Group judgements are often – implicitly or explicitly – influenced by their members’ individual expertise. However, given that expertise is seldom recognized fully and that some distortions may occur (bias, correlation, etc.), it is not clear that differential weighting is an epistemically advantageous strategy with respect to straight averaging. Our paper characterizes a wide set of conditions under which differential weighting outperforms straight averaging and embeds the results into the multidisciplinary group decision-making literature.

Subjective probability plays an increasingly important role in many fields concerned with human cognition and behavior. Yet there have been significant criticisms of the idea that probabilities could actually be represented in the mind. This paper presents and elaborates a view of subjective probability as a kind of sampling propensity associated with internally represented generative models. The resulting view answers to some of the most well known criticisms of subjective probability, and is also supported by empirical work in neuroscience and (...) behavioral psychology. The repercussions of the view for how we conceive of many ordinary instances of subjective probability, and how it relates to more traditional conceptions of subjective probability, are discussed in some detail. (shrink)

Research on “improper” linear models has shown that predetermined weighting schemes for the linear model, such as equally weighting all predictors, can be surprisingly accurate on cross-validation. We review recent advances that can characterize the optimal choice of an improper linear model. We extend this research to the understanding of fast and frugal heuristics, particularly to the ecologically rational goal of understanding in which task environments given heuristics are optimal. We demonstrate how to test this model using the Recognition Heuristic (...) and Take the Best heuristic, show how the model reconciles with the ecological rationality program, and discuss how our prescriptive, computational approach could be approximated by simpler mental rules that might be more descriptive. Echoing the arguments of van Rooij et al., we stress the virtue of having a computationally tractable model of strategy selection, even if one proposes that cognizers use a simpler heuristic process to approximate it. (shrink)

A cursory glance at the list of Nobel Laureates for Economics is sufficient to confirm Stanovich’s description of the project to evaluate human rationality as seminal. Herbert Simon, Reinhard Selten, John Nash, Daniel Kahneman, and others, were awarded their prizes less for their work in economics, per se, than for their work on rationality, as such. Although philosophical works have for millennia attempted to describe, explicate and evaluate individual and collective aspects of rationality, new impetus was brought to this endeavor (...) over the last century as mathematical logic along with the social and behavioral sciences emerged. Yet more recently, over the last several decades, propelled by the emergence of artificial intelligence, cognitive science, evolutionary psychology, neuropsychology, and related fields, even more sophisticated approaches to the study of rationality have emerged. (shrink)

Cognitive scientists have revealed systematic errors in human reasoning. There is disagreement about what these errors indicate about human rationality, but one upshot seems clear: human reasoning does not seem to fit traditional views of human rationality. This concern about rationality has made its way through various fields and has recently caught the attention of philosophers. The concern is that if philosophers are prone to systematic errors in reasoning, then the integrity of philosophy would be threatened. In this paper, I (...) present some of the more famous work in cognitive science that has marshaled this concern. Then I present reasons to think that those with training in philosophy will be less prone to certain systematic errors in reasoning. The suggestion is that if philosophers could be shown to be less prone to such errors, then the worries about the integrity of philosophy could be constrained. Then I present evidence that, according to performance on the CRT (Frederick 2005), those who have benefited from training and selection in philosophy are indeed less prone to one kind of systematic error: irrationally arbitrating between intuitive and reflective responses. Nonetheless, philosophers are not entirely immune to this systematic error, and their proclivity for this error is statistically related to their responses to a variety of philosophical questions. So, while the evidence herein puts constraints on the worries about the integrity of philosophy, it by no means eliminates these worries. The conclusion, then, is that the present evidence offers prima facie reasons to ascribe a mitigated privilege to philosophers' ability to rationally arbitrate between intuitive and reflective responses. (shrink)

How much irrationality should we ascribe to human cognition? Psychological evidence suggests that people’s reasoning is largely inaccurate, but according to an evolutionary argument for rationality, we have good reasons to believe that this is not so. To solve the conflict between psychological evidence and EAR, commentators have usually put the blame either on the psychological evidence, arguing that inaccurate reasoning appears only in the context of lab studies, or on the premises of EAR, charged with not being in line (...) with the concepts and findings of evolutionary biology. I argue that Hammond’s distinction between two distinct criteria of rationality, namely coherence and correspondence, might shed new light on this apparent conflict. I show that EAR might be interpreted in two different ways, and that EAR and psychological evidence might in fact be both correct if they appeal to different criteria of accurate reasoning. Moreover, evolutionary considerations have been recently used not to oppose the existence of violations of norms of coherence but rather to explain it. (shrink)

