When we turn to research on the development of effective procedures for using the language competencies in acquiring new knowledge, we again encounter a formidable literature-although not many useful results for instructional purposes. Rothkopf (1972) has questioned the value of studies from verbal learning laboratories to learning by reading in realistic settings. In turn, Carver(1972) has critically reviewed Rothkopf's and other research on "mathemagenics" and concluded that "The tasks used in the experiments were not valid when generalizing to any applied situation and were not valid for generalizing to any situation of theoretical interest" (p. 116). Carroll (1971) thinks that the concept of mathemagenics is useful, yet he agrees that "Carver's criticisms have some force; many of the points he raises should be made the basis for further experimental investigations" (p. 166).
Various attempts to improve learning from listening materials have been reviewed elsewhere (Sticht, 1972) with the general conclusion that such efforts have, to date, been largely ineffective.
The question of the relative effectiveness of spoken versus written discourse, or a combination of the two (audio-visual media), on learning outcomes has most recently been considered by Carroll (1971). His comments on past reviews in this area suffice to bring the reader up to date: "All these reviews suggest that the matter is an extremely complicated one; research seems to present conflicting evidence on numerous points" (p. 129).
As the number of empirical studies grows larger and larger, with seemingly little progress being recognized on how people learn to read and how they read to learn, we find more and more often the call for theory-based research. Anderson (1972), Bormuth (1970), Carroll (1972), and Simons (1971) have called for theories of language comprehension. Geyer (1971) reports more than 40 models or partial models of the reading process or its components. Bloom (1971), commenting on operant conditioning and cognitive psychology approaches to reading (behavior?), feels that we are witnessing a paradigm clash a la Kuhn (1962) between behaviorist and cognitive psychologists, from which the better paradigm will eventually emerge dominant. Frase (1969), speaking about the discipline of instructional psychology in general, and the problem of learning from meaningful verbal discourse in specific, has succinctly summarized what many have felt is a major problem. Our empirical data are not adequately organized, hence they do not provide knowledge-just facts-and many of these facts are tenuous at best.