Could corpus analysis be used to solve some of the problems surrounding non-native speakers' understanding of seemingly synonymous words? This research is needed because there is no single universal unified lexical theory that can be applied to the study of semantics or of language as a whole. Instead, there are a variety of views than can be condensed into two competing positions: (1) language is a system, formal in nature, that does not need to take into account the real world (Chomsky); or (2) the sharing perspective of language, one that does take into consideration the link between shared signs and shared concepts present in daily life. Another way to put it: Do we want our students to be able to communicate with others (sharing), or simply to manipulate words (system). Each word has a meaning (system), but also a sense and way that it is used that is important for communication (sharing). Furthermore, words are rarely if ever wholly synonymous; the most that can be claimed is that they have varying degrees of congruence. Six third-year Japanese high school students made up the study group for this paper. Six pairs of target synonymic pairs (words) in English were chosen for the experiment. The investigation was framed into five distinct stages. In each stage the students investigated the meaning/sense of the word using different methods, beginning with the dictionary and moving on to more complicated exercises, including the use of concordances. It is concluded that this series of exercises succeeds in allowing students to discover and learn the different nuances of words needed for successful second language learning. Five appendices with the exercises used in this study are included. (Contains 28 references.) (KFT)