The aim of this dissertation is to create a naturalistic philosophical picture of creative capacities that are specific to our species, focusing on artistic ability, religious reflection, and scientific study. By integrating data from diverse domains within a philosophical anthropological framework, I have presented a cognitive and evolutionary approach to the question of why humans, but not other animals engage in such activities. Through an application of cognitive and evolutionary perspectives to the study of these behaviors, I have sought to (...) provide a more solid footing for philosophical anthropological discussions of uniquely human behavior. In particular, I have argued that art, religion and science, which are usually seen as achievements that are quite remote from ordinary modes of reasoning, are subserved by evolved cognitive processes that serve functions in everyday cognitive tasks, that arise early and spontaneously in cognitive development, that are shared cross-culturally, and that have evolved in response to selective pressures in our ancestral past. These mundane cognitive processes provide a measuring rod with which we can assess a diversity of cultural phenomena; they form a unified explanatory framework to approach human culture. I have argued that we can explain uncommon thoughts in terms of interactions between common minds. This dissertation is subdivided into four parts. Part I outlines the problem of human uniqueness, examining theories on how humans conceptualize the world, and what their mental tool box looks like. Part II discusses the evolutionary and cognitive origins of human artistic behavior. Part III focuses on the cognitive science of religion, especially on how it can be applied to the reasoning of theologians and philosophers of religion. Part IV considers the cognitive basis of scientific practice. (shrink)

Intractability and optimality are two sides of one coin: Optimal models are often intractable, that is, they tend to be excessively complex, or NP-hard. We explain the meaning of NP-hardness in detail and discuss how modem computer science circumvents intractability by introducing heuristics and shortcuts to optimality, often replacing optimality by means of sufficient sub-optimality. Since the principles of decision theory dictate balancing the cost of computation against gain in accuracy, statistical inference is currently being reshaped by a vigorous new (...) trend: the science of simplicity. Simple models, as we show for specific cases, are not just tractable, they also tend to be robust. Robustness is the ability of a model to extract relevant information from data, disregarding noise.Recently, Gigerenzer, Todd and the ABC Research Group (1999) have put forward a collection of fast and frugal heuristics as simple, boundedly rational inference strategies used by the unaided mind in real world inference problems. This collection of heuristics has suggestively been called the adaptive toolbox. In this paper we will focus on a comparison task in order to illustrate the simplicity and robustness of some of the heuristics in the adaptive toolbox in contrast to the intractability and the fragility of optimal solutions. We will concentrate on three important classes of models for comparison-based inference and, in each of these classes, search for that to be used as benchmarks to evaluate the performance of fast and frugal heuristics: lexicographic trees, linear modes and Bayesian networks. Lexicographic trees are interesting because they are particularly simple models that have been used by humans throughout the centuries. Linear models have been traditionally used by cognitive psychologists as models for human inference, while Bayesian networks have only recently been introduced in statistics and computer science. Yet it is the Bayesian networks that are the best possible benchmarks for evaluating the fast and frugal heuristics, as we will show in this paper. (shrink)

"It is difficult not to notice a curious unrest in the philosophic atmosphere of the time, a loosening of old landmarks, a softening of oppositions, a mutual borrowing from one another on the part of systems anciently closed, and an interest in new suggestions, however vague, as if the one thing sure were the inadequacy of extant school-solutions. The dissatisfactions with these seems due for the most part to a feeling that they are too abstract and academic. Life is confused (...) and superabundant, and what the younger generation appears to crave is more of the temperament of life in its philosophy, even though it were at some cost of logical rigor and formal purity." - William James (1904). -/- Abstract: The study of model-based reasoning (MBR) is one of the most interesting recent developments at the intersection of psychology and the philosophy of science. Although a broad and eclectic area of inquiry, one central axis by which MBR connects these disciplines is anchored at one end in theories of internal reasoning (in cognitive science), and at the other, in C.S. Peirce’s semiotics (in philosophy). In this paper, we attempt to show that Peirce’s semiotics actually has more natural affinity on the psychological side with ecological psychology, as originated by James J. Gibson and especially Egon Brunswik, than it does with non-interactionist approaches to cognitive science. In particular, we highlight the strong ties we believe to exist between the triarchic structure of semiotics as conceived by Peirce, and the similar triarchic stucture of Brunswik’s lens model of organismic achievement in irreducibly uncertain ecologies. The lens model, considered as a theory of creative abduction, provides a concrete instantiation of at least one, albeit limited, interpretation of Peirce’s semiotics, one that we believe could be quite fruitful in future theoretical and empirical investigations of MBR in both science an philosophy. (shrink)

Decision making theory in general, and mental models in particular, associate judgment and choice. Decision choice follows probability estimates and errors in choice derive mainly from errors in judgment. In the studies reported here we use the Monty Hall dilemma to illustrate that judgment and choice do not always go together, and that such a dissociation can lead to better decision-making. Specifically, we demonstrate that in certain decision problems, exceeding working memory limitations can actually improve decision choice. We show across (...) four experiments that increasing the number of choice alternatives forces people to collapse choices together, resulting in better decision-making. While choice performance improves, probability judgments do not change, thus demonstrating an important dissociation between choice and probability judgments. We propose the Collapsing Choice Theory (CCT) which explains how working memory capacity, probability estimation, choice alternatives, judgment, and regret all interact and effect decision quality. (shrink)

