Truth, language, and reality: an essay on Donald Davidson’s philosophy of truth

Download 104.19 Kb.
Size104.19 Kb.
1   2   3
Objections: absurd and sceptical

There are various arguments which aim to show epistemic theories end up re-defining our concept of truth so much is no longer resembles the concept we were originally interested in. One version of this worry, voiced by Russell against the early Joachim-Bradley version of coherence, regards the inability for pure coherence alone to account for truth. He argues “there is no reason to suppose that only one coherent set of beliefs is possible”39 (a possibility known as incommensurability). We can imagine two different people with two separate systems of belief which are individually coherent but mutually incompatible40. For example we can imagine that there is one system of beliefs in which the sentence “Jane Austen died in her bed” coheres and another system of beliefs in which the sentence “Jane Austen did not die in her bed” coheres, leading the coherence theorist to assert paradoxically that both sentences are true thus violating the law of excluded middle.

               However it’s not clear this objection poses much concern for coherence theories properly understood. At least in contemporary philosophy Quine and Davidson’s arguments for indeterminacy suggest that it’s unavoidable that there could be many coherent systems ways of describing the world (and that this is not so unintuitive after all). So the fact that there could be more than one set of coherent beliefs may not be enough of an objection.

The intuitive crux of Russell’s objection is that multiple logically contradictory sets of belief could exist. But in the context of the early 20th century Russell’s objection can be met. Joachim and Bradley’s definition coherence amongst a system of beliefs ensures that each sentence would be entailed by the system and that the system was itself entailed by that sentence. Given we run with this definition of entailment we can simply take the conclusion of Russell’s objection, invert it, and demonstrate its false premises following a proof by contraposition:


Definition of coherence: if a proposition is contained within a coherent set, then that proposition is entailed by that set such that “p  ╞ p”

Starting with the premises:

1. 1╞ p

2. 2╞ p

3. {1,2} ╞ P & {1,2} ╞ P

However this is a contradiction and so we must infer that:

4. Either (1╞ p) is not the case, or (2╞ p) is not the case.

Premise 4 implies it can’t be the case for there to be two sets which are both coherent and which prove contrary propositions. The jump from 1&2 to 3 may be protested by the correspondence theorist due to the fact it assumes the final criterion for coherence is the larger combination set rather than the two original ones. But the coherentist can simply respond that a coherent system would in fact entail that no matter which beliefs were added they would either be entailed (if they were true and coherent) or they would be shown to be incoherent (and false). Joachim and Bradley made explicit their notion of truth entailed that there was conceivably only one coherent set of beliefs possible in the end.

Russell acknowledged that this response was available, and that Joachim and Bradley both made use of it, but argued it was a deus ex machina41. He remarks this move is "in general discredited and that idealists are themselves rather ashamed of [it]"42 , and argues the whole notion is empty for coherentists hold "'what is true in the end' as though what is true 'in the end' were anything different from what is true"43. Although Russell doesn’t provide a reason why the converse (why what is true is not simply what is true in the end) is implausible, this point is remarkably close to the charge of explanatory vapidity his own account faces. He seems to have demonstrated that all this talk of “in the end” is merely another way of saying “is true”.

A response to Russell which evades reaching the pitfall above can be found in Blanshard44 who argues that it is not as if we ourselves arbitrarily construct a system of beliefs out of sheer fantasy. He argues that perceptual beliefs are caused by the world and our concepts are a complex and interdependent network of the means by which we organize, predict, explain, and understand the world. The objection Russell raises requires that coherence is purely internal and subjective; something Blanshard, Dummett, Putnam, and the Vienna Circle all reject. Furthermore if we are willing to grant Davidson’s rejection of the dualism of scheme and content Russell’s objection can be characterized as presupposing this untenable dualism.

  Davidson’s rejection of the scheme-content dualism seems to provide a well articulated culmination of the two threads running through this discussion; that deep seated conceptual relativism is impossible and in the same token our relation to the world is causal. Russell’s objection requires not simply that there are two possible sets of beliefs about the same world (say one in which distance is measured in yards and the other metres), but that there can be radically different conceptual schemes. But Davidson urges us to reject this move. Perhaps as suggested by Bradley and Joachim, the idea we could have multiple schemes which don’t themselves contain reference to a single one for truth itself is incoherent. Davidson urges us to abandon the entire notion of having a conceptual scheme that organizes unconceptualized empirical content. The second aspect of Davidson’s rejection of the scheme-content dualism brings Blanshard’s point to the fore. Davidson also urges us to accept the relationship between the world and our beliefs as causal. The consequence of this line of thought Davidson makes clear;

“In giving up dependence on the concept of an uninterpreted reality, something outside all schemes and science, we do not relinquish the notion of objective truth – quite the giving up the dualism of scheme and world, we do not give up the world, but re-establish unmediated touch with the familiar objects whose antics make our sentences and opinions true or false”45.

