Lingware

Revision as of 20:03, 13 December 2010

The UNL System comprises three different types of language resources: lexica, grammars and corpora. These resources are being provided mainly through the UNLarium, a linguist-friendly dabatase management system to provide dictionary entries and grammar rules to the several language modules participating in the UNL Programme.

Lexica

In the UNL framework, there are two major types of lexical databases: monolingual and bilingual.

The monolingual resources are the following:

• in UNL
  • The UNL Dictionary, or simply UNL^dic, which is a flat list of UWs along with non-relational semantic features
  • The UNL Ontology, which is a hierarchy of UWs
  • The UNL Knowledge Base, or simply UNL^kb, which is a network with systematic relations between UWs
  • The UNL Example Base, or simply UNL^eb, which is a network with episodic relations between UWs
• in natural language
  • The NL dictionary, or simply NL^dic, which is a list of natural language headwords
  • The NL Example Base, or simply NL^eb, which is a list of colocations and

The bilingual resources are the following

• The UNL-NL Dictionary, or simply UNL-NL^dic, which is list of systematic lexical mappings between UNL and a given natural language
• The UNL-NL Memory, or simply UM, which is a list of past (episodical) mappings between UNL and a given natural language

Grammars

Sets of rules to UNL-ize natural language documents and to NL-ize UNL documents, to be provided according to the Grammar Specs.

Corpora

Documents written in UNL, to be provided according to the UNL Document Structure

• UNL Corpora