charity. An empirical semantic theory, on this line of argument, makes the meaning of sentences held true depend on the observable circumstances under which those sentences are uttered. Davidson holds, accordingly: “The causal relations between the world and our beliefs are crucial to meaning not because they supply a special sort of evidence for the speaker who holds the beliefs, but because they are often apparent to others and so form the basis for communication” (1990c, p. 76). Davidson holds, more specifically, that “linguistic communication, the indispensable instrument of fine-grained interpersonal understanding, (p.231) rests on mutually understood utterances, the contents of which are finally fixed by the patterns and causes of sentences held true" (1990a, p. 326). This is a kind of communitarian causal holism about prepositional content.”
Davidson claims that “without communication propositional thought is impossible,” and that “until a base line has been established by communication with someone else, there is no point in saying a person's thoughts or words have a propositional content” (1991, p. 160). Davidson holds that this “follows at once if we suppose that language is essential to thought, and we agree with Wittgenstein that there cannot be a private language” (1991, p. 157). Davidson maintains that only communication can supply the objective check needed to distinguish between correct and incorrect language use. Interpersonal sharing of verbal reactions to common sensory stimuli, according to Davidson, is crucial to locating the cause of a thought and defining its contents. Two people can observe each other's verbal reactions, and thereby each can correlate the other's reactions to his or her own sensory stimuli with an eye toward identifying a common cause. This commonly identified cause can serve as a basis for objective correctness in belief. Apart from standards set by intersubjectively shared verbal reactions, according to Davidson, talk of objective truth and of determinate belief-contents makes no sense.
We now have the rationale for Davidson's twofold principle of charity before us. Can the principle survive scrutiny?
2. Misplaced Charity?
Let us distinguish between the following:
(a) We must interpret sentences held true according to the actual events and objects in the outside world that cause those sentences to be held true (cf. Davidson 1983, p. 435).
(b) We must interpret sentences held true according to what we believe to be the events and objects in the outside world that cause those sentences to be held true.
Thesis (a) entails that we must interpret sentences affirming the existence of external physical objects (for example, chairs, tables, and books) as being about the actual external physical objects that cause those sentences to be believed. This thesis rests on Davidson's assumption that “the causality [actually responsible for the holding of beliefs] plays an indispensable role in determining the content of what we say and believe” (1983, p. 435). Thesis (b) is more modest; its key implication is that we must interpret sentences affirming the existence of external physical objects as being about what we believe to be the (p.232) objects that cause those sentences to be believed. This thesis presupposes that ascribed content is determined by what we take to be causally responsible for a belief, whereas (a) depends on the view that ascribed content is determined by the actual causal basis of a belief.
The key issues now are these: Why must we interpret one's beliefs according to their actual causes or even according to what we believe to be their causes? In addition, why must one's beliefs be largely true in virtue of being about what actually causes them? Until we have answers to these questions, we lack support for Davidson's twofold principle of charity. We can now begin to assess that principle.
Must we interpret a person's beliefs according to their actual causes? The following scenario seems coherent; at least, Davidson has not shown it to be incoherent. Smith's beliefs regarding the external world are actually caused by discrete physical objects (for example, chairs, tables, and books), and we somehow come to know this to be so. Smith himself, however, believes that there are no discrete physical objects, and holds that there actually is, in accord with (Spinozistic) unrestricted neutral monism, only a single nonmental nonphysical thing that at times merely appears to be discrete physical objects. Those fundamental ontological commitments determine the content of virtually all of Smith's beliefs, including all his beliefs about the objects of experience; and we interpret Smith's utterances and inscriptions accordingly. We could, nonetheless, come to know that those fundamental ontological commitments are incorrect (say, on the basis of truths learned from our best experimental science), and, therefore, that Smith's beliefs are largely incorrect. We could grant that Smith's beliefs are largely consistent, but this would not threaten, in any way, our knowledge that his beliefs are largely incorrect. Massive consistency and massive error sit well together. In this case, contrary to Davidson's thesis (a), we do not interpret a person's beliefs according to their actual causes.
