Methods of Text and Discourse Analysis

In Search of Meaning

Stefan Titscher, Michael Meyer, Ruth Wodak, Eva Vetter

• Publisher: SAGE Publications Ltd
• Year: 2000
• Print ISBN: 9780761964834, 9780761964827
• eBook ISBN: 9781446232842
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

?This volume is the most comprehensive overview to date of sociologically orientated approaches to text and discourse analysis and is worth reading even for those who are interested only in purely linguistiv approaches to text and discourse. Its main merit, I think, is that it intorduces approaches which up to now have hardley been admitted into the universe of scientific discourse? - Discourse Studies

Methods of Text and Discourse Analysis provides the most comprehensive overview currently available of linguistic and sociological approaches to text and discourse analysis.

Among the 10 linguistic and sociological models surveyed in this book some of the more important are Grounded Theory, Content Analysis, Conversation Analysis and Critical Discourse Analysis. The book presents each approach according to a standardised format, which allows for direct systematic comparisons. The fully annotated lists of sources provide readers with an additional means of evaluation of the competing analytical methods. Interdisciplinary and international in its aims, Methods of Text and Discourse Analysis suggests the benefits both linguists and sociologists will derive from a more intimate knowledge of each others? methods and procedures.

PART ONE

---

Methods and Texts

CHAPTER 1

---

ON SOCIAL-SCIENTIFIC METHODS OF TEXT ANALYSIS

Before we embark on the critical and comparative presentation of individual methods (see Part 2), it seems both necessary and wise to clarify, from an interdisciplinary viewpoint, certain basic concepts that relate to our theme of methods of text analysis. Even the simple question of what a text is permits no easy answer, but leads us immediately into the highly varied theoretical approaches of text linguistics and discourse analysis where even the concepts of text and discourse are used in a multiplicity of ways and where they are anchored in very different research traditions. A similar situation is found with the term method, and with our understanding of what social scientists actually do when they analyse texts: what procedures, rules and instruments do they use for this task? The two introductory chapters should be seen as a basis for all the expositions that follow; they are intended as an illustration of the interdisciplinary nature of our project.

Methods are not isolated in space, but are either explicitly or implicitly related to theoretical assumptions and structures. Quite often methods are applied without due reflection and without taking account of such theoretical roots. Our map of theories and methods aims to assist in a reflective approach by creating an awareness of the interconnection of traditions and the proximity or distance between the individual methods. It may be seen what theoretical preconditions are associated with the application of a particular method.

The final part of the book is devoted to a bibliometric comparison of methods, and a comparison of the frequencies of citation and reference in various literature databases. We are of course aware that this kind of comparison cannot pass judgement on the quality of a particular method. It merely reflects the extent of its diffusion within the ‘scientific community’: to what degree a method has gained acceptance, has been adopted and applied. Together with our comparison and the criteria to be discussed, the frequency of citation provides an additional perspective which may also be significant for the selection of a particular method.

1.1 WAYS OF ACCESSING DATA

---

The term ‘method’¹ normally denotes research pathways: from the researcher’s own standpoint or from point A (theoretical assumptions), another point B (observation) is reached by choosing a pathway which permits observations and facilitates the collection of experiences. If one proceeds systematically wrong turnings are avoidable. Methodical procedure can, like Ariadne’s thread, guarantee the researcher a safe route back. By giving them experience along the way, methodical procedure may also assist those investigators who look over their shoulders and see their starting point differently, even deciding not to go back but to find other more interesting starting points. No matter how the investigative journey may turn out, methodical procedure will make it easier to record findings and to compile reports of experience.

In so-called empirical social research a distinction is made between elicitation and evaluation methods: between ways of collecting data (in the laboratory or by fieldwork) and procedures that have been developed for the analysis of collected data. Methodical procedures for the collection of data organize observation; evaluation methods regulate the transformation of data into information and further restrict the opportunities for inference and interpretation.² In the context of some empirical research, fieldwork implies those stages in a task that permit the collection of data in situ. In most cases this requires a direct discussion between the researchers and the carriers or representatives of the patterns and structures that are sought or being investigated.

