2018 Sat Session C 1100

Saturday, November 3, 2018 | Session C, Terrace Lounge | 11am

On a developmental delay in the L1 acquisition of the Japanese nominative object
T. Sano, H. Shimada, Y. Fujiwara

In the Japanese nominative object construction (1), an NP with Theme role receives nominative Case, which is canonically assigned to the subject position; however, it is an object. Only subjects can be antecedents of zibun,[1] and thus in (1) the nominative object buta ‘pig’ cannot be the antecedent of zibun.[2] This paper examines the development of the objecthood of the Japanese nominative object.

Background: Wexler’s Universal Phase Requirement (UPR) explains developmental delays in non-actional passives and raisings.[3][4]   Under this account, vP is a phase for children until age 7, and operations crossing  it are illicit. Hence, such movements in the non-actional passive (2a) and raising are illicit until age 7.  In the nominative object construction (2b), Case checking takes place from T to the object, crossing vP.[5] Hence, UPR predicts that the nominative object is illicit for children until 7; however, in natural production, surface “nominative objects” are widely observed even at age 2.[6] Are such early “nominative objects” structurally objects or subjects? We addressed this question with a comprehension experiment (3) using the zibun test.

Experiment: First, to ensure that children chose the antecedent of zibun based on the subjecthood/objecthood of the antecedent, we tested a “full unaccusative sentence” (4), which is known to develop early despite involving an A-chain,[7] [8] and an accusative object sentence (5). In (4), the Theme NP buta can be the antecedent of zibun since it is the subject, but not in (5) since it is an object. As shown in Table 1, 6-year-olds correctly accepted (4) for the matching situation and rejected (5) for the mismatching situation more than 80% of the time.

As the target condition, we examined whether 6-year-olds disallow the Theme nominative object in (1) as antecedent of zibun. The acceptance rate of 48% for (1) (cf. 97.5% for adults) shows that 6-year-olds are not adult-like in generating the nominative-marked NP in (1) syntactically in the object position.

Discussion: Thus, Japanese 6-year-olds cannot reliably treat nominative objects (cf. accusative objects) as objects, as predicted by UPR. Given this and the many examples of early “nominative objects” on the surface level, we speculate that Japanese children base-generate the nominative phrase in (1) in vP Spec, a subject position, [9] as in (6). If so, then even in 6-year-olds’ grammar T can Case-check the nominative phrase in (6) without violating UPR, since vP Spec is “outside” the vP phase.[10] Thus, there are two subject phrases in (6) for 6-year-olds, explaining the acceptance rate of (1) near 50%: As there are two possible antecedents of zibun for 6-year-olds, the acceptance rate is around chance level. Lastly, we would like to add that our observation can be explained by UPR but not by other developmental hypotheses on the delay of non-actional passives and/or raising, such as the Universal Freezing Principle,[11] Canonical Alignment Hypothesis,[12] and Argument Intervention Hypothesis,[13] because they make predictions based only on movement, not on Case checking of the nominative object as discussed in this paper.