Could both be right? Children’s developing understanding of subjectivity in language

• In Event: 1-042 - Epistemic reasoning in pragmatic interpretation

Integrative Statement

To understand many phrases, we must consider the perspective of the speaker. When two people disagree about whether a sweater “is pretty," for example, not only is 'pretty' interpreted relative to them as speakers, but it seems the disagreement is 'faultless': neither is wrong (Barker, 2013). This may be true for SUBJECTIVE adjectives (e.g., pretty, funny), but also RELATIVE adjectives (e.g., tall, expensive) (Syrett, 2006), where speakers may have different standards due to distinct past experiences, as well as the inherent vagueness of the predicate. We may disagree about whether 60_F is cold both because we are accustomed to different temperatures, and because what counts as “cold" is ambiguous. This is in contrast to disagreements over ABSOLUTE adjectives (e.g., spotted, striped), where it seems only one speaker may be right. Here, we study children's understanding of how different adjectives participate in faultless disagreement as a window onto their understanding of subjectivity in language, and diversity of belief more broadly. In two experiments, we ask whether adult and child judgments of faultless disagreement are modulated by their knowledge of speakers' experiences. Given suggestions in the developmental literature that children begin life as 'naive realists,' (Holubar & Markman, 2013), we predicted children would initially make judgments egocentrically, endorsing the speaker who asserted the child's own belief.

Experiment 1: Adults saw puppets exposed to either DISTINCT or IDENTICAL distributions of novel objects. The puppets then disagreed about an additional, intermediate object of the same kind, using adjectives from the three categories (ABSOLUTE, RELATIVE, & SUBJECTIVE). There were significant effects of condition and adjective type on faultless disagreement judgments: adults permitted it more in the DISTINCT condition, and for relative and subjective adjectives, rather than absolute. An interaction between condition and adjective type indicated that the manipulation of speakers' distributions by condition affected absolute and relative, but not subjective, adjectives.

Experiment 2: Children 4-9.5 participated in the DISTINCT condition. 4-7-year-olds rarely permitted faultless disagreement, siding instead with the speaker who accorded with their own perceptions. When asked to explain the disagreements, however, children showed increasing sensitivity to how the meanings of different adjective-noun phrases are constructed, citing speakers' distinct experiences to explain disagreements over “tall," and personal tastes to explain disagreements over “pretty." Subjective disagreements proved the most difficult for children. Children did not reliably make adult-like judgments until 8-9 years of age, consistent with claims that around this age children achieve an 'interpretive theory of mind,' understanding that individuals with equated experiences may still interpret the world differently (Carpendale et al., 1996).

Our results suggest that adults interpret the meanings of many phrases relative to their speakers, and are sensitive to multiple sources of subjectivity in word meanings. These sources vary in their accessibility to children, and the process by which children come to master them is remarkably protracted. Results from a newly-collected follow-up indicate young children reject divergent subjective assertions, even when they are explicitly marked as “opinions."

Authors

• Ruthe Foushee, UC Berkeley
  Presenting Author

• Mahesh Srinivasan, UC Berkeley
  Non-Presenting Author