This study aims to explore if it is possible to form separate categories of aspirated voiceless stops in a second language, distinct from the equivalent categories in the native language, for native speakers of a language with an intermediate degree of aspiration, and if such category formation is eased by long-term exposure to another language in which aspirated voiceless stops exist. Two groups of adult native Japanese speakers who had lived in Sweden for a long and short time respectively were recorded when reading a list of sentences containing word-initial, utterance-medial /p t k/ in Japanese and English. Both groups produced higher VOT values for the English stops than for the Japanese stops. The results were significant for /t/ and /k/ and for the long-term residents' /p/, but not for the short-term residents' /p/, presumably because of a low number of tokens. The results are nevertheless interpreted as evident of the possibility of phonetic category formation even though there is only a small difference from the corresponding category in the native language. Since both groups had established new phonetic categories, degree of exposure to Swedish is interpreted as likely not to be a relevant factor.