banner
SIL International Home

SIL Electronic Book Reviews 2003-023

URL: http://www.sil.org/silebr/silebr2003-023

Evaluation in text: Authorial stance and the construction of discourses

Edited by Susan Hunston and Geoff Thompson

New York: Oxford University Press, 2001. Pp. 240. paper $26.00. ISBN 0198299869.

Reviewed by E. Lou Hohulin

SIL International

To introduce this review, I cannot do better than to quote the aim of the editors found in the Preface:

Our aim in selecting papers was to represent as wide a range of approaches as possible to the notion of evaluation…. We wanted to present a multiform view of language: as grammar, as lexis and as text, and we wanted to highlight the role of language in reflecting and constructing ideas and opinions.

Although there is a wide range of approaches to the notion of evaluation among these papers, the compilation is unified in the following ways:

In spite of the fact that the authors seem to be narrowly focused on a single topic in text analysis and a particular theory, they relate the notion of evaluation to many topics that are of concern to all analysts of text and discourse.

Chapter 1. “Evaluation: An Introduction” by Susan Hunston and Geoff Thompson. The first paper seems to describe the framework for the other papers. For that reason, the review of this paper will be more thorough. First of all, the two authors define evaluation (p. 5):

For us…evaluation is the broad cover term for the expression of the speaker or writer’s attitude or stance towards, a viewpoint on, or feelings about the entities or propositions that he or she is talking about…

According to them, evaluation in discourse performs three functions: (1) It expresses the speaker’s or writer’s opinion, and in doing so it reflects the value system of that person and their community; (2) it constructs and maintains relations between the speaker or writer and hearer or reader; and (3) it organizes the discourse.

Explaining the “expressing opinion” function, the authors say that every act of evaluation expresses a communal value-system, and every act of evaluation continues to build that value system. In turn that value system is a component of the ideology that lies behind every text. One example given to illustrate “expressing an opinion” within a communal value-system is an excerpt from a text in which a wife describes her husband as a “workaholic,” and then attaches negative events in their lives to that evaluation.

The authors explain the “maintaining relations” function of evaluation by discussing the three main areas that have been studied in understanding the building and maintaining of relations between writer and reader. Those three main areas are (1) manipulation, (2) hedging, and (3) politeness. When a writer is using manipulation or persuasion, the authors claim that it takes a conscious effort of detachment on the part of the reader not to identify with the writer’s point of view or the ideology that underlies it. They also indicate that the less obtrusively an evaluation is placed in a clause, the more likely it is to successfully manipulate the reader. (Hedging, they believe, may be used to moderate claims or to effect politeness in maintaining writer-reader relations.)

Note that Hunston and Thompson have the extralinguistic situation of a text in mind in regard to the first two functions of evaluation. That is not surprising. However, the final point made about the function of evaluation was the most interesting. In the section, “Organization of discourse,” the authors discuss the notion that evaluations occur at boundary points in discourse, and therefore indicate its organization. They describe and illustrate with a textual excerpt to show how the study of evaluation along with clause relations will reveal the organization of discourse saying “Evaluation which both organizes the discourse and indicates its significance might be said to tell the reader the ‘point’ of the discourse” (p. 12).

Hunston and Thompson have three more sections in their paper. In the section, “How do we recognize evaluation,” they say that in order to recognize evaluation, it helps to understand that conceptually it is: (1) comparative, (2) subjective, (3) value laden.

The task then is to identify those forms and structures that signal these concepts. Examples are given under the headings of lexis, grammar, and text. Examples of evaluative lexical units are adjectives (splendid, terrible), adverbs (happily, unfortunately), nouns (success, failure), and verbs (succeed, fail). To explain evaluative grammatical structure, Labov is quoted, who claimed that any departures from basic narrative syntax has marked evaluative force. Under the heading “text” the authors give an example of two paragraphs with a relationship of evaluation; also a paragraph that describes a situation, a series of events in a narrative episode, followed by one that expresses an evaluation of the situation.

The point being made in these sections is that there are comparative, subjective, value-laden evaluations at all levels of linguistic structure. The paper ends with a section in which four parameters of evaluation are identified: (1) good-bad, (2) certainty, (3) expectedness, and (4) importance. But they say:

We would wish to argue, however, that evaluation is essentially one phenomenon rather than several, and that the most basic parameter, the one to which the others can be seen to relate, is the good-bad parameter. (p. 25)

Chapter 2. “Persuasive Rhetoric in Linguistics: A Stylistic Study of Some Features of the Language of Noam Chomsky” by Michael Hoey. In this paper, Hoey presents four answers to the question of why transformational-generative grammar took such a strong hold of the linguistic world. His fourth answer is the one that relates to this compilation of papers, reflecting a particular view of the relation between language, knowledge, and the world (i.e., the study of language and ideology) as these papers do.

