Poster B34, Thursday, August 16, 3:05 – 4:50 pm, Room 2000AB
Not all “visual knowledge” are created equal: Blind individuals’ judgments about animal appearance.
Judy Sein Kim1, Giulia Elli1, Erin Brush1, Marina Bedny1;1Johns Hopkins University
How does the way we acquire information affect what we know and how we represent it? Are learning that tigers have stripes by seeing a picture or by reading about them in a book equivalent? We explore this question by comparing knowledge of animal appearance in individuals who are congenitally blind and those who are sighted. Previous studies show that blind individuals acquire rich information about vision. For example, blind children understand the difference between ‘look’ and ‘see’ (Landau & Gleitman, 1985). Blind adults know the relationships among colors (e.g., that red is similar to orange but not blue), as well as the meanings of visual verbs (e.g., glance vs. stare; Shepard & Cooper, 1992; Bedny et al., unpublished data). Are there types of information that are nevertheless uniquely or preferentially acquired through vision? In Experiment 1, 20 congenitally blind and 20 sighted control participants sorted cards with animal names (Braille or print) based on their shape, texture, or color. Blind participants’ sorting based on shape and texture were correlated with the sighted (shape: rho=0.83, texture: rho=0.77), although not as correlated as sighted participants were with each other (comparing blind-to-sighted vs. sighted-to-sighted correlations, shape: t(37)=1.71, p=0.09; texture: t(36)=4.16, p=0.0002). In contrast, blind participants’ sorting based on color was not correlated with that of the sighted (rho=0.34, p=0.25), although sighted participants were correlated with each other (B-S vs. S-S: t(37)=11.66, p<0.0001). For color, only blind participants’ sorting was correlated with taxonomic similarity (blind: rho=0.47, sighted: rho=0.08), suggesting that blind participants use knowledge about taxonomy to infer animal colors. One possible explanation for blind individuals’ lack of agreement with the sighted on color sorting is that animal colors are difficult to describe using language. To test whether sighted individuals can verbalize distinctions in shape, texture, and color, a second group of sighted participants (Experiment 2, n=14) were asked to provide descriptions of animals for each dimension (e.g., “Describe the shape of the animal so that someone who knows nothing about animals can pick it out from a book of photographs”). Contrary to predictions, sighted subjects were able to describe best the dimension about which blind participants know least: color. Sighted participants’ descriptions closely matched the sighted groups’ sorting from Experiment 1 for color (rho=0.65) and texture (rho=0.8). By contrast, shape descriptions deviated from sighted participants’ shape sorting (rho=0.4). Thus, verbalizability does not fully account for which aspects of animal appearance blind individuals acquire. Arbitrary object-color mappings may be unlikely to be learned, even though they are linguistically available. These results suggest that not all “visual knowledge” are created equal. Sighted but not blind individuals may learn animal colors because they are behaviorally relevant for referent identification or because they are easier to learn through the visual modality. By contrast, blind individuals may preferentially acquire knowledge about visual appearance that is structured (e.g. into a similarity space or schema), and behaviorally or inferentially relevant (e.g. look vs. see and stare vs. peek).
Topic Area: Meaning: Lexical Semantics