38

Intuitions without the Understanding So we appear to have many reasons for doubting that the understanding can play the roles that Sosa has assigned it, either in our theory of the nature of intuitions or in our account of their justificatory powers. Can we do better? I will try here to sketch the outlines of an alternative view. We are agreed that an intuition is an intellectual seeming: a state that non-sensorily presents a proposition as true. The phenomenon of a proposition’s intellectually seeming true to you is undoubtedly real. To use an example of George Bealer’s: When you first consider one of de Morgan’s laws, [for example, (3) Not (p and q) is equivalent to (Not p) or (Not q)] often you draw a blank; after a moment’s reflection, however, something happens: it now really seems obvious to you.14 Elijah Chudnoff’s (2013: 50) example can elicit a similarly compelling impression. (4) Two circles can have at most two common points. But it is very implausible that its so seeming simply consists in some species or other of assent. Rather, the phenomenon seems pre-doxastic: it compels assent from you, and explains why you are attracted to assent to it. It’s not the assent or the attraction to assent itself. And it may well be that there is no good reductive account of this phenomenon. Why should there be, when there are so few good reductive accounts of other phenomena? (p.146) As these examples bring out, in addition to there being conscious episodes in which a proposition seems true to you, there are conscious episodes in which a proposition seems obviously true to you, conscious episodes in which a proposition seems necessarily true to you, and, indeed, conscious episodes in which a proposition seems both things to you at once. Often, though not always, when we report on having the intuition that p we mean not merely that it seems true to us that p, but that it seems obvious to us that p is (necessarily) true. I think this is probably what Gödel had in mind in talking about certain propositions (he had in mind some of the axioms of set theory) that ‘‘force themselves upon you’’ when you contemplate them. They are not forced upon you by sensory evidence, or by following easily from other things that you already recognize yourself to have reason to believe. Rather, when you contemplate one of these propositions, you cannot help but have the impression that it is obviously true.15 A Priori Justification without the Understanding This, then, is the somewhat minimalist account of intuitions that I favor. An intuition is an intellectual seeming: there need be no further reductive characterization. Such an intellectual seeming could be one of obvious truth; or one of necessary truth; or just a seeming of plain truth; or of some combination of these different types. And in all these flavors, the intuition can come in gradable quality: it can be more or less strong or more or less vivid. The question now before us is whether we can account for the justificatory role of such states without appealing to the understanding. We have noted at least four features of that justificatory role that need to be accounted for: (5) Intuitions are treated as data—that is, they are treated as providing justifiers that are themselves beyond justification. (6) Intuitions are regarded as providing a priori justification. (7) Intuitions are regarded as providing such strong justification that they are capable of overturning entrenched and highly justified theories. (8) Intuitions are regarded as a source of reliable truth. By contrast: (9) Intuitions are not regarded as infallible or indefeasible. (10) If p is obvious to someone, it doesn’t follow that everyone else will find p obvious. (p.147) (11) If p is obvious, it doesn’t follow that p cannot also be supported by argument. Intuitions can mislead; and they can be either defeated or supported by further considerations. All that is perfectly compatible with their having features (5)–(8).16 What we need to explain is not why intuitions are infallible, but only why they can be the source of enormously strong, but defeasible, evidence. Along with many others, I have emphasized that there is no mystery about why a seeming, sensory or otherwise, can be the source of some prima facie justification. If we cannot start with seemings, we cannot start anywhere. If it seems to me that p, I am prima facie justified in believing that p. That’s how it is with vision. And that’s how it is with intellectual seemings, too.17 Why should the justification provided by intellectual seemings be thought to be a priori? The short answer is that when we say that a justification is a priori we mean that its source does not depend on perception (broadly understood) but may depend on intuition or conceptual competence. The trickier question is what epistemologically interesting principle underlies the distinction illustrated by these lists. That is a much harder question to which I won’t try to give an answer on this occasion.18 The crucial question right now is whether this view is able to explain why intuitions are accorded as much weight as they are. How could a deeply entrenched view like the JTB view of knowledge be overturned on the basis of its simply intellectually seeming to us that Mr. Smith, in a hypothetical scenario, has a justified true belief but does not know? Part of the explanation is already contained in the fact that the justification that an intuition provides is foundational, so not dependent on any other beliefs for its plausibility. But that’s not a full explanation by itself, since justification can both be foundational and, in a given context, quite weak. (For example, with respect to the proposition that there are some tomatoes in front of me, vision might be thought to supply quite strong justification, touch somewhat weaker justification, smell perhaps weaker still, and hearing very little.) When we get strong justification from our intuitions, I believe that