Appraisal theory is concerned is concerned with the linguistic resources for by which a texts/speakers come to express, negotiate and naturalise particular inter-subjective and ultimately ideological positions. Within this broad scope, the theory is concerned more particularly with the language of evaluation, attitude and emotion, and with a set of resources which explicitly position a text's proposals and propositions interpersonally. That is, it is concerned with those meanings which vary the terms of the speaker's engagement with their utterances, which vary what is at stake interpersonally both in individual utterances and as the texts unfolds cumulatively.
The paper is intended to provide an overview of appraisal theory by way of an introduction. It therefore omits some of the detail and some of the more problematic areas of the analysis. As well, it excludes any extended exemplification of appraisal theory in action in authentic text analysis. More detail can be obtained on the appraisal website ( www.languageofevaluation.info/appraisal ) in the "Introductory Course in Appraisal Analysis" and in, for example, Iedema, Feez, and White 1994 or White 1998 (available as an e-mail attachment from Peter White at p.r.white@bham.ac.uk ).
Some of the key references on Appraisal include (in chronological order): Iedema et al. 1994, Martin 1995a, Martin 1995b, Christie and Martin 1997, Martin 1997, Coffin 1997, Eggins and Slade 1997 (especially chapter 4), White 1998, Martin 2000, Coffin 2000, White 2000, Körner 2001, Rothery and Stenglin in press, and a special edition of the journal Text to appear in 2002.
The following set of notes relies primarily upon Iedema et al. 1994, Christie and Martin 1997, Martin 2000, White 1998and White to appear from which most of the material is taken.
It must be noted that appraisal theory is very much an on-going research project - many problems are still to be solved and many lexicogrammatical and semantic issues have yet to be addressed. There are numerous registers and discourse domains to which the theory has not yet been applied. (Past experience indicates that analyses of new discourse domains typically lead to significant extensions to and elaborations of the appraisal framework since each domain will typically operate with at least some unique semantic features.) The community of researchers using the theory in some way, however, continues to grow and therefore we anticipate continuing breakthroughs in the mapping of this semantic domain.
Appraisal theory divides evaluative resources into three broad semantic domains:
Values by which speakers pass judgements and associate emotional/affectual responses with participants and processes (see underlined items)
Well, I've been listening to the two guys who are heroes [value judgement] and I admire [affect] them both.
Pop Group Republica - super-schlock stinkers only a Pepsi executive could ever love
Resources for positioning the speaker's/author's voice with respect to the various propositions and proposals conveyed by a text; meanings by which speakers either acknowledge or ignore the diversity of view-points put at risk by their utterances and negotiate an interpersonal space for their own positions within that diversity.
For example:
Values by which (1) speakers graduate (raise or lower) the interpersonal impact, force or volume of their utterances, and (2) by which they graduate (blur or sharpen) the focus of their semantic categorisations.
[ABC redio interview]
MITCHELL:
There is an argument, though, is there [attribution], the banks have been a bit [graduation: force] greedy [attitude] I mean, the profits are high and good on them [attitude], they're entitled to have high profits, but at the same time the fees are bordering on [graduation: focus] the unreasonable [attitude] now.
PRIME MINISTER MR HOWARD:
Well, there's a lot of [graduation: force] anger [attitude: affect] about many [engagement: force] of the fees and this is really why, I say again, [engagement: proclamation] the more competition we can have the better [attitude]. And there's no doubt that [engagement: probability] home loan interest rates, in particular, are lower now because of competition.
Appraisal theory is, of course, located within the framework of Systemic Functional Linguistics. The primary impetus for its development has come from work conducted in the 80s and 90s for the Write It Right project of the NSW Disadvantaged Schools Program. Under Write It Right, researchers explored the literacy requirements of the discourses of science, technology, the media, history, English literature studies, geography and the visual arts. Much of what is presented here comes directly from that research (see for example Iedema, Feez, and White 1994, and Christie and Martin 1997, Rothery and Senglin in press).
Predictably, issues to do with the semantics of the interpersonal proved to be central the various Write It Right projects. For example, across all the discourse domains it proved necessary to explore in what contexts, by what linguistic means and to what rhetorical ends writers pass value judgements, attribute their propositions to outside sources or modalise their utterances.
As indicated above, the researchers starting point was the model of Tenor and the interpersonal provided by the established systemic literature. That literature provided a relatively detailed account of the lexicogrammar of the interpersonal which includes, for example, accounts of
Additionally, work by Poynton in particular (1985, 1990), had provided a model of the interpersonal with respect to social context, that aspect of context of situation which is termed Tenor, and which is concerned with the constitution of social roles and relationships and the negotiation of these roles and relationships by speakers. Under this model, three dimensions are identified by which social relationships may be organised - power/status, contact and affect. A certain amount of work has been carried out to explore the lexicogrammatical reflexes by which power, contact and affect are realised. Thus the principles of reciprocity, proliferation and contraction have been identified by which,
Similarly, various correlations between choices from the interpersonal lexicogrammar and these Tenor variables have been observed. Thus a consistent preference for high values of modals of obligation (you must/should, it's necessary that etc) and for high values of probability (definitely, I'm certain that etc) are linked with the more powerful speaker in an unequal status relationship. In contrast, a preference for modal values of inclination ( I'm keen, I'm willing etc) and for low values of probability (perhaps, may, I guess...) are linked with the less powerful speaker in an unequal status relationship. Likewise, the use of reduced expression forms, colloquial lexis and a diversity of forms of personal address are associated with contexts of higher involvement/contact between interactants. Heightened affective involvement, similarly, has its own set of indicators - the presence of exclamation, repetition, intensification and attitudinal lexis, and so on. (See Martin 1992: 523-535)
While these insights are of obvious relevance to key questions within the interpersonal semantics, they nevertheless were not formulated to answer the types of new questions arising from the Right it Write research. A need soon emerged for new linguistic accounts with which could, for example,
In retrospect we can now say that the Write It Right research was discovering a need for a better, or at least more delicate understanding of the discourse semantics of Tenor. The researchers needed to better understand the rhetorical and social-positioning consequences associated with the various options provided by the interpersonal lexicogrammar.
Perhaps most particularly, a need was emerging for a revision or at least a broadening Poynton's notion of contact. For Poynton, contact tracks the `frequency of interaction' between the interactants in the communicative exchange, and the `extent of time' that those interactants have been involved in some social relationship. Such a formulation is obviously directed towards the relationships between individuals who come into direct social contact - it is most obviously directed towards the interactants in dialogical exchanges. But the `contact' established between, say, a monologic media, history or scientific text and its audience must obviously be understood in less concrete, interactional and contingent terms. It emerged that a key issue here turned, not on the degree of social familiarity or intimacy between interactants, but on the way that texts went about constructing certain degrees of what might be termed evaluative or ideological contact with their prospective readerships. That is to say, it became clear that to understand the rhetorical potential and communicative properties of the types of text under consideration, it was necessary to explore how the evaluative positions conveyed by a text were constructed as more or less compatible, convergent and in sympathy with the anticipated positions of the text's prospective readerships. It was necessary to explore the linguistic means by which a text's authorial voice was constructed as more or less open to alternative or divergent viewpoints, as more or less willing to enter into negotiation with these alternative positions. These developments suggested that, rather than `contact', the term `solidarity' was to be preferred for this mode of social positioning (returning to the original terminology adopted by Brown & Gillman, 1960, in their classic analysis of the pronouns of address in European languages).
In conclusion to this section, therefore, we can say that the Right It Write research had revealed a need to understand more fully the rhetorical consequences associated with choosing one interpersonal value over another, and a need to identify ideological or evaluative solidarity as a key parameter along by which the interpersonal aspects of the social context may vary. In some instances these needs have motivated the identification of discourse semantic subsystems not previously recognised in the literature. Thus the Right It Write research gave rise to an account of Judgement as a set of meanings by which speakers appraise the behaviour of human individuals and to Appreciation as a set of meanings for making aesthetic and related assessments of objects and products. In other instances, these needs gave rise to new approaches to modelling the rhetorical potential of particular lexicogrammatical choices and the relationships between choices.
In this section I will offer a relatively brief outline of the three appraisal sub-systems: ATTITUDE , ENGAGEMENT and GRADUATION . The purpose here is to give an overall sense of how the appraisal system is formulated and the types of semantic issues it equipped to deal with. A more detailed account of all the sub-systems will be provided in later sections, The practical objective of this section is to provide the basic text-analytical tools by which the three appraisal systems can be identified and distinguished.
