United States

In deciding “Which has more?” of some items, we rely on cues in the environment to decide how to answer: both in the form of linguistic cues and object form. Linguistic cues may appear as a classifier or “quantizer” like “ounces” as in “Which has more in ounces?” On the other hand, object cues have to do with the form of the entity and how object-like it is. Both linguistic and object cues can present ambiguities and leave the desired dimension up for interpretation. In the two experiments presented here, we examine how people decide whether to count or measure when presented with two item sets that vary independently in both numerosity and total size. Our findings suggest that dimension selection in quantifying is a process that relies not only on one dimension, but on effects of both linguistic indications and object factors.

CogSci 2025

Linguistic and visual cues in counting and measuring 

decision making

psychology

reasoning

### Welcome to CogSci Conference 2025!

The 47th Annual Meeting of the Cognitive Science Society was a hybrid meeting held in San Francisco. 

<div style="position:relative;padding-top:0;width:900px;height:500px;"><iframe style="position:absolute;border:none;width:100%;height:100%;left:0;top:0;" src="https://online.fliphtml5.com/ebtyf/amvr/"  seamless="seamless" scrolling="no" frameborder="0" allowtransparency="true" allowfullscreen="true" ></iframe></div>

#### About

The Cognitive Science Society brings together researchers from around the world who hold a common goal: understanding the nature of the human mind. The mission of the Society is to promote Cognitive Science as a discipline, and to foster scientific interchange among researchers in various areas of study, including Artificial Intelligence, Linguistics, Anthropology, Psychology, Neuroscience, Philosophy, and Education.

The Society is a non-profit professional organization and its activities include sponsoring an annual conference and publishing the journals Cognitive Science and TopiCS.

#### Our History 

* **Society Creation**<br>
The Society was incorporated as a 501(c)(3) non-profit professional organization in Massachusetts in 1979. The organizing committee included Roger Schank, Allan Collins, Donald Norman, and a number of other scholars from psychology, linguistics, computer science, and philosophy. 
<br><br>
* **Conference Creation**<br>
The first conference on cognitive science was held at La Jolla, California in August, 1979, and has occurred annually since then. The proceedings of each conference are published, and those from most years are available through Lawrence Erlbaum Associates, Inc. The annual proceedings of the Cognitive Science Conference represent a major source of information on new work and new ideas in the scientific study of thinking. In 1990, the Society, with help from an anonymous donor, established the David Marr Prize for the best student paper at each annual meeting.
<br><br>
* **Journal Creation**<br>
The Journal, Cognitive Science, began publication in 1976, and is now published by Wiley-Blackwell. The Executive Editor is currently Richard P. Cooper of Birkbeck, University of London, and there are 18 Associate Editors and a 30-member editorial board. It serves as the premier outlet for research reports that intersect two or more disciplines. Copyrights for articles published in the journal are held by the Society. The Governing Board of the Cognitive Science Society voted in late 2006 to found a new journal, Topics in Cognitive Science (topiCS). The Editor in Chief is Wayne Gray, Cognitive Science Department, Rensselaer Polytechnic Institute. The journal seeks to fill a niche not occupied by Cognitive Science Journal or other cognitive science journals. Membership in the Society includes a subscription to Cognitive Science and TopiCS. Copyrights for articles published in the journal are held by the Society.
<br><br>

#### Code of Conduct

By attending the CogSci 2025 Conference, you are required to adhere to the society’s **[Code of Conduct](https://drive.google.com/file/d/1ChPuihLy6jE_BWqfO7J2KKgX35JW2zsM/view?usp=sharing)**.
<br><br>


You need to log in with the email address you registered with. 

Login credentials were sent to you from Underline -  subject line "Welcome to the CogSci 2025 Conference". Please be sure to check your spam/promotional inbox  if you do not see an email confirmation right away.





Please log in to join this event.

To access the site, please register [**here**](https://cognitivesciencesociety.org/registration/).

If you are registered and feel like you are seeing this message by mistake, please make sure you are logged in with the same email that you registered with. 

Please register!

