edit IntroductionParaphrase is one of the core concepts of the Meaning-Text Theory. Cf. the definition of meaning as the invariant of paraphrases.
(1) a. John’s comfortable income enables him to travel a lot. b. Thanks to his comfortable income, John is able to travel a lot. c. Since John earns a lot of money, he is in the position to travel frequently. d. With the kind of money he’s making, John can afford frequent travel. e. John has a comfortable income, so he can travel a lot. Manipulating paraphrases (i.e. producing and recognizing them, see below) is an important part of the linguistic competence of a speaker:
The paraphrases of (1) are linguistic paraphrases, i.e. their recognition and production requires only linguistic knowledge (i.e. roughly the knowledge of words and constructions of the language without world knowledge). Extralinguistic paraphrases or Cognitive paraphrases are paraphrases where world knowldge is needed in addition to linguistic knowledge. Linguistic paraphrases share the same linguistic meaning, whereas extralinguistic paraphrases share the same information content, but their linguistic meaning differs. Sentences in (2) are examples of extralinguistic paraphrases. (2) a. John has a comfortable income, so he can travel a lot. b. John earns 100.000 a year so he can travel a lot. In order to establish the informational equivalence between sentences in (2), the Speaker needs some knowledge about the situation of utterance: whether the income John is making can indeed be called comfortable depends on the currency involved, the place where John lives, etc. Extralinguistic paraphrases are quite frequent in speech and deserve attention. However, their modeling requires substantial non-linguistic resources and must be done in an interdisciplinary setting. We are concerned here only with linguistic paraphrases. (Reference for extralinguistic paraphrases ???) edit Recognition and Production of ParaphrasesThe study of (linguistic) paraphrases consists in studying either
In a MTT approach, we are more concerned with production than with recognition of paraphrases. Recognition of paraphrases is studied for instance in Fuchs 19??, Lévrat 19?? Two varieties of paraphrasing can be distinguished: edit Virtual paraphrasingOn the one hand, paraphrasing is needed in order be able to speak fluently. It is easy to understand why this must be so if we think about numerous pitfalls (in the form of restricted lexical cooccurrence, word-order constraints, lexical holes, defective paradigms, etc.), awaiting anyone who is trying to utter a sentence. In order to be able to avoid a dead-end in sentence production, the Speaker vitally needs spare paraphrastic variants. Thus, speaking can be thought of as virtual paraphrasing—a process involving multiple choices between possible paraphrases. edit Reformulative paraphrasingOn the other hand, paraphrasing is needed in order to speak/write better: to make oneself clear, to avoid unwanted repetitions, to produce a particular artistic effect, etc. This conscious paraphrasing, by which the Speaker modifies an existing sentence, can be termed reformulative paraphrasing; it is exploited namely in complex linguistic task such as writing and translation (= interlinguistic paraphrasing; see below, p. 9).
edit Deep-Syntactic parphrasing rules(known from the 1960) edit Semantic paraphrasing rules(Cf. Milicevic 2003, and to appear) edit ExamplesPenelope is certain that Ulyssis will come back. Penelope is sure that Ulyssis will come back. That Ulyssis will return is a sure thing for Penelope.
|