From an evolutionary standpoint, a default presumption is that true beliefs are adaptive and misbeliefs maladaptive. But if humans are biologically engineered to appraise the world accurately and to form true beliefs, how are we to explain the routine exceptions to this rule? How can we account for mistaken beliefs, bizarre delusions, and instances of self-deception? We explore this question in some detail. We begin by articulating a distinction between two general types of misbelief: those resulting from a breakdown in (...) the normal functioning of the belief formation system (e.g., delusions) and those arising in the normal course of that system's operations (e.g., beliefs based on incomplete or inaccurate information). The former are instances of biological dysfunction or pathology, reflecting “culpable” limitations of evolutionary design. Although the latter category includes undesirable (but tolerable) by-products of “forgivably” limited design, our quarry is a contentious subclass of this category: misbeliefs best conceived as design features. Such misbeliefs, unlike occasional lucky falsehoods, would have been systematically adaptive in the evolutionary past. Such misbeliefs, furthermore, would not be reducible to judicious – but doxastically1noncommittal – action policies. Finally, such misbeliefs would have been adaptive in themselves, constituting more than mere by-products of adaptively biased misbelief-producing systems. We explore a range of potential candidates for evolved misbelief, and conclude that, of those surveyed, onlypositive illusionsmeet our criteria. (shrink)

Naive physics beliefs can be systematically mistaken. They provide a useful test-bed because they are common, and also because their existence must rely on some adaptive advantage, within a given context. In the second part of the commentary we also ask questions about when a whole family of misbeliefs should be considered together as a single phenomenon.

Creating agents that proficiently interact with people is critical for many applications. Towards creating these agents, models are needed that effectively predict people's decisions in a variety of problems. To date, two approaches have been suggested to generally describe people's decision behavior. One approach creates a-priori predictions about people's behavior, either based on theoretical rational behavior or based on psychological models, including bounded rationality. A second type of approach focuses on creating models based exclusively on observations of people's behavior. At (...) the forefront of these types of methods are various machine learning algorithms. This paper explores how these two approaches can be compared and combined in different types of domains. In relatively simple domains, both psychological models and machine learning yield clear prediction models with nearly identical results. In more complex domains, the exact action predicted by psychological models is not even clear, and machine learning models are even less accurate. Nonetheless, we present a novel approach of creating hybrid methods that incorporate features from psychological models in conjunction with machine learning in order to create significantly improved models for predicting people's decisions. To demonstrate these claims, we present an overview of previous and new results, taken from representative domains ranging from a relatively simple optimization problem and complex domains such as negotiation and coordination without communication. (shrink)

Many believe that values are crucially dependent on emotions. This paper focuses on epistemic aspects of the putative link between emotions and value by asking two related questions. First, how exactly are emotions supposed to latch onto or track values? And second, how well suited are emotions to detecting or learning about values? To answer the first question, the paper develops the heuristics-model of emotions. This approach models emotions as sui generis heuristics of value. The empirical plausibility of the heuristics-model (...) is demonstrated using evidence from experimental psychology, evolutionary anthropology and neuroscience. The model is used then to answer the second question. If emotions are indeed heuristics of value, then it follows that emotions can be an important and useful source of information about value. However, emotions will not be epistemically superior in the sense of being the highest court of appeal for the justification of axiological beliefs (the latter view is referred to as the Epistemic Dependence Thesis, or EDT for short). The paper applies the heuristics-model to celebrated cases from the philosophy of emotions literature arguing that while the heuristics-model offers a good explanation of typical patterns of emotional reactions in such cases, advocates of EDT will have a hard time accounting for these patterns. The paper also shows that the conclusions drawn from special cases generalize. The paper ends by arguing that skepticism about the metaethical significance of emotions is compatible with a commitment to the importance of emotions in first-order normative ethics. (shrink)

Many cognitive scientists, having discovered that some computational-level characterization f of a cognitive capacity φ is intractable, invoke heuristics as algorithmic-level explanations of how cognizers compute f. We argue that such explanations are actually dysfunctional, and rebut five possible objections. We then propose computational-level theory revision as a principled and workable alternative.

This article introduces the topic ‘‘Production of Referring Expressions: Bridging the Gap between Computational and Empirical Approaches to Reference’’ of the journal Topics in Cognitive Science. We argue that computational and psycholinguistic approaches to reference production can benefit from closer interaction, and that this is likely to result in the construction of algorithms that differ markedly from the ones currently known in the computational literature. We focus particularly on determinism, the feature of existing algorithms that is perhaps most clearly at (...) odds with psycholinguistic results, discussing how future algorithms might include non-determinism, and how new psycholinguistic experiments could inform the development of such algorithms. (shrink)

Research on moral judgment has been dominated by rationalist models, in which moral judgment is thought to be caused by moral reasoning. The author gives 4 reasons for considering the hypothesis that moral reasoning does not cause moral judgment; rather, moral reasoning is usually a post hoc construction, generated after a judgment has been reached. The social intuitionist model is presented as an alternative to rationalist models. The model is a social model in that it deemphasizes the private reasoning done (...) by individuals and emphasizes instead the importance of social and cultural influences. The model is an intuitionist model in that it states that moral judgment is generally the result of quick, automatic evaluations (intuitions). The model is more consistent than rationalist models with recent findings in social, cultural, evolutionary, and biological psychology, as well as in anthropology and primatology. (shrink)

Inductive generalization, where people go beyond the data provided, is a basic cognitive capability, and it underpins theoretical accounts of learning, categorization, and decision making. To complete the inductive leap needed for generalization, people must make a key ‘‘sampling’’ assumption about how the available data were generated. Previous models have considered two extreme possibilities, known as strong and weak sampling. In strong sampling, data are assumed to have been deliberately generated as positive examples of a concept, whereas in weak sampling, (...) data are assumed to have been generated without any restrictions. We develop a more general account of sampling that allows for an intermediate mixture of these two extremes, and we test its usefulness. In two experiments, we show that most people complete simple one-dimensional generalization tasks in a way that is consistent with their believing in some mixture of strong and weak sampling, but that there are large individual differences in the relative emphasis different people give to each type of sampling. We also show experimentally that the relative emphasis of the mixture is influenced by the structure of the available information. We discuss the psychological meaning of mixing strong and weak sampling, and possible extensions of our modeling approach to richer problems of inductive generalization. (shrink)

What are the causes of the evolution of complex cognition? Discussions of the evolution of cognition sometimes seem to assume that more complex cognition is a fundamental advance over less complex cognition, as evidenced by a broad trend toward larger brains in evolutionary history. Evolutionary biologists are suspicious of such explanations since they picture natural selection as a process leading to adaptation to local environments, not to progressive trends. Cognitive adaptations will have costs, and more complex cognition will evolve only (...) when its local utility outweighs them. (shrink)

Our programmatic article on Homo heuristicus (Gigerenzer & Brighton, 2009) included a methodological section specifying three minimum criteria for testing heuristics: competitive tests, individual-level tests, and tests of adaptive selection of heuristics. Using Richter and Späth’s (2006) study on the recognition heuristic, we illustrated how violations of these criteria can lead to unsupported conclusions. In their comment, Hilbig and Richter conduct a reanalysis, but again without competitive testing. They neither test nor specify the compensatory model of inference they argue for. (...) Instead, they test whether participants use the recognition heuristic in an unrealistic 100% (or 96%) of cases, report that only some people exhibit this level of consistency, and conclude that most people would follow a compensatory strategy. We know of no model of judgment that predicts 96% correctly. The curious methodological practice of adopting an unrealistic measure of success to argue against a competing model, and to interpret such a finding as a triumph for a preferred but unspecified model, can only hinder progress. Marewski, Gaissmaier, Schooler, Goldstein, and Gigerenzer (2010), in contrast, specified five compensatory models, compared them with the recognition heuristic, and found that the recognition heuristic predicted inferences most accurately. (shrink)

We argue that word meanings are not stored in a mental lexicon but are generated in the context of working memory from long-term memory traces that record our experience with words. Current statistical models of semantics, such as latent semantic analysis and the Topic model, describe what is stored in long-term memory. The CI-2 model describes how this information is used to construct sentence meanings. This model is a dual-memory model, in that it distinguishes between a gist level and an (...) explicit level. It also incorporates syntactic information about how words are used, derived from dependency grammar. The construction of meaning is conceptualized as feature sampling from the explicit memory traces, with the constraint that the sampling must be contextually relevant both semantically and syntactically. Semantic relevance is achieved by sampling topically relevant features; local syntactic constraints as expressed by dependency relations ensure syntactic relevance. (shrink)

How can we help people develop judgement and decision skills? One approach is to teach formal methods such as decision analyses, but these are difficult to apply in ill-structured settings, and the methods are unworkable when one is under time pressure and uncertain conditions. If we regard these skills as types of expertise that can be developed, then in a given domain we may attempt to define the cues, patterns, and strategies used by experts, and develop a programme to teach (...) people how to think like experts. However, in many settings this can be impractical. A different approach to decision skills training is to teach people how to learn like experts. By identifying a set of strategies used by experts to develop their proficiency at decision making, we can develop a programme aimed at helping people become reflective practitioners. This article describes such a training programme. (shrink)

