Douglas Bilber, Stig Johansson, Geoffrey Leech, Susan Conrad, Edward Finegan: Longman Grammar of Spoken and Written English

[Originally posted to the CETEFL list 18/1/00]

Douglas Bilber, Stig Johansson, Geoffrey Leech, Susan Conrad, Edward Finegan: Longman Grammar of Spoken and Written English, Pearson Education Limited, 1999, Hardback ISBN: 058223725-4, 1204 pages, English price 73.83 pounds sterling (roughly 443 US dollars).

(I have called this a notice rather than a review to emphasise that this is an account more than an evaluation).

One of the most fascinating things about this book - apart from the price - is how it was written. Descriptive and based on research it is a truly collaborative enterprise. Editorial meetings took place at Cambridge, London, Feusisberg (Switzerland), Flagstaff (Arizona) and Lancaster (UK). The authors are American, English and Norwegian. Email enabled the authors, Douglas Bilber, Stig Johansson, Geoffrey Leech, Susan Conrad, Edward Finegan and their research teams to keep in close contact.

For their analysis of spoken English the authors used the British National Corpus produced for the British publisher Longman. The British Market Research Bureau was subcontracted by Longman to gather data on spoken English in England, that is to collect natural, everyday conversation from a representative sample of the British population, which they did across age, sex, social group and regional spread. (A similar procedure was followed in the USA). In both countries the informants tape recorded all their conversational interaction over a week using a high-quality tape recorder. The conversations were then manually transcribed and later analysed lexicographically and grammatically. These procedures generated 329 American "texts" - "text" is used in the context of conversation to mean a unit of conversation - (2,480,800 words) and 3,436 British "texts" (3,929,500 words). These were supplemented by different dialects and non-conversational speech e.g. lectures.

To build up the corpus of written English they took texts from the registers fiction, newspaper language and academic prose. For the most part the vast majority of texts were written after 1980, with the exception of a few classical fiction texts from the first half of the twentieth century. These were supplemented by some general prose. This produced 34,137 texts (38,577,000 words).The texts varied in length from short articles to whole novels.

The total corpus, conversation plus written language, produced 37,244 texts representing 40,025,700 words. (For comparison, take the COBUILD Grammar, which is based on something like just over 20 million words).

The whole of the LSWE corpus was then analysed grammatically. First the corpus was tagged automatically i.e. each running word - a word separated by a space on both sides - was given a grammatical label such as "plural common noun" or "past tense lexical verb". The automatic tagger achieved an accuracy of 90-95%.

Specially developed computer programmes then investigated selected features. A by-product of these processes was the creation of data bases, for example frequency lists of all the adjectives or verbs found in the corpus, or the 50 most common verbs found in a particular register. Where the description of grammatical features involved not just form but meaning, or in ambiguous cases, the analysis was carried out by humans. (For details of further procedures used I refer you to the introduction to LGSWE).

The two most important basic conclusions that are drawn from an examination and analysis of this gigantic amount of data are:

(1) the confirmation that it is necessary to distinguish between the grammar of spoken and written English (and even between different registers).

(2) the necessity to account for grammatical choices in terms of use.

"The vocabulary and grammar that we use to communicate", write the authors in the introduction, "are influenced by a number of factors, such as the reason for the communication, the context, the people with whom we are communicating, and whether we are speaking or writing. Taken together, these choices give rise to systematic patterns of use in English." Traditionally, such patterns have not been dealt with by grammar, partly because only structural matters were considered, but partly because the necessary information was not available. LGSWE changes this by using up-to-date technology to "chart new domains with new high-resolution detail." It describes not only the available range of grammatical features in English but their actual use. "Actual use" means how features are used, to what extent they are used, their variability in relation to other features and the factors that favour or disfavour each variant.

The authors acknowledge their indebtedness to A Comprehensive Grammar of the English Language (Quirk, Greenbaum, Leech and Svartik, Longman., 1985) adopting its descriptive framework and terminology and seeing the two grammars as complementing each other. On matters of frequency of use, differences across registers , the new grammar supplants native speaker intuition (rarely accurate the authors state) which up to now, has been the only resource available.

Each section of the grammar, typically, has the following divisions:

1. Exposition of the point under consideration

2. Frequency information

3. Discussion

It should be emphasised that all the examples quoted in the course of the expositions are authentic examples drawn from the corpus and not specially written examples constructed to illustrate separately stated rules. In other words , the LGSWE is the grammar of the 40 million words of the corpus used in naturally occurring contexts which are taken to be representative of the whole English language.

I have chosen three sections to give an impression of how the grammar proceeds - the use of the definite article, the survey of lexical verbs, and the progressive aspect.

The use of the definite article, as described in LGSWE distinguishes between:

(1) Anaphoric reference (the referent of the noun phrase is assumed to be known) e.g. "A doctor was allowed to carry on working after telling (colleagues) he had Aids... The doctor , who died last summer....

(2) indirect anaphoric reference (knowledge inferred) e.g. "The Mercedes took a hard bounce from a pothole.."

(3) cataphoric (referring to something that comes later) e.g. "Another potential votes tells him about the car that went through his wall.."

