Naturalism and Epistemic Norms

Here's a question which I've been thinking about for a while but which I got a bit clearer on this weekend: have we successfully naturalized epistemic normativity if, for every epistemically normative claim, we can identify the fact which makes true, and it turns out to be a natural fact?

I'm tempted to say yes. The view could be that (for example) claims like:
(1) S's belief that p is epistemically correct
are made true by the same facts as claims like:
(2) (S believes that p and) given S's information, p is probably true.

This sort of view appears to raise an analogue of the Open Question Argument. It could be objected that the normativity of (1) is not adequately captured by (2) because one can agree that p is probably true given S's situation, but still wonder whether S's belief that p is epistemically correct.

One can respond that it need not be part of the view that (1) means the same as (2), and that therefore the identity between the facts which make these two claims true need not be at all obvious. If there is meant to be more to the objection than this, it's not clear what it is. To insist that the question of the correctness of S's belief remains genuinely open when you know that (2) is true is just to beg the question against the view that the fact which makes (1) true is the fact that makes (2) true.

Epistemic Value and Epistemic Norms

I'll be off to Stirling this weekend for a workshop on Epistemic Value. The main focus seems to be on the value of knowledge, but I'm hoping there's going to be plenty of discussion of issues around epistemic normativity too, which I'm interested in at the moment - I'm currently drafting a paper on why it looks to be one of the more difficult kinds of mental normativity for naturalists to accommodate (although there's still reason to be optimistic that they can do it).

The thought is that a line naturalists might take with normativity claims in general is that they are claims about the maximization of some value or other which can be naturalistically characterized. (Not that this line's not fraught with difficulties, but I think it has some appeal as a strategy.) The trouble is that it's difficult to recast claims of epistemic normativity in a simple way as claims about the maximization of some value (e.g. true belief), as I note in another current draft paper, which delves into epistemic consequentialism as part of one of its attempts to see how Crispin Wright's views on entitlement can be motivated. However, I also argue there that a subtle epistemic consequentialism might have something going for it (though it wouldn't rescue Wright-style entitlement as a form of warrant), and I now want to argue that this subtler consequentialist view might be appealed to as the basis for a naturalized account of epistemic normativity.

Experience and Concepts

During a helpful conversation with Amber Carpenter today I realized that an argument from Kant that I've been interested in for a while may have an analogue in Plato.

The Kantian version of the argument is that experience cannot provide grounds for (certain of) our most fundamental concepts, since we need those concepts to be in place before we can have experiences of the right kind (this is most prominent in the Transcendental Aesthetic). The Platonic analogue (I'm looking at Phaedo 74c-75c here) is that experience cannot provide grounds for our grasp of (certain) forms, since whenever we experience something(s) as possessing the relevant quality or standing in the relevant relation we compare it/them (unfavourably) with the form itself, which wouldn't be possible unless we already knew the form.

Plato seems to take it that the forms in question are real, although they cannot be empirically known. Whereas Kant, believing that certain basic concepts cannot be empirically grounded, concludes that they do not correspond to (transcendentally) real features of the world.

(For the record, I myself am tempted to think both that our basic concepts are empirically grounded and that they reflect real structural features of the independent world.)

Don't Give Up On The Given

Larmore (2002, ‘Attending to Reasons’, in N. Smith (ed.) Reading McDowell: On Mind and World, London: Routledge, pp. 193-208) writes:

.. the Given cannot serve as a justification for a belief, if it lacks conceptual articulation; yet to the extent that it is shaped by an understanding of the world we already possess, it cannot count as simply "given". (p. 196)

Larmore may be right that something like this is supposed to be 'the well-known trouble with [the] "foundationalist" approach'. But I don't see the force of the dilemma. We can accept the truth of the first horn, the claim that conceptual articulation is required in order for sensory input to justify a belief. But we do not thereby impale ourselves upon second horn. For provided this conceptualized sensory input is ‘shaped’ by an ‘understanding of the world’ which is itself shaped in response to what is given to us in (unconceptualized) sensory input, there is no reason to think that this understanding in any way distorts our experience of the world. Rather, we can hope that the conceptual shape of our conceptualized sensory input reflects the structure of the world.

Why We Do This

There's an article here on academic blogging which people reading this might find interesting. (Thanks to Daniel for the pointer.) I agree with many of the sentiments expressed, though if 'younger scholars' are really worried that 'blogging would eat up time that could be devoted to publishing articles or working on a book', that would seem to be a mistake - at least in philosophy (although I would have thought in other discipines too) ideas floated as blog entries can later become bits of papers or books, having benefited from commentators' feedback. (Or that sort of feedback could help you realize that an idea isn't worth pursuing sooner than you would have otherwise - also very useful if you want to spend more time getting your better ideas into print ...) Blogging isn't 'conventional academic writing' but it's capable of being as much a preamble to it as a good seminar is. Particularly if you're someone who finds (as I do) that making yourself write something down is a good way of getting it clear(er).

OK, that's enough self-justification. Back to my conventional academic writing.

Restricted Precisification

Something I've wondered for a while is whether supervaluationists are right to think of precisifications as ways of making *every* term in our language precise. Instead, mightn't it be helpful (at least sometimes) to talk about precisifications-for-a-term, i.e. ways of making *that term* precise? Here are two prima facie reasons to think so.

1. Robbie Williams shows in a forthcoming paper that on the usual supervaluationist story there is a problematic interaction between two kinds of vagueness: how-tall-must-it-be-to-be-a-mountain vagueness and problem-of-the-many vagueness. Suppose we resolve the problem of the many by saying that on each precisification just one of the Kilimanjaro candidates is a mountain. Then there is nothing which is a mountain on every precisification. But that undermines something supervaluationists (and the rest of us) typically want to say about how-tall vagueness, namely that some things are definitely mountains.
We could deal with this if we said that there are some things (Kilimanjaro-candidates) which are definitely mountains where what this means is that they are mountains on every acceptable way of making precise how tall something has to be to be a mountain, but which are not definitely mountains where what this means is that they are mountains on every way of specifying precisely which of the Kilimanjaro-candidates counts as a mountain.

2. Precisifying everything at once makes it always false to say of a vague term that it is definitely vague. (No term is vague on every complete precisification, because no term is vague on *any* complete precisification.) But we can truly assert that a term is definitely vague if what it means is that *on all precifications of 'vague'* the term in question counts as vague.

If precisification worked this way, 'Definitely ...' would function differently on different occasions of use. It would always have the same truth-conditions as 'On all precisifications of the salient term(s) ...', but which term(s) are salient will change. Presumably context would help us determine which term(s) are salient on any particular occasion.