Keith Stenning

Computer-based logic proofs are a form of unnatural language in which the process and structure of proof generation can be observed in considerable detail. We have been studying how students respond to multimodal logic teaching, and performance measures have already indicated that students' pre-existing cognitive styles have a significant impact on teaching outcome. Furthermore, a large corpus of proofs has been gathered via automatic logging of proof development. This paper applies a series of techniques, including corpus statistical methods, to the proof logs. The results indicate that students' cognitive styles influence the structure of their logical discourse, via their differing methods of handling abstract information in diagrams, and transferring information between modalities.

In the late summer of 1998, the authors, a cognitive scientist and a logician, started talking about the relevance of modern mathematical logic to the study of human reasoning, and we have been talking ever since. This book is an interim report of that conversation. It argues that results such as those on the Wason selection task, purportedly showing the irrelevance of formal logic to actual human reasoning, have been widely misinterpreted, mainly because the picture of logic current in psychology and cognitive science is completely mistaken. We aim to give the reader a more accurate picture of mathematical logic and, in doing so, hope to show that logic, properly conceived, is still a very helpful tool in cognitive science. The main thrust of the book is therefore constructive. We give a number of examples in which logical theorizing helps in understanding and modeling observed behavior in reasoning tasks, deviations of that behavior in a psychiatric disorder (autism), and even the roots of that behavior in the evolution of the brain.

Machine generated contents note: -- Preface -- Acknowledgements -- Notes on Contributors -- PART I: COMPLEXITY IN ANIMAL MINDS -- Introduction: M.McGonigle-Chalmers -- Relational and Absolute Discrimination Learning by Squirrel Monkeys: Establishing a Common Ground with Human Cognition; B.T.Jones -- Serial List Retention by Non-Human Primates: Complexity and Cognitive Continuity; F.R.Treichler -- The Use of Spatial Structure in Working Memory: A Comparative Standpoint; C.De Lillo -- The Emergence of Linear Sequencing in Children: A Continuity Account and a Formal Model; M.McGonigle-Chalmers&I.Kusel -- Sensitivity to Quantity: What Counts Across Species?; S.T.Boysen&A.M.Yocom -- PART II: COMPLEXITY IN ROBOTS -- Editorial Introduction; D.McFarland -- Towards Cognitive Robotics: Robotics, Biology and Developmental Psychology; M.Lee, U.Nehmzow&M.Rodriguez -- Structuring Intelligence: The Role of Hierarchy, Modularity and Learning in Generating Intelligent Behaviour; J.J.Bryson -- Epistemology, Access, and Computational Models; G.Luger -- Reasoning About Representations in Autonomous Systems: What P´Olya and Lakatos Have To Say; A.Bundy -- PART III: LANGUAGE, EVOLUTION AND THE COMPLEX MIND -- Editorial Introduction; K.Stenning -- How to Qualify for a Cognitive Upgrade: Executive Control, Glass Ceilings, and the Limits of Simian Success; A.Clark -- Private Codes and Public Structures; C.Allen -- The Emergence of Complex Language; W.Hinzen -- Language Evolution: Enlarging the Picture; K.Stenning&M.Van Lambalgen -- Epilogue: Reminiscences of Brendan McGonigle -- Index.

Classical symbolic computational models of cognition are at variance with the empirical findings in the cognitive psychology of memory and inference. Standard symbolic computers are well suited to remembering arbitrary lists of symbols and performing logical inferences. In contrast, human performance on such tasks is extremely limited. Standard models donot easily capture content addressable memory or context sensitive defeasible inference, which are natural and effortless for people. We argue that Connectionism provides a more natural framework in which to model this behaviour. In addition to capturing the gross human performance profile, Connectionist systems seem well suited to accounting for the systematic patterns of errors observed in the human data. We take these arguments to counter Fodor and Pylyshyn's (1988) recent claim that Connectionism is, in principle, irrelevant to psychology

The notion of “bounded rationality” was introduced by Simon as an appropriate framework for explaining how agents reason and make decisions in accordance with their computational limitations and the characteristics of the environments in which they exist (seen metaphorically as two complementary scissor blades).We elaborate on how bounded rationality is usually conceived in psychology and on its relationship with logic. We focus on the relationship between heuristics and some non-monotonic logical systems. These two categories of cognitive tools share fundamental features. As a step further, we show that in some cases heuristics themselves can be formalized from this logic perspective. We have therefore two main aims: on the one hand, to demonstrate the relationship between the bounded rationality programme and logic, understood in a broad sense; on the other hand, to provide logical tools of analysis of already known heuristics. This may lead to results such as the characterization of fast and frugal binary trees in terms of their associated logic program here provided.

