This review of Bever's psycholinguistics survey is for the most part favorable. Commentary is centered on sections 1, 2, 4, and 6 of the report. The survey's first part is judged significant in that Wundt's pioneering work in psycholinguistics is discussed. The second section, on grammar as a psychological process, is found obscure in its denunciation of the competence-performance distinction. Evidence presented in section 4, on language perception processes, is not found convincing enough to refute the theory of the psychological reality of linguistically defined surface structure. The failure to present relevant data is cited as another fault of the report as well as forcing conclusions based on weak evidence. Section 6 presents three implications of the theory for research in automatic language processing. The arguments, that psycholinguistic research is not a constraint on computational linguistics, and that automatic language processing has as a goal the development of man-made communication, are held to be unsound. The argument that the development of computational algorithms based on grammars generates advances in linguistic theory is considered more valid. (Author/AM)