Skip to Main content Skip to Navigation
Journal articles

Modéliser ce qui résiste à la modélisation : De la sémantique à la sémiotique

Aurélien Bénel 1 
1 Tech-CICO - TECHnologies pour la Coopération, l’Interaction et les COnnaissances dans les collectifs
ICD - Institut Charles Delaunay, CNRS - Centre National de la Recherche Scientifique : FRE2732
Abstract : To deal with Artificial intelligence applied to Digital humanities, this article boldly focuses its state of the art on the 1970s. We discover that modelling archaeological artifacts at that time with “analysis languages” and “domains” competes with current projects based on RDF and OWL, even with subtleties that only the best ones achieve today. But projects from the 1970s are especially interesting because of the debates they inspired in the archaeologists community, debates with such a theoretical depth that they still could have an impact, more than 45 years later. Among the critics of that time, the most insightful and constructive is P. Bruneau (an archaeologist). According to him, the objects of Human Sciences, because they are already meaningful (unlike the objects of Nature), must be described with semiotic (rather than semantic) methods. Coping with context cannot be reduced to an elusive statement, neither to an additional model: it requires the systematic refusal of broad scope models. Models should take into account that among all the possible distinctive features of an object, only a few are pertinent: they are highlighted by the differences with the neighbouring objects of the “technical universe”. Several current works in Knowledge engineering, consciously or not, fit into those perspectives at least partially. As illustrated with our own software and experiments, taking into account this semiotic approach offers promising prospects for the instrumentation of the daily practice of researchers in Humanities as well as for scientific mediation. But a greater prospect could be in the bunch of design open issues that this approach raises: mundane issues at first sight, but related indeed with a better comprehension of sense and sense making. This could lead us back to the roots of Artificial intelligence and Cognitive sciences, when their main aim was less to replace human intelligence than to better understand it.
Complete list of metadata

Cited literature [31 references]  Display  Hide  Download
Contributor : Jean-Baptiste VU VAN Connect in order to contact the contributor
Submitted on : Wednesday, September 9, 2020 - 2:43:32 PM
Last modification on : Wednesday, August 31, 2022 - 6:55:58 PM
Long-term archiving on: : Thursday, December 3, 2020 - 1:08:27 AM


Publisher files allowed on an open archive


Distributed under a Creative Commons Attribution 4.0 International License



Aurélien Bénel. Modéliser ce qui résiste à la modélisation : De la sémantique à la sémiotique. Revue Ouverte d'Intelligence Artificielle, Association pour la diffusion de la recherche francophone en intelligence artificielle, 2020, Intelligence Artificielle et Humanités Numériques, 1 (1), pp.71-88. ⟨10.5802/roia.4⟩. ⟨hal-02313535⟩



Record views


Files downloads