ATTITUDE includes those meanings by which texts/speakers attach an intersubjective value or assessment to participants and processes by reference either to emotional responses or to systems of culturally-determined value systems. ATTITUDE itself divides into three sub-systems
The general outlines of the grammar and semantics of affect are well understood. AFFECT is concerned with emotional response and disposition and is typically realised through mental processes of reaction (This pleases me, I hate chocolate, etc) and through attributive relationals of AFFECT (I'm sad, I'm happy, She's proud of her achievements, he's frightened of spiders, etc). Through ideational metaphor, they may, of course, be realised as nouns - eg His fear was obvious to all. Martin has developed a system for a fine-grained analyses of this semantic (See Martin 1997). I observe at this point, however, that values of affect occur as either positive or negative categories (love versus hate, please versus irritate, be bored versus be intrigued) and that each meaning is located along a sliding scale of force or intensity from low to high - thus like, love, adore; to be troubled by, the be afraid of, to be terrified of etc.
The attitudinal sub-system of JUDGEMENT encompasses meanings which serve to evaluate human behaviour positively and negatively by reference to a set of institutionalised norms. Thus JUDGEMENT is involved when the speaker provides an assessment of some human participant with reference to that participant's acts or dispositions (The exposition here relies primarily on the work from the media project of the Disadvantaged Schools Program (DSP) detailed in Iedema, Feez and White 1994.) The social norms at risk with these JUDGEMENT assessments take the form of rules and regulations or of less precisely defined social expectations and systems of value. Thus, under JUDGEMENT we may assess behaviour as moral or immoral, as legal or illegal, as socially acceptable or unacceptable, as laudable or deplorable, as normal or abnormal and so on. The DSP materials propose two broad categories of JUDGEMENT and five narrower sub-types within these two categories, which will explored in a later section. Values can be realised as,
Like AFFECT, values of JUDGEMENT have either positive or negative status - virtuous versus immoral, honest versus deceitful, brave versus cowardly, smart versus stupid, normal versus weird.
Like AFFECT, meanings can be located on a sliding scale of force or intensity from low to high values - he's an OK player, a skilled player, a brilliant player
In such instances, the value of JUDGEMENT is explicitly expressed by means of a particular lexical choice - skilfully, corruptly, lazily etc. Following the DSP material, such are classed as `inscribed' expressions of JUDGEMENT since the evaluation is overtly `inscribed' in the text through the vocabulary choice. The picture is complicated, however, by the possibility that the JUDGEMENT assessment may be more indirectly evoked or implied - rather than explicitly inscribed - by what can be termed `tokens' of JUDGEMENT. Under such tokens, JUDGEMENT values are triggered by superficially neutral, ideational meanings which nevertheless have the capacity in the culture to evoke judgemental responses (depending upon the reader's social/cultural/ideological reader position). Thus a commentator may inscribe a JUDGEMENT value of negative capacity by accusing the government of `incompetence' or, alternatively, evoke the same value by means of a token such as `the government has not laid the foundations for long term growth'. The question of `tokens' of JUDGEMENT will be taken up when we return to JUDGEMENT in detail.
I am disappointed [affect] and ashamed [affect] that two of our most admired and respected [affect] sportsmen could behave in such a manner. To play for your country is an honour and a privilege, not a right.
Those who are chosen to represent Australia should not only be talented [JUDGEMENT] but they should be above reproach [JUDGEMENT]. Sport is supposed to teach honour, fair play, teamwork, leadership and social skills [JUDGEMENT]. It is not supposed to "create" or support greed and egos [JUDGEMENT]. Gambling is not what we want our children to be learning from their heroes [JUDGEMENT] and mentors. [The West Australian - 11/12/98: 12, letter to the editor, Jennifer Black, Riverdale]
APPRECIATION1 is the system by which evaluations are made of products and processes, It encompasses values which fall under the general heading of aesthetics, as well as a non-aesthetic category of `social valuation' which includes meanings such as significant and harmful. While JUDGEMENT evaluates human behaviours, APPRECIATION typically evaluates natural objects, manufactured objects, texts as well as more abstract constructs such as plans and policies. Humans may also be evaluated by means of APPRECIATION, rather than JUDGEMENT, when viewed more as entities than as participants who behave - thus, a beautiful woman, a key figure.
Values of APPRECIATION may focus on the compositional qualities of the evaluated entity - how well formed it is. For example - harmonious, symmetrical, balanced, convoluted. Or they may focus on the aesthetically-related reaction with which the entity is associated. That is, the APPRECIATION is formulated in terms of the entity's aesthetic impact - for example, arresting, captivating, boring, dreary, beautiful, lovely etc
Like both AFFECT and JUDGEMENT, values of APPRECIATION have either positive or negative status- harmonious versus discordant, beautiful versus ugly etc
They also can be located on the cline of low to high force/intensity: eg pretty, beautiful, exquisite.
Under GRADUATION, we are concerned with values which act to provide grading or scaling, either in terms of the interpersonal force which the speaker attaches to an utterance or in terms of the preciseness or sharpness of focus with which an item exemplifies a valeur relationship,. These two dimensions are variously be labelled `FORCE' (variable scaling of intensity) and `FOCUS' (sharpening or blurring of category boundaries).
force includes values which have elsewhere been labelled, intensifiers, down-tones, boosters, emphasisers, emphatics etc. Perhaps this category's most obvious mode of expression is through the adverbs of intensification - slightly, a bit, somewhat, rather, really, very, completely etc. Somewhat more problematically, this principle of scaling also applies to those values which act to measure quantity, extent, and proximity in time and space - small, large; a few, many; near, far etc. force may also be expressed through lexical items in which the scaling value (typically a high value of intensity) is fused with some ideational meaning. This mode of force is found widely in the media - for example, the temperature plunged, prices skyrocketed, they've axed the entire accounting division, the storm cut a swathe through...
It should also be noted that this principle of grading for force operates intrinsically across value of attitude in the sense that each particular attitudinal meaning represent a particular point along the scale of low to high intensity. Thus, for example, to like represents a lower scaling of force , to love a higher scaling and to adore the highest scaling - similarly for the judgement values represented by she plays competently / skilfully / brilliantly etc.
FOCUS covers those meanings which are elsewhere typically analysed under the headings of `hedging' and `vague language'. Typical values are, he kind'v admitted it; he effectively admitted it, he as good as admitted etc; a whale is a fish, sort'v. Under appraisal theory, values which sharpen rather than blur the focus are also included - for example a true friend, pure folly, he drank his friend under the table, literally.
The operating of the principle of scaling/graduation is, perhaps, somewhat more problematic here than in the context of FORCE . As we have seen, under FORCE, scaling operates in the context of gradable categories - values which admit of different degrees of some core meaning. In contrast, under FOCUS, scaling operates in contexts which are not gradable in this sense, or where the communicative objective is not to grade in this way. For example, the state of having made a `break' with someone or something, indicated in `a clean break', is not typically construed as gradable. A similar case applies in `a true friend' and `pure folly'. Nevertheless, there is a strong sense that such values have been `scaled up' by the application of the value of FOCUS - there is a sense even of intensification. We find the inverse - a down scaling - operating in the context of values which soften the FOCUS. Thus `kind'v', in `it was kind'v nerve-wracking', lowers the scaling of intensity. From this perspective, FOCUS can be seen as the domain of the application of scales of intensity to ungraded categories. Thus under FOCUS, the scaling, and hence the lowering and raising of intensity, is realised through the semantics of category membership, through a process of narrowing or broadening the terms by which category memberships is determined, through the sharpening or softening of semantic focus.
Under ENGAGEMENT, we are concerned the linguistic resources which explicitly position a text's proposals and propositions inter-subjectively. That is, this set of rhetorical resources is concerned with those meanings which vary the terms of the speaker's engagement with their utterances, which vary what is at stake interpersonally both in individual utterances and as the texts unfolds cumulatively. To put this in terms of questions of communicative functionality and rhetorical potential, the paper is concerned with the resources for by which a text comes to express, negotiate and naturalise particular inter-subjective and ultimately ideological positions.
The meanings analysed here under ENGAGEMENT are typically analysed in the literature under the various headings of evidentiality, epistemic modality and hedging - lexico-grammatical resources such as modals of probability and usuality, reality phase (it seems, apparently), projection/attribution, hearsay and so on. Under appraisal theory, however, the resources included under ENGAGEMENT are rather more extensive than those included in the traditional categories and include negation, counter-expectation (concessives) modal adjuncts of what Halliday terms `presumption' or `obviousness ( 1994: 83) as well as intensifiers such as, `I contend that ...', `He did leave the door open'..
The modelling of ENGAGEMENT has been shaped by the specific research objectives of the projects out of which APPRAISAL theory emerged. It has been shaped by projects which shared a concern for what we might term the rhetorical potential of texts - with exploring how texts are constructed not only to persuade explicitly but also to influence and ultimately to naturalise attitudes, beliefs, and assumptions by more indirect, more implicit means. The modelling was also shaped by our observation that there was systematic, text-type and discourse type related variation in the way that such persuasion and/or influence was approached. That is, we needed a model which could describe and explain the various styles or strategies of inter-subjective positioning that we had observed operating recurrently within different discourse domains. Here the notion of style or strategy relates to the favouring of a particular sub-set of values from the APPRAISAL system and a disfavouring of other values and is somewhat similar to Biber and Finnegan's notion of a `style of stance'. A number of specific objectives follow from these concerns:
We will return to the details of the ENGAGEMENT system in a later section.
The general outlines of the grammar and semantics of affect are well understood. AFFECT is concerned with emotional response and disposition and is typically realised through mental processes of reaction (This pleases me, I hate chocolate, etc) and through attributive relationals of AFFECT (I'm sad, I'm happy, She's proud of her achievements, he's frightened of spiders, etc). Through nominalisation, they may, of course, be realised as nouns - eg His fear was obvious to all.
Values of affect provide one of possibly the most obvious ways that a speaker can adopt a stance towards some phenomenon - they provide the resources by which the speaker can indicate how that phenomenon affected them emotionally, to appraise that phenomenon in affectual terms. This functionality is illustrated by the following extract from a newspaper feature article in which the author describes her own experiences as the adoptive mother of an Australian Aboriginal baby. (AFFECT values are in underline/bold).
As an adoptive family we have had pain and trauma, tears and anger, and sometimes despair. There has also been love and laughter and support from friends and extended family. My children have added richness to my life and taught me much about myself. (Sydney Morning Herald 4/6/97.)
Such evaluations or responses are, of course, inter-subjectively charged and put at risk solidarity between speaker and audience. By appraising events in affectual terms, the speaker/writer invites their audience to share that emotional response, or at least to see that response as appropriate and well motivated, or at least as understandable. When that invitation is accepted, then, solidarity or sympathy between speaker and listener will be enhanced. Once such an empathetic connection has been established, then there is the possibility that the listener will be more open to the broader ideological aspects of the speaker's position. When the invitation to share the emotional response is not taken up - when the affectual value is seen as inappropriate, or bizarre or dysfunctional etc - then solidarity or sympathy will most probably be diminished and the chance of ideological concord diminished.
We can see this strategy at work in the extract above. The article appeared at a time when Australian Aborigines were calling for a public apology and financial compensation for the Australian government's previous policy of forcibly removing aboriginal children from their families and placing them with adoptive white parents. The policy had been described as a form of cultural genocide. A position generally supportive of the Aboriginal perspective had been widely adopted by the media and the political left and centre. The world view of the author of the extract was obviously at odds with this position, at least to the extent that for her the experience of raising two Aboriginal children had nothing to do with genocide and had not been grounds for shame and guilt. Her inclusion of AFFECT values of the type cited above can be seen as part of a strategy by which she was at least able to negotiate some space for her alternative, divergent social perspective. Her construing the issue in terms of basic human emotional responses could be expected to establish, at least in some readers, a sense of sympathy, a sense of common experiences and hence to enhance the possibility that her overall position in the article might be seen by readers as legitimate and well motivated.
The functionality of the author's own emotional responses in the construction of an interpersonal position is therefore relatively unproblematic. The formulation of APPRAISAL adopted here, however, takes into account not only authorial AFFECT but also emotional responses attributed to other social actors. The analysis relies on an observation of the way emotional reactions generally attract social evaluation as appropriate or inappropriate, as natural or unnatural and the way that description of emotions can be expected to trigger sympathetic or unsympathetic responses in the reader/listener. As well, we see the human participants introduced into a text not as isolated individuals but, potentially, as more generalised social types who will be seen to associate with a given socio-semiotic position according to their social characteristics. A reader who sympathises with the emotional response attributed to a given socio type is thus predisposed to legitimate the social position that socio type represents. We can see this dynamic at work in the following extract, taken from a letter to the editor of the Australian newspaper by an Australian of Vietnamese background. She was writing at a time when racism had become a hot media topic following the recent rise of an anti-Asian, anti-immigration and covertly racist political movement under the leadership of the independent parliamentarian, Pauline Hanson.
LAST week, Pauline Hanson attacked Footscray, labelling it an ethnic enclave that makes her feel like a foreigner in her own country.
Has Pauline Hanson been to Footscray? Is she aware of its proud tradition of struggle and hard work? Does she know about the waves of immigrants who have worked in its quarries, factories, workshops and businesses? Immigrants who have been part of the backbone of Australia's labour force and thankful for the opportunity to work and start a new life in this country. (The Australian, 4/6/97)
Here the writer is obviously concerned to negotiate intersubjective space for a social position sympathetic to the interests of immigrant Australians, in contradistinction to that advanced by Pauline Hanson and her followers. Accordingly the immigrants of one of Australia's most multicultural areas, the Melbourne municipality of Footscray, are evaluated positively through emotional responses attributed to them. Thus, they are declared to be `proud' of their hard work and struggle and to be ` thankful' for their opportunities in their new home. The writer establishes a stance towards a particular socio-semiotic reality via the affectual values she attributes to representatives of that reality.
The System of JUDGEMENT encompasses meanings which serve to evaluate human behaviour positively and negatively by reference to a set of institutionalised norms. The exposition here relies on work from the media project of the New South Wales Disadvantaged Schools Program (DSP) detailed in Iedema, Feez and White 1994, and on White 1998
The social norms at risk with JUDGEMENT take the form of rules and regulations or of less precisely defined social expectations and systems of value. Thus, under JUDGEMENT we may assess behaviour as moral or immoral, as legal or illegal, as socially acceptable or unacceptable, as normal or abnormal and so on.
We propose two broad categories of JUDGEMENT and five narrower sub-types within these two categories. It is necessary to stress, however, that since JUDGEMENT is so highly determined by cultural and ideological values, it should not be assumed the same sub-categorisations will apply in other cultural contexts, especially beyond the Western, English-speaking, essentially middle-class setting of the media analysis upon which the theory is based
We proposes the two broad categories of social sanction and social esteem. JUDGEMENTS of social sanction involve an assertion that some set of rules or regulations, more or less explicitly codified by the culture, are at issue. Those rules may be legal or moral and hence JUDGEMENTS of social sanction turn on questions of legality and morality. From the religious perspective, breaches of social sanction will be seen as sins, and in the Western Christian tradition as `mortal' sins. From the legal perspective they will be seen as crimes. Thus to breach social sanction is to risk legal or religious punishment, hence the term `sanction'.
JUDGEMENTS of social esteem involve evaluations under which the person judged will be lowered or raised in the esteem of their community, but which do not have legal or moral implications. Thus negative values of social esteem will be seen as dysfunctional or inappropriate or to be discouraged but they will not be assessed as sins or crimes. (If you breach social sanction you may well need a lawyer or a confessor but if you breach social esteem you may just need to try harder or to practice more or to consult a therapist or possibly a self-help book.)
We divide social esteem into the following three subcategories: normality or custom (how unusual someone is, how customary their behaviour is), capacity (how capable someone is) and tenacity (how dependable someone is, how well they are disposed emotionally or in terms of their intentionality).
The full system of JUDGEMENT, is set out below in Figure 1.
Social Esteem |
positive [admire] |
negative [criticise] |
normality (custom) `is the person's behaviour unusual, special, customary?' |
standard, everyday, average...; lucky, charmed...; fashionable, avant garde... |
eccentric, odd, maverick...; unlucky, unfortunate...; dated, unfashionable ... |
capacity `is the person competent, capable?' |
skilled, clever, insightful...; athletic, strong, powerful...; sane, together... |
stupid, slow, simple-minded...; clumsy, weak, uncoordinated...; insane, neurotic... |
tenacity (resolve) `is the person dependable, well disposed?' |
plucky, brave, heroic...; reliable, dependable...; indefatigable, resolute, persevering |
cowardly, rash, despondent...; unreliable, undependable...; distracted, lazy, unfocussed... |
Social Sanction |
positive [praise] |
negative [condemn] |
veracity (truth) `is the person honest?' |
honest, truthful, credible...; authentic, genuine...; frank, direct ...; |
deceitful, dishonest...; bogus, fake...; deceptive, obfuscatory... |
propriety (ethics) `is the person ethical, beyond reproach?' |
good, moral, virtuous...; law abiding, fair, just...; caring, sensitive, considerate... |
bad, immoral, lascivious...; corrupt, unjust, unfair...; cruel, mean, brutal, oppressive... |
Figure 1: JUDGEMENT (after Iedema, Feez, and White 1994)
It is vital, additionally, to distinguish between what can be termed `inscribed' (or explicit) JUDGEMENT and `tokens' of JUDGEMENT (or implicit JUDGEMENT). Under the inscribed category, the evaluation is explicitly presented by means of a lexical item carrying the JUDGEMENT value, thus, skilfully, corruptly, lazily etc. It is possible, however, for JUDGEMENT values to be evoked rather than inscribed by what the authors label `tokens' of JUDGEMENT. Under these tokens, JUDGEMENT values are triggered by superficially neutral, ideational meanings which nevertheless have the capacity in the culture to evoke judgemental responses (depending upon the reader's social/cultural/ideological reader position). Thus a commentator may inscribe a JUDGEMENT value of negative capacity by accusing the government of `incompetence' or, alternatively, evoke the same value by means of a token such as `the government did not lay the foundations for long term growth'. There is, of course, nothing explicitly evaluative about such an observation but it nonetheless has the potential to evoke evaluations of incompetence in readers who share a particular view of economics and the role of government. Similarly, a reporter might explicitly evaluate the behaviour of, for example, a Californian suicide cult as `bizarre' or `aberrant' or they might evoke such appraisals by means of tokens such as `They referred to themselves as "angels"' or `They filled the mansion with computers and cheap plastic furniture.' Such tokens, of course, assume shared social norms. They rely upon conventionalised connections between actions and evaluations. As such, they are highly subject to reader position - each reader will interpret a text's tokens of judgement according to their own cultural and ideological positioning. They are also subject to influence by the co-text, and an important strategy in the establishment of interpersonal positioning in a text is to stage inscribed and evoked evaluation in such a way that the reader shares the writer's interpretations of the text's tokens.
APPRECIATION is the system by which evaluations are made of products and processes. The account set out here relies entirely on the work of Rothery, developed initially during research into the language of the visual arts for various NSW Disadvantaged Schools Program projects, as well as subsequent analysis by Rothery and Stenglin of the role of evaluation in secondary school English essays (Rothery and Stenglin in press). (For a review see Martin 1997: 24-26). APPRECIATION encompasses values which fall under the general heading of aesthetics, as well as a non-aesthetic category of `social valuation' which includes meanings such as significant and harmful. APPRECIATION can be thought of as the system by which human feelings, either positive or negative, towards products, processes and entities are institutionalised as a set of evaluations. Thus, whereas JUDGEMENT evaluates human behaviours, APPRECIATION typically evaluates texts, more abstract constructs such as plans and policies, as well as manufactured and natural objects. Humans may also be evaluated by means of APPRECIATION, rather than JUDGEMENT, when viewed more as entities than as participants who behave - thus, a beautiful woman, a key figure.
Rothery and Stenglin (in press) propose three subcategories under which appreciations may be grouped: reaction, composition and valuation. According to Rothery & Stenglin, reaction is `interpersonally tuned. It describes the emotional impact of the work on the reader/listener/viewer.' Thus, under reaction, the product/process is evaluated in terms of the impact it makes or its quality. For example:
Under composition, the product or process is evaluated according to its makeup, according to whether it conforms to various conventions of formal organisation. As Rothery and Stenglin state, `Composition is textually tuned. It describes the texture of a work in terms of its complexity or detail.' For example:
Under the subcategory of `social value', the object, product or process is evaluated according to various social conventions. This domain is very closely tied to field in that the social valuation of one field will not be applicable or relevant in another. Thus we would expect that the set of social values which have currency in, for example, the visual arts, might not have extensive application in the world of politics. In the context of the media texts under which much of this theory was developed, the key values were those of social significance or salience (whether the phenomenon was important, noteworthy, significant, crucial etc) and of harm (whether the phenomenon was damaging, dangerous, unhealthy etc).
ENGAGEMENT, as already indicated, covers all those resources by which the textual or authorial voice is positioned inter-subjectively. Lexicogrammatically it encompasses a diverse array of resources:
On what basis is it proposed, therefore, that such lexico-grammatical diversity should be grouped together within the one system? As indicated above, there are several well-established traditions within the literature by which at least a sub-set of these resources are analysed as serving a similar rhetorical functionality. Thus analyses under the headings of `evidentiality', `modality' and `hedging' will often include modal verbs and adjuncts, reality phase and at least some types of attribution/reported speech. It is rather less usual for such analysis to include values of negation and expectation/counter-expectation, but these values do nevertheless sometimes receive and analysis under these headings. The insight at work here is, as Lyons for example has argued (Lyons 1977), that all such meanings serve to indicate an attitude towards the proposition or proposal by the speaker/author.
I concur to the extent that I see such values as attitudinal in the broadest sense of the term. I differ from these established analyses, however, in that, I see as inadequate the truth-functional orientation of traditional modality theory, the epistemic reliability orientation of the evidentiality approach and the negative/positive face approach typically adopted by the `hedging' literature. I will propose an alternative analysis of the rhetorical functionality of these resources, following suggestions from Lemke ( 1992), Fairclough ( 1992), Thibault ( 1997) and Fuller (1998), based on Bakhtin's inter-connected notions of heteroglossia and dialogism (1973, 1978, 1981, 1986). I will argue that this heteroglossic understanding of the semantics of ENGAGEMENT is much more compatible with Halliday's characterisation of the functionality of MOOD in that it understands inter-subjective positioning in social rather than individualistic terms and in that it attends to the way all utterances are centrally concerned with the negotiation of interactional and informational meanings.
We can say that traditional approaches have construed these resources as either being concerned with the speaker's commitment to the truth-value of their utterances or with allowing the speaker to characterise their utterances as less than factual or as less than certain `knowledge', as having a diminished epistemic reliability. Thus Lyons contrasts what he terms the `subjectivity' of the modal meaning with the `objectivity' of `bare assertions' and describes such modalised utterances as `non-factive', in contrast to these `factive' utterances which are `straightforward statements of fact [which] may be described as "epistemically non-modal" because the speaker commits himself to the truth of what he asserts' (1977: 794). In a somewhat similar vein, Chafe observes,
People are aware, though not necessarily consciously aware, that some things they know are surer bets for being truer than others, that not all knowledge is equally reliable. Thus one way in which knowledge may be qualified is with an expression indicating the speaker's assessment of its degree of reliability. (Chafe 1986: 264)
Under these formulations, therefore, the semantics at issue is represented as emerging from meaning making in which individual speakers apply a `subjective' coloration or slant to the propositional content of their utterances so as to hedge the truth value of that content or to indicate doubts about its reliability. The semantics are construed as turning on whether individual speakers present themselves as willing or unwilling to commit to the truth of what they assert. Frequently the choice is construed as one between objective `facts' and the subjective uncertainty of the modal or the evidential value _ hence Lyons contrast between `non-factive' and tellingly what he terms the `straightforward statements of fact' (Lyons 1977: 794, emphases mine).
There is an implication, therefore, in the various formulations that the overriding purpose of communication is to exchange truth values or certain knowledge and that these modal, evidential or hedging values are introduced only in communicatively non-optimal circumstances. Thus the speaker is represented as inserting modal values and hence adopting an interpersonal position when they have failed to achieve an absolute, and hence `straightforward' (following Lyon's citation above), commitment to the truth of their utterances. These are thus values to be used, so to speak, when facts fail you. The term `hedge', I believe, reflects this perspective, suggesting as it does some form of evasion or even deceit, some sense of improperly `having it both ways' .
Such approaches are also informed by a view of communication in which either speaker or speaker and listener are constructed in individualised terms, rather than as social subjects dealing with meanings informed by and reflecting social structures and conditions. Thus the presence of the modal/evidential/hedging value is seen primarily to reflect the speaker's individual state of mind - speakers insert a modal qualification, for example, as a way of signalling their uncertainty, as a way of coding their individual lack of commitment to some propositional content.
Under the model developed here, however, I adopt an approach to these values which gives a greater role to the audience, or at least to the way texts can be seen to negotiate meanings with actual and potential audiences. As well, I construe meaning making in social rather than individualised terms and will not give priority to ideational content and its associated truth value.
In this I am reflecting general systemic functional assumptions about language and language use. I am, however, more specifically influenced by Bakhtin's notions of `heteroglossia' and `intertextuality' (1973, 1981, 1986) Under these notions, Bakhtin insists upon the intertextual nature of all texts, observing that all texts necessarily reference, respond to, and to greater or lesser extents incorporate other texts both actual and prospective.
The desire to make one's speech understood is only an abstract aspect of the speaker's concrete and total speech plan. Moreover, any speaker is himself a respondent to a greater or lesser degree. He is not, after all, the first speaker, the one who disturbs the eternal silence of the universe. And he presupposes not only the existence of the language system he is using, but also the existence of preceding utterances-his own and others'-with which his given utterance enters into one kind of relation or another (builds on them, polemicizes with them, or simply presumes that they are already known to the listener). Any utterance is a link in a very complexly organized chain of other utterances. (Bakhtin 1986: 69)
Thus we might say that no utterance is an island, as it were.
The heteroglossic perspective emphases the role of language in positioning speakers and their texts within the heterogeneity of social positions and world views which operate in any culture. All texts reflect a particular social reality or ideological position and therefore enter into relationships of greater or lesser alignment with a set of more or less convergent/divergent social positions put at risk by the current social context. Thus every meaning within a text occurs in a social context where a number of alternative or contrary meanings could have been made, and derives its social meaning and significance from the relationships of divergence or convergence into which it enters with those alternative meanings. As Lemke observes, in his interpretation of Bakhtin,
Lexical choices are always made against the background of their history of use in the community, they carry the `freight' of their associations with them, and a text must often struggle to appropriate another's word to make it its own. (Lemke 1992: 85)
Thus texts are `heteroglossic' - they directly address or at least implicitly acknowledge a certain array of more or less convergent and divergent socio-semiotic realities. They address those alternative realities as expressed in previous texts and as they are expected to be realised in future texts. As a consequence, every meaning within a text occurs in a social context where a number of alternative or contrary meanings could have been made, and derives its social meaning and significance from the relationships of divergence or convergence into which it enters with those alternative meanings.
(This notion of heteroglossia is also reflected in Foucault's account of intertextuality. Thus Foucault states, `there can be no statement that in one way or another does not reactualize others' - Foucault 1972: 98 /d. This notion is also fundamental to Fairclough's analysis of intertextuality and orders of discourse. See for example Fairclough 1989, 1992)
When informed by this view of text as heteroglossic, our approach to these linguistic resources will be rather different from the individualistic approach exemplified by Lyons' definition. Rather than seeing these values as necessarily oriented to coding a speaker's individual position or attitude, I will see them as operating to reflect the process of interaction or negotiation within a text between alternative socio-semiotic positions.
Under the individualistic (what Lemke terms `social interactionist') model, a modal value such as `maybe' or `I think that ..' is seen as acting to indicate uncertainty or lack of commitment to, or confidence in the truth values by the individual speaker - it is seen as epistemological, as a reflex of the speaker's current state of knowledge with respect to some propositional content. Under the heteroglossic perspective, rather than necessarily reflecting the speaker's state of knowledge, it can additionally or alternatively be seen as signalling that the meanings at stake are subject to heteroglossic negotiation. It may have no connection at all with doubt or vagueness, being used, instead, to acknowledge the contentiousness of a particular proposition, the willingness of the speaker to negotiate with those who hold a different view, or the deference of the speaker for those alternative views.
The terms of that negotiation will vary according to the context of situation and, in particular, the social relationships between speaker and audience. Thus, within academic discourse, the speaker may use a modal of probability to acknowledge the contentiousness or novelty of a given meaning, thereby coding a willingness to recognise and negotiate with divergent heteroglossic positions over that meaning. Such functionality is exemplified below by an extract from an article in which the writer seeks to advance the novel, contentious proposition that Marx was a precursor of contemporary anthropological theories of culture. In the course of this opening paragraph, the writer goes from characterising the proposition as extremely improbable, to asserting it forcibly. The movement is not from actual doubt, vagueness or epistemological unreliability to certainty. It is a rhetorical move designed to deal with the novelty and contentiousness of the author's primary proposition. (I have firstly underlined the various wordings which characterise various meanings in these modal terms, and then the final affirmative statement, where the author declares his position without qualification.)
This consideration of Marx as a precursor, though a largely unacknowledged one, of the modern anthropological theory of culture is situated on somewhat improbable terrain. It lies in a no-man's-land between two rather unlikely propositions: first, that there can be anything much new to be said about Marx; and second, that, having been enthusiastically cited now for a century by those who would entirely conflate human history with natural history and culture, into its occasioning circumstances, Marx had anything at all of value to say to his contemporaries - still less has anything to offer us about culture. Yet such a consideration is neither absurd nor untimely, as Raymond Williams' recent discussion cited above demonstrates. (Kessler 1987: 35)
In other contexts, the same general semantic resources may be used towards rather different rhetorical ends. For example modals of probability may function to enable speakers to avoid indicating a firm preference for one heteroglossic position, not because they entertain genuine epistemological doubt over the issue or because they wish to show deference to alternative positions, but because they choose, for whatever interpersonal reasons, to resist being positioned in this way. The following extract from the stage play, Educating Rita, illustrates such a strategy. (The character Rita is a mature age university student from a working class background. Frank is her university tutor. The pair are engaged in a one-to-one tutorial session.)
Rita: That's a nice picture, isn't it Frank?
Frank: Uh yes, I suppose it is.
Rita: It's very erotic.
Frank: Actually I don't think I've looked at this picture in 10 years, but, yes, it is, I suppose so.
Rita: Well, there's no suppose about it.
The extract demonstrates a clash in the interpersonal styles (what we might term codes, following Bernstein 1970) between Rita's monoglossic and Frank's heteroglossic rhetorical strategy. Presumably the audience doesn't interpret Frank's lines as indicating that the character has a great deal invested epistemologically or interpersonally in the painting. Rather, the Frank character here seems to be using values of probability (I suppose, I don't think etc), not out of either doubt, deference or a desire to save Rita's `face', but as almost a passive aggressive tool for insisting upon his heteroglossic mode and for denying or seeking to suppress the Rita character's monoglossic mode. Rita, of course, is alive to this strategy and confronts it through what amounts to a rejection of heteroglossia in this particular context - `Well, there's no suppose about it.' (See Martin to appear to appear/c for an extended discussion of interpersonal positioning in Educating Rita.)
A crucial feature of these values, therefore, is their context-dependent polysemous functionality. In a sense, this multi-functionality can be seen as analogous to that of the smile as a communicative device. In one context, a smile may act or be read as genuinely signalling a mental state of happiness or pleasure in the person smiling. In other contexts the smile is a politeness marker, exchanged between acquaintances as they pass in the corridor, for example, as an indicator of recognition or acknowledgement, and thus carrying no affectual value at all. Similarly, a modal value of probability may, in one context, signal genuine epistemological doubt in the speaker. Equally, it may have no connection at all with doubt, being used, rather to acknowledge the contentiousness of a particular proposition, the willingness or unwillingness of the speaker to negotiate with those who hold a different view, or the deference the speaker wishes to display for those alternative views.
From this Bakhtinian perspective, therefore, I characterise as too narrowly-based those formulations which would construe such values exclusively in negative terms as `hedges', as deviations from `straightforward' factuality, or as points of epistemological unreliability. We should, rather, see them as acting to open up, or to extend the semantic potential available to the text - in some contexts enhancing the possibility of a continued heteroglossic negotiation between divergent positions, and in others acting to forestall or fend off that negotiation.
In the previous sections I have set out in general terms the heteroglossic orientation with informs the remodelling of inter-subjective positioning from which the formulation of ENGAGEMENT is derived. In the following sections I describe some of the key elements of the ENGAGEMENT system.
As I have indicated, under a heteroglossic approach, we see utterances as necessarily invoking, acknowledging, responding to, anticipating, revising or challenging a range of more or less convergent and divergent alternative utterances and hence social positions. This perspective thus provides a potent counter to the common-sense notion that certain utterances are interpersonally neutral and hence `factual' or `objective' while others are interpersonally charged and hence `opinionated' or `attitudinal'. Under systemic functional perspectives, of course, all utterances are analysed as simultaneously ideational and interpersonal - there is no utterance which is without interpersonal value. Nevertheless, the influence of the common-sense notion of the `fact' is widespread and it may be tempting to see some utterances as more interpersonal than others. Under the heteroglossic orientation, however, we are reminded that even the most `factual' utterances, those which are structured so as to background interpersonal values, are nevertheless interpersonally charged in that they enter into relationships of tension with a related set of alternative and contradictory utterances. The degree of that tension is socially determined. It is a function of the number and the social status of those alternative socio-semiotic realities under which the utterance at issue would be problematised. Accordingly, the difference between the utterance `Francis Bacon was the author of The Tempest.' and `In my view, Francis Bacon was the author of The Tempest.' is not one of `fact' versus `opinion' but of the degree to which the utterance acknowledges the intertextual or dialogic context (in Bakhtin's sense) in which it operates. The distinction, therefore, can be represented in terms of heteroglossic negotiation - the first utterance (the so-called `fact') ignores or down-plays the possibility of heteroglossic diversity by dint of its lexico-grammar while the second actively promotes that possibility. Alternatively, we can say the first denies or ignores the intertextual heterogeneity in which it operates while the second asserts it.
The key insight here is that the positive declarative is not any less interpersonal and positioning than the array of alternatives by which the heteroglossic diversity is invoked, acknowledged or fended off . Rather, in using the positive declarative, the writer adopts a particular rhetorical strategy towards the possibility of heteroglossic diversity, namely of either choosing not to directly acknowledge that possibility, or of assuming a homogeneous rather than a heterogeneous speech community. This is not, of course, a novel position and there is a long-standing debate in the literature over the `modal' status of the so-called `bare declarative'. Thibault, for example, adopts a similar position to mine in his `Mood and Eco-social dynamics' paper (1997). He states, for example, `no utterance is free from subjective presencing of the speaker'(1997: 53).
The fundamental choice in the system of ENGAGEMENT, therefore, is between what we might term the mono-glossic option (the so-called `bare' declarative) and the network of resources by which we directly inscribe the possibility of social heterogeneity into the text.
Figure 2: ENGAGEMENT entry point
A basic distinction, then, under engagement is that between meanings which acknowledge in some way the heteroglossic diversity associated with all utterances (the heteroglossic) and those which ignore that diversity (the monoglossic). Within this broad space of heteroglossic acknowledgement there is an array of alternative meanings, each of which has its own distinctive rhetorical properties in that each differs in the terms by which it acknowledges or invokes the heteroglossic context.
Let's consider some of the options.
Under the heteroglossic approach, attribution is unproblematically a key resource for acknowledging heteroglossic diversity - it provides a resource by which the heteroglossic diversity of the current social context can be directly inscribed in the text. By attributing an utterance or proposition, the text explicitly marks it as contingent, as but one of a number of possible utterances, as located at a specific point in the network of inter-subjective positions put at risk by the texts current context of situation. The utterance is marked as sourced by a particular individualised social-subjectivity. Since that source is external to the text, attribution serves as one mechanism for multi-vocalising the text, for constructing the text, in Bakhtin's terms as dialogic, as emanating from multiple sources and reflecting multiple points of view. I term such attributing resources `extra-vocalisation', in recognition of the way the introduce externally sources voices into the text.
Figure 3: extra-vocalisation
One key resource, therefore, for acknowledging the heteroglossic diversity is provided by extra-vocalisation. Extra-vocalisation contrasts with an array of resources by which the heteroglossic diversity is construed as more internal to the text, where the dialog (in Bakhtin's terms) is essentially internal rather than external. Thus,
Extra-vocalisation
Some scholars say Francis Bacon wrote The Tempest.
Intra-vocalisation
Francis Bacon probably wrote The Tempest. (probability)
It's clear Francis Bacon wrote The Tempest. (appearance)
I contend, therefore, that Francis Bacon did write The Tempest. (proclamation)
All these intra-vocalising resources act to multiply the voices of the text, to cast it as dialogic in Bakhtin's sense by explicitly subjectivising the voice of the author, by bringing it forward from the interpersonal background and casting it as contingent, as just one of a number of possible voices.
I will explore this semantic in the context of values of probability.
To probabilise propositional content is to cast it in terms of a particular, individualised inter-subjective position. In some cases of probability this is made highly explicit through the resources of grammatical metaphor in structures such as `I think/suppose/guess... .' Such metaphors operate at two stratal levels - at both the level of the lexicogrammar and the discourse semantics. At the discourse semantic level, the meanings are interpersonal - coding an inter-subjective value of probability, as Halliday's analysis of the tagging behaviour of such structures indicates. (Tags target the Subject of the projected clause, not the projecting clause, indicating that modal responsibility is assigned to the MOOD element of the projected clause, not the projecting Sensor - `I think he's left already, hasn't he'.) At the lexicogrammatical level, in contrast, such structures exploit the ideational metafunction to make their meanings. The meanings at issue are construed in terms of an experiential mental process which projects the propositional content at issue. By this mechanism, the authorial role is explicitly represented in the text as the Sensor who does the projecting. The intertextual role of the authorial voice as source is thereby foregrounded. (For a full account of this semantic see Fuller 1998, from which this account is derived.)
But even when the inter-subjectivity is less foregrounded, the modal still operates to interpolate the internal authorial voice and thereby cast the utterance as explicitly contingent, as one of a possible set of heteroglossic alternatives. The voice that opines, for example, that `the government may be corrupt' simultaneously evokes voices who make different statements with respect to the government. The modal locates the proposition at some point in the semantic space between the polar opposites of absolute `Yes' (the positive) and absolute `No' (the negative). By so locating the utterance, the modal brings into play all the other points along the cline between the polar absolutes.
Thus while attribution and probability differ in the terms by which they complicate and diversify the voicing of the text (attribution introduces an external, probability an internal voice), they nevertheless share this functionality of multiplying the voices in the text and thereby establishing each voice as representing but one of a number of possible heteroglossic positions. Insights from Bakhtin are once again helpful in clarifying this point. He talks as about `our-own-ness' and the `other-ness' of a text's multiple voices.
Our speech, that is, all our utterances (including creative works), is filled with others' words, varying degrees of otherness and varying degrees of `our-own-ness', varying degrees of awareness and detachment. These words of others carry with them their own expression, their own evaluation tone, which we assimilate, rework and reaccentuate. (Bakhtin 1986: 89) Bakhtin 1986: 89)
Thus both probability and attribution multiply voices, with extra-vocalising attribution associated with `other-ness' and intra-vocalising probability associated with `our-own-ness'.
Values of proclamation (I declare, it is my contention etc) are, perhaps, less problematic in this context, in the sense that they quite obviously act to interpolate the authorial voice. By means of such rhetorical gestures, the authorial voice is explicitly foregrounded, declaring its role as the inter-subjective source of the utterance in question. We'll look further at the underlying communicative motivation for such an overt declaration of subject-hood in the following section.
As indicated above, the intra-vocalising values include the following options (this is only a partial sub-set of the total repertoire of intra-vocalisation)
One further dimension of semantic optionality remains to be explored. This divides probability, appearance and hearsay, on the one side, from proclamation, on the other. It turns on the distinction between what I term `open' and `close'.
The resources which I include under `open' can be understood as acting to `open up' the heteroglossic dialogue, to extend the text's potential for construing heteroglossic diversity. These include probability, appearance (reality phase) and hearsay.
The semantics of probability and appearance (reality phase) are widely referenced in the literature. As discussed above, under the heteroglossic perspective, meanings such as `I think ...', `probably', `It seems...', `Apparently ...' etc are not construed as evasions of truth values but rather as resources by which the speaker `opens up' their potential for interacting with the heteroglossic diversity.
The second option under `open', `appearance', serves a similar rhetorical function to probabilise. It `opens up' the potential for negotiating the heteroglossic diversity by reference not to probability but by foregrounding and making explicit the evidential process upon which all propositions rely.
The final option under `open' - `hearsay' - entertains heteroglossic diversity in a similar way, although its semantics perhaps require some additional explication. Under `hearsay' the possibility of heteroglossic alternation remains open because the utterance is marked as based on what some unspecified person said. In many languages, of course, hearsay is coded grammatically rather than lexically - hearsay is an integral part of the MODALITY system. Within English and similar languages, hearsay is coded by means of wordings which derive from the grammar of verbal projection but in which the projecting sayer is absent or cannot be specified - thus, `Reportedly, she viewed the papers', `Her alleged viewing of the documents.', `It's said she viewed the papers.', `I hear she viewed the documents.'. The semantic consequence of such structures is not to introduce an alternative voice into the text and hence `hearsay' is included within `intra-vocalise' and not within `extra-vocalise'. Rather, it functions to indicate that the meanings qualified by the hearsay are negotiable in heteroglossic terms.
Figure 4: ENGAGEMENT - open versus close
There are a number of grammatical grounds for, firstly, identifying `hearsay' as a distinct discourse semantic category separate from the extra-vocalisation categories which it superficially resembles and, secondly, for aligning it in the semantics with `probability' and `appearance'.
Where the `open' options operate to extend the possibilities for heteroglossic interaction, those under `close' act in some way to limit the range or possibility of interaction with the diversity. They nevertheless remain fundamentally heteroglossic - though acting to `close down' the range of that interaction, they nevertheless address or invoke that diversity in some way. Under this semantic of `close', an alternative, typically contrary meaning is referenced or at least entertained (and hence the author enters into a heteroglossic dialogue) but is then suppressed, replaced, rejected or challenged in some way, and any heteroglossic dialogue thereby `closed down'.
Accordingly, the types of `proclamation' I have previously sited increase the strength of the speaker's engagement with the utterance/proposition in question. They thereby act to challeng the reader/listener to question/reject/doubt their propositions by increasing the interpersonal cost, in some way, of that proposition being rejected. They do this by adding some additional support or motivation for the proposition.
Previous I have listed proclamations which take the form of an explicit interpolation of the speaker into the text: `I'd say that / it's my contention that the Premier saw the documents.' Alternatively, `pronouncement' may take the an intensifying comment adjunct (`Really, the Premier saw the papers'), stress on the auxiliary (`The Premier did view the documents'), or through structures such as `It's a fact that...'. (See Fuller 1995: Chapter 4 for a discussion of `interpolation'.) The author thereby increases the interpersonal cost of any rejection/doubting of their utterance, rendering such a direct challenge to the author's dialogic position. Of course, through such a strategy, by confronting the possibility of rejection, the author integrates that possibility into the text and thereby acknowledges the heteroglossic diversity of meaning making in socially diverse social contexts.
The third dimension within APPRAISAL is that of GRADUATION. Within this semantic space, we are concerned with values which scale other meanings along two possible parameters - either locating them on a scale from low to high intensity, or from core to marginal membership of a category.
This semantics of those which scale according to intensity is most transparently exemplified by the set of adverbials which have typically been explored in the literature under headings such as `intensifiers', `amplifiers' and `emphatics. The set includes slightly, a bit, somewhat, quite, rather, really, very, and extremely. Via these values, the speaker raises or lowers the intensity of a wide range of semantic categories - thus `very' in `a very smart fellow' acts to heightens the intensity of the JUDGEMENT value (capacity) of `smart'. and `a bit' in `I'm a bit worried' acts to lower the intensity of the affectual value of `worried'. Under the system set out here, this dimension of scaling with respect to intensity will be labelled FORCE.
The values of FORCE contrast with those that are labelled FOCUS. Here the scaling operates in terms of the sharpness or softness of the valeur relationship represented by the item. Values at the `sharp' end of the focus scale are exemplified by true friend, pure evil, a clean break, a genuine mistake, a complete disaster, par excellence. Here FOCUS values operate to indicate that the valeur represented has core or prototype status - that the valeur relationship is sharply focussed. Values at the `soft' end of the focus scale are exemplified typically by examples of what Lakoff ( 1972) termed `hedges' - `all day, it was kind'v nerve-wracking', `a whale is fish, sort'v', `he as good as killed his brother' etc. Here the value operates to indicate that the item in question has marginal status in the category or that the valeur relationships are blurred or have imprecise boundaries.
Before turning to these two dimensions in more detail, I will address some general features of the semantics of scaling. It should firstly be noted that scaling is not confined to cases where the value is explicitly carried by some independent, isolating lexical item such as very or somewhat. We need, additionally, to consider implicit scaling. Once we allow for an implicit semantic, we discover that scaling, in terms of the raising or lowering of intensity, operates across the APPRAISAL system and is not confined within a specific sub-domain. We discover that most values of APPRAISAL are scaled for intensity, in the sense that are located somewhere on a cline between high and low degrees. This feature has already been demonstrated in the context of AFFECT. For example, in dealing with the general affectual value of `antipathy', the speaker must choose either a low value, (dislike, for example), a median value (hate) or a high value (abhor). Accordingly, some scale of intensity (from low to high) is an integral part of this semantic and to deal with such meanings is necessarily to down-tone or intensify. The operation of this implicit scaling for intensity across the gradable values of APPRAISAL is exemplified by the following,
In this sense, scaling can be seen as an interpersonal coloration or tonality across the APPRAISAL system.
This notion of scaling as a semantic orientation which may operate implicitly across semantic domains leads us to understand more precisely the relationships between values of FORCE (raising and lowering of intensity) and values of FOCUS (sharpening and softening the focus of the valeur relationship). As we have seen, under FORCE, scaling operates unproblematically in the context of gradable categories - values which admit of different degrees of some core meaning. In contrast, under focus, scaling operates in contexts which are not gradable in this sense, or where the communicative objective is not to grade in this way. For example, the state of having made a `break' with someone or something indicated in `a clean break' is not typically construed as gradable. A similar case applies in `a true friend' and `pure folly'. Nevertheless, there is a strong sense that such values have been `scaled up' by the application of the value of FOCUS - there is a sense even of intensification. We find the inverse - a down scaling - operating in the context of values which soften the FOCUS. Thus `kind'v', in `it was kind'v nerve-wracking', lowers the scaling of intensity. From this perspective, FOCUS can be seen as the domain of the application of scales of intensity to ungraded categories. Thus under FOCUS, the scaling, and hence the lowering and raising of intensity, is realised through the semantics of category membership, through a process of narrowing or broadening the terms by which category memberships is determined, through the sharpening or softening of semantic focus.
The system of GRADUATION so far set out is illustrated below.
Figure 5: GRADUATION - Force versus Focus
The different subtypes of FORCE constitute a semantically complex network of meanings. Work is continuing in accounting for the communicative and rhetorical properties of the various options within this general semantic space. The relationship between the options and is highly complex the semantics of various options difficult to define adequately. For one highly provisional mapping of these options you might like to refer to, see Roach 1991: chapter 3, section III.5 - available as an e-mail attachment upon request. The work is continuing and you might like to keep in touch with the appraisal research group for any more recent developments.
Given the complexity of this semantics, therefore, I will not attempt here to provide anything more than a listing of the key options.
· Graders: (adverbials) - slightly, a bit, somewhat, quite, rather, really, very, completely; (adjectivals) - slight, severe (applies, for example, to `headache'); slight, steep (applies to `incline')
· Measure: small, medium, large; narrow, wide; light, heavy etc).
(For a more detailed discussion of these values see White 1998: chapter 3, section III.5)
By way of conclusion to this discussion, it is worth reviewing the way that scaling for intensity provides for a broad semantic which operates trans-systemically. In particular, we note the way the distinction between high and low values can provide for two broad, opposed groupings of values - that is to say, sets of values which, though from different sub-systems, nevertheless are alike in realising either high or low intensity . The two trans-systemic groupings are exemplified below in Table 1.
Low INTENSITY |
High INTENSITY | |
Probability |
Perhaps he's a post-modernist |
He's definitely a post-modernist |
Appearance |
He seems to be a post-modernist |
It's obvious he's a post-modernist |
Proclaim |
I'd say he's a post-modernist |
I declare he's a post-modernist |
Extra-vocalise |
She say's he's a post-modernist |
She insists he's a post modernist |
Affect |
He likes post-modernists |
He adores post-modernists |
Judgement |
He's a satisfactory post-modernist |
He's a brilliant post-modernist |
Appreciation |
An attractive post-modernist work. A minor post-modern work |
An exquisite post-modernist work. A major post-modern work |
Focus |
It's a post-modern work, kind'v |
It's genuinely post-modern |
Bakhtin, M. 1973. Problems of Dostoevsky's Poetics, Michigan, Ann Arbor.
Bakhtin, M. 1978. Esthétique Et Théorie Du Roman, Paris, Gallimard.
Bakhtin, M. 1981. The Dialogical Imagination, M. Holquist, (ed.), C. Emerson & M. Holquist, (trans.), Austin, University of Texas Press.
--- 1986. 'The Problem of Speech Genres', in Speech Genres and Other Late Essays, McGee, V.W. (trans), Austin, University of Texas Press: 60-101.
Bernstein, B. 1970. Class, Codes, and Control. Volume 1: Theoretical Studies Towards a Sociology of Language, London, Routledge & Kegan Paul.
Chafe, W.L. 1986. 'Evidentiality in English Conversation and Academic Writing', in Evidentiality: the Linguistic Coding of Epistemology, Chafe, W.L. & Nichols, J. (eds), Norwood, New Jersey, Ablex Publishing Corporation.
Christie, F. & Martin, J.R. (eds) 1997. Genres and Institutions: Social Processes in the Workplace and School, London, Cassell.
Eggins, S. & Slade, D. 1997. Analysing Casual Conversation, London, Cassell.
Fairclough, N. 1989. Language and Power, London, Longman.
--- 1992. Discourse and Social Change, Cambridge, UK, Polity Press.
Foucault, M. 1972. The Archaeology of Knowledge, London, Tavistock Publications.
Halliday, M.A.K. 1994. An Introduction to Functional Grammar, London, Edward Arnold.
Iedema, R., S. Feez, and P.R.R. White. 1994. Media Literacy, Sydney, Disadvantaged Schools Program, NSW Department of School Education.
Kessler, C.S. 1987. 'Marx As Cultural Theorist: the Prehistory of Modern Anthropology', in Creating Culture - Profiles in the Study of Culture, Austin-Brooks, D.J. (ed.), Sydney, Allen & Unwin: 35-49.
Lakoff, G. 1972. 'Hedges: A Study in Meaning Criteria and the Logic of Fuzzy Concepts', Proceedings of the Chicago Linguistics Society 8: 183-228.
Lemke, J.L. 1992. 'Interpersonal Meaning in Discourse: Value Orientations', in Advances in Systemic Linguistics. Recent Theory and Practice, Davies, M. & Ravelli, L. (eds), London, Pinter Publishers.
Lyons, J. 1977. Semantics, Cambridge, UK., Cambridge University Press.
Martin, J.R. in press. 'Beyond Exchange: APPRAISAL Systems in English', in Evaluattion in Text, Hunston, S. & Thompson, G. (eds), Oxford, UK, OUP.
--- to appear. 'Inter-Feeling: Gender, Class and Appraisal in Educating Rita', in Evaluation in Text, Hunston, S. & Thompson, G. (eds), London, Oxford University Press.
Rothery, J. & Stenglin, M. in press. 'Interpreting Literature - The Role of APPRAISAL', in Researching Language in Schools and Functional Linguistic Perspectives., Unsworth, L. (ed.), London, Cassell.
Thibault, P. 1997. 'Mood and Eco-Social Dynamics', in On Subject and Theme: A Discourse Functional Perspective, Hasan, R. & Fries, P. (eds), The Hague, Benjamins: 51-90.
White, P.R.R. 1998. 'Telling Media Tales: the News Story As Rhetoric'. University of Sydney, Sydney.
Bakhtin, M. 1973. Problems of Dostoevsky's Poetics, Michigan, Ann Arbor.
Bakhtin, M. 1978. Esthétique Et Théorie Du Roman, Paris, Gallimard.
Bakhtin, M. 1981. The Dialogical Imagination, M. Holquist, (ed.), C. Emerson & M. Holquist, (trans.), Austin, University of Texas Press.
--- 1986. 'The Problem of Speech Genres', in Speech Genres and Other Late Essays, McGee, V.W. (trans), Austin, University of Texas Press: 60-101.
Bernstein, B. 1970. Class, Codes, and Control. Volume 1: Theoretical Studies Towards a Sociology of Language, London, Routledge & Kegan Paul.
Chafe, W.L. 1986. 'Evidentiality in English Conversation and Academic Writing', in Evidentiality: the Linguistic Coding of Epistemology, Chafe, W.L. & Nichols, J. (eds), Norwood, New Jersey, Ablex Publishing Corporation.
Christie, F. & Martin, J.R. (eds) 1997. Genres and Institutions: Social Processes in the Workplace and School, London, Cassell.
Coffin, C. 1997. 'Constructing and Giving Value to the Past: an Investigation into Second School History', in Genre and Institutions - Social Processes in the Workplace and School, Christie, F. & Martin, J.R. (eds), London, Cassell.
Coffin, C. 2000. 'Unpublished Ph.D Thesis'. University of New South Wales.
Eggins, S. & Slade, D. 1997. Analysing Casual Conversation, London, Cassell.
Fairclough, N. 1989. Language and Power, London, Longman.
--- 1992. Discourse and Social Change, Cambridge, UK, Polity Press.
Foucault, M. 1972. The Archaeology of Knowledge, London, Tavistock Publications.
Halliday, M.A.K. 1994. An Introduction to Functional Grammar, London, Edward Arnold.
Iedema, R., S. Feez, and P.R.R. White. 1994. Media Literacy, Sydney, Disadvantaged Schools Program, NSW Department of School Education.
Kessler, C.S. 1987. 'Marx As Cultural Theorist: the Prehistory of Modern Anthropology', in Creating Culture - Profiles in the Study of Culture, Austin-Brooks, D.J. (ed.), Sydney, Allen & Unwin: 35-49.
Körner, H. 2001. 'Unpublished Ph.D Thesis'. University of Sydney.
Lakoff, G. 1972. 'Hedges: A Study in Meaning Criteria and the Logic of Fuzzy Concepts', Proceedings of the Chicago Linguistics Society 8: 183-228.
Lemke, J.L. 1992. 'Interpersonal Meaning in Discourse: Value Orientations', in Advances in Systemic Linguistics. Recent Theory and Practice, Davies, M. & Ravelli, L. (eds), London, Pinter Publishers.
Lyons, J. 1977. Semantics, Cambridge, UK., Cambridge University Press.
Martin, J.R. 1995a. 'Interpersonal Meaning, Persuasion, and Public Discourse: Packing Semiotic Punch', Australian Journal of Linguistics 15: 3-67.
--- 1995b. 'Reading Positions/Positioning Readers: JUDGEMENT in English', Prospect: a Journal of Australian TESOL 10 (2): 27-37.
--- 1997. 'Analysing Genre: Functional Parameters', in Genres and Institutions: Social Processes in the Workplace and School, Christie, F. & Martin, J.R. (eds), London, Cassell: 3-39.
--- 2000. 'Beyond Exchange: APPRAISAL Systems in English', in Evaluation in Text, Hunston, S. & Thompson, G. (eds), Oxford, Oxford University Press.
Roach, P. 1991. English Phonetics and Phonology - A Practical Course, Cambridge, Cambridge University Press.
Rothery, J. & Stenglin, M. in press. 'Interpreting Literature: The Role of APPRAISAL', in Researching Language in Schools and Functional Linguistic Perspectives., Unsworth, L. (ed.), London, Cassell.
Thibault, P. 1997. 'Mood and Eco-Social Dynamics', in On Subject and Theme: A Discourse Functional Perspective, Hasan, R. & Fries, P. (eds), The Hague, Benjamins: 51-90.
White, P.R.R. 1998. 'Telling Media Tales: the News Story As Rhetoric'. unpublished Ph.D dissertation, University of Sydney, Sydney.
--- 2000. 'Dialogue and Inter-Subjectivity: Reinterpreting the Semantics of Modality and Hedging', in Working With Dialog, Coulthard, M., Cotterill, J., & Rock, F. (eds), Neimeyer.
White, P.R.R. to appear. 'Attitude and Arguability: Appraisal and the Linguistics of Solidarity', Text - Special Edition on Appraisal.
1 The account set out here relies entirely on the work of Rothery, developed initially during research into the language of the visual arts for various DSP projects, as well as subsequent analysis by Rothery and Stenglin of the role of evaluation in secondary school English essays (Rothery and Stenglin in press). (For a review see Martin 1997: 24-26).
