182
Of course, I have left many serious questions unsettled, such as the range of topics to be read and discussed, the problem of requiring equal levels of proficiency in cognate (and hence more comprehensible languages) and more exotic languages, and how the passing level is determined. Some applied research may eventually help to solve them; at the moment, however, the format seems clear.
Second language acquisition theory, as presented in this volume, gives no magical and obvious answer to the difficult question of ESL testing at the university level. The goal of such testing is to determine whether students know enough English to study in English. As is well known, the "backwash" effect has been a problem in this area for years: many foreign students study for the TOEFL examination exclusively, and are helped to do so by special courses designed to do just this (see discussion in Wiggon, 1979).
Applying the same arguments here that we used earlier, emphasizing the instructional value of tests, we come to the conclusion that subject matter testing would be of benefit at this level as well. This is more easily said than done; it would be prohibitively expensive to design standardized subject matter tests in all disciplines for international students. International courses, as outlined in the previous sections, may be a step in this direction, as long as they use subject matter tests as finals; a students' release from the ESL requirement could be at least partially dependent on his ability to pass international courses.
If the conclusions we have reached in this volume are correct, it implies that we have some fairly serious gaps in our materials. Before listing where I think these gaps are, let me first of all note that materials need to meet the same requirements that methods do, as listed in Chapters III and IV. If materials are supposed to help students in language acquisition, they should either supply input that is comprehensible, interesting/relevant, and not grammatically sequenced themselves, or they should provide students with the means of obtaining such input. If materials are meant to help language learning, they
182