Teaching and Language Corpora

Anne Wichmann, Steven Fligelstone

• Publisher: Routledge
• Year: 2014
• ISBN:
  9781317889571
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Languages
• Index: Languages & Linguistics

Corpora are well-established as a resource for language research; they are now also increasingly being used for teaching purposes. This book is the first of its kind to deal explicitly and in a wide-ranging way with the use of corpora in teaching. It contains an extensive collection of articles by corpus linguists and practising teachers, covering not only the use of data to inform and create teaching materials but also the direct exploitation of corpora by students, both in the study of linguistics in general and in the acquisition of proficiency in individual languages, including English, Welsh, German, French and Italian. In addition, the book offers practical information on the sources of corpora and concordances, including those suitable for work on non-roman scripts such as Greek and Cyrillic.

1

_________________________

Teaching and Language Corpora: a Convergence

GEOFFREY LEECH

There is every reason to believe that language corpora will have a role of growing importance in teaching. This book, and the workshop (TALC94) which gave birth to it, are testimonies to the richness of the interest and experience which are already being applied to the convergence of language teaching and language research, through the link of corpus-based methods.

1 Up to now

Until recently, teaching had little connection with the momentum behind the evolution of corpus-based methods in linguistics. There were other forces in play. But one of the functions of the TALC94 workshop, as the first-ever international (or even national) conference on corpora and teaching, was that it enabled us to learn, for the first time, about the whole range of largely unpublicized pedagogical activities making use of corpora.

The experience of ICAME (International Computer Archive of Modern English) has been especially indicative. For seventeen years, ICAME, with its annual conference and journal,¹ has spearheaded research developments in corpus linguistics, with particular reference to the English language. But it was not until 1992 that there was an item in the ICAME conference programme referring to the use of corpora in teaching. This was the paper by Steve Fligelstone, ‘Some reflections on the question of teaching from a corpus linguistics perspective’. Fligelstone led a workshop at the 1992 conference in Nijmegen on the topic, and the paper was reworked for publication (Fligelstone 1993).

Future historians tracing the history of computer corpora in linguistics might easily assume that ICAME members had given no serious thought to the educational use of corpora up to that time. But this would be a false conclusion. Most of the members of ICAME were then, and still are, university teachers, and most of them will have increasingly been using their corpora and corpus-based techniques in teaching, as well as in research, for a number of years. In my own case, for example, I began using an incomplete prototype LOB (Lancaster-Oslo/Bergen) Corpus for postgraduate teaching as early as 1976, and this use of corpora in teaching has continued, and gradually been extended to new areas of the curriculum at Lancaster, ever since. The original ‘trickle down’ from research to teaching is now becoming a torrent!

The notion of ‘trickle down’ from research to teaching seems particularly appropriate to corpus linguistics. This is because the computer corpus, as a resource for finding out about language and texts, is totally neutral as to these two major interconnecting activities of universities. The corpus, purely as a resource, is rather like a shelf in a university library: it is there to be exploited, and the same resources are equally usable for research and teaching. The history of computer corpora, on the other hand, has been tied to the history of computer technology. Inevitably, while computers were limited to large mainframes available to the initiated few, computer corpora were largely restricted to research use. But as computers have grown smaller, cheaper, and massively more powerful, their use in teaching has grown immeasurably. It is natural that the movement from research to teaching has taken place in this way, as the information revolution in the use of computers has more and more extended itself from the laboratory to the classroom.

It is also evident that the corpus, as an information source, fits in very well with a dominant trend in university teaching philosophy over the past twenty years, which is the trend from teaching as imparting knowledge to teaching as mediated learning (cf. Laurillard 1993: 13–15). In this context, there is no longer a gulf between research and teaching (cf. Knowles 1990), since the student is placed in a position similar to that of a researcher, investigating and imaginatively making sense of the data available through observation of the corpus. As Tim Johns has said (quoted by Gavioli in this volume), [Georges Clemenceau (1841–1929): ‘War is much too serious a thing to be left to the military.’] ‘research is too serious to be left to the researchers’: teaching is a natural extension of research. The student-centred paradigm of ‘discovery learning’ — or what Johns has called ‘data-driven learning’ — can scarcely be better exemplified than through the use of the computer corpus. Almost uniquely, among the information resources of which students make use in education, a text corpus is of primary interest because of what it is. Other resources, such as databases, are of interest indirectly, because of what they are ‘about’. But a corpus is, of itself, a rich resource of authentic data containing structures, patterns and predictable features that are waiting to be ‘unlocked’ by the human intelligence. Perhaps the nearest equivalent, in other disciplines, is in the direct confrontation with data that occurs in the scientific laboratory, or in fieldwork. It is this experiential confrontation with the material of study that can make corpus work so rewarding for the student. And it often happens that a student working on a relatively small corpus assignment comes up with original observations and discoveries which have probably never been brought to notice before, even in the most detailed dictionaries and grammars of a language.

