Positioning Mispositioned: Theorising ‘Position’ at the Semantics–Context Interface

Doran, Martin & Zappavigna (2025: 56): 

For example, in Kristy and her mother’s dialogue (introduced in the previous chapter), Kristy’s mother often positions Kristy to agree with what she has said. For example, in (1), she uses the tag ‘don’t you?’, to suggest that she expects Kristy will support her position.

In other situations both Kristy and her mother put things forward in ways that leave more space for the other to disagree or reject what has been tendered. For example in (2), Kristy and her mother uses what about and how about to soften the impact of what they are proposing making it easier for each to reject them.

These examples also illustrate that what is being negotiated can vary. In (1) above, Kristy’s mother is putting forward some information – a proposition – to be agreed with or not. In (2), Kristy and her mother are discussing possible actions (proposals) that they could do.

Reviewer Comments:

Guest commentary by ChatGPT

On the surface, these observations are plausible enough. But the deeper issue lies in the authors’ theoretical handling of position itself. What is being positioned, and where? What kind of construct is a position?

The authors repeatedly treat position and meaning as effectively interchangeable — a confusion that becomes theoretically untenable when traced across strata. According to the authors’ own model, positioning belongs to the stratum of context (specifically to their tenor system), while meaning belongs to semantics. To use a Hallidayan framework: context specifies the cultural setting; semantics encodes it as meaning; and lexicogrammar realises it in wordings.

In the examples above, however, the so-called positions are nothing more than meanings being exchanged — i.e. interpersonal semantic selections. The tag don’t you? is a demand for a confirming response; how about I get you dressed? is a proposal realised in mood and modality. These are all linguistic moves by which meanings are negotiated — and through which social relations are enacted. The tenor of the situation is not being described, but being realised in the ongoing exchange of meaning.

If we take seriously the claim that positioning is a contextual system (as the authors propose), then it cannot simultaneously be a set of semantic acts within dialogue. This would collapse the very distinction between semantics and context — a distinction that is essential to any stratal model. Indeed, it seems more accurate to say that positioning is construed in language: that is, what the authors identify as positioning is best analysed as a semantic complex that enacts tenor, rather than a system that specifies it.

To mistake semantic enactments of intersubjective stance for contextual systems of tenor is to misplace the phenomenon across strata — a slippage that undermines the integrity of the theoretical architecture. Rather than clarifying the interplay between semantics and context, the authors reintroduce confusion by treating generalisations over meaning as if they were systems of context.

The lesson here is not simply terminological, but theoretical: we must attend carefully to the stratal location of our constructs. Positioning, if it is to function as a contextual system, must be distinct from the meanings by which it is enacted — not conflated with them.

Rebranding And Circular Reasoning

Doran, Martin & Zappavigna (2025: 46):

As these examples show, the use of engagement can produce a text that is very interpersonally nuanced – rendering a proposition that may or may not have been stated, while at the same time tendering a new proposition. 

This gives some insight into why humanities texts, which regularly draw heavily on engagement, can at times be such ‘heavy going’ (i.e., why they can have such strong interpersonal mass (Martin, 2017/2020) – they are responding to a range of stances in their academic community while at the same time trying to put forward a stance themselves (Doran, 2020a, 2020b; Hood, 2010, 2022).

Reviewer Comments:

[1] To be clear, these examples have simply illustrated the interpersonal semantic systems of ENGAGEMENT and SPEECH FUNCTION. The authors, however, have recategorised such propositions using the paradigmatic features 'rendering' and 'tendering', and misunderstood these semantic moves as parameters of the contextual system of tenor, despite 'tenor' referring to 'the statuses and role relationships; who is taking part in the interaction' (Halliday 1994: 390).

[2] To be clear, the authors’ explanation here is circular: humanities texts are said to be ‘heavy going’ because they exhibit strong ‘interpersonal mass’, which in turn is defined by their reliance on ENGAGEMENT resources — the very thing that supposedly makes them heavy going. No independent criterion is offered for identifying or measuring ‘interpersonal mass’ apart from the difficulty it is invoked to explain. As such, the term functions as a rebranding of the problem, not an explanation of it — offering semantic tautology in place of theoretical insight.