Interpretation is the process whereby a hearer reasons to an interpretation of a speaker's discourse. The hearer normally adopts a credulous attitude to the discourse, at least for the purposes of interpreting it. That is to say the hearer tries to accommodate the truth of all the speaker's utterances in deriving an intended model. We present a nonmonotonic logical model of this process which defines unique minimal preferred models and efficiently simulates a kind of closed-world reasoning of particular interest for human cognition. Byrne's “suppression” data (Byrne, 1989) are used to illustrate how variants on this logic can capture and motivate subtly different interpretative stances which different subjects adopt, thus indicating where more fine-grained empirical data are required to understand what subjects are doing in this task. We then show that this logical competence model can be implemented in spreading activation network models. A one pass process interprets the textual input by constructing a network which then computes minimal preferred models for (3-valued) valuations of the set of propositions of the text. The neural implementation distinguishes easy forward reasoning from more complex backward reasoning in a way that may be useful in explaining directionality in human reasoning.

We review the various explanations that have been offered toaccount for subjects'' behaviour in Wason ''s famous selection task. Weargue that one element that is lacking is a good understanding ofsubjects'' semantics for the key expressions involved, and anunderstanding of how this semantics is affected by the demands the taskputs upon the subject''s cognitive system. We make novel proposals inthese terms for explaining the major content effects of deonticmaterials. Throughout we illustrate with excerpts from tutorialdialogues which motivate the kinds of analysis proposed. Our long termgoal is an integration of the various insights about conditionalreasoning on offer from different cognitive science methodologies. Thepurpose of this paper is to try to draw the attention of logicians andsemanticists to this area, since we believe that empirical investigationof the cognitive processes involved could benefit from semanticanalyses.

Executive function has become an important concept in explanations of psychiatric disorders, but we currently lack comprehensive models of normal executive function and of its malfunctions. Here we illustrate how defeasible logical analysis can aid progress in this area. We illustrate using autism and attention deficit hyperactivity disorder (ADHD) as example disorders, and show how logical analysis reveals commonalities between linguistic and non-linguistic behaviours within each disorder, and how contrasting sub-components of executive function are involved across disorders. This analysis reveals how logical analysis is as applicable to fast, automatic and unconscious reasoning as it is to slow deliberate cogitation.

Executive function has become an important concept in explanations of psychiatric disorders, but we currently lack comprehensive models of normal executive function and of its malfunctions. Here we illustrate how defeasible logical analysis can aid progress in this area. We illustrate using autism and attention deficit hyperactivity disorder (ADHD) as example disorders, and show how logical analysis reveals commonalities between linguistic and non-linguistic behaviours within each disorder, and how contrasting sub-components of executive function are involved across disorders. This analysis reveals how logical analysis is as applicable to fast, automatic and unconscious reasoning as it is to slow deliberate cogitation.

Modern logic provides accounts of both interpretation and derivation which work together to provide abstract frameworks for modelling the sensitivity of human reasoning to task, context and content. Cognitive theories have underplayed the importance of interpretative processes. We illustrate, using Wason's [Q. J. Exp. Psychol. 20 (1968) 273] selection task, how better empirical cognitive investigations and theories can be built directly on logical accounts when this imbalance is redressed. Subjects quite reasonably experience great difficulty in assigning logical form to descriptively interpreted rules in this task and materials, though not to deontically interpreted ones. The main empirical contrasts in reasoning performance are precisely between these interpretations, so better semantic theory can explain the main empirical effects. However, a more adequate notion of logical form than that typically employed in the psychology of reasoning is required, and we offer a richer alternative. Prima facie evidence that subjects do in fact experience the predicted difficulties in interpreting descriptive materials is provided by analyses of socratic tutoring dialogues. Experimental verification that these difficulties do in fact affect subjects' performance in the standard task is provided by six novel experimental conditions each designed to test different aspects of the semantic predictions. The results bear out the predictions. The semantic distinction between descriptive and deontic rules interacts with the task specifics to provide powerful generalisations about reasoning which surface in detailed explanations of many disparate observations. We conclude that semantic analyses have more direct benefits for psychological investigation than is usually credited, and conversely, that the extraordinary pragmatic circumstances of psychological experiments yield much thought provoking data for semantic analysis. © 2004 Cognitive Science Society, Inc. All rights reserved.Subjects quite reasonably experience great difficulty in assigning logical form to descriptively interpreted rules in this task and materials, though not to deontically interpreted ones. The main empirical contracts in reasoning performance are precisely between these interpretatins, so better semantic theory can explain the main empirical effects. Howevers, a more adequate nation of logical from than that typically employed in the psychology of reasoning in required, and we offer a richer alternative.Prima facie evidence that subjects do in fact experience the predicted difficulties in interpreting descriptive materials is provided by analyses of socratic tutoring dialogues. Experimental verification that these difficulties do in fact affect subjects' performance in the standard task is provided by six novel experimental conditions each designed to test different aspects of the semantic predications.The resluts bear out the predictions. The semantic distinction between descriptive and deontic rules interacts with the task specifics to provide powerful generalisations about reasoning which surface in detailed explanations of many disparate observations. We conclude that semantic analyses have more direct benefits for psychological investigation than is usually credited, and conversely, that the extraordinary pragmatic circumstances of psychological experiments yield much thought provoking data for semantic analysis.

