How do Young Children Think the Blind See?

• In Event: 3-236 - Linguistic and Cognitive Representations of Information Access

Integrative Statement

Two-year-olds already know much about vision (Flavell, 2004), understand how seeing leads to knowing (Pillow, 1989), and use visual verbs to talk about perceptual experiences (Bretherton & Beeghly, 1982). How do children learn the mapping between visual verbs and their conceptual referents — the acts of looking and seeing? Locke famously argued that knowing a word’s meaning requires first-person experience. Therefore, someone born blind would not understand words such as “look” or “see”. However, data show that blind people do acquire a rich visual lexicon, suggesting that first-person experience is not necessary (Landau & Gleitman, 1985; Bedny et al., in prep.). Here we reverse Locke’s question: how do sighted children come to understand words such as look and see when they are applied to blind agents?
In Experiment 1, we asked what children know about the canonical modalities of perception verbs. We asked 4-, 6-, and 9-year-olds (N=12 per group) and adults (N=51) which body part they would use to “see”, “look”, “hear”, etc. (25 verbs, Table 1) — the eyes, ears, hands, mouth, or nose? We also asked about two fictional characters, one sighted and one blind, each described with key details (e.g. “her eyes don’t work, she is blind”, repeated every 5 verbs). Four-year-olds responded with the canonical body part (e.g. eyes for visual verbs) for both the sighted and blind characters, whereas 6-, 9-year-olds and adults either said that the blind character “cannot” [visual verb] or extended the meaning of the verbs to non-canonical body parts (e.g. hands) (Figure 1A). Thus, 4-year-olds appear to associate eyes with visual verbs, even for blind (non-canonical) agents.
Experiment 2 tested whether children’s restricted understanding of modality generalizes to other verbs – namely, motion verbs. New participants (N=12 per group) were tested as in Experiment 1, with the addition of a fictional wheelchair-bound character and leg-related motion verbs (Table 1). We replicated Experiment 1 findings and extended them to motion verbs (Figure 1B): 4-year-olds answered that all characters — including the wheelchair-bound one — would use the leg to [motion verb]. Thus, young children’s representation of verb meanings seems strongly tied to their canonical modality.
In Experiment 3, we asked whether 4-year-olds, despite their rigid use of visual verbs, distinguish between look (active, direct the eyes) and see (stative, achieve visual perception) (Viberg, 1983). Participants from Experiment 2 were shown pictures of the blind and wheelchair-bound characters pointing their eyes towards an object, and were asked, “Is she looking at X? Does she see X?” Four-year-olds answered “yes” for both characters, still asserting that the blind could look and see. However, like older children, they were more likely to say yes for looking than seeing, thus preserving the active/stative distinction.
In sum, we found that children’s verb meanings are initially rigidly tied to the canonical modality (i.e. seeing is done with the eyes), with adult-like generalization developing slowly between 6- and 9-years-of-age. However, children seem to understand the active/stative distinction quite early, suggesting that these aspects of perception verbs might be governed by different learning mechanisms.

Authors

• Giulia V. Elli, Johns Hopkins University
  Presenting Author

• Marina Bedny, Johns Hopkins University
  Non-Presenting Author

• Barbara Landau, Johns Hopkins University
  Non-Presenting Author