A second line of objection to epistemic theories, also from Russell46, is that there is no reason for the epistemic theorist to abide by the laws of logic. As we have seen from the above argument bivalence makes it so that a great many beliefs are incoherent, so why not get rid of bivalence? Actually, it seems that we can make every single belief perfectly cohere if we simply abandon all logic. Moreover, if coherence is defined in such logical terms, doesn't this presuppose that the logical framework it is constructed on is true outside of coherence?

             This objection, however, is quite ambiguous as to what it is trying to argue against. Russell seems to be presupposing that there are laws of logic or axioms which are self-evident or prima facie true; which is (as we saw from Kant’s second point) exactly the type of position the epistemic theory is arguing against. Within a model there are certainly axioms and rules of inference but, as Blanshard47 makes a large point of, these aren’t necessarily taken to be true. The “laws” of logic are merely the means by which we adjudicate the truth of sentences within that model. If we want to prove these axioms or rules true or justify our endorsement of them it would seem absurd to dogmatically assent to their truth; especially considering the dispensability of many previous axioms and rules. Russell himself dispelled the previously axiomatic principle of sufficient reason stating "I cannot see why we should expect a reason for everything"48 which from Anaximander onwards had been taken as a pre-requisite to science and philosophy.

There is also the response that the pragmatic and verificationist theories of truth can utilize which is that the reason we keep laws of logic, or logic at all, is because it is so incredibly useful. As Duhem, Quine, and Neurath49 have shown via underdetermination of theory by evidence, we can always construct multiple theories explaining any given empirical observation. The reason why this doesn’t lead us to scepticism about hypotheses is because the system of judgment we endorse, science, is adopted in virtue of its usefulness, simplicity, digestibility, and law- and truth-conduciveness, not by giving us an unambiguous criterion of truth which places it on higher ground than any other theory.

The intuitive engine behind Realist theories in general poses more of a problem to epistemic theories. Although epistemic theories can address the sceptical worry that we can never know what it is that makes our judgements true, it looks as if they replace this by not being able to address the sceptical worry that we can never get things objectively right. The dialectical ground the realists find themselves so firmly rooted in is that, at least for most, what’s essential to the intuitive concept of truth is its objectivity and mind-independence.

We can accept the Kantian point that justification must always be articulated and supported by entities with judgeable contents (as Davidson puts it “nothing can count as a reason for holding a belief except another belief”50; or as Sellars claims, by endorsing something as knowledge we place it “in the logical space of reasons”51). But this need not commit us to the idea that truth is epistemic. As we saw earlier Realist theories can easily argue that concerns about justification need not motivate our metaphysical account of truth. Unless the epistemic theory gives us reason to redefine objectivity and mind-independence (in their terms) for epistemological concerns it appears to be on the same dialectical footing as the Realist theories.

 Taking stock, it seems that the historical debate over epistemic theories leads in the same direction as the debate over Realist theories. As Russell’s scruples with Joachim and Bradley indicates, if taken in a robust manner, coherence theories end up as explanatorily vacuous as robust correspondence theories. In addition the dialectic between the more modern variants of epistemic theories and Realist theories indicates that we have two equally powerful intuitions which have theories of truth as their battleground.

Davidson on truth:
It would be worthwhile at this point to take a step back and think of just what the philosophical interest of truth is; what were the motivations for offering a theory of it in the first place? Having an answer to scepticism and an explanation for justification are what I have been arguing are two major motivations. But history as we have seen indicates each theory only offers an answer only to its preferred sceptic. Realist theories derive their intuitive pull from the idea that truth is essentially mind independent and objective, epistemic theories' intuitive pull is in the equally entrenched sentiment that truth is related to knowledge, justification, and assertions, but each theory leaves the other’s intuitions behind.

Over the course of his career Davidson has pressed52 that the concept of truth is necessary for communication. A grasp of the concept of truth, he argues, is essential to the idea that utterances have truth conditions and that interpretation of utterances requires the concept of truth. He reasons a predominant motivation for providing a theory of truth in the first place is to give content to the means by which we interpret and assess the various assertions, descriptions, predictions and so on made by speakers in a language. Provided we accept Davidson’s point, if we can provide an account of truth such that it meets the demands it has as an element of communication, we won’t need to provide a robust definition of knowledge and fall prey to scepticism or irreconcilable disputes between equally powerful intuitions. From the perspective of language and communication, truth is both objective53 and related to justification.

