Paraphrases can be described at various levels of utterance representation in the Meaning-Text Model (MTM). We analyze different kinds of paraphrases that are found in a parallel corpus of German translations of a Russian novel and assign them to the different representation levels.
Doing so, some rules of the deep syntactic paraphrasing system of MTM are modified. For paraphrases that need a deeper analysis to be recognized as such a division into two groups is proposed:
paraphrasings that can be described at the semantic level and those that can be described at the transition between the semantic and the deep syntactic level. Some difficulties in the formal recognition of paraphrases will be discussed and the frequency of different types of paraphrase rules will be evaluated.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados