Joshua Edward Pearson
Belief Revision Revised
A Puzzle about Weak Belief
Composing Composers
Under Review - comments welcome!
Might-Counterfactuals, Might-by-Cases and Causal Independence
Email for a draft; handout will soon be available
A Project on Counterfactual Knowledge
Email me if you're interested and would like to chat. As a taster, here's a work-in-progress handout from early on in this project.
Weak Belief and Heuristics
On the back-burner, but I'm still very interested. Email me if you'd like to chat!
Eastern APA, 2025. Thanks to Abe Matthew!
I outline a novel counterexample to the principle of belief revision, Anticipation: if both learning e and learning not-e would render belief in p unjustified, you cannot now be justified in believing p. If I’m right, not only is the leading theory of belief revision false, so are various recently proposed weakenings. I defend a new theory that correctly predicts the failures of Anticipation I argue for, predicated on the simple idea that one is justified in ruling out possibility just in case that possibility is sufficiently improbable.
I present an intractable puzzle for currently popular view that belief is weak — the view that expressions like 'S believes p' ascribe to S a doxastic attitude towards p that is rationally compatible with low credence in p. The puzzle concerns issues that arise on considering beliefs in conditionals. I show that proponents of weak belief either cannot consistently apply their preferred methodology when accommodating beliefs in conditionals, or they must deny that beliefs in conditionals can be used in reasoning.
Suppose you believe on independent grounds that Verdi is Italian, Bizet is French, and Satie is French. To your surprise, you then learn that all three composers are compatriots. What should you believe? Some argue you should be cautious and become ambivalent as to whether all three composers are French or all three are Italian. Surprisingly, if that's right, Composers wreaks epistemological anarchy: a wide variety of epistemological principles turn out to be false. Those resistant to this anarchy instead argue you should be bold and conclude that all three composers are French. I endorse the anarchy. But I do so here in a unique way. Existing approaches to Composers side exclusively with either the cautious or bold reaction, ruling the other out as irrational. This is undesirable: both judgements look reasonable. I outline a new approach to Composers that successively captures this permissive element of the case.
This paper concerns a distinctive type of case concerning counterfactuals, introduced by Pollock (1981, p. 254) and recently discussed by Boylan and Schultheis (2021) (henceforth ‘B&S’). These cases challenge a principle I call ”Possibility Preservation”, validated on many leading theories of counterfactuals, including Lewis’s (1973). In response, B&S propose a Lewisian semantics of counterfactuals that rejects comparability among possible worlds. I argue their approach comes at too high a cost: it not only gives up Possibility Preservation, but also two further intuitive principles—Might- by-Cases and Specificity—and leads to counterintuitive verdicts about the probabilities of counterfactuals. I offer an alternative. Drawing on sequence semantics for conditionals and the ”Epistemic Thesis”—that might-counterfactuals express uncertainty about their corresponding would-counterfactuals—I develop an alternative model of these cases that captures our probabilistic intuitions, preserves Specificity and renders failures of Might-by-Cases not problematic but independently motivated.
I am developing a longer project on counterfactuals, that can be split into three parts:
​
Part (1) concerns counterfactual Skepticism. Hájek famously defends the view that most counterfactuals are false, and so not known. His arguments now deploy premises which are widely contested. However, I'll argue that new arguments for counterfactual skepticism can be constructed without using these assumptions. Indeed, I'll develop two such arguments, one for each side on a central debate in the literature: whether the controversial "Conditional Excluded Middle"(CEM) is true.
​
However, I'm not a counterfactual skeptic. Parts (2) and (3) will develop theories that can respond to these new skeptical arguments. (2) develops a modified version of Lewis's semantics, directed to those who think CEM is false. Part (3), directed at those sympaethic to CEM, proposes to respond to the new skeptical arguments by synthesizing two cutting-edge theories: sequence semantics for counterfactuals, and probabilistic normality theories for knowledge. ​
The idea that there is a weak sense of 'belief' — by which belief in p is rational compatible with low credence that p — is becoming increasingly popular. At the same time, this idea has generated various puzzles, including those given by myself, Helena Fang, and Richard Teague. I suspect that a satisfactory response to them requires a more radical departure from our understanding of weak beliefs than has thus far been realized. The key is to realize that weak beliefs are rationally formed using heuristics. It is well known that applying multiple good heuristics to the same problem can lead to inconsistent results; these puzzles are merely a demonstration of this.
I'm interested in the interaction between belief, knowledge, conditionals and probability. Much of my research is about how and when your full beliefs can be defeated by learning new information. Right now, I'm examining the consequences of this research to how beliefs should relate to action and inquiry. Further, I am developing a project on counterfactuals that explores analogies to my work on defeat. See my research statement here and papers below.
​
Asymmetric Inductive Knowledge
On the back-burner, but I'm still very interested. Email me if you'd like to chat!
Consider Alice, who is unsure whether the urn in front of her is ‘Uniform’ and contains 1,000 marbles, or is ‘Mixed’ and contains 500 green and 500 blue marbles. She randomly samples 100 marbles, sees that they are all green, and believes on this basis that the urn is Uniform. Anti-skeptics should concede that in at least some version of this case, Alice’s belief can constitute knowledge. I argue that this anti-skeptical position is only tenable if we maintain that Alice’s knowledge is asymmetric in various surprising ways. One such striking asymmetry implies a kind of ego-centrism; for instance, I argue that while Alice can know before looking at her sample that if all the marbles in her sample are green, the urn is Uniform, she is, surprisingly, not able to know of some other equally-sized random sample S that if all the marbles in S are green, the urn is Uniform. This raises a puzzle: can any plausible theory of knowledge accommodate such bizarre asymmetries? I outline a theory that can.