Internal Purview and Tone: Repackaging Mood Without Gain

Doran, Martin & Zappavigna (2025: 75):

A further contrast in purview for open propositions is shown in (39), from Halliday (1970: 22), where a wh-interrogative is used with different tones – a falling tone (tone 1) in the first example (39a.) and a rising tone (tone 2) in the second example (39b.).

(39) a. //1 where are you going//

        b. //2 where are you going//

As Halliday notes, “the first [39a with a falling tone] is a normal question, neither abrupt nor deferential, while the second is deferential: it is a question accompanied by a request for permission to ask ‘where are you going, may I ask’.” (Halliday 1970: 22). We can understand this by considering again the distinction between internal and external readings of these positions. Looking externally, in terms of the ‘information’ being asked for, both simply pose a question – they have the same purview, namely – speaker: + listener. Viewed internally in terms of their role as speech acts initiating exchanges on the other hand, they differ in their purview. Example (39a.), with a falling tone, does not seek permission in any way to ask the question – internally it can be read as an assert (+internal speaker purview; – internal listener purview). By contrast (39b.), with a rising tone, indicates some tentativeness (Halliday & Matthiessen, 2014, p. 169) – in some sense asking permission to initiate the exchange. This example can be read internally as a pose (– internal speaker purview; + internal listener purview). 

Similarly, in the Australian English dialect spoken by the first and third authors of this book, open propositions can be internally shared (+ internal speaker purview; + internal listener purview) by using non-lexical tags such as ay (/æɪ/) on a falling tone (tone 1).

(40) But what do we do, ay?

This example is still asking the question (externally speaking); but through the tag it is making it clear that this is a question which the interlocutors probably both share.

ChatGPT Comments:

 Several points stand out:

---

1. Redundant re-description of mood and tone

• The authors use Halliday’s falling/rising tone distinctions to motivate internal purview, but SFL already captures this through Mood + Key + intonation patterns.

• The “internal vs external” distinction is a semantic overlay on phenomena that are already systematically handled by Mood: falling tone = normal declarative/interrogative force; rising tone = tentative or deferential. Their framework does not add explanatory power — it merely renames these distinctions in terms of ± internal speaker/listener purview.

---

2. Introduction of “internal purview” adds analytic complexity without gain

• By splitting each proposition into “internal” vs “external” purview, the authors create a nested system on top of the existing SFL apparatus.

• This doubling is not justified empirically: there is no demonstration that these “internal” values capture patterns beyond what tone, key, and mood already explain.

---

3. Conflation of surface markers and systemic categories

• Tone, lexical tags, and intonation contours are used as evidence of purview categories.

• But purview is positioned as a systemic feature of context, not a phonetic or prosodic feature. By grounding purview in tone and tags, they blur the line between language and context, which was already a problematic aspect of the earlier discussion.

---

4. Circular reasoning

• They argue: rising/falling tone distinguishes internal purview, then show examples with rising/falling tone to demonstrate internal purview.

• This mirrors the circularity seen in previous critiques: the data are interpreted to fit the framework rather than to test it.

---

Overall assessment

• The excerpt exemplifies the recurring problem: the “purview” system often renames or overlays phenomena already described by SFL (Mood, intonation, key) without adding genuine analytic leverage.

• The introduction of internal/external purview, and the use of sociolectal features like ay, highlights the fragility and non-generalisability of the framework.

Purview Collapsing into Mood: The Misdescription of Wh-Questions

Doran, Martin & Zappavigna (2025: 74-5):

The examples above all focused on complete propositions and proposals. For proposition or proposals that are open, often involving wh– questions, the default purview is pose (–speaker purview; + listener purview). We can see this in (37), where Kristy poses a wh-question to her mother.

Though as example (30) above shows, open propositions can have their purview varied – in this case as aired, through I wonder (replayed in (38)).

 

ChatGPT Comments:

 This continues the same issues we’ve seen earlier:

1. Circularity of “default” — The claim that open wh– questions default to pose (– speaker purview; + listener purview) is not an explanation but a tautological restatement of their own invented system. They say “the default is pose” because their grid requires it, not because any descriptive evidence demands it.

2. Category slippage — They oscillate between proposition type (wh– interrogative) and purview value (pose vs air), treating one as if it naturally “defaults” into the other. But they provide no systemic grounding: in SFL, the default mapping between interrogatives and exchange roles is already accounted for in mood structure. Their “default purview” adds nothing except redundancy.

3. Recycling the same example — Example (37) is simply a restaging of Kristy’s earlier wh– question (already discussed in 34). And (38) is again the coconut cream hoppers example we’ve seen twice already. This recycling suggests they don’t have enough genuinely distinct data to sustain the typology, so they keep dressing the same examples in new conceptual clothing.

4. Misuse of systemic concepts — In SFL, defaults are structural (e.g. interrogatives defaulting to “demand information”). Here, the “default purview” is not a systemic default but an analytic convenience. Calling it a “default” risks misleading readers into thinking the model is grounded in the grammar, when it is in fact grounded in their newly constructed grid.

---

More specifically:

Excerpt under review:

“For proposition or proposals that are open, often involving wh– questions, the default purview is pose (– speaker purview; + listener purview). … Though as example (30) above shows, open propositions can have their purview varied – in this case as aired, through I wonder.”