In this article we describe research methods that are used for the study of individual multiattribute evaluation processes. First we explain that a multiattribute evaluation problem involves the evaluation of a set of alternatives, described by their values on a number of alternatives. We discuss a number of evaluation strategies that may be applied to arrive at a conclusion about the attractiveness or suitability of the alternatives, and next introduce two main research paradigms in this area, structural modelling and process (...) tracing. We argue that the techniques developed within these paradigms all have their advantages and disadvantages, and conclude that the most promising technique to detect the true nature of the evaluation strategy used by a judge seems to be the analysis of verbal protocols. At the same time we think it is wise not to rely on just one technique, but to use a multimethod approach to the study of multiattribute evaluation processes whenever that is possible. (shrink)

In this response, we clarify several misunderstandings of the understanding/acceptance principle and defend our specific operationalization of that principle. We reiterate the importance of addressing the problem of rational task construal and we elaborate the notion of computational limitations contained in our target article. Our concept of thinking dispositions as variable intentional-level styles of epistemic and behavioral regulation is explained, as is its relation to the rationality debate. Many of the suggestions of the commentators for elaborating two-process models are easily (...) integrated into our generic dual-process account. We further explicate how we view the relation between System 1 and System 2 and evolutionary and normative rationality. We clarify our attempt to fuse the contributions of the cognitive ecologists with the insights of the original heuristics and biases researchers. (shrink)

The understanding of decision-making systems has come together in recent years to form a unified theory of decision-making in the mammalian brain as arising from multiple, interacting systems (a planning system, a habit system, and a situation-recognition system). This unified decision-making system has multiple potential access points through which it can be driven to make maladaptive choices, particularly choices that entail seeking of certain drugs or behaviors. We identify 10 key vulnerabilities in the system: (1) moving away from homeostasis, (2) (...) changing allostatic set points, (3) euphorigenic signals, (4) overvaluation in the planning system, (5) incorrect search of situation-action-outcome relationships, (6) misclassification of situations, (7) overvaluation in the habit system, (8) a mismatch in the balance of the two decision systems, (9) over-fast discounting processes, and (10) changed learning rates. These vulnerabilities provide a taxonomy of potential problems with decision-making systems. Although each vulnerability can drive an agent to return to the addictive choice, each vulnerability also implies a characteristic symptomology. Different drugs, different behaviors, and different individuals are likely to access different vulnerabilities. This has implications for an individual's susceptibility to addiction and the transition to addiction, for the potential for relapse, and for the potential for treatment. (shrink)

This paper reviews the largely psychological literature surrounding apparent failures of human rationality (sometimes referred to as 'the Rationality Wars') and locates it with respect to concepts and issues within more traditional epistemological inquiry. The goal is to bridge the gap between these two large and typically disconnected literatures – concerning rationality and the psychology of human reasoning, on the one hand, and epistemological theories of justified or rational belief, on the other – and to do so in such as (...) way as to expose interesting points of contact and convergence between them. (shrink)

Selection explanations explain some non-accidental generalizations in virtue of a selection process. Such explanations are not particulaizable - they do not transfer as explanations of the instances of such generalizations. This is unlike many explanations in the physical sciences, where the explanation of the general fact also provides an explanation of its instances (i.e. standard D-N explanations). Are selection explanations (e.g. in biology) therefore a different kind of explanation? I argue that to understand this issue, we need to see that (...) a standard D-N explanation of some non-accidental generalization (al Fs are Gs) may also ipso facto explain its contrapositive (all non-Gs are non-Fs), but the explanation is particularizable with respect to the former but not to the latter. This can be seen by noting that the Raven Paradox counterexample to the H-D model of confirmation also generates a counterexample to the D-N model of explanation (all ravens are black does not explain why the non-black shoe is a non-raven). In such cases it is natural to take the generalization with the positive predicates to have a particularizable explanation. However, this need not be the case, and in selection explanations it is the generalization with the positive predicates whose explanation is no particularizable. Thus there is no need to suppose that selection explanations are fundamentally different. (shrink)

In the debate between simple inference heuristics and complex decision mechanisms, we take a position squarely in the middle. A decision making process that extends to both naturalistic and novel settings should extend beyond the confines of this debate; both simple heuristics and complex mechanisms are cognitive skills adapted to and appropriate for some circumstances but not for others. Rather than ask `Which skill is better?'' it is often more important to ask `When is a skill justified?'' The selection and (...) application of an appropriate cognitive skill for a particular problem has both costs and benefits, and therefore requires the resolution of a tradeoff. In revisiting satisficing, we observe that the essence of satisficing is tradeoff. Unlike heuristics, which derive their justification from empirical phenomena, and unlike optimal solutions, which derive their justification by an evaluation of alternatives, satisficing decision-making derives its justification by an evaluation of consequences. We formulate and present a satisficing decision paradigm that has its motivation in Herbert Simon''s work on bounded rationality. We characterize satisficing using a cost–benefit tradeoff, and generate a decision rule applicable to both designing intelligent machines as well as describing human behavior. (shrink)

