50

Clearly, this is not the assertion of someone who is prepared to say, “Relative to my political perspective, it is justified to say that recent work has indicated the need to review and revise traditional views about knowledge; however, if you happen not to share that perspective, you may ignore everything I say, for I am not claiming that it is justified relative to your perspective.” On [End Page 219] the contrary, the intended force of the assertion is, rather, this: Recent work has shown that conventional ideas about knowledge are flawed, so anyone, regardless of his or her prior commitments, who wants to believe what is true had better change his or her mind about the nature of knowledge. If this were not the intended force of the assertion, why the tone of rebuke? I conjecture that one of the reasons why a relativism about reasons can sometimes seem compelling to those who hear it is that they hear it as making the rather different claim that what people think of as reasons for believing something can vary with their politics. This latter view, which is really an empirical conjecture about mental dispositions rather than a philosophical view about the nature of good reasons for belief, may or may not be true. It is certainly no part of classical epistemology to deny that it is true. But whether or not it is true has nothing to do with the philosophically controversial claim that lies at the heart of constructivism, namely, that what is a good reason—not merely what is thought to be a good reason—can vary with one’s background political values. At various points in her essay Smith suggests that there is something question-begging about the classical self-refutation arguments that are directed against relativism, that they depend in some unspecified way on the very system—classical epistemology—that relativism is supposed to place in question. But what do these arguments presuppose? They presuppose nothing more than the assumption that when we believe something, we believe it because we think there are reasons to think it is true— reasons that we perceive as general enough to be obvious even to people who do not share our perspective, so that we feel entitled to recommend it to them. In other words, they presuppose nothing more than the very attitude that Smith adopts toward her own claim that there is something seriously wrong with classical epistemology. Can we really imagine a way of thinking about belief that will leave no room for that attitude? It is time we turned to the third possible way of prosecuting a constructivist view of knowledge, the option I labeled (C) above. According to this family of views, political values play an indispensable role in explaining why we believe what we believe, not because reasons are constituted by political values but, rather, because we cannot be moved to belief purely by reasons. On this account of the matter, we preserve the distinction between factual evidence and political interests, but we insist that one cannot fully explain [End Page 220] why something is believed without adverting to political factors. Once more, the view comes in two flavors, one stronger than the other. The stronger, more extreme view would have it that factual evidence plays no role whatsoever in the explanation of belief, that although values do not justify a belief, we are not in any way moved to belief by considerations that justify; we are only moved by our political interests. Bizarre as this view may seem, it is practically orthodoxy among sociologists of science. On the most charitable reading, it stems from an innocent confusion about what is required by the enterprise of treating knowledge sociologically. Some twenty years ago, when the idea of a sociology of science was first beginning to take root, David Bloor argued as follows: If we wish to explain why certain beliefs come to be accepted as knowledge at a given time, we must not bring to bear our views about which beliefs are true and which false. If we are trying to explain why they came to hold that some belief is true, it cannot be relevant that we know it not to be true. This is one of the so-called Symmetry Principles of the sociology of knowledge: treat true and false beliefs symmetrically in explaining why they came to be believed. One can debate the merits of this principle, but for present purposes I propose to grant it. However, it is one thing to say that true and false beliefs should be treated symmetrically and quite another to say that justified and unjustified ones should be so treated as well. While it may be plausible to ignore the truth or falsity of what I believe in explaining why I came to believe it, it is not plausible to ignore whether I had any evidence for believing it. For some reason that is never explained, however, Bloor and his colleagues seem to think that the two principles are on a par and that both are equally required by the enterprise of treating scientific belief sociologically. Bloor builds both principles into the very foundation of the subject: The sociology of knowledge “would be impartial with respect to truth and falsity, rationality or irrationality, success or failure.” 5 The intuitive view, however, is that I sometimes believe something precisely because I have a good reason for believing it—as, for example, when I come to believe that there is someone at my door because I hear a knock. To insist that for every belief that I have there must be some alternative way of explaining why I came to have it, one that does not mention my evidence for it, is to insist that reasons can never constitute the real explanation for why I came to believe what I believe. [End Page 221] But what is the argument for this astonishing claim? And how could there be an argument for it that did not immediately undermine itself? Surely anyone promoting the view that reasons cannot move people to belief would presumably have to represent themselves as having come to that view precisely because it is justified. None of this is to say, of course, that all belief—or even all scientific belief—must be explained in terms of the compelling evidence assembled for it. The history of science is replete with theories— phrenology, for example—for which there never was any good evidence. It is simply to insist that scientific belief is sometimes to be explained in terms of compelling evidence and that the history and sociology of science, appropriately conceived, need have no stake in denying that. Smith has a number of strange things to say about this nexus of issues. First, there is her account