Startseite Linguistik & Semiotik Dependency, Corpora and Cognition
Kapitel
Lizenziert
Nicht lizenziert Erfordert eine Authentifizierung

Dependency, Corpora and Cognition

  • Richard Hudson
Veröffentlichen auch Sie bei De Gruyter Brill

Abstract

Using Google N-grams as a resource, I review the history of dependency analysis in quantitative linguistics, then address a number of general issues: (1) How corpus studies relate to cognition: I present three connections, and argue that a corpus can’t be used as direct evidence either for the language system or for processing difficulty. (2) The nature of syntactic relations: I contrast n-grams with both phrase structure and dependency structure, arguing that dependency structure is most compatible with what we know about cognition. (3) The simplicity of syntactic structure: I argue that, in terms of cognitive reality, syntactic structure is too complex for simple tree diagrams and is formally a network in which words may depend on several other words. (4) The universality of syntactic structure: I argue that languages vary almost without limit, so if we respect cognitive reality we can’t assume a universal set of categories or dependency patterns. On the other hand, some features are shared by some languages, so some cross-language comparison is in fact possible.

Abstract

Using Google N-grams as a resource, I review the history of dependency analysis in quantitative linguistics, then address a number of general issues: (1) How corpus studies relate to cognition: I present three connections, and argue that a corpus can’t be used as direct evidence either for the language system or for processing difficulty. (2) The nature of syntactic relations: I contrast n-grams with both phrase structure and dependency structure, arguing that dependency structure is most compatible with what we know about cognition. (3) The simplicity of syntactic structure: I argue that, in terms of cognitive reality, syntactic structure is too complex for simple tree diagrams and is formally a network in which words may depend on several other words. (4) The universality of syntactic structure: I argue that languages vary almost without limit, so if we respect cognitive reality we can’t assume a universal set of categories or dependency patterns. On the other hand, some features are shared by some languages, so some cross-language comparison is in fact possible.

Heruntergeladen am 1.10.2025 von https://www.degruyterbrill.com/document/doi/10.1515/9783110573565-001/html
Button zum nach oben scrollen