Contemporary Debates in Epistemology

Matthias Steup, John Turri, Ernest Sosa, Matthias Steup, John Turri, Ernest Sosa

• Publisher: Wiley-Blackwell
• Year: 2013
• ISBN:
  9781118328125
• Edition: 2
• Topic: Philosophy
• Subtopic: Epistemology in Philosophy

Fully updated with new topics covering the latest developments and debates, the second edition of this highly influential text retains its unique combination of accessibility and originality.

• Second edition of a highly influential text that has already become a standard in the field, for students and professional researchers alike, due to its impressive line-up of contributors, and its unique combination of accessibility and originality
• Twenty-six essays in total, covering 13 essential topics
• Features five new topics that bring readers up to speed on some of the latest developments in the field, and give them a glimpse of where it's headed: Should knowledge come first? Do practical matters affect whether you know? Is virtuous motivation essential to knowing? Can knowledge be lucky? Can evidence be permissive?
• Substantially updates two other debates: Is there immediate justification? Can belief be justified through coherence alone?

Chapter One

Should Knowledge Come First?

What is the place of knowledge within epistemology? This is a methodological question of first importance. Timothy Williamson argues that knowledge should come first. Methodologically, this means that we shouldn’t expect an informative analysis of knowledge in terms of belief, truth, plus some further set of necessary, non-circular ­conditions. Nor should we accept an impoverished conception of evidence or epistemic normativity that would be acceptable to a skeptical interlocutor. Rather, we should begin doing epistemology by focusing on epistemic access itself, and knowledge itself is the most natural candidate for this access. We can then understand other notions such as evidence and justification in terms of knowledge, which is a factive, world-involving state. In opposition to Williamson’s view, Trent Dougherty and Patrick Rysiew argue for an experience-first approach to epistemology. Experience is where it all begins – indeed, where it must begin. Understanding experience to include ­perceptual ­awareness, introspective awareness, and rational insights, Dougherty and Rysiew argue that it is our basic evidence – that in virtue of which all else is made evident – that must come first in any intellectual project. On this approach, it is experience that ­ultimately justifies belief, guides rational thinkers, signifies truth, and settles disputes.

Knowledge First

Timothy Williamson

Epistemology matters.¹ It is not just fascinating in itself; its concerns arise in every serious form of human inquiry. How much does this evidence support that claim? Is this alleged evidence just another claim, itself in need of evidence? Is the category of ­evidence even relevant here? Such disputes are often resolved by the internal ­standards of the inquiry. But sometimes they go deeper, raising issues adequately addressed only at a level of abstract reflection characteristic of philosophy. It is all the more striking that in the second half of the twentieth century epistemology acquired a reputation (of which many practitioners still seem unaware) as sterile and inward-looking. Some of that was just the usual complaint about analytic philosophy, that it is boring and inaccessible to the untrained. But much of it came from other analytic philosophers, well acquainted with analytic epistemology, who still found it small-minded and old-fashioned in relation to its proper task.

One aspect was the post-Gettier industry of trying to analyze knowledge in terms of belief, truth, and non-circular further conditions. The increasingly gerrymandered ­definitions were obvious signs of a degenerating research program. Most of them, if correct, seemed to make knowledge too grue-like to be worth analyzing. But in any case they succumbed one after another to counterexamples. Moreover, no prior reason to expect knowledge to have such an analysis withstood scrutiny. Evidence accumulated that few if any words of natural language are understood by means of complex ­definitions. Nor does the nature of knowledge provide any clear evidence that it has such an analysis. That knowing entails believing truly does not show that for some non-­circular condition C, knowing is equivalent to believing truly and meeting C. In consequence, the project of analyzing knowledge has lost its importance to analytic epistemology.

Another common charge is that epistemology is obsessed with the problem of ­skepticism, wasting its time on an imaginary opponent. The charge is partly unfair. Not only is analytic epistemology not mainly concerned to answer skepticism, the issue usually arises when apparently legitimate ways of criticizing ill-founded beliefs, applied more systematically, lead to skeptical conclusions. Epistemologists did not willfully introduce the skeptical tendency from outside; it was already in us. That does not mean that the skeptic is right, just that we should consider the issue.

