Keith Stenning

Reasoning with conditionals containing negated constituents

Three experiments investigated matching bias in conditional reasoning tasks. Matching bias occurs when Ss ignore negations and match named items. Experiment 1 used an abstract and a thematic version of Evanss (1972) construction task. Results showed that matching may be due to an interaction between task demands and constructing contrast classes when interpreting negations. Experiment 2, which used Wasons (1968) selection task, introduced a manipulation to ease contrast-class construction. Confirmation plus falsification dominated over matching. Experiment 3 introduced two other manipulations to aid contrast-class construction with abstract material. Confirmation was facilitated, matching was suppressed, and falsification remained unchanged. These results suggest that matching occurs only when insufficient or ambiguous information prevents the intended interpretation of negations.

Why observing a dialogue may benefit learning

The Vicarious Learner project is investigating the fundamental role of dialogue for learning. More specifically, the project is exploring the benefits to learners of being able to observe others participating in discussion. Such opportunities are becoming fewer with the proliferation of computer-based courses and distance learning, as well as growing student numbers. This paper presents the theoretical aspects of the work. A high-level ‘process model’ of learning is presented and then a more detailed model of what happens in educational dialogues. This ‘logic model’ of dialogue breaks discussions into two parts: the introduction of new premises and the derivation of new premises by the application of rules of reasoning in the domain. It is argued that in ordinary conversation the first aspect dominates while in educational dialogues, many misunderstandings arise from the need for more emphasis on explicit demonstration of use of rules.

Contrasting the cognitive effects of graphical and sentential logic teaching: Reasoning, representation and individual differences

Abstract Hyperproof is a computer program created by Barwise and Etchemendy (1994) for teaching logic using multimodal graphical and sententid representations. Elsewhere, we have proposed a theory of the cognitive impact of assigning information to different modalities. The theory predicts that Hyperproofs devices for graphical abstraction will play a pivotal role in determining learning outcomes. Here, the claims are tested by a controlled comparison of the effects of teaching undergraduate classes using Hyperproof and a traditional syntactic teaching method. The results indicate that there is significant transfer from the logic courses to a range of verbal reasoning problems. There are also significant interactions between theoretically motivated precourse aptitude measures and teaching methods, and these interactions influence post-course reasoning performance in transfer domains. As well as being theoretically significant, the results provide support for the important practical conclusion that indivi...

The Role of Representation in Teaching and Learning Critical Thinking

Curriculum design in recent years reflects the growing belief in the importance of developing thinking skills. In this paper, we focus on a particular theoretical approach to the study and teaching of thinking: cognitive science. We first give a very brief review of the recent research on critical thinking. We then concentrate on what cognitive science can add by looking at models of how people learn and how they transfer what they learn from one context to another. The main concept we focus on is representation and the crucial aspects of systems of representation and the meaningful transformation of information. We present examples of how representations can support, or hinder, problem solving and communication. We also discuss the social aspects of representation, the challenges of language use, and the paradoxes thrown up by attempting to guide students to be more critical thinkers.

On the construction of representations for individuals from descriptions in text

Abstract Subjects read texts describing pairs of individuals sentence by sentence in a self-paced reading time task and then answered questions about the individuals and recalled them. The task is designed to explore how people represent the binding of attributes to individuals. The reading time data show that the construction of representations is organised around what is known about the currently referenced individual. The more that is known the more slowly subjects read. This slowing is not due to articulatory rehearsal. A regression model of reading times describes the partitioning of working memory resources across the semantic structures being processed. The construction processes yield redundant representations consisting of sets of feature values encoding aspects of the information in the text and contributing independently to memory performance. Modelling in terms of feature representations enables prediction of the patterns of error in recall. The reading time and recall error models are interpr...

