28

V. RATIONAL INSIGHT AND CARROLLIAN CIRCULARITY Even if we waived worries about the very cogency of the capacity for rational insight, however, it’s not clear how rational insight into the validity of MPP could help vindicate Simple Inferential Internalism, and this for reasons that are reminiscent of Lewis Carroll’s famous note What the Tortoise Said to Achilles.' 13 According to Simple Inferential Internalism, a thinker's move to (3) on the basis of (1) and (2) counts as justified only if he is in a position justifiably to believe that his inference is valid. Let us suppose, for the sake of argument, that a capacity for rational insight explains how he is in a position justifiably to believe that, if his MPP premises are true, then so must be his conclusion. How might this explanation go? For obvious reasons, it's not plausible to think of this capacity for rational insight as operating on individual inferences one by one, generating for each of them the insight that if its premises are true, then so is its conclusion. Rather, we suppose that rational insight equips the thinker to arrive at the wholly general insight that MPP is valid, that is that: For all p, q: Necessarily: If both p andp q’, then q. Now, however, we need to ask how such a justified belief in the validity of all arguments of the form MPP could help a thinker be justified in performing any particular MPP inference, for example, the Argerich inference with which we began. To bring this knowledge to bear on the justifiability of that inference will, it would seem, require the thinker first to establish its relevance to that inference, by reasoning as follows: (i) Any inference of the form MPP is valid. (ii) This particular inference, from (1) and (2) to (3) is of MPP form. Therefore, (iii) This particular inference from (1) and (2) to (3) is valid.14 Rational insight, we are conceding, gets us as far as the general propositional knowledge that all arguments of MPP form are valid. However, to bring this knowledge to bear on the justifiability of any particular inference will require the thinker to be able justifiably to infer the validity of that particular inference from the validity of all arguments of MPP form. And this will require him to be able to reason according to MPP justifiably. Now, however a fatal circularity looms. To infer from (1) and (2) to (3) justifiably, I must be able justifiably to believe that the inference from (1) and (2) to (3) is valid. To be able justifiably to believe that this inference is valid, I must be able justifiably to infer that it is valid from the general proposition that all inferences of its form are valid. To be able justifiably to infer that it is valid from the general proposition that all inferences of its form are valid, I must be able justifiably to infer according to MPP. So, on the picture on offer, my inference from (1) and (2) to (3) will count as justifying only if I am already able to infer according to MPP justifiably. The very ability we are trying to explicate is presupposed by the internalist account on offer. At this point, an internalist might be tempted by the following thought. So long as we are being so concessive about rational insight, why can’t we grant thinkers rational insight into the validity of specific inferences, and not require that this be derived from some general knowledge of the validity of all inferences of the form MPP. Perhaps this more general knowledge could be arrived at later, by using the knowledge gained through these acts of individual insight? It’s important to appreciate that this manoeuvre will not help overcome Simple Inferential Internalism’s difficulties with the problem of circularity. For, once again, we can ask how my knowledge of the validity of the inference from (1) and (2) to (3) is supposed to bear on my warrant to infer (3)? According to Simple Inferential Internalism, this inference will be justified only if I am able justifiably to believe that that my premises provide me with a good reason for drawing the conclusion. But it is very hard to see, once again, how my putatively justified judgment that my premises entail my conclusion could bear on my entitlement to draw the conclusion in anything other than inferential form, thus: (iv) This particular inference from (1) and (2) to (3) is valid. (v) If an inference is valid, then anyone who is justified in believing its premises and knows of its validity is justified in inferring its conclusion. Therefore, (vi) Anyone who is justified in believing the premises of this inference is justified in inferring its conclusion. (vii) I am justified in believing the premises (1) and (2). Therefore, (viii) I am justified in inferring (3).15 Even if we conceded, then, that we have rational insight into the validity of specific inferences, we do not escape the threat of circularity that afflicts the internalist account. Once again, an ability to infer justifiably according to MPP is presupposed. Commenting on an earlier presentation of this argument, Crispin Wright observes: It is clear how the simple internalist must reply to Boghossian. To staunch her view against all threat of Carrollian regress, she must insist that recognition of the validity of a specific inference whose premises are known provides a warrant to accept a conclusion not by providing additional information from which the truth or warrantedness of the conclusion may be inferred, but in a direct manner … In effect, and paradoxically, her view must be that warrants acquired by inference are, in a way, a subspecies of non-inferential warrant in general: that an appreciation that a conclusion follows from warranted premises confers, when it does, a warrant for an acceptance of that conclusion in no less direct a fashion than that in which a visual appreciation of the colour of the sky confers warrant for the belief that it is blue.16 I agree with Wright that this represents the only escape route available to the simple internalist. Wright himself does not endorse it, or even present it as an especially attractive option. He claims for it only that it remains undefeated by the sorts of considerations adduced so far. No doubt there remains scope for discussion. But it is very difficult to see, it seems to me, how the inferential case is to be plausibly assimilated to the admittedly non-inferential warrant provided for the belief that the sky is blue by the observation of a blue sky under favourable circumstances. Admittedly, we don’t have very refined ways of deciding when a warrant for a particular belief is direct and when it is fundamentally inferential in nature. What we seem to operate with is a rough-and-ready criterion which says in effect: when the gap between the content of an apparent observation and the content of the belief that it is supposed to justify is too large, the justification must be inferential in nature, even if that may not be apparent from its presentation in everyday conversation. So, for example, in response to the question How do you know it is going to rain?' I may simply point to the dark and threatening clouds. But as everyone would agree, the observation of the dark and threatening clouds doesn't justify the belief in rain all by