was concerned, all errors whatever their origin were dealt with byessentially the same technique of further drilling and exercise. Thefirst paper in this collection dates to a time when this essentially behaviouristic account of second language learning was coming to be seriously questioned. This was the result of the interest which psycholinguists, influenced by Chomsky, were beginning to show infirst language acquisition. It was natural that one should ask whether the cognitive processes which came into play in firstlanguage acquisition were the same as those used in secondlanguage learning, and indeed in the early seventies, which markedthe beginning of serious empirical research into second languageacquisition, this was the question which was uppermost in theminds of the researchers. With the general abandonment of the belief in a specific language acquisition device, this is no longer aquestion which actively engages the interest of investigators. At thesame time the role of the first language in second languageacquisition has become a more interes'ting question. The term
interlanguage
was coined by Selinker in the belief that the languagelearner's language was a sort of hybrid between his
LI
and the targetlanguage. The evidence for this was the large number of errorswhich could be ascribed to the process of transfer. But when secondlanguage acquisition researchers began to collect data from learnersnot receiving formal instruction, particularly children, the pro- portion of transfer errors was found generally to be quite small.Furthermore, these errors seemed to be found in most learners atthe same stage of development and largely independent of thenature of their mother tongue. Clearly interlanguage was not ahybrid language and had a developmental history of its own. Thespeculation"about a built-in syllabus for second language learningmade in the first paper in this collection in 1967 seemed to bereceiving empirical support. The notion of a 'natural sequence' for second language learning is now widely accepted with considerablesupport from experimental evidence. The relevance of thesefindings for language teachers is clear: that if we could establish thenatural order in which a knowledge of the second language isgradually built up by the learner, then the materials, particularly thestructural syllabus, could be graded upon a more solid basis thanthe current one, which is a mixture of some concept of usefulnessand some idea of linguistic dependency, but certainly not on any psycholinguistic evidence of language learning.While establishing the presence and nature of a 'natural sequence'of development may be the principal objective of second languageacquisition research, the field has broadened out in its scope toinclude other topics. There has been in recent years a shift
Introduction 3
of emphasis in language teaching from a preoccupation with thelearning of the language as a system towards the functional use of that system for communicative purposes. This has had its influence insecond language acquisition research. There is now a greater awareness that under natural circumstances languages are acquiredthrough the need and attempt to communicate, that is throughconversation. But what is the nature of that conversation? Is it justlike that between native speakers or does it have special character-istics? Native speakers of a language in fact adapt their use of language in a number of ways when interacting with learners andthis is the data on which a learner works to create for himself hisknowledge of the language system and its use in communication.But the learner is himself hampered in his attempt to use hisinterlanguage for communicative purposes by its relative simplicityand poverty. How does he overcome these disabilities? Whatstrategies does he adopt to minimize the disabling effect of hisignorance? These too are topics dealt with in papers in thiscollection. An understanding of interlanguage is no longer narrowly bounded by a consideration of the structural properties; we also wantto know the communicative circumstances under which it developsand how it is manipulated by its speakers in their attempts tocommunicate. We want to know these things because they too may be relevant to language teaching and learning