In previous posts, I discussed the problem generated by the case of Jack and Jill. When Jill’s fear influences her visual experience that presents Jack as angry, does Jill get as much reason from her experience to believe her eyes, as she could if her fear didn’t influence her experience? The problem is that both Yes and No can seem plausible. In this post I focus on potential reasons for answering No.
If the answer is No, then something about the relationship between Jill’s fear and her experience has to explain what weakens the epistemic power of Jill’s experience. In answering a question raised by John, I noted two potential explanations.
Defeat: The influence on Jill’s experience gives her a defeater, where a defeater is a consideration Jill has access to, and that reduces the experience’s power. For instance, Jill is in a position to notice that her experience is congruent with her fear, so perhaps that should give her pause.
Reliabilism: Jill’s experience is formed in an unreliable way, on the assumption that it’s not generally truth-preserving to believe that what you fear is actual.
A more subtle explanation distinguishes between two kinds of epistemic power. First, there’s the power an experience (perhaps together with other factors) might have to provide a reason for a belief. An experience (and its helpers) might harbor this power, independently of whether the subject ever forms a belief on their basis. We can call this kind of power ‘reason-giving power’.
Second, there’s a power the experience might have to (help) make it the case that a belief formed on the basis of that experience is ill-founded (or well-founded). We can call this second kind ‘forward-looking power’, since it is defined in terms of the epistemic status of a belief formed in response to the experience.
The subtle explanation for answering No then says that Jill’s experience retains its reason-giving power but loses (only) its forward-looking power. The influence of her fear renders the reason-power un-usable.
Un-usable reason-power: Jill’s experience retains its power to provide a reason for subsequent beliefs, but this power can’t be used to make subsequent beliefs well-founded.
In think there are strong objections to all three of these explanations. As I mentioned in replying to John’s question on the previous post, the idea that the cases must always generate a defeater has the consequence that every time you notice that an experience is congruent with what you antecedently wanted, feared, or suspected, it would lose epistemic power. And the proposal that reason-power is un-usable waters down the notion of reason-giving power so much that it seems fair to question whether there is any epistemic property the notion is meant to be tracking.
Reliabilist proposals seem to lack the resources to identify the problem with Jill’s confirming her suspicion that Jack is angry. (To keep things simple, let’s consider a case where her fear goes with a suspicion). Compare a case where Jill starts out with belief X, and then infers belief Y. Then she uses belief Y to strengthen her original belief X. Although this process of strengthening belief X is circular, it seems just as truth-conducive as the process by which belief X was formed originally.
Inferentialism is a fourth potential explanation. The three other explanations outlined here challenge inferentialism indirectly, by providing alternatives to it. But I find challenges to the very coherence of inferentialism much more powerful than the objections to it inherent in alternatives. That’s why I devote more time in the book to making sense of what inferentialism could be. How could experiences be conclusions of epistemically appraisable inferences? And how could inferences issue in epistemically appraisable experiences?
One question might be whether there is additional evidence available to Jill if she attends more to these perceptions, She might decide her initial impression was faulty. In the case of the literature on unconscious bias, one idea is to increase self-awareness to improve behavioural outcomes.
Thanks, David. It’d be interesting to find out which modes of self-awareness improve behavioral outcomes and which don’t.
Your comment raises a really interesting epistemological question. Suppose Jill is disposed to decide that her initial impression was faulty if she considers the matter, but she doesn’t actually consider the matter. Is her merely having the disposition is enough to weaken the power of her experience? If it is, that would be another potential alternative to inferentialism. It wouldn’t speak to the cases where Jill lacks the disposition, though.
Hi Susanna, that’s an intriguing point about reliabilism not having sufficient explanatory resources to adequately accommodate the No answer. I wonder if you don’t mind saying a bit more about that.
It does seem that the issue for reliabilism remains even if we add some resources to it beyond truth-conduciveness, such as factoring in for epistemic luck. On the one hand, there may be an element of bad luck in Jill’s situation which the reliabilist could exploit to accommodate the weakening of epistemic power. But on the other hand, the sort of way in which Jill is (un)lucky may very well be the norm when it comes to perceptual beliefs. So there does seem to be an interesting question of how the reliabilist might go about differentiating the varieties of truth-conducive processes.
Hi Katia, thanks for your comment. If there are cases where a process that the reliabilist would count as justification-conferring is also a case like Jill’s, then that suggests they lack the resources needed to analyze the case.
I wasn’t sure what the element of bad luck in Jill’s situation was, on your reading of it. Are you thinking of Jill as being subject to bad luck when her fear influences her experience? And is it bad luck because her fear might not have influenced her experience, or because she might have had the same experience without fear influencing it? Those contingencies do seem like a kind of bad luck. Interested to hear more about what epistemic significance you think they have.
If the influence of fear on experience is a form of epistemic bad luck, inferentialism provides a way to identify the bad-making features of those contingencies. “Bad luck” seems like a label for a contingency that’s bad, but on its own it doesn’t seem to identify the bad-making feature. So I’m not sure that appealing to bad luck gives the reliabilist an extra resource for identifying the bad-making feature. It’s more like a label for a phenomenon needs a theory of what the bad-making features are.
Thanks for your response, Susanna. I had in mind exactly the sort of “bad luck” you mentioned – various contingencies that line up in an unfortunate way. One bit of bad luck for Jill is that her fear influences her experience, but another bit of bad luck could be that she formed the (unjustified) belief that Jack is angry at her in the first place. There may be a third place for bad luck to enter in, if we think that Jill had a chance to figure out that something has gone wrong from the first-person perspective but didn’t (if, for instance, she routinely and knowingly assumes Jack to be angry).
But as you say, enumerating way of being (un)lucky may not be adding genuine explanatory resources to reliabilism. It allows the reliabilist to claim that a generally truth-conducive process may deliver misleading results on specific occasions, but that’s not yet explaining why. Indeed, calling the bad-making feature in Jill’s case “bad luck” may preclude a satisfying explanation of it; after all, it’s just bad luck, or so the reliabilist story would go. So inferentialism is in a much better explanatory position.