Full text
| |
| Author(s): |
Antonielle Cantarelli Martins
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Paulo. |
| Institution: | Universidade de São Paulo (USP). Instituto de Psicologia (IP/SBD) |
| Defense date: | 2017-06-09 |
| Examining board members: |
Fernando Cesar Capovilla;
Felipe Venâncio Barbosa;
Lineu Norió Kohatsu;
Maria Isabel da Silva Leme;
Janice Gonçalves Temoteo Marques
|
| Advisor: | Fernando Cesar Capovilla |
| Abstract | |
This doctoral dissertation springs from the Research and Development Program on Brazilian Sign Language (Libras) Dictionaries conducted by Professor Capovilla and students at the University of Sao Paulo over the last 25 years. The dissertation pertains to Libras lexicology, lexicography, and metalexicography. The dissertation centers on the lexicographic research that the author has conducted as coauthor of the Dic-Brasil: Libras Dictionary (Capovilla, Raphael, Temoteo, & Martins, 2017a, 2017b, 20917c). It describes the strategies used for studying Libras lexicon, and compares lexicographic strategies used in six classic dictionaries, three of which on Libras, and the other three on American Sign Language. The dissertation analyzes the role of sign iconicity in sign dictionaries. It compares different ways in which iconicity is dealt with in six classic sign language dictionaries. In order to increase the pragmatic efficacy with which dictionaries represent signs, the present dissertation analyzes the role played by both illustration and description of both sign form and sign meaning, as well as by descriptions of sign iconicity, etymology and morphology. The dissertation describes a study on sign iconicity that analyzes the relationship between sign admissibility and sign guessability. Signs are iconic when their meaning is admissible and can be guessed. Paradoxically, while only 10% of sign meaning can be guessed, the remaining 90% of signs that cannot be guessed are considered admissible, once the observer is informed about what they mean. In the study, 70 naive subjects (Groups 1 and 2) examined 201 signs (Sets A and B). Group 1 subjects rated the admissibility of Set A signs (using a Likert scale: -2, -1, +1, +2) and guessed the meaning of Set B signs. Group 2 did the opposite. Results showed that: Subjects rated 28 signs as inadmissible (average rating -1), 77 signs as admissible (average rating +1), and 96 signs as neutral (-.99 average rating +.99). Only 24 of 201 signs were effectively guessed (average rating +1), all of which had been regarded as being admissible by the other group subjects. No sign that had been regarded as inadmissible ended up being guessed. Therefore, admissibility seemed necessary for guessability and positively correlated with it. Yet it seemed insufficient for guessability. The data suggests that, apparently, sign admissibility can only predict sign guessability when admissibility surpasses a given threshold of strong admissibility, which is located somewhere in between 1.5 and 1.75 point (1,50 average rating 1,75) in the Likert scale (AU) | |
| FAPESP's process: | 14/02420-0 - Brazilian Sign Language lexicography in the South and Federal District |
| Grantee: | Antonielle Cantarelli Martins |
| Support Opportunities: | Scholarships in Brazil - Doctorate |