This reply to Oaksford and Chater’s ’s critical discussion of our use of logic programming to model and predict patterns of conditional reasoning will frame the dispute in terms of the semantics of the conditional. We begin by outlining some common features of LP and probabilistic conditionals in knowledge-rich reasoning over long-term memory knowledge bases. For both, context determines causal strength; there are inferences from the absence of certain evidence; and both have analogues of the Ramsey test. Some current work shows how a combination of counting defeaters and statistics from network monitoring can provide the information for graded responses from LP reasoning. With this much introduction, we then respond to O&C’s specific criticisms and misunderstandings.

Alternative logics have been invoked periodically to explain the systematically different modes of thought of the subjects of ethnography: one logic for ‘us’ and another for ‘them’. Recently anthropologists have cast doubt on the tenability of such an explanation of difference. In cognitive science, [Stenning and van Lambalgen, 2008] proposed that with the modern development of multiple logics, at least several logics are required for making sense of the cognitive processes of reasoning for different purposes and in different contexts. Alongside Classical logic — the logic of dispute), there is a need for a nonmonotonic logic which is a logic of cooperative communication. Here we propose that all people with various cultural backgrounds make use of multiple logics, and that difference should be captured as variation in the social contexts that call forth the different logics’ application. This contribution illustrates these ideas with reference to the ethnography of divination.

Interpretation is the process whereby a hearer reasons to an interpretation of a speaker's discourse. The hearer normally adopts a credulous attitude to the discourse, at least for the purposes of interpreting it. That is to say the hearer tries to accommodate the truth of all the speaker's utterances in deriving an intended model. We present a nonmonotonic logical model of this process which defines unique minimal preferred models and efficiently simulates a kind of closed‐world reasoning of particular interest for human cognition. Byrne's “suppression” data (Byrne, 1989) are used to illustrate how variants on this logic can capture and motivate subtly different interpretative stances which different subjects adopt, thus indicating where more fine‐grained empirical data are required to understand what subjects are doing in this task. We then show that this logical competence model can be implemented in spreading activation network models. A one pass process interprets the textual input by constructing a network which then computes minimal preferred models for (3‐valued) valuations of the set of propositions of the text. The neural implementation distinguishes easy forward reasoning from more complex backward reasoning in a way that may be useful in explaining directionality in human reasoning.

"A structured connectionist model using temporal synchrony has been proposed by Shastri & Ajjanagadde. This model has provided a mechanism which encodes rules and facts involving n-ary predicates and handles some types of dynamic variable binding. Their model, however, has problems in dealing with important knowledge representation issues such as binding generation, consistency checking, unification, etc. This paper presents a restructuring of the original model which overcomes many of its limitations while, at the same time, reducing the number of types of node [sic] required and maintaining the merits of the original model."

We discuss external and internal graphical and linguistic representational systems. We argue that a cognitive theory of peoples' reasoning performance must account for (a) the logical equivalence of inferences expressed in graphical and linguistic form, and (b) the implementational differences that affect facility of inference. Our theory proposes that graphical representation limit abstraction and thereby aid “processibility”. We discuss the ideas of specificity and abstraction, and their cognitive relevance. Empirical support both comes from tasks which involve the manipulation of external graphics and tasks that do not. For the former, we take Euler's (1772) circles, provide a novel computational reconstruction, show how it captures abstractions, and contrast it with earlier construals and with Johnson‐Laird's (1983) mental models representations. We demonstrate equivalence of the graphical Euler system, and the nongraphical mental models system. For tasks not involving manipulation of external graphics, we discuss text comprehension, and the mental performance of syllogisms. By positing an internal system with the same specificity as Euler's circles, we cover the mental models data, and generate new empirical predictions. Finally, we consider how the architecture of working memory explains why such specific representations are relatively easy to store.

Our long term goal is an understanding of human communication in terms which would provide the basis for rational design. The kernel would be a theory of the cognitive consequences of allocating the same information to different media and modalities, based on the user's information processing characterised in computational terms. Our theory of the cognitive consequences of media/modality allocation starts from an analysis of differences in logical expressiveness of graphical and linguistic representations (Stenning \& Oberlander (1994, 1995)). This semantic approach requires conceptualisations of {\it medium} and {\it modality} that can be related to representation systems. We propose that media are the physical/perceptual aspects of representations; modalities are classes of interpretation function which map media onto meanings. These interpretations of the terms contrast with existing HCI usage. The curtailments of abstraction observed in graphics arise from interactions between medium and modality. Graphical modalities are distinguished from sentential languages by the nature of their interpretation functions. A hierarchy of expressiveness of interpretations of graphics is defined, and compared with interpretations of sentential languages. Using the expressiveness of representations to predict their cognitive properties also requires reference to the availability of constraints of their interpretation to users. Further contrasts between graphics and language emerge in the availability of their constraints. Three example domains of graphical representations are analysed from this perspective---matrix graphics; logic diagrams; and semantic networks. Some empirical evidence of the usability of these notations is reviewed as evidence that the proposed conceptualisation offers powerful generalisations.