of the symmetry postulate itself. She rebukes Lorraine Code for misunderstanding the frequently mentioned but evidently rarely read “symmetry postulate” of Edinburghbased . . . sociology of science. The postulate, which has nothing to do with tolerance in any of the usual senses of the term, maintains not that all constructions of reality are equally worthy but, rather, that the credibility of all constructions of reality, including those now commonly accepted as true or reasonable, should be regarded as equally needful of explanation and as explicable, in principle, by the same general types of causes. A key methodological point of departure—not epistemic judgment—in contemporary science studies, the symmetry postulate is routinely transformed into a fatuous egalitarianism (everything is equally true, good, worthy, valid, and so forth) by those who encounter it primarily through hearsay. However, the formulation of the symmetry postulate quoted above, which I am happy to say is taken from Bloor’s actual text and is not based on mere hearsay, commits the sociology of knowledge to the view that it must always be possible to explain any belief without having to mention either its truth or its reasonableness. It is not the much milder view that Smith seems to think it amounts to, that no belief can be presumed intrinsically credible or self-evident. This latter claim insists only that no belief be adjudged reasonable without some supporting evidence; it has nothing to do with the radical proposal, clearly endorsed by Bloor and other supporters of the “Strong Programme,” [End Page 222] that it should always be possible to explain any belief in terms that do not mention its reasonableness. Another criticism of the symmetry postulate is forcefully expressed by philosopher of science John Dupré: By asserting that all scientific belief should be explained in terms of the goals, interests, and prejudices of the scientist, and denying any role whatsoever for the recalcitrance of nature, it leaves no space for the criticism of specific scientific beliefs on the ground that they do reflect such prejudices rather than being plausibly grounded in fact. The uncongeniality of the sociology of science program to thinkers genuinely concerned with political influences on scientific belief is nicely stated by the feminist philosopher of science Alison Wylie . . . : “Only the most powerful, the most successful in achieving control over the world, could imagine that the world can be constructed as they choose.” 6 Smith takes Dupré to task for trying to “validate” his “intellectual traditionalism” by appealing to gestures of “solidarity with political radicalism”: Notable in that connection is Dupré’s curious allusion to “the uncongeniality of the sociology of science program to thinkers genuinely concerned with political influences on scientific belief”—curious because the work of that program’s most eminent practitioners, for example, David Bloor, Barry Barnes, Andrew Pickering, or Steven Woolgar, would certainly seem to be extensively concerned with “political influences on scientific belief.” (Indeed, in the outraged view of various recent detractors, it is quite menacingly concerned with nothing else.) Has Dupré failed to notice this concern? Or is he suggesting that the sociologists are just faking it . . . ? Smith seems to have missed the important but fairly obvious point that Dupré is making, which is that if you say that the correct explanation for all scientific belief is in terms of the political goals, interests, and prejudices of the scientist, you make it impossible to criticize a specific scientific belief as merely reflective of such prejudice. I take it that all of us would want to be able to distinguish between the conclusions of The Bell Curve, for example, and Maxwell’s equations for electromagnetism by saying precisely that the first merely or largely reflects the prejudices of its authors whereas the second does not, that it is grounded in solid, persuasive evidence. How [End Page 223] are we to make this distinction, however, if we simultaneously insist that all scientific belief be explicable in terms of political goals, interests, and prejudices? As I mentioned above, this third constructivist option has a more interesting, milder version, one that Smith does not discuss but which it would be good to have on the table for the sake of completeness. On this milder view, although evidence does enter the explanation for why a particular view is believed, it can never be enough to explain it. Any evidence we might possess always underdetermines the specific belief that we arrive at on its basis. Something else must close the gap between what we have evidence for and what we actually believe, and that something else is provided by the thinker’s background values and interests. This idea, that the evidence in science always underdetermines the theories that we believe on its basis, has exerted considerable influence in the philosophy of science, even in nonconstructionist circles. In its modern form it originated in the thought of turn-of-the-century French physicist and philosopher Pierre Duhem. 7 Suppose that an experimental observation is inconsistent with a theory that you believe: the theory predicts that the needle will reach 10, but the needle does not budge from 0. What Duhem pointed out is that this does not necessarily refute the theory. For the observational prediction is generated not merely on the basis of the theory; in addition one must assume auxiliary hypotheses about how the experimental apparatus works. In light of the recalcitrant observational result, something has to be revised, but so far we do not yet know exactly what. Perhaps it is the theory; perhaps it is the auxiliary hypotheses. Perhaps, indeed, it is the very claim that we recorded a genuinely recalcitrant result, as opposed to merely suffering some visual illusion. Duhem argued that evidence alone could never decide which revisions are called for and, hence, that belief fixation in science could not be a purely rational matter. Something else had to be at work as well. What the constructivist adds is that this extra element is something political. There are, however, several well-known problems with this ingenious argument. One problem is that it is subtly begging the question. In trying to show that we can never have reason to modify one belief as opposed to another, it must assume that we can never have such reason. Consider Duhem’s example of an astronomer peering through his telescope at the heavens and being