Misunderstanding Interpersonal Semantics as Context

Doran, Martin & Zappavigna (2025: 45-6):

Turning to the note option, these can be enacted through the heteroglossic option for entertain, where there is an acknowledgement of the possibility of different voices, but no explicit stance is taken.

A noting move in this kind of example will typically occur for the low and median positions of modality (Halliday & Matthiessen, 2014, pp. 178-179), e.g., using possible, perhaps, maybe, can, may etc. Higher modality, e.g., probably, likely, must, will, certainly etc., is more likely to realise some sort of support or reject (though see Chapter 3 for how this interacts with speaker and listener purview).

Reviewer Comments:

The authors propose ‘note’ as a subtype of ‘tender’, situated within their system of POSITIONING — which they treat as contextual (tenor). However, the example they give is clearly a move in an exchange — that is, a semantic choice in the interpersonal metafunction. 

‘Noting’ is the authors' rebranding of ‘entertaining’ of the ENGAGEMENT system (White 1998), as realised lexicogrammatically through modal expressions such as possible, perhaps, maybe, can, may, etc. These are resources of language, not of context. Once again, the authors conflate language and context, misunderstanding the stratified architecture of SFL.

Misunderstanding Interpersonal Meaning as Tenor

Doran, Martin & Zappavigna (2025: 45, 55):

Support can be done through a range of heteroglossic proclaiming resources (Martin & White, 2005). In this case a position is endorsed, pronounced or concurred with, as in (26), but like (25) it renders the position it tenders.¹² 

Similarly in (27) Kristy’s mother supports the position that she will be home about the same time as Dee’s big kids get home by using probably.

 

¹² This example also illustrates the role of engagement in realising internal rendering. Here the of course is not rendering an opinion on whether they support or reject Indigenous lands being stolen, but rather is supporting this as a linguistic act – i.e. supporting it as true. We can contrast this with an external support – Fortunately, Indigenous lands were stolen, which is definitely not what is being said. Heteroglossia when used for rendering is often used for internal rendering in this sense, functioning as it does to manage the play of voices.

Reviewer Comments:

[1] Here the authors again rebrand the semantic system of ENGAGEMENT (White 1998) as their contextual system of tenor. In terms of Halliday's semantic system of SPEECH FUNCTION, each instance of 'support' is an initiating move in an exchange: a statement. On the other hand, as Halliday (1994: 390) puts it:

Tenor refers to the statuses and role relationships; who is taking part in the interaction.

i.e. not to the language that enacts those statuses and role relationships.

[2] To be clear, in SFL theory the difference between of course and fortunately is not one of internal vs external ‘rendering’, but one of interpersonal meaning — specifically, between asseverative and qualificative comment Adjuncts (Halliday & Matthiessen 2014: 190–2). Of course realises the meaning ‘obviously’ (asseverative), while fortunately realises the meaning ‘luckily’ (qualificative). These are clearly linguistic resources — part of the speaker’s meaning potential — not contextual roles. The authors misattribute these semantic choices to their model of tenor.

Category Errors, Faulty Analogies, and Confusions Between Semantics and Context

Doran, Martin & Zappavigna (2025: 44-5):

Here we are emphasising the similarities in tenor between rejection and support done dialogically and monologically.  

In terms of discourse semantics, this indicates similarities between heteroglossia, which offers resources for managing multiple voices, and negotiation (dialogica), which offers resources for managing multiple turns.  

Put another way, we are suggesting a parallel between (23), where the tendering and rendering is established across two turns, and (24) where the tendering and rendering occurs in a single turn.

Reviewer Comments:

[1] Here again the authors misunderstand language choices (tendering and rendering, rejection and support) as the cultural context of language (tenor). As Halliday (1994: 390) puts it:

Tenor refers to the statuses and role relationships; who is taking part in the interaction.

[2] Here the authors conflate the speaker’s semiotic resources — heteroglossia and negotiation — with the linguist’s modelling of those resources. To be clear, heteroglossia is the speaker’s resource for acknowledging other possible points of view (‘voices’), not a resource for “managing” them. Likewise, negotiation is the speaker’s system of potential moves in an exchange, not a resource for “managing” multiple turns.