(4) situational - the shared situational context of the speaker and hearer e.g. I think there is someone at the door.

The corpus findings are as follows:

(1) There are marked differences across the registers of patterns of noun phrases with the definite article.

(2) Anaphoric reference accounts for less than a third of the definite noun phrases in all the registers.

(3) Situational reference is found for more than 50% of the cases in conversation, but only 10% in written registers.

(4) Cataphoric references represent 30-40% of the definite noun phrases in news and academic prose.

(5) The distribution of the indefinite article is relatively similar across registers: approximately 20,000 occurrences per million words in the written registers and about 13,000 per million words in conversation.

(6) With the distribution of the definite article it is different. There are much greater difference across  registers. The definite article is more than twice as common as the indefinite article in written registers though in conversations the frequencies of the indefinite and definite are more equal.

The grammar's survey of lexical verbs illustrates the use of frequency statistics. For the purposes of classification, although many verbs have multiple meanings, verbs are divided into seven major semantic domains: activity verbs, communication verbs, mental verbs, causative verbs, verbs of simple occurrence, verbs of existence or relationship and aspectual verbs.

Table 5.1 gives the total number of common lexical verbs from each semantic domain (verbs that occur at least 50 times per million words).

Semantic domain  number of lexical verbs  % of all  common verbs
Activity         138                 49
Mental           53                 19
Communication      36                 13
Existence         22                  8
Occurrence        14                  5
Causative          10                  4
Aspectual          8                  3

The four registers use verbs from the seven semanticdomains quite differently.

Activityverbs are most common in conversation, fictionand news.

Mentalverbs are particularly common in conversationand fiction.

Existenceverbs are notably common in fiction, news and academic prose. In academic prose, existence verbs are almost as commonas activity verbs.

The lexical diversity within each semantic domain differs across registers (occurrences per million words).

Put, pay, bring, play, meet and run, for example, occur between + 200 and + 500 times per million words in conversation, fiction and news - but under 200 times in academic texts.

Describe occurs over 500 times suggest 200 in academic texts, but under 200 in all other varieties. Suggest occurs over 200 times in academic texts but under 200 times in all other varieties. Include and involve only appear significantly in news texts - contain, exist, indicate and represent only in academic texts.

The most interesting information about the use of the progressive comes under the heading "Lexical associations of progressive aspect." There we learn that:

Progressive aspect is most common in conversation and in  fiction. In conversation, most progressive verb phrases are in the present tense (c. 70%). In fiction, most progressive verb phrases are in the past tense (also c. 70%).

Corpus findings

Common verbs in the progressive aspect come fromdynamic verbs marking physical or communicativeactivities, stative verbs describing physical situations and mental, attitudinaland perceptual states.

Severalverbs have a strong lexical association with the progressive aspect, occurring over 50% of the timeas progressives. Most of the verbs refer to physical or communication activity.

Many verbs from the same semantic domains hardly everoccur in the progressive aspect.

Verbs referring to activities and physical events occurringover 50% of the time with the progressive, in conversation and/or fiction:

dance, drip head(for) march, pound, rain, stream, sweat.

40 other verbs that frequently occur with the progressive aspect (more than ten times per millionwords) e.g. bring, buy, carry, come, cry, do, drive, eat, give, go, laugh, leave, make, move, pay,play, run, take, walk, work.

rarely occurring with the progressive aspect (less than20% of the time):
arrest, attain, award, dissolve,find, frighten, invent, rule, shut, shrug, smash,suck, suspend, swallow, throw trap.

The contents of the book are as follows:

Section A: Introductory
Introduction: a corpus-based approach to Englishgrammar

Section B: Basic grammar: description and distribution
Word and phrase grammar
Clausegrammar

Section C: Key word classes and their phrases
Nouns,pronouns and the simple noun phrase
Verbs
Variation in verb phrase: tense, aspect, voice and modality
Adjectives and adverbs

Section D: More complex structures
Complex noun phrases
The form and function of complement clauses
Adverbials

Section E: Grammar in a wider perspective
Word order and related syntactic choices
The grammatical marking of stance
Lexical expressions in speech and writing
The grammar of conversation

Is the kind of information outlined above really important for the teacher in the classroom, the instructor teaching adults or the teacher trainer? In my opinion the answer is an emphatic: "Yes". The LGWSE provides the most widely researched, most up-to-date description of real English in use in existence, or likely to exist for years to come. It provides basic information that is irreplaceable in making decisions about the relative importance of which aspects of grammar to teach and concentrate on in the writing of text books, syllabuses and examinations.

Given the price, it is not possible to say: "rush out and  buy a copy". But it is unquestionably a book that should be in every institutional and public reference library in countries where the English language is taught.

Dennis Newson has been teaching at the University of Osnabrueck, Germany, for over 20 years. In the past he taught in schools and colleges in Ghana and Sierre Leone (West Africa), Qatar (The Arabian Gulf) and at the Technical University in Trondheim, Norway.

Dennis (Newson) Fachbereich 7 University of Osnabrueck GERMANY
http://www.dennisnewson.de