The 47th Annual Meeting of the Cognitive Science Society presents the latest research across cognitive science and highlights the theme of Cognition in Context.

poster

As large language models (LLMs) advance, there is growing interest in using them to simulate human social behavior through generative agent-based modeling (GABM). However, validating these models remains a key challenge. We present a systematic two-stage validation approach using social dilemma paradigms from psychological literature, first identifying the cognitive components necessary for LLM agents to reproduce known human behaviors in mixed-motive settings from two landmark papers, then using the validated architecture to simulate novel conditions. Our model comparison of different cognitive architectures shows that both persona-based individual differences and theory of mind capabilities are essential for replicating third-party punishment (TPP) as a costly signal of trustworthiness. For the second study on public goods games, this architecture is able to replicate an increase in cooperation from the spread of reputational information through gossip. However, an additional strategic component is necessary to replicate the additional boost in cooperation rates in the condition that allows both ostracism and gossip. We then test novel predictions for each paper with our validated generative agents. We find that TPP rates significantly drop in settings where punishment is anonymous, yet a substantial amount of TPP persists, suggesting that both reputational and intrinsic moral motivations play a role in this behavior. For the second paper, we introduce a novel intervention and see that open discussion periods before rounds of the public goods game further increase contributions, allowing groups to develop social norms for cooperation. This work provides a framework for validating generative agent models while demonstrating their potential to generate novel and testable insights into human social behavior.

Validating Generative Agent-Based Models of Social Norm Enforcement: From Replication to Novel Predictions

Planning, or the ability to simulate and execute a sequence of steps toward a goal, is crucial for success in many activities. However, common tasks used to measure planning often fail to correlate with one another, suggesting they may not assess the same underlying skill. To explore a novel measurement of planning, this study examined performance on four planning mini games, a non-planning control game, and three standard planning tasks. Results revealed that the planning mini games exhibited stronger intercorrelations than traditional tasks, suggesting they may capture a more consistent and unified planning construct. Notably, two of the selected mini games emerged as particularly promising paradigms for assessing planning skills with reduced confounds from processing speed. These findings provide initial evidence that mini games such as those explored here could complement or replace traditional cognitive planning tasks, offering an appropriately complex evaluation of the multifaceted skill of planning.

Evaluating Planning Through Play: Exploring the Use of Mini Games to Assess Planning Abilities

Intrinsic motivation plays a crucial role in shaping exploration and learning, yet its specific contributions to causal discovery remain underexplored. This study examines the impact of three intrinsic motivation metrics—entropy, information gain, and empowerment—on causal learning outcomes. Across two experiments, participants engaged in interactive tasks requiring them to infer causal structures through exploration. Results indicate that information gain and empowerment significantly predict learning success, whereas broad, undirected exploration (entropy) does not. These findings suggest that learners optimize causal discovery by prioritizing actions that maximize information and control, rather than engaging in indiscriminate exploration. Our study offers insights into how strategic exploration facilitates causal reasoning and how these principles can be applied to machine learning.

Understand Intrinsic Motivation in Causal Learning Through Exploration

People often rely more on their prior beliefs than the presented evidence when evaluating arguments. We investigate the cognitive mechanisms underlying this phenomenon. We hypothesise that when individuals encounter an argument that is congruent with their beliefs, it activates related information in memory. For belief-congruent arguments, people should therefore be more likely to both correctly recognise previously encountered information and incorrectly recognise new information as previously seen. To test this, we first investigated the effect of participants' beliefs about political claims on their evaluation of corresponding arguments that varied in quality. We then employed a surprise memory test to assess participants’ recognition memory for these arguments. While we replicated the finding that prior beliefs drive argument evaluations, prior beliefs did not affect memory performance for all arguments in the same way. Our results indicate that individuals may use prior beliefs to aid memory only when the memory task is difficult.

Differential Memory for Belief-Congruent versus Belief-Incongruent Arguments Cannot Explain Belief-Driven Argument Evaluation

We report an experiment that examines the relationship between insight and analytic learning and relational concept acquisition. This paper introduces a novel paradigm in which warmth judgments were included throughout a relational category learning task. Specifically, some subjects completed this category learning task through classification, whereas others completed it through inference. Additionally, some subjects made warmth judgments throughout this task, whereas others did not (control). The results revealed no evidence that warmth judgments impacted concept learning. More importantly, we find that subjects who engaged in classification reported greater and more rapid increases in warmth judgments than subjects who engaged in inference. These findings directly paralleled subjects’ learning patterns, wherein classification subjects showed evidence of more rapid learning, whereas inference subjects displayed more gradual, stepwise learning. Taken together, the present results suggest that classification involves more insight-based learning, whereas inference seems to involve more analytic learning.

The Role of Insight and Analytic Learning During Concept Acquisition

Multiple learning processes contribute to successful goal-directed actions in response to changes in physiological states and environments. Among them, explicit strategies play a crucial role, enabling rapid and flexible sensorimotor adaptation. Yet, how the training target distribution impacts strategy discovery remains poorly understood. To address this, we conducted a visuomotor adaptation reaching task that isolated explicit strategy. We manipulated two training distribution features in a 2×2 between-participants design (N = 50/group): the spatial arrangement (dense vs. distributed) and the target number (2 vs. 8). To pinpoint the strategies participants adopted, all groups periodically reached to a shared generalization target without feedback. Learning was faster with fewer and denser targets. Strikingly, the training target number had no effect on generalization, but those trained with densely arranged targets adopted simpler yet flawed strategies, leading to poor generalization. These findings suggest that training distribution influences strategy discovery—potentially steering participants toward overfitting, non-generalizable solutions.

