| Title |
Using Descriptive Generalisations in the Acquisition of Lexical Data forWord Formation |
| Authors |
Ulrich Heid (Institut f¸r Maschinelle Sprachverarbeitung Universität Stuttgart Azenbergstrafle 12, 70174 Stuttgart, Germany) Bettina Säuberlich (Institut f¸r Maschinelle Sprachverarbeitung Universität Stuttgart Azenbergstrafle 12, 70174 Stuttgart, Germany) Arne Fitschen (Institut f¸r Maschinelle Sprachverarbeitung Universität Stuttgart Azenbergstrafle 12, 70174 Stuttgart, Germany) |
| Session |
WO2: Acquisition Of Lexical Information |
| Abstract |
This paper presents a method for acquiring data for a word formation analyser. There are several approaches to the analysis of complex words in German. As all of them have theoretical and/or practical drawbacks, we opt for a different approach: Instead of using linking elements, we make use of three different stem types, simplex, derivational, and compounding stems. Candidates for these can be generated automatically using knowledge about linguistic processes in German word formation. Based on the analysis of only a few phenomena we have gathered about 14.000 stems in a short time frame, all of them manually checked. As a result, certain wrong analyses can be avoided and ambiguities can be solved. |
| Keywords |
Lexical data |
| Full Paper |