Having quoted Tim Johns, I should celebrate the nature of his contribution, as a prime example of a university teacher who has exploited the computer corpus mainly for teaching. Indeed, the above quotation ‘research is too serious to be left to the researchers’ reverses, by implication, the traditional donnish assumption that research or scholarship is the more important thing, and that teaching is just a spin-off from it, the vehicle whereby students are permitted to participate in the don’s world of recherché knowledge. Tim Johns’s earlier work in CALL (Computer-assisted Language Learning) — see Higgins and Johns (1984), Johns (1988) — naturally availed itself of the corpus-rich atmosphere of Birmingham, and he became among the first to advocate and to explore the use of corpora in teaching. Perhaps it is significant that Johns, being a teaching-oriented rather than research-oriented lecturer, never became a habitué of ICAME! And it is also significant that he was the first to insist that the use of computer corpora in teaching was itself a topic for research (Johns, this volume). (See also Johns 1991a, 1991b, 1993.)

Those, like Johns, who have been placing teaching with corpora at the forefront of our attention as a matter of primary interest, may well find my ‘trickle down’ metaphor unhelpful, or even offensive. ‘Trickle down’ implies that research is ‘up there’ as an élite activity, and teaching is ‘down here’ in a lower, subservient role. But, in the experience of many, there is not a one-way dependence of this kind. One finds that ‘trickle up’ from teaching to research can be just as important.

The convergence mentioned in my title is a natural coming together of teaching with research from various points of view. This is natural whether we consider it from the ‘trickle down’ point of view, where the resources and techniques used in research progressively become available for teaching, or from the Johns ‘trickle up’ point of view, where the development of language-teaching techniques naturally appropriates to itself the resources available for research, and becomes a topic for research in its own right. The convergence is aided by the increasing similarity, in higher education, of the paradigm for research and the paradigm for teaching (using and analysing resources in a self-access mode).

Research with corpora, over the past twenty years, has become an amazingly fertile development. Whereas as recently as ten years ago corpus-based methodology was the fringe activity of a tiny minority of eccentrics, it has now become the mainstream of computational linguistics, and has increasingly established itself in mainstream linguistics. This in itself means that corpus linguistics is appearing more and more as a part of the university curriculum in linguistics, both at undergraduate and postgraduate levels.

But the thing to avoid, if we can, is treating the use of corpora in teaching as a bandwagon. Teaching bandwagons, if driven too far and too fast, can do much harm to those on the receiving end. Some will remember that ten years ago, when the new educational possibilities of CALL were very much in the ascendancy, many warned heavily against too great an enthusiasm for this new toy — the computer in the language classroom. False expectations of the powers of technology, it was recalled, had been raised by an earlier innovation, the introduction of language laboratories. The warning was salutory: since the computer entered the classroom, students have learned a lot about how to handle computers. But has their knowledge of languages taken a great leap forward? The educational benefits of technologies are still far from fully understood and acknowledged.

At that time, Higgins and Johns (1984: 12) were among those who warned against a revolutionary zeal for computers. As a conception of the role of the computer in the classroom, Higgins (1988: 12–15) preferred the metaphor of the ‘pedagogue’ (in ancient Greece, the slave who accompanied the pupil to school) to that of the ‘magister’ (the Roman ‘master’ to whom the pupil submitted in obedience). Unlike the magister, the pedagogue was merely a humble facilitator of the learning process. This view of the computer certainly comes to the fore when we think of corpora. The computer is simply the device that gives access, the intermediary between the learner and the corpus as a fountain of knowledge and understanding. But we may go even one step further, and say that the corpus itself has no more than the facilitative ‘pedagogue’ role. It enables the learner/student to explore, to investigate, to generalize, to test hypotheses; but it does not itself initiate or direct the path of learning.

