Rambling thoughts on rambling topics


World, causality, belief

In the ‘Afterthoughts’ to ‘A coherence theory of truth and knowledge’ Davidson makes clear that he agrees with Rorty that his concept of ‘mild correspondence’ is better characterised as a form of pragmatism (although, he hastens to add, not in the particular sense of Rorty …) In later papers Davidson is even more explicit and argues that truth is such a fundamental concept that any attempt to define it is bound to fail. Cf., his 1990 paper ‘The structure and content of truth’, and the 1996 paper ‘The folly of trying to define truth’. What a truth theory such as Tarski’s does, Davidson maintains, is provide us with minimal conditions that our use of the concept should satisfy. In that way a truth theory is part of a broader theory of rationality, but, and this is crucial for the (later) Davidson, truth cannot be defined, and hence, neither can rationality. These concepts can be circumscribed, we can lay down rules for their proper use, but strict definition is impossible.

It will be clear that from that perspective any theory that appeals to ‘truth makers’ is on the wrong track. It is already quite far removed from the position that Davidson takes in ‘A coherence theory of truth and knowledge’ and he has gone further in the opposite direction thereafter. For a proper appreciation we need to take into account the role a truth theory is supposed to play. It is not to define truth, but it is to lay down rules that we are bound to when dealing with truth (as long as we want to count at being rational in our dealings with it). From that perspective, assigning truth conditions to atomic sentences is not making any claims about the relation between such sentences and particular aspects of the world. That part of the truth ”definition” (scare quotes used deliberately) only serves to provide a basis for the clauses that deal with connectives and quantifiers (and perhaps other logical constants). It is in the latter clauses that real constraints are being formulated: to be rational is to accept that A and B is true iff A is true and B is true; it is to reject A if and only if one accepts not-A; and so on. But in the atomic case no such constraints are forthcoming. At this point a theory of truth is unhelpful, and something else should take its place. And that, according to Davidson, is to be found in a pragmatist outlook.

Note that this concerns the role of a truth theory vis à vis our concepts of truth and rationality. The case of a theory of meaning is different: when I am given truth conditions for atomic sentences of a language that I do not understand, as part of a theory that is formulated in a language that I do understand, I do learn something. But I learn something about meaning, not about truth. This will be clear from the fact that no information is forthcoming if the theory is formulated in a language I do not understand, or if the theory is about and formulated in the same language.

Note that the empirical content of, e.g., my belief that the entity over there is a dog does not derive from my trusting my senses, it is reality itself that is the cause of me having that belief, and thus also of what is the content of the belief. Trust in my senses comes in, in exceptional cases, if this belief conflicts with some other beliefs and if one of the ways of removing the conflict is to assume that at that specific point my senses were not, as they usually are, ‘transparant’. Recall that Davidson does not want any epistemic intermediaries, neither language, nor conceptual schemes, not perceptual capacities. Thus, there must be a reason for doubting the belief that the thing over there is a dog, and this reason will always be another belief (or set of such). Why and how a belief that is caused by reality may shed doubt on another belief, that by default we assume is also caused by reality, is a complex question that has different answers in different circumstances. This is were rationality of procedures kicks in, and although some of these procedures are ‘hard wired’ in our language (logical consistency, and so on), others may be of a more acquired and changeable nature, that reflect changes over time in, e.g., what counts as proper scientific procedure. I think that Davidson, although he focusses on the former, should be able to allow for the latter as well.

Martin Stokhof
from: Radical Discussion Board
date: spring 2020