What about thesis (b)? We believe (and know), according to the example at hand, that Smith's beliefs regarding the external world are caused by discrete physical objects, but we do not interpret Smith's beliefs according to causation by such physical objects. Instead, we interpret Smith's beliefs as affirming, and thus committing him to, unrestricted neutral monism: the Spinozistic view that there is actually only a single nonmental nonphysical thing that at times merely appears to be discrete physical objects. Our best explanation of Smith's verbal behavior is that he does not share our true beliefs about what actually causes his beliefs. Consequently, we may reasonably interpret his beliefs as differing from what we believe (and, by hypothesis, know) to be their actual causes. Thesis (b), then, suffers the same unhappy fate as (a).
One might reply that my example actually supports (a) and (b). If we know that discrete physical objects cause Smith's beliefs about the external world, we shall interpret his beliefs as actually being about such objects, even if Smith himself claims otherwise. On this reply, what a belief is “actually about” is determined by what actually causes that belief; and since, by hypothesis, discrete physical objects actually cause Smith's beliefs, those beliefs are actually about such objects. Knowing what actually causes Smith's beliefs, we shall (p.233) interpret those beliefs accordingly, with respect to what they are actually about. We do not have here, then, a problem for theses (a) and (b), for my example has us knowing that Smith's beliefs are actually caused by discrete physical objects.
The anticipated reply assumes that what actually causes a belief determines what that belief is about. If what a belief is about is determined by what its constituents refer to, and if what a beliefs constituents refer to is determined by what causes them, then what a belief is about is determined by its cause. This causal approach to what a belief is about will be acceptable so long as we define ‘about’ and ‘refer’ in strictly causal terms. If we understand talk of the “meaning” and the “content” of a belief via a strictly causal approach to aboutness and reference, we may say that a beliefs meaning, or content, issues from what causes that belief. On this view, meaning is reference, and reference arises just from causal relations between (constituents of) beliefs and their causes.
Davidson cannot plausibly adopt the reply at hand, since his own view of meaning is not strictly causal in the way the reply assumes. Following Grice (1957), Davidson acknowledges a crucial role for intention in meaning and communication, as follows:
Davidson's view of the role of intention in meaning conflicts with any strictly causal view implying that a beliefs meaning depends simply on what causes one to hold that belief. Communication-based intended truth-conditions determine meaning, on Davidson's view; a beliefs content—what it is “about”— depends specifically on how the pertinent believer intends the belief (or accepted sentence) to be interpreted. The anticipated reply is, then, unavailable to Davidson.
An utterance has certain truth conditions only if the speaker intends it to be interpreted as having those truth conditions. … A malapropism or slip of the tongue, if it means anything, means what its promulgator intends it to mean. … What matters to successful linguistic communication is the intention of the speaker to be interpreted in a certain way, on the one hand, and the actual interpretation of the speaker's words along the intended lines through the interpreter's recognition of the speaker's intentions, on the other. (1990a, pp. 310–11)
Davidson's account of meaning, in terms of intended truth-conditions, allows for cases where the actual cause of one's belief differs from how one intends the belief to be interpreted. Smith, for instance, intends his beliefs regarding the external world to be interpreted in accord with unrestricted neutral monism; and we, as interpreters, can comply with his intention, even while knowing that discrete physical objects actually cause his beliefs. Smith's intended interpretation, in this case, runs afoul of the actual cause of his beliefs. Once we acknowledge that meaning depends on intended interpretation, as Davidson does, we cannot guarantee agreement between belief-contents and the actual causes of those beliefs or even between belief-contents and what we believe to be the actual causes of those beliefs. The example of Smith makes this clear.
(p.234) Having allowed intended interpretation to open a gap between belief-contents and their actual causes, we cannot close that gap in all cases of meaningful belief. One might, we have seen, seek refuge in a purely causal account of contents; but this move will depart not only from Davidson's own approach to meaning, but also from our ordinary notion of belief-contents involving what an individual actually understands.