In this chapter I seek to demonstrate that Chomsky is a skilled rhetorician whose chief rhetorical device is to make it difficult for a reader to support an alternative or opposing view to Chomsky’s own without looking foolish, a claim first made in outline form by Botha (1973).

According to Botha, the “persuasive power” of transformationalists’ arguments was found in eight strategies of persuasion. Hoey chose just two to demonstrate his main point about Chomsky’s tactics:

  1. Inflate the apparent merit of your own arguments by emphatically calling them “striking,” “powerful,” “strong,” “forceful,” “convincing,” and so on. Deflate your opponent’s arguments by means of the corresponding antonyms.

  2. Warn your opponent that if he did not accept your theoretical viewpoint, your data, or your arguments, then he would be guilty of irrationality and/or then your common field, as a field of research, would be destroyed. (p. 30)

Hoey quotes passages from Chomsky’s writings, analyzes them, and then describes what was found:

One of the first characteristics of these passages to strike the reader is the high traffic in evaluations of various sorts. The clause relation of Situation-Evaluation can be seen as the most fundamental relation in discourse organization; the use of evaluations does not therefore distinguish Chomsky’s style from those of other writers. What does make Chomsky’s use of evaluations significant are (i) the quantity of them; (ii) the interweaving of these with the situational elements; and (iii) the presentation of them without basis. (p. 32)

In the conclusion of this chapter he states:

I have examined two factors, closely related, that enter into Chomsky’s success as a rhetorician. Both had been delineated by Botha (1973), though neither was illustrated or described in any detail. It has been found that Chomsky uses evaluation both as a running supportive commentary on his own arguments and as a device for cowing opposition. In connection with the latter point I have argued that an absence of basis to back his more controversial evaluations is disguised by means of embedding and references to unspecified work. While these features are intermittent in all his later writings, they are almost wholly absent in Syntactic Structures; it is clear therefore that Chomsky’s initial influence was not dependent on them. It is all the more to be regretted therefore that he felt it necessary to resort to such measures in his subsequent writings, particularly as his influence has been felt as much in matters of debating style as in content. (p. 37)

Hoey’s paper was exceptionally interesting. His analysis of Chomsky’s rhetoric is very convincing and helped me understand my own defensive reactions to much of Chomsky’s writings.

Chapter 3. “Corpus-Based Analysis of Evaluative Lexis” by Joanna Channell. Channel’s title gives a clue to the main claim in her paper, i.e., the analysis of evaluation in text does not need to be based on “linguistic intuitions”; instead it can be based on what she calls the “systematic observation of naturally occurring data.” Her own work in the analysis of the evaluative function of a word or expression is derived from concordanced examples—data that was drawn from the 200 million words of spoken and written English in the Bank of English corpus, used in the preparation of Collins Cobuild publications.

Mainly two techniques are used: (1) collocational analysis that used computational techniques to identify words which typically co-occur with the lexical item under investigation, and (2) an adaptation of conversation analysis to analyze the particular effect the meaning or function of the item had on the following context.

Channell’s work on fat, using data from Collins Cobuild English Dictionary (1995), illustrates her techniques as well as her conclusions about the “evaluative” function of the word fat in context. She says, “In a British context, it is clear that the word, fat, because of the learned prejudices of British culture in regard to body weight, is neither a neutral descriptor, nor a compliment.” Eighteen instances in context are shown and then she gives her analysis:

  1. fat collocates with words which are negative descriptors (old, ass, bald, slob, crafty, pompous);

  2. that to be fat is undesirable is shown by structures which allude to not being fat (end up feeling fat, it is a wonder you don’t get fat, he saw I’m no longer fat);

  3. fatness is humorous (best-dressed fat woman in town; your mom is so fat, she … ).

When the attribute is applied to animals, the evaluation is positive (“dozens of large, fat sows surrounded by their little piglets—oh, how cute! …my duck, Jeffrey, sleek and fat”) (p. 43).

Channell’s paper has implications for theoreticians. Convincing evidence is given from her research on “evaluative” function in text that corpus-based analysis is likely to introduce greater rigor into theory testing than using isolated data obtained from deductive introspection.

