Abstract
To investigate the effect of harmonization between words' meaning and their expression style (typography and prosody), we conducted a series of multi-modality priming experiments using word-fragment completion task (WFC). At encoding, words were presented both visually as the letters and auditorily as speech, and their typography and prosody were manipulated to harmonize or not to harmonize with the words' meaning. The words' script (hiragana/kanji) at encoding was also manipulated. At test, all WFC items were written using hiragana. The results showed no effect of harmonization between words' meaning and typography on implicit memory, which did not corroborate with a previous study. We conducted another experiment, manipulating typography but keeping all speech sounds neutral at encoding. Results showed an effect of harmonization. These results suggest that when speech information is manipulated to harmonize with the words' meaning, the processing of speech is given the priority over that of visual character information.
