beliefs about your present sensations (e.g. your belief that your nose itches or that your head aches), beliefs about your present thoughts (e.g. your belief that you are currently thinking of New Jersey), and your perceptual beliefs (e.g. your belief that the page you're reading right now is covered with black marks). Even if you cannot think of any independent considerations to offer in support of these beliefs, these beliefs look like good candidates for justification and knowledge. Consider the belief that you're in pain, for example. It wouldn't be an arbitrary thing to believe if it's formed in response to the kinds of experiences you'd have when touching a hot iron or skinning your knee.27
1.38 One of the oddities of the infinitist view is that it will try to account for the fact that the beliefs just mentioned can be justifiably held by positing an infinite series of reasons where the reasons it posits seem to be more epistemically problematic than the belief that they're supposed to support. If you believe that you're in pain and someone asks you to identify a good reason to think that you are, you might come up with something. You might say that you're sweating and showing the standard physiological responses to pain, and you might point out that you need painkillers. But even if we have these reasons, they might seem otiose.
1.39 While considerations such as these might be good reasons for someone to believe that you're in pain, why would you need them to justifiably form this belief? How could such considerations account for the fact that it would be right for you to be much more confident that you're in pain than you are confident that any of the supporting reasons you've just mentioned are true?
1.40 This general idea is captured famously by Ludwig Wittgenstein in his posthumous On Certainty (1969) in the following passages, where Wittgenstein suggests it would not be promising to adduce what is less certain to one in the service of supporting what is more certain to one.
My having two hands is, in normal circumstances, as certain as anything that I could produce in evidence for it. That is why I am not in a position to take the sight of my hand as evidence for it.
(OC, §250)
If a blind man were to ask me “Have you got two hands?” I should not make sure by looking. If I were to have any doubt of it, then I don't know why I should trust my eyes. For why shouldn't I test my eyes by looking to find out whether I see my two hands? What is to be tested by what?
(OC, §125)
Wittgenstein, in making these remarks, is taking the proposition that one has hands to be (like the proposition that one is in pain) the sort of thing that you know if you know anything at all.28 And such bedrock certainties (what Wittgenstein calls “hinges”) are difficult to support by appealing to any kind of evidence that's more certain to one than these bedrock certainties themselves.
1.41 At some point – in the case of what is most obvious to us – it seems that the ability to just see that something is so using the finite reasons at hand should be enough for knowledge and justification if these epistemic standings are attainable. The intuitive force of this point is hard to ignore, and if this point is conceded, then it's unclear why there would have to be some further infinite set of reasons waiting in the wings for our beliefs to constitute knowledge or to be justifiably held. If you can just see that something is true, it wouldn't be right to describe your belief as being held on an arbitrary basis.
1.4 Coherentism
1.42 Think about your two books, Your Book of Beliefs and Your Book of Justified Beliefs. The Supporting Justified Belief Rule tells us that a belief in the first book, B1, earns a place in the second iff it's supported by other beliefs that have a place in the second book. If we adhere to this rule and there's a finite (but non‐zero) number of entries in Your Book of Justified Beliefs, infinitism must be mistaken. Let's now consider an alternative to infinitism, coherentism.
1.43 Imagine that you have Your Book of Beliefs in hand. A team of epistemologists has promised to send you a copy of Your Book of Justified Beliefs once they finish a thorough investigation of you, your beliefs, and your belief‐forming habits. Curiosity gets the better of you, so you start to wonder which entries in your book will be entries in theirs. The Supporting Justified Belief Rule tells you this much: if any of your beliefs is included in both books, it's because there's something in both books that supports it. You start to look for connections between the entries. You discover that you can group the entries into categories like this:
Entries that fit with a significant number of other entries and do not conflict with any other entries.
Entries that conflict with other entries and are not supported by a significant number of entries.
Entries that neither conflict with other entries nor fit with other entries.
Entries that fit with a significant number of other entries but also conflict with other entries.
According to the Supporting Justified Belief Rule, the entries that fit into the second and third categories won't be justified. You won't expect to find these entries in Your Book of Justified Beliefs. The fourth category is tricky. On the one hand, some of these entries might receive strong support from other entries and so you might think that the conflict doesn't really threaten them. Some of these entries might receive weak support and look bad in light of well‐supported entries. Let's set these aside for the time being.
1.44 The best candidates for entries in Your Book of Justified Beliefs will be those in the first group – viz. entries that fit with a significant number of other entries and do not conflict with any other entries. Question: could it be that all it takes for a belief to be justified is for it to fit into the first category? Could mutual support between beliefs be all that's required for these beliefs to be justified? This is indeed what the coherentist thinks. As Catherine Z. Elgin (1996) states the idea, beliefs that are justified are parts of a system where the parts are “reasonable in light of one another” (1996, p. 13).
1.45 There is are two key components to this core idea: (i) items that aren't contained in Your Book of Beliefs simply won't have a direct bearing on whether beliefs that belong to a coherent system are really justified or not; and, second, (ii) justified beliefs are justified because of their place in a system of mutually supporting items. Instead of thinking of your justified beliefs as forming a structure like a tower or pyramid with foundational beliefs at the bottom and inferential beliefs at the top, think about your system of justified beliefs as forming a piece of woven cloth. The strength of a piece of woven fabric has all to do with the interlocking warp and weft strings. It doesn't require any unsupported supporters.
1.46 As you examine your beliefs, you might find that some beliefs aren't supported by any further beliefs. These beliefs, according to the coherentist, lack the kind of support required for justified beliefs. There is not some other source of rational support that isn't some further belief. As Donald Davidson remarked, “nothing can count as a reason for holding a belief except another belief” (1986, p. 141). The coherentist will say that there are various causes of our beliefs (e.g. experiences, sensations, apparent memories, etc.), but will insist that it's just relations between beliefs that determine whether they're justified.29
1.47 Even the critics of coherentism will acknowledge that there are cases where coherence seems to play an important role in the justification of our beliefs, but the crucial question isn't whether the overall coherence of a system of beliefs plays some role in the justification of belief, but whether the justification of our beliefs could be wholly determined by the coherence of the system. Coherentists like to think of justified beliefs as part of a web of belief that has a sufficiently high degree of coherence, in part because we