CORPUS OF SEMIOTIC ANALYSIS OF SIGN LANGUAGE.

The Brazilian Sign language (LIBRAS) is of visual-spatial manner with lexical items that denote objects, actions, abstract concepts and all the other possibilities of human language, necessary for Intellectual development and human traits. Some of the Sign Language's utterances are iconic and/or metaphorical signs. Such categorization is useful in scaffolding the learning of the language. But it also precludes a more complex meaning signification mechanism for those signs whose meanings only indirectly are derived from their apparent iconicity. Additionally, there is a lack of adequate corpus in Libras to be used by Natural Language Processing applications. A richer semiotic approach, along with an information system, is proposed to consider sign meaning attribution, knowledge production and computational treatment. A corpus was created to serve as basis for a richer Discourse Analysis of LIBRAS. A Case Study of the use of an information system to create corpus for future research as per the framework is presented with classroom utterances from students of graduate course in linguistics/LIBRAS. [ABSTRACT FROM AUTHOR]