Chapter 4. “Adverbial Marking of Stance in Speech and Writing” by Susan Conrad and Douglas Biber. Conrad and Biber also apply corpus-based methods to study the ways in which speakers and writers use adverbials to mark their personal “stance.” The term is defined as follows:

A cover term for the expression of personal feelings and assessments in three major domains:

1. epistemic stance, commenting on the certainty (or doubt), reliability, or limitations of a proposition, including comments on the source of information;

2. attitudinal stance, conveying the speaker’s attitudes, feelings, or value judgements;

3. style stance, describing the manner in which the information is presented. (p. 57)

Subclasses are defined in the major domains. For example, epistemic stance has five adverbial classes: doubt/certainty (certainly, perhaps), actuality (actually, in fact), imprecision (like, sort of), source of information (according to), limitation/perspective (generally).

The authors restrict their study of stance to grammatical devices used to frame a proposition: “Adverbial stance markers can be characterized with respect to three major parameters: (1) semantic class, (2) grammatical realization, (3) placement in the clause” (p. 58). Data from three registers of spoken and written English: conversation, academic prose, and news reports are analyzed. Research included frequency counts on the adverbs, prepositional phrases, finite subordinate clauses, and their placement in clauses occurring in their data; these counts are graphically displayed. The parameters of register and number of occurrences per 100,000 words are used in the graphs. Finally, the authors’ conclusions about adverbial marking of stance in texts are described. For example:

there are twice as many stance adverbials in conversation as in written registers. This distribution fits well with the expectation that conversational partners are personally involved with their messages and therefore commonly frame propositions with their personal attitudes and assessments. (p. 64)

Conrad and Biber are convinced that their paper and research shows the valuable contribution that corpus-based analyses can make to the adverbial marking of stance in English and that their findings show the importance of register to analyses of stance marking. It is likely that linguists could further this type of corpus-based research by doing some cross-language comparisons of the marking of stance related to register in languages other than Indo-European ones.

Chapter 5. “A Local Grammar of Evaluation” by Susan Hunston and John Sinclair. In this paper, Hunston and Sinclair argue that it has long been known that every language has phenomena that are not handled by a general grammar, and yet have recognizable patterns. They, as well as other linguists, believe that “these structures are neglected because they do not exhibit interesting generalities—many of them are social conventions reflected in the language, highly specific and unpredictable from any general perspective” (p. 76).

Hunston and Sinclair begin their argument by discussing how electronic text corpora and software are now being used to check how reliable grammars are at “providing a categorization of anything that arises in the actual business of speaking and writing” (p. 75). The authors continue:

Even when using the latest parsers, corpus grammarians continue to find that, once the software has done its best with open text, there is still a lot left over. By this they do not mean the errors and problems which can in principle be resolved by further development of the system, but the observation that there remain segments of text which will never be adequately described by whatever parser is developed for the main body of the text. (p. 76)

Hunston and Sinclair say that given the fact that evaluation is, as consistently argued in this volume, a central function of language, and since general grammars have been unable to present a coherent account of it, it is clearly an ideal area to try out the concept of local grammars. In the remainder of the paper they seek to answer the question “How do we recognize evaluation, and describe it in a way that brings out its discourse function?” To quote, “There are two aspects to the categorization of language units—their constituency and their structural position” (p. 80). In their research, they focussed on the constituency aspect of evaluation, and that they “offer a description of sentences that are likely to be interpreted as evaluations without recourse to evidence of their structural roles” (p. 80). They identify and parse evaluations in text corpora and define their patterns. For example, one pattern identified is:

(1)  + Link verb + adjective group + clause

Following are some examples taken from table 5.3 (p. 85):

    Evaluative category Thing evaluated
it link verb adjective group finite or nonfinite clause
It was certain that he was much to blame.
It was surprising how many on that course had disabled children
It seemed important to trust her judgment

Six evaluative patterns were defined and illustrated with each pattern displayed in parsed sentences on tables.

This paper illustrates some excellent techniques for parsing sentences to reveal syntactic patterns that would help to more accurately describe the use of evaluation in text whether or not it is agreed that there is a need for “local” grammars.

Chapter 6. “Evaluating Evaluation in Narrative” by Martin Cortazzi and Lixian Jin. In the introduction, it is stated, “Evaluation is a major criterion of narrative,” and that evaluation is held to be the key to narrative because it is through evaluation that speakers show how they intend the narrative to be understood and what the point is. Three layers of evaluation in narrative are proposed and three prepositions are used to classify these layers: in, of, and through.

