ABSTRACT. Over the last decade, evidence has been accumulating that even if they give biased responses, many people show signs of ‘conflict detection’: They are less confident on tasks on which their intuitive responses are at conflict with logical principles (i.e., incorrect), than to tasks on which their intuitive responses are correct. The five contributions to this symposium present new empirical evidence on the robustness and the relevance of this phenomenon, expanding it to a wider range of tasks, across languages, and into educational/training practice (* = presenter): 1. Haen*, Janssen, Verkoeijen, De Neys, & Van Gog test (consistency in) conflict detection in probabilistic reasoning tasks of varying complexity; 2. Purcell & Bagó* investigate conflict detection in evaluating arguments in reasoning about climate change, and address the role of priors and partisanship; 3. Janssen*, Mutis, & Van Gog show that conflict detection also occurs when evaluating authenticity of real and deepfake videos; 4. Białek*, Paruzel-Czachura, & Domurat show that conflict detection is found both when reasoning in the native and a foreign language; and 5. Franiatte*, Boissin, & De Neys present evidence that the conflict detection phenomenon has both scientific and practical relevance, as it predicts the effectiveness of debiasing training.

ABSTRACT. There are two main approaches to studying insight problem solving: The first addresses the cognitive mechanism of restructuring, i.e., fundamental changes in problem representations occurring after initial failure to solve a problem. The second approach focuses on the phenomenology of insight, i.e., the Aha! experience and other accompanying emotions. In our view, the mechanism of interest is sudden restructuring which has been postulated as defining feature of insightful solutions. Restructuring is, however, challenging to grasp empirically. What is needed are reliable measures of solvers’ internal problem representations and, importantly, how they change during the solving process. In this symposium, we seek to find common ground regarding possible methods to assess restructuring by bringing together researchers from Germany, Belgium, Estonia, Lebanon and the USA. Topics to be discussed are trace measures such as multiple importance-to-solution ratings (Danek) and online nearness-to-solution ratings (Stuyck) as well as the effect of sleep on memory restructuring (Sanders). The symposium will be rounded off with the presentation of a theoretical framework looking at the core processes underlying insight events in diverse contexts (Tulver), and a high-level perspective on different types of insight and creativity (Dietrich).

ABSTRACT. In a December 2021 interview, Francis Collins, the departing director of the NIH noted "to have now 60 million people still holding off of taking advantage of lifesaving vaccines is pretty unexpected. It does make me, at least, realize, ‘Boy, there are things about human behavior that I don't think we had invested enough into understanding.’ “

Decision-making models--intended to explain and predict volitional behavior-- are extreme abstractions from the biology of Homo sapiens. They invariably pick out only cognitive/rational mechanisms. To the extent that an abstraction captures salient features it is valuable. To the extent that it fails to do so, it can be misleading. It is proposed that by picking out only cognitive/rational mechanisms models of decision-making are far too abstract and removed from the biology to accurately capture behavior. A case is made for tethering cognitive/rational decision-making models to “lower level” noncognitive systems such as the autonomic, instinctive, and associative systems. Volitional behavior is then a blended response of these various systems.

To make this case I appeal to (i) data from cooperative economic decision-making tasks to support the blended response hypothesis; (ii) evolutionary and anatomical evidence for the tethered brain; and (iii) the neuroscience literature on affect and arousal to propose a lingua franca of communication and a control structure for the tethered mind. I conclude by explaining some real world behaviors with tethered rationality.

ABSTRACT. Preference reversals and other forms of incoherence violate some Von Neumann-Morgenstern axiom of rationality.

This has led some to conclude that incoherence is irrationality.

Our thesis is that incoherence is a feature, not a bug.

A useful way of thinking about incoherence is that it arises when System 1 (intuition) recommends a different course of action than does System 2 (deliberation).

Coherence implies one of two states of the world: both System 1 and System 2 are correct. both System 1 and System 2 are wrong.

One explanation for incoherence then is that a cognitive system is flagging to the decision maker that they are risk of a wrong decision. In this view, having two cognitive systems is an error minimization strategy and incoherence is a useful error detection mechanism. The paper reconsiders some results that have been described as demonstrating irrationality e.g. a judge’s tendency to become more likely to grant parole as the hearings approach lunchtime. I reconsider these results through an error minimization criterion and conclude that incoherence is precisely the outcome we would hope for in a world where people suffer cognitive fatigue. Incoherence is not necessarily an indication of irrationality.

ABSTRACT. Effective communication depends on reasoning about what others know and believe, and failures in executive functioning can disrupt the way adults reason about mental states. No study has revealed errors traceable to mental state reasoning processes, such as failures in interpreting premises, simulating possibilities, and formulating conclusions. We investigated a family of mental state reasoning errors. Online participants (N = 323) considered conditional inferences couched in terms of a mental state, e.g., Alia knows that if it’s rainy then the café is closed; It’s rainy. What follows? They generated natural conclusions using a novel sentence construction interface. Many participants spontaneously generated erroneous responses such as, Alia knows that the café is closed. This pattern reflects an “omniscience” error: reasoners erroneously imputed knowledge of a deductive consequence to an individual. We discuss the results in the context of recent proposals on epistemic inference.

ABSTRACT. The recent intense interest in conspiratorial thinking, among scientists and the public alike, largely comes about because people perceive (rightly or wrongly) that belief in conspiracies is irrational. How then, does conspiratorial thinking relate to measures of rationality? There have been few studies that have examined the associations of conspiracy belief with a comprehensive battery of rational thinking tasks that tap both epistemic and instrumental rationality. The Comprehensive Assessment of Rational Thinking (CART) provides an opportunity to do just that because one of the subtests on the CART assesses the tendency to endorse conspiracy beliefs. Converging analyses using the other 18 subtests and 4 thinking dispositions measured on the CART indicated that three variables were key predictors of the endorsement of conspiracy beliefs: superstitious thinking, actively open-minded thinking, and probabilistic reasoning. Theoretical consideration of these best predictors led us to rethink the classification of conspiracy belief as contaminated mindware and move instead toward a conception of conspiratorial thinking as a cognitive style. Conspiracy mentality measures should not be conceptualized and discussed as if the optimal score on them is zero. Like other style measures, optimality on conspiracy belief measures displays an inverted-U function.

