Paraphrase and parallel treebank for the comparison of French and Chinese syntax

Abstract

This paper proposes to study the contrastive syntax of French and Chinese through the lens of syntactic mismatches, and by making use of parallel treebanks. A syntactic mismatch is the non-similarity between the syntactic structures of one linguistic unit and its translation. Syntactic mismatches are formalized using the notion of paraphrase from the Meaning-Text Theory, which allows for capturing mismatches at different levels of the linguistic description (e.g. Semantic, Deep-Syntactic, and Surface-Syntactic). In this paper, we report in details on the types of paraphrases found in the seed corpus used, demonstrating that the Deep-Syntactic paraphrases constitute the best starting point for our study. Then, we show how, starting from the seed corpus, we semi-automatically constructed a multi-layer parallel treebank with the alignment and annotation of paraphrases.

Igor Mel'čuk. 2021. Ten studies in Dependency Syntax. Berlin: De Gruyter Mouton. Pp. 444. US $160.99 (hardcover).

“…Moreover, there have already been several studies that make comparisons between the SUD and UD (Osborne and Gerdes 2019, Yan and Liu 2019, 2022). Also, apart from SUD, another approach based on MTT can be found in Mille et al (2012) and in Poiret et al (2021). In spite of having written a large number of publications on MTT, Mel'čuk himself seems not to have developed large-scale annotated treebanks, as have those Praguian researchers who follow the Functional Generative Description (FGD) framework.…”