This paper reports on the results of two naming experiments that investigated lexical neighborhood effects in Japanese word recognition. Two naming experiments (normal and noisy conditions) were conducted with Japanese adult listeners. Each participant responded to 700 words that had varying neighborhood density. The noun lexicon was based on the NTT Japanese psycholinguistic databases. The RT results of Experiment 1 (normal condition) demonstrated that (1) neighborhood density was a facilitatory effect (fast namers only) and (2) word frequency effect was a facilitatory effect(slow namers only). The RT results of Experiment 2 (noisy condition) demonstrated that (1) neighborhood density was an inhibitory effect (slow namers only) and (2) word frequency effect was a facilitatory eflect (fast namers only). Word identification results of Experiment 2 further replicated the results reported in Luce and Pisoni (1998). These results support the claim by Luce & Pisoni that words in the lexicon are grouped in part in terms of sound similarity. We further discuss the time course of word recognition based on our naming experiments.