An Argument for Epistemic Purity

There is a growing number of those who smell the scent of pragmatic factors in epistemic justification and knowledge. On the former, it might be interesting to see how defenders of the view want to avoid the following argument, an argument that relies on a constraint on a theory of justification. The constraint is in the following ballpark: learning some epistemology shouldn’t undermine justification. I’ll make this more precise below.

I’ll give a first argument that even the pragmatists should be happy to accept, and then extend it in a way they’ll have to reject. After that, I’ll give a couple of ways of resisting the argument.

Regarding the epistemology of justification, a theory will delineate factors that supply justification. In delineating factors, a theory will have to be sensitive to various constraints, one of which is the problem of the truth connection. The kind of justification epistemologists are interested in is one that is appropriately related to truth. It is controversial which approach to the truth connection is correct, but we need not settle that for the purposes of my argument.

Suppose then that we have a theory that claims that F is the set of factors that supply epistemic justification for a specific kind of belief B. Then take minimal subsets of these factors, subsets that generate a minimal amount of justification needed for the claim in question (according to the theory), and imagine a person S who believes the claim on the basis of some such subset G.

To each member of this particular subset there corresponds a set of propositions involves claims to the effect that each member of G is a truth indicator for B (on whatever is the correct account of the truth connection regarding epistemic justification–maybe it’s reliability, maybe it’s probability or likelihood of truth, maybe it’s my preferred view discussed here).

Here’s the application of the principle that learning a little epistemology shouldn’t undermine justification. In particular, I think two claims are true:

1. Add belief in the truths in G, and the justificatory status of B won’t decrease.

2. For the false ones, add the belief that they are false and justification decreases.

Clause 2 is the important one, since pragmatic factors are not truth indicators (if they are, then the desideratum that a good epistemological theory satisfy the truth connection has no bite to it). Once S adds the belief that such factors are not relevant to getting to the truth, that additional belief constitutes an undercutter for the set (remember that the set is minimal, so that if the justifying power of any member is blocked, the rest of the set can’t sustain justification by itself), and thus prevents justification from obtaining.

So pragmatic factors violate the condition above, if we make the claim precise as follows: factors that make for epistemic justification shouldn’t be able to be undermined by learning whether or not they satisfy the truth connection.

For various reasons, this is an argument only a foundationalist could love. More carefully, coherentists need to reformulate it to preserve the holistic character of justification they favor, and to do so, they will want to reject the idea of anything less than a holistic system supplying justification for a belief. Here’s how they can do it. Instead of talking of factors that supply justification, they need to talk of factors that do not by themselves supply justification, but are a non-redundant part of a larger whole that is itself unnecessary for justification but sufficient for it. Mackie fans will recognize this as a description of INUS conditions, which Mackie used to define causation. Holistic coherentists can’t abide the idea of the factors in question generating a bit of warrant for a belief, but they can talk about such factors being INUS conditions of warrant or justification.

Their holism also prevents them from requiring that each individual INUS condition satisfies the truth connection, but they have to say something. So here’s what they can, and should, say: the larger condition of which a given INUS condition is a part satisfies the truth connection, which it does only when each of the INUS conditions included in it is subject to no undefeated defeaters, either rebutting or undercutting defeaters.

One undercutter of an INUS condition is the claim that the larger condition minus this particular condition would do just as well, or better, at satisfying the truth connection. That is, one defeater is the claim that this present condition is irrelevant to the satisfaction of the truth connection by the larger condition.

So, now we have a collection of factors conceived, not as items that supply justification for a given belief, but as INUS conditions of justification for that belief. The particular collection of factors is supposed to be minimal, meaning that if the contribution of any member is defeated, then there is insufficient justification for the belief in question. And one way of neutralizing a factor in this collection is to come to believe that it contributes nothing to the satisfaction of the truth connection.

So we have two arguments, one fit for the foundationalists among us and a more general one that even a coherentist can embrace. So the latter argument is the more general one, of which the first one is an instance in virtue of involving some distinctively foundationalist claims.

There are two things to notice from the pragmatists’ point of view. The first point is that the first argument doesn’t pose a threat to their view. That argument would undermine their view only if they thought that pragmatic factors supply or generate justification for a belief. But pragmatists do not think that. They can allow that purely truth-related factors supply or generate justification, but that pragmatic factors come into play in determining how much justification these factors must supply.

That only works, though, with the foundationalists’ argument. When we look at the more general argument employing INUS conditions, things are different.

