Representing a location in space requires two things: an anchor point, and a code (or coordinate system) to define other locations relative to that anchor point. Recent work has shed light on the latter, providing evidence that the default ‘format’ of visuospatial representation may be polar coordinates (i.e., angle/distance relations). Yet the former remains a topic of debate. For example, a classic distinction in the realm of spatial navigation research pits representation relative to landmarks against representation relative to boundaries. Here, we exploit the polar format of spatial representations to propose a new method for assessing the locus of spatial representation. Specifically, we show that from simple localization errors we can infer the anchor point from which observers localized a target point. We highlight a few basic demonstrations of this method and discuss possible applications for further research on spatial representation.

Our world is filled with statistical information: from dice rolls to lotteries, we often act based on our impressions of probability. Yet the human mind is not wired to reason about truly probabilistic events, often imposing structure on data or events where no such structure exists (as in ‘illusory correlations’). Here, we consider a case study in intuitive statistics: disjunctive events. For example, participants are asked to imagine a button that, when pressed, has a 1 in 100 chance of yielding a prize. They are told to imagine pressing that button 100 times. Across several paradigms, we show that people overestimate the probability of this disjunctive event — in stark contrast to classic demonstrations where people underestimate such probabilities (e.g., when iteratively selecting marbles from jars with replacement). These results reflect a tendency to view events as causally connected in illusory ways; implications for other domains of reasoning are discussed.

A variety of phenomena related to the oblique regions of space have been observed across modality and across domain. For instance, the classic ‘oblique effect’ describes a deficit in visual acuity for oriented lines in the oblique regions of space, and classic ‘prototype effects’ describe a bias to mis-localize objects towards the oblique regions of space. While there has been speculation that some ‘oblique-related effects’ share a common mechanism, many of these effects are explained in very different terms. The visual oblique effect itself is often understood as arising from coding asymmetries in orientation-selective neurons in the brain, whereas motor oblique effects have been described as arising from gravitational cues and/or physical limitations of the arm. Are these really distinct effects? Here, we show that individuals show stable oblique biases across these two modalities, suggesting that these effects may have a common cause.

Mechanistic explanations reveal the rich causal structure of the world we inhabit. For instance, an explanation like “A clock ticks because an internal motor turns a gear which moves the hands” explains a feature of the clock (i.e., the fact that it ticks) by describing the parts and actions that cause it. People often seek out such explanations, as they may be particularly valuable to understanding the world. However, are mechanistic explanations truly a single class of explanation? Here, we distinguish between two subtypes of mechanism: constitutive and etiological. We argue that this distinction, long made by philosophers of science, has cognitive consequences: People treat these two kinds of explanation differently and prefer one kind over the other. We discuss implications for understanding mechanism and for explanation preferences more broadly.

People do not think in isolation. Whether purchasing a new product on Amazon, deciding what movie to watch, or evaluating scientific evidence, we often rely on aggregated sources of information (e.g., product ratings or reviews) to make decisions. Indeed, the internet has given rise to unprecedented levels of aggregated information, to the extent that it is difficult to imagine anything for which we might not be able to find summary information. In other words, what we know (or think we know) is constrained not just by our own knowledge, but by the knowledge of our community (Sloman & Rabb, 2016). Yet this raises a question: what happens when a community of knowledge is not in agreement? Here, we assess this question by pitting cases of high confidence against cases of high consensus. Results from two experiments show that 1) individuals are sensitive to both confidence and consensus; 2) individuals utilize such information in a predictable but context-dependent manner; and 3) perceptions of confidence and consensus influence judgments and decisions in a substantial way, even when individuals are not aware of the contrast between them. Taken together, the findings suggest that individuals are highly sensitive to variability in aggregated information – rather than merely an average – and that these ‘summary statistics’ of aggregated information have a substantial, reliable impact on decision-making.

Visual number adaption is a widely accepted phenomenon. This paper advances an alternative explanation for putative cases of the phenomenon. We propose that such cases may simply reflect observers adapting to the items in perceived displays, rather than their numerical quantity. Three experiments motivate consideration of this novel proposal and call into question the evidential basis for received formulations of the number adaptation hypothesis.

What makes an explanation satisfying? Much work hasinvestigated explanatory preferences for things like animalsand artifacts, but how do explanation preferences manifest ineveryday life? Here, we focus on the criminal justice system asa case study. In this domain, outcomes critically depend on howmembers of the system (e.g., lawyers, jurors) generate andinterpret explanations. We investigate lay preferences for twodifferent classes of explanations: those that appeal to‘mechanistic’ aspects of a crime (i.e., how the culpritcommitted the crime) vs. ‘teleological’ aspects of that crime(i.e., the purpose of the crime). In two experiments, wedemonstrate that people have a systematic preference for'motive' accounts of crimes (analogous to a teleologypreference) at different stages of the investigative process. Wediscuss these findings in light of a broad literature on thecognitive basis of explanation preferences. We also discussimplications for the criminal justice system.

Over the past two-and-a-half decades, numerous empiricalstudies have demonstrated a relationship between numbersand space. A classic interpretation is that these spatial-numerical associations (SNAs) are a product of a stablemental number line (MNL) in the mind, yet others haveargued that SNAs are a product of transient mappings thatoccur in working memory. Importantly, although the latterinterpretation has no implications for the representation ofnumber, the former suggests that the representation ofnumber is inherently spatial. Here, we tease apart questionsof spatial representation (à la an MNL perspective) andspatial strategy (à la alternative accounts). In a novel place-the-number task, we demonstrate that numbersautomatically bias spatial attention whereas other ordinalsequences (i.e., letters) do not. We argue that this isevidence of an inherently spatial representation of numberand explore how this work may help answer futurequestions about the relationship between space andnumber.

Does the human mind contain a task-general ‘randomization machine’? Stable biases of randomization have been identified that span multiple domains and modalities, in both lower-level perceptual tasks and in higher-level cognitive tasks. The stability of such biases indicates that the mind may rely on a stable set of properties to create and perceive randomness. But what computational principles support randomization? Here, we approach this question by building a computational model of human randomization that generalizes across spatial and numerical tasks. We show that simple computational heuristics capture higher-order properties of human-generated random sequences, in both numerical and spatial randomization tasks each with many possible options. Furthermore, we show that human behavior in both types of tasks can be approximated by the same low-dimensional model, implying that a domain-general set of computational principles may underlie randomization behavior in general.

Many studies have investigated the roles that area and number play in visual quantity estimation. Yet, recent work has shown that perceived area is not equal to true, mathematical area. This simple fact calls into question many findings in numerical cognition and suggests a new theoretical perspective: that area estimation plays a dominant role in visual quantity estimation. We examine two ‘case studies’: (1) a ‘general magnitude’ account of visual quantity estimation, which posits bi-directional influences between area and number. In contrast with prior work, controlling for perceived area reveals a unidirectional relation between area and number (Experiments 1 and 2), and (2) acuity of area and number estimation (Experiment 3). We show how an understanding of the perception of area forces a reevaluation of several findings concerning the relative acuity of number and area estimation. Combined, and in contrast to many prior studies, our findings suggest a dominant role of area in visual quantity estimation.