Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Perceptual categorization in the visual domain: evidence from three sign languages
Poster Session B, Friday, September 12, 4:30 - 6:00 pm, Field House
This poster is part of the Sandbox Series.
Alexandra Navarrete-González1, Lyke Esselink2, Floris Roelofsen2, Brendan Costello1,3; 1Basque Center on Cognition, Brain and Language, 2University of Amsterdam, 3Ikerbasque, Basque Foundation for Science, Bilbao, Spain
Category is a fundamental feature of language architecture. Perceptual categorization (PC) has been widely studied for spoken languages, but only a few studies have been conducted in sign languages (see Baker et al., 2005; Best et al., 2010; Emmorey et al., 2003; Mertz, 2024). Previous studies have shown mixed results and were limited by types of stimuli used (e.g. still images, disembodied hands or unnaturally synthesized signs). In this study, we use realistic avatars signing four handshape contrasts to examine PC in users of Catalan Sign Language, Spanish Sign Language and Sign Language of the Netherlands. To explore whether differences in categorization depend on the type of phonological feature, contrasts involved changes in selected fingers or in finger position (curving and spreading of the fingers). The contrasts included existing and non-existing feature values (i.e. handshapes) in the phonological repertoire of the sign languages to examine the role of linguistic knowledge in the perception of categories. Finally, we test signers and non-signers to examine how much PC in sign language is driven by familiarity with this linguistic modality. We present pseudosigns as name signs of members of a family. Stimuli were recorded by a signer using motion capture and then synthesized on a metahuman avatar. For each con-trast, we created a set of eight stimuli equally distributed between the two target signs. Par-ticipants perform an ABX (categorization) task and a same-different (discrimination) task in two orders (ABX-BAX, AB-BA), and provide confidence ratings for the responses on the ABX task. For each contrast, the ABX task has 64 trials (8 contrast variants x 4 repeti-tions x 2 orders), and the same-different task 96 (6 contrast variants x 4 repetitions x 2 conditions (same-different) x 2 orders). If the sign contrasts are perceived categorically, the categorization task should show the typical sharp between-category shift on the contin-uum, and discrimination should be greater at the category boundary. The confidence rat-ings will reflect tolerance to (phonetic) variability in the feature value. Based on previous findings and pilot data, we expect results to depend on contrast type. For changes in select-ed fingers, we expect to find evidence of PC, while changes in finger position may be a matter of degree rather than category. Category may depend on anatomical features that serve as reference points: change in spread is contrastive between +/- contact between fin-gers. We will examine group and individual differences to probe whether experience with a specific sign language alters one’s categorical boundaries. By including non-signers, we will be able to test whether any PC effects are due to (sign) language experience or to more general visual-domain classification of the human body. Data collection is under way, and we will present the results at the conference. The find-ings will reveal commonalities and differences in the auditory and visual linguistic do-mains. This work contributes to understanding better how PC of phonological contrasts is processed in our minds and paves the way to investigating how linguistic categories are represented in the brain across different modalities.
Topic Areas: Signed Language and Gesture, Phonology