Computational extraction of lexico-grammatical information for generation of Swedish intonation

By using a minimal amount of syntactic information in combination with information on lexical word class and morphological and lexico-semantic coreferential (identity of sense) relations, it is possible to generate an appropriate prosodic structure for Swedish texts. The structure of the algorithms involved in the prosodic preprocessing are presented. These include a referent tracker, a word-class tagger, a complex-word identifier, a clause boundary identifier, and a prosodic constituent parser.