Abstract

In this paper, the incidental clauses delimited by commas are inquired from a computational linguistic perspective. Some theoretical aspects based on grammatical criteria, and a discussion about its nature and definition are proposed. For this objective, we consider (i) the possibility that incidental clause has (or not) a syntactic function, (ii) the possibility that the incidental clauses interrupts (or not) the regular order of the sentence where it is inserted, and (iii) the different levels where the incidental clauses could be situated. From these issues, a classification and a definition of incidental clauses delimited by commas are proposed, and a method for automatic detection of these expressions using the Smorph and MPS softwares are described; this method got adequate accuracy and coverage percentages. Finally the contribution of this paper for the grammatical studies and the computational linguistics are presented, and new topics for future works are proposed