However, skepticism plays a further role in defining the framework of much analytic epistemology. Many practitioners take the key epistemological notion to be not knowledge but justification, in a specifically epistemic sense (a qualification ­henceforth understood). They typically explain the difference between knowledge and justification by contrasting an everyday situation with a skeptical scenario in which everything appears the same to one but one is a brain in a vat. In the good case, one knows that one has hands. In the bad case, since one lacks hands, one does not know that one has them; one merely appears to oneself to know that one has hands. By contrast, such epistemologists claim, in the two cases one is justified to exactly the same degree in believing that one has hands. They intend this notion of justification for general ­epistemological purposes, not only for handling skepticism. In particular, they treat it as the appropriate normative standard for criticizing beliefs.

That view casts appearances in a leading role: justification supervenes on them. But what is epistemologically so special about appearances? One answer is that if cases appear the same to one, then one cannot discriminate between them, and cannot fairly be criticized in one case with respect to a feature it does not share with the other, since one cannot discern it. The underlying principle is that justification is exactly the same in cases indiscriminable to the subject. But that principle falls to an objection from the non-transitivity of indiscriminability. Consider a long sorites series of cases α₀, … , α_n, where the subject α_i and α_i+1 are so similar in appearance that they are ­indiscriminable (i = 0, … , n − 1), but α₀ and α_n are so different in appearance that they are easily discriminable. By the principle, justification is exactly the same in α_i and α_i+1 (i = 0, … , n − 1). Therefore, by the transitivity of exact sameness in a given respect, justification is exactly the same in α₀ and α_n. But that is absurd, for justification differs between easily discriminable cases for some proposition. To avoid such objections, one must stick to the weaker principle that when appearances are exactly the same, so is justification. But why predict so tight a link between justification and appearances, if it is not mediated by indiscriminability?

Another idea is that in some sense one is always fully acquainted with present appearances to one, even if one cannot discriminate slight differences between them. This may help to explain the epistemological privileging of appearances (if one is never fully acquainted with anything else). The upshot is some sort of phenomenal conception of evidence. But by now, alarm bells should be ringing. The idea of full acquaintance with appearances has no basis in contemporary psychology or cognitive science. Nor has it much phenomenological plausibility. When I ask myself what I am acquainted with, the physical objects in front of me are far more natural candidates than their appearances. If I try to introspect or otherwise catch how things appear to me, I ­experience confusion, characteristic of embarking on an ill-defined task. Rather, full acquaintance with appearances is a wild postulate of a specific type of ­epistemological theory, one that requires something to be fully and unproblematically given to the subject to serve as the basis for justification. This is a barely updated Cartesianism: however vulnerable I am to doubt, ignorance, and error, something in me is clear and distinct.

To a depressing extent, epistemology has served as the refuge of an otherwise ­discredited philosophy of mind, supporting and supported by the definition of ­epistemic normativity in terms of a skeptical challenge. For justification was explained by the contrast between the ordinary case and the skeptical scenario. There may even be a distorting selection effect, by which those inclined to think along such lines are ­disproportionately drawn to, and rewarded in, epistemology rather than other branches of philosophy. It is Cartesianism that makes epistemology the starting point.

Suspicious of full acquaintance with sensory appearances, we might strip them out of the picture. What that leaves of the inner is a formal structure of beliefs, for which the norm of justification above reduces to mere internal coherence. Subjective Bayesianism is the best developed such view. Despite its mathematical virtues, it fails to make most distinctions of epistemological significance. It treats alike you and someone with the same credences but a radically different perceptual experience of the world.

The starting point of Cartesian epistemology is the comparison between the good case and the corresponding bad case. From a contemporary perspective, what they most obviously share is an internal microphysical state S. Consequently, they also share any mental state that supervenes on S. But S has no privileged epistemological status. Our internal microphysical states are typically unknown to us, and can only become known through arduous scientific investigation. Similarly, any mental state that supervenes on S has as such no epistemological privilege. It may be a state of depression that we cannot introspect ourselves to be in. What matters is whether we are aware of being in the given state, whether we know that we are in it.

Epistemologies that explicitly make knowledge a secondary phenomenon may ­nevertheless implicitly put it first, because they select the mental states to which they officially assign a privileged epistemological status for their supposed amenability to being known. They typically take that amenability for granted, rather than subjecting their crucial choice of starting point to open reflection. The less amenable the selected mental states turn out to being known, the more arbitrary becomes their promotion to a privileged epistemological status. Once we give up Cartesian fantasies about the mind, we can recognize that no special sort of fact is automatically amenable to being known, although many sorts of fact often are known. Rather than seeking a domain to which we have privileged epistemic access, we should concentrate on epistemic access itself. By far the clearest explication of “epistemic access” is simply knowledge. Thus attempts to start epistemology with something much more internal than knowledge nicely illustrate the naturalness of starting with knowledge itself.

