Study-guide summary of Davidson’s “On the Very Idea of a Conceptual Scheme”

Philosophy 405: Contemporary

Spring, 2005

Wilburn

405 main page

Purposes and Significance: Davidson here is deflating claims of Conceptual Scheme Relativism (CSR). This is the proposal that different people, cultures, historical epochs or intelligent species might have, not merely different conceptual schemes, but incommensurable ones, i.e., schemes which are so radically different from each other that they lack a basis for mutual comparison.

The proposal that there might be such incommensurable conceptual schemes has been used in contemporary philosophy to a number of ends. One such end is metaphysical antirealism, i.e., the view that there is no one determinate way the world is or that there is no privileged language in terms of which we should describe it. Another such end is epistemological skepticism, i.e., the idea that, for all we know, the world might be radically different in general character than we ordinarily take it to be. Thus, much may seem to depend upon whether or not we take CSR seriously.

Davidson’s Motivating Suspicion: Before presenting his arguments against Conceptual Scheme Relativism, Davidson articulates a general suspicion against the view. His article can be reasonably viewed as one long systematic attempt to support this initial suspicion. His suspicion is this: The aspirations of CSR are betrayed by the fact that CSR advocates’ most common operative metaphor is that of variant points of view. However, he notes, it only makes sense to speak of variant points of view if “there is a common coordinate system on which to plot them.” (p. 329) And to concede such a common coordinate system, he notes, it to effectively deny claims of radical incomparability. This, he feels, explains why the examples offered by CSR advocates always seem so unconvincing. Kuhn, for example, describes Aristotelian physics and Ptolemaic astronomy as part of a radically different paradigm, but then proceeds to elegantly describe them in exacting post-revolutionary terms (p. 328)  

Davidson’s Argument:

Let’s now consider the turns of Davidson’s argument in support of the above-described suspicion. Arguing that thought presupposes language, he suggests that we associate conceptual schemes with sets of intertranslatable languages. (p. 330) This is a substantive suggestion, as it simultaneously denies us any critical metaphysical stance outside of language from which to assess conceptual schemes and provides us with a test for differences between conceptual schemes, i.e., incommensurable conceptual schemes are not intertranslatable.

Now, the possibility of such non-translatability, he grants, can be either total or partial. His concern in the rest of his paper is to consider and reject each of these two possibilities in turn. (p. 330)

Consider Total Non-Translatability: Once again, Davidson expresses a suspicion. We could never conceive of a non-translatable language, he suggests, because our best available criterion for languagehood is translatability itself. But this, he claims, should be demonstrated, not simply assumed.

To the end of providing such a demonstration, Davidson asks the following question: How could incommensurable conceptual schemes be even theoretically possible? How would two conceptual schemes have to operate differently, and what would they have to operate differently on, for such incommensurability to arise? Davidson concludes that to answer such questions in a way conducive to CSR, two elements must be present in our account of the relation between conceptual schemes and reality. We must view language as a regimenting force operating on a shared realm of data. (p. 332) There must be, in effect, as difference between scheme and content, where the scheme somehow proves adequate to the content. But this idea, in turn, is analyzable in terms at least two broadly different analogies. On the first, schemes (languages) organize (systematize, divide up) content, and on the second, schemes fit (predict, account for, face or cope with) content. As for the data involved, that which gets organized must be either reality (physical evidence) or experience (sensory promptings). (p. 333)

Organization Metaphor: Consider the organization metaphor. How can schemes be viewed as organizing the data of nature or experience. The data, Davidson notes, cannot be simply undifferentiated. To be subject to organization or fit by schemes, it must be divided into parts. But to be divided into parts, Davidson maintains, it must be individuated “according to familiar principles,” that is, through the resources of our existing and familiar conceptual scheme: “A language that organizes such entities must be a language very like our own.” (Ibid.) And in being like our own, it must contain a familiar conceptual repertoire ranging, not only over experiences, but familiar physical entities as well (“knives and forks, railroads and mountains, cabbages and kingdoms”). (Ibid.) As a result, Davidson concludes, we cannot substantively understand the notion of the data to be organized or fit by schemes except in the prosaic and familiar terms characteristic of our home language. But, since we can speak of the data only in terms of our own familiar language, we cannot make substantive sense of the idea that this data might be differently organized by another language untranslatable into our own. Since we are construing schemes as languages, we get the following consequence: If schemes are in the business of organizing the data, we cannot substantively imagine two incommensurable schemes both effectively organizing the same overall empirical data. [Note: This is one of the sketchiest, and thus one of the weakest, parts of Davidson’s paper. You would do well to ask yourself whether or not Davidson really gives an argument here at all that “experience would have to be individuated according to familiar principles.”]