ABSTRACT. Cognitive science is shaped by debates about the right paradigm to approach human rationality. This is exemplified by an ongoing dispute about Bayesian cognitive science. Some researchers emphasise the merits of Bayesianism in understanding and unifying different areas of human thinking. Others complain about fundamentalism, lack of falsifiability and mistaken notions of support. In my talk, I address this issue from a general perspective by asking about the status of Bayesian cognitive science: Is it a framework or a theory? Depending on the answer, what are the fitting criteria for evaluating it?

I will argue that frameworks and theories concern different aspects of models (or rather collections thereof). Roughly speaking, frameworks are about the language of models, while theories are about their content. According to the standard view (predictivism), theories are best evaluated by the ability to predict novel data. Frameworks, however, need to just model data and allow formulating theories. Bayesian scholars take quite diverse approaches in whether they use Bayesianism as a framework or propose a Bayesian theory. This observation helps to clarify some misconceptions that have come into this debate from critics but also some defenders of Bayesian cognitive science.

ABSTRACT. Modern societies are crucially dependent on the opinion of experts. The asymmetric relationship between experts and non-experts poses many socially impactful problems that have been subjected to the methodological lens of disciplines as diverse as philosophy of science, social epistemology, argumentation theory, and social psychology. We focus on one fundamental problem that lies at the intersection of these approaches: how lay reasoners should update their beliefs in some hypothesis or claim H given that some expert asserts that H. First, we propose to treat expert opinion as evidence from testimony, developing a broadly Bayesian model of both the experts’ reliability and the evidentiary impact of their testimony (in terms of Bayesian confirmation measures). Second, we show how our model impacts the discussion concerning expert opinion both in social epistemology and in argumentation theory. As for the former, we show how the idea of “epistemic deference” amounts to ignoring the fallibility of real experts and fails as a general strategy of belief updating in the face of expert opinion. As for the latter, we elaborate on the recent discussion of “ab auctoritate” reasoning to clarify and evaluate this much-disputed argumentative strategy. By adopting a model of expert reliability based on Jeffrey conditionalization, we argue for a new characterization of ab auctoritate, more faithful to the complex epistemic interplay between laymen and experts.

ABSTRACT. Creativity and association are known to be closely related. The association of creative people tends to diverge, but it remains unclear how the tendencies and strategies of one’s free associations are determined. Our study examined how exposure to word pairs that have a far or short semantic distance can influence people’s thinking space. We used distributed word representations based on large-scale language models to examine semantic divergence in a chained free association (CFA) task (Gray et al., 2019). In Experiment 1, we examined the effects of the level of relevance between pairs of two words and the number of word pairs presented. Participants repeated CFA tasks twice. Between the tasks, they encountered one or five pairs of words, either strongly related (S condition: e.g., publication/book) or weakly related (W condition: e.g., typhoon/skin), and they assessed the association between paired words. Our result showed that simply evaluating the relevance of word pairs with low association levels can trigger divergent thinking. However, the amount of score that the W condition increased was smaller than the amount of score that the S condition decreased. Therefore, in Experiment 2, we explore the influence of both word relevance strength and diversity on divergent thinking.

ABSTRACT. The conjunction fallacy (CF) describes a pattern where individuals disregard the principles of probability by deeming certain conjunctive statements more probable than the individual parts of those statements. The most well-known scenario (due to Tversky and Kahneman, 1982, 1983) involves a progressive young woman called Linda who is consistently assessed as more likely being a feminist bank-teller than a bank-teller. The fallacy may be fruitfully reconstructed as the normatively correct assessment of something else than probability, for instance of inductive confirmation (as suggested by Tentori, Crupi and their colleagues in their, e.g., 2007, 2013, 2017) or coherence (Jönsson and Shogenji, 2017). We show that these accounts are, however, formally incapable of accounting for CF when information involved in the CF-scenario may be best represented by a probabilistic chain network. We therefore suggest a novel account of CF according to which the fallacious reasoning is due to an assessment of explanatory power. In other words, we suggest that CF arises due to a reasonable abductive inference. Our approach not only accounts for several variations of CF, but also enables us to formulate novel predictions regarding human reasoning within similar scenarios.

ABSTRACT. According to a popular account in epistemology, a set of propositions is justified, if it is coherent (Olsson, 2022). Similarly, a new proposition should be accepted if it coheres with the accepted body of beliefs. But what is coherence? And what justifies the above claims? To address these questions, various Bayesian measures of coherence have been introduced. These measures are all based on the prior probability distribution over the corresponding propositional variables. Criticizing this ``static'' account of coherence, we propose that the coherence of an information set is related to how well the information set responds if each of the propositions is confirmed by an independent and partially reliable information source. Working out this idea will show that the proposed ``dynamic'' perspective has various advantages and solves some outstanding problems of coherentist epistemology with implications for the study of reasoning and argumentation.

ABSTRACT. In this paper, I defend an extended objective Bayesian account of rational belief formation. In epistemic utility theory, rational beliefs should minimise the agent's expected inaccuracy (their expected 'distance from the truth'). Thus, given a measure of inaccuracy (a 'score'), the agents should be able to compute their rational degrees of belief by minimising the expected distance from the truth, given available evidence. However, there are many non-equivalent measures of inaccuracy. Based on concepts from information theory, I argue that for a learning agent the logarithmic score is the most adequate measure of inaccuracy. In particular, the value of the logarithmic score can be interpreted as the length of the shortest possible inquiry. By re-interpreting the source coding theorem for symbol codes, which establishes the entropy-limit (theoretical lower bound) for optimal compression codes, I show that the expected value of the logarithmic score (the entropy of the agent's beliefs) corresponds exactly to the most efficient inquiry (i.e. the smallest amount of questions needed to determine the true state of the world, given available evidence). This creates tight connections between rational belief formation, Bayesian inference, optimal inquiry, and confirmation.

ABSTRACT. To investigate, in an ecological way, how people evaluate the truthfulness of universal and probabilistic statements we introduce a modified version of the Wason Selection Task. Participants see four decks of cards (instead of four cards), with each card representing a unique observation, and they are asked to turn as many cards as they deem necessary to judge if a given statement is true or false, both for the observed sample (deductive task) and for an imaginary reference population (inductive task). Participants encounter universal (“All P are Q”) or probabilistic statements (“more/less than x% of P are Q”; between-subjects). We use abstract, realistic neutral, and realistic polarizing statements (within-subjects) to gauge the effect of content and motivated cognition on evidence search and judgments. We also investigate the effect of incentives: half of the participants receive an endowment for each turn, correct (incorrect) deductive judgments are rewarded (penalized), and turning a card incurs a cost (other half: fixed participation fee). We develop prescriptive models of evidence search (universal and probabilistic statements imply different normative strategies) and contrast them with actual behaviors in two online experiments. We find significant effects of incentives, statement type and content on evidence search and judgments.

ABSTRACT. Which social decisions are favoured by intuitive processes? And which by deliberative processes? The dual-process approach to human sociality has emerged in recent decades as a vibrant and exciting area of research. Scholars have studied the cognitive underpinnings of various social decisions, including cooperation, altruism, truth-telling, reciprocity, deontological judgments, aggression, fake news sharing, and engagement with hateful content online. Simultaneously, theoretical frameworks have been developed to organize the experimental regularities. In this symposium, we sit at the intersection of theory and experiments, bringing together leading experts to illuminate the latest developments in this research domain. Our discussions will cross various methodologies, from experiments with adults, adolescents, and even three-year-old children, to the broad perspectives offered by meta-analytic approaches, as well as the development of novel theoretical frameworks. Our objective is to provide a platform that not only introduces attendees to the exciting challenges in this field but also fosters the synthesis of diverse perspectives. We aim to bridge gaps in current knowledge, stimulate new lines of inquiry, strengthen existing collaborations, and encourage new ones. This symposium is an opportunity to contribute to and shape the evolving narrative of how intuition and deliberation influence human sociality.

ABSTRACT. Jonathan Schooler A theoretical discussion of the various factors that impact the state of openness to experience, such as exposure to art, that may impact the creative process by changing individuals’ sensitivity to input and ideas that they might otherwise overlook.

Yoed Kenett Insight relates to the creation of new links between unrelated concepts, or the creation of “shortcuts” in memory that facilitate analogical mapping and memory restructuring. Computational tools from network science and natural language processing are enabling quantitative and empirical research into insight-related memory restructuring effects.

Laura Macchi The creative act of restructuring required in insight problem solving involves an interpretative function cross-cutting language, thought and perception, depending on an interplay between many factors including intentions and context. An important role is played by the unconscious analytic thought that includes a relevant, analytic, and goal-oriented search.

Moshe Bar Both the content and the progression pattern of mindwandering are determined by our overarching state of mind (SoM), a construct encompassing the pilar components of cognition, including perception, attention, creative thinking, openness to experience and mood. These elements are clustered together, depending on internal processes as well as external context.

ABSTRACT. In Skovgaard-Olsen and Klauer (in review), we present a new Conflict model of moral judgments which builds on the CNI model of Gawronski et al. (2017) but extends it to solve two problems affecting the latter. First, the CNI model makes an invariance assumption in its model parameters that was found to be problematic in Skovgaard-Olsen and Klauer (2023). Second, the CNI model does not have a built-in mechanism for handling cases where participants are conflicted between deontological and utilitarian response tendencies. The new Conflict model solves both problems.

In this talk, a pre-registered study will be presented which serves three goals. The first is to establish discriminant validity of the conflict detection / resolution path of the Conflict model by introducing a between-participants manipulation affecting these model parameters. The second is to establish construct validity of its conflict detection (Con) parameter by correlating this Con parameter with a conflict detection index. The third goal is to compare the Conflict model with a model that reduces the complexity of the processing tree by removing the conflict resolution (Res) branches of the Conflict model and restricting its Con parameter to governing Skip options alone.

Gawronski, B., Armstrong, J., Conway, P., Friesdorf, R., & Hütter, M. (2017). Consequences, norms, and generalized inaction in moral dilemmas: The CNI model of moral decision-making. Journal of Personality and Social Psychology, 113, 343–376.

Skovgaard-Olsen, N., & Klauer, K. C. (2023). Invariance Violations and the CNI Model of Moral Judgments. Personality and Social Psychology Bulletin. Doi: 10.1177/01461672231164888

Skovgaard-Olsen, N., & Klauer, K. C. (in review). Norm Conflicts and Morality: The CNIS Conflict Model of Moral Decision-Making.

ABSTRACT. Moral identity refers to the importance individuals place on morality as part of their self-concept and was introduced to bridge the gap frequently emerging between judgment and action. Previous research has examined measures attempting to capture moral identity and criticized most instruments for underestimating its complexity. Following this reasoning, we explored the multidimensionality of moral identity. In a series of five studies (total N = 1442), we collected a comprehensive list of moral traits and analyzed their underlying structure. The resulting measure, the Moral Identity Profile (MIP), consists of four components that embody distinct dimensions of a person's moral identity: (1) Connectedness describes the desire to be connected and responsive to others, (2) Truthfulness represents the importance of truth and authenticity, (3) Care indicates the willingness to help and support others, and (4) Righteousness captures the importance of ethical values and justice. We examined the MIP dimensions in relation to personality dimensions (HEXACO), indicators of prosocial and antisocial behavior, and moral beliefs (Moral Foundations Theory and Morality-as-Cooperation Theory). The results showed the presence of distinctive relationships for the four dimensions, and even yielded opposing effects, demonstrating the conceptual richness of our approach and challenging previous unidimensional conceptualizations of moral identity.

