| This dissertation describes an approach to the automatic morphological analysis of French verbs. The function of the analysis routine is to restore a verb to the infinitive by determining the correct suffix and analyzing the tense and mood of this suffix. The design and implementation of the computer program and the lexicon is described from a linguistic as well as computational point of view.;The verbs used as the test data were extracted from 1980 and 1981 issues of Paris Match. The program, consisting of 1,251 PL/I statements, was debugged and tested on an IBM 370 Model 148 running under OS/VS1/VM at Georgetown University. The logic was developed using principles of structured programming, including top down design and modular programming techniques. The appendix includes listings of the source program and lexicon, examples of the input data, and results achieved by the program.;The program has been designed for easy modification by the user and possible inclusion in a larger automatic analysis system. Some of the practical applications of a system of this nature include machine translation and artificial intelligence.;The program generates two reports. The first, a listing of the verbs that were matched with an irregular verb stored in memory or a regular verb contained in the lexicon; and the second, a listing of the verbs that were not matched but for which an apparently successful attempt was made at restoring them to the infinitive. These infinitives would normally be manually verified before being entered in the lexicon along with other syntactic and/or semantic codes. For both reports, the output consists of the verb and its infinitive; and optionally, the tense, mood, stem, and suffix of the verb. |