The Role of Parallelism in Strategies of Pronoun Comprehension

Two experiments investigated the role of parallelism in strategies of pronoun assignment and tested the proposition that a parallel function strategy and the subject assignment strategy jointly contribute to the interpretation of pronouns. The parallel function strategy interprets a pronoun as coreferential with a preceding noun phrase in the same grammatical role, while the subject assignment strategy interprets a pronoun as coreferential with a preceding subject noun phrase. The results of Experiment I showed that the subject assignment and parallel function strategies jointly constrained assignment. When both strategies yielded the same interpretation, as was the case for subject assignments with subject pronouns, subject assignment preferences were greater than when only the subject assignment strategy was available, as was the case with nonsubject pronouns in partially parallel sentences (i.e., sentences in which the pronoun and antecedent differed grammatically). When the two strategies yielded conflicting assignments, as was the case with nonsubject pronouns in fully parallel sentences, subject assignments were reduced even further and nonsubject assignments were preferred. Experiment 2 replicated the results of Experiment 1 in isolated sentences rather than in short texts and ruled out the idea that a third “parallel order” strategy might be used in partially parallel sentences. We suggest that the subject assignment strategy reflects the topic status of the subject noun phrase, while parallel function reflects the use of correspondences between sentence structures in comprehension. We also suggest that the joint operation of heuristic strategies implies a model of discourse processing in which a number of constraints compete in the interpretation of noun phrases.

Learning how to tell a good story: The development of content and language in children's telling of one tale

This paper describes the development of explanation and its expression in the storytelling of 5‐ to 10‐year‐old children. The children were asked to tell the story contained in a picture book as they looked through it, and afterward were asked questions about the key sequences. These sequences in the childrens narratives were coded as Confused, Descriptive, or Explanatory and the language was coded for several features important for establishing textual coherence. In this task, explanation appears earlier than the literature suggests. Nevertheless, in the youngest children it is limited by their comprehension of the story, while many older children do not do justice in their stories to the complexity of understanding revealed in their responses to questions. The linguistic variables each show different patterns of correlation with the cognitive measures and only low intercorrelations with one another. One stylistic feature, the inclusion of connectives other than and/then is a good predictor of explanati...

Visual/verbal-analytic reasoning bias as a function of self-reported autistic-like traits: A study of typically developing individuals solving Raven’s Advanced Progressive Matrices

People with autism spectrum condition (ASC) perform well on Raven’s matrices, a test which loads highly on the general factor in intelligence. However, the mechanisms supporting enhanced performance on the test are poorly understood. Evidence is accumulating that milder variants of the ASC phenotype are present in typically developing individuals, and that those who are further along the autistic-like trait spectrum show similar patterns of abilities and impairments as people with clinically diagnosed ASC. We investigated whether self-reported autistic-like traits in a university student sample, assessed using the Autism-Spectrum Quotient (AQ; Baron-Cohen, Wheelwright, Skinner, et al., 2001), predict performance on Raven’s Advanced Progressive Matrices. We found that reporting poorer social skills but better attention switching predicted a higher Advanced matrices score overall. DeShon, Chan, and Weissbein (1995) classified Advanced matrices items as requiring a visuospatial, or a verbal-analytic strategy. We hypothesised that higher AQ scores would predict better performance on visuospatial items than on verbal-analytic items. This prediction was confirmed. These results are consistent with the continuum view and can be explained by the enhanced perceptual functioning theory of performance peaks in ASC. The results also confirm a new prediction about Raven’s Advanced Progressive Matrices performance in people with ASC.

Reconnecting interpretation to reasoning through individual differences

Computational theories of mind assume that participants interpret information and then reason from those interpretations. Research on interpretation in deductive reasoning has claimed to show that subjects’ interpretation of single syllogistic premises in an “immediate inference” task is radically different from their interpretation of pairs of the same premises in syllogistic reasoning tasks (Newstead, 1989, 1995; Roberts, Newstead, & Griggs, 2001). Narrow appeal to particular Gricean implicatures in this work fails to bridge the gap. Grices theory taken as a broad framework for credulous discourse processing in which participants construct speakers’ “intended models” of discourses can reconcile these results, purchasing continuity of interpretation through variety of logical treatments. We present exploratory experimental data on immediate inference and subsequent syllogistic reasoning. Systematic patterns of interpretation driven by two factors (whether the subjects model of the discourse is credulous, and their degree of reliance on information packaging) are shown to transcend particular quantifier inferences and to drive systematic differences in subjects’ subsequent syllogistic reasoning. We conclude that most participants do not understand deductive tasks as experimenters intend, and just as there is no single logical model of reasoning, so there is no reason to expect a single “fundamental human reasoning mechanism”.

Connectionism, classical cognitive science and experimental psychology

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 Pylyshyns (1988) recent claim that Connectionism is, in principle, irrelevant to psychology.