Fitting Metaphor: Now, consider the fitting metaphor. Here, Davidson invokes a fair amount of technical background, but his largest point can be understood without excessive reliance upon it. According to the fitting metaphor, acceptable conceptual schemes or theories cope with, predict or account for all available sensory evidence. Here the putative connection between language and reality differ from that of the organization metaphor in its points of contact. Predicates organize, but sentences predict. But Quine showed in “Two Dogmas of Empiricism” that no sentence, considered in isolation, predicts or is confirmed by any given bit of experience. Thus, the fitting relation is supposed to stand between our total collection of sentences (Quine’s “whole of science” and our total evidence, rather than between our predicates and individuable parts of reality). 

But to say that our total conceptual scheme (language) fits reality in this sense is just to say that it is true. And to say that it is true is just to say that a consistent and systematic set of “Tarski-sentences” can be offered of the form “s is true iff p” where s is a sentence of our language and p is either that sentence itself (without quotation marks) or else a translation of that sentence into our language (e.g., “Schnee is weiss” is true iff Snow is white.)

But, since on this account the notion of truth is itself explicated in terms of our procedures of translation, we are left unable to make sense of the idea of a conceptual scheme (language) being true without being translatable into our own. In short, if we assume that our best current theory concerning the nature of truth is provided by Tarski’s account (on which we understand truth for a language when we can generate a T-sentence for every single sentence of that language), we get the result that no conceptual scheme can fit a language and yet be untranslatable into, and thus incommensurable with, our own. [Note: Here you might want to take exception to Davidson’s description of the significance of this finding when he writes “That experience takes a certain course, that our skin is warmed or punctured, that the universe is finite, these facts, if we like to talk that way, make sentences and theories true. But this point is put better without mention of facts.”]

Consider Partial Non-Translatability: Above, Davidson allegedly rules out the possibility that we could ever make sense of the idea of an alternative conceptual scheme that is totally incommensurable into our own. Now, let’s consider his more modest claim, that we cannot even make sense of the idea of an alternative conceptual scheme which is only partially, but significantly, non-translatable into our own. (p. 335) Remember that Davidson rejects the idea that thought is possible without language or, alternatively stated, that mind is prior to language. (p. 329) But this, Davidson maintains, has serious consequences for translational procedure, as it implies that we cannot attribute mental states to a speaker prior to attributing meanings to his sentences. There is a close relationship between belief attribution and sentential interpretation. Or, as Davidson writes,

Now we can put the point in a somewhat sharper way. Allow that a man's speech cannot be interpreted without knowing a good deal about what he believes (and intends and wants). And that fine distinctions between beliefs are impossible without understood speech; how then are we to interpret speech or intelligibly to attribute beliefs and other attitudes? Clearly we must have a theory that simultaneously accounts for attitudes and interprets speech, a theory that rests on evidence which assumes neither. (p. 335)

What must such a theory be like? For Davidson, it must be one concerning which sentences a subject speaker accepts as true. This is the point of his ketch and yawl example. (p. 335) To even begin the process of interpretation, you must assume that his companion who utters, “Look at that handsome yawl” believes that the ketch is a yawl. This is an application of the principle of charity, and interpretation requires it. To interpret another’s speech, or even interpret another’s behavior as speech behavior, we must begin with the assumption that he holds his sentences to be true and thus largely agrees with us concerning what the relevant facts are. Thus, disagreement is only interpretable against a wide background of agreement, and his conceptual scheme cannot, after all, be all that different from our own.