There has been a great deal of discussion in the literature on second-language acquisition about the relationship between first-(L1) and second-language (L2) acquisition. I believe that the dust is beginning to settle and that it is now possible to move beyond this debate. I would like to do so in this paper, and examine in particular the role of linguistic input in L1 and L2 and the nature of conversational strategies in both processes.

The term “interlanguage” is generally used in connection with the performance of second language learners in naturalistic learning situations, that is, outside the classroom and in situations where message CONTENT rather than message FORM is in focus. This is a reflection of the generally healthy separation that has taken place between investigations into learning and research into teaching or instruction. It remains true, however, that teaching methods and techniques can hardly be formulated without taking into account our current knowledge of what goes on or what seems to go on in learning. This, of course, has to be done with all due caution, that is, with a full appreciation of the limited nature of this current knowledge (Tarone et al. 1976). Seeing teaching from a learner's point of view is, in itself, a desirable attitude and one which is likely to be fostered if one searches through the literature on acquisition for possible insights, pointers, and even techniques of relevance to language instruction. Over the past four years, instructional techniques for teaching writing to advanced learners of English as a foreign language have been developed at the English Department, Utrecht University. These techniques are directly inspired by research into learning and have in turn proved to be of value for such research even though their primary aim is instructional. This paper will concentrate on the instructional aspects.

In this paper I shall argue that FEEDBACK constitutes a central but unjustly neglected concept in the study of institutionalized foreign language learning. Since feedback is the crucial element in distinguishing cybernetic from linear models in general, I shall also be adopting the position that language learning, and foreign-language learning in particular, is best represented in cybernetic terms. To date, however, as Crothers and Suppes (1967:19) point out, feedback in psychological literature (and the same is true, I think, for the literature on foreign-language teaching) has usually been characterized simply as a 0, 1 event-that is, as a binary opposition which could be represented as “yes/no.” Our discussion will show, however, that even in the limited communication situation of the foreign-language classroom the students can and do extract much more from the behavior of the teacher than we might suppose – even if what they learn is somewhat different in nature from what the teacher intends. In proposing an explicitly cybernetic approach as a basis for modeling foreign-language learning, we must be careful to point out that this approach has most in common with many models put forward to explain particular processes within biology and, in the study of speech, with those postulated to explain speech degeneration (Arnold 1960, 1961; Gerard 1959; Milisen 1966; West 1957; Wolf and Wolf 1959; Wood 1945) and also some aspects of speech production and perception (Laver's 1970 model is implicitly cybernetic, as are Corder's speculations on the role of error in foreign-language learning. The analysis-by-synthesis model of speech perception, in that it can be reduced to the form HYPOTHESIS + FEEDBACK, is also essentially cybernetic in nature.)

The following paper is an account of empirical research in lexical acquisition, an ongoing study whose preliminary findings were reported at the Neuchâtel colloquium May 1976 (Levenston and Blum 1977). A discussion of the research methodology has also been published (Levenston and Blum 1978), as well as an attempt to link this work with studies of lexical simplification in other linguistic contexts (Blum and Levenston 1977). Before discussing in detail the current findings, it seems appropriate to summarize briefly the preceding research, of which they form an integral part.

The data presented in this paper were gathered using a French cloze test in the context of a longitudinal evaluation of a number of French immersion programs in Ontario conducted by the Bilingual Education Project of the Ontario Institute for Studies in Education (see, for example, Barik and Swain 1978(a), 1978(b); Swain and Barik 1976; Swain and Lapkin 1977). One of the major research questions to which these studies have been addressed concerns the degree of second language proficiency attained by the immersion students: how much French is acquired by English-speaking students in these bilingual programs compared to that acquired by students in a regular program taking 20 to 40 minutes daily of core French or conventional French as a second language training? Through the use of standardized French achievement tests and other French measures developed by the Bilingual Education Project (e.g. Barik 1975, 1976; Barik, Swain, and Schloss 1978), it was determined early in the longitudinal study of early French immersion that immersion pupils were attaining a level of second language proficiency far superior to that of the comparison pupils taking core French. It became clear that a more appropriate reference group for measuring French language proficiency would consist of native French-speaking pupils of the same age and grade level. In other words, a shift of emphasis in the original research question has occurred over the past several years so that now the focus is on the degree to which immersion students can attain native-like ability in the second language (Lapkin 1978).

This paper presents the results of an experiment which investigated the syntactic parsing strategies used by L2 learners at different stages of acquisition. This represents a shift in emphasis from most studies in the field of language learning which deal almost exclusively with production data. The goal of acquisition studies is to understand both the nature of the linguistic system that a learner has acquired and the processes involved in the acquisition of that system. This linguistic knowledge is generally inferred from the perceptive or productive behavior of the learner. There are several reasons why production has been emphasized over perception in second language acquisition research: 1) Research in this area has been greatly affected by studies on child language acquisition in which traditional perception experiments are inappropriate due to a child's lack of cognitive maturity and verbal abilities. 2) Many perception experiments require a paraphrasing ability beyond that of L2 learners in the early stages of acquisition. 3) Problems involving production are more apparent in most pedagogical encounters and, therefore, are of more immediate concern to researchers. Yet, the investigation of perceptual strategies of L2 learners is important, if only because it affords us the possibility of viewing a different corpus of L2 data, thereby giving us additional insight into the nature of the acquisition process.

The present paper touches slightly on theoretical issues concerned with the notion of linguistic redundancy but has primarily a very practical scope, i.e. the facilitation of the foreign language learning process. Specifically, this paper suggests the utilization of redundancy features in aiding learners to develop their recognition knowledge in the new language. As such, the paper is concerned with course objectives and the learner's linguistic needs.