Knowledge: A Human Interest Story

Imagine a scene from a detective movie. The detective and their team are stumped. They’ve pursued all their leads and gotten nowhere, and all their suspects have been cleared. What do they do? In many such movies, it’s time to go back to the drawing board. “Let’s put on the left side of the board just what we know,” the detective might say. “This revolver was found at the murder scene. There is a palm print on the gun. The victim had a wallet on their person. The name on the driver’s license was scratched off”—and so on. Once they put on the board only what they know, the hope is, it will reveal potential paths of inquiry that may have been unjustly closed off when they started focusing on specific suspects.It seems to me that this picture of the relationship between knowledge and inquiry—that knowledge is the beginning of inquiry—is very appealing and natural. But, it might surprise some readers to learn, it is not the governing picture of the relationship in much contemporary epistemological literature on the relationship between knowledge and inquiry. Rather, epistemological work focuses (favorably or not) on the question of whether knowledge is the end of inquiry or whether inquiry aims at knowledge, or whether knowing the answer to a question does or does not preclude the possibility of properly inquiring into that question (or genuinely inquiring at all). All of these, it also seems to me, are excellent questions. But it’s surprising how much contemporary epistemology has neglected this other appealing picture of the most important relationship between knowledge and inquiry: that knowledge is the starting point.This neglect is commendably remedied in Brian Weatherson’s excellent 2024 Knowledge: A Human Interest Story. In this book, Weatherson argues for an interest-relative view about knowledge. This is a kind of view that Weatherson has endorsed for twenty years, though the details of the view receive considerable updates in the recent volume. Weatherson has changed his mind on a number of key issues, some of which make his view deviate in new ways from other versions of interest-relativity in the literature. The book is a clear, rigorous, and rich advancement of the literature on interest-relativity and pragmatic encroachment; it ably showcases the evolution of the literature over the past two decades and contributes to its next developments.The primary locus of debate in the literature is over whether knowledge is interest-relative, and that is the central debate that Weatherson intercedes in. According to interest-relative theories, in Weatherson’s formulation, “what a person knows is sensitive to what their interests are, and in particular to what inquiries they are engaged in” (7). Note that this formulation can be contrasted with some others Weatherson might have adopted. For example, Weatherson does not formulate interest-relativity, as others do, in terms of the denial of a supervenience thesis—something like, it’s possible for two subjects to be in the same epistemic position with respect to a proposition but to differ in whether they know that proposition because their interests differ. Nor does he say that whether a person knows something can vary merely with changes in the person’s interests. I’ll discuss below why these differences matter and why it makes sense for Weatherson to rely on his formulation.Weatherson spells out his main argument in chapter 2. You read in a reliable history book that the Battle of Agincourt took place in 1415, and you are then offered a choice: (1) two plus two equals four, or (2) the Battle of Agincourt took place in 1415. You get fifty dollars if the statement you choose is true, and there is no payout (or loss) otherwise. Dominance arguments seem to rationally require choosing option 1. After all, in this situation you know that option 1 will perform no worse, and might perform better, than option 2. There is a very strong intuition that the uniquely rational play is to pick option 1.But why? The natural explanation is that you don’t know that the Battle of Agincourt took place in 1415, while you do know that two plus two equals four. If you knew both, then you’d also know that choosing either option 1 or option 2 will have the same payout: fifty dollars. And that means you should be indifferent between option 1 and option 2, rather than rationally required to pick option 1. Denying this requires also “denying some very simple principles connecting knowledge and action—and in particular denying that it is rational to take the action one knows to be best” (19).So it looks like we need to say that, when offered this game, you should pick option 1 and that this must mean that you don’t know that the Battle of Agincourt took place in 1415. Barring an implausibly strong form of skepticism, though, you can sometimes know such things. In ordinary situations you can know mundane facts you learn from reliable history books. Notice that nothing hinges on having a firm intuition about this specific case, and Weatherson takes pains not to rest his argument on case-specific intuitions about whether knowledge is present. The issue is that it is implausibly over-skeptical to respond to this case by saying that you can never know such things as that the Battle of Agincourt took place in 1415.So, when you first read that the Battle of Agincourt took place in 1415, you know it and can use it in reasoning. For example, if you’re just trying to answer a trivia question, you can check a history