Speech shadowing is a dual-task paradigm that could reveal certain features of speech processing, where the shadowing latency between the onsets of heard and reproduced speech has often been a key investigation tool. The present study investigated the shadowing latencies in native Japanese adults who do (AWS) and do not stutter (AWNS), with relevant analysis of speech errors. Fifteen AWS and fourteen AWNS participated in the study. They were required to shadow two meaningful Japanese passages of approximately 1.4 min. Fifty phrase onsets were chosen for measuring shadowing latencies. The resultant latencies were longer than previously reported for English in both groups, which most likely reflects the larger mean syllable numbers per word of Japanese than of English. The AWS group had a significantly shorter latency than the AWNS. Besides, it was significantly more error-prone than the AWNS. The eleven AWS whose latency was more than 500 ms showed a significant negative correlation (trade off) between speech errors and latencies, which was not the case with the AWNS group. Those results imply that only the AWS may have hit the limit of their working memory capacity during shadowing. Thus the shadowing paradigm brought new insights into speech processing and stuttering.