Jorion, Paul (1990) Summary in English of PRINCIPES DES SYSTEMES INTELLIGENTS. [Book Chapter]

Full text available as:

[img] HTML


This book, published in a series called “Cognitive Science”, aims -- as suggested in the title -- at defining general properties of « intelligent systems ». The question of the intelligence of a computer system running a piece of software can be considerably simplified if no consideration is given to the concept of a « subject » displaying intelligent behavior. In everyday life we content ourselves with simply assuming that our human interlocutors are inhabited by a « subject »: we never demand any proof of the actual existence of such a subject. Searle’s « Chinese room » thought experiment shows a human subject being fluent in Chinese but being unaware of such capacity for being solely aware of manipulating skillfully sets of symbols. For all practical purpose all persons dealing with the prisoner of the Chinese room would be justified to regard themselves as dealing with a fluent speaker of Chinese. In the circumstance the « subjective » feeling of the subject that he understands or not Chinese is irrelevant to his actual fluency. The same would apply to a computer system: as long as the sentences it produces are indistinguishable from those that a human being would generate, it would justifiably be regarded as intelligent. The question reduces then to that of designing a function and data structure that generates algorithmically specific sentences by traveling through the word-space of a particular language’s lexicon. The first constraint to be observed is that of grammatical acceptability of the clauses generated. The second is that of the plausibility of these sentences within a general context of human cultures (common sense view). The third constraint is that of topicality. The fourth constraint is that of non-contradiction between consecutive clauses. Little by little a picture emerges of a network of memory traces capable of growing in an organized manner associated with an affect dynamics capable of generating from these memory traces, clauses displaying emergent logical features.

Item Type:Book Chapter
Keywords:Aristotle, Freud, Lacan, truth, learning, self-organization, logic, belief, connectionism, emotional dynamics, affect, semantics, syntax, Douglas Hofstadter, taxonomy, intelligence, intention, lexicon, memory, pragmatics, neuron, PDP, relevance, ambiguity, plausibility, folk psychology, reasoning, reminiscence, knowledge representation, semantic networks, meaning, memory storage, syllogism, synonymy, Wittgenstein, Merleau-Ponty
Subjects:Computer Science > Artificial Intelligence
Computer Science > Language
ID Code:483
Deposited By: Jorion, Paul
Deposited On:29 Jun 1998
Last Modified:11 Mar 2011 08:53


Repository Staff Only: item control page