Evolutionary theories of morality, beginning with Darwin, have focused on explanations for altruism. More generally, these accounts have concentrated on conscience to the neglect of condemnation. As a result, few theoretical tools are available for understanding the rapidly accumulating data surrounding third-party judgment and punishment. Here we consider the strategic interactions among actors, victims, and third-parties to help illuminate condemnation. We argue that basic differences between the adaptive problems faced by actors and third-parties indicate that actor conscience and third-party condemnation (...) are likely performed by different cognitive mechanisms. Further, we argue that current theories of conscience do not easily explain its experimentally demonstrated insensitivity to consequences. However, these results might be explicable if conscience functions, in part, as a defense system for avoiding third-party punishment. If conscience serves defensive functions, then its computational structure should be closely tailored to the details of condemnation mechanisms. This possibility underscores the need for a better understanding of condemnation, which is important not only in itself but also for explaining the nature of conscience. We outline three evolutionary mysteries of condemnation that require further attention: third-party judgment, moralistic punishment, and moral impartiality. (shrink)

Making decisions can be hard, but it can also be facilitated. Simple heuristics are fast and frugal but nevertheless fairly accurate decision rules that people can use to compensate for their limitations in computational capacity, time, and knowledge when they make decisions [Gigerenzer, G., Todd, P. M., & the ABC Research Group (1999). Simple Heuristics That Make Us Smart. New York: Oxford University Press.]. These heuristics are effective to the extent that they can exploit the structure of information in the (...) environment in which they operate. Specifically, they require knowledge about the predictive value of probabilistic cues. However, it is often difficult to keep track of all the available cues in the environment and how they relate to any relevant criterion. This problem becomes even more critical if compound cues are considered. We submit that knowledge about the causal structure of the environment helps decision makers focus on a manageable subset of cues, thus effectively reducing the potential computational complexity inherent in even relatively simple decision-making tasks. We review experimental evidence that tested this hypothesis and report the results of a simulation study. We conclude that causal knowledge can act as a meta-cue for identifying highly valid cues, either individual or compound, and helps in the estimation of their validities. (shrink)

I argue that bounded agents face a systematic accuracy–coherence trade-off in cognition. Agents must choose whether to structure their cognition in ways likely to promote coherence or accuracy. I illustrate the accuracy–coherence trade-off by showing how it arises out of at least two component trade-offs: a coherence–complexity trade-off between coherence and cognitive complexity, and a coherence–variety trade-off between coherence and strategic variety. These trade-offs give rise to an accuracy–coherence trade-off because privileging coherence over complexity or strategic variety often leads to (...) a corresponding reduction in accuracy. I conclude with a discussion of two normative consequences for the study of bounded rationality: the importance of procedural rationality and the role of coherence in theories of bounded rationality. (shrink)

Algorithm Decision Making (ADM) systems designed to augment or automate human decision-making have the potential to produce better decisions while also freeing up human time and attention for other pursuits. For this potential to be realised, however, algorithmic decisions must be sufficiently aligned with human goals and interests. We take a Principal-Agent (P-A) approach to the questions of ADM alignment and trust. In a broad sense, ADM is beneficial if and only if human principals can trust algorithmic agents to act (...) faithfully on their behalf. This mirrors the challenge of facilitating P-A relationships among humans, but the peculiar nature of human-machine interaction also raises unique issues. The problem of asymmetric information is omnipresent but takes a different form in the context of ADM. Although the decision-making machinery of an algorithmic agent can in principle be laid bare for all to see, the sheer complexity of ADM systems based on deep learning models prevents straightforward monitoring. We draw on literature from economics and political science to argue that the problem of trust in ADM systems should be addressed at the level of institutions. Although the dyadic relationship between human principals and algorithmic agents is our ultimate concern, cooperation at this level must rest against an institutional environment which allows humans to effectively evaluate and choose among algorithmic alternatives. (shrink)

This paper develops a novel approach to the question of the normativity of logic, which we reinterpret as a clash between two intuitions: the direct normativity intuition and the unfeasibility intuition. The standard response has been to dismiss the direct normativity intuition, bridging logic and reasoning via principles that relativize the normative import of logic to pragmatic and feasibility considerations. We argue that the standard response is misguided. Building upon theories of bounded rationality, our approach conceptualizes reasoning as constrained by (...) multiple, independent normative factors, logical and non-logical ones. These different factors can conflict with one another, to the effect that logically sound inferences might not coincide with what is feasible for an agent to infer. From this perspective, we will argue that logic gives us only prima facie, i.e. contributive and defeasible, obligations on what to believe, but that such obligations do not always coincide with what an agent ought to believe all things considered. These distinctions will dissolve the alleged clash between the direct normativity and the unfeasibility intuition. (shrink)

Generative AI enables automated, effective manipulation at scale. Despite the growing general ethical discussion around generative AI, the specific manipulation risks remain inadequately investigated. This article outlines essential inquiries encompassing conceptual, empirical, and design dimensions of manipulation, pivotal for comprehending and curbing manipulation risks. By highlighting these questions, the article underscores the necessity of an appropriate conceptualisation of manipulation to ensure the responsible development of Generative AI technologies.

