Grammaire I Description Transformationnelle D’un Sous-Ensemble Du Francais

Antonio A.m. QueridoCOLING 1969

TL;DR

This paper presents "GRAMMAIRE-I," a transformational grammar model for a subset of French. It introduces a novel system for handling anaphoric relations and a unified approach to determiners, aiming to provide a more comprehensive and integrated framework for French syntax compared to existing models like Chomsky's.

What Problem Does This Paper Address?

The paper addresses the challenge of creating a comprehensive and internally consistent transformational grammar for French. Specifically, it aims to:

  • Differentiate from Chomsky's model: While building on Chomsky's work, GRAMMAIRE-I seeks to address perceived limitations or offer alternative approaches, particularly in areas like the absence of a lexical component and subcategorization rules.
  • Develop a unified system for determiners: Traditional grammars often treat articles, demonstratives, and interrogatives separately. GRAMMAIRE-I proposes a single system to account for these, aiming for greater elegance and explanatory power.
  • Formalize anaphoric relations: The paper tackles the complex issue of how pronouns and other referring expressions relate to their antecedents within and across sentences, proposing a unified schema for different types of anaphora.
  • Integrate phrase types: It seeks to categorize and handle various types of embedded clauses (relative, infinitive, interrogative) within a coherent framework.

Methods

GRAMMAIRE-I employs a transformational grammar approach, building upon Chomsky's framework but with key modifications:

  • No lexical component or subcategorization rules: Unlike Chomsky's model, GRAMMAIRE-I omits these, relying solely on phrase structure rules to generate deep structure "phrase-markers."
  • Feature-based system: It utilizes binary features (e.g., +hum, -pers) to mark syntactic and semantic properties of elementary syntagmatic categories.
  • Unified determiner system: The paper introduces a single system for determiners, encompassing articles, demonstratives, and interrogatives, marked by features like DEF (definite), ANAPH (anaphoric), and DEM (demonstrative).
  • Anaphora resolution: A novel, unified schema is proposed for handling reflexive, relative, and inter-phrasal anaphora, marking the relation with an anaphoric adjective.
  • Phrase structure rules and transformations: The grammar is defined by a set of phrase structure rules (categorization rules) and transformational rules (cyclic and post-cyclic) that operate on these structures.

Datasets

Not discussed in the paper. The paper focuses on theoretical linguistic modeling rather than empirical data analysis or corpus-based training.

Key Insights and Ideas

  • Unified Determiner System: The proposal to treat articles, demonstratives, and interrogatives under a single DET category with specific features offers a more parsimonious and potentially more explanatory account of their behavior.
  • Anaphora as a Unified Phenomenon: Treating reflexive, relative, and inter-phrasal anaphora under a single schema, marked by anaphoric adjectives, simplifies the analysis and highlights common underlying principles.
  • Rejection of Lexical Component: The decision to omit a lexical component and subcategorization rules suggests a focus on deriving grammatical structure primarily from phrase structure rules and transformations, potentially leading to a more abstract and generative model.
  • Integration of Phrase Types: The categorization of embedded clauses into five groups and their systematic analysis within the grammar provides a structured approach to complex sentence structures.

Key Results

The paper presents a theoretical framework for analyzing a subset of French grammar. It details:

  • A set of phrase structure rules for generating deep structures.
  • A comprehensive list of cyclic and post-cyclic transformational rules designed to derive surface structures.
  • A system of features for classifying verbs, nouns, and adjectives.
  • A detailed account of the proposed determiner system and anaphoric relations.

Quantitative results are not presented as this is a theoretical linguistic paper.

Limitations and Future Work

  • Limited Scope: The grammar is explicitly stated to cover only a "sous-ensemble" (subset) of French, implying that a full grammar would require significant expansion.
  • Absence of Lexical Component: While a deliberate choice, the lack of a lexical component might limit the model's ability to account for the full range of lexical variations and semantic nuances.
  • Manual Lexical Insertion: The paper notes that lexical insertion is done manually, suggesting that a fully automated or generative lexicon is not yet developed.
  • Future Work: The authors mention that future phases of their research will involve elaborating on morpho-phonological components, introducing more complex quantifiers, and potentially considering indirect interrogatives as underlying forms. They also plan to introduce anaphoric adjectives in a later phase.

Summary updated

BibTeX
@inproceedings{querido1969grammaire,
  author = {Antonio A.M. Querido},
  title = {{Grammaire I Description Transformationnelle D’un Sous-Ensemble Du Francais}},
  booktitle = {International Conference on Computational Linguistics},
  year = {1969},
}
Figure