Looks like Pettigrew has written about logical omniscience/ignorance and logical learning, which I’m guessing is relevant to value ignorance (unranked worlds?) and value learning, as per his choosing for changing selves.
Choosing for changing selves is also very relevant for the preferences of past selves, and may give some alternative approaches to the TRIA around weighting care for other time slices preferences differently depending on the strength of the psychological connection.
How do hyperintensions play in to the four solutions discussed this week?
Wouldn’t it be weird if value ignorance/learning - logical ignorance/learning - hyperintensions - value ignorance/learning. Or if there is some simple representation that solves all of them.
Egre - Logical Omniscience Summary
-
Core Problem: Belief models assume beliefs are closed under logical consequence, leading to unrealistic implications
-
Two Main Issues:
- Closure under equivalence: If expressions are synonymous, believing one means believing the other
- Closure under consequence: Believing p means believing all logical consequences of p
-
Four Solutions:
- Primitive Propositions: Treat propositions as primitives rather than sets of possible worlds
- Impossible Worlds: Allow logically impossible worlds to be epistemically possible
- Structured Meanings: Represent propositions as n-tuples of constituent denotations
- De Re Belief with Acquaintance Relations: Use different counterpart relations for different modes of presentation