The authors briefly describe the six-part structure of narrative text that has been developed by Labov and his co-workers. Evaluation, the sixth element of the structure is a salient feature of the model. Also, according to them, Labov and Waletsky (1967) propose evaluation as one of the social functions of narratives:

The model proposes that there are two social functions of narratives: a referential function, which gives the audience information through the recapitulation of the teller’s experience, and an evaluative function, which communicates the meaning of the narrative by establishing some point of personal involvement. (p. 105)

Two types of evaluation device are discussed: external and internal. In external evaluation the teller interrupts his recounting in order to tell his listeners what the point is. Internal evaluation marks the evaluation grammatically, semantically, or prosodically. Evaluations, according to Cortazzi and Jin, may appear anywhere in a narrative and be realized by any level of linguistic structure.

“Evaluation of discourse” may vary cross culturally, and an example from research done by Deborah Tannen is given in comparing the stories of Greek and American women told after they had viewed a silent film:

The Greeks focused on personal involvement; they were concerned with characters’ motives and offered many judgments. The Americans, in contrast, focused on context and gave detailed objective reports. It is not difficult to imagine how the two groups might evaluate each others’ stories. (p. 111)

It is obvious that this type of evaluation has implications for all who communicate cross-culturally.

“Evaluation through discourse” is described as the evaluation of tellers, hearers, or their situations through the way the narration is given. “In educational settings, for example, children have often been evaluated for fluency, confidence, or their learning of content knowledge through their ability to tell stories” (p. 114). Other examples given of evaluation through discourse are in therapeutic settings, interviews, and courtrooms.

The authors conclude: “Evaluation in narrative, of narrative, and through narrative is clearly complex” (p. 120). Then, after a brief summary of the layers of evaluation and the linguistic and socio-cultural elements that may be interwoven across the layers, they say: “Evaluating evaluation benefits from multi-disciplinary views. At least, that is our evaluation” (p. 121). I recently used this chapter as a reading during a text and discourse analysis workshop in India. Several of the participants found the 3-part evaluation framework helpful in the analysis and interpretation of their texts.

Chapter 7. “Evaluation and Organization in Text: The Structuring Role of Evaluative Disjuncts” by Geoff Thompson and Jianglin Zhou. In this paper, it is argued that both coherence and cohesion depend on evaluation. Thompson and Zhou are particularly interested in adverbial disjuncts (p. 124):

We wish to argue that the function of disjuncts can often only be fully understood if they are seen as serving to show how elements of a text are linked, and that their interpersonal function is inextricably combined with a textual, cohesive function. Our basic view of these items is therefore that in many—though by no means all—cases they function as “conjuncts with attitude.”

Concessive, expectancy, hypothetical-real, and alternative are four of the cohesive relations that Thompson and Zhou illustrate with textual excerpts that use words such as admittedly, true, certainly, and plainly (concessive), surprisingly, curiously, unfortunately, sadly (expectancy), ostensibly, actually, in fact, (hypothetical-real), and maybe, perhaps (alternative).

Their paper concludes:

that coherence in text can only be adequately understood if the concept of propositional coherence is complemented by that of evaluative coherence, and that, amongst other things, this involves recognition of the conjunctive function of disjuncts. (p. 139)

Chapter 8. “Beyond Exchange: Appraisal Systems in English” by J. R. Martin. Martin has examined and describes the system of choices in an evaluative lexis that expresses a speaker's or writer’s opinion in text. The area of meaning described is called appraisal. According to Martin, there are subsystems of appraisal. Affect deals with the expression of emotion, judgment deals with moral assessments of behavior, and appreciation deals with aesthetic assessments. His evaluative system is linked with other meaning systems. For example, evaluation to the concept of “engagement”; this is the indication of a speaker’s degree of commitment to the appraisal expressed. Martin also links the appraisal system to the concept of redundancy, i.e., by using other parts of the lexicogrammatical system that cover the same semantic area, a speaker or writer may reinforce their evaluative appraisal.

The editors, Hunston and Thompson, make a particularly interesting observation about Martin’s paper: … “appreciative meanings (the film was very sad) are close in semantic terms to mental processes (the film moved me to tears) …’ (p. 142). They, then, claim that Martin has postulated an important distinction between inscribed and evoked appraisal:

Inscribed appraisal is explicitly expressed in the text (a bright kid, a vicious kid), whereas with evoked appraisal an evaluative response is projected by reference to events or states that are conventionally prized (a kid who reads a lot) or frowned on (a kid who tears the wings off butterflies). (p. 142)

This is a particularly telling point because a study of the distinction between inscribed and evoked appraisal would aid us in understanding those things in a text that trigger inferences and hearers’ interpretation of text.

Martin has excellent data charts showing some helpful classifications of the English evaluative lexis. On table 8.4, “A framework for analyzing judgement in English’, Martin classifies the English judgment vocabulary on the basis of two parameters: social esteem (normality—“is he or she special,” capacity—“is he or she capable,” tenacity—“is he or she reliable, dependable”), and social sanction (veracity—“is he or she honest” and propriety—“is he or she beyond reproach”). I intend to use his charts for doing some comparative work on the evaluative lexical items used in texts of the languages I am studying.

Chapter 9. “Evaluation and the Planes of Discourse: Status and Value in Persuasive Texts” by Susan Hunston. In the research underlying this paper, Hunston studied texts belonging to different genres, but all of them had a persuasive function—“Persuasive texts have been chosen because, in them, evaluation is important to the purpose of the text” (p. 177).

H uses Sinclair’s model of “planes of discourse” and develops the distinction between entities and propositions as the objects of evaluation. When considering the interactive plane of discourse, the text is considered to be constructed so as to reflect the ongoing interaction between writer and reader. On this plane, evaluation relates to the function of the proposition in the text. When considering the autonomous plane discourse, the text is studied in terms of its content rather than of its construction and evaluation relates to the expression of the writer’s perception of the world. Hunston explains that the distinction between “evaluating a part of the discourse” itself, and “evaluating something else” corresponds to Sinclair’s two planes.

Hunston makes another helpful distinction in exploring the complexity in evaluation. That distinction is between attribution and averral:

If a piece of language—spoken, written, or thought—is attributed, it is presented as deriving from someone other than the writer. If a piece of language is averred, the writer him or herself speaks. The distinction between averral and attribution is important to the study of evaluation, because it can be used to position the reader to attach more or less credence to the various pieces of information. (p. 178)

Within the interactive plane, Hunston differentiates between status and value—“Every statement is understood as having a particular orientation with respect to the world outside the text’ (p. 185)—and proposes that every clause or statement in a text carries with it a particular status:

On the interactive plane, each statement is of a particular type (e.g. a fact or an assessment) and has a source (e.g. averred by the writer, or attributed to someone else): these determine its status. At the same time many of the statements are given a positive or negative value (e.g. that it is supported by evidence, or that it is not true). (p. 177)

Hunston explores in depth the ways in which the status of a statement constrains the kind of value that it can be given. Statements are classified as either world-creating (assumptions, hypothetical, recommendations) or world-reflecting (fact/event, interpretation/hypothesis, assessment). Huston illustrates each type with excerpts from texts.

The role of evaluation in discourse is summarized as follows:

Evaluation is important to discourse for two reasons: it plays a vital role in constructing the ideological basis of a text, thereby locating writer and reader in an ideological space; and it plays a vital role in organizing a text. Evaluation on both the autonomous and the interactive planes take part in both functions. (p. 205)

Summary. This book is highly recommended to anyone doing text or discourse research for a number of reasons:

  1. Every author describes the methodology and/or techniques used in research.

  2. Every author used a text corpus, and all examples are taken from natural texts; also, the lexical units and syntactic structures being discussed are shown within an adequate context so that readers can determine whether there is substantial evidence for the claims made about evaluation, authorial stance, and the construction of discourse.

  3. Many of the papers have excellent graphic illustrations, i.e., charts and figures that could be used for comparative work.

References

Botha, R. P. 1973. The justification of linguistic hypotheses. The Hague: Mouton.

Labov, W., and Waletsky, J. 1967. Narrative analysis: oral versions of personal experience. In J. Helm (ed.), Essay on the Verbal and Visual Arts. Seattle: American Ethnological Society, 12-44.

Sinclair, J. M. 1981. Planes of discourse. In S. N. A. Rizvi (ed.), The Two-Fold Voice: Essays in Honour of Ramesh Mohan, 70–89. Salzburg: University of Salzburg.

Tannen, D. 1980. A comparative analysis of oral narrative strategies: Athenian Greek and American English. In W. L. Chafe (ed.), The Pear Stories: Cognitive, Cultural and Linguistic Aspects of Narrative Production. Norwood, NJ: Ablex, 51-87.