It is timely, none the less, to welcome the emergence of the computer corpus as a linguistic learning resource. The convergence of research and teaching is already taking place — it is a fait accompli. Our task is to make the best use of it, and to exchange ideas on how the computer corpus can be exploited to the best advantage in the future. This means, first, exchanging experience on how we have used corpora in teaching in the past, and how we are developing these techniques at the present time. The time is right for taking stock. My plan is to do this by surveying the activities connecting corpora and teaching, and their motivations.

2 And now?

Like many fields of endeavour, the corpus-aided language teaching field can be thought of as containing a core — a central or focal area — and an expanding periphery. The core, which can be seen as the main concern of this book, is the direct use of corpora as resources for teaching. The periphery can be seen as a set of corpus applications which indirectly contribute to teaching. Both the core and the periphery are important for the way we think about the field, and for realizing its potential. Yet a further set of activities, more peripheral still, takes the form of teaching-oriented corpus developments. The follow list summarizes the activities I have in mind:

■ Direct use of corpora in teaching:

Teaching about

Teaching to exploit

Exploiting to teach

■ Use of corpora indirectly applied to teaching:

Reference publishing

Materials development

Language testing

■ Further teaching-oriented corpus development:

LSP corpora

L1 and L2 developmental corpora

Bilingual/multilingual corpora.

The three main headings above can be viewed as three concentric circles, starting with the innermost one, which will occupy our attention in the following section.

2.1 Direct use of corpora in teaching

What is the nature of the interaction between corpora and teaching? The three different ways in which corpora may be used in teaching, as listed above, are distinguished by Fligelstone (1993): teaching about [corpora], teaching to exploit [corpora] and exploiting [corpora] to teach.

Teaching about

The first of these is probably the least interesting or innovative. As I have already said, corpus linguistics, seen as a subdiscipline within linguistics, has now come of age (see Svartvik 1992), and is beginning to find its way into curricula, both postgraduate and undergraduate. One symptom of the ‘arrival’ of corpus linguistics is that introductory textbooks on the subject are already being written: I am aware of more than one such publication under preparation at present.

What does teaching corpus linguistics mean? Just as a student studying for a linguistics degree (or for that matter any language-related degree — say, in English or Italian) takes courses in such subjects as phonetics, syntax, sociolinguistics, or discourse analysis, there are now beginning to be courses on corpus linguistics, or courses containing corpus linguistics as a substantial component. I recently received a letter from an East European university asking me for a basic reading list on corpus linguistics, and some information about corpora and the software available. The letter enclosed an outline of a syllabus for a corpus linguistics course the author was planning to introduce. This was of interest to me, in showing that, for an academic who had not reached corpus linguistics directly through research involvement, the teaching of corpus linguistics was nevertheless becoming an important part of the curriculum.

As with other courses, the curriculum will tend to cover main areas of the subdiscipline: say, its history, its data and subject-matter, its methods of investigation, the models or theories it employs. In the case of corpus linguistics, inevitably important topics are: (a) what corpora exist?, (b) can they be accessed, analysed or exploited?, (c) what software can be used for this purpose?, (d) what are the applications of corpus linguistics? And arising out of these is a more philosophical or theoretical question: (e) what view of language and methods and goals of linguistic study is presented through corpus linguistics? How does this view compare with other views? (Here the Chomskian distinction between ‘internalized’ and ‘externalized’ language comes to the fore: corpus linguistics very much identifying its domain as the latter — see Chomsky 1988.)

In principle, of course, corpus linguistics could be taught as a purely academic subject, in which the students never get their hands on a computer, or gain access to a corpus. But, I would strongly suggest, almost more than any other branch of linguistics, corpus linguistics requires that students have ‘hands on’ experience of the subject: of the use and exploration of corpora. A course which did not provide access in this way would be like an astronomy course in which the students were never allowed access to a telescope: it would be a dull course indeed. Only through using corpora can one gain a first-hand sense of their potential. For example, in using a grammatically tagged corpus, one starts asking intelligent questions about how it is possible to build an automatic tagger. And one also considers questions of linguistic content su...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Preface
• Editors’ acknowledgements
• Publisher’s acknowledgements
• Contributors
• General Introduction
• 1 Teaching and Language Corpora: a Convergence
• Section A Why Use Corpora?
• Section B Teaching Languages
• Section C Teaching Linguistics
• Section D Practical Perspectives
• Appendices
• References
• Index