Cohen on Boghossian Book Reviews/Comptes Rendus 229 Argumentation has its own Trinity—Rationality, Universality, and Autonomy— but there is no escaping the suspicion that it might be an unholy trinity. Argumentation is constitutive of rationality. It is universal in that we can argue about anything at all. And argument is autonomous insofar as it establishes its own rules. Individually, these are profoundly important for understanding argumentation. Together they tell us that part of what it is to be rational is to reason about being rational, and therein lies a paradox. Universality means that we can even argue about the principles of argumentation, while autonomy means that there is no further court of appeal for the principles of argument than argument itself. At the same time, rationality demands that we reason about reason—apparently, with No Exit from argument anywhere in sight. But surely it would be the height of irrationality to try to argue about something by first arguing about how to argue, then arguing about how to argue about how to argue, ad infinitum. Rationality also bids us stop. Paul Boghossian’s recent book, Fear of Knowledge, has something to say about this problem that is worth listening to. While the primary subject of his book is exactly as the subtitle advertises, the subject of rationality is never far away. What Boghossian offers is an extended argument against some forms of contemporary anti-realism and, by implication, an argument for realism. The intended audience is philosophers with metaphysical and epistemological interests. In the end, however, it could be argumentation theorists who might be most engaged by it and have the most to say about it because while the book is seriously flawed as an argument, it makes a positive contribution when read as a discourse about argument. Let me dispense with the negatives first. The greatest failings of Boghossian’s argument are dialectical and rhetorical rather than logical. His readings of Wittgenstein, Kuhn, Rorty, and his other opponents are sometimes so uncharitable as to elicit strong objections from anyone familiar with their work. Even the readers who are sympathetic with Boghossian’s own positions will get the feeling that there is much more to be said on behalf of his opponents, and I expect those readers could well be moved to try to defend his opponents! Rather than reinforcing realist inclinations, the dialectical effect is to move the reader closer to the anti-realist camp. Readers’ objections cannot be answered, of course, unless they are anticipated, so it is incumbent on the author to make that effort. Boghossian occasionally makes something of an effort, but generally falls short. One example is especially illustrative. In characterizing the relativist position as it applies to the justification of our beliefs, rather than those beliefs directly, he writes: “It is also hard to explain why anyone should care about what follows from a set of propositions that are acknowledged to © Informal Logic Vol. 27, No. 2 (2007): pp. 229-232. FEAR OF KNOWLEDGE: AGAINST RELATIVISM AND CONSTRUCTIVISM BY PAUL BOGHOSSIAN. Oxford: Clarendon Press, 2006. 152pp. ISBN: 019928718X US$24.95 (hbk) Book Review 230 be uniformly false” (p. 87). Why should we assume that relativists who refuse to regard any propositions as expressing objective truths thereby acknowledge them as false? When the subject is truth itself, the dichotomy of truth and falsity would surely be a false dichotomy. Boghossian makes an attempt at the appropriate dialectical move: “for judgments… to be untrue, it is not strictly speaking required that we think of those judgments as false” (p. 88, his italics). A promising start, but he undoes whatever good will the concession earned him by his awkward follow- up: “There are two ways, of course, for a judgment to be untrue, only one of which consists in its being false. The other way is for the judgment to be untrue because incomplete.” The other way? There is a failure to imagine the possibility of other measures for our assertions about the world besides that of True-or-False, or the possibility that there could be reasons for thinking the True-or-False scale is as inappropriate for judgments as Black-or-White or Odd-or-Even. For comparison, we might ask whether the only other way for a proposition to be not-male, besides being female, is to be incompletely gendered. To be fair, Boghossian does argue directly for his claim that epistemic judgments, at least, have to be absolute. It is also true that relativists have a remarkable record as enabling accomplices for Straw Man argumentation directed against them because of their predilection for hyperbole. Kuhn’s comment that scientists working under different paradigms “live in different worlds” is just one example, albeit a notorious one, of something that could (and I think, should) be taken as metaphorical over- statement but has more often been taken as evidence of metaphysical silliness. The result has been that Kuhn’s main insights were often overshadowed by the attendant brouhaha. Reading Rorty also requires some care because despite his consistent rejection of the concept of truth, he is comfortable enough to use the word “true.” From his pen, however, that word has to be understood as just an honorific that we apply to sentences that, for any of a number of different possible reasons, we think are good to say and believe (see, e.g., Consequences of Pragmatism, p. xiii). Boghossian’s arguments against relativism concerning the truths about the world culminate with a critique of relativism concerning the principles governing our reasoning about how we should reach those truths. This stage of his argument is where the real fun begins for those whose interests are less metaphysical and more focused on the theory of argumentation. Boghossian tackles the problem of the “inevitable norm-circularity of justifications of our epistemic systems” (p. 95) head on. He recognizes that while reasoning about reasoning is an integral part of being rational, it is a special case posing special problems. Part of what makes the problem particularly difficult for Boghossian is his desideratum that “epistemic facts” must be modally strong. The evidence for our knowledge may be contingent; its relation to our knowledge is not: “facts about what belief would