Almost all of the methods that are brought together in this book may be ascribed to the area of social research³ in so far as their results are obtained from tangible reality and this reality is acknowledged as a test case of their (theoretical) conclusions. None of these methods is a field-research method in the sense of observing actual behaviour of participants in a meeting which one wishes to analyse, since text analyses can begin only after the material has been collected. Although text analyses may precede fieldwork – if for instance answers from interviews are being analysed – they may also take place without this preparatory work, for example if generally accessible texts are being investigated and no special phase of field data collection is required.

The routes to be followed in empirical research will be decided initially by the general research questions – and these are, to a certain extent, determined by the theoretical approach one has decided to follow. If one believes it is vital to investigate attitudes, then questioning will seem more suitable than observation. If concrete non-verbal modes of behaviour are being examined, then observation is indispensable. If one wishes to study behaviour from some period of time in the past, one must make use of texts already elicited or documented in writing. If everyday behaviour is of primary interest, laboratory conditions are excluded. These simple examples should demonstrate that the initial framework is determined by the research questions to the extent that, while certain basic limitations derive from them, on the other hand they help in the selection of economic ways of clarifying the basic research problem. Not every mode of procedure is equally suited to handling every question. Data do not always need to be collected: use can often be made of materials that are already available.

Every piece of research begins with theoretical assumptions: ‘The field worker cannot begin to describe any social event without some specification of his scientific theory, i.e. his theory of objects, his model of the actor, or the kind of social order presupposed’ (Cicourel 1964: 51).

Using these models the research question can be clarified and from this can be derived a particular ‘research strategy’. This term incorporates those decisions that need to be taken when planning a piece of research before the concrete stages of empirical work can be derived.⁴

First, at a fundamental level it depends on what the research aims to discover: (a) does one wish to explore and find explanations for the facts to be investigated; (b) are theoretical concepts or hypotheses to be tested; or (c) is it a matter of finding a description of a particular field or defined population?

Investigative processes (as referred to in (a) above) require heuristic or interpretative procedures, and have as their goal the clarification of ideas or concepts (consider the numerous investigations on the theme of ‘politeness’) and/or the development of theoretical assumptions (such as action-determining structures or historically and socially conditioned patterns of meaning). Such undertakings are often organized in the form of individual case studies. One example of this is the study by Bensman & Gerver (1973), in which deviant behaviour in a production department is investigated and interpreted, on the basis of observer participation, as an important element in the maintenance of the social system. Exploratory studies – in the traditional view⁵ – are predominantly for the purpose of preparing research that tests hypotheses. If theoretical assumptions are to be tested (as suggested in point (b) above), then experimental or quasi-experimental research designs are required,⁶ that is to say, research instruments which can be used to vary systematically independent variables in order to assess their possible influence on dependent variables (see in this connection Kleining’s (1994) suggestions for experiments with texts). For the purpose of description (see point (c) above), it is, on the one hand, typical to use the types of study which in extreme cases may be conducted by counting distinctive features in a sampling and estimating procedure. Examples of this are public opinion surveys and – with reference to linguistic analysis – the assembly and evaluation of a ‘Wendekorpus’ (‘corpus of change’) to investigate the lexicon of the German language in the years 1989–90. On the other hand, description may have additional hidden aims and seek to describe the behaviour of actors in particular social fields. Such investigations do not aim to account for quantifiable distributions but rather to document spheres of existence, and sometimes also to go beyond description and explain the rules which determine them.

Secondly, at the level of the connection between the research process and the affected or possible consumers, one must enquire about the form of the contact between investigators and those who will provide the necessary data (responses, documents, etc.). The various possible positions in this respect may be located between the two extremes of greatest possible involvement and total withdrawal of the investigators. Typical of the first approach would be projects that feel obliged to use action-research, while the second approach is seen in non-reactive methods, experimental designs and all standardized procedures.

In connection with this, thirdly, some commitment must be made about the approach of the investigators. If they approach their field of research ‘openly’, this implies dispensing with standardized instruments or predetermined categories. It also implies a willingness to distance themselves from any prior understanding and thereby to change the mode of procedure in the course of the investigation. In contrast, the epistemological approach would require the setting up and testing of hypotheses, since procedural changes in the course of research can only lead to results that are difficult to verify.⁷

Fourthly and finally, under decisions concerning research strategy, there is the question whether the piece of research should provide a snapshot (generalizable over time) or whether it should investigate changes. The first type is considerably more frequent and implies data collection at a particular point in time or during a particular phase of the investigation. In the second case, the investigator must opt for one of the various types of sequential or ‘panel’ procedures⁸ and either collect data on a number of occasions or analyse material from different periods.

