Skip to main content
Log in

Why Machines Don’t (yet) Reason Like People

  • Technical Contribution
  • Published:
KI - Künstliche Intelligenz Aims and scope Submit manuscript

Abstract

AI has never come to grips with how human beings reason in daily life. Many automated theorem-proving technologies exist, but they cannot serve as a foundation for automated reasoning systems. In this paper, we trace their limitations back to two historical developments in AI: the motivation to establish automated theorem-provers for systems of mathematical logic, and the formulation of nonmonotonic systems of reasoning. We then describe why human reasoning cannot be simulated by current machine reasoning or deep learning methodologies. People can generate inferences on their own instead of just evaluating them. They use strategies and fallible shortcuts when they reason. The discovery of an inconsistency does not result in an explosion of inferences—instead, it often prompts reasoners to abandon a premise. And the connectives they use in natural language have different meanings than those in classical logic. Only recently have cognitive scientists begun to implement automated reasoning systems that reflect these human patterns of reasoning. A key constraint of these recent implementations is that they compute, not proofs or truth values, but possibilities.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1

Similar content being viewed by others

References

  1. Baratgin J, Douven I, Evans J, Oaksford M, Over D, Politzer G (2015) The new paradigm and mental models. Trends Cogni Sci 19(10):547–548

    Article  Google Scholar 

  2. Bledsoe W (1977) Non-resolution theorem proving. Artif Intell 9:1–35

    Article  MathSciNet  MATH  Google Scholar 

  3. Bonacina MP (1999) A taxonomy of theorem-proving strategies. In: Artificial intelligence today. Springer, Berlin, Heidelberg, pp 43–84

    Chapter  Google Scholar 

  4. Bonacina MP, Furbach U, Sofronie-Stokkermans V (2015) On first-order model-based reasoning. In: Martí-Oliet N, Ölveczky P, Talcott C (eds) Logic, rewriting, and concurrency. Springer, Berlin

    Google Scholar 

  5. Braine MDS (1978) On the relation between the natural logic of reasoning and standard logic. Psychol Rev 85:1–21

    Article  Google Scholar 

  6. Brewka G, Dix J, Konolige K (1997) Nonmonotonic reasoning: an overview, vol 73. CSLI publications, Stanford

    MATH  Google Scholar 

  7. Elqayam S, Over DE (2013) New paradigm psychology of reasoning: an introduction to the special issue edited by elqayam, bonnefon, and over. Think Reason 19(3–4):249–265

    Article  Google Scholar 

  8. Garey MR, Johnson D (2002) Computers and intractability. W.H. Freeman, New York

    Google Scholar 

  9. Gentzen G (1969) Investigations into logical deduction. The collected papers of Gerhard Gentzen, pp 68–131

  10. Ginsberg ML (1994) AI and nonmonotonic reasoning. In: Handbook of logic in artificial intelligence and logic programming, vol 3. Oxford University Press, Inc., pp. 1–33

  11. Girle R (2009) Modal logics and philosophy. Routledge, Abingdon

    Google Scholar 

  12. Goodrich MA, Schultz AC (2008) Human-robot interaction: a survey. Found Trendsin Hum Comput Interact 1(3):203–275

    Article  MATH  Google Scholar 

  13. Goodwin GP, Johnson-Laird P (2005) Reasoning about relations. Psychol Rev 112(2):468

    Article  Google Scholar 

  14. Halpern JY, Vardi M (1991) Model checking vs. theorem proving: a manifesto. Artif Intell Math Theory Comput 212:151–176

    Article  MathSciNet  MATH  Google Scholar 

  15. Hattori M (2016) Probabilistic representation in syllogistic reasoning: a theory to integrate mental models and heuristics. Cognition 157:296–320

    Article  Google Scholar 

  16. Van der Henst JB, Yang Y, Johnson-Laird PN (2002) Strategies in sentential reasoning. Cogn Sci 26(4):425–468

    Article  Google Scholar 

  17. Hinterecker T, Knauff M, Johnson-Laird P (2016) Modality, probability, and mental models. J Exp Psychol 42(10):1606

    Google Scholar 

  18. Jalal S (2015) Non-monotonic reasoning: mimicking human thought process through argumentation. University of California, Davis

    Google Scholar 

  19. Jeffrey R (1981) Formal logic: its scope and limits, 2nd edn. McGraw-Hill, New York City

    MATH  Google Scholar 

  20. Jiang Y, Papapanagiotou P, Fleuriot J (2018) Machine learning for inductive theorem proving. In: International conference on artificial intelligence and symbolic computation. Springer, Cham, pp 87–103

