Maurice Gross

Gross worked on automatic translation at the École Polytechnique without prior training in linguistics. This led in 1961 to a scholarship to the Massachusetts Institute of Technology (MIT), where he met Noam Chomsky and became acquainted with Generative grammar.[4] After returning to France, he worked as a computer scientist at the Centre National de la Recherche Scientifique (CNRS). In 1964 he went a second time to the United States, this time to the University of Pennsylvania, where he worked with Zellig S. Harris, whom he thereafter regarded as his linguistic foster father. He received his research PhD at the Sorbonne in 1967 with his dissertation L'Analyse formelle comparée des complétives en français et en anglais ("Comparative formal analysis of complements in French and English"). He went on as a lecturer at the University of Aix-en-Provence, where he worked with Jean Stefanini. In 1969 he received his teaching doctorate (habilitation) at the University of Paris with defense of his thesis Lexique des constructions complétives, published under the title Méthodes en syntaxe (Paris : Hermann, 1975). He was appointed professor at the new University of Vincennes (which was later Paris VIII), then at the University of Paris VII.

In 1968, he founded the Laboratoire d'Automatique Documentaire et Linguistique (LADL, the CNRS), and in 1977 the journal Lingvisticae Investigationes.

At the age of 67, while completing an essay explicating a fundamental principle in the work of his mentor,[5] Maurice Gross succumbed to bone cancer.

Gross's work, and that of the LADL, gives priority to the principles of methodological rigor, respect for data, empirical observation, comprehensive coverage of a language, and reproducibility of experiments.[6][7] A systematic description of simple sentences of French[8] yielded a dictionary based on the syntax[1][7] identifying properties of words salient for parsing and grammatical tagging,[1] and providing a reasoned and detailed classification of most of the elements of the French language.[6][7] Indeed, before generative grammar adopted the Projection Principle or the Theta criterion, Gross had undertaken the systematic investigation of the interdependence of lexical entries and grammatical rules.[9] It was for this reason that his methods and results were given the name lexicon-grammar. His students have verified this working hypothesis in many typologically diverse languages, including not only Romance languages and German, but also Modern Greek, Korean, Arabic,[9] Malagasy, and other languages.

The work of the LADL was greatly enhanced by the use of computers, beginning in the 1980s.[2] One result was a morpho-syntactic "electronic dictionary" of French.[8] In parallel, taking finite automata as the competence model of language,[1] Gross developed the concept of local grammar.[10] Local grammars, consisting of finite automata coupled with morpho-syntactic dictionaries, support automatic text analysis[6][10] by the closed-source Intex software developed by Max Silberztein and, after Maurice Gross's life by the open source Unitex conceived by the Gaspard-Monge Computer Science Laboratory (LIGM) and the open source NooJ software developed at the Université de Franche-Comté.

Concurrently, Gross was working on problems that he considered fundamental to linguistics, although they had long been neglected in the field, such as lexical ambiguity, idioms and collocations, and "support verb" constructions.[10] In 1976 he discovered the "double scan" property of certain support verb constructions,[10] which systematically identifies idioms.[1][10] Gross's computer-assisted research on large amounts of linguistic material led to a picture of language as an instrument that is freely manipulated yet highly constrained idiomatically,[10] a result that is consistent with the distinction between the Idiom Principle and the Open Choice Principle found by corpus linguist John McHardy Sinclair (1933 -2007).[11] Gross described the organization of language as a lexicon-grammar,[12] and argued that any grammar must fail if its formalization fails to take into account its dependence on the lexicon.[13] He demonstrated that to fully describe a language one must collect a huge quantity of tagged word combinations.

The facts registered in the dictionaries and grammars resulting from such collection are useful for natural language processing and in particular for deep linguistic processing.

Gross's students include Alain Guillet, Christian Leclère, Gilles Fauconnier, Morris Salkoff, Joëlle Gardes, Bertrand du Castel, Annibale Elia, Jean-Pierre Sueur, Laurence Danlos, Hong Chai-song, Cheng Ting-au, Claude Muller, Eric Laporte, Denis Maurel, Max Silberztein, Tita Kyriacopoulou, Elisabete Ranchhod, Anne Abeillé, Mehryar Mohri, Emmanuel Roche, Nam Jee-sun, Jean Senellart, and Cédrick Fairon.[14]

A complete bibliography of the writings of Maurice Gross is available. Below is a brief selection.

• Chevalier, Jean-Claude. 2001. "Maurice Gross, un grand linguiste", Le Monde, 12 décembre 2001. Archive.
• Delamar, Michel. 2001. "Maurice Gross", 12 décembre 2001. Cache.
• Dougherty, Ray. 2001. Maurice Gross Memorial Letter. Archive.
• Ibrahim, Amr Helmy. 2002. "Maurice Gross (1934-2001). À la mémoire de Maurice Gross". Hermès 34.
• Lamiroy, Béatrice, [www.cairn.info/revue-travaux-de-linguistique-2003-1-page-145.htm "Maurice Gross (1934-2001)", Travaux de linguistique 1/2003 (no46), p. 145-158]. Cache.
• Perrin, Dominique. 2002. "In Memoriam Maurice Gross." Cache.