The class of pattern languages was introduced by Angluin (1980), and a lot of studies have been undertaken on it from the theoretical viewpoint of learnabilities. However, there have been few practical studies except for the one by Shinohara (1982), in which patterns are restricted so that every variable occurs at most once. In this paper, we distinguish repetitive variables from those occurring only once within a pattern, and focus on the number of occurrences of a repetitive-variable and the length of strings it matches, in order to model the rhetorical device based on repetition of words in classical Japanese poems. Preliminary result suggests that it will lead to characterization of individual anthology, which has never been achieved, up till now.