A Linguistic Description and Computer Program for Children's Speech

Geoffrey J. Turner, Bernard A. Mohan

• Publisher: Routledge
• Year: 2014
• eBook ISBN: 9781317933106
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

This book presents a framework for the linguistic analysis of speech and a computer program to process the results of this analysis. The model of description for the linguistic analysis is that known as 'scale-and-category' grammar. It is particularly suited for a study of how people use their language, and especially for a sociologically-oriented study of linguistic behaviour. By incorporating a concept of 'delicacy', it enables the investigator to vary, according to his particular interests, the amount of detail he enters into at various points in the description. The present authors have made use of this facility and discuss the special interests, sociological and psychological, that influenced their choice of detail.

The computer program analyzes the grammatical structures written in a linear notation. A second version has been written which allows easy modification to handle a variety of grammatical schemes, and the program has application to the processing of the analysis of sequential behaviour in general, especially where there are complex relations between the units analyzed.

PART ONE

A LINGUISTIC APPROACH TO CHILDREN’S SPEECH

Chapter 1

AIMS AND APPROACH

In a paper called ‘Syntax and the Consumer’, Halliday (1964) puts forward the view that ‘different coexisting models in linguistics may best be regarded as appropriate to different aims, rather than as competing contenders for the same goal’. The aim of the grammar developed by Halliday and his colleagues is ‘to show the patterns inherent in the linguistic performance of the native speaker’. The description involves ‘a characterisation of the special features, including statistical properties, of varieties of the language used for different purposes (“registers”), and the comparison of individual texts, spoken and written, including literary texts. This in turn is seen as a linguistic contribution towards certain further aims, such as literary scholarship, native and foreign language teaching, educational research, sociological and anthropological studies and medical applications.’ As Halliday points out: ‘The interest is focused not on what the native speaker knows of his language but rather on what he does with it; one might perhaps say that the orientation is primarily textual and, in the widest sense, sociological.’

Our own particular aim was to make a statistical comparison of linguistic features in the speech of a sample of five-year-old working-class and middle-class children. Our main interest was in isolating social class differences in usage, those social class differences which Bernstein’s theory of elaborated and restricted codes (e.g. Bernstein, 1965) implies are of educational consequence. It was necessary to take fairly large samples of children (450 children in all) in order to be able to say anything useful about the influence of social class. Furthermore, it was necessary to obtain speech from each subject in a variety of situations, in order to obtain some kind of check on the influence of the situation and to see whether particular situations favoured one or other of the social classes. It can be seen that there is a basic compatibility between our aims and those for which Halliday regards his type of grammar as appropriate.

Marshall and Wales (1966). in a critique of Halliday’s paper, argue that the study of linguistic performance should be preceded by a study of linguistic competence, competence being the tacit knowledge of the language user. They write: ‘It has always been accepted that, once the outlines of a competence model are fairly clear, we can then explore the manner in which this competence is realised in performance (Chomsky and Miller, 1963).’ If we carry on performance studies without a competence theory, ‘we are left without a norm against which to evaluate the results of such performance studies’. For a study of certain aspects of language performance a prior study of competence would indeed be necessary or at least highly desirable, for example, for a full examination of ‘maze’ behaviour, that is, false starts, substitutions etc., of grammatical deviations, and so forth. However, for most of our work it is not necessary to attempt to characterise the underlying competence of the speaker. Bernstein (1965) writes: ‘The code which the linguist invents in order to explain speech events is capable of generating n number of speech codes.’ The language code is a set of options. It is ‘a set of rules to which all speech codes must comply, but which speech codes are generated is a function of the system of social relations’. The form of the social relation regulates the options which the speakers take up. Speech codes are distinguished in terms of the relative frequencies with which particular options available in the language code are taken up. As Bernstein (1964b) puts it: ‘Speech … is constrained by the circumstances of the moment, by the dictate of the local social relation and so symbolises not what can be done, but what is done with different degrees of frequency.’ The social class differences in speech and writing that have been found by Bernstein (1962b) and Lawton (1963, 1964) have been differences in relative frequency. There has been no suggestion in the above studies that there is any fundamental difference in the tacit knowledge of the middle-class and working-class language user. What is of primary importance to Bernstein is not the difference between ‘competence’ and ‘performance’ but the difference between performances that have been influenced by different social relations.

We might point out, theoretical considerations apart, that it would have been an enormously difficult task, in terms of time and manpower, to have attempted a systematic study of underlying competence in samples of subjects large enough to be representative of the different social classes. The difficulties of obtaining evidence about underlying competence may be gathered from this quotation from Chomsky (1964): ‘If anything farreaching and real is to be discovered about the actual grammar of the child, then rather devious kinds of observation of his performance, his abilities and his comprehension in many different kinds of circumstances will have to be obtained, so that a variety of evidence may be brought to bear on the attempt to determine what is in fact his underlying linguistic competence at each stage of development.’ It was not feasible for us to approach large samples of subjects in this way.

Almost all of the linguistic description to be presented in this Monograph consists of a modified and much simplified version of the description made and at present being developed by Professor M. A. K. Halliday, University College, London. It should be pointed out that the categories for the present description were worked out in 1965. For practical purposes it was necessary to ‘freeze’ the description at that time; as we were coding a large body of data there was not time available for incorporating revised or new categories into the description and for re-working the data once the main analysis had begun. Considerable changes have been made in Halliday’s description since that time. Some suggestion of these changes may be gauged from the name that is now used to refer to the description, ‘systemic’, rather than ‘scale-and-category’ (relating to the scales ‘rank’, ‘exponence’. ‘delicacy’ and later ‘depth’ and the categories ‘unit’, ‘structure’, ‘class’ and ‘system’) or ‘system-structure’. Briefly, much greater emphasis has been placed on paradigmatic relations (the realm of the category ‘system’) and much less emphasis on syntagmatic relations (‘structure’), the latter now being fully derived from the former. Our work came too early to draw fully on the new forms of description. So whilst we describe systems, not all the structural representations we give are related to underlying systemic choices. For this reason we shall refer to the grammatical description we used as ‘scale-and-category’ rather than ‘system-structure’ or ‘systemic’ description. Although our main purpose is to describe the grammatical categories that we used, we shall mention, wherever it seems appropriate, the more recent categor...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Original Title Page
• Original Copyright Page
• Table of Contents
• Acknowledgements
• Introduction
• Part One A Linguistic Approach to Children's Speech
• Part Two A Computer Program for Assisting the Linguistic Analysis of Children's Speech
• Bibliography