At this point in the discussion, I am beginning to suspect that Livingston and I have different conceptions of what Davidson’s “framework” is.  I take it to be quite a bit more than the idea that a theory of meaning is a recursive truth-definition. So this response will set out what I take Davidson’s view and framework to be, and explain why, on that understanding, the inadequacies Livingston and McDowell ascribe to Davidson’s framework are not genuine inadequacies. Of course it may well be that I am just not getting something.

I will begin by discussing two peripheral points: First, my understanding of Davidsonian indeterminacy makes it something more than ambiguity that can be cleared up by determining the intention of the speaker or author. Second, addressing the remark at the bottom of footnote 14,  I give an explanation of what I take to be the Quinean-Davidsonian approach to ontology. The main part of this response discusses the central issue of whether a Davidsonian account of language is adequate as a philosophical account of language.

I. Indeterminacy and “Being Supervenes on Truth”

A) Indeterminacy

Livingston characterizes Davidson’s limitation of indeterminacy as follows:

 …Any indeterminacy or ambiguity that actually arises in the course of interpretation, either across languages or within a single one, will be relatively local, limited, and fairly easily resolved. Disagreements may arise about which of the various senses of an ambiguous term is intended, but this kind of disagreement will evidently no longer deeply threaten our pre-existing intuitions about the relative fixity and determinacy of meaning. (5 pages in)

I take both Davidson and Quine to be committed to something more radical than that.  I take their view to be that there are nothing but language-like items underlying language—there are no meanings in the sense of interpretation-enders. Indeterminacy is not epistemological, but ontological.  I take Quine and Davidson to agree with Derrida on this point. Indeterminacy is a feature of any account that takes language-like signs as basic to thought. When there is indeterminacy, it is not always a matter of lack of information or mis-guessing the intention—the intention is itself in a notation about which problems of interpretation equally arise. To deny meanings in Quine’s way of denial is to say that thought and intention, rather than resolving the interpretation of signs that are subject to interpretation, are themselves language-like and so subject to interpretation. So language and the thoughts expressed by language, in Derrida’s terms, are writing-like.

An analogy used by both Quine and Davidson is helpful here: Assigning numbers to lengths is indeterminate in a similar sense. There are no real numbers corresponding to lengths that we are somehow in the dark about—there are only numbers relative to a choice of unit.

B) Being supervenes on truth       

The metaphor of assigning numbers to lengths is also appropriate to the Quinean-Davidsonian perspective on the question of being. Like Carnap in “Empiricism, Semantics and Ontology,”[1] they treat entities as organizational devices for sets of true sentences. The primary connection to the world is not reference, but truth—in Quine’s terms, we respond to the world with occasion-sentences, not by naming. If sentences are the mode of contact with the world, and sentences do not refer, as versions of the Frege argument show, then a theory is primarily a set of truths. Those truths, as Quine showed, can be organized by very different singular terms. Davidson retains the idea that truths are primary, but given his generalized “maximize agreement” principle, odd ontologies are not candidates.

II. The Principle of Charity and Rules of Language

The “principle of charity” for Davidson is not only a transcendental condition on language-interpretation, but also a condition on interpreting and understanding intentional actions generally. “Charity” is part of understanding what it is to be an agent. Speech acts and text-producing acts are special cases of actions. In their case, charity interprets speech actions as reflecting mostly true beliefs and mostly accurate conceptions of the good. To treat events as intentional, purposive actions just is to suppose that the Other is in a shared environment with us and is by and large aware of that environment, broadly construed, and sharing evaluations of that environment. Being largely aware of a shared environment, for Davidson, is believing mostly truths and mostly valuing the good.

If the environment contains squares, and we take the speech act to be sincerely commenting on this aspect of the environment, charity suggests as the default hypothesis that the person is referring to squares by means of a term that expresses the concept “square.” The rule for applying the concept “square” is quite simple: Call something “square” if and only if it is a square. The truth-conditions for “Fred is square” are as follows: “`Fred is square’ is true if and only if Fred is square.”

Grasping truth-conditions is not coming to have an analysis of a concept or coming to have a definition of the application-conditions of a word in other terms. Rather, a necessary and completely sufficient condition for grasping what “square” means is being able to specify in one’s own language what it takes for “square” to apply to an object. The most direct way of doing that is by the sentence, “‘Square’ is true of an object if and only if it is square.” Now, this assumes that the interpreter is using “square” rather than citing “square.” Admittedly, Davidson gives little helpful information about what it takes for a speaker to use a term, that is, for a term to be in the speaker’s language.[2]

It is difficult to see how the problem that any amount of data is compatible with an infinity of projections by the speaker being interpreted can get a purchase, on a Davidsonian account. If the other is an agent in a shared environment aware of that environment, then, if the environment contains squares, the agent is aware of squares. If the agent is aware of squares, then the agent has the concept of a square. There is nothing to having the concept square beyond having a term whose extension is all and only the squares. For that to be the case, there must be a term F in the person’s idiolect such that F is true of an object if and only if the object is square.

III. How do we come to have a language?

I think the real issue for Livingston is expressed in his footnote 5, where he writes, “[Davidson’s framework]..does not help us to understand how we can enter into a linguistically shaped understanding of the world from a position innocent of that understanding.”

Now Davidson, following Quine, is a holist about what it takes to have concepts and beliefs. He is also a holist about how we achieve agency, since there is no agency without language and no language without agency. His articles on triangulation and
agency[3] give an account of the connection between recognition of an Other and coming to have a language that notes that genuine intentionality, and so agency, requires having a language. Davidson notes how some aspects of animal behavior, such as responding to other organisms responding to the environment, a kind of pre-linguistic proto-triangulation, are required for the introduction of full intentionality with language. The background of shared practices goes very far back indeed.

Davidson’s work explaining how the process of becoming an agent and a language-user gets off the ground, as it were, seems to me to go as far as philosophical reasoning can go in giving an account of the coeval origins of language and the attribution of beliefs and desires to oneself and others. So, how is it possible to go “from a position innocent of that understanding” to a position where one has language, recognizes agency in oneself and others, and has become an interpreter? This seems to me to be a question with which pure philosophy, in isolation from developmental psychology and linguistics, is ill-equipped to deal. Beyond setting out the sort of a priori considerations Davidson has in the above-mentioned articles, one would seek empirical theories. There are empirical theories, supported by actual data, that humans have innate dispositions both to attribute intentional states to others and to acquire a human language. Experimental results are clarifying the stages by which infants acquire language and an understanding that there are Others in the world. How philosophy, of whatever subtlety and depth, can intervene on this topic, beyond what Davidson has done, is not clear to me.