Partial mental simulation explains fallacies in physical reasoning

Abstract

People can reason intuitively, efficiently, and accurately about everyday physical events.Recent accounts suggest that people use mental simulation to make such intuitive physical judgments. But mental simulation models are computationally expensive; how is physical reasoning relatively accurate, while maintaining computational tractability? We suggest that people make use of partial simulation, mentally moving forward in time only parts of the world deemed relevant. We propose a novel partial simulation model, and test it on the physical conjunction fallacy, a recently observed phenomenon (Ludwin-Peery, Bramley, Davis, & Gureckis, 2020) that poses a challenge for full simulation models. We find an excellent fit between our model's predictions and human performance on a set of scenarios that build on and extend those used by Ludwin-Peery et al. (2020), quantitatively and qualitatively accounting for deviations from optimal performance. Our results suggest more generally how we allocate cognitive resources to efficiently represent and simulate physical scenes.

An approximate representation of objects underlies physical reasoning.

“…The use of body approximations is in line with the general proposal that people’s intuitive physics is not a perfect simulation, but rather relies on principled shortcuts and workarounds (Bass et al, 2021; Battaglia et al, 2013; Smith & Vul, 2013; T. D. Ullman et al, 2017).…”

Prospective search time estimates reveal the strengths and limits of internal models of visual search.

“…As we discuss above, internal models of visual search may alternatively produce search time estimates based on an approximate probabilistic model, and without any reference to target-distractor similarity. More work is needed to determine where internal models of visual search fall on the continuum between being simulation-based and being rule-based, noting that much of human knowledge may lie somewhere in between these two ends (Bass et al, 2021; Hegarty, 2004).…”

Children’s active physical learning is as effective and goal-targeted as adults’.

“…While this provides a valuable benchmark, examining more computationally frugal and fallible models of physical reasoning has the potential to add nuance to our notions of about what constitutes informative dynamics for bounded learners (cf. Bass, Smith, Bonawitz, & Ullman, 2021;Ullman et al, 2017). That is, it is interesting to consider how computational limitations might shape what constitutes a physically informative action for human learners.…”

“…to add nuance to our notions of about what constitutes informative dynamics for bounded learners (cf. Bass et al, 2022;Ullman et al, 2017). That is, it is interesting to consider how computational limitations might shape what constitutes a physically informative action for human learners.…”