Using the Seoul Corpus (Yun et al. 2015), or the Korean Corpus of Spontaneous Speech, this paper shows that a categorical account of Korean lateralization in /ln/ sequences cannot explain the variability of real pronunciation of underlying /ln/ sequences in Korean. Previous accounts of underlying /ln/ sequences in Korean have considered them to be realized with progressive assimilation, which is also defined as a phonological rule in Korean grammar by the National Institute of Korean Language. Herein, by examining the Seoul Corpus, underlying /ln/ sequences within word boundaries were found to not always be realized with progressive assimilation ([ll] 53.47%), but were found to sometimes be realized with non-assimilation ([ln] 39.1%, [øn] 7.43%). Moreover, nonassimilation in underlying /ln/ sequences was found to be statistically significant for certain sociological factors, including gender (P<.05) and age (P<.001). This shows that women apply assimilation more than men, and older generations apply assimilation more than younger generations in the same contexts. Finally, Korean lateralization was found to be sensitive to speech rate (i.e., the difference between [ln] and [ll] in duration was found to be statistically significant (P<.05)) based on the phonetic measurement of the duration of each syllable. These results indicate that lateralization occurs more than nonassimilation in fast speech. (Keimyung University and Kyung Hee University)