[3] Here the authors set up the following proportionalities:

dialogue is to monologue as

heteroglossia is to negotiation as

two turns is to single turn

To be clear, proportional analogies rely on structural or functional equivalence: if A is to B as C is to D, then the relation between A and B must be of the same kind as the relation between C and D — whether it’s one of scale, category, realisation, or function.

But in this case, the analogy fails on all counts. Dialogue and monologue differ in interactional structure, not functional type; negotiation and heteroglossia belong to different semantic systems with distinct organising principles; and turns and voices are not even comparable units — the former are structural units of exchange, the latter semantic projections of alignment. No relational consistency holds across the three pairs.

[4] In terms of SPEECH FUNCTION, (23) is an initiating statement followed by a discretionary response — a contradiction — while (24) is simply an initiating statement with no response (Halliday & Matthiessen 2014: 137). The authors rebrand these moves in the exchange structure as ‘tender’ and ‘reject’, and misclassify them as features of context (tenor) rather than of language (semantics).

Confusing Language With Theory

Doran, Martin & Zappavigna (2025: 39):

More generally, the relatively simple distinctions we have introduced so far help us deal in broad terms with two of our main concerns with regard to dialogue. First, the options in rendering offer a set of resources for negotiating feelings in dialogue, and so bring together attitude and exchange in discourse semantics. Second, the conflation of tendering and rendering offers a resource for chaining stretches of dialogue together and realising prosodic and indefinitely extended phases of chat.

Reviewer Comments:

[1] As previously demonstrated, the relatively simple distinctions the authors have introduced are those of Halliday's interpersonal semantics, rebranded and misunderstood as tenor.

[2] This confuses language with theory. These resources are those of language. The conflation of tendering and rendering is a way of modelling these resources. Moreover, this conflation merely acknowledges that responses can initiate further responses.

Problems With The Authors' 'Resource Guide'

Doran, Martin & Zappavigna (2025: 26-7):

As will be clear form this introduction, tenor is one module in the overall model of language and social context assumed here. Because of its position in the architecture of SFL (between genre and language and alongside field and mode) we draw on a wide range of descriptions of English – some of which may not be familiar to readers. Below we note some useful introductory resources for key descriptions, as well as the key reference books for each area.

Paralanguage
Martin, J. R. & M. Zappavigna (2019) Embodied meaning: a systemic functional perspective on body language. Functional Linguistics 6(1). https://doi.org/10.1186/s40554-018-0065-9 

Ngo, T., Hood, S., Martin, J.R., Painter, C., Smith B.A. and Zappavigna, M. (2022) Modelling Paralanguage Using Systemic Functional Semiotics. London: Bloomsbury. …

Discourse semantics
Chapters 2 (appraisal) and 7 (negotiation) of Martin and Rose (2007). Working with Discourse: Meaning beyond the Clause. London: Continuum. 

Martin, J.R. (1992) English Text: System and Structure. Amsterdam: John Benjamins.

Reviewer Comments:

[1] To be clear, the misunderstanding of strata and metafunctions as modules has persisted since Martin (1992: 390, 488):

Each of the presentations of linguistic text forming resources considered above adopted a modular perspective. As far as English Text is concerned this has two main dimensions: stratification, and within strata, metafunction. …

The problem addressed is a fundamental concern of modular models of semiosis — namely, once modules are distinguished, how do they interface? What is the nature of the conversation among components?

[2] For a meticulous review of Martin & Zappavigna (2019) that identifies many of its misunderstandings of SFL Theory, see Martin & Zappavigna's Model Of Paralanguage.

[3] For a meticulous review of Ngo et al. that identifies many of its misunderstandings of SFL Theory, see Modelling Paralanguage Using Systemic Functional Semiotics.

[4] For a meticulous review of Martin and Rose (2007) that identifies many of its misunderstandings of SFL Theory, see Working With Discourse: Meaning Beyond The Clause.

[5] For a meticulous review of Martin (1992) that identifies many of its misunderstandings of SFL Theory, see English Text: System And Structure.