In this work, I investigate the implications of reversing the common assumption of rationality on behalf of human agents typically underlying philosophical research. Instead, I assume that human agents can become rational only if they learn to edge against their dumbness. Specifically, I show that intelligence cannot be considered the opposite of dumbness. To this end, I embrace the difference among System 1, System 2, and System 1.5. On these grounds, I argue that System 2 can be considered the system (...) of intelligence, whereas Systems 1 and 1.5 can be viewed as the systems of rationality. More precisely, I argue that System 2 concerns theoretical reasoning, whereas System 1 and 1.5 concern practical reasoning. Accordingly, I show that people are dumber whenever they fail to have rational control over System 1 via System 1.5, whereas people become more rational whenever they manage to have rational control over System 1 via System 1.5. Interestingly, both cases can be considered the causes of two distinct social processes. On one side, when dumbness increases among individuals, social groups achieve coordination that detriments general welfare. On the other, when dumbness decreases among individuals, social groups achieve coordination that maximizes general welfare. (shrink)

In this letter, we argue that an economic perspective on the mind has played—and should continue to play—a central role in the development of cognitive science. Viewing cognition as the productive application of mental resources puts cognitive science and economics on a common conceptual footing, paving the way for closer collaboration between the two disciplines. This will enable cognitive scientists to more readily repurpose economic concepts and analytical tools for the study of mental phenomena, while at the same time, enriching (...) our understanding of the modern economy, which is increasingly driven by mental, rather than physical, production. (shrink)

There is now a significant body of literature on consequentialist ethics that propose satisficing instead of maximizing accounts. Even though epistemology recently witnessed a widespread discussion of teleological and consequentialist theories, a satisficing account is surprisingly not developed yet. The aim of this paper is to do just that. The rough idea is that epistemic rules are justified if and only if they satisfice the epistemic good, i.e., reach some threshold of epistemic value (which varies with practical context), and believing (...) is justified if and only if it follows said rules.I argue that this alternative to the implicitly established way of thinking in maximizing terms has significant advantages. First, maximizing epistemic value can be unreasonably demanding; second, a satisficing theory can make finding reasonable rules for belief formation and sustenance much more accessible; and third, a satisficing approach is a better alternative to both general subjectivist and maximizing objectivist attempts to spell out epistemic blame. (shrink)

Common-sense (or folk) psychology holds that (generally) we do what we do for the reasons we have. This common-sense approach is embodied in claims like ‘I went to the kitchen because I wanted a drink’ and ‘She took a coat because she thought it might rain and hoped to stay dry’. However, the veracity of these common-sense psychological explanations has been challenged by experimental evidence (primarily from behavioural economics and social psychology) which appears to show that individuals are systematically irrational—that (...) often we do not do what we do because of the reasons we have. Recently, some of the same experimental evidence has also been used to level a somewhat different challenge at the common-sense view, arguing that the overarching aim of reasoning is not to deliver better or more reason-governed decisions for individual reasoners, but to improve group decision making or to protect an individual’s sense of self. This paper explores the range of challenges that experimental work has been taken to raise for the common-sense approach and suggests some potential responses. Overall, I argue that the experimental evidence surveyed should not lead us to a rejection of individual rationality. (shrink)

This paper argues that emotions can play an epistemic role as justifiers of evaluative beliefs. It also presents the heuristics theory of emotion as an empirically informed explanation of how emotions can play such a role and why they in practice usefully complement non-affective evaluative judgments. As such, the heuristics theory represents a form of moderate rationalism: it acknowledges that emotions can be epistemically valuable, even privileged in some sense, but denies that they would be uniquely privileged. I argue that (...) judgments and emotional responses pick out different but correlated kinds of evaluative properties and therefore emotional responses and non-affective evaluative judgments play mutually complementary rather than mutually exclusive roles. It follows that emotional responses can be legitimately drawn upon to support evaluative beliefs, but they lack supreme epistemic authority. (shrink)