itself, but only by way of an inference in which the content of that observation serves as a premise. The gap between the content of the observation and the content of the belief it is supposed to ground is simply too large. Similarly, I say, in the case before us. The gap between the content of the apparent observation If (1) and (2) are true, then (3) must be true and the belief: I am justified in believing (3) is simply too large for the warrant to be direct, even if in most conversational contexts the inference could be left unsaid.17 To sum up. In order to ensure that a thinker's inference from particular MPP premises to a particular MPP conclusion not be blameworthy, the simple inferential internalist insisted that the inference's justifiedness be transparent to the thinker-the thinker has to be in a position reflectively to appreciate that his inferring this conclusion from these premises is justified. But this runs into two major problems. First, it requires us to take seriously a notion of rational insight, a notion that no one has been able to render respectable. Second, and even if we waived this first worry, the aimed-for transparency will still be unattainable, since the only way to attain it will require that the thinker use such knowledge as rational insight is able to afford him as the basis for an inference to the justifiedness of his conclusion. So no matter how concessive we are about rational insight and about the knowledge of logical implication that it is supposed to engender, there seems to be no way to satisfy the transparency insisted upon by Simple Inferential Internalism. VI. BLIND YET BLAMELESS INFERENCE: DEFLATIONARY OPTIONS The question is where we go from here. Simple Inferential Externalism is false. Simple Inferential Internalism, construed as requiring some form of reflectively accessible warrant, is unsatisfiable. We know, furthermore, that we cannot say that deductive inferences do not transfer warrant; that would be not merely implausible but self-undermining. Hence, unless we are to admit that our epistemic system is subject to deep and crippling paradox, there had better be a stable and coherent account of what the conditions for warrant transfer are. In searching for a solution, we must respect the following facts. On the one hand, the failure of Simple Inferential Internalism teaches us that it must be possible for certain modes of reasoning to be entitling without our knowing, or being able to know, anything about them. I'll put this by saying that it must be possible for certain inferences to be blind but justifying.'8 On the other hand, the counterexamples to Reliabilism teach us that the way not to accommodate this phenomenon is through Simple Inferential Externalism. So our question is: How should we construe warrant transfer consistent both with Simple Internalism's and Simple Externalism's falsity? If there is to be a way forward, the following had better be true: the antiReliabilist examples, properly understood, don't really motivate Access Internalism, even though they seem to do so. Rather, they motivate something weaker which can be reconciled both with the falsity of Reliabilism and with the falsity of Access Internalism. What could that intermediate position be? The minimal lesson of the anti-Reliabilist examples, as we saw, is that being justified cannot coexist with being epistemically blameworthy. To get from here to Access Internalism you need to assume furthermore that what makes a belief epistemically blameworthy is the absence of a reflectively appreciable warrant for it. And although all the known examples uniformly support this construal, it's not actually forced. For all that the examples show, in other words, it is possible that there should be some other way in which a belief might be held blamelessly other than by being supported by some reflectively appreciable warrant. Some philosophers are inclined to think that there isn't much of a problem here because they think that it isn't all that hard to be epistemically blameless. Gilbert Harman, for example, thinks that just about any belief, or method for forming beliefs, that one cares to have is blameless, at least initially. He writes: What I take to be the right theory of justification goes something like this (Goodman, 1995; Quine, 1960a; Quine and Ullian, 1978; Rawls, 1971). In deciding what to believe or what to do, you have to start where you are with your current beliefs and methods of reasoning. These beliefs and methods have a privileged status. You are justified in continuing to accept them in the absence of a serious specific challenge to them, where the challenge will typically involve some sort of conflict in your overall view. Conflict is to be resolved by making conservative modifications in your overall view that makes your view more coherent in certain ways. Your goal in resolving conflict is to reach what Rawls calls areflective equilibrium’, in which your various views are not in tension with each other … The crucial point is that, to a first approximation, continuing to accept what you accept does not require justification. What requires justification is making changes in your view. 19 On this view, which Harman dubs General Conservatism', we have a quick and painless answer to our question. We are now justified in using MPP because MPP is one of the methods with which westart’ and we have, so far, encountered no incoherence in our overall view to which the best response would have been to reject or modify it. The principal thought behind general conservatism is an innocent until proven guilty' model of epistemic justification. It doesn't matter what beliefs or methods one starts with-they are all prima facie justified. What matters is how one changes one's view in response to a developing incoherence. But this is all very misleading, for the notion ofcoherence’, is empty unless it embeds a specific conception of logical consequence and logical consistency. (Actually, it would probably have to include not just that but conceptions of probabilistic consistency, and a great deal more, but I’ll let that pass.)20 That in effect implies, however, that talk about coherence' presupposes an answer to our question, rather than providing one. You need to have figured out which deductive rules are justifying in order to have a substantive coherence theory rather than the other way around. This brings us to the seconddeflationary’ answer to our question: a list. You want to know which inference patterns are permitted to be blind? These ones: Modus Ponens, Non-Contradiction, and a few others. Don’t ask why it is precisely these inference patterns that are sanctioned. There is no deep answer to that question; there is just the list. What makes this brand of deflationary answer unsatisfactory is that it is hard to believe that the property of being warrant-transferring is simply a primitive property that an inference pattern either has or fails to have. Surely, if an inference pattern is warrant-transferring there must be some property by virtue of which it is warrant-transferring. And our question is: What, in the most basic cases, in which reflectively available support is not possible, could that property be?