In the previous seven chapters, we have considered some of the leading theories of language–brain relationships which characterize the first century of work in this field. There are several features of these theories which we shall consider before going on to more modern linguistic and psycholinguistic approaches.

The first point about this work is that it is mainly clinically based. The investigators who developed the theories we have outlined are clinicians, for the most part neurologists. With few exceptions, experimental psychologists and theoretically oriented linguists have not figured in our review. Even when the work of non-clinicians is represented in this survey, it is mainly derived from clinical observations. For instance, Jakobson's data-base for his theories of the dissolution of the sound system was clinical observation. Careful examination of the papers of Broca, Wernicke, Lichtheim, Jackson, Head, Marie, Luria, and other investigators we have cited will reveal that the vast majority of the observations upon which theories are based are clinical.

Clinical observations certainly have an important role to play, but they also have their limitations. In clinical reports, as a rule, the conditions of observation and the nature of the observations are not well defined or controlled. It is rare for the authors we have reviewed to report quantitative data regarding patients' performances, or to describe the exact conditions of testing, the exact nature of the stimuli presented to a patient, or the exact way of scoring a patient's responses. The lack of detail and precision in case reports can create important problems for theory construction.