ABSTRACT. Moral dilemmas often juxtapose a utilitarian resolution (for the maximization of the overall welfare) to a deontological resolution (consistent with prescriptive moral norms). Moreover, dilemmas are often sacrificial: whichever the choice made, the incurrence of at least one human life loss is required. When these dilemmas also include the moral agent as an involved character of the scenario, the typical utilitarian option allows to protect themself along with the majority of the characters involved and at the expense of a third person. The present study reviews the role of the moral agent within sacrificial moral dilemmas, reflecting of the implications of framing the utilitarian resolution as a self-sacrificial or a self-protective act. Sourcing from several academic databases, we focused on experimental papers investigating moral judgment with the use of sacrificial moral dilemmas. Following a sequential skimming procedure based on keywords, titles, and abstracts content analysis, 18 papers were finally taken into consideration. This contribution provides an overview of the use of textual dilemmas in experimental moral psychology, questioning if (and how) the utilitarian sacrifice framing may influence the description of moral judgment in different contexts and applications.

ABSTRACT. Philosophers have long been arguing the precise semantics of different deontic terms (e.g. may, obligated) within normative statements, e.g. the war ought to be stopped. There are subtle differences between deontic terms which often lead to conflation of meaning. These conflations hinder the development of logic of normative concepts. While various schemes have been proposed in deontic logic research to resolve this conflation issue, little research has been done on the human reasoning side of understanding such deontic terms. In this talk, we are going to introduce our proposed normative scheme with bitstring semantics that is expressive enough to cover the basic normative concepts in most mainstream normative schemes. Even though further confirmation is needed, our explorative experiments on human deontic reasoning have shown results that are consistent with our proposed scheme. Nonetheless, human reasoning can sometimes deviate from what is expected in the scheme. Hence, we are also going to discuss what are the possible causes (e.g. scalar implicatures & scalar diversity) of such deviation, what insight this explorative study has brought us for future research, and what is the implication of our proposed scheme in the development of deontic logic and knowledge representation of rules on human behaviours.

ABSTRACT. Scholars have been using hypothetical dilemmas to investigate moral decision-making for decades. While this approach has been invaluable, reflecting on a hypothetical dilemma is not the same as responding to a dilemma situation in real life. Whether people’s responses to hypothetical dilemmas truly reflect the decisions they would make in real life is unclear. We discuss a series of studies where participants had to make the real-life decision to administer an electroshock to either a single victim or allow a group of victims to receive a shock instead. Our results indicate both synergies and divergences between hypothetical dilemma judgment and real-life dilemma behaviour. We argue that hypothetical-dilemma research is invaluable for understanding moral cognition, but that more research should explore how and when real-life moral behaviour diverges from hypothetical judgment.

ABSTRACT. We designed a new experimental paradigm to investigate moral learning from consequences of previous decisions. Participants (total N=1601) faced a series of realistic moral dilemmas between two conflicting choices: one prescribed by a moral rule (typically 'deontologist') and the other favored by cost-benefit reasoning (CBR; typically 'utilitarian'). They observed the consequences of each decision before making the next one. In one condition, CBR-based decisions consistently led to good outcomes and rule-based decisions to bad outcomes. In the other condition, this contingency was reversed.

Across three studies, we consistently found systematic, experience-dependent changes in people's moral decisions over just 13 choices: participants adjusted how much they followed CBR versus moral rules according to which produced better consequences. Further, we use computational modelling to quantify individual variability in learning style and find that (1) most learning is model-based and (2) only some proportion of participants show evidence for metacognitive learning (i.e., learning about following rules vs. CBR as decision-making strategies rather than situation-specific behaviours). These learning effects transferred to general measures of moral attitudes and decisions in a donation task; however, this transfer was only robust for metacognitive learners, demonstrating the importance of this learning style for generalising to novel situations.

Overall, our findings demonstrate adaptive moral learning from consequences of previous decisions, which generalises to novel decisions with real-world implications. Individual differences in morality may thus be more malleable than previously thought.

ABSTRACT. When making inductive inferences, environmental restrictions can permit the sampling of specific items while systematically excluding others. These sampling frames may cause items to be collected based on their category membership (category frame) or their possession of a target property (property frame). Supported by Bayesian principles, individuals are less likely to generalise a property to novel items under a property frame compared to a category frame. The current work investigated whether sensitivity to this frames effect could be augmented via training with worked examples and feedback. Experiment 1 (N = 201) found that such training in either a property or category frame increased sensitivity to that frame relative to a no-training control. In Experiment 2 (N = 170), participants completed a property induction task before and after receiving training in both frames. A positive training effect was found, but only in participants with a poor baseline understanding of sampling frames. Experiment 3 (N = 197) also utilised a pre-post design but presented training on each frame in a separate block. Here, a positive training effect was found across all participants. This work emphasises how individual understanding of the effects of sample selection mechanisms can be enhanced with appropriate training.

