Abstract
In this paper, we defend Davidson's program in truth-theoretical semantics against recent criticisms by Scott Soames. We argue that Soames has misunderstood Davidson's project, that in consequence his criticisms miss the mark, that appeal to meanings as entities in the alternative approach that Soames favors does no work, and that the approach is no advance over truth-theoretic semantics.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
A different misunderstanding is expressed in Horwich (2005, p. 4 & Ch. 8) who suggests Davidson aimed to analyze sentence meaning in terms of truth conditions. Why this is a mistake will become clear in the course of our discussion. This misunderstanding is also widespread. Burge gives it as if it were the standard account of Davidson in Burge (1992, pp. 20–1).
In his initial characterization of the project, Davidson talks of meanings in the plural. What emerges from his discussion is that the kernel of truth in the following claim does not require the assignments of meanings to sentences or words.
In “Semantics for Natural Languages” (1970, p. 62), Davidson says, “Making a systematic account of truth central in empirical semantics is in a way merely a matter of stating old goals more sharply.” He goes onto say, “Some problems that have dominated recent work on semantics would fade in importance: the attempt to give ‘the meaning’ of sentences…”; but here the inclusion of ‘the meaning’ in quotation marks serves to indicate he has in mind producing entities that will serve as meanings and a standard for translation; it is also Davidson’s view that the standard for translation emerges only in the context of radical interpretation.
Alternatively, assign an individual concept to ‘Theatetus’ and read ‘Theatetus flies’ as the list: the individual concept of Theatetus, the property of flying, and instantiation. This only renders the problem more evident.
It has been suggested to us that the problem raised here applies to the approach we sketch below. The line of thought seems to be this: “You say that the theory that assigns the function and the argument to expressions does not say how to put them together to get the meaning of ‘Theatetus flies’. But the sorts of theories you recommend consist of axioms and rules and the theory doesn’t say how to put them together to get theorems. So it’s the same thing!” The tu quoque charge misunderstands the point of the argument, which is that (i) the information we need to understand is not given by the assignment of entities to expressions; and (ii) knowledge of those assignments, excepting in the case of ordinary referring terms, is not included in the information that is needed. Furthermore, it is not, as we will explain, part of our view that an axiomatic reference or truth theory contains all the information that is needed, though it is part of our view that information about assignments of meanings to all expressions in a language is not needed.
We need both (i) and (ii) because (ii) does not ensure that we know the syntactic forms of each axiom, which is required if we are to use them to derive canonical reference theorems, as opposed to knowing what they are by way of names or descriptions. We need to know what the forms of the axioms are to use the canonical proof procedure to isolate the right set of theorems. We are assuming that knowledge of a grammar for the language of the truth theory that sorts expressions into grammatical categories such as noun phrase, verb phrase, determiner, noun, verb, adverb, adjective, etc., relative to which we understand the rules of inference. This plays a role in putting the knowledge specified in (i) and (ii) together in a way that enables us to understand the language of the reference theory, which is essential if we are to use it to understand object language expressions.
A clear potential problem with this is that it assumes the meaning of t is fixed by its referent. The Slingshot in fact exploits this in one of its assumptions.
It is obvious that for a context sensitive language the syntactic criterion won’t work; we cannot give context relative truth conditions for ‘I am hungry’, e.g., using a sentence with context sensitive features. This requires a generalization of Tarski’s Convention T. We want theorems to provide context relative truth conditions which interpret utterances of object languages sentences relative to those contexts.
See Cummins (2002) for an example of an author who takes Davidson to be reifying truth conditions and assigning them as meanings to sentences.
Quine describes the contrast he has in mind as follows: “The main concepts in the theory of meaning, apart from meaning itself, are synonymy (or sameness of meaning), significance (or possession of meaning), and analyticity (or truth by virtue of meaning). Another is entailment, or analyticity of the conditional. The main concepts in the theory of reference are naming, truth, denotation (or truth-of), and extension. Another is the notion of values of variables” (Quine 1953, p. 130).
