If you think emotional states have representational content, it seems reasonable to think that there are rational constraints between the having of a certain emotion (say, feeling regretful that one has dropped something on one’s foot) and the having of a certain belief (say, believing that one has dropped something on one’s foot). Now, I imagine that some would want to question such a connection, but it seems at least a decent position to consider something like:
- it is rational to regret that p only if it is rational to believe that p.
But now suppose that we think that for theoretical purposes (say in characterizing instrumentally rational action) we should really be talking in terms of partial beliefs rather than all-or-nothing beliefs. In the official idiom, it seems, we dispense with talking about “believing that one has dropped something on one’s foot” and instead talk of things like “believing to degree d that one has dropped something on one’s foot”. (I’ll come back later to the question about whether we just ditch the all-or-nothing belief talk).
What then should we say about the rational connections between doxastic and emotional states? How are emotions rationally constrained by belief? Here’s a very natural thing to write down:
- it is rational to regret that p to degree d only if it is rational to believe that p to degree d.
The trouble with this is that I’m not sure I really understand the notion of “partial regret” that is now being talked about. Of course, I understand the idea of intensity of regret: I might regret insulting someone with a greater intensity than I regret forgetting my umbrella this morning. But “degrees of regret” in the intensity sense aren’t obviously what we want in this connection (I’m tempted to say they’re obviously not what we want). But do we really any other grip on the notion of a degreed emotion?
Of course, some people are likely to have similar sceptical thoughts about the notion of belief–finding all-or-nothing belief talk familiar home turf, and talk of partial belief rather mysterious. But the cases, to me, seem only superficially similar. To begin with, I think I had some pre-theoretic grip on the notion of degree of belief/confidence (though even here I think that there is a phenomenological intensity sense of “degree of confidence” that needs to be cleared out of the way). And even if I were to give up the pre-theoretic grip, I’ve at least got a theoretical/operational grip on the notion of degree of belief through decision-theoretic connections to action.
With partial emotions, I’m all at sea—things like regret seem to me, pre-theoretically, all-or-nothing (setting aside differences of felt intensity). And neither do I have a natural operational/theoretical grip on such partial emotions to reach for. I’d be very glad if someone could convince me that I do understand the notion, or point to literature where such issues are discussed!
Another strategy, I suppose, is to think of all-or-nothing belief as distinct from the degreed notion. If that’s the case, then we could formulate the connection between beliefs and regret just as originally stated. This’d be interesting to me, since previously I’ve never really been clear what’s lost if we ditch all-or-nothing belief-talk (and ensuing puzzles like the lottery paradox) and only appeal in our theories of mind to the partial beliefs. But if other emotional states with intentional content have rational connections to all-or-nothing beliefs, then it seems we’ve got a real theoretical role for such states.
Of course, this line of thought gives urgency to puzzles about how to relate partial beliefs and all-or-nothing beliefs — (e.g. all-or-nothing belief as partial belief above a certain threshold). That’s a whole literature in itself.
What do people think? Am I being really naive in thinking there are rational connections like the above to get worried about? Do they require reformulation when we introduce partial beliefs, or (as suggested at the end) is this a way of arguing the importance of retaining all-or-nothing belief talk as well as partial belief talk? Can anyone make sense of the notion of a partial emotion (when distinguished from the phenomenological-intensity reading)?
—where the modal expresses an idealized ignorance). The second is an interesting epistemic constraint on true counterfactuals I call “Bennett’s Hypothesis”.
‘ is read as ‘it is determinate that’:
be defined to be 
. This quantity—how little one accepts the negation of a proposition—might be thought of as the upper bound of an interval whose lower bound is the probability of A itself. So rather than describe one’s doxastic attitudes to known indeterminate A as being “zero credence” in A, one might prefer the description of them as themselves indeterminate—in a range between zero and 1.
J Robert G Williams 