Beth Cherryman

In what follows I will argue that Davidson’s theory of truth does not entirely solve the problem of radical interpretation.  I shall outline the problem of radical interpretation as the problem of correctly attributing belief, and explain that Davidson thinks a theory of truth is required because truth does not refer to meaning, detailed beliefs, or interpretations.  I will present three possible criticisms of a theory of truth as a theory of interpretation.  First, what the radical interpreter takes as true may in fact be the simplest coherent web of belief (and may be false).  Second, the indeterminacy of translation mitigates Davidson’s theory of interpretation.  Last, conceptual relativism may well apply to rationality.

All understanding of speech involves a radical interpretation.  To interpret the phrase ‘Es regnet’ as ‘it is raining’ requires the interpreter to go from a non-interpreting description (‘Es regnet’) to an interpreting description (the speaker is saying that it is raining)1.  Further, the simultaneous role of belief and meanings2 implies that even in the same language, using the same words, an interpretation may not align to what the speaker meant to say with his words.  Radical interpretation is necessary for understanding domestic, as well as foreign, speech1.  Without a correct interpretation we cannot correctly attribute intentions and meanings and therefore cannot provide intentional explanations.

In order to interpret a language, we must know certain things1.  Knowledge of what each meaningful expression means does not suffice.  We might understand that in German ‘Es regnet’ means it is raining, and we might know that the speaker is speaking German.  Therefore we know that the speaker said it was raining.  However, this does not inform the interpreter what it is to know what an expression means1.  Moreover, knowing the causal relation between utterances and behaviour does not suffice to interpret a language1.  This approach reduces interpretable speech to nothing but (identical with) a kind of behaviour, and this behaviour is nothing but movement of the lips and vocal cord.  This approach fails because it gives no intelligible account of what we might know that would allow us to re-describe utterances that have not been interpreted as correctly interpreted ones1.  To say the behaviour is right we would have to know the intention, which is not behaviourally accessible.  Similarly, knowing the meaning of basic elements, such as words, does not suffice to give us an interpretation of language1.  The meaning of words cannot be specified before their use in sentences1.  There is no chance of giving an account of words before giving one of sentences.  Neither knowledge of the meaning of each meaningful expression in the language, nor knowledge of the causal relations between utterances and behaviour, nor knowledge of the basic elements of a language serve for interpretation.

Davidson puts forward a theory of truth to solve the problem of radical interpretation.  The method constructs a truth definition, in the style of Tarski, in which the systematic contribution of elements of sentences to their overall meaning is revealed4.  Such a theory entails that for every sentence S of the object language there is a T-sentence of the following form1:

S is true iff p

For example, ‘Es regnet’ if and only if it is raining.  A theory of truth generates T-sentences for every sentence of the language, and they correspond in a one-to-one sentence-to-sentence relation.  Davidson has applied a common logical structure1.

The theory of truth is empirically adequate if the T-sentences are true1.  To know whether or not a T-sentence is true evidential support must be provided.  This evidential support cannot appeal to the speaker’s intentions or beliefs as these attributions of attitudes demand a theory that must rest on much the same evidence as interpretation (infinite regress)1.  Rather, Davidson suggests that the interpreter assumes the sentence he is interpreting to be true1.  Davidson thinks this is a plausible attitude to take, as the interpreter may know the speaker intends to express a truth in uttering a sentence without having any idea what truth.  Now, each T-sentence provides a true biconditional1.  The ‘S’ and the ‘p’ either side of the ‘iff’ must have the same truth-value.  So, ‘Es regnet’ if and only if ‘it is raining’ is a T-sentence, for example, if ‘Es regnet’ is true and ‘it is raining’ is true. To prevent T-sentences like ‘Es regnet’ is true if an only if ‘2+2=4’, where both sides of the biconditional may have the same truth-value, the construction takes place within a generally holistic theory of knowledge and meaning4.  The totality of T-sentences should optimally fit the evidence about sentences held true by native speakers1.  That is, we can use a true T-sentence to interpret a language only when we have an empirically adequate theory of truth for the whole language.  We must have T-sentences for all (or most) of the other sentences in the language as well.  Thus, a radical interpreter can tell when a subject holds a sentence true, and using the principle of charity (constraint to maximise the truth in the subject’s sayings) ends up making an assignment of truth conditions to individual sentences4.

Thus, by implementing a theory of truth as a theory of interpretation, Davidson contends he has evaded the trap of assuming knowledge of meanings or detailed knowledge of beliefs that the three insufficient approaches required to some degree.

