Previous studies have focused on the frequency and types of recurrent accentual mistakes made by Japanese learners. However, in the practical teaching session, there remains a debate as to whether it is crucial to give instruction on accentual mistakes. Particularly, errors that mistake the unaccented pattern for the final- accented pattern have been frequently overlooked. To explore this issue, the present study investigates native Japanese speakers' evaluations of six types of accentual mistakes
on a five-point rating scale of naturalness using five hundred and forty sentences produced by non-native Japanese learners. The results revealed that, contrary to expectations, although mistaking the initial-accented pattern for the unaccented and
the final-accented pattern led to significantly lower ratings, mistaking the unaccented pattern for the final-accented pattern also lowered the naturalness significantly. Overall, these results indicate the necessity of giving explicit instruction on the error of
mistaking the unaccented pattern for the final-accented pattern.

Researches on learners’ perception of Japanese pitch accents have focused mainly
on the accuracy rate during accent kernel marking. However, whether listeners use this
perceived pitch changing to identify words and, if they do so, whether language experience influences this process has been barely discussed. To bridge this gap, this study uses dichotic listening tasks to investigate the accuracy rate and reaction time during word identification processing in perception of Japanese pitch accents by native Japanese speakers (JNs) and two groups of native Mandarin Chinese speakers, namely those with high proficiency in Japanese (CJs) and those with no prior experience of Japanese (CNs).
The results reveal that JNs and CJs process Japanese pitch accents more stably than
CNs. To be more exact, neither the pitch accent patterns nor the stimuli difference among items influences the accuracy rate or the reaction time of JNs or CJs; however, stimuli difference among items changes both the accuracy and the reaction time of CNs. To summarize, the findings suggest that during the word identification processing in perception of Japanese pitch accents, JNs and CJs are reliant on the conceptualized pitch categories of Japanese accent patterns, while CNs are reliant on the conceptualized pitch categories from Mandarin tones and affected further by the phonetic features of stimuli.

First language experience affects the processing of Japanese pitch accents; however, the influence of second language learning experience is unclear. The present study utilized identification experiments to compare the difference among native Japanese speakers (JN), native Chinese speakers with no experience of Japanese (CN), and native Chinese speakers with high Japanese proficiency (CJ). The results showed that identification responses to high-low vs. low-high accent differed between CN and CJ, indicating that Japanese learning experience alters listeners' perception of Japanese pitch accent. CN are more sensitive to stimuli with gradual decreases in pitch, and tend to identify them as high-low accent. On the other hand, CJ are not sensitive to such stimuli, and tend to identify them as low-high accent.

A previous study suggests that native Chinese speakers may experience pronunciation difficulty in learning Japanese when the utterances include successive high or low moras. This kind of pronunciation difficulty has been examined in long compound words; however, no attempt has been made to ascertain whether it occurs in phrases. To verify this hypothesis, the current study performed an experiment using sixty phrases including six different disposition types of word accent patterns. The results indicated that native Chinese speakers do experience pronunciation difficulty in phrases with successive high moras. However, difficulties in phrases with successive low moras are influenced by the accent pattern of the first word.