ABSTRACT. This Stage 2 Registered Report concerns the relationship between rumination, a repetitive style of negative thinking, and conspiracy beliefs (Stage 1 protocol: https://osf.io/y82bs, date of in-principle-acceptance: 23/05/2023). Based on four pilot studies, we tested in a fifth, registered study whether brooding, a particularly dysfunctional form of rumination, contributes to conspiracy beliefs using a repeated-measures within-person experiment (N = 1,638). Mean difference scores (conspiracy beliefs at T2 minus conspiracy beliefs at T1) were significantly greater in the brooding condition than in the control condition. However, we could neither confirm that this effect was larger than the specified smallest effect size of interest of d = 0.20, nor conclude that it was too small to be of interest (i.e., smaller than d = 0.20). We explored how reflection, an analytic form of rumination, impacted conspiracy beliefs. We further discuss implications for theories about the formation of conspiracy beliefs, and efforts aimed at preventing or reducing unfounded conspiracy beliefs. Hopefully, this article sparks a discussion among conspiracy belief researchers about how smallest effect sizes of interest could be determined in a principled way based on real-world outcomes.

ABSTRACT. In this study, we investigated human causal learning in a continuous time and space setting. We find participants to be capable active causal structure learners, and with the help of computational modelling explore how they mitigate the complexity of continuous dynamics data to achieve this. We propose participants combine systematic interventions with a narrowed focus on causal dynamics that occur during and directly downstream of their interventions. This task-decomposition approach achieves comparable accuracy to attending to all the dynamics, while discarding almost half of the data. We argue this strategy makes sense from a resource rationality perspective: ignoring dynamics outside of interventions saves computational cost while the interventions naturally decompose the global learning problem into a series of more manageable sub-problems. We also find that when the causal relata are given real world labels, participants will use their domain specific priors to guide their structure inferences. In particular, individuals with accurate prior expectations were less likely to make the common local computations error of mistaking an indirect for a direct relationship. Overall, our experiments reinforce the idea that humans are frugal and intuitive active learners who combine actions and inference to optimize learning while minimizing effort.

ABSTRACT. When evaluating hypotheses in the light of some evidence, people show a well-documented tendency to place too much importance on the immediate impact of that evidence and comparatively neglect the base-rates, in apparent violation of the standards of Bayesian rationality. One perspective holds that this tendency stems from people’s inherent interest in the impact a piece of evidence has on belief in a hypothesis, rather than in conditional probabilities, effectively holding their judgments to non-normative standards. On another account, it stems from humans’ preference for representing probabilistic information via causal models. Base-rates rarely lend themselves to a causal interpretation, which leads them to be overlooked when computing posteriors. A major difficulty for comparing these perspectives is that in most usual situations, causality and evidential impact are closely related and often confounded. We propose a new experiment designed specifically to tease apart their respective contribution to subjects’ judgments. Our findings concur with the evidential-impact perspective in showing that subjects overall neglect base-rates even when those are given a causal interpretation. The causal status of the updating evidence does however appear to modulate subjects’ proclivity to engage in evidential-impact reasoning, as they do so more readily when that evidence is causal.

ABSTRACT. Analytic thinking dispositions are relatively stable, trait-like individual differences. Several different analytic dispositions have been described in the literature. They are known to correlate moderately, but it is unclear to what extent they reflect a common higher-order dimension, and to what extent they occur in varying combinations. In ongoing analyses, the present study (N = 400) explores the question whether there are different types of analytic thinkers, that is, different profiles based on combinations of the known dispositions. The study takes a person-centered approach and uses latent profile analysis to identify groups with different profiles. The study investigates whether the groups with different profiles differ in terms of demographics and a range of thinking and reasoning tasks. The findings will further our understanding of the construct of analytic thinking dispositions.

ABSTRACT. Decision-making styles are mostly measured by self-reported instruments which raises questions about their psychometric characteristics. The current study aims to address two of the frequent concerns related to these measures. Firstly, the consistency of decision-making across situations and secondly, a comparison of self-reported styles with an assessment provided by other people. For the first aim, 100 older adults (45-83 years) assessed their decision-making styles using the General Decision-Making Styles (GDMS) inventory and their decision-making in three specific decision situations. The correlation analysis indicates strong associations not only between the styles and specific decisions, but also across the three particular situations. For the second aim, the self-assessment of 84 participants was compared with an assessment provided by three other persons – a colleague, a close person and a family member. The scores of the self-reported decision-making styles were found to be closely related to the assessment by other people. Moreover, the assessments by other people were even more inter-associated than with the self-assessment. The two studies indicate that self-reported decision-making styles, as measured by the GDMS, are a reliable and valid indicator of the way in which people make decisions.

ABSTRACT. There is conflicting evidence regarding whether sophistication exacerbates ideological divisions on scientific topics. While some data suggest that it amplifies ideological divides, others argue that, except for the most contentious issues, sophistication leads people to align with scientific positions. We pit these two accounts—motivated reasoning versus lazy thinking—against one another and investigate whether sophistication heightens the influence of ideology on scientific beliefs within a novel context: Eastern Europe. Our findings reveal that while cultural conservatives were less inclined to adopt scientific positions on five out of seven issues—namely, climate change, the Big Bang, evolution, vaccines, and GMOs; ideological effects were non-significant for nuclear energy and homeopathy. However, we observed that sophistication, although positively associated with embracing scientific positions on most issues, also accentuated ideological divisions, albeit only for two topics: climate change and GMOs. Our subsequent goal is to understand the cognitive mechanisms underlying these effects. We designed two studies that examine whether people engage in effortful information processing to reach desirable conclusions. Effort is assessed using subjective, behavioral, and physiological indicators. These findings will contribute to clarifying whether the mechanisms of motivated reasoning or lazy thinking better account for reasoning concerning scientific beliefs.

ABSTRACT. High-level cognitive abilities, such as intelligence or creativity, can be viewed as the emergent property of a complex system of mental processes. This view departs from the usual perspective of intelligence as a variable that reflects an underlying psychological reality, determining performance on reasoning tasks. Instead, performance on reasoning tasks would be determined by the combination of various cognitive processes, mechanisms and abilities. In this perspective, individual differences in intelligence are not primarily interesting in and of themselves: the focus becomes modeling the response process that leads to performance, and understanding the multiplicity of variables that contribute to performance. This talk will summarize the perspective of intelligence as an ermerging construct with multiple determinants, old and new arguments in its favor (from Thomson 1916 to network models of intelligence tests), and its consequences for research, teaching, and clinical practice with intelligence tests. It will also present the draft of an integrative model of processes contributing to intelligent performance, and outline its possible uses in practice.

