to me to be fundamentally different, although it may meet many of the goals ESP is designed for. While ESP requires a detailed analysis of the syntax, vocabulary, and discourse of a subfield, to be developed into a syllabus and presented bit by bit, subject matter teaching focusses only on the topic, the information or skill to be learned, the assumption being that much of the syntax, vocabulary, and discourse style will be acquired along with the subject matter. (This idea is not entirely foreign to ESP; several ESP courses emphasize "authentic activities". See, for example, Robinson, p. 39; Widdowson, cited in Robinson, p. 23.)

(d) Evidence for subject matter teaching: the immersion programs

Immersion bilingual programs have demonstrated what is possible in second language acquisition using subject matter. In immersion programs, initially monolingual majority children are schooled in a minority language (French in Anglophone Canada; Spanish in the United States). They are taught their academic subjects totally in the second language. In what is known as "total early immersion", input in the second language begins in kindergarten. Late immersion programs may begin later, after the children have had at least a year of instruction in the second language.

The immersion programs appear to be successful in many ways. The many reports that have been published confirm over and over that immersion students acquire high levels of competency in the second language (while they may not reach native-like levels, they outperform peers who have had standard foreign language classes), they make normal progress in school, doing as well in subject-matter as monolinguals, and they do not fall behind peers in first language development (for reviews, see Lambert and Tucker, 1972; Swain, 1974).

Cohen and Swain (1976) discuss these successes in light of the lack of success of many other types of bilingual programs. Among the differences between immersion and other programs, these characteristics of immersion may help to explain its success. Cohen and Swain point out that in early immersion "all kindergarten pupils are unilingual in L1. In essence, the successful program starts out as a segregated one linguistically" (p. 47). As mentioned above, this raises the students' chances of

170