This monograph investigates aspects of language processing that are not specialized in the left hemisphere, and claims that there are "levels" (such as pitch functions) and "subsets" (such as phrase structuring) which are different in essential ways from each other, and from the aspects of speech and language which are typically lateralized. Chapter 1 presents the assumptions that underlie the rest of the monograph: that varieties of evidence are valuable to the formulation of models of language; and that hypotheses about language should be not only testable, but tested. The claim is made that neurolinguistics provides one such context for hypothesis-testing. Neurophysiological foundations for studies of language and speech in the brain are surveyed; and facts are presented about hemispheric specialization for language. It is claimed that some functions of pitch in speech are not specialized in the language hemisphere, and evidence is presented for cerebral processing of pitch. A description follows of types of speech performance more or less affected by brain damage, and a series of dichotic listening studies on "automatic" speech are described. Finally, it is concluded that speech is made up of graded, heterogeneous subclasses and that, therefore, continua are needed in descriptive models of language. A 54-page bibliography is provided. (Author/AM)