Each of our theories of mental representation provides some insight into how the mind works. However, these insights often seem incompatible, as the debates between symbolic, dynamical, emergentist, sub-symbolic, and grounded approaches to cognition attest. Mental representations—whatever they are—must share many features with each of our theories of representation, and yet there are few hypotheses about how a synthesis could be possible. Here, I develop a theory of the underpinnings of symbolic cognition that shows how sub-symbolic dynamics may give rise (...) to higher-level cognitive representations of structures, systems of knowledge, and algorithmic processes. This theory implements a version of conceptual role semantics by positing an internal universal representation language in which learners may create mental models to capture dynamics they observe in the world. The theory formalizes one account of how truly novel conceptual content may arise, allowing us to explain how even elementary logical and computational operations may be learned from a more primitive basis. I provide an implementation that learns to represent a variety of structures, including logic, number, kinship trees, regular languages, context-free languages, domains of theories like magnetism, dominance hierarchies, list structures, quantification, and computational primitives like repetition, reversal, and recursion. This account is based on simple discrete dynamical processes that could be implemented in a variety of different physical or biological systems. In particular, I describe how the required dynamics can be directly implemented in a connectionist framework. The resulting theory provides an “assembly language” for cognition, where high-level theories of symbolic computation can be implemented in simple dynamics that themselves could be encoded in biologically plausible systems. (shrink)

People draw subtle distinctions in the normative domain. But it remains unclear exactly what gives rise to such distinctions. On one prominent approach, emotion systems trigger non-utilitarian judgments. The main alternative, inspired by Chomskyan linguistics, suggests that moral distinctions derive from an innate moral grammar. In this article, we draw on Bayesian learning theory to develop a rational learning account. We argue that the ‘size principle’, which is implicated in word learning, can also explain how children would use scant and (...) equivocal evidence to interpret candidate rules as applying more narrowly than utilitarian rules. (shrink)

ABSTRACTProbability judgements entail a conjunction fallacy if a conjunction is estimated to be more probable than one of its conjuncts. In the context of predication of alternative logical hypothesis, Bayesian logic provides a formalisation of pattern probabilities that renders a class of pattern-based CFs rational. BL predicts a complete system of other logical inclusion fallacies. A first test of this prediction is investigated here, using transparent tasks with clear set inclusions, varying in observed frequencies only. Experiment 1 uses data where (...) BL makes dominant predictions; Experiment 2's predictions were less clear, and we additionally investigated judgements about the second-most probable hypotheses. The results corroborated a pattern-probability account and cannot be easily explained by other theories of CFs. IFs were not limited to conjunctions, but rather occurred systematically for several logical connectives. Thus, pattern-based prob... (shrink)

Despite their success in describing and predicting cognitive behavior, the plausibility of so-called ‘rational explanations’ is often contested on the grounds of computational intractability. Several cognitive scientists have argued that such intractability is an orthogonal pseudoproblem, however, since rational explanations account for the ‘why’ of cognition but are agnostic about the ‘how’. Their central premise is that humans do not actually perform the rational calculations posited by their models, but only act as if they do. Whether or not the problem (...) of intractability is solved by recourse to ‘as if’ explanations critically depends, inter alia, on the semantics of the ‘as if’ connective. We examine the five most sensible explications in the literature, and conclude that none of them circumvents the problem. As a result, rational ‘as if’ explanations must obey the minimal computational constraint of tractability. (shrink)

In many learning or inference tasks human behavior approximates that of a Bayesian ideal observer, suggesting that, at some level, cognition can be described as Bayesian inference. However, a number of findings have highlighted an intriguing mismatch between human behavior and standard assumptions about optimality: People often appear to make decisions based on just one or a few samples from the appropriate posterior probability distribution, rather than using the full distribution. Although sampling-based approximations are a common way to implement Bayesian (...) inference, the very limited numbers of samples often used by humans seem insufficient to approximate the required probability distributions very accurately. Here, we consider this discrepancy in the broader framework of statistical decision theory, and ask: If people are making decisions based on samples—but as samples are costly—how many samples should people use to optimize their total expected or worst-case reward over a large number of decisions? We find that under reasonable assumptions about the time costs of sampling, making many quick but locally suboptimal decisions based on very few samples may be the globally optimal strategy over long periods. These results help to reconcile a large body of work showing sampling-based or probability matching behavior with the hypothesis that human cognition can be understood in Bayesian terms, and they suggest promising future directions for studies of resource-constrained cognition. (shrink)

How are natural symbol systems best understood? Traditional “symbolic” approaches seek to understand cognition by analogy to highly structured, prescriptive computer programs. Here, we describe some problems the traditional computational metaphor inevitably leads to, and a very different approach to computation (Ramscar, Yarlett, Dye, Denny, & Thorpe, 2010; Turing, 1950) that allows these problems to be avoided. The way we conceive of natural symbol systems depends to a large degree on the computational metaphors we use to understand them, and machine (...) learning suggests an understanding of symbolic thought that is very different to traditional views (Hummel, 2010). The empirical question then is: Which metaphor is best? (shrink)

