Poster #110 - Four-year-olds' On-Line Use of Referential Context: The Effect of Speaker Conventionality

• In Event: 3-159 - Poster Session 12
  In Poster Session: PS 12 Section - Language, Communication

Integrative Statement

Previous research has demonstrated that adult listeners resolve ambiguous reference in real time by relying on expectations about conversational norms. One such conversational norms suggests that rational speakers provide sufficient but not redundant information in utterances (Grice, 1975). For example, in a situation where a modified noun phrase (e.g., “Look at the tall glass”) is used to distinguish between two objects of the same kind (e.g., two glasses), adults can locate the target even before hearing the noun, even though the adjective also describes another object in the display (e.g. a tall pitcher). This phenomenon is referred to as the contrast effect. However, this effect appears to be dependent on the expectation that the speaker is behaving typically; Grodner and Sedivy (2011) found that by telling adults that the speaker had “an impairment that caused language and social problems”, the contrast effect disappeared.

In the current study, we used an eye-tracking method to examine whether 4-year-olds share these same expectations about the contrastive nature of adjectives and whether these expecations are modulated by speaker conventionality. Children were assigned to one of two conditions. In the conventional condition, a speaker was introduced by a confederate (i.e., “This is Monica”). During the labeling trials, this speaker labeled objects in a conventional manner (e.g., “This is a banana”). In the unconventional condition, the speaker was introduced as someone who used language unconventionally (i.e., “Monica says things in a very strange way”), and used unusual descriptions during the labeling trials (e.g., “This is a long, yellow fruit you can peel that monkeys like to eat”). On the critical trials, children in both conditions were presented with displays containing a target (e.g., a big cookie), a contrast object (e.g., a small cookie), a competitor (e.g., a big clock) and a distractor (e.g., a small butterfly). If children make the contrastive inference and use the information about the speaker, children in the conventional condition would be more likely to look at the member of the pair that matched the adjectival phrase than those in the unconventional condition.

Figure 1 shows the probability of fixations to the target and the competitor during the critical instruction. Children’s eye-gaze were analyzed during during critical trials using a mixed-model logistic regression. The analyses examined the proportion of looks to the target in pre-adjective, adjective, and noun periods separately. There was no significant predictor of the proportion of looks to the target in all periods (ps > .07). The current data shows that children’s attention to the target did not vary as a function of speaker conventionality. We are currently conducting a follow-up study to replicate this effect.

Authors

• Narae Ju, University of Calgary
  Presenting Author

• Justine Thacker, University of Calgary
  Non-Presenting Author

• Craig G Chambers, University of Toronto
  Non-Presenting Author

• Susan Graham, University of Calgary
  Non-Presenting Author