  21. Johnson-Laird PN (1983) Mental models: towards a cognitive science of language, inference, and consciousness. Harvard University Press, Cambridge

    Google Scholar 

  22. Johnson-Laird PN (2006) How we reason. Oxford University Press, Oxford

    Google Scholar 

  23. Johnson-Laird PN, Byrne R (1991) Deduction: essays in cognitive psychology. Laurence Erlbaum Associates, Mahwah

    Google Scholar 

  24. Johnson-Laird PN, Byrne RM, Schaeken W (1992) Propositional reasoning by model. Psychol Revi 99(3):418

    Article  Google Scholar 

  25. Johnson-Laird PN, Girotto V, Legrenzi P (2004) Reasoning from inconsistency to consistency. Psychol Rev 111(3):640

    Article  Google Scholar 

  26. Johnson-Laird PN, Khemlani SS, Goodwin GP (2015) Logic, probability, and human reasoning. Trends Cogn Sci 19(4):201–214

    Article  Google Scholar 

  27. Keene S (1989) Object-oriented programming in Common LISP: a progammer’s guide to CLOS. Addison-Wesley, Boston

    MATH  Google Scholar 

  28. Khemlani S, Johnson-Laird P (2013) Cognitive changes from explanations. J Cogn Psychol 25(2):139–146

    Article  Google Scholar 

  29. Khemlani S, Johnson-Laird PN (2013) The processes of inference. Argum Comput 4(1):4–20

    Article  Google Scholar 

  30. Khemlani S, Lotstein M, Trafton JG, Johnson-Laird P (2015) Immediate inferences from quantified assertions. Q J Exp Psychol 68(10):2073–2096

    Article  Google Scholar 

  31. Khemlani SS, Byrne RM, Johnson-Laird PN (2018) Facts and possibilities: a model-based theory of sentential reasoning. Cogn Sci 42(6):1887–1924

    Article  Google Scholar 

  32. Khemlani SS, Harrison AM, Trafton JG (2015) Episodes, events, and models. Front Hum Neurosci 9:590

    Article  Google Scholar 

  33. Khemlani SS, Johnson-Laird P (2017) Illusions in reasoning. Minds Mach 27(1):11–35

    Article  Google Scholar 

  34. Khemlani SS, Johnson-Laird PN (2011) The need to explain. Q J Exp Psychol 64(11):2276–2288

    Article  Google Scholar 

  35. Khemlani SS, Johnson-Laird PN (2012) Hidden conflicts: explanations make inconsistencies harder to detect. Acta Psychol 139(3):486–491

    Article  Google Scholar 

  36. Khemlani SS, Lotstein M, Johnson-Laird PN (2015) Naive probability: model-based estimates of unique events. Cogn Sci 39(6):1216–1258

    Article  Google Scholar 

  37. Khemlani SS, Mackiewicz R, Bucciarelli M, Johnson-Laird PN (2013) Kinematic mental simulations in abduction and deduction. In: proceedings of the national academy of sciences, p. 201316275

  38. Kinyon M (2019) Proof simplification and automated theorem proving. Philos Trans R Soc A 377(2140):20180034

    Article  MathSciNet  Google Scholar 

  39. Kowalski R (2011) Computational logic and human thinking. Cambridge University Press, Cambridge

    Book  MATH  Google Scholar 

  40. Kowalski R, Hayes PJ (1983) Semantic trees in automatic theorem-proving. In: Automation of Reasoning. Springer, Berlin, Heidelberg, pp. 217–232

    Chapter  Google Scholar 

  41. Lake BM, Ullman TD, Tenenbaum JB, Gershman SJ (2017) Building machines that learn and think like people. Behav Brain Sci 40

  42. Lassiter D (2017) Graded modality: Qualitative and quantitative perspectives. Oxford University Press, Oxford

    Book  Google Scholar 

  43. Loveland DW (2016) Automated Theorem Proving: a logical basis. Elsevier, Amsterdam

    MATH  Google Scholar 

  44. Marek VW, Truszczynski M (2013) Nonmonotonic logic: context-dependent reasoning. Springer, Berlin

    MATH  Google Scholar 

  45. McCarthy J (1960) Programs with common sense. In: Proceedings of the teddington conference on the mechanization of thought processes. H.M. Stationery Office

  46. McCarthy J (1986) Applications of circumscription to formalizing common sense knowledge. Artif Intell 28:89–116

    Article  MathSciNet  Google Scholar 

  47. McCune W (1997) Solution of the Robbins problem. J Autom Reason 19(3):263–276

    Article  MathSciNet  MATH  Google Scholar 

  48. McDermott D (1987) A critique of pure reason. Comput Intell 3(1):151–160

    Article  Google Scholar 

  49. McDermott D, Doyle J (1980) Non-monotonic logic i. Artif intell 13(1–2):41–72

    Article  MATH  Google Scholar 

  50. Minsky M (1975) Frame-system theory. In: proceedings of the 1975 workshop on theoretical issues in natural language processing, Association for Computational Linguistics, pp. 104–116.