the reason is nearly always that those intuitions are impressions not merely of truth but of obvious truth. If I take it that p is obvious, then I’m taking it that any competent, rational person would be willing to consent to p on no evidence at all. I may be wrong to take p to be obvious; I may even be wrong to take it to be true. But in the absence of defeaters, I am prima facie entitled to believe p and to insist that it should take a great deal to defeat p. That just is what it is to find p obvious. So, unless we are to question the probative value of finding certain propositions to be obvious, we seem to have all the explanation we need. (p.148) This brings us, finally, to the question about reliability. According to the view I am defending, we have impressions of obviousness that help us reasonably answer such questions as: (a) Is the Twin Earth scenario metaphysically possible? (b) What would the extension of ‘water’ have been on Twin Earth? And we take it that these impressions of obviousness are reliable. How could we be justified in taking these intuitions to be reliable? Even though I am not a reliabilist about justification, I take this question seriously. Even if reliability is not what makes our judgments justified, evidence of the unreliability of a given putative source of justification can undermine the justification that the source is presumed to provide. Do we have evidence of the unreliability of our intuitions? Philosophers who have engaged in empirical work on intuitions say yes. I think this work is flawed, but will not engage it here. I am more concerned about an a priori argument that purports to show that our intuitions could not be reliable because there could be no plausible mechanism in virtue of which they could be reliable. The worry here is just an application of Hartry Field’s version of Paul Benacerraf’s problem about mathematical knowledge, on a Platonistic view of mathematics. We assume that the subject matter of mathematics is abstract. We also assume that we have a reliable capacity for forming true mathematical beliefs. But how could we have such a capacity on a Platonistic view? Wouldn’t such a capacity necessarily involve our being able to track the mathematical facts? But how could we possibly be doing that, if there could not, in the nature of things, be a causal channel open between us and the abstract mathematical facts? This worry applies equally to facts about logic, modality, epistemic rationality, and morality. And it raises a quite general concern, which applies equally to intuitions: how could any capacity we have, including that of intuitions, be a reliable guide to facts in these domains? The dialectic here is that the skeptic thinks he has a proof that would show, on completely a priori grounds, that states of intuition could not reliably track the modal and abstract facts. Sosa, as we have seen, deals with this challenge by postulating that the relevant abstract facts are somehow or other encoded in our understanding of concepts, so all we have to do is exercise our capacity to say what is in our concepts. But he doesn’t explain why we should be confident that the relevant facts are encoded in our concepts; nor how it is that we access what is in them. In any case, we have seen that even if those problems could be overcome, explanations that lean entirely on our understanding of concepts cannot provide a sufficiently general explanation of a priori knowledge. If our reliability cannot be explained entirely in terms of understanding, how could it be explained? Recall, what’s needed is a ‘‘proof of concept,’’ not an actual detailing of (p.149) the mechanisms that underlie our putative reliability. We need to show that there is a possible story that could explain our reliability, consistent both with the absence of a causal channel between the subject matter and us and with scientific scruple. I think that there clearly is such a story. On the kind of explanation I have in mind, the problem of reliability is split into two parts: on the one hand, into a scientifically respectable account of why we have certain concepts, and judgments involving them; and, on the other, into a scientifically respectable account of why, given that we have such concepts and make such judgments, we would be reliable about them. Once the problem is split up in this way, we can easily imagine a scientifically respectable answer to each of its two parts. Thus, it seems plausible to suppose that having the capacity to think about what follows from what would be evolutionarily advantageous. And it also seems plausible to suppose that once we develop the capacity to make such judgments and deploy them in our thinking, that it would be evolutionarily advantageous for us to be at least fairly reliable in how we arrive at them. If these judgments were arrived at on the basis of intuition, it would be evolutionarily advantageous for our intuitions to be at least fairly reliable. Obviously, there are many outstanding issues for such a style of explanation. But its availability is sufficient to refute the suggestion that we have an a priori proof at hand that shows that there can be no scientifically respectable explanation of the reliability of our intuitions.19 Conclusion I am increasingly inclined to the view that we cannot adequately explain a priori justification without appeal to intuitions. If such an appeal to intuitions is to help, it must provide epistemological resources that go beyond those provided by explanations in terms of epistemological analyticity (appeals to our understanding of concepts). Accounts, like Sosa’s, which reduce intuitions to attractions to assent, and which give the understanding an indispensable role in explaining the justificatory powers of such attractions, cannot provide such a resource. As a result, such accounts must be rejected. I have briefly presented an alternative account of these issues, one that seems to me to hold greater promise.