In this context we can also treat a passage in a later paper which seems to echo the theme here. In “In Defense of Convention T,” Davidson says, “In the previous paragraph, the notion of meaning to which appeal is made in the slogan ‘The meaning of a the sentence depends on the meanings of its parts’ is not, of course, the notion that opposes meaning to reference, or a notion that assumes that meanings are entities” (1973a, pp. 70–71). He goes onto say, “The slogan reflects an important truth, one on which, I suggest, a theory of truth confers a clear content. That it does so without introducing meanings as entities is one of its rewarding qualities” (p. 71). There are two points here. First, the truth theory itself employs only semantic concepts drawn from the theory of reference; second, it makes no use of assignments of meanings as entities. This is not to disclaim the aim of giving the meaning of each sentence of the language by way of identifying for it a theorem in which the truth conditions for it are given using a sentence alike in meaning (that is required by Convention T!), or by doing so on the basis of axioms that use expressions alike in meaning to those that give their truth, reference and satisfaction conditions (for that is, after all, the idea about how it is to be done). It is only to take not of the fact that what the theory shows it does not say.
We assume a grammar for the truth theory as a background for understanding the rules of inferences that sorts terms into categories such as noun phrase, verb phrase, determiner, verb, noun, adverb, adjective, connective, etc. As noted in connection with theory S, this is needed if we are to put the information in (i) and (ii) together so as to come to understand the language of the truth theory, which is essential to using it to interpret object language sentences. It should be noted in connection with this that the object is to gain insight into the rules that attach to expressions in the object language which govern how they contributed to fixing the truth conditions of object language sentences by way of their meaning. It no part of this that we represent what the theorist knows as a model of how the speaker’s competence with respect to the rules is realized, and it could not be, given that we presuppose the speaker has a language as a condition on stating what he could know. We thank Miguel Hoeltje for helpful discussion in connection with these issues.
Interestingly, if we identify a meaning theory with what we must know, then we need not know the truth theory: we need to know what it is and what its axioms mean, but knowledge that its axioms are true is not part of what we must know to do the work that needs to be done. This is important because it might enable us to apply this technique to a language with vague terms and semantically defective terms without worrying about the truth theory not being true. At the same time, though, relative to the knowledge that the terms in the object language are not semantically defective, we can infer that the truth theory is true, and so, relative to this additional bit of knowledge, come to know what the truth theory states. See Lepore and Ludwig (2005, Ch. 10).
Soames declares (p. 6) that Davidson thought “truth theories that are both true and compositional will end up deriving only those statements [‘S’ is true iff P]” in which what replaces ‘P’ translates ‘S’. Again, he is mistaken. Davidson’s idea was that context sensitive elements in natural languages would play a crucial role. As he himself notes (Footnote 10 on p. 26 of Inquiries into Truth and Interpretation; Davidson 2001), “[c]ritics have often failed to notice the essential proviso mentioned in this paragraph. The point is that (S) could not belong to any reasonably simple theory that also gave the right truth conditions for ‘That is snow’ and ‘This is white’. (See the discussion of indexical expressions below).”
We are convinced it is not. See Lepore and Ludwig (2005, Ch. 11).
In “Radical Interpretation” (Davidson 1973b, p. 25), Davidson says, in motivation for asking the question “What could we know that would enable us to understand any potential utterance of a speaker?”, that “… it is not altogether obvious that there is anything we actually know which plays an essential role in interpretation.” In “A Nice Derangement of Epitaphs” (Davidson 1986, p. 438), he says, “To say that an explicit theory for interpreting a speaker is a model of the interpreter’s linguistic competence is not to suggest that the interpreter knows any such theory…They are rather claims about what must be said to give a satisfactory description of the competence of the interpreter.” In “The Structure and Content of Truth” he says that the aim of the theory is, inter alia, to describe a certain complex ability: “it at once describes the linguistic abilities and practices of the speaker and gives the substance of what a knowledgeable interpreter knows which enables him to grasp the meaning of the speaker’s utterances. This is not to say that either speaker or interpreter is aware of or has propositional knowledge of the contents of such a theory” (Davidson 1990, pp. 311–312). In “The Social Aspect of Language” he says, “let me say (not for the first time): I do not think we normally understand what others say by consciously reflecting on the question what they mean, by appealing to some theory of interpretation, or by summoning up what we take to be the relevant evidence” (Davidson 1994, p. 3).
By ‘the structure of their dispositions’ we mean no more than ‘what the rules are with respect to the use of the words which the dispositions express’.
