The use of picture and sound mnemonics is a common method in teaching hiragana as a second language (L2) or a foreign language (FL). However, there is only one published study that investigated their effects on recalls by L2 learners of Japanese in Japan, and it found their effectiveness not on short-term recalls but on long-term recalls (Quackenbush, Nakajo, Nagatomo, & Tawada, 1989). Would the results be the same for FL learners of Japanese in the US? Would there be any performance difference between those who have prior experience in learning non-Roman scripts (non-Roman group) and those who do not (Roman group)?
The present paper reports a study which attempted to answer these questions. In this study, 66 beginning learners of Japanese at an American university participated in four computer-generated tutorial sessions of 40 hiragana. In each session, one of the four sets of ten hiragana was introduced in one of the four teaching methods: (a) picture and sound mnemonics (P+S), (b) picture cues (P), (c) sound cues (S), and (d) flash cards (F). Each session consisted of (a) a learning phase, (b) a review phase, and (c) a self-test phase. Each session was followed immediately by an oral interview, and by recall tests two to five days later.
The results from the self-test showed positive effects of the P+S method and S method only for the Roman group. The results from the long-term recall test indicated no overall effects of mnemonics and no group effects. These results are discussed in relation to the Quackenbush et al. (1989) study, and transfer of L1 script recognition strategies (Chikamatsu, 1996; Koda, 1989; Mori, 1998). In addition, an observation is made as to why the P+S method was effective, and future studies are suggested.