This chapter describes a method of disambiguating multi-sense words in a sentence by using example sentences in which such words are already disambiguated, and by using taxonym and synonym hierarchies. As a knowledge base, we developed a small-scale text database containing 730 example sentences in English that include the verb “take,” and prototyped a program that resolves 12 senses of the verb “take” in the input sentences. Our test results show the feasibility of our approach. The advantages of the approach are: (1) it does not require special semantic categorization; (2) the knowledge base is easy to create and maintain; (3) closely related senses, as in polysemous words, can be disambiguated; and (4) the approach is robust.