A widespread need to explain the behavior and outcomes of AI-based systems has emerged, due to their ubiquitous presence. Thus, providing renewed momentum to the relatively new research area of eXplainable AI (XAI). Nowadays, the importance of XAI lies in the fact that the increasing control transference to this kind of system for decision making -or, at least, its use for assisting executive stakeholders- already affects many sensitive realms (as in Politics, Social Sciences, or Law). The decision-making power handover to (...) opaque AI systems makes mandatory explaining those, primarily in application scenarios where the stakeholders are unaware of both the high technology applied and the basic principles governing the technological solutions. The issue should not be reduced to a merely technical problem; the explainer would be compelled to transmit richer knowledge about the system (including its role within the informational ecosystem where he/she works). To achieve such an aim, the explainer could exploit, if necessary, practices from other scientific and humanistic areas. The first aim of the paper is to emphasize and justify the need for a multidisciplinary approach that is beneficiated from part of the scientific and philosophical corpus on Explaining, underscoring the particular nuances of the issue within the field of Data Science. The second objective is to develop some arguments justifying the authors’ bet by a more relevant role of ideas inspired by, on the one hand, formal techniques from Knowledge Representation and Reasoning, and on the other hand, the modeling of human reasoning when facing the explanation. This way, explaining modeling practices would seek a sound balance between the pure technical justification and the explainer-explainee agreement. (shrink)

Deliberative analysis enables us to weigh features, simulate futures, and arrive at good, tractable decisions. So why do we so often eschew deliberation, and instead rely on more intuitive, gut responses? We propose that intuition might be prescribed for some decisions because people’s folk theory of decision-making accords a special role to authenticity, which is associated with intuitive choice. Five pre-registered experiments find evidence in favor of this claim. In Experiment 1 (N = 654), we show that participants prescribe intuition (...) and deliberation as a basis for decisions differentially across domains, and that these prescriptions predict reported choice. In Experiment 2 (N = 555), we find that choosing intuitively vs. deliberately leads to different inferences concerning the decision-maker’s commitment and authenticity—with only inferences about the decision-maker’s authenticity showing variation across domains that matches that observed for the prescription of intuition in Experiment 1. In Experiment 3 (N = 631), we replicate our prior results and rule out plausible confounds. Finally, in Experiment 4 (N = 177) and Experiment 5 (N = 526), we find that an experimental manipulation of the importance of authenticity affects the prescribed role for intuition as well as the endorsement of expert human or algorithmic advice. These effects hold beyond previously recognized influences on intuitive vs. deliberative choice, such as computational costs, presumed reliability, objectivity, complexity, and expertise. (shrink)

Topics in Cognitive Science, Volume 14, Issue 3, Page 467-491, July 2022.

YouTube has been implicated in the transformation of users into extremists and conspiracy theorists. The alleged mechanism for this radicalizing process is YouTube’s recommender system, which is optimized to amplify and promote clips that users are likely to watch through to the end. YouTube optimizes for watch-through for economic reasons: people who watch a video through to the end are likely to then watch the next recommended video as well, which means that more advertisements can be served to them. This (...) is a seemingly innocuous design choice, but it has a troubling side-effect. Critics of YouTube have alleged that the recommender system tends to recommend extremist content and conspiracy theories, as such videos are especially likely to capture and keep users’ attention. To date, the problem of radicalization via the YouTube recommender system has been a matter of speculation. The current study represents the first systematic, pre-registered attempt to establish whether and to what extent the recommender system tends to promote such content. We begin by contextualizing our study in the framework of technological seduction. Next, we explain our methodology. After that, we present our results, which are consistent with the radicalization hypothesis. Finally, we discuss our findings, as well as directions for future research and recommendations for users, industry, and policy-makers. (shrink)

De Neys (2021) argues that the debate between single- and dual-process theorists of thought has become both empirically intractable and scientifically inconsequential. I argue that this is true only under the traditional framing of the debate—when single- and dual-process theories are understood as claims about whether thought processes share the same defining properties (e.g., making mathematical judgments) or have two different defining properties (e.g., making mathematical judgments autonomously versus via access to a central working memory capacity), respectively. But if single- (...) and dual-process theories are understood in cognitive modeling terms as claims about whether thought processes function to implement one or two broad types of algorithms, respectively, then the debate becomes scientifically consequential and, presumably, empirically tractable. So, I argue, the correct response to the current state of the debate is not to abandon it, as De Neys suggests, but to reframe it as a debate about cognitive models. (shrink)

Many fields (social choice, welfare economics, recommender systems) assume people express what benefits them via their 'revealed preferences'. Revealed preferences have well-documented problems when used this way, but are hard to displace in these fields because, as an information source, they are simple, universally applicable, robust, and high-resolution. In order to compete, other information sources (about participants' values, capabilities and functionings, etc) would need to match this. I present a conception of values as *attention policies resulting from constitutive judgements*, and (...) use it to build an alternative preference relation, Meaningful Choice, which retains many desirable features of revealed preference. (shrink)