What we need now is to begin to trace the intricate connections between truth and communication. One place to start is with an intuitive point made by Grice that “what words mean is a matter of what people mean by them”54. One of Grice’s main insights55 about communication is that our means of discerning what a speaker means by a particular utterance will largely depend on our ability to interpret them with having certain intentions or propositional attitudes. It follows that what gives content to and justifies our ascription of intentions and propositional attitudes is the world of which they are directed and which is publicly assessable. It is here that the concept of truth becomes relevant. The necessary third vertex in triangulating between a speaker and an interpreter is the objective world, and this objective world is that which we can be right or wrong about, where truth is essential. Davidson expresses the relation between this Gricean thought and truth theories in saying:
“A theory of truth links speaker with interpreter; 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.”56
Now what is needed is a more detailed account of how it is that a theory of truth links a speaker an interpreter and how it is that truth in this role taps into the intuitions of the Realist and epistemic theorist.

This is where Davidson will point to Tarski57 as demonstrating how to capture the original Aristotelian definition of truth and in the process formalize the last point on which both major theories were agreed on. This is because Tarski’s definition of truth is extensional and recursive. An intensional definition of truth provides necessary and sufficient conditions for truth and has proven problematic. However an extensional definition merely characterizes the range of things which can be said to be true; it tells us what sorts of things are true. A recursive definition is one which does this by utilizing the primitive notion of truth. Given this Tarksi’s definition is minimal enough for all parties to agree on but substantial enough to map out where to start looking for connections between truth and other concepts.

Tarski’s definition of truth works, very roughly, as follows. Firstly we take the notion satisfaction to be basic and defined recursively e.g. satisfaction for predication of a variable in a model U under variable assignment S: ╞U Px [s]  xPU. We can then give the extension of truth in a metalanguage which contains sentences called T-sentences of the form “’s’ is T in L iff p” where ‘s’ is a structural description of a sentence in an object language containing all the sentences in L and ‘p’ is a logically identical sentence to s expressed in the metalanguage. The extension of truth in L will then be given by all the true T-sentences for L.

Davidson argues58 we now must provide a means of demonstrating how Tarski's theory links into a theory of communication. Davidson does this by utilizing his notion of radical interpretation. The role of the radical interpreter is this: without having prior knowledge of meanings or the propositional attitudes of a speaker construct a Tarski-style truth theory for that speaker’s idiolect. This is done by noticing the interconnections between this speaker’s utterances and the various features of the world of which these utterances are directed and which is available to both speaker and interpreter. Davidson’s larger role for radical interpretation is to shed light on the nature of meaning, but for the present purposes it is not necessary we accept or explicate these points. What is necessary is Davidson’s point that there would be no way for a speaker’s utterances to be interpreted without the interpreter having a grasp on the concept of truth which is given content by Tarksi’s definition.

One way in which truth is necessary for radical interpretation, Davidson emphasises a point tapped into by Realist theories, is that truth must function as a mind-independent intersubjective standard. Understanding that “apples” refers to apples or that “is green” is true of green things requires that one understands that things may or may not be apples or green. But recognizing the possibility of error entails one acknowledges that there is something of which we can be wrong, namely, the way things are. This is why our concept of truth must be objective. The entire notion of normative practices like language requires a concept of getting things right or wrong.

A similar point has been noted by Sellars59 and Brandom60 who have made clear the fact that meaningful communication requires the application of a concept, and that the application of a concept is essentially normative. For Davidson one crucial concept which distinguishes uniterpretable activity of everything from thermometers to fish from interpretable activity discursive rule-following humans is the acquisition of the concept of objective truth and falsity.

A second point, in which truth is necessary for communication, Davidson stresses, is the fact that a theory of truth for natural languages must be an empirical theory of truth.61 An empirical theory of truth entails that whatever evidence or data that is entered into the second half of a T-sentence is publicly ascertainable, law-like, universally quantified, and applied to utterances which are relativised to a speaker, time, and place. For example, in the T-sentence “’snow is white’ is T in English iff snow is white”, what suffices to determine whether or not snow is white will be subject to however it is that we empirically determine that (in normal observational conditions) whatever it is that snow refers to by the speaker satisfies the truth-conditions for the predicate “is white”. It should be stated here the empirical element of the theory is minimal; it makes no commitments to the nature of experience, the existence of sense data, and so on. It is merely a result from the point emphasized by Wittgenstein, Dewey, Grice, Quine and the whole of post-war Oxford philosophy of language: that language and meaning is and must be public.