Research in education and cognitive development suggests that explaining plays a key role in learning and generalization: When learners provide explanations—even to themselves—they learn more effectively and generalize more readily to novel situations. This paper proposes and tests a subsumptive constraints account of this effect. Motivated by philosophical theories of explanation, this account predicts that explaining guides learners to interpret what they are learning in terms of unifying patterns or regularities, which promotes the discovery of broad generalizations. Three experiments provide (...) evidence for the subsumptive constraints account: prompting participants to explain while learning artificial categories promotes the induction of a broad generalization underlying category membership, relative to describing items (Exp. 1), thinking aloud (Exp. 2), or free study (Exp. 3). Although explaining facilitates discovery, Experiment 1 finds that description is more beneficial for learning item details. Experiment 2 additionally suggests that explaining anomalous observations may play a special role in belief revision. The findings provide insight into explanation’s role in discovery and generalization. (shrink)

Mental representations remain the central posits of psychology after many decades of scrutiny. However, there is no consensus about the representational format(s) of biological cognition. This paper provides a survey of evidence from computational cognitive psychology, perceptual psychology, developmental psychology, comparative psychology, and social psychology, and concludes that one type of format that routinely crops up is the language-of-thought (LoT). We outline six core properties of LoTs: (i) discrete constituents; (ii) role-filler independence; (iii) predicate–argument structure; (iv) logical operators; (v) inferential (...) promiscuity; and (vi) abstract content. These properties cluster together throughout cognitive science. Bayesian computational modeling, compositional features of object perception, complex infant and animal reasoning, and automatic, intuitive cognition in adults all implicate LoT-like structures. Instead of regarding LoT as a relic of the previous century, researchers in cognitive science and philosophy-of-mind must take seriously the explanatory breadth of LoT-based architectures. We grant that the mind may harbor many formats and architectures, including iconic and associative structures as well as deep-neural-network-like architectures. However, as computational/representational approaches to the mind continue to advance, classical compositional symbolic structures – that is, LoTs – only prove more flexible and well-supported over time. (shrink)

This paper presents the Rational Action, Noisy Choice for Habituation (RANCH) model. The model was evaluated with adult looking time collected from a paradigm analogous to the infant habituation paradigm. And the model captured key patterns of looking time documented in developmental research: habituation and dishabituation.

Psychology has benefited from an enormous wealth of knowledge about processes of cognition in relation to how the brain organizes information. Within the categorization literature, this behavior is often explained through theories of memory construction called exemplar theory and prototype theory which are typically based on similarity or rule functions as explanations of how categories emerge. Although these theories work well at modeling highly controlled stimuli in laboratory settings, they often perform less well outside of these settings, such as explaining (...) the emergence of background knowledge processes. In order to explain background knowledge, we present a non-similarity-based post-Skinnerian theory of human language called Relational Frame Theory (RFT) which is rooted in a philosophical world view called functional contextualism (FC). This theory offers a very different interpretation of how categories emerge through the functions of behavior and through contextual cues, which may be of some benefit to existing categorization theories. Specifically, RFT may be able to offer a novel explanation of how background knowledge arises, and we provide some mathematical considerations in order to identify a formal model. Finally, we discuss much of this work within the broader context of general semantic knowledge and artificial intelligence research. (shrink)

A semantic universal, which we here dub the Veridical Uniformity Universal, has recently been argued to hold of responsive verbs (those that take both declarative and interrogative complements). This paper offers a preliminary explanation of this universal: verbs satisfying it are easier to learn than those that do not. This claim is supported by a computational experiment using artificial neural networks, mirroring a recent proposal for explaining semantic universals of quantifiers. This preliminary study opens up many avenues for future work (...) on explaining semantic universals more generally, which are discussed in the conclusion. (shrink)

The prominence of Bayesian modeling of cognition has increased recently largely because of mathematical advances in specifying and deriving predictions from complex probabilistic models. Much of this research aims to demonstrate that cognitive behavior can be explained from rational principles alone, without recourse to psychological or neurological processes and representations. We note commonalities between this rational approach and other movements in psychology – namely, Behaviorism and evolutionary psychology – that set aside mechanistic explanations or make use of optimality assumptions. Through (...) these comparisons, we identify a number of challenges that limit the rational program's potential contribution to psychological theory. Specifically, rational Bayesian models are significantly unconstrained, both because they are uninformed by a wide range of process-level data and because their assumptions about the environment are generally not grounded in empirical measurement. The psychological implications of most Bayesian models are also unclear. Bayesian inference itself is conceptually trivial, but strong assumptions are often embedded in the hypothesis sets and the approximation algorithms used to derive model predictions, without a clear delineation between psychological commitments and implementational details. Comparing multiple Bayesian models of the same task is rare, as is the realization that many Bayesian models recapitulate existing (mechanistic level) theories. Despite the expressive power of current Bayesian models, we argue they must be developed in conjunction with mechanistic considerations to offer substantive explanations of cognition. We lay out several means for such an integration, which take into account the representations on which Bayesian inference operates, as well as the algorithms and heuristics that carry it out. We argue this unification will better facilitate lasting contributions to psychological theory, avoiding the pitfalls that have plagued previous theoretical movements. (shrink)

