Procedure and Subjects. In order to see the correlation between the acquisition of number morphology and that of the number agreement under consideration, two experiments were conducted. The first experiment investigated whether children could interpret number morphology on nominals in an adult-like manner. This experiment was intended to gather base-line information before the second experiment. In accordance with children’s responses in this experiment, they were divided into two groups. This experiment had three test conditions, which differed in the morphological property of the object; the indefinite singular condition, the bare plural condition, and the definite plural condition. Children were shown two pictures sequentially and asked a simple yes-no question. For example, while a picture was presented on a screen of a lap-top computer, an instruction about the picture was given as in (16a). Then, the second picture showed up, and a question about the picture was given as in (16b). The question included any one of the following: the indefinite singular, bare plural, or definite plural object. In all the second pictures, the person was holding one of the two items given (for example, in (16b), Suzi was holding one of the two dolls). Thus, the expected response should be “yes” in the indefinite singular condition, but “no” in the bare plural and definite plural conditions.5 Each condition had two items, and the test items amounted to six in total.6
Procedure and Subjects. The second experiment was intended to investigate when children acquire fully adult-like knowledge about the agreement between possessive pronouns and head nouns. Children were given a picture in which two people were doing some action and were asked a yes-no question about the picture. This experiment had eight test items and they were divided into two types: four were target items which were intended to induce no-responses from children and four were control items which were biased toward yes-responses. The target items had singular marking on potential antecedents, verbs, and possessive pronouns (the singular condition) and the control items had plural marking on them (the plural condition). All eight questions had body-part nouns as head nouns, but they were divided into two sub-types by the semantic properties (or pragmatic presuppositions) associated with the head nouns: four of them had body-part nouns which human beings have one of (the non-pair condition) and the remaining four had those which human beings have two of (the pair condition). The examples of our target and control items are given in (17). As shown in (17), all the entities depicted in the pictures were doing the actions denoted by the predicates. In the singular conditions ((17a) and (17b)), the subjects, verbs and possessive pronouns had singular morphology. In order to make singular marking felicitous, the distributive quantifier each was used. Only the head nouns had plural morphology, which is not permissible in adult grammar. In the plural conditions ((17c) and (17d)), all the subjects, verbs, possessive pronouns, and head nouns had plural morphology, which is perfectly fine in adult grammar. In the non-pair condition ((17a) and (17c)), body-part nouns which presuppose one-to-one correspondence between body-parts and individuals were used. In the pair condition ((17b) correspondence were used. and (17d)), those which presuppose two-to-one