surprised at what he finds there, perhaps a hitherto undetected [End Page 224] star in a galaxy he has been charting. Upon this discovery, according to Duhem, the astronomer may revise his theory of the heavens or he may revise his theory of how the telescope works. Rational principles of belief fixation do not tell him which to do. The idea, however, that in peering at the heavens through a telescope we are always testing our theory of the telescope just as much as we are testing our astronomical views seems absurd. The theory of the telescope has been established by numerous terrestrial experiments and fits in with an enormous number of other things that we know about lenses, light, and mirrors. It is better confirmed than our views about the contents of a given galaxy. As a result, it is simply not plausible that, in coming across an unexpected observation of the heavens, a rational response might be to revise what we know about telescopes. In the first instance it would be more rational to revise our beliefs about the galaxy. The point, of course, is not that we might never have occasion to revise our theory of telescopes; one can certainly imagine circumstances under which that is precisely what would be called for. The point is that not every circumstance in which something about telescopes is presupposed is a circumstance in which our theory of telescopes is being tested, and so the conclusion that rational considerations alone cannot determine how we respond to recalcitrant experience is blocked. A second difficulty with the general thesis of the underdetermination of belief by evidence is of a familiarly reflexive nature: it must exempt itself from the general claim that it is putting forward. To say that the evidence underdetermines what we tend to believe on its basis is to say that the evidence alone does not take us all the way to the belief that we actually form; something else must close the gap, and that something else is political according to the constructionist. But let us ask this: How is this established? How do we know that our beliefs always outstrip the evidence and that it is something political that fills the gap? The answer, we are invited to believe, is that this is the lesson of the close study of the history of science engaged in by our colleagues in science studies. Careful, painstaking investigation of the actual course of the development of science reveals that what we believe must always in part be explained by appeal not only to the evidence at our disposal but by our underlying political values as well. But if the evidence always underdetermines what we believe on its basis, [End Page 225] then presumably the historical evidence assembled by science scholars also underdetermines the claim that evidence always underdetermines what we believe on its basis. If the constructionist is being consistent, he or she must admit that the evidence alone does not settle whether belief is always underdetermined by the evidence, that something about his or her own political agenda is making the constructionist say that it is. But if that is the case, why do I have to accept what he or she says, if I do not happen to share that agenda? * * * A number of difficult questions are left dangling of which the most important is this: Suppose someone raises an entirely global skeptical challenge to the norms of correct reasoning that we employ. How could we go about adjudicating which of us is correct? It is hard to see how the debate would be conducted. We will expect a reasoned argument demonstrating the incorrectness of our norms. But a reasoned argument that will impress us would have to conform to our norms, which is precisely what the skeptic is challenging. On the other hand, if we are not given a reasoned argument that conforms to our norms, we will remain unmoved. In trying to get us to change our minds, the skeptic would be better off threatening us with physical harm than with giving us reasons that we refuse to recognize as reasons. There is a deep question about how our specific conception of what counts as a good reason is to be justified in response to a hypothetical global challenge. If the challenge is piecemeal, to some specific part of our conception, then it is entirely conceivable that we should be able to use one part of our conception to defend or attack another. But it is hard to see how the argument is to be conducted if the challenge is global. What is crucial, it seems to me, is to see that the varieties of constructivism that we have been exploring are totally inadequate responses to the possibility of such skepticism. They are not coherent attitudes by which a belief can be maintained or an assertion put forward, as is shown by the fact that they will be violated even by the skeptic’s own attitudes toward his or her own assertions. I do not see how we are to think about belief and assertion without the idea that some beliefs are backed by reasons for taking them to be true and others are not. I do not see how to hang on to that idea without drawing some sort of principled distinction between those things that count as reasons and those that do not. And I do not see how, in the most general case, I [End Page 226] can think of these reasons as mere reflections of my own perspective and not ones that may equally have a grip on someone else. In saying all this, I am, of course, showing myself to be one of those “scholars . . . whose conviction of the adequacy of traditional views remains unshaken and who, accordingly, will continue to reaffirm the traditional positions as such, staunchly rehearsing the classic justifications and refutations, unruffled by what colleagues experience as the logical bite of current challenges.” But as Martin Luther is reputed to have said, here I stand, I can do no other. Paul Boghossian is a professor of philosophy and chairman of the Department of Philosophy at New York University. His research interests are in the philosophy of mind, philosophy of language, and epistemology. He has written on a variety of topics, including color, rule-following, eliminativism, naturalism, self-knowledge, a priori knowledge, analytic truth, realism, and the aesthetics of music. He is currently at work on a book on the notion of objectivity. Some recent publications include “How Are Objective Epistemic Reasons Possible?” in Philosophical Studies; “Inference and Insight” in Philosophy and Phenomenological Research; “On Hearing the Music in the Sound” in the Journal of Aesthetics and Art Criticism; and “The Gospel of Relaxation” in the New Republic.