Identifying empty subjects by modality information: the case of the Japanese sentence-final particles -yo and -ne

The present study conducted four experiments to investigate how modality information provided through the sentence-final particles -yo and -ne were utilized in identifying an empty subject by native Japanese speakers. Experiment 1 conducted a whole-sentence anomaly decision task, finding that base sentences without -yo and -ne attached were processed more quickly than sentences with either -yo or -ne and that sentences with -yo were processed more quickly than the same sentences with -ne. A delay in processing sentences with -ne was created by the ambiguity of an empty subject identified by -ne as either ‘I’ or ‘you’. In Experiment 2, the auxiliary verb -ou ‘let us’ was added to the base sentence before -yo and -ne, providing a cue to identify the empty subject as ‘we’. Although the base sentences were processed more quickly than those containing the particles -yo and -ne, no other difference resulted from the attachment of these particles. To eliminate the possibility of orthographic-length effects, Experiment 3 compared base sentences with -ou, -ou-yo, and -ou-ne, finding no difference among them (i.e., no ortho- graphic-length effects). Experiment 4 was conducted to further eliminate the possible involvement of discourse-level computation by utilizing base sentences with overt subjects, past tense verbs, and the auxiliary verb -rasii ‘appear to’. Once subjects of sentences were clearly shown, there was no difference among base sentences and those with either -yo or -ne attached (i.e., no discourse-level computation effects). Thus the present study proved that the modality information inherent in the particles -yo and -ne was used for identifying empty subjects.