  51. Minsky M (1985) The Society of Mind. Simon and Schuster, New York City

    Google Scholar 

  52. Moosavi-Dezfooli SM, Fawzi A, Frossard P (2016) Deepfool: a simple and accurate method to fool deep neural networks. In: proceedings of the IEEE conference on computer vision and pattern recognition, pp. 2574–2582

  53. Newell A, Shaw JC, Simon HA (1963) Empirical explorations with the logic theory machine. In: Feigenbaum E, Feldman J (eds) Computers and Thought. McGraw-Hill, New York City

    Google Scholar 

  54. Nguyen A, Yosinski J, Clune J (2015) Deep neural networks are easily fooled: high confidence predictions for unrecognizable images. In: proceedings of the IEEE conference on computer vision and pattern recognition, pp. 427–436

  55. Oaksford M, Chater N (2007) Bayesian rationality: the probabilistic approach to human reasoning. Oxford University Press, Oxford

    Book  Google Scholar 

  56. Peirce CS (1931–1958) Collected papers of Charles Sanders Peirce. In: Hartshorne C, Weiss P, Burks A (eds) vol 8. Harvard University Press, Cambridge, MA

  57. Pelletier FJ (1986) Seventy-five problems for testing automatic theorem provers. J Autom Reason 2:191–216

    Article  MathSciNet  MATH  Google Scholar 

  58. Pfeifer N (2013) The new psychology of reasoning: a mental probability logical perspective. Think Reason 19(3–4):329–345. https://doi.org/10.1080/13546783.2013.838189

    Article  Google Scholar 

  59. Ragni M, Eichhorn C, Bock T, Kern-Isberner G, Tse APP (2017) Formal nonmonotonic theories and properties of human defeasible reasoning. Minds Mach 27(1):79–117

    Article  Google Scholar 

  60. Ragni M, Johnson-Laird P (2018) Reasoning about possibilities: human reasoning violates all normal modal logics. In: proceedings of the 40th annual conference of the Cognitive Science Society

  61. Ragni M, Knauff M (2013) A theory and a computational model of spatial reasoning with preferred mental models. Psychol Rev 120(3):561

    Article  Google Scholar 

  62. Reiter R (1980) A logic for default reasoning. Artif Intell 12:81–132

    Article  MathSciNet  MATH  Google Scholar 

  63. Rips L (2019) Cognitive processes in propositional reasoning. Psychol Rev 1:90

    Google Scholar 

  64. Robinson JA (1979) Logic: form and function. Edinburgh University Press, Edinburgh

    MATH  Google Scholar 

  65. Smullyan RR (2012) First-order logic, vol 43. Springer, Berlin

    MATH  Google Scholar 

  66. Su J, Vargas DV, Sakurai K (2019) One pixel attack for fooling deep neural networks. IEEE Transactions on Evolutionary Computation. https://doi.org/10.1109/TEVC.2019.2890858

  67. Sutcliffe G (2015) The 9th IJCAR automated theorem proving system competition–CASC-J9. AI Communications, (Preprint), pp 1–13

  68. Tessler M, Goodman N (2014) Some arguments are probably valid: syllogistic reasoning as communication. In: Proceedings of the annual meeting of the cognitive science society (vol. 36 No. 36)

  69. Touretzky D (1986) The mathematics of inheritance systems. Morgan Kaufmann, Burlington

    MATH  Google Scholar 

  70. Wittgenstein L (1953) Philosophical investigations. Macmillan, London

    MATH  Google Scholar 

  71. Wos L (1988) Automated reasoning: 33 basic research problems. Prentice-Hall, Upper Saddle River

    MATH  Google Scholar 

  72. Wos L, Pieper GW (2003) Automated reasoning and the discovery of missing and elegant proofs Automated reasoning and the discovery of missing and elegant proofs. Rinton Press, Princeton

    MATH  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Sangeet Khemlani.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Khemlani, S., Johnson-Laird, P.N. Why Machines Don’t (yet) Reason Like People. Künstl Intell 33, 219–228 (2019). https://doi.org/10.1007/s13218-019-00599-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s13218-019-00599-w

Keywords

Navigation