be justified by a given piece of evidence are facts that must be thought of as absolute” (p. 111). We are in a bind. Our epistemic principles must be justified as absolute and necessary, but they can only be justified by reference to themselves. The problem of circularity immediately confronts us. That problem is followed by another: modality. If our epistemic principles are socially constructed, culturally relative, or contingent in any other way, then their justification as necessary apparently becomes impossible. Even if we accept circularity as non-vicious, the contingency remains. There is a complicated admixture of pragmatic elements and hard-line realism in Boghossian’s three-part strategy to escape circularity, justify our beliefs, and ascend Book Reviews/Comptes Rendus 231 to necessity. The pragmatic fulcrum Boghossian uses against relativism (and skepticism) is our “blind entitlement” to certain epistemic principles. “Each thinker is blindly entitled to his own epistemic system—each thinker is entitled to use the epistemic system he finds himself with” (p. 99). We are, after all, like those sailors on Neurath’s ship. Or we could say that we have to go to epistemic war with the epistemic beliefs we have. The fact that there might be extraordinary occasions for calling those principles into question—in which case, they could not be legitimately deployed in their own justification—does not mean that they cannot be taken for granted in the normal course of events. Not all circular justifications are equally objectionable. However, when the goal is more than just acceptable, consensus beliefs but absolutely objective, universal, and necessary principles, benign circularity cannot be enough. The next step is to blur the boundaries that are often drawn (but not by all constructivists!) between explaining our beliefs and justifying them, and between evidential reasons and inferential reasons. Naturalized epistemologies invoke causes to explain our beliefs; normative epistemologies refer to reasons to justify our beliefs. Boghossian passes easily from one project to the other, rejecting “the constructivist’s claim… [that] exposure to the relevant evidence can… never suffice to explain why we form the beliefs that we form” (p. 112). Instead, for some beliefs, exposure to the relevant evidence suffices to provide the proper explanation for them. “It is impossible to see what would prevent our epistemic reasons from sometimes causing our beliefs” (p. 117). Since Rorty’s version of the history of modern philosophy pinpoints exactly this conflation of reasons and causes as the undoing of classical empiricism, Boghossian might have offered his readers more in support of his claim. Similarly, perhaps more care could have been taken to avoid any hint of the implication that there is such a thing as the (one and only) correct explanation for each phenomenon. Relativists and constructivists, at least, will question this for physical phenomena; I would have thought that everyone would question it when the explanandum is belief, as it is here. The final step is to reconsider the possibility of radically different epistemic systems. With an argument reminiscent of Davidson’s critique of alternative conceptual schemes, Boghossian analyzes the case of Azande reasoning, Evans- Pritchard’s candidate for an alternative conceptual scheme. He reaches the conclusion that if they appear to eschew modus ponens, then we must somehow be mis-translating their logical particles. Any connective that does not support modus ponens necessarily cannot be what we mean by “if.” Does this line of reasoning make the necessity of modus ponens as an epistemic principle for inferences simply an artifact of the methodology of translation? Not entirely. “In doing epistemology, we not only assume that [epistemic facts] are knowable, we assume that they are knowable a priori” (p. 76). Apparently, we are entitled to both the truth of our epistemic principles and to their necessity! (And, by the way, so much the worse for any form of naturalized epistemology.) There is more to this than merely the method of assertion. Boghossian can be helped out here by a favorite distinction among informal logicians, viz., the difference between narrow and broad uses of the word logic. In its narrow sense, logic is about formal inferential systems; more broadly, it is about good reasoning generally. 232 The same distinction can be applied with good effect to reason and rationality. When discussing paradigm shifts, incommensurability, and constructivism in connection with what is often called the “Quine-Duhem” thesis of the under- determination of theory by evidence. Boghossian writes: Duhem argued that reason alone could never decide which revisions are called for and, hence, that belief revision in science could not be a purely rational matter: something else had to be at work as well. What the social constructivist adds is that this extra element is something social. (p. 126) In this passage, “reason alone” and what is “purely rational” must be taken as narrow. It leads us to accept only what is included in the given data and what can be deductively inferred from that data. He rejects Duhem’s conclusion, and exculpates Quine: Quine never endorsed the view that reason alone cannot tell us which revisions to make in the face of recalcitrant evidence. His was the much more limited claim that any evidence we might collect for a given generalization is logically consistent with the falsity of that claim. (p. 126) Of course, Quine most certainly did endorse the view that reason in the narrow sense cannot tell us which revisions to make. It is reason in the broad sense that is sensitive to such pragmatic considerations as the modesty and conservatism of our theories, along with their utility, convenience, and even aesthetics. Web management is a rational affair—a broadly rational affair. In the end, it is this same distinction that provides us with our exit. It is rationality in the narrow sense that invites us to argue about arguing. Sometimes we need to do just that. But it is rationality in the broad sense that counsels us to stop. It is an epistemic virtue of the first rank to be able to keep things in balance. And with that sense of proportion, there is no need to fear either knowledge or reason. DANIEL H. COHEN Colby College