Seriously Misunderstanding Stratification

Doran, Martin & Zappavigna (2025: 17, 10):

Realisation is a hierarchy of abstraction, with higher strata realised by patterns of meaning at lower ones. … The realisation hierarchy we assume here was outlined in Figure 1.3 above (with genre as a more abstract pattern of register patterns, register as a more abstract pattern of discourse semantic patterns and so on).

 Blogger Comments:

This seriously misunderstands stratification. Less importantly, realisation is not a hierarchy, but the relation between levels of symbolic abstraction, as between axes or between strata. The hierarchy in this case is the hierarchy of strata: stratification.

Most importantly, higher strata are not "realised by patterns of meaning at lower ones". On the one hand, the function of the stratification hierarchy is to relate meaning, semantics, to other levels of symbolic abstraction: lexicogrammar and phonology below, and the context of language above. The authors' misunderstanding reflects Martin's mantra 'all strata make meaning' which confuses stratification (realising meaning) with semogenesis (making meaning).

On the other hand, the "patterns" of each stratum are the systems of each stratum. The term "patterns" suggests a confusion with instantiation: the patterns of instantiation that distinguish the variants on the cline of instantiation from each other.

Problems With The Notion Of Context Skewing Language System Probabilities

Doran, Martin & Zappavigna (2025: 11-12):

The resonance between discourse semantic systems and register variables is annotated in Figure 1.5 – with field skewing probabilities in IDEATION and CONNEXION systems, mode skewing probabilities in PERIODICITY and IDENTIFICATION systems and tenor skewing probabilities in NEGOTIATION and APPRAISAL systems.

Reviewer Comments:

[1] As previously explained, context-metafunction resonance means

• differences in field are realised by differences in ideational meaning,
• differences in tenor are realised by differences in interpersonal meaning, and
• differences in mode are realised by differences in textual meaning.

As previously demonstrated, the authors misunderstand context-metafunction resonance as an exclusive relation between context and language on the basis of metafunction. 

[2] As previously explained, this misunderstands the relation between context and registers of language in SFL Theory. Halliday (2005 [1991]: 60):

Register variation can be defined as the skewing of (some of) these overall probabilities, in the environment of some specific configuration of field, tenor and mode. It is variation in the tendency to select certain meanings rather than others, realising variation in the situation type.

That is, different registers are defined by having different probabilities of selection of semantic and lexicogrammatical features, with each register realising a different configuration of field, tenor and mode variables, and each of these defining a different situation type. Importantly, the skewing of probabilities is not caused by the contextual variables of field, tenor and mode, if only because the relation between context and language is not causal (enhancement), but one of intensive identity (elaboration).

Misunderstanding Tenor And The Architecture Of Language

Doran, Martin & Zappavigna (2025: 11):

Through the history of SFL, a range of scholars have tried to make tenor do a very wide range of things. But with a more elaborated model of language and context that includes discourse semantics, genre and, as we will discuss below, instantiation and individuation, the division of semiotic labour can be more comfortably distributed across the model.

Reviewer Comments:

[1] This is the logical fallacy known as ipse dixit: a bare assertion unsupported by evidence. Here it also serves as an instance of the straw man fallacy, since it sets up an imaginary position to be argued against in the work of the authors; cf. tilting at windmills. 

To be clear, tenor merely concerns who is taking part in a speech situation. It is simply the theoretical projection of the interpersonal metafunction onto context. It will be seen in this review that it is the authors themselves who try "to make tenor do a very wide range of things", by extending it to include interpersonal semantics and to serve as a 'principle of instantiation' for language.