ABSTRACT. When reasoning about a claim, it makes sense to be more persuaded if lots of other people agree. But, there are many factors that make weighing the evidence behind a consensus complicated. For example, a consensus might be more or less informative depending on the type of claim, or whether each consensus member formed their opinions independently. These factors might also influence people differently depending on their own assumptions or preferences. In this study we used a mock social media platform to assess how persuaded people were by two factors: the presence of consensus (no consensus vs. consensus), and source independence (a consensus based on independent information sources vs. a consensus formed off shared, dependent sources). We varied these factors at both the group and individual level. At the group level, we assessed a third factor: whether people were influenced by the type of claim being reasoned about (we assessed 60 different claims divided into 4 categories). Almost everyone was more persuaded by consensus trials compared to no consensus trials. However, the strength of this effect was credibly stronger if the claim was likely to have a ground truth. We found that around one third of participants were sensitive to source independence. Of these, three quarters were more persuaded by a consensus based on independent sources, but the quarter who were more persuaded by dependent sources were persuaded just as strongly.

ABSTRACT. Idea generation is pivotal in creative cognition, problem-solving, and reasoning, especially when emerging as a sudden insight. The last two decades brought significant progress elucidating insight, by understanding its cognitive aspects through the study of neural correlates. Such state-of-the-art research on insight via cognitive and neural methodologies is the focus of this symposium. A recent breakthrough reveals that insight contributes to long-term memory retention, termed the 'insight memory advantage.' Jasmin Kizilirmak will present the neuro-cognitive underpinnings of this effect and its relationship with aging. Additionally, REM sleep has been shown to facilitate insight problem-solving, presumably by enhancing greater associativity that could favor memory restructuring. Théophile Bieth will present how changes in problem-related semantic associations in memory can prompt restructuring during insight and how sleep can influence such processes. The neural mechanism of such associative thinking, and especially generating free- vs. goal-directed associations, will be presented by Yoed Kenett. Insights are also associated with internal attention allocation. Carola Salvi will discuss the role of the visual cortex in this relationship, summarizing past and new results obtained via diffusion tensor imaging. Drawing on neurophysiological results, Ruben Laukkonen will introduce the 'Eureka Heuristic’ model, explaining insight within a computational neuroscience framework.

ABSTRACT. People frequently debate topics important to them on social media, and usually, arguments made by an individual are convincing to some but not others. We are interested in why people differ in their evaluations of such ‘everyday’ arguments. In two experiments, participants rated their beliefs about various political claims (e.g., ‘abortion should be legal’) and the quality of ‘good’ and ‘bad’ arguments about these claims. These ‘everyday’ arguments were designed to resemble those seen on social media, with good and bad arguments manipulated to have either strong or weak evidence supporting their conclusions, respectively. Experiment 1 finds a belief bias for everyday arguments; participants consistently rated good arguments as better than bad arguments, but argument quality ratings were also strongly correlated with individual beliefs. The effect of belief was around 3 times larger than the effect of argument quality. Experiment 2 replicates these results using a more transparent procedure; participants used their beliefs to assess the quality of the arguments despite being instructed not to, irrespective of whether they saw the arguments before or after rating their beliefs about the political claims. Our findings highlight that the same information can be perceived differently depending on a person’s prior beliefs.

ABSTRACT. The theory of mental models holds that they are end products of both language and perception. This talk shows how an integrated treatment of human thinking based on the construction of perceptual mental models can motivate smart and stable computer vision systems. I introduce a novel visual reasoning architecture that perceives the world by dynamically constructing and updating spatial models. The models represent the iconic structure of observations encoded in images and streaming video. The system can be queried in real-time with natural language spatial relations, "e.g., focus on what is to the left of the ___" to focus attention on portions of the input imagery. It handles restrictive adverbials ("to the left" vs. "directly to the left") and superlative expressions (e.g., "the uppermost object"), and it explains pragmatic spatial inferences. It is built on mReasoner, a general computational cognitive model of thinking and reasoning. I describe how it's been used for recent practical applications on an embodied robotic platform.

ABSTRACT. Cognitive scientists treat verification as a computation in which descriptions that match the relevant situation are true, but otherwise false. The claim is controversial: The logician Gödel and the physicist Penrose have argued that human verifications are not computable. In contrast, the theory of mental models treats verification as computable, but the two truth values of standard logics, true and false, as insufficient. Three online experiments (n = 208) examined participants’ verifications of disjunctive assertions about a location of an individual or a journey, such as: ‘You arrived at Exeter or Perth’. The results showed that their verifications depended on observation of a match with one of the locations but also on the status of other locations (Experiment 1). Likewise, when they reached one destination and the alternative one was impossible, their use of the truth value: could be true and could be false increased (Experiment 2). And, when they reached one destination and the only alternative one was possible, they used the truth value, true and it couldn’t have been false, and when the alternative one was impossible, they used the truth value: true but it could have been false (Experiment 3). These truth values and those for falsity embody counterfactuals. We implemented a computer program that constructs models of disjunctions, represents possible destinations, and verifies the disjunctions using the truth values in our experiments. Whether an awareness of a verification’s outcome is computable remains an open question.

