The New Brunswick Experiment - first post

Vincent Morissette suggests that I may have been over-pessimistic in my judgement as to the capacities of a list such as this to do more than sloganize when addressing complex questions. This seems to be a challenge to which I should at least attempt to rise, so I will try to write a short series of postings in which I will look more closely at the questions and doubts raised by Lightbown's paper on language acquisition through pure input methods. In the first three postings, I will set out some of what I see as the important background questions to be borne in mind when considering her results.

Lightbown's paper was published in the context of an evaluation of 'input- based' language acquisition approaches. Such approaches are most strongly associated with the work of Stephen Krashen. Krashen believes that, to a large extent, adults and adolescents will learn an L2 successfully to the extent that they model their learning behaviour upon that which leads to what he refers to as 'natural language learning', which in turn is closely modeled upon the process whereby a child is held to acquire her L1. This process is not conceived as requiring any direct intervention on the part of parents or teachers - any child who hears a minimum of language, whether directly addressed to her or not, will acquire that language. Acquisition is achieved through the Universal Grammar, which can be thought of as containing all the possible rules of all possible human languages - the child's task is limited to discovering which of those rules apply to the language which is spoken in her milieu. (This is, of course, a vast oversimplification).

In Krashen's version of the basic theory, the child does not need to speak in order to acquire the language - language production is the end product of language acquisition and not part of the process. He also claims that in 'natural' L2 acquisition contexts, adults will sit and listen for a long period - up to six months - before actually attempting to speak. He deduces that what is important for language acquisition is that there should be a sufficient amount of input, and that this input should be delivered in such a way that, even though it may be structurally and lexically somewhat beyond the present competence of the learner, it is nevertheless *comprehensible*. We may ensure that the input is comprehensible through :

a) simplification. Parents, when talking to their children, simplify their language. So do other caretakers, and so, also, do language teachers.

b) contextualization. The language is set within a total context which makes use of gesture, of images, of topic maintenance and so on. Krashen was at one stage very keen on Asher's Total Physical Response approach, in which instructors give orders - 'Touch your nose' - while demonstrating the associated behaviour.

Krashen's model has been much criticized ; it has been suggested, for one thing, that it is intellectually vacuous, being founded upon a distinction between 'learning' on the one hand, and 'acquisition on the other which cannot be sustained. However, here I shall focus upon one strand of criticism, which, while accepting that a greater variety of input than has traditionally been offered to learners is a positive factor, holds that full command of an FL cannot be attained without some insistence on output. The argument here is basically that, if the learner is not forced to actually produce complex utterances, she will not feel called upon to analyse the structure of the language, and will never achieve more than fragmentary command of the syntax. Those who hold to this argument point to the immersion experiments in Canada, where anglophone learners of French are reported to have developed a kind of 'classroom pidgin', in which brief utterances of one or two words suffice to achieve communication.

Lightbown's study has been cited as support for Krashen's input only approach. There is some irony in this, as Lightbown herself has gone on record as a critic of Krashen's. One of the questions we may raise about the study, then, is 'does it in fact support an input- only oriented approach to language acquisition?'.

In a second posting, I will go on to consider the question of whether teaching an FL to primary- school children is an efficient use of educational resources. I will suggest that doubts as to its utility have to taken seriously, and that it is only in the context of the rather limited success of attempts to teach a language to young children in school that we can evaluate Lightbown's findings.

Timothy Mason

Next post