The preliminary decisions outlined here narrow the spectrum of usable procedures but perhaps force a combination of particular modes of research. It may therefore seem necessary – or at least wise – to examine the content of texts first and then to enquire about their effect on recipients. Whatever the case, such questions should always be clarified before a decision is reached about the method which will actually be used in a particular project. An empirical method should be understood as a set of procedural rules which has available a set of principles governing how investigators should gather experiences and how they should organize their observations if they wish to proceed scientifically. Proceeding scientifically, in this respect, is understood as systematic, rule-governed work. Adhering to schematized modes of procedure makes it possible for investigators to remain uninvolved. It enables them to maintain the required distance from the phenomena (action-field, behavioural structures) that are being studied and, again despite appropriate sensitivity towards the field of research, to adopt the role of a neutral (as opposed to partisan) observer. Only in this way can scientists maintain their own ‘meaning and relevance structures’, which they use first to inform themselves about the object of study and to observe it, and then to interpret and classify.⁹

Among the methods of data collection most frequently used we find questioning, observation¹⁰ or sociometry. The more highly elaborated a method is, the more differentiated are the various procedures with which data collection can be carried out. Distinctions can therefore be made between written/oral, individual/group questioning, participant or non-participant, and overt and covert observation. (Most of these procedures or fieldwork techniques may be applied in more standardized or less standardized variants.) Data for the analysis of group structures can be collected by means of sociometric questioning or with the help of a living sociogram. In objective hermeneutics (see Chapter 14), designated by its creator Oevermann et al. (1979) as a synthetic approach, there are also different procedures (such as sequential and detailed analysis, that is to say interpretative procedures) that can be applied according to the research goal.

Methods are therefore families of related procedures whose relationship is determined by one or more common features: by a common theoretical base (as in the case of objective hermeneutics), by their relationship to the object of study (sociometry is used in the analysis of group relations, content analysis for investigating the contents of communication), by their efficiency and limitations. For instance, observation techniques may also be used to approach non-verbal behaviour, but only behaviour that is currently observable or recorded on film. Questioning can be used to collect data on attitudes and intentions, ‘internal’ behaviour, or past activities. In many cases the individual methods make use of differently standardized procedures. Through a predetermined structuring of the collection process an attempt is made to minimize the interpersonal influences of those collecting the data: interviewers have predetermined questionnaires, observers have a standardized framework for classifying observations, content analysts use a system of categories to classify textual elements as uniformly as possible. In addition, highly structured modes of data collection have economic advantages in that they simplify the subsequent evaluation procedures. Whether standardized procedures can be used, however, depends to a considerable extent on prior knowledge, on the subject area and on the research goal. Any predetermination of categories presupposes knowledge of events that may possibly occur (for example, textual contents) or of reactions (such as answers to questions). The field of investigation and the subject area (such as daily rituals in spontaneously occurring face-to-face interactions¹¹) can render pointless standardized modes of procedure. The development of a framework of answers, observations or categories requires clear theoretical assumptions. For this reason alone exploratory investigations (see above) have little or no structure.

For each of the procedures mentioned there is a range of more or less widely accepted rules that researchers are obliged to follow. If they do not follow the rules they may be accused of not operating ‘cleanly’. In interviews, for example, the interviewees should not be over-questioned (in content or style of questions), the ordering of the questions should avoid ‘halo effects’, and so on. In sociometric questioning the questions selected should be positively formulated and not ask for negation, and so on. As a final example, in content a...

Table of contents

• Cover Page
• Title
• Copyright
• Contents
• Acknowledgements
• Introduction
• PART I METHODS AND TEXTS
• PART 2 OUTLINE OF METHODS OFTEXT ANALYSIS
• PART 3 OVERVIEW AND COMPARISON
• Glossary
• Appendix: Publications and Keywords for Bibliometry
• Bibliography
• Index