2018 Friday Poster 6623

Friday, November 2, 2018 | Poster Session I, Metcalf Small | 3pm

Understanding language-specific anaphora in children
M. Ma, Y. Oshima-Takane, Y. Kayama

Previous research has reported that Japanese-speaking children understand the anaphoric function of null forms earlier than pronouns (Okumura et al., 2016). Some researchers suggested that Japanese-speaking children understand null forms earlier because they are exposed to null forms more often than pronouns in reference to given information (Guerriero et al., 2006; Okumura et al., 2016). To examine the effect of input on the acquisition of language-specific anaphora, we investigated whether English-speaking children understand the anaphoric function of pronouns but not null forms when they are in the null argument stage (i.e., null argument sentences are often produced), because in English, pronouns are typically used in reference to given information, whereas null forms are used only in limited contexts (Goldberg, 2001).

Thirty-two English-speaking children in the null argument stage (mean: 33 months; range: 31-38 months) were assigned to either the pronominal or null object condition. Six familiar transitive verbs (carry, cut, push, drink, throw, kick) were used in a comprehension task and each verb item consisted of three phases (Context, Dialogue, and Test phase; Figure 1). In the Context phase, children saw a still image of one animal and two objects on one screen (e.g., a dog, a carrot, and a potato) while listening to the audio describing the target object: Miss Dog found a carrot.

The non-target object (e.g., potato) was not linguistically mentioned. In the Dialogue phase, an animation of two puppets talking about the target object was shown, e.g., Puppet A: What is she doing now? Puppet B: Look, she is cutting it/Ø! Puppet A: Really? Where? In the Test phase, to determine whether children can correctly identify the referents of the pronominal or null object sentences based on the previous contextual information, they were presented with two test animations (e.g., target animation: the dog cutting the carrot; non-target animation: the dog cutting the potato) side by side. The children in the pronominal condition heard Find the picture of Miss Dog cutting it. (Test1) and Where is the picture of Miss Dog cutting it? (Test2). Those in the null object condition heard Find the picture of Miss Dog cutting Ø. (Test1) and Where is the picture of Miss Dog cutting Ø? (Test2). Children were expected to look at target animations significantly longer than non-target animations if they understand the anaphoric use of pronouns or null forms. The proportion scores were calculated for three different periods (prior to, during, and after the test sentence) for each item by dividing the time looking at the target animation by the sum of the time looking at the target and the non-target animations.

The results showed that only the mean proportion score for the time period after the Test2 sentence in the pronominal object condition was significantly higher than chance (.05), t(14) = 3.524, p = .004; Figure 2). The present study demonstrated that English-speaking children in the null argument stage interpret the pronominal objects correctly but not null objects, suggesting that the acquisition of language- specific anaphora depends on the input.