The English prepositional phrase often has multiple candidates for att
achment. Usually, it is difficult to determine uniquely the attachment
of the prepositional phrase only by the syntactic analysis. Since the
prepositional phrase appears with a high frequency, it is a very impo
rtant problem in English analysis to disambiguate the attachment. In t
he machine translation, for example, the translation of the prepositio
n and consequently the translation of the whole sentence become unnatu
ral if the attachment of the prepositional phrase is not correctly ide
ntified. This paper proposes the Example-Based Disambiguation (EBD) me
thod, which is composed of the following processes: (1) examples (prep
ositional phrase and its attachment) are extracted from the corpus; (2
) the semantic distance between the input expression and the example i
s calculated according to the thesaurus; and (3) the attachment of the
prepositional phrase is disambiguated using the example with the mini
mum distance. Through the comparison experiment between the proposed m
ethod and the conventional method, it is shown that the former is bett
er in terms of the success rate in the determination of the prepositio
nal phrase attachment.