Suppose the role of practical matters is to raise or lower the amount of evidence one needs for justification. That is, suppose a given theory claims that a pragmatic element plays a role in determining what counts as a minimal subset composed of INUS conditions in a given set of circumstances. Is the obtaining of these pragmatic elements itself a member of such a minimal subset?

If so, there’s a problem, since they are not truth-conducive. Since they are not, we can imagine an epistemological learner adding the belief that they are not, and inferring that a member of the minimal subset G has been neutralized, and hence that the remainder of G is no longer a minimal subset. But that violates the condition that epistemological learning shouldn’t undermine justification, just as above.

Not surprising, since the pragmatists will want to keep the account of justification separate from the pragmatic factors that simply set how good a justification needs to be to justify a belief in a given set of circumstances. So we must assume that the obtaining of the pragmatic elements is not itself a member of G.

But remember what constitutes G: INUS conditions of justification for the belief B, where G is a minimal set of such. The obtaining of the pragmatic factors is Insufficient for justification–that’s obvious. This obtaining is also Nonredundant–without this obtaining, we’d lose our explanation of whether B is justified. There is also a larger condition that is Unnecessary for justification, but also Sufficient for it–but sufficient only because it includes the obtaining of the pragmatic factors in question. So, the obtaining of the pragmatic factors is an INUS condition for justification for the belief B, and as such, is a member of G.

So, bottom line, what does the above show? I think the most obvious lesson is that you’ll have to make some theoretical commitments in order to defend pragmatic encroachment. One way is obvious: endorse foundationalism. There is another way too. If one adopted F&C-style internalism, so that only mental states can explain justificatory status, then the INUS conditions a coherentist must appeal to must be mental states as well.


An Argument for Epistemic Purity — 2 Comments

  1. First of all, Jon, there must be something weird here, in that you end up claiming that a milder version of a theory has entailments (the rejection of the epistemology-shouldn’t-hurt principle) that a stronger version (i.e., mild version-plus-theoretical-commitments) does not. That sort of thing is possible with various kinds of nondeductive inferences, especially inference to the best explanation, but the argument here looks like it’s meant to be pretty much deductive. So I have to wonder whether you’ve misconstrued the state of play that obtains after we consider your two arguments. Basically, since it seems that some further theoretical commitments are consistent with pragmatism, and others are not, one might just as well place the burden on the purist to articulate more clearly what commitments would rule out pragmatism.

    If that’s right, then I think the place to look to diagnose this oddity in your argument has to be your claim that the coherentist-friendly argument is strictly less theoretically burdened than the foundationalist-flavored version. Here’s a possible commitment that the proponent of the former might have that the latter might not: that it is possible to formulate INUS conditions for being justified that include only truth-conducive conditions. I don’t think that the pragmatist should deny that one could formulate such conditions for ‘having justification to such-and-such a degree’ — but they can & should deny that one could do so for the threshhold-involving concept ‘is justified’. The pragmatist might argue that, since the epistemically-pure materials by themselves do not include any machinery to set that threshhold, there just cannot be any larger condition that fails to include a pragmatic, threshhold-setting condition. There just are no minimal subsets G of INUS conditions for being justified that do not include some merely practical (i.e., non-truth-conducive) machinery.

    One possible direction to upshoot from here would be to go towards a kind of skepticism or don’t-care-ism about ‘is justified’: all that matters, epistemically, is having degrees of justification, but there’s no epistemic point to the further question of having sufficient justification. This is not an unheard-of view in the literature. But another direction one might go would be to inspect the original epistemology-shouldn’t-hurt principles, and see if they can reasonably be weakened. If we re-gloss “learning some epistemology shouldnâ??t undermine justification” as “learning some epistemology shouldn’t reduce your degree of justification”, then the pragmatism you’re considering here is completely consistent with it.

  2. Jonathan, I don’t think I said that the two versions of the argument were related so that one was logically stronger than the other. If I did, I recant! The foundationalist argument has implications that the coherentist will find unacceptable, so the coherentist needs to reformulate the argument so that it lacks these implications. That doesn’t make the foundationalist argument imply the coherentist one, and that’s the sense of “logically stronger” in which there’d be a problem.

    The version of pragmatism you describe in the second paragraph is just the version I was thinking of. And it needs either to find a mistake with the INUS condition argument (such as denying the “learning some epistemology” principle or by restricting the kinds of INUS conditions that can be used in the theory.

Leave a Reply

Your email address will not be published. Required fields are marked *