More than 50 years ago, Bongard introduced 100 visual concept learning problems as a challenge for artificial vision systems. These problems are now known as Bongard problems. Although they are well known in cognitive science and artificial intelligence, only very little progress has been made toward building systems that can solve a substantial subset of them. In the system presented here, visual features are extracted through image processing and then translated into a symbolic visual vocabulary. We introduce a formal language (...) that allows representing compositional visual concepts based on this vocabulary. Using this language and Bayesian inference, concepts can be induced from the examples that are provided in each problem. We find a reasonable agreement between the concepts with high posterior probability and the solutions formulated by Bongard himself for a subset of 35 problems. While this approach is far from solving Bongard problems like humans, it does considerably better than previous approaches. We discuss the issues we encountered while developing this system and their continuing relevance for understanding visual cognition. For instance, contrary to other concept learning problems, the examples are not random in Bongard problems; instead they are carefully chosen to ensure that the concept can be induced, and we found it helpful to take the resulting pragmatic constraints into account. (shrink)

Quilty-Dunn et al. defended the reemergence of language-of-thought hypothesis (LoTH). My commentary builds up implications for the study of the development of our logical capacities. Empirical support for logically augmented LoT systems calls for the investigation of their logical primitives and developmental origin. Furthermore, Quilty-Dunn et al.'s characterization of LoT helps the quest for the foundation of logic by dissociating logical cognition from natural language.

Where does a probabilistic language-of-thought (PLoT) come from? How can we learn new concepts based on probabilistic inferences operating on a PLoT? Here, I explore these questions, sketching a traditional circularity objection to LoT and canvassing various approaches to addressing it. I conclude that PLoT-based cognitive architectures can support genuine concept learning; but, currently, it is unclear that they enjoy more explanatory breadth in relation to concept learning than alternative architectures that do not posit any LoT.

Bayesian approaches to human cognition have been extensively advocated in the last decades, but sharp objections have been raised too within cognitive science. In this paper, we outline a diagnosis of what has gone wrong with the prevalent strand of Bayesian cognitive science (here labelled pure Bayesian cognitive science), relying on selected illustrations from the psychology of reasoning and tools from the philosophy of science. Bayesians’ reliance on so-called method of rational analysis is a key point of our discussion. We (...) tentatively conclude on a constructive note, though: an appropriately modified variant of Bayesian cognitive science can still be coherently pursued, as some scholars have noted. (shrink)

Cognitive Science, Volume 46, Issue 4, April 2022.

Algorithms for approximate Bayesian inference, such as those based on sampling (i.e., Monte Carlo methods), provide a natural source of models of how people may deal with uncertainty with limited cognitive resources. Here, we consider the idea that individual differences in working memory capacity (WMC) may be usefully modeled in terms of the number of samples, or “particles,” available to perform inference. To test this idea, we focus on two recent experiments that report positive associations between WMC and two distinct (...) aspects of categorization performance: the ability to learn novel categories, and the ability to switch between different categorization strategies (“knowledge restructuring”). In favor of the idea of modeling WMC as a number of particles, we show that a single model can reproduce both experimental results by varying the number of particles—increasing the number of particles leads to both faster category learning and improved strategy‐switching. Furthermore, when we fit the model to individual participants, we found a positive association between WMC and best‐fit number of particles for strategy switching. However, no association between WMC and best‐fit number of particles was found for category learning. These results are discussed in the context of the general challenge of disentangling the contributions of different potential sources of behavioral variability. (shrink)

Identifying patterns in the world requires noticing not only unusual occurrences, but also unusual absences. We examined how people learn from absences, manipulating the extent to which an absence is expected. People can make two types of inferences from the absence of an event: either the event is possible but has not yet occurred, or the event never occurs. A rational analysis using Bayesian inference predicts that inferences from absent data should depend on how much the absence is expected to (...) occur, with less probable absences being more salient. We tested this prediction in two experiments in which we elicited people's judgments about patterns in the data as a function of absence salience. We found that people were able to decide that absences either were mere coincidences or were indicative of a significant pattern in the data in a manner that was consistent with predictions of a simple Bayesian model. (shrink)

