University Paris-Sorbonne

 (Centre for Theoretical and Applied Linguistics)

CELTA Laboratory


The CELTA, Centre for Theoretical and Applied Linguistics (EA 3553) was created in 2002 after the gathering of several teams studying the linguistics of modern languages. There are 23 researchers and professors from the Paris-Sorbonne University and other French universities working for it, and the CELTA maintains close relationships with research institutes from other countries of the European Union – old members and new members from Central Europe included – and in the world (Japan, Russia, USA).

Seven project leader professors oversee and direct the works of about twenty young researchers following a PhD curriculum at the Paris-Sorbonne University. The languages or groups of languages studied are English, Dutch, Slavic languages (Croatian, Serbian, Bosnian, Montenegrin, Polish and Russian), and Japanese.

The CELTA has an interdisciplinary approach to research, encouraging a cooperation between most specialized studies on various languages and computer-assisted formalisation which is achieved by a specific adaptation of computer tools to research in linguistics. Computer sciences provide a useful help in the conception and formalisation of linguistic theories in different areas by using discovery tools (exploration methods conceived in artificial intelligence) on revisable databases using parameters defined by the linguists themselves, who do not need to have basic skills in computer sciences. These databases are revisable as research progress goes along. Computer sciences are on the one hand heuristic tools and on the other hand a way to store data in a reusable and transmittable form.

However, the computer-assisted approaches are inseparable from descriptive approaches : scholarship and linguistic skills of specialists of numerous languages are crucial to submit to criticism old and new a priori knowledge in grammar and produce linguistic theories the internal coherence and adequacy of which may be verified and validated by formal methods.

The CELTA has been accepted in March 2008 as a member of the pan-European initiative CLARIN (Common Language Resources and Technology Infrastructure) thanks to the CASK project and the computer-assisted semantic research software SEMANA (accessible from the CELTA website after registering). By using this software, we are initiating a research on the grammatical metalanguage of European languages. Indeed, as all of the languages studied by the CELTA team have a rich tradition of grammatical and linguistic thinking, the formalisation enterprises must incorporate this heritage while taking care to systematise theoretical and terminological knowledge.

 The study of the form of languages – from phonology to syntax from both synchronic and diachronic points of view – aims at revealing the cognitive processes underlying the construction of utterances and discourse. The purpose being to construct a linguistic theory which takes its roots in the European analytic school of thoughts. The formal theories of Associative Semantics (AS) and Meta-Informative centring (MIC) elaborated by CELTA researchers coherently modify and unify both kinds of theoretical frames currently used in linguistics : the Predicate-Argument Structures (PAS) theories (conceived as models of semantic relations as expressed in natural languages, but often through a certain confusion between semantics proper and syntax) and the informative structure (IS) of utterances theories (which actually concern meta-information or the presentation of the informative content as it is conveyed by the linguistic forms, but which have the defect to omit an important part of the meta-informative means of languages, usually considered as strictly pertaining to syntax).

            The AS & MIC theories are grounded in ontology, however clearly distinguishing between an universal ontology and the particular semantics of each language. That is why these theories may serve as bases of comparison and contrast allowing to identify informative and meta-informative content which are similar and opposed, expressed by various morpho-syntactic forms in different languages.