book to confirm your intuition that the answer is 1415. This is normally just the kind of thing that we think grants you knowledge. But then, in the abnormal situation in which you’ve been offered the gamble, it’s no longer certain enough to rely on in reasoning, and so you must not know it. This means that your knowledge is sensitive to your interests, for example, in winning fifty dollars in this gamble. That’s just interest-relativity about knowledge.So far, as Weatherson emphasizes, this general argument strategy has more similarities to than differences from his own previous positions and other recent defenses of interest-relativity. But much of the book is focused on developing ways that Weatherson’s current position differs from other interest-relative views. I’ll mention two.One distinctive feature of Weatherson’s new view is his contention that interest-relativity isn’t limited to knowledge. Instead, there is no domain of purely epistemic evaluation: “Everything in epistemology is interest-relative. Knowledge, rational belief, and evidence are all interest-relative” (3). Because, as he argues in chapter 9, evidence is interest-relative, “probability, justification, rationality, and evidential support are interest-relative too” (51). If this is right, then this explains why Weatherson does not opt for definitions of interest-relativity as suggested above—in terms of denials of supervenience theses or whether “mere” changes in interests can impact knowledge. If “everything” is interest-relative, then arguably there are no “mere” changes in interests. Changes in interests—at least, changes important enough to destroy knowledge—might very well force other important epistemic changes.If they inevitably do, it looks like any time two subjects differ with respect to knowledge that p, they’ll also differ with respect to evidence, probability, truth, or reliability with respect to p; for every change in knowledge, there has to be a change in evidence, probability, truth, or reliability. Is this still an interest-relativity about knowledge? On the one hand, as long as evidence, probability, truth, or reliability are themselves interest-relative, knowledge will be (indirectly) sensitive to interests. On the other hand, it’s hard to distinguish this collection of views from a view that is interest-invariant about knowledge but interest-relative about, say, evidence.A second way that Weatherson’s new view is distinctive is that, on his view, it’s not just your practical interests that your knowledge is sensitive to. Your merely theoretical interests can impact your knowledge as well. Even if you haven’t been offered the gamble described above, but you’re simply curious about what the best option would be to take in such a situation, that can impact knowledge. Knowledge can be used in reasoning—theoretical or practical. So, if you can’t use some proposition in some purely theoretical matter you’re inquiring into, then you don’t know it.This is a fairly major deviation from the recent interest-relative tradition. Interest-relative theorists, Weatherson included, rely on a principle that whether you know that p is a matter of whether p is “certain enough” (14), has “high enough” evidential probability (51), or generally is “warranted enough” (Fantl and McGrath 2009: 66) to be used in reasoning. Because what is warranted enough to use in reasoning can vary from one practical situation to another, what is known also can vary. What determines what practical situation you’re in is what the expected payoffs are of various courses of action. So far, Weatherson is right in line with that tradition.But when we turn to theoretical reasoning, it’s harder to see how this is supposed to work. Because the practical situation remains constant in cases of different purely theoretical reasoning, payoffs remain constant across different theoretical situations. Why, then, do propositions go from warranted enough to not warranted enough across those situations? Consider some putative reasoning from a premise, p, to a conclusion, q. Is p warranted enough to draw q as a conclusion? Why would the answer to this question depend on whether you are engaged in inquiry into this issue (whether q is true, or whether p supports q), or even whether you should be inquiring into it (144)? If you’re not thinking about the issue or have no interest in the question of whether q is true or whether p supports q, then you of course won’t conclude q on the basis of p. But it’s hard to see how that affects whether p is warranted enough to draw that conclusion.For example, suppose you have a gun but haven’t checked to see if it’s loaded. You’re not going to point it at someone and pull the trigger; you’re not inquiring into whether to do so; the question of doing so isn’t even on your radar. You’re not currently interested in the question of whether to point the gun at someone and pull the trigger. Regardless, that the gun is empty isn’t warranted enough to justify doing that. Changes in what you’re inquiring about or have an interest in inquiring about, by themselves, don’t affect whether p is warranted enough to draw conclusions from. But when it comes to purely theoretical reasoning, changes in what you’re inquiring about or have an interest in inquiring about are the only changes there are. In short, Weatherson’s modification here deviates from other interest-relative views not just in the kind of reasoning it applies to. It also counts as a relevant “interest” what reasoning you happen to be interested in.