One problem with Davidson’s theory of truth as a theory of interpretation is he assumes a correspondence theory of truth, in which an empirical statement is true if and only if it corresponds to the facts3.  Yet, because a theory of interpretation must deal with language, and interpretation of language is equivalent to the correct attribution of belief, when we say ‘facts’ we are actually talking about a web of belief.  When the subject utters “Es regnet” perhaps what he is expressing is his dislike of the rain, which he happens to utter whenever it is raining near him.  The radical interpreter mistakenly interprets this as ‘it is raining’ for the very reason that the speaker utters it whenever it is raining near to him.  Perhaps every speaker of the subjects language also utters ‘Es regnet’ when it is raining near them to express their dislike of the rain.  We might well dispense with a correspondence theory of truth and replace it with some equivalence between what is ‘true’ and what amounts to the simplest coherent web of belief or when it is finally useful to belief3.

Extrapolating this point, Davidson concedes there is an indeterminacy of translation.  The totality of a subject’s behaviour does not determine whether one translation of their sayings or another is the right one.  But there is nothing more than the totality of behaviour to fix one interpretation as the true one.  What the radical interpreter sees of someone’s linguistic behaviour may consistently give rise to different interpretations of the individual terms they utter4.  ‘Es regnet’ may be uttered every time it rains near the subject, but some times the subject means to inform that it is raining and at other times the subjects means to express his dislike of the rain, or that it is April, or that it is miserable outside and it is better to remain indoors, etc.  Or even the subject may utter ‘Es regnet’ when it is not raining near him to imply that someone is a terrible singer (it may be a cultural reference that bad singing brings on the rain).  These differing interpretations could infect the whole language4.  When it is the case that two different interpretations of what the speaker is saying, that is two different theories as to what the subject means to say, are adequate to the whole of experience, the question of which one the speaker really holds seems to lapse4.  This seems to mitigate the interpretive power of Davidson’s theory of interpretation (and any other).  Moreover, as radical interpretation is a problem even in the same language, meaning terms can become indeterminate.  There may be no real truth of the matter whether one uses the word ‘rain’ to refer to rain or something else, say bad singing.  Indeed this may be the case for one’s own meaning terms, which seems not to correspond with an understanding of oneself as capable of thought at all4.

Finally, the notion of conceptual relativism suggests that beliefs and meaning are too context dependent to have a truly general theory of interpretation.  Davidson accounts for limited relativity with his principle of charity, but his theory rests on the assumption there is an underlying shared basis of rationality.  Rationality is universal, but rationality is also linked with beliefs and meanings and as a consequence context dependent.  For example, religious beliefs or adherence to superstitious rituals (not walking under ladders) may be an acceptably rational way to behave is some cultures, particularly past cultures.

To conclude, the problem of radical interpretation is that language is attribution of the speaker’s beliefs, which are not directly accessible to the interpreter.  Knowledge of what each meaningful expression means, knowledge of the causal connection between utterances and the behaviour of human agents, and knowledge of the meanings of basic elements like words are all insufficient for interpretation of language.  This is because radical interpretation should rest on evidence that does not assume knowledge of meanings or detailed knowledge of beliefs.  Instead Davidson proposes a theory of truth can be used as a theory of interpretation.  That is constructing a truth definition in the style of Tarski in which the systematic contribution of elements of sentences to their meaning is revealed.  The construction takes place within a generally holistic framework.  A radical interpreter assumes a subject holds a sentence true and, using the principle of charity, makes an assignment of truth conditions to individual sentences.  Evidence makes this theory empirically adequate.  However, a theory of truth as a theory of interpretation does not seem to entirely solve the problem of radical interpretation.  First, what the radical interpreter takes as ‘true’ may in fact only be the simplest coherent web of belief.  Second, the indeterminacy of translation suggests the holistic approach may give rise to several true and consistent interpretations of a particular utterance.  The radical interpreter may not know which interpretation to go with.  It may not matter which interpretation the radical interpreter decides is correct.  Lastly, conceptual relativism may apply to rationality as well as belief and meaning.  This would infringe the scope of Davidson’s (or anyone’s) theory of interpretation.

References

1.Davidson, D (1973) ‘Radical Interpretation’ in Inquiries into Truth and Interpretation. [Courtesy of] Oxford University Press (2003).

2.Davidson, D (1974) ‘Psychology as Philosophy’ in Readings in the Philosophy of Social Science, eds. Martin, M and McIntyre, L. (1994) MIT.

3.Hollis, M (1994) The Philosophy of Social Science: an introduction. Cambridge University Press.

4.Blackburn, S (2005) Oxford Dictionary of Philosophy. Oxford University Press.