Community Research and Development Information Service - CORDIS


GRAMPLUS Result In Brief

Project ID: 249520
Funded under: FP7-IDEAS-ERC
Country: United Kingdom

Towards a more inclusive grammatical theory

EU-funded researchers have worked to restore grammatical theory to its necessary place in the theory of human language behaviour. The project addressed shortcomings related to the issue of formalism in semantically-based functional and cognitive theories of grammar.
Towards a more inclusive grammatical theory
Current psycholinguistic theories mostly ignore formal linguistic theory. Also, dominant models in computational linguistics are generally low-level finite-state or context-free systems known to be incomplete with respect to the full range of human language. The project GRAMPLUS (Grammar-based robust natural language processing) sought to provide a more restricted theory of constructions, with a formalism that is both efficiently parsable and expressive enough to support semantic interpretation.

The team aimed to extend linguistic theory and its existing computational applications in several different theoretical, computational and applied directions using combinatory categorial grammar (CCG). While CCG has been widely adopted for computational applications, parsers based on CCG are limited by the bottleneck of labelled data.

GRAMPLUS therefore proposed a number of extensions to CCG itself and to related computational applications. The approach enabled a number of important advancements, including successful parser generalisation using a number of semi-supervised methods training on unlabelled text. The team also delivered new techniques for parsing and for automatic semantic parser induction from sentences paired with database queries. The latter have been successfully applied in a psychologically and linguistically plausible model of child language learning based on exposure to meaning-revealing context.

The work has also resulted in improved parsers for under-resourced languages, including Hindi, and combined logical and distributional semantics with state-of-the-art performance in application to question answering. Another achievement was a demonstration that musical harmony can be analysed using the same kind of CCG grammar, with the same parsing algorithm, and statistical model.

Project outcomes have implications for machine learning and tasks requiring semantic interpretation. GRAMPLUS methods and results are also of general interest to linguists, psychologists and other cognitive scientists, as well as those interested in robust practical applications of natural language processing.

Related information


Grammatical theory, human language, linguistic theory, computational linguistics, GRAMPLUS, semantic
Record Number: 188531 / Last updated on: 2016-09-14
Domain: Industrial Technologies