next up previous contents
Next: TypeWh-Type Up: Pronoun Previous: Pronoun

Preliminary Recommendations

Comments

The NERC and the Leech & Wilson schemes propose a multilayered treatment of Pronouns, thus permitting different levels of granularity in annotation. The reader should note that the table above describes the most fine-grained level of linguistic description of the two proposals, where Pronouns are recognised as a category of their own.

At a less granular level, in these two systems Pronouns appear merged together with Determiners. The background reason for this merging in the NERC survey was the necessity of meeting the requirements, on the one hand, of a number of English tagsets (e.g. Penn Treebank, Brown, Lancaster) where, for example, Demonstratives are undistinguished as to their pronominal and determiner functions and receive a unique tag, and, on the other hand, of tagsets which include Articles among the Determiners. Hence, a multilayered approach was adopted, where three different fine-grained levels of linguistic distinctions offer the possibility for each existing practice to be placed at the appropriate level, thus permitting its reusability (see Monachini & Östling (1992b)).

This solution has also been adopted by the EAGLES Corpus group for linguistic annotation (EAGLES, 1996).

A first version of the present document also proposed the same treatment of Pronouns and Determiners. However, after the first cycle of applications, and in particular after concrete testing by MULTEXT, it seemed better to distinguish between different functions and, therefore, to have different categories for Pronouns and Determiners, at least at the lexical level. Lexical descriptions should be independent from applications and should aim at a general description of each language; corpus tags, depending on the capabilities of state-of-art tagging techniques, may underspecify lexical specifications, collapsing many distinctions and presenting broader categories (Calzolari & Monachini, to appear).

Furthermore, following the TEI proposal, it has also been decided to have Articles as a separate category.





next up previous contents
Next: TypeWh-Type Up: Pronoun Previous: Pronoun