Temporalis muscle biomarkers from routine brain MRI and risk of dementia in two independent cohorts.

To excel as epistemic agents, people must recognize their epistemic limitations, including the bounds on their beliefs—for instance, whether beliefs are based on limited evidence (e.g., Cusimano & Lombrozo, , ) or generated by fallible mechanisms (e.g., Plunkett, Buchak, & Lombrozo, ). Accurately representing one's epistemic weaknesses is sometimes characterized in terms of “intellectual humility,” a virtue that involves appropriate attention to and recognition of one's intellectual limitations (e.g., Bowes et al., ; Davis et al., ; Lynch, ; Sgambati & Ayduk, , ; Leary et al., ; Porter et al., ; Whitcomb, Battaly, Baehr, & Howard‐Snyder, ). Over the last few years, both philosophers and psychologists have refined characterizations of intellectual humility (IH) and its measurement (e.g., Ballantyne, ; Davis et al., ; Hoyle, Davisson, Diebels, & Leary, ; Krumrei‐Mancuso & Rouse, ; Leary et al., ; Porter et al., , ), including its connection to intra‐ and interpersonal dimensions of belief. For instance, if individuals recognize their own ignorance and fallibility, they might be more likely to subject their own beliefs and reasoning to appropriate levels of scrutiny (Bowes, Ringwood, & Tasimi, ), and to constructively engage with other views and with the people who hold them (Bowes, Blanchard, Costello, Abramowitz, & Lilienfeld, ). Here, we explore one way in which IH could shape belief. Beyond a potential impact on the of people's beliefs, might IH be associated with how beliefs themselves are construed? The hypothesis we explore is that IH might be associated with a tendency to construe beliefs in terms. This is because probabilistic beliefs implicitly acknowledge both ignorance and fallibility. To illustrate, consider two individuals, Blake and Ping, who both believe that nuclear power is safe. For Blake, this belief is binary: Blake believes that nuclear power is safe (as opposed to believing it is not). For Ping, this belief is probabilistic: Ping thinks there is some probability (greater than 50%) that nuclear power is safe. The critical contrast between Blake and Ping is not in belief precision, as we need not assume that Ping can offer a precise number rather than a range (e.g., 51%–100%). And crucially, this contrast is not in belief strength (see Ballantyne, ; Baron, ): if both individuals were prompted to assign some probability to the claim that nuclear power is safe, they could both choose 80%. Yet, Blake and Ping's beliefs could still differ in a substantive way. We propose that because Ping's belief is expressed probabilistically, it has the following two properties that differentiate it from Blake's: (i) it recognizes the uncertainty associated with whether the proposition is true, and thus the ignorance of the believer, and (ii) by virtue of entailing some probability associated with the proposition being false, it builds in the possibility of being wrong (see Buchak, ). In other words, probabilistic beliefs might reflect the believer's recognition of their ignorance and fallibility—two notions central to IH. The distinction between binary and probabilistic belief finds precedent in prior theoretical and empirical work. Philosophers debate whether beliefs are fundamentally binary (or potentially trinary, with belief, disbelief, and suspension; e.g., Byrne, ; Harman, ; Horgan, ) or continuous, corresponding to something like strength of belief or credence (e.g., Pettigrew, ; Jeffrey, ; for hybrid views see Buchak, ; Jackson, ; Weisberg, ). These debates have partial analogues within psychology, with some research documenting binary notions of belief (e.g., Johnson, Merchant, & Keil, ), and other work exploring continuous notions, often involving subjective probability (e.g., Chater, Tenenbaum, & Yuille, ; see also Mellers et al., ; Yaniv & Foster, ). In more recent work, Vesga, Van Leeuwen, and Lombrozo () presented participants with vignettes in which a character arrives at a belief for epistemic or nonepistemic reasons. For example, a character could come to believe that someone is innocent of a crime based on the evidence (an epistemic reason) or out of loyalty (a nonepistemic reason). They found that participants were significantly more likely to classify the belief as probabilistic (e.g., that the character believes the other is innocent with some probability) in the epistemic case than in the nonepistemic case. This finding supports the psychological reality of the binary‐probabilistic belief distinction, and moreover suggests that people might find it natural to classify beliefs in both ways, with classifications potentially reflecting the epistemic versus nonepistemic basis for belief.