Hiroshi Nagata

Judgments of Sentence Grammatically with Differentiation and Enrichment Strategies

30 students rated the relative grammaticality of isolated sentences twice, with sentences presented repeatedly between the two judgments. Subjects were asked to adopt either a differentiation or an enrichment strategy during a repetition phase. The former strategy involved differentiation by structurally analyzing the sentences, while the latter involved enrichment by supplying situational contexts for the sentences. Analysis showed an interaction between a strategy adopted and the time of judgments such that the subjects instructed to adopt a differentiation strategy tended to use a more stringent criterion on judgments after than before repetition, whereas the subjects asked to adopt an enrichment strategy showed a slight tendency to use a more lenient criterion. These findings indicated that these two types of mental processes are involved in part in the change in criterion for judgments of grammaticality as shown in previous studies.

Repetition Effect in Judgments of Grammaticality of Sentences: Examination with Ungrammatical Sentences:

Nagata (1988) found that repeated presentation of sentences made the criterion more stringent in judgments of grammaticality of both grammatical and ungrammatical sentences. Whether this finding was externally valid was examined here by using sentences of differing levels of ungrammaticality. In Exp. 1, 11 students judged the grammaticality of ungrammatical sentences twice, receiving a repetition treatment between the two judgments. The test items involving two levels of ungrammaticality were drawn from among the work of Kuno from 1973 to 1983. Judgments alter repetition became more stringent for both levels of ungrammaticality. 11 different students in Exp. 2 judged ungrammatical sentences embodying three types of Chomskys 1965 rule violations. Results showed no shift in judgments after repetition for either of the three types of ungrammatical sentences. These findings indicate that Nagatas finding is of limited external validity. Discussion emphasized that susceptibility of sentential materials to the repetition treatment depends on the degree to which ungrammaticality is detectable and explainable.

LONG-TERM EFFECT OF REPETITION ON JUDGMENTS OF GRAMMATICALITY

11 students were asked again to judge the relative grammaticality of the sentences which had been presented repeatedly 4 mo. previously. The repetition treatment involved repeated presentation of the sentences. The findings show that the stringent judgment criterion adopted 4 mo. before is still applied when judging the grammatical sentences. These findings cannot be explained by a satiation theory which does not predict the long-term effect of repetition on judgments of grammaticality.

Effectiveness of word order and grammatical marker as syntactic indicators of semantic relations

The effectiveness of word order and grammatical markers as syntactic indicators of semantic relations was explored using an artificial linguistic system. The results showed: (1) semantic interpretation of novel sentences was possible not only for Ss given fixed word order sentences with or without the markers, but also for Ss given the markers in random word order sentences; (2) Ss given neither indicator showed chance performance; (3) word order tended to be more efficient than grammatical markers as clues to semantic relations. The results suggest that the critical condition for language acquisition is the constant covariation of linguistic structure with semantic contexts. A process approach is discussed in accounting for the findings.

Change in the Modulus of Judgmental Scale: An Inadequate Explanation for the Repetition Effect in Judgments of Grammaticality

Nagata reported that repeated presentation of sentences made the judgment criterion more stringent in judgments of grammaticality of the sentences. This study examined whether the shift in judgments could be explained by change in the modulus of the scale of judgments, i.e., the change produced when the same rating scale was applied for the two successive judgments. 12 students were asked to rate the grammaticality of sentences twice, without the sentences being presented repeatedly. No shift in judgment criterion was found between the two ratings. This finding suggests the repeated presentation of sentences rather than the change in scale modulus produced the observed shift in judgments in the previous study.

Effectiveness of word order and grammatical markers as syntactic indicators of semantic relations in opaque input conditions

Using an artificial linguistic system an attempt was made: (1) to estimate the relative opacity between a meaning-form correspondence which would not interfere with the acquisition of a linguistic system and (2) to determine the relative effectiveness of word order and grammatical markers as syntactic indicators of semantic relations in less transparent input conditions. The results showed that syntax was induced even under 25% opaque conditions and that word order exhibited a slight tendency to be more effective than grammatical markers. Discussion emphasized the minimum amount of structural invariance necessary for a meaning-form mapping under opaque input conditions.

Faulty Japanese Sentences are Judged More Grammatical When Punctuation is Used: Negative Implications for Chomsky's Principle of Full Interpretation

88 adult Japanese speakers judged the grammaticality of isolated simple bitransitive sentences involving an uninterpretable extra argument in addition to three legitimate arguments. The sentences thus violated Chomskys principle of Full Interpretation which prohibits the structure building of a sentence including uninterpretable elements. The primary variable of interest was the presence or absence of punctuation, i.e., commas, which enclosed the extra argument. Findings showed that sentences with punctuation were judged more grammatical than the ones without punctuation, with an average score of judged grammaticality exceeding 3 on a 7-point scale (1= least grammatical; 7 = most grammatical). This score would not be expected if the speakers possess and judge the sentences in conformity with the principle of Full Interpretation.

Recognition of Concrete and abstract Pictures as a Function of Poststimulus Interval

This study examined picture recognition of concrete common objects and abstract random shapes under two presentation conditions: The durations of the unfilled intervals (off-time duration) that followed the presentation of the pictures were varied either randomly or in blocks. The major findings were: recognition of the concrete pictures was better than recognition of random shapes, the off-time duration did not influence recognition of either type of picture under the random presentation condition, and under the block presentation condition, recognition of the concrete pictures increased as a function of off-time duration, whereas the effect of off-time duration was not found for the abstract shapes. The results were interpreted as suggesting that a longer off-time duration is necessary for abstract shapes so that stimulus differentiation, verbal encoding, visual analogizing, and rehearsing may be utilized in processing.

OFF-LINE REFLEXIVE RESOLUTION IN PARSING JAPANESE LOGOPHORIC SENTENCES: AN EXTENSION

This study examined the external validity of Nagatas 1995 finding on off-line resolution of a Japanese reflexive. jibun, for logophoric and nonlogophoric sentences. The subject marker, -ga, was attached to both a matrix-subject and a subordinate-subject as contrasted with the previous study in which the matrix-subject received a topic marker, -wa. Analysis again showed no effect of sentence type, thus making invalid Kurodas (1973) thesis and Chomskys (1981) binding principle as applied to parsing logophoric and nonlogophoric sentences.

Effectiveness of word order over grammatical markers as a syntactic indicator of semantic relations in an opaque partial description situation

Using an artificial language system, an attempt was made (1) to examine the effect of relative completeness in describing the reference situations on acquisition of the system and (2) to determine the effectiveness of word order and grammatical markers as syntactic indicators of semantic relations in an opaque partial description situation. The results showed that syntax was induced faster in the mixed description condition, where partial three-word sentences and two-word phrases were given mixed with complete five-word sentences, than in the partial description condition, where only three-word sentences and two-word phrases were given. It was further found that word order was more effective than grammatical markers as a syntactic indicator. It is suggested that language acquisition under partial description conditions is possible provided a sufficient number of experiencing exemplars and/or pragmatic cues as to meaning-form correspondence are given.