This serves as a response to another objection we have heard: that the account of how to provide a compositional meaning theory would mean that a child learning its first language would have to have the concept of truth, which is implausible. This would be so only if we were making the claim that the way the competence that the theory represents is realized in speakers is by way of their having propositional knowledge of the theory itself. But we have no such commitment, and it cannot be taken to be a general commitment of giving a compositional meaning theory because every compositional meaning theory that aims at an explicit statement of the meaning of every sentence will deploy concepts which a child learning a first language will not plausibly have, such as the concept of meaning itself.
Soames says, “Davidson originally held that a truth theory for L qualifies as a theory of meaning, if knowledge of what it states is sufficient for understanding L” (2008, p. 5). It is fair to call a body of knowledge sufficient to interpret any sentence of a language on the basis of rules governing its semantical primitives a meaning theory for the language, and it is fair to say this is what Davidson sought in an account of how the meanings of sentences depend on the meanings of their contained words. It would then follow that if knowledge of what a truth theory states sufficed to understand the language for which it was a theory, then it would be a meaning theory. This conditional thesis is trivial and nonsubstantive, and there is no reason to think Davidson would not have been happy to endorse it. But since it is trivial and nonsubstantive, it is also not a thesis that he would have had to give up. The substantive question is what knowledge one could have (propositional knowledge) that would suffice to interpret a speaker’s language. And the exegetical question about “Truth and Meaning” is what Davidson’s position on that was. In “Truth and Meaning,” he does speak as if the meaning theory is the truth theory. But even there it seems clear he was thinking that the required knowledge went well beyond what was stated by the truth theory, for he offers an argument to show that a true truth theory would meet an appropriate analog of Convention T for a context sensitive language because it would track correctly the use (actual and potential) of predicative terms in combination with demonstratives. Isn’t it transparent that that is something one would have to know in order to use the truth theory for interpretation? In a retrospective remark in “Reply to Foster” Davidson says, “That empirical restrictions must be added to the formal restrictions if acceptable theories of truth are to include only those that would serve for interpretation was clear to me even when I wrote ‘Truth and Meaning’. My mistake was not, as Foster seems to suggest, to suppose that any theory that correctly gave truth conditions would serve for interpretation; my mistake was to overlook the fact that someone might know a sufficiently unique theory without knowing that it was sufficiently unique. The distinction was easy for me to neglect because I imagined the theory to be known by someone who had constructed it from the evidence, and such a person could not fail to realize that his theory satisfied the constraints” (Davidson 1976, p. 173). If we take Davidson at his word (why wouldn’t we?), he misexpresses his underlying thought in “Truth and Meaning” when he speaks as if the truth theory were a meaning theory itself.
Davidson makes this point in explaining why a theory of truth consisting just of the infinite class of (T) biconditionals in which the right hand side translates the sentence mentioned on the left is inadequate: “Such a theory would yield no insight into the structure of language and would thus provide no hint of an answer to the question how the meaning of a sentence depends on its composition” (Davidson 1970, p. 56).
Soames says, “One could … take propositions to be inherently and intrinsically representational, and so sui generis. However, this is a council of despair. Davidson would not accept such obscuritanism, and we should not either. If we posit structured propositions as meanings of sentences, we ought to explain what they are, and how they are able to play the roles we assign to them” (Soames 2008, p. 17).
For all we have said, natural language may involve terms that (putatively) refer to or denote propositions (just as it may contain terms that (putatively) refer to or denote unicorns). In giving the semantics, then, we may need terms in the metalanguage that correspond to those object language terms. Our point is that the machinery of the compositional meaning theory need not itself introduce such terms or endorse the existence of any entities of that type.
Davidson says, “nothing strictly constitutes a theory of meaning,” because the truth theory is not a theory of meaning and the “statement that a translational theory entails certain facts is not, because of the irreducible indexical elements in the sentences that express it, a theory in the formal sense” (Davidson 1976, p. 179). Davidson says this because of his paratactic analysis of ‘that’-clauses; if we do not follow him, we need not shy from saying there is something that constitutes the meaning theory.
References
Barwise, J., & Perry, J. (1981). Semantic innocence and uncompromising situations. Midwest Studies in Philosophy, 6, 387–403.
