Abstract
Ten native English learners of Japanese, ten intermediate native English learners of Japanese and ten native Japanese speakers of English each were requested to report what they thought was the subject or actor of a series of English NVN word strings, in which case marking and lexical-semantics cues were systematically manipulated. These NVN strings were aurally presented first alone, and subsequently the same strings were presented for the second time together with noncanonical NNV and VNN strings. Similarly, their counterpart Japanese NNV strings were first presented alone, and secondly with noncanonical VNN and NVN strings.
The results revealed that 1) a greater animacy effect (‘animacy noun as a subject’ bias) was detected when the sentence verb was see rather than eat(or each of their Japanese counterparts); 2) English accusative pronouns generally created greater case biases than nominative ones; and 3) native English speakers interpreting Japanese word strings responded differently under the two presentation conditions.