Abstract
The present study investigated collocation patterns of frequently-used 22
onomatopoeias and verbs, taken from Korean corpus and 40 native Korean
speakers. Using the indexes of entropy and redundancy, type and token
frequencies of collocational appearances of each onomatopoeia and its multiple
verb combinations were calculated using the Sejong corpus
(http://www.sejong.or.kr) and verbs recalled by the native Korean speakers. A
significant difference was found in neither entropies nor redundancies between
corpus frequencies and recall frequencies. This result suggests that collocation
patterns obtained by corpus data must reflect those recalled by native Korean
speakers. The cluster analysis over entropies and redundancies of the 22
onomatopoeias showed three clusters, (1) onomatopoeias with mild restriction of
verb-collocation (various verbs co-occurring with these onomatopoeias), (2) those
with semi-mild restrictions, and (3) those with severe restriction (regularly cooccurring
with a few verbs). The present study depicted a wide range of
collocation patterns of the 22 onomatopoeias and verbs.