Burge, T. (1992). Philosophy of language and mind: 1950–1990. The Philosophical Review, 101(1), 3–51.
Chihara, C. S. (1975). Davidson’s extensional theory of meaning. Philosophical Studies, 28, 1–15.
Church, A. (1951). The need for abstract entities in semantic analysis. In Proceedings of the American Academy of Arts and Letters (p. 80).
Cummins, R. (2002). Truth and meaning. In J. K. Campbell, M. O’Rourke, & D. Shier (Eds.), Meaning and truth: Investigations in philosophical semantics. New York: Seven Bridges Press.
Davidson, D. (1966). Theories of meaning and learnable languages. In Y. Bar-Hillel (Ed.), Proceedings of the 1964 international congress for logic, methodology and philosophy of science. Amsterdam: North Holland Publishing Co.
Davidson, D. (1967). Truth and meaning. Synthese, 17, 304–323.
Davidson, D. (1970). Semantics for natural languages. In Linguaggi nella Societa e nella Tecnica. Milan: Comunita.
Davidson, D. (1973a). In defence of Convention T. In H. Leblanc (Ed.), Truth, syntax and modality. Dordrecht: North-Holland Publishing Company.
Davidson, D. (1973b). Radical interpretation. Dialectica, 27, 314–328.
Davidson, D. (1976). Reply to Foster. In G. Evans & J. McDowell (Eds.), Truth and meaning: Essays in semantics. Oxford: Oxford University Press.
Davidson, D. (1986). A nice derangement of epitaphs. In E. Lepore (Ed.), Truth and interpretation: Perspectives on the philosophy of Donald Davidson. Cambridge: Blackwell.
Davidson, D. (1990). The structure and content of truth. The Journal of Philosophy, 87(6), 279–328.
Davidson, D. (1994). The social aspect of language. In B. McGuinness & G. Oliveri (Eds.), The philosophy of Michael Dummett. Dordrecht: Kluwer.
Davidson, D. (2001). Inquiries into truth and interpretation (2nd ed.). New York: Clarendon Press. Original edition, 1984.
Glock, H.-J. (2003). Quine and Davidson on language, thought, and reality. Cambridge, UK: Cambridge University Press.
Horwich, P. (2005). Reflections on meaning. Oxford: Oxford University Press.
Katz, J. (1982). Common sense in semantics. Notre Dame Journal of Formal Logic, 23(2), 174–218.
Lepore, E., & Ludwig, K. (2005). Donald Davidson: Truth, meaning, language and reality. New York: Oxford University Press.
Quine, W. V. O. (1953). Notes on the theory of reference. In From a logical point of view. Cambridge: Harvard University Press.
Quine, W. V. O. (1960). Word and object. Cambridge: MIT Press.
Soames, S. (1992). Truth, meaning, and understanding. Philosophical Studies, 65(1–2), 17–35.
Soames, S. (2008). Truth and meaning: In perspective. Truth and Its Deformities: Midwest Studies in Philosophy, 32, 1–19.
Stich, S. (1976). Davidson’s semantic program. Canadian Journal of Philosophy, 6, 201–227.
Tarski, A. (1983). The concept of truth in formalized languages. In Logic, semantics, metamathematics. Indianapolis: Hackett Publishing Company. Original edition, 1934.
Wallace, J. (1978). Logical form, meaning, translation. In M. Guenthener-Reutter (Ed.), Meaning and translation. London: Duckworth.
Wittgenstein, L. (1961). In D. F. Pears & B. F. McGuinness (Eds.), Tractatus logico-philosophicus. London: Routledge & Kegan Paul Ltd.
Acknowledgements
We would like to thank John Burgess, Gil Harman, and Miguel Hoeltje for discussion and comments that have led to improvements in this paper.
Author information
Authors and Affiliations
Corresponding author
Additional information
Affiliation of K. Ludwig after July 1, 2010: Department of Philosophy, Indiana University, Bloomington, IN 47405-7005, USA.
Rights and permissions
About this article
Cite this article
Lepore, E., Ludwig, K. Truth and meaning redux. Philos Stud 154, 251–277 (2011). https://doi.org/10.1007/s11098-010-9536-x
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-010-9536-x