On one knowledge-first view, our total evidence consists of facts we know, ­irrespective of whether they are facts about our mental states. We are in no position to use facts we don’t know as evidence. When we acquire new evidence in perception, we do not first acquire unknown evidence and then somehow base knowledge on it later. Rather, acquiring new evidence is acquiring new knowledge. That knowledge need not itself be based on further evidence, nor is it evidence for itself in some non-trivial way. But it is evidence for or against potential answers to questions to which we do not yet know the answer.

Equating evidence with knowledge helps reconnect epistemology with other fields. For one of the ways in which it marginalized itself was by depicting evidence as utterly unsuited to its role in science. The evidence for a well-confirmed scientific theory is typically a matter of public record. At least to a first approximation, it consists of facts intelligibly related to the theory and available to be known by anyone of suitable ­intelligence and training who takes the trouble to find out. It does not consist of facts about the present mental states of scientists or anyone else, facts that are no matter of public record and whose evidential relation to the theory itself has never been properly explained. Although the fact that a physicist believes a physical theory may raise its probability, the link requires an auxiliary sociological hypothesis and is hardly an ­evidential relation of the sort with which physics typically deals.

If our evidence is what we know, the evidence differs between the good and bad cases, contrary to what skeptics and many other epistemologists assume. For in the good case but not the bad, the subject’s evidence includes the fact that she has hands. Although one can stipulate an alternative sense for the word “evidence” in which the evidence is the same, the challenge is to give epistemological significance to the new sense. The preceding reflections suggest that any idea that the two cases are ­evidentially equal is no basic insight but a product of misconceived epistemological theorizing. In particular, the fact that for all one knows in the bad case one is in the good case does not entail that one has the same evidence; it just means that for all one knows in the bad case one has the same evidence as in the good case. One is not always in a position to know whether one’s evidence includes a given proposition. Although we might prefer a notion of evidence that does not work like that, we have no right to expect one.

A knowledge-first approach discourages trying to explain knowledge in terms of belief. We may even try the reverse, explaining belief in terms of knowledge. Here is a simple picture. Beliefs are the products of cognitive faculties whose function is to ­produce knowledge. When and only when all goes well, beliefs constitute knowledge. Even if something goes wrong, the belief may still be true, just as someone’s scheme for getting rich may fail while they become rich by an unintended chain of events. Thus knowing is the successful state, believing the more general state neutral between success and failure. Knowing corresponds to doing, believing to trying. Just as trying is naturally understood in relation to doing, so believing is naturally understood in relation to knowing.

If justification is the fundamental epistemic norm of belief, and a belief ought to constitute knowledge, then justification should be understood in terms of knowledge too. Indeed, a belief is fully justified if and only if it constitutes knowledge. Although your belief that you have hands is fully justified, the corresponding brain in a vat’s belief is not. But the brain in a vat has a good excuse for believing that it has hands, because, for all it knows, its belief that it has hands is justified, since, for all it knows, it knows that it has hands. Confusion between justifications and excuses undermines much talk of epistemic justification.

Some beliefs fall shorter of justification than others. In that respect we can grade beliefs by their probability on the subject’s evidence, that is, on the subject’s knowledge. A theory of evidential probability can be developed along such lines. It fills a gap ­between purely subjective probabilities, Bayesian credences (“degrees of belief”), and purely objective chances in indeterministic physics. When we ask how probable a theory is on our evidence, we want something less dependent on our doxastic state than a credence but more dependent on our epistemic state than a chance.

The most salient feature of knowing as the focus of epistemology is that it is a world-involving state. For it is factive: knowing that P, unlike believing that P, entails that P. Thus the state of the external environment constitutively constrains one’s ­epistemic state. More specific factive epistemic states include perceiving that P and remembering that P. But even believing involves the world in another way. For the external environment constitutively constrains the contents of most intentional states; belief and knowledge are no exceptions. Croesus knew and believed that he was rich in gold. Despite being in the same internal microphysical states, Twin-Croesus on Twin-Earth neither knew nor believed that he [Croesus] was rich in gold. He had no knowledge or beliefs about Croesus, since he never had any suitable contact with him, however indirect. Nor did he know or believe that he [Twin-Croesus] was rich in ...