Hostname: page-component-84b7d79bbc-2l2gl Total loading time: 0 Render date: 2024-07-28T11:14:26.540Z Has data issue: false hasContentIssue false

Intelligence et intentionalité. Critique de l'argument de Searle contre l'intelligence artificielle

Published online by Cambridge University Press:  13 April 2010

Maryvonne Longeart
Affiliation:
Université du Québec à Hull

Extract

Dans son célèbre article de 1950 «Computing Machinery and Intelligence», Alan Turing suggère qu'un critère d'intelligence pour une machine serait qu'elle puisse se faire passer pour un interlocuteur humain. Ce test, appelé «Test de Turing», a été critiqué par John Searle, qui a remis en cause sa pertinence en s'appuyant sur une expérience mentale connue sous le nom de l'argument de la chambre chinoise (Chinese Room Argument). L'expérience tend à montrer que l'intelligence artificielle au sens propre du terme est impossible, parce que les symboles manipulés par les programmes de prétendue «intelligence artificielle» (IA) sont dépourvus de signification. L'intentionalité propre à certains êtres vivants peut seule faire naître le sens. L'imitation d'un comportement intentionnel ne le peut pas. Le test de Turing serait done insatisfaisant, parce qu'il ne permettrait pas de distinguer les systèmes intentionnels de systèmes qui ne le sont pas, et que par conséquent il ne permettrait pas de distinguer l'intelligence authentique (intentionnelle) de la simulation de l'intelligence (non intentionnelle).

Type
Articles
Copyright
Copyright © Canadian Philosophical Association 1991

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

Notes

1 Turing, Alan, «Computing Machinery and Intelligence», Mind, vol. 59 (1950), p. 433461.CrossRefGoogle Scholar

2 Searle, John, «Minds, Brains, and Programs», Behavioral and Brain Sciences, vol. 3 (1980), p. 424.CrossRefGoogle Scholar

4 Turing, A., «Computing Machinery and Intelligence», p. 442. Ce résultat a été publié par Turing dans son article «On Computable Numbers, with an Application to the Entscheidungsproblem», Proceedings of the London Mathematical Society, series 2 (1936), p. 230265.Google Scholar

5 J. Searle, «Minds, Brains, and Programs», p. 424.

6 A. Turing, «Computing Machinery and Intelligence», p. 436.

7 Dennett, D., «The Milk of Human Intentionality» [commentaire de J. Searle, «Minds, Brains, and Programs»], Behavioral and Brain Sciences, vol. 3 (1980), p. 429.Google Scholar

8 Goldman, Alvin J., Epistemology and Cognition, Cambridge, MA, Harvard University Press, 1986, chap. 6.Google Scholar

9 Quine, W. V. O., Word and Object, Cambridge, MA, The MIT Press, 1960, p. 46.Google Scholar

10 Ibid., p. 225.

11 Ceci ne préjuge pas de l'indépendance de la fonction (produire une démarche intelligente) par rapport à la structure physique susceptible de la réaliser (l'architecture requise pour implanter un algorithme intelligent). Patricia P. Churchland fait très justement remarquer que la distinction structure / fonction est trop simpliste. «It may be that if we had a complete cognitive neurobiology, we would find that to build a computer with the same capacities as the human brain we had to use as structural elements things that behaved very like neurons.» (Neurophilosophy: Toward a Unified Science of the Mind / Brain, Cambridge, MA, The MIT Press, 1986.)Google Scholar Qu'il soit impossible de spécifier complètement un algorithme d'un point de vue fonctionnel, indépendamment d'une implantation physique particulière, n'est pas douteux. Cette contrainte est la pierre d'achoppement de tout fonctionnalisme radical à la Fodor.

12 Minsky, M., «Decentralized Minds» [commentaire de J. Searle, «Minds, Brains, and Programs»], Behavioral and Brain Sciences, vol. 3 (1980), p. 439.CrossRefGoogle Scholar

13 J. Searle, «Minds, Brains, and Programs», p. 420.

14 Lind, R., «The Priority of Attention: Intentionality for Automata», The Monist, vol. 69 (1986), p. 609619.CrossRefGoogle Scholar

15 Cf. W. V. O. Quine, Word and Object, p. 221.

16 Dennett, D., The Intentional Stance, Cambridge, MA, The MIT Press, 1987, p. 29Google Scholar: «All there is to being a true believer is being a system whose behavior is reliably predictable via the intentional strategy.»

17 P. Churchland, Neurophilosophy.

18 Winston, P. H., Intelligence Artificielle, Paris, Inter Éditions, 1988, p. 252.Google Scholar

19 Rapaport, W. J., «Philosophy, Artificial Intelligence and the Chinese Room Argument — Using the Theory of Abstract Data Types to Contradict John Searle's Argument», Abacus, vol. 3, no 4 (1986), p. 717.Google Scholar

20 Thaggard, P., «The Emergence of Meaning: How to Escape Searle's Chinese Room», Behaviorism, vol. 14, no 2.Google Scholar

21 À ce sujet, voir T. Winograd et F. Flores, Understanding Computers and Cognition, Ablex, 1986.

22 Kripke, S., Wittgenstein on Rules and Private Language, Oxford, Blackwell, 1980Google Scholar. Voir aussi Bouveresse, J., La force de la règie. Wittgenstein et l'invention de la nécessité, Paris, Édition de Minuit, 1987.Google Scholar

23 L. Wittgenstein, Investigations philosophiques, § 202.

24 Laurière, J.-L., Intelligence artificielle. Résolution de problème par l'homme et la machine, Paris, Eyrolles, 1987, p. 273Google Scholar. Voir aussi «Intelligence artificielle et jeux d'échecs», La Recherche (1983).

25 Pour une analyse approfondie du problème de l'intériorité chez Wittgenstein, voir J. Bouveresse, Le mythe de l'intériorité. Expérience, signification et langage privé chez Wittgenstein, Paris, Éditions de Minuit, 1976. Bouveresse cite en page de garde cette remarque de Wittgenstein: «The difficulty in philosophy is to say no more than we know.»

26 Smith, B., «Prologue to “Reflection and Semantics in a Procedural Language”», dans R. Brachman et H. Levesque, dir., Readings in Knowledge Representation, Kaufmann, 1985.Google Scholar

27 Voir en particulier H. Dreyfus, «From Micro-Words to Knowledge Representation: AI at an Impasse», dans R. Brachman et H. Levesque, dir. Readings in Knowledge Representation. Sur le connexionnisme et les limites du modèle représentationnel, voir aussi P. Churchland, Neurophilosophy.

28 Pour un aperçu de l'état de la recherche sur les modèles TPD, voir Artificial Intelligence Review, vol. 3, no 1 (1989), qui est entièrement consacré à cette approche.