In order for us to feel warranted in applying an empirical theory of truth to speakers in such a way that they agree with publicly ascertainable evidence we must accept a certain primacy in the principle of charity. Davidson argues62 this has two aspects, one dealing with correspondence and the other coherence. The correspondence principle requires the interpreter take the interpreted to be responding and describing a world both have access to. The coherence principle requires the interpreter attribute a degree of logical consistency in the interpreted and consequently assume the agent being dealt with is intrinsically rational. Interpretation requires as a prerequisite both that we are dealing with rational agents and that both of us are responding to the same stimuli.

What justifies our application of the principle of charity is Davidson’s adherence to the idea that human beings are essentially rational63. This idea is perhaps one of the few that many philosophers from Plato through Leibniz to Kant and then in the modern era Sellars and McDowell either implicitly or explicitly endorse. This constitutive ideal of rationality is where Davidson can meet the intuitions driving epistemic theories. The reason why we feel that truth is necessarily tied to our means of understanding, predicting, and describing the world is because communication and interpretation require we assume we are dealing with rational agents and thus require we attribute truth-conditions to utterances based upon our most justified, pragmatic, and coherent methods of interacting with the world.

This is where it should become apparent that Davidson’s position offers us a way of accommodating the powerful intuitions behind the Realist and epistemic theories. As things stand Davidson has characterized the concept of truth as both mind-independent and connected to justification and knowledge. The reason why Davidson is able to do this is because, unlike the traditional theories, he makes explicit the role language plays in our means of utilizing the concept of truth by taking into account Tarski’s work in formal languages, 20th century philosophy of language, and the principle of charity.

Objections and concluding remarks:

Objections to Davidson’s programme come from a few directions. It has been argued that the principle of charity, or at the very least some ramifications of Davidson’s use of it, commits us to absurdities. Foley and Fumerton64 and Dalmiya65 have argued that the idea of an omniscient interpreter, which was introduced by Davidson66 in order to respond to the worry of global falsity of interpretation, leads to the absurd consequence that Davidson’s approach can meet the sceptical worries the Realist and epistemic theories face only if the claim that there is an omniscient being that exists can be defended (Foley and Fumerton) or that Davidson’s argument ends up with a vicious circularity (Dalmiya).

Brueckner has responded the challenges above on their own terms67; however, as suggested by Davidson himself, the idea that is being conveyed in the argument from the omniscient interpreter can do better without an omniscient interpreter. The role of the omniscient interpreter is merely meant to respond to the familiar worry we addressed in the first sections of this paper, namely could it be the case that all our most researched, justified, and useful beliefs are coherent and shared and yet false?

This worry can only be given a cursory answer here. On one hand if we accept the idea, voiced by Russell, Blanshard, Sellars, and Davidson, that the relationship between belief and the world is causal, we must assume that if as a community we in general respond similarly and agree on most aspect of the world, we are in general right about the way things are. On the other, and closely related, if we are willing to grant Davidson’s rejection of the scheme-content dualism, the charge that global falsity is possible in Davidson’s account simply undermines itself by presupposing the idea of world beyond conceptualization which we have failed to properly conceptualize.

Another objection which poses concern for Davidson’s theory is that it goes too far. In the work of Frege, Ramsey, Quine, and Horwich it has been argued that truth must be taken to be conceptually basic and undefinable (a point Davidson will agree with) and that nothing more can be said about truth (a point which poses a challenge for Davidson). The arguments for the latter thesis are largely that (a) as demonstrated above the concept of radical interpretation and charity is not altogether unproblematic and (b) what about all the problems posed for Davidson’s semantic project at large?

Again I can only point out lines of response which might be pursued. Davidson will argue that deflationary theories aren’t justified in their negative claim about truth merely on grounds that it fails to account to the rather plausible intuitions driving epistemic theories. Furthermore it could be argued that deflationary theories would then need to provide an account of language which treats communication, interpretation, and truth as wholly independent, which has yet to be given.

Rorty68 often comments on the delicate balance philosophers must strike between appealing to new ways of thinking to get over old problems and retaining enough old intuitions to feel like we’ve quelled our reasons for philosophizing in the first place. His sympathies lie on the more polemic side of getting over old intuitions and moving on to new ones. Davidson’s work, although largely dependent on getting over old dogmas, does not altogether ignore the weighty historical concerns I have spent most of this paper illuminating. Davidson can respond to the arguments of Berkeley, Kant, Russell, and Gödel without needing to address one form of scepticism over another, or one form of explanation over another. While not entirely free from objections Davidson’s position has already established itself on a more tenable dialectical ground than Realist or epistemic theories.

Share with your friends:
1   2   3

The database is protected by copyright © 2020
send message

    Main page