Humans and many other species selectively attend to stimuli or stimulus dimensions—but why should an animal constrain information input in this way? To investigate the adaptive functions of attention, we used a genetic algorithm to evolve simple connectionist networks that had to make categorization decisions in a variety of environmental structures. The results of these simulations show that while learned attention is not universally adaptive, its benefit is not restricted to the reduction of input complexity in order to keep it (...) within an organism's processing capacity limitations. Instead, being able to shift attention provides adaptive benefit by allowing faster learning with fewer errors in a range of ecologically plausible environments. (shrink)

Constructing an intuitive theory from data confronts learners with a “chicken‐and‐egg” problem: The laws can only be expressed in terms of the theory's core concepts, but these concepts are only meaningful in terms of the role they play in the theory's laws; how can a learner discover appropriate concepts and laws simultaneously, knowing neither to begin with? We explore how children can solve this chicken‐and‐egg problem in the domain of magnetism, drawing on perspectives from computational modeling and behavioral experiments. We (...) present 4‐ and 5‐year‐olds with two different simplified magnet‐learning tasks. Children appropriately constrain their beliefs to two hypotheses following ambiguous but informative evidence. Following a critical intervention, they learn the correct theory. In the second study, children infer the correct number of categories given no information about the possible causal laws. Children's hypotheses in these tasks are explained as rational inferences within a Bayesian computational framework. (shrink)

This paper studies the learnability of natural concepts in the context of the conceptual spaces framework. Previous work proposed that natural concepts are represented by the cells of optimally partitioned similarity spaces, where optimality was defined in terms of a number of constraints. Among these is the constraint that optimally partitioned similarity spaces result in easily learnable concepts. While there is evidence that systems of concepts generally regarded as natural satisfy a number of the proposed optimality constraints, the connection between (...) naturalness and learnability has been less well studied. To fill this gap, we conduct a computational study employing two standard models of concept learning. Applying these models to the learning of color concepts, we examine whether natural color concepts are more readily learned than nonnatural ones. Our findings warrant a positive answer to this question for both models employed, thus lending empirical support to the notion that learnability is a distinctive characteristic of natural concepts. (shrink)

Meta-learning is even more broadly relevant to the study of inductive biases than Binz et al. suggest: Its implications go beyond the extensions to rational analysis that they discuss. One noteworthy example is that meta-learning can act as a bridge between the vector representations of neural networks and the symbolic hypothesis spaces used in many Bayesian models.

The causal view of categories assumes that categories are represented by features and their causal relations. To study the effect of causal knowledge on categorization, researchers have used Bayesian causal models. Within that framework, categorization may be viewed as dependent on a likelihood computation (i.e., the likelihood of an exemplar with a certain combination of features, given the category's causal model) or as a posterior computation (i.e., the probability that the exemplar belongs to the category, given its features). Across three (...) experiments, in combination with computational modeling, we offer evidence that categorization is better accounted for by assuming that people compute posteriors and not likelihoods, though both probabilities are closely related. This result contrasts with existing analyses of causal-based categorization, which assume that likelihood computations give a good approximation of human judgments. We also find that people are able to compute likelihoods in a closely related task that elicits judgments of consistency rather than category membership judgments. Our analyses show that people do use causal probabilistic information as prescribed by a Bayesian model but that they flexibly compute likelihoods or posteriors depending on the task. We discuss our results in relation to the relevant literature on the topic. (shrink)

According to standard linguistic theory, the meaning of an utterance is the product of conventional semantic meaning and general pragmatic rules on language use. We investigate how such a division of labor between semantics and pragmatics could evolve under general processes of selection and learning. We present a game‐theoretic model of the competition between types of language users, each endowed with certain lexical representations and a particular pragmatic disposition to act on them. Our model traces two evolutionary forces and their (...) interaction: (i) pressure toward communicative efficiency and (ii) transmission perturbations during the acquisition of linguistic knowledge. We illustrate the model based on a case study on scalar implicatures, which suggests that the relationship between underspecified semantics and pragmatic inference is one of coevolution. (shrink)

We were encouraged by the broad enthusiasm for building machines that learn and think in more human-like ways. Many commentators saw our set of key ingredients as helpful, but there was disagreement regarding the origin and structure of those ingredients. Our response covers three main dimensions of this disagreement: nature versus nurture, coherent theories versus theory fragments, and symbolic versus sub-symbolic representations. These dimensions align with classic debates in artificial intelligence and cognitive science, although, rather than embracing these debates, we (...) emphasize ways of moving beyond them. Several commentators saw our set of key ingredients as incomplete and offered a wide range of additions. We agree that these additional ingredients are important in the long run and discuss prospects for incorporating them. Finally, we consider some of the ethical questions raised regarding the research program as a whole. (shrink)