Collectivity, Distributivity, and the Interpretation of Plural Numerical Expressions in Child and Adult Language

• Kristen Syrett, Ph.D. and Rutgers, The State University of New Jersey, Linguistics, New Brunswick, United States
• Julien Musolino : Rutgers, The State University of New Jersey, Psychology, Piscataway, United States

Intro

• Sentences containing plural numerical expressions (e.g., two boys) can give rise to two interpretations (collective and distributive), arising from the fact that their representation admits of a part-whole structure
• designed to explore children’s understanding of this distinction and its implications for the acquisition of linguistic expressions with number words.
• preschoolers access both interpretations, indicating that they have the requisite linguistic and conceptual machinery to generate the corresponding representations.
• hift their interpretation in response to structural and lexical manipulations.
• unlike adults, they are drawn to the distributive interpretation, and are not yet aware of the lexical semantics of each and together, which should favor one or another interpretation.
• Here, we take a different approach, and use numerically quantified expressions to study how children acquire a fundamental semantic property shared by a range of plurality-denoting expressions

Findings

• Our findings demonstrate that the ability to generate collective and distributive interpretations of sentences such as (1) is part of the semantic repertoire of children as young as three (Experiment 1). However, we also uncover intriguing differences in the preferences preschoolers and adults have for resolving the collective/distributive ambiguity: whereas adults strongly prefer the collective interpretation, preschoolers prefer the distributive one (Experiment 2).
• Following analyses by Link (1983, 1987) and others more recently, we will assume that these sentences are truly ambiguous, and not merely underspecified, and that the source of the ambiguity in our target sentences is the VP predicate. Here, we adopt a default semantics approach in order to illustrate how two different interpretations may be generated.
• When the predicate in our example sentence is applied to the individuals, the derived reading is the distributive one
• When the predicate is applied to the group, however, a collective reading is derived, and the extension is an atomic joint ‘car pushing’ event in which the boys collectively push the car.
• Beginning with the latter, Musolino (2009) was primarily concerned with the range of readings arising from the interaction of two numerically quantified expressions in so-called relational plural sentences such as (3).

Judgment Task with Ambiguous Sentences

• The results demonstrate that both children and adults were able to access both the collective and distributive interpretations of the target sentences. While there was a significant main effect of age (p= .02), there was no main effect of context (p=.75) and no interaction between age group and context (p=.26).
• This difference stems from the fact that while four-year-olds were near ceiling in their acceptance of the sentences in the distributive context, adults’ acceptance rates were slightly suppressed.

Ambiguous Sentences That Yield Either Interpretation.

• As the results demonstrate, adults overwhelmingly preferred the collective version of the event
• In this experiment, we found that while adults robustly prefer the collective context as a match for the ambiguous target sentences, children display a slight preference in the opposite direction, leaning towards preference for the distributive context.
• structural manipulation of passivization will lead participants to prefer the collective context.
• As predicted, adults consistently accepted the passive test sentences in the collective context, but largely rejected them in the distributive context. Most children also followed this pattern, although the difference between acceptances in the two contexts was not as striking for children as it was for the adults.
• whether participants can recruit lexical semantic information provided by individual words to disambiguate the target sentence and assign either a collective or distributive interpretation, depending on the lexical item.
• As predicted, adults were guided by the presence of the additional lexical item in their interpretation of these sentences, accepting the test sentences with each in the distributive context, but rejecting them in the collective context
• In place of the ambiguous sentences, children heard sentences with a post-verbal together (
• Interestingly, despite children’s acceptance of the together sentences in both contexts of the judgment task of Experiment 4, children appeared to be aware of the collectivizing force of together in the current preference task.