Crossed Semantic Analysis of Literary Texts with DeSeRT

Abstract : DeSeRT (acronym for Détection Sémantique de Reformulations et de Topiques) has been designed to identify and compare rewritings, paraphrases or borrowings. It is based on the already developed idea according to which, even if the reformulations cannot be reduced to paraphrases, they conserve the meaning of original texts by reusing words of which meaning is close. As a consequence, the detection of co-occurrences of a few semantically equivalent lemmas in small blocks of texts is sufficient to capture the meaning and, therefore to identify the reformulations of the same ideas. Based on these theoretical principles, the DeSeRT system is intended to spot and track reformulations and allusive references. This contribution describes a few results obtained on the corpus of the “Haine du theater” (the Haine of Theater) project, to detect arguments exchanged during the 17th and 18th about the theater controversies.
Complete list of metadatas

https://hal.sorbonne-universite.fr/hal-01483367
Contributor : Jean-Gabriel Ganascia <>
Submitted on : Sunday, March 5, 2017 - 4:53:57 PM
Last modification on : Thursday, March 21, 2019 - 2:43:08 PM

Identifiers

  • HAL Id : hal-01483367, version 1

Citation

Jean-Gabriel Ganascia, Chiara Mainardi. Crossed Semantic Analysis of Literary Texts with DeSeRT. Digital Humanities, Jul 2016, Krakow, Poland. ⟨hal-01483367⟩

Share

Metrics

Record views

179