The case of Smith counts against not only Davidson's requirements on interpretation, but also his requirement that one's beliefs must (in virtue of their “nature”) be generally true (1983, p. 437; 1975, p. 169). Smith's beliefs would not be generally true in the envisaged situation where discrete objects cause his beliefs but he holds an unrestricted neutral monism that bears on virtually all of his beliefs. We do not need any particular standard for individuating beliefs to make this point; we need only grant that neutral monism figures, directly or indirectly, in most of Smith's beliefs, and (following Davidson) that a theory of truth conditions for a speaker describes “how the speaker intends his utterances to be interpreted” (1990a, p. 312). Smith intends his utterances to be interpreted in a way that runs afoul of what actually causes his beliefs. Consequently, Smith's beliefs are, contrary to Davidson's principle of charity, not generally true. (This result is compatible with Davidson's view (1990a, p. 314) that “it is a mistake to look for … any … sort of explicit definition or reduction of the concept of truth.”)
Even if we, as interpreters, typically ascribe our own (semantic) standards of truth as being satisfied by speakers (including ourselves), we can ask this question: What reason have we to think that those standards always successfully identify the actual causes of one's beliefs? This question is highlighted by Davidson's observation that “we can't get outside our skins to find out what is causing the internal happening of which we are aware,” and “introducing intermediate steps or entities into the causal chain, like sensations or observations, serves only to make the epistemological problem more obvious” (1983, p. 429). The actual causes of one's beliefs can be misrepresented by the belief-contents resulting from how one intends one's accepted sentences to be interpreted; and this can hold for most of one's belief-contents. Witness the case of Smith, the unbridled Spinozist.
Davidson has invoked two considerations to block the uncharitable result at hand. First, he appeals to the possibility of an omniscient interpreter to argue that a speaker and an interpreter cannot understand one another on the basis of shared but generally false beliefs. The imagined omniscient interpreter, by hypothesis, knows everything about the world and about what does and would cause a speaker to assent to a sentence. Davidson explains how this possibility bears on fallible speakers and interpreters:
The omniscient interpreter, using the same method as the fallible interpreter, finds the fallible speaker largely consistent and correct. By his own standards, of course, but since these are objectively correct, the fallible speaker is seen to be largely correct and consistent by objective standards. We may also … let the omniscient interpreter turn his attention to the fallible interpreter (p.235) of the fallible speaker. It turns out that the fallible interpreter can be wrong about some things, but not in general. … (1983, p. 435; cf. Davidson 1977, p. 201)
This appeal to an omniscient interpreter does nothing to support Davidson's principle of charity. We now need independent support for Davidson's assumption that the omniscient interpreter would find fallible speakers largely correct; but Davidson has given no such support. So far as Davidson's argument goes, the omniscient interpreter could find fallible speakers largely incorrect. In particular, this interpreter, for all Davidson has shown, could find our beliefs largely incorrect, inasmuch as our presumed causes of our beliefs misrepresent virtually all our beliefs' actual causes.
Davidson's second strategy contends that “global confusion, like universal mistake, is unthinkable, not because imagination boggles, but because too much confusion leaves nothing to be confused about and massive error erodes the background of true belief against which alone failure can be construed” (1970, p. 221). The pertinent claim now is that universal mistake is unthinkable because it removes the background crucial to construing mistake. This claim is arguably true in one respect. One's genuinely thinking that a mistake has occurred arguably requires one's being correct in thinking that M: Mistake consists in conditions C. One's understanding what mistake is, it is arguable, requires correctness about M. If one must be correct about M, in one's genuinely thinking that a mistake has occurred, one who thinks that a mistake has occurred cannot be a victim of universal mistake. This line of argument is more promising than Davidson's.
