Kim, Seon-Jung. 1999. The Notion of Adjacency and Korean Phonological Phenomena. Studies in Modern Grammar 18, 91-109. In this paper I discuss adjacency and its role in Korean Phonology, focusing on umlaut and some related processes. In the previous analyses, the relationship between palatalisation and umlaut has not been treated. However, blocking of umlaut before a palatal or palatalised consonant can be explained by the notion of adjacency. When the element I spreads to the preceding onset, it is not able to spread again to the preceding nucleus. But in the case that the preceding onset cannot accept the element I, then umlaut happens. Because of the notion of adjacency, umlaut does not take place when two consonants intervene the two vowels. However, when the two consonants turn to a single consonant through a phonological process such as aspiration and so-called place assimilation, umlaut takes place. The theoretical background of this paper is Government Phonology, proposed by Kaye, Lowenstamm & Vergnau d (1985, 1990), Charette and Kaye (1994).