The IMSLex dictionary database is our central lexicon repository. It covers information on inflection, word formation, and valence for several ten thousand German base forms (see details below). From the IMSLex database, we derive specialized lexicon data for various applications in natural language processing, information retrieval, and information extraction. Where necessary, semantic information can be added from the Tübingen GermaNet lexical-semantic dictionary.

Technology

The lexical data itself has been built up semi-automatically with the help of specialized text mining methods from corpus linguistics.

Please refer to the dissertation of Arne Fitschen: Ein Computerlinguistisches Lexikon als komplexes System (ps) (pdf), the IMSLex related publications and the list of IMSLex related projects for further information.

Applications

• creation of full form lexicons

  With the help of the AMOR generator of inflected forms, one can create full form lexicons like the one, which is incorporated in the statistical lexicon for the TreeTagger part-of-speech tagger.
  A sample full form lexicon (adjectives, nouns, and verbs starting with 'p') can be downloaded here.

• on-line morphological and syntactic analysis

  For example, the German ParGram grammar uses IMSLex as its lexical database for deep syntactic analysis.

Current lexicon size

(May 2003)

	inflection stem	derivation stem	composition stem	valence info
adjectives	11,000	23	80	2,000
adverbs	1,000	n/a	n/a	n/a
nouns	22,500	1,000	12,500	10,000
particles	300	n/a	n/a	n/a
proper nouns	10,000
verbs	6,000	350	160	6,000
167 derivation suffixes

For the size and contents of the GermaNet lexical-semantic dictionary, please refer to the GermaNet homepage.