One's beliefs, on whatever topic, cannot be universally, or altogether, false if any genuine belief one has requires a correct (semantic) belief analogous to M: a belief about relevant truth-conditions. In that case, any genuine belief that X is F requires a correct (semantic) belief that being F consists in conditions C. Semantic beliefs determine truth-conditions, and thus meaning, for one. If one's genuinely understanding what it is for something to be F requires such a correct semantic belief, not all one's beliefs can be false, given that belief that P requires understanding that P, and understanding that P requires true semantic beliefs regarding certain truth-conditions that determine meaning. In this case universal mistake is unthinkable.
We cannot conclude now that Davidson's principle of charity is finally substantiated. That principle is not just the thesis that universal mistake is impossible; nor is that principle entailed by the latter thesis. The principle's relevant implication now is that “it must generally be the case that a sentence is true when a speaker holds it to be” (1975, p. 169). Davidson construes this as the view that one's beliefs must be “largely correct” (1991, p. 160). If, however, one intends almost all one's beliefs to be interpreted in accord with unrestricted neutral monism, but those beliefs' being interpreted thus misrepresents the actual causes of one's beliefs (say, discrete physical objects), then one's beliefs could indeed be largely incorrect. This will be the case so long as one's intention to have one's beliefs interpreted in a certain way does not (p.236) determine, directly or indirectly, what actually causes one's beliefs. This consideration fits, moreover, with the previous view that genuine belief requires true (semantic) belief regarding truth-conditions, and thus that universal mistake is impossible. We have no avenue, then, from the impossibility of universal mistake to Davidson's requirement that one's beliefs must, by “nature,” be largely correct. Consequently, Davidson's principle of charity does not compel.
3. Beyond Charity
If our intentions to be interpreted in a certain way determine the content of our beliefs, our beliefs can generally misrepresent their actual causes. Interpreters of our beliefs can, at least in principle, come to know that such widespread misrepresentation occurs, and can interpret our beliefs accordingly. Davidson's twofold principle of charity runs afoul of these considerations. That principle will survive only if (contrary to Davidson's Gricean view of meaning) we eliminate a role for intention in meaning or (contrary to Davidson's non-idealism) we regard intentions as determining what actually causes a belief. Davidson's theory of meaning cannot consistently accommodate either of these options.
Davidson proposes that both causation and intention figure in the determination of meaning. His view bases meaning on how one intends to be interpreted (1990a), but seeks to block massive error in belief while allowing some error. I have argued that we cannot block massive, less-than-universal error once we base meaning on intention. The simplest way to correct Davidson's semantic theory is to make meaning a function not of actual causation, but of the causal relations one intends to be interpreted as endorsing. This revision preserves the Gricean component central to Davidson's account, but jettisons the causal requirements on meaning that lead to Davidson's troubled principle of charity. One can still regard meaning as dependent on observable circumstances, but only on observable circumstances acknowledged by one's interpretive intentions. The latter observable circumstances can, nonetheless, be characterized by a “distal theory” of the stimulus, like Davidson's (1990c), that appeals to shared causes salient for speakers and interpreters.
We still lack compelling support for Davidson's striking thesis that “what stands in the way of global skepticism of the senses is … the fact that we must, in the plainest and methodologically most basic cases, take the objects of a belief to be the causes of that belief” (1983, p. 436). We have seen that we need not take the actual causes of one's beliefs to be the objects of one's beliefs (even generally), when intentions fix objects of belief. In addition, we have seen that our own beliefs' objects, when determined by intentions, need not be fixed by the actual causes of our beliefs. Even if universal mistake is impossible, owing to the indispensable role of correct beliefs about truth-conditions, one's beliefs can be generally incorrect. Such is the fate of beliefs whose objects are determined by interpretive intentions rather than actual (p.237) causes. Davidson's interpretive charity, then, is misplaced, and skepticism about whether our beliefs are largely correct remains a live, if bothersome, option. Such skepticism dies hard, if at all; at least, we need to go beyond considerations of interpretive charity to challenge it. The epistemological lessons of Chapter 1, in particular, are not threatened at all by interpretive charity. Davidson has not shown that our beliefs about the external world must be, or even are, largely correct. (p.238)