ABSTRACT. Spatial reasoning plays a crucial role in our daily lives. The results of previous studies suggest that spatial mental representations involved in reasoning are rich in motor components, indicating a possible involvement of motor resources. Parallel studies on action observation suggest that humans are able to extract and simulate specific motor information when observing others moving. The present study investigates whether the observation of actions involving physical effort triggers a mental representation that slows down spatial reasoning. Participants observed actions that involved physical effort (heavy condition) or actions that did not involve physical effort (light condition) before tackling spatial reasoning problems. With this manipulation we can investigate the influence of different types of motor information on participants kinematic mental models and thus on the resulting inferences. Participants watching the video in the light condition can easily build a kinematic mental model based on motor information, which should improve their performance in the inference task. In contrast, participants watching the video in the heavy condition can build a kinematic mental model based on effort and fatigue; this type of motor information should influence performance compared to the light condition in terms of reaction times. The results confirmed this prediction.

ABSTRACT. Counterfactual thoughts are ubiquitous in daily life, we are drawn to imagine how past events could have turned out differently, and to construct alternatives to reality based on our “if only” conjectures. We rely on counterfactuals to understand the causes of past events, and we sometimes use them to justify past decisions. We also rely on them to form intentions for the future, and we sometimes use them to avoid repeating the mistakes of the past. Counterfactuals can amplify emotions such as regret and guilt, and social ascriptions such as blame and fault. This symposium contains research that demonstrates the far-reaching effects of counterfactuals in our everyday lives: Patrizia Catellani (Milan Catholic University, Italy) will describe how counterfactual messages can be used to combat fake news, compared to messages promoting analytical thinking. Kai Epstude (University of Groningen, Netherlands) will outline the role of counterfactuals in political ideology in the context of motivated counterfactual thinking. Ruth Byrne (Trinity College Dublin, Ireland) will discuss how counterfactual explanations lead people to switch from a risky to a safe decision (or vice versa) following a recommendation by an Artificial Intelligence decision support system compared to a human expert. Rachel Smallman (Texas A&M University, US) will consider the impact of counterfactual thoughts on health behavior. And Paul Henne (Lake Forest College, US) will describe how counterfactuals are generally used to make inferences about causality in a given situation. The symposium will end with a panel discussion on the practical uses and misuses of counterfactuals.

ABSTRACT. The “Wason selection task” is still one of the most studied tasks in cognitive psychology. We argue that the low performance originally obtained seems to be caused by how the information of the task is presented. By systematically manipulating the task instructions, making explicit the information that participants are required to infer in accordance with the logical interpretation of the material implication “if, then”, we found an improvement in performance. In Experiment 1, the conditional rule has been formulated within a relevant context and in accordance with the conversational rules of communication, hence transmitting the actual meaning of the material implication. In Experiment 2, a similar improvement has been obtained even without the realistic scenario, only by making explicit the unidirectionality of the material implication. We conclude that task instructions are often formulate neglecting the conversational rules of communication, and this greatly reduces the possibility to succeed in the task.

ABSTRACT. Which aspects of the description of a criminal event tend to be altered in fraudulent testimony? Is there a connection between the altered aspects and the words that describe them? Literature revealed a tight connection between lying and counterfactual-thinking, an embodied process that relies heavily on the motor system. Moreover, studies on language processing indicate that action verbs elicit stronger motor system activation than attention verbs, suggesting deeper cognitive processing. This investigation is based on the assumption that truth-telling involves a mental simulation of past events, whereas lying and counterfactual-thinking involve a counterfactual simulation in which the elements and/or relations of the factual simulation are altered, resulting in greater activation of the mental simulation network than truth-telling. Two predictions descend from this assumption: 1) Truth-tellers, Liars, and Counterfactual-Thinkers should focus more on events described with action verbs and this effect should be more pronounced in Liars and Counterfactual-Thinker; 2) Liars and Counterfactual-Thinkers should use more action verbs to alter the event. Results confirm participants' focus on events described with action verbs and the more use of action verbs to alter the event for Liars and Counterfactual-Thinkers.

ABSTRACT. Russian invasion of Ukraine is accompanied with considerable volume of fake news, which may undermine trust in the motives and actions of Ukraine and Western countries. It is therefore essential to search for and test interventions capable of reducing trust in such fake news. In one laboratory and two online experiments, we measure trust in true, false and fake news related to Russo-Ukrainian war and examine the effects of providing fake news refuting information. By manipulating the timing of such information, we effectively test whether it is more effective to Debunk (i.e., to refute fake news after people are already exposed to them) or Prebunk (i.e., to intervene before fake news start spreading). Across our three experiments we consistently find that while Debunk intervention significantly reduces the trust in fake news, and the effect lasts for at least a couple of weeks, Prebunk intervention is effective only (a) for a very short period time or (b) if it occurs just moments before exposure to fake news.

ABSTRACT. Should a health campaign emphasise the benefits of adopting a health behaviour (“If you give up smoking, you reduce your risk of lung cancer”) or the costs of not adopting it (“If you don’t give up smoking, you won’t reduce your risk of lung cancer”)? What factors should determine the choice of message? A large literature on so-called ‘message framing’ or ‘goal framing’ has held that these messages are equivalent but should be used in different contexts: for instance, with different types of behaviour (see, e.g., Rothman & Salovey, 1997; van’t Riet et al., 2016). That literature leaves unexplored how audiences understand the messages, which is likely an important factor in audiences’ response. We report a novel interview study, inspired by ‘think aloud’ protocols (e.g., Ericsson & Simon, 1993), which explores how participants interpret materials from a typical message-framing study. Interviews probe how participants understand central elements of the materials - including risk language, conditionals, and author intentions – and how persuasive they find the messages. Interview data are analysed using Template Analysis, pointing to participants’ conscious interpretations and responses. We draw from the analysis lessons for the message/goal framing literature to break an acknowledged theoretical impasse (van’t Riet et al., 2016).

Ericsson, K., & Simon, H. (1993). Protocol Analysis: Verbal Reports as Data (2nd ed.). MIT Press.

King, N. (1998) ‘Template analysis’, in G. Symon and C. Cassell (eds.) Qualitative Methods and Analysis in Organizational Research. Sage.

