We discussed a range of things subjects "do" in response to being in an epistemic position. (The language of "doing" and "choices" here is not meant to prejudge questions about what's voluntary.) I divided things people do into things that were here-and-now doxastic responses of some sort, and things that were instead activities (perhaps purely deliberative activities) aimed at improving one's epistemic position, by uncovering/remembering more evidence, or gaining new reflective insights, or clearing things up in your mind, etc.
Examples of doxastic responses include all-out commitments, like belief and disbelief, and also withholding (aka suspending or being agnostic). I mentioned doubting and being certain as other responses, perhaps not explicable in terms of the others. Or doxastic responses may include grades of confidence, with Bayesian/probabilist models being a familiar paradigm but consider also views that posit less structure, or more structure, to our confidence. If we have both the graded attitudes and the all-out attitudes, as many think, what is the relation between them? A hard problem. One strategy is to think there is some threshhold, no doubt dependent on context, and perhaps also dependent on what proposition is in question, and believing is a matter of being confident above that threshhold. I mentioned another strategy, which is to think that there is some psychological act of "settling" or making up your mind, which needn't be accompanied by any change in your degree of confidence. Believing would be a matter of having settled in one distinctive way. (Would withholding be a matter of settling in another way, or a matter of not having settled?)
We presume there are epistemic reasons or justification for being such that you doxastically respond in some ways right now rather than others. We presume there are also often epistemic reasons or justification for changing your doxastic response in some ways. These are sometimes labeled your "static" normative properties and your "dynamic" normative properties. On some views the relation between these is straightforward, on other views it's more complex. Perhaps there are also distinctively epistemic reasons for settling, if this is a notion we end up making use of. Perhaps there are also distinctively epistemic reasons for putting things on your "to-do list" of helpful activities (including purely deliberative activities). But even if there are distinctively epistemic reasons of these latter sorts, insofar as they come apart from the first two notions, they're outside our focus in this class.
With the all-out attitudes, we can talk of having any amount of reason or justification for having them (or, in the same neighborhood, having some more reason or justification for having one of them rather than another). Or we can talk of having "enough" or "sufficient" reason or justification for having them. I tend to think in terms of the former.
What are the normative statuses a doxastic response can have? Some philosophers find it helpful to begin theorizing with the notion of knowledge. Others find it helpful to begin with the notion of a well-founded, properly based response (e.g., a well-founded belief). Others find it helpful to begin with the notion of "propositional" or prospective justification. (We discussed distinctions that Crispin and others draw within this category, using terms like "warrant" and "entitlement". I use "justification" in the generic way that Crispin uses "warrant.") Others find it helpful to begin with the notions of "reasons" or "evidence." In some philosophers' mouths, these terms are just synonyms for "justification." But I pointed out some commitments or biases that may go togther with these terms, that might not be innocent, and that we should be attentive to at the start of our theorizing. With "reason" the issues have to do with a thread in ethics that thinks of reasons as facts you might not yet be aware of. With "evidence" the issues have to do with a tendency to think of evidence as constituted by a set of propositions. Trying to model all epistemic differences between two subjects as deriving from a difference in such sets is a serious constraint on our theories. Perhaps it's a constraint we should embrace, but that's a substantive choice; and we should be on guard against being led into such choices just by our vocabulary.
I find it most natural and productive to begin theorizing with the notion of "propositional" or prospective jutification, but you may start from a different place.
I discussed why the label "propositional" might mislead. One motivation for the label is that the output side of the relation is a proposition: the proposition you ought to believe. On some theories of belief, though, propositions are not the objects of belief (even after allowing for the variety of things called "propositions"). Another awkwardness here is how this motivation carries over to doxastic responses other than belief, such as withholding or graded confidence. When you have prospective justification to withhold belief in p, or to be 2/3 confident in p, what are the propositions that are justified in those cases? It's hard to identify a single, non-higher-order proposition to be the output. It's much easier to think of the output as being some prospective attitude, rather than a proposition.
What makes me hesitate most with the label "propositional" is not the picture it encourages of the output side as always a proposition, but rather the picture it can encourage of the input side as also just being some propositions. It's common to hear propositional justification described as some kind of (presumably non-deductive) "logic of support." This is a substantial commitment in two respects: first, it buys into the same picture discussed above with "evidence", where the determinants of your normative status are just a set of propositions. Second (I neglected to mention this in the meeting), it buys into a picture of the normative status as one that abstracts from a subject's capacities and limitations. If p deductively entails q, then it ipso facto will propositionally justify q, no matter how complex the relation between p and q. That's a kind of normative relation it may be useful to think about. But when I talk about "propositional" or prospective justification, I don't want to be forced to idealize in that way. I want to be able to say that two subjects both have justification to believe p, but only one of them also has justification to believe q, because only she "sees" that q follows from p. (It isn't promising to think of the additional leverage that subject has in terms of some further proposition in her evidence set.)
(This wasn't discussed in seminar, but I'll mention it here and we'll come back to it in later weeks.) Tim Willenken, who was a Pitt grad, articulated an interesting question about the relation between prospective and doxastic or dynamic justification: if a subject has prospective justification to believe p, must there be available to her some epistemically permissible way to form that belief? Willenken thought the answer had to be yes (see also the "Generalized Internalism Requirement" in Kolodny's "Why Be Rational?"). I think the answer may be no. It can happen that there's some doxastic response that your epistemic position makes appropriate, but that there's no epistemically good dynamic route from your current attitudes to that response. (Buridan's Ass may be a helpful analogue: the ass has reason to be walking to the left or to the right, but getting himself moving in one of the directions rather than the other has to be non-rational.)
We discussed a schematic case where you're justified in believing q because of inferring (or being in a position to infer) it from some other proposition(s) p you're also justified in believing. When asked what in such cases is or constitutes or provides your justification for q, philosophers give different answers: It's the proposition p. Or: it's the reasons you have for p. Or: it's the fact that (or event where) you have some reasons or other for p. (Perhaps we need to include in your justification for q also something about what makes this inference available to you, when others are not.)
We introduced the notion of "immediate justification" as justification for a q where there is no such p. The notion of "basing," or whatever we call the relation that holds between prospective justification and well-founded beliefs, should also be applicable when the prospective justification is immediate. So we shouldn't conflate "basing" a belief on some justification with inferring that belief from some other beliefs. (The latter may provide one example of the former.)