[2] This misunderstands the architecture of language proposed by SFL Theory. The "semiotic labour" is already distributed along all the dimensions proposed. For example, Halliday & Matthiessen (2014: 32):

Problems With The Discourse Semantics Framework

Doran, Martin & Zappavigna (2025: 11):

The model of discourse semantics we assume here is also organised by metafunction (Hao, 2020; Martin, 1992; Martin & Rose, 2003/2007; Martin & White, 2005). As outlined in Figure 1.5, its systems comprise: 

• the ideational systems of IDEATION, which map how we construe our experience in terms of the occurrences, states, entities and qualities of language, and CONNEXION, which articulates how we connect these ideational meanings and stretches of text together into larger sequences of experience or rhetoric; 
• the textual systems of PERIODICITY, which are concerned with how we foreground and background information as waves of prominence, and IDENTIFICATION, which map how we introduce and track people, places and things; and most significantly for this book, 
• the interpersonal systems of NEGOTIATION, which present the resources we have for dialogue, and APPRAISAL, which map the resources we have for evaluating, amplifying and arranging interpersonal meanings.

Reviewer Comments:

[1] To be clear, the experiential discourse semantic system of IDEATION in Martin (1992) is Martin's rebranding of his misunderstandings of (textual) LEXICAL COHESION in Halliday & Hasan (1976). However, it has more recently become a rebranding of the ideational semantics — both experiential and logical — of Halliday & Matthiessen (1999) by Martin's former student Hao. See, for example:

Misrepresenting Halliday & Matthiessen's Semantics As Martin And Hao's Discourse Semantics 

[2] To be clear, the logical discourse semantic system of CONJUNCTION in Martin (1992), renamed CONNEXION by Martin's former student, Hao, is Martin's rebranding of his misunderstandings of (textual) grammatical cohesive CONJUNCTION in Halliday & Hasan (1976), confused with (logical) grammatical CLAUSE COMPLEXING in Halliday (1985).

[3] To be clear, the textual discourse semantic system of PERIODICITY in Martin & Rose (2007) is Martin's rebranding of writing pedagogy. No system networks have ever been devised, and the structures each consist of only one functional element (e.g. hyper-Theme). Moreover, in Martin (1992: 393), these rebrandings were modelled as interstratal interaction patterns, rather than discourse semantic systems.

[4] To be clear, the textual discourse semantic system of IDENTIFICATION in Martin (1992) is Martin's rebranding of his misunderstandings of grammatical REFERENCE in Halliday & Hasan (1976). 

[5] To be clear, the interpersonal discourse semantic system of NEGOTIATION in Martin (1992) is Martin's rebranding of his misunderstandings of SPEECH FUNCTION in Halliday (1985). 

[6] To be clear, for Halliday, APPRAISAL is a system between grammar and lexis, in which attitudinal assessments are realised by lexical items.

∞

For detailed evidence of the myriad misunderstandings of SFL theory in the 'discourse semantic' framework, see here (Martin 1992) and here (Martin & Rose 2007).

The Models Of Discourse Semantics And Paralanguage In This Book

Doran, Martin & Zappavigna (2025: 9, 10):

For language, this book assumes a model of language comprising the strata of phonology (or graphology for written language or the embodied signing of sign languages), lexicogrammar and discourse semantics (Figure 1.3).

We call the most abstract stratum of language discourse semantics, not semantics, to emphasise its orientation to larger text patterns (Martin, 1992; Martin & Rose, 2003/2007); this distinguishes it from the clause semantics of say Halliday and Matthiessen (1999) or Hasan (e.g. Hasan, Williams, Cloran, & Lukin, 2005). For grammatical analysis of English examples we rely on Halliday (1985 and subsequent editions); for phonological analysis we draw on Halliday and Greaves (2008); and for paralanguage (i.e., body language, gesture and the like) we reference Ngo et al. (2022).

Reviewer Comments:

[1] To be clear, Martin's discourse semantics arose as his rebranding of lexicogrammatical cohesion (Halliday & Hasan 1976), the non-structural system of the textual metafunction. For the myriad misunderstandings of SFL theory in this model, see here (Martin 1992) and here (Martin & Rose 2007).

[2] To be clear, Martin's ideational discourse semantics is now a rebranding of the semantics of Halliday & Matthiessen (1999) by Martin's former student, Hao. See, for example:

Misrepresenting Halliday & Matthiessen's Semantics As Martin And Hao's Discourse Semantics

[3] For evidence of the theoretical misunderstandings and plagiarism in this work, see:

Modelling Paralanguage Using Systemic Functional Semiotics