Rothman, A. J., & Salovey, P. (1997). Shaping perceptions to motivate healthy behavior: the role of message framing. Psychological Bulletin, 121(1), 3-19. https://doi.org/10.1037/0033-2909.121.1.3

Van’t Riet, J., Cox, A. D., Cox, D., Zimet, G. D., De Bruijn, G. J., Van den Putte, B., ... & Ruiter, R. A. (2016). Does perceived risk influence the effects of message framing? Revisiting the link between prospect theory and message framing. Health Psychology Review, 10(4), 447-459. https://doi.org/10.1080/17437199.2016.1176865

ABSTRACT. In lists of verbal probability expressions (unlikely, possible, very likely) “certain” has a privileged position. Most people place it at the top of a 0-100 percent probability scale. But statements with such epistemic qualifiers have rarely been compared with plain statements where such qualifiers (including certainty) are omitted. We show, in seven experimental studies (N = 2,141) that plain statements without a qualifier (“Henry made four errors”) are often held to be more credible than statements explicitly claimed to be certain (“It is certain that Henry made four errors”). We also demonstrate that certainty-statements are perceived to be less precise than just plain statements, by indicating the lower bound of a range of estimates whose upper bound is not known. So the certainty-statement about Henry’s errors is judged to be approximately correct even if he made five errors. Finally, by stating explicitly that an estimate, or a speaker, is certain, a speaker reveals that the estimate is based on judgments (even guesswork) rather than on evidence. So if you know an exact amount or quantity, and want your expertise to be acknowledged, don’t spoil it by saying that you are certain.

ABSTRACT. This paper addresses the non-linear effect that money has on cooperative behavior. In economic theory, money is assumed to have a positive effect on cooperation, by providing incentives to agents. The evidence from field experiments indicates instead that small positive incentives can be detrimental to cooperation, crowding out intrinsic motivation, while larger incentives crowd it back in. The same happens, in the opposite direction, with negative incentives. By reviewing the existing qualitative evidence from economics and psychology, the paper proposes a possible mechanism which can lead to this non-linear effect, based on the methodology and the results from economics and psychology. Money increases mutual benefits but decreases altruism, having a negative effect on cooperation when introduced in lesser amounts.

ABSTRACT. In value-based decision-making tasks, we tend to perform overt visual search for visually displayed offers during the sampling of options, followed by alternation between them, until a choice is committed. For this kind of tasks, neurons in orbitofrontal cortex (OFC) have been reported to encode offer value, bringing up questions about the dynamics of value-based computation. However, the neural basis of how gaze aids at value-based decisions is unknown. We recorded simultaneous gaze and OFC activity of two macaque monkey subjects performing a two-alternative reward gambling task. The offers were sequentially presented at opposite sides of the screen, each followed by a blank screen delay time. Interestingly, we found that the looking time of either offer was predictive of the final choice during the whole task time, including delay times. We found that cells encode expected value (EV) of the offers, predominantly during their respective presentation and at subsequent delay time. We found that fixation gates the encoding of ipsilateral EV, even when the offer is not visible. In addition, looking back to the first offer side during second delay re-activated the encoding of first offer EV, even if the subjects looked to opposite side during second offer presentation.

ABSTRACT. The complexity of everyday choices make them difficult to formally study. We address this challenge by constructing a dataset of over 100K real-life decision problems based on a combination of social media and large-scale survey data. Using large language models (LLMs) for automated coding, we are able to extract hundreds of choice attributes at play in these problems and map them onto a common representational space. This representation allows us to quantify both the content (e.g. broader themes) and the structure (e.g. specific tradeoffs) inherent in everyday choices. We also present subsets of these decision problems to human participants, and find consistency in choice patterns, allowing us to predict naturalistic choices with established decision models. Overall, our research provides new insights into the attributes and tradeoffs that underpin important life choices. In doing so, our work shows how LLM-based structure extraction can be used to study real-world cognition and behavior.

ABSTRACT. Understanding human behaviour in decision problems and strategic interactions has wide-ranging applications in psychology, economics, and artificial intelligence. Game theory offers a robust foundation for this understanding, based on the idea that individuals aim to maximize a utility function. However, the exact factors influencing strategy choices remain elusive. While traditional models try to explain human behaviour as a function of the outcomes of available actions, recent experimental research reveals that linguistic content significantly impacts decision-making, thus prompting a paradigm shift from outcome-based to language-based utility functions. This shift is more urgent than ever, given the advancement of generative AI, which has the potential to support humans in making critical decisions through language-based interactions. We propose sentiment analysis as a fundamental tool for this shift and take an initial step by analyzing 61 experimental instructions from the dictator game, an economic game capturing the balance between self-interest and the interest of others, which is at the core of many social interactions. Our meta-analysis shows that sentiment analysis can explain human behaviour beyond economic outcomes. We discuss future research directions. We hope this work sets the stage for a novel game theoretical approach that emphasizes the importance of language in human decisions.

ABSTRACT. Endowment effect (EE) is one of the main phenomena in the psychology of decision-making. In this reasoning bias, people attribute more value to an owned object than to another object with the same economic value that they do not own. Several competing theories have been proposed to explain EE. Studying the influence of depression on the EE could contribute to evaluating the merits of those theories. Under the loss aversion account, we could expect that depression increases EE, because loss aversion could be heightened with depression. Conversely, under the pragmatic account, we could expect that depression decreases EE. Pragmatic theory proposes that social norms associated with giving are implicitly activated by the endowment of the object and should usually lead to the EE. However, depression is associated with impaired pragmatic skills. Thus, the current study aimed to assess the relationship between depressive symptoms and the EE. We measured depressive symptoms with the Patient Health Questionnaire-9 and the EE with the exchange paradigm. Results showed that depressive symptoms were associated with a decrease in the EE. Those findings might not be consistent with loss aversion theory. We discuss results in the context of competing theories concerning EE.