Trends in Cognitive Sciences
Volume 5, Issue 10, 1 October 2001, Pages 434-442
Journal home page for Trends in Cognitive Sciences

Review
Mental models and deduction

https://doi.org/10.1016/S1364-6613(00)01751-4Get rights and content

Abstract

According to the mental-model theory of deductive reasoning, reasoners use the meanings of assertions together with general knowledge to construct mental models of the possibilities compatible with the premises. Each model represents what is true in a possibility. A conclusion is held to be valid if it holds in all the models of the premises. Recent evidence described here shows that the fewer models an inference calls for, the easier the inference is. Errors arise because reasoners fail to consider all possible models, and because models do not normally represent what is false, even though reasoners can construct counterexamples to refute invalid conclusions.

Section snippets

The theory of mental models

Thinking depends on tacit processes that are guided by constraints: the thinker's goal, if any, and pertinent knowledge and beliefs. The idea that it depends on mental models goes back to the Scottish psychologist, Kenneth Craik, who suggested that perception constructs ‘small-scale models’ of reality that are used to anticipate events and to reason 8. Mental models might originally have evolved as the ultimate output of perceptual processes. They can represent spatial relations 9, events and

One model is better than many

The fewer the number of models needed for an inference, and the simpler they are, the easier the inference should be. It should take less time, and be less prone to error. This prediction is a consequence of the limitations of working memory 17: multiple models can overload its processing capacity and lead to errors in which reasoners fail to consider some models of the premises. Halford and his colleagues have shown that the fewer the entities in a model of a relation, the easier inferences

Truth, meaning, and knowledge

Reasoners focus on what is true and neglect what is false. One consequence is the difficulty of the selection task (Box 3). Another consequence is the occurrence of systematic fallacies (Box 4). And because meaning is central to models, the content of inferences and background knowledge can modulate reasoning. They influence the interpretation of premises 37. The following inference, for example, is valid in form:

  • Eva is in Rio or she's in Brazil;

  • She's not in Brazil.

  • Therefore, she's in Rio.

But

Strategies in reasoning

An important recent discovery is that when individuals carry out a series of inferences, they develop strategies for coping with them. Deduction itself can be a strategy 44, and Western cultures might resort to it more than East Asian cultures 45. However, deduction in turn elicits a variety of strategies 46. An earlier version of the mental-model theory implied that reasoners start reasoning with the most informative premise but this claim is not always true 47. Reasoners’ strategies determine

Conclusions

Deductive reasoning is under intense investigation 59. The field is fast moving and controversial. This article has reviewed just one theory: that reasoning depends on imagining the possibilities compatible with the premises, and drawing conclusions from these mental models. The theory makes five main predictions, which have been corroborated experimentally:

  • 1.

    One model is better than many. That is, the fewer models needed for an inference, and the simpler they are, the easier the inference.

  • 2.

Questions for future research

  • What underlies the vast differences in reasoning ability from one individual to another? The processing capacity of working memory accounts for some, but not all, of these differences.

  • Reasoning takes into account relevant general knowledge. What process triggers its recovery from long-term memory?

  • What is the origin of the different strategies for reasoning, and how do reasoners develop them?

  • Logic means never having to be sorry about a valid conclusion. In daily life, however, individuals

Acknowledgements

Preparation of this article was supported by a grant from the National Science Foundation (Grant 0076287) to study strategies in reasoning. It was made possible by the community of reasoning researchers. There are too many individuals to name, but the author thanks them all.

References (59)

  • J. Harr

    A Civil Action

    (1995)
  • L.J. Rips

    The Psychology of Proof

    (1994)
  • R. Jeffrey

    Formal Logic: Its Scope and Limits

    (1981)
  • P.N. Johnson-Laird et al.

    Deduction

    (1991)
  • B.G. Bara

    In favour of a unified model of deductive reasoning

  • K. Craik

    The Nature of Explanation

    (1943)
  • J. Glasgow

    Array representations for model-based spatial reasoning

  • N. Moray

    Mental models in theory and practice

  • P.N. Johnson-Laird

    Imagery, visualization, and thinking

  • A. Garnham et al.

    The mental models theory of language comprehension

  • P.N. Johnson-Laird

    Mental Models: Towards a Cognitive Science of Language, Inference, and Consciousness

    (1983)
  • K. Stenning et al.

    Image and language in human reasoning: a syllogistic illustration

    Cognit. Psychol.

    (1997)
  • G.S. Halford

    Processing capacity defined by relational complexity: implications for comparative, developmental, and cognitive psychology

    Brain Behav. Sci.

    (1998)
  • J. Evans et al.

    Human Reasoning: The Psychology of Deduction

    (1993)
  • P.N. Johnson-Laird

    Deductive reasoning

    Annu. Rev. Psychol.

    (1999)
  • W.S. Schaeken

    Mental models and temporal reasoning

    Cognition

    (1996)
  • M. Carreiras et al.

    Reasoning about relations: spatial and nonspatial problems

    Think. Reason.

    (1997)
  • A. Vandierendonck et al.

    Working memory constraints on linear reasoning with spatial and temporal contents

    Q. J. Exp. Psychol.

    (1997)
  • Cited by (267)

    • If and only if people were logical! The effect of pragmatic enrichment on reasoning with abstract and realistic materials

      2022, Journal of Pragmatics
      Citation Excerpt :

      On the one hand, pragmatic theories of language tacitly assume that the perfection of conditionals happens in the environment of natural language, that is, in everyday meaningful contexts, such as in (6) rather than (5). On the other hand, Bonnefond and Van der Henst (2009) observed that in the domain of the psychology of reasoning the two most popular Models of deduction – Mental Logic (Braine and O'Brien, 1998) and Mental Model Theory (Johnson-Laird and Byrne, 1991; Johnson-Laird, 2001) – treat reasoning as an uniform system, that is, a system that works in a similar way independently of the type of premises given as input. However, as Bonnefond and Van der Henst (2009) point out, fMRI studies have shown that distinct brain networks are involved depending on the content of the premises in a given reasoning task.

    • Reasoning and decision making

      2022, International Encyclopedia of Education: Fourth Edition
    View all citing articles on Scopus
    View full text