Part 2: Some key issues in applied linguistics
Applied linguistics and descriptions of language
What does it mean to describe 'language'? What does such a description need to include? 'Language' may be conceptualized variously as a form of human behavior, sounds in the air and marks on the page, a system of symbolic communication, an innate cognitive competence; and the answers to these questions may depend on who is asking them and why: the researcher recording the characteristics of a nearly extinct language is likely to need different kinds of information from that required by the minister devising the language policy of a newly independent territory, or a language teacher, or a translator and so on. For some of these purposes applied linguistics draws on linguistics, especially when what the applied linguist needs to do is to give some account of the properties of a language, for example in the context of courses for the teachers of language learners or of children learning literacy. It might be assumed that the descriptions of the properties of languages which linguistics has made available would be ideal for such purposes, but this does not always turn out to be so, partly because the aims of the two enterprises are far from identical.
Linguistics is sometimes referred to as 'autonomous' linguistics, a phrase which signals its concern with the formal system of language as independent from other disciplines (Crystal 1991). The goal of linguistic enquiry, in this tradition (which is associated particularly with Chomsky), is the abstraction from the ubiquitous and polymorphous phenomenon which is language, towards greater understanding of the universals of the human mind. Describing a language is synonymous with 'describing part of the minds' of its speakers (Salkie 2001: 107, 110). The data of speakers' actual behavior (their 'performance'), in this context, '... are disregarded ... because they are of little theoretical interest: they do not provide reliable evidence for the essential nature of human language' (Widdowson 1996: 70). A summary statement of this essential nature, of what is common to human languages, is provided by Pinker (1999: 92): '... the ingredients of language are a list of memorized words, each an arbitrary pairing between a sound and a meaning, and a set of productive rules that assemble words into combinations'. Units such as individual sounds, which are normally meaningless on their own, are elements in a combinatorial system. The finite number of elements- such as the distinct sounds which contrast with each other in anyone language- can combine to generate exponentially larger numbers of units of meaning. According to the Chomskyan tradition, human beings, endowed as they are with linguistic 'competence', are genetically programmed to be able to process language- any language- and what is of interest is the interface between language and mind.
The area of second language acquisition provides one overlap between the concerns of autonomous linguists and those of applied linguists. Some would claim, with Gregg (1993), that the basic focus of this research is '... L2 competence, in the Chomskyan sense of the term: How do L2 learners acquire competence in a second language!, or, Why do learners not acquire the same competence as native speakers of the L2?' (p.278). However, the more 'applied' this interest is, and the more pressing the need not just to investigate these questions but to actually teach particular learners, the greater the priority given to speakers' linguistic performance. Applied linguistics thus necessarily concerns itself with language behavior, with speaking, listening, reading and writing. As conceived within theoretical- or autonomous- linguistics, such behavior involves those features of language which are 'secondary ..., incidental, and peripheral' in contrast to those which are 'essential and primary' (Widdowson 1996: 24-5).
The question of how far the study of language should involve attention to empirical evidence, but at this point it may be useful, particularly for those readers who are less familiar with the linguistic disciplines, to provide a very basic, and necessarily greatly simplified, outline of what is thought to be involved in human linguistic competence.
Language 'rules'
One academic who has made some efforts to popularize the study of language, Jean Aitchison, maintains that 'almost all linguists, not just psycholinguists, are trying to find out about a speaker's mental "grammar"- the internalized set of rules which enables someone to speak and understand their language' (1989: 164). It is so common for rules of this kind to be conflated with the social, extralinguistic rules which are invoked to pronounce an utterance in a non-standard dialect, for example, as 'ungrammatical', that we feel it is worth pointing out the distinction, as introductory textbooks about linguistics always do. Yet in our own experience, academics from a range of disciplines lack confidence in their students'- and often their own- knowledge of linguistic (particularly grammatical) 'rules'. In this they share with the public at large some of the attitudes about local accents and class dialects which within linguistics are classified as norms of a quite different kind from the descriptions which indicate which combinations are possible and which not within a given language variety. At the risk of laboring the point, here it is made by the author of one of the textbooks of the kind referred to above:
Sometimes you will hear people object that certain expressions or constructions are 'not English' or 'ungrammatical'. Some teachers still like to say this about ain't or the use of the double negative, as in lain't got no money. But this is not so. Something is only ungrammatical if it fails to follow a rule in the way it is formed. I ain't got no money doesn't follow the same rule in its construction as I haven't any money but it's not without one. People who use this construction wouldn't dream of saying got I have money n't no, which would be uninterpretable. (Finch 1998: 3)
This injunction to draw a distinction between extralinguistic and intralinguistic 'rules', the latter being 'internalized in us as native speakers of the language' (ibid.: 9), is by no means the end of this discussion. It will resurface in our consideration of lay discourse and epistemic authority, as well as in our exploration of natural and social realism, but for now it is sufficient to allow that languages do at least appear to be subject to rules which fall somewhere between the evaluative norms of the social world and the 'universal' laws of the natural world.
Traditional description of how the units of language are combined assumes a basic pattern that can be visualized as a vertical axis and a horizontal axis which intersect. Within a sentence, the vertical axis is the 'syntagm', which represents the sequential order of words in a language, created by its syntax. The horizontal axis is the 'paradigm', which represents 'the set of alternative words which could be used equally legitimately in a particular position in the syntagm' (Graddol et a1. 1994: 72). This pattern can be illustrated by presenting some of the variations on a simple sentence which are possible in English.
The pattern going across the horizontal axis illustrates some features of English syntax, linking in a familiar order words which can be classified as having particular properties. So we are able to process 'one simple question', but would be puzzled by 'question simple one'. The pattern going down each of the vertical axes illustrates alternatives among words which have something in common with each other. Thus 'long', 'interesting', 'difficult' and 'simple' are all examples of words said to belong to the same 'word class' (also known, more traditionally, as 'part of speech').
A similar pattern can account for the relations between sounds. In English, for example, it is very common for words of one syllable to consist of the sequence consonant-vowel-consonant, and several examples illustrated together (transcribed here- and thus idealized as letters) can contribute to a grid which can vary horizontally and vertically.
pat
bid
top
Variations both across and down the elements of this three by three square can generate multiple combinations.
Traditionally, linguistics is concerned to describe this system of patterns (or rules) among the elements of language, working from sounds, through words to syntax. (Another level of analysis is 'morphology', concerned with the internal structure of words, but it is not important to discuss this here.) Single sentences are often held to be the largest unit of analysis for the 'core' of linguistics, as the patterns involved in longer stretches of language are nowhere near as susceptible to description of this kind.
Methodologically, the appeal to internalized, mental representations of 'rules' such as these is via 'native speaker intuition'. In other words, if you as a reader grew up with English in your linguistic repertoire, you would be expected to know that got I have money n't no violates the rules of English grammar, even if you are not able to parse the construction and explain in technical terminology why this is. The emphasis, as we have remarked, is on competence rather than performance. However, this view competes with alternative approaches to language description which put much greater weight on performance data.
