The main aim of this paper is to clarify the syntactic and semantic properties of the cognate object expressions talk the talk and walk the walk, so as to give a correct Japanese subtitle to a US soldier’s utterance in the film Full Metal Jacket (1987). So-called cognate object expressions consist of two types, the event-dependent type and the event-independent type (Kitahara 2010, 2011). While the cognate object of the former type just repeats the notion denoted by the main verb, the cognate object of the latter type functions as a representation-source theme referring to a pre-existing type (Dowty 1991, Langacker 1991). Through an examination of data from large-scale corpora, it is shown that the expressions talk the talk and walk the walk belong to the event-independent type. In parallel with these linguistic analyses, the paper argues that authentic data from movies can serve as an effective introduction to specialized education in English Linguistics for university students.