Overfitting of Explicit Strategies during Sensorimotor Learning

Statistical learning (SL) is a domain-general learning mechanism necessary for multiple areas of cognitive development. The present study investigates whether children can simultaneously track temporal and spatial visual statistics and how individual differences in cognitive abilities and early language experience relate to SL. Fifty-eight hearing children aged 4–6 years (mean = 5.8) completed a novel visual SL paradigm, tracking the spatiotemporal statistics of four cartoon alien triplets. Cognitive control, receptive vocabulary, and auditory SL were also assessed to measure individual differences. Children achieved 56% accuracy on 2AFC test trials, performing above chance and demonstrating learning of complex patterns. For children under 6.5 years (n = 28), visual SL performance was positively associated with receptive vocabulary (r = 0.65) and cognitive control (r = 0.56). Future testing with deaf children in oral-speech or bilingual (ASL/English) programs will explore how language experience shapes SL capacities, offering insights into early cognitive development.

How language shapes learning: Visual statistical learning in deaf and hearing children

Minsky and Papert's (1969) book *Perceptrons* is often remembered as the book that (counter-productively) ended neural network research for nearly two decades. One of the authors' main results was that perceptrons (under reasonable limitations) cannot detect if a pattern is fully connected. Perhaps less known, to their initial surprise, the authors also showed that if guaranteed there are no holes in an image, perceptrons *can* detect if a pattern is fully connected. Given the simplicity of perceptrons, it seems reasonable to think that they might suggest a lower bound for what humans can visually detect without moving their eyes. If so, the results on connectedness suggest some counter-intuitive findings about human perception, namely that we should be able to learn to solve 2D mazes at a glance and detect how many objects are in an image at a glance (i.e., subitize) even when the number is large.

What Perceptrons Might Tell Us About Our Own Abilities

Explaining why events occurred involves solving different information-processing problems: inferring what actually happened (causal inference) but also highlighting a subset of the causes that contributed to the outcome (causal selection). While past research has investigated causal inference and causal selection separately, we report results of an experiment (N=284) examining how people solve both problems jointly, as is the case in real-world explanation settings. We find evidence that participants infer the state of unobserved variables on the basis of available evidence, and observe common behavioral signatures of causal selection. However, explanation preferences deviate in important ways from the predictions of a computational model combining existing theories of causal inference and causal selection. In particular, participants were disproportionately likely to select inferred over observed variables. We suggest a possible preference for producing explanations that allow the explainee to benefit from inferential work performed by the explainer.

Unifying inference and selection in singular causal explanation

Creativity is a complex and multifaceted construct that has been linked to various cognitive, emotional, and personality factors. This study explores the interplay between fluid intelligence, creative reasoning, personality traits, and mood states in predicting divergent thinking performance. We hypothesize that extraversion, openness to experience, positive moods, and creative reasoning would predict performance on divergent thinking tasks. A total of 120 young adults participated in the study, completing assessments on fluid intelligence (APM), creative reasoning (CRT-Reasoning), creativity (CRT-Creativity, TCT-DP), personality traits (Big Five Inventory 2), and moods (Current Mood Scale). Results revealed a negative relationship between fluid intelligence and divergent thinking performance, suggesting that higher fluid intelligence may be associated with more convergent thinking strategies. Extraversion emerged as a significant positive predictor of divergent thinking, supporting the idea that sociability and external engagement foster creativity. Openness to experience did not significantly predict divergent thinking, indicating that its impact may vary across domains of creativity. Mood states, especially hopelessness, were negatively associated with creativity, but the frequently found association between positive moods and creativity could not be replicated. These findings underscore the importance of considering cognitive, emotional, and personality factors in the study of creativity and suggest potential pathways for future research into how these elements interact to shape creative thinking.

Downloads

Next from CogSci 2025

Validating Generative Agent-Based Models of Social Norm Enforcement: From Replication to Novel Predictions

Similar lecture

The paradox of trait impressions in naturalistic contexts: rich information, sparse predictive cues

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from CogSci 2025

Validating Generative Agent-Based Models of Social Norm Enforcement: From Replication to Novel Predictions

Similar lecture

The paradox of trait impressions in naturalistic contexts: rich information, sparse predictive cues

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads