I will briefly demonstrate the heteroglossic value in the context of modals of what Halliday terms probability (Halliday 1994). Under traditional analyses, such meanings are typically construed either as indicating either lack of commitment to truth value by the speaker, some uncertainty, or as reflecting negatively on the reliability of the proposition as knowledge. Thus Lyons contrasts what he terms the `subjectivity' of the modal meaning with the `objectivity' of `bare assertions' and describes the 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).
Thus 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. There is an implication that the overriding purpose of communication is to exchange truth values or certain knowledge and that these modal values are introduced only in communicatively non-optimal circumstances. Accordingly 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 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.
The literature which explores the functionality of modals, for example, in their actual context of use has demonstrated overwhelmingly, I believe, that that truth-functional explanations are inadequate. The literature reveals such modals to be multi-functional and to be just as likely to reflect politeness or deference as they are to indicate epistemological unreliability. I will demonstrate this point briefly by reference to the two texts set out below.
The heteroglossic approach provides a principled account of this multi-functional semantic. Rather than seeing these values as oriented to coding a speaker's individual state of mind, as a reflex of the speaker's state of knowledge, we see them as operating to reflect the process of interaction or negotiation within a text between the alternative socio-semiotic positions put at risk by that text. The value signals that the meanings at stake are subject to heteroglossic negotiation and 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.