---

1. The illusion of a “default”

The authors claim that open wh– questions “default” to pose. But this “default” is only true by stipulation: they have defined pose as (– speaker purview; + listener purview), then observe that wh– interrogatives are typically structured as demands for information from the addressee. Rather than discovering something new, they are redescribing the already well-established systemic fact that interrogatives enact demands for information. The novelty of “purview” collapses into circular restatement.

---

2. Redundancy with SFL mood system

In Halliday & Matthiessen’s model, the correlation between mood type and speech function is already systemic:

• interrogative ↔ demand for information

• declarative ↔ statement of information

• imperative ↔ demand for goods-&-services.

This account already captures Kristy’s wh– question in (37). To claim that “the default purview is pose” adds no explanatory depth; it simply repackages mood semantics into the “purview” grid, with no new descriptive reach.

---

3. Category slippage: proposition ↔ purview

The authors conflate two distinct dimensions:

• proposition type: whether the clause is open (wh–) or closed (polar).

• purview type: who is said to “own” or “share” responsibility for knowing.

They present the mapping as if the clause type “naturally” entails a purview, but in fact the “purview” category is artificially imposed. This slippage generates the illusion of systemic necessity where there is only analytic projection.

---

4. Reliance on recycled examples

The data presented are not fresh demonstrations but recycled:

• Example (37) repeats Kristy’s wh– question from (34).

• Example (38) repeats the “coconut cream hoppers” case already presented twice.

This recycling betrays the fragility of the framework. If the typology were descriptively robust, one would expect a rich array of distinct examples across Hasan’s corpus. Instead, the same handful of cases must be dressed up in different theoretical guises to sustain the typology.

---

5. Misuse of “default” as systemic category

In SFL, “default” is a technical concept tied to systemic probability: certain features are unmarked, carrying higher systemic probability in context. Here, “default purview” is neither probabilistic nor systemic; it is merely an analytic shorthand for “what their table already predicts.” By borrowing SFL’s terminology, the authors create an illusion of system-internal motivation that their model cannot supply.

---

Conclusion

This passage illustrates the fundamental weakness of the purview typology: it offers no explanatory insight beyond what mood semantics already covers, yet it recycles SFL’s terminology (“default”) and Hasan’s data to generate the appearance of theoretical novelty. The result is circular, redundant, and analytically fragile.

Interpersonal Fog: When Linguistic Resources Are Mistaken For Social Roles

Doran, Martin & Zappavigna (2025: 56-7):

Negotiating meaning is a two-way affair. People can affiliate, align, gossip, attack, chat etc. by rendering meanings, but they can also do so by tendering meaning. The way people tender meanings implicates a range of interpersonal systems – in particular those of NEGOTIATION, concerned with how people exchange meanings in dialogue (Berry, 1981a; Martin, 1992; Ventola, 1987) and ENGAGEMENT, concerned with how people manage the play of different voices (Martin & White, 2005). As we mentioned in the previous chapter, we are particularly concerned with how these two resources interact as they work together. Work by Muntigl (2009), Kim et al. (2023) and Zhang (2020a, 2020b, 2020c, 2021) in particular have highlighted how people position each other as they work toward consensus. These systems in turn implicate a wide range of interpersonal grammatical systems – including the core system of MOOD as well as systems often positioned as supporting MOOD choices, such as TAGGING and MODALITY (Halliday & Matthiessen, 2014). The way all of these systems engage with tone choices in phonology is also part of the picture (Halliday & Greaves, 2008). 

In short, to understand how people organise social relations, it is not enough to just understand how people react; we also have to look carefully at how people put meaning forward – we draw on a full set of interpersonal resources in language.

Reviewer Comments:

[1] This continues the confusion between context and language. Specifically, the authors mistake the interpersonal linguistic resources for realising tenor for tenor itself, the roles and relationships of the interlocutors in the cultural setting. The authors simply do not understand what different levels of symbolic abstraction actually means, and the sense in which levels are distinct from each other. It is from this failure of understanding that this entire publication proceeds.

[2] The authors here demonstrate their lack of understanding of stratification by using the term 'implicate', which does not describe the relation of symbolic abstraction, instead of the term 'realise', which does. See Halliday & Matthiessen (2014: 269).

[3] To be clear, SFL Theory models language in context. The notion of understanding 'how people organise social relations' by just understanding 'how people react' is a rhetorical straw man.

Misreading Modality

Doran, Martin & Zappavigna (2025: 47-8):

Looking at the I think now, as (Halliday & Matthiessen, 2014, pp. 686-687) notes, the I think here is unlikely to be providing a full proposition. We can see this by the fact that if we were to tag the clause, we would more likely tag in relation to the you’re upset, rather than the I think; We would more likely say: I think you’re upset because the TV wasn’t working, aren’t you; as opposed to I think you’re upset because the TV wasn’t working, aren’t I. Similarly, if Kristy was to support what her mother was saying, she would likely support the ‘you’re upset…’, rather than the ‘I think’ with Yeah, I am, rather than Yes, you do. This is because the I think is not presenting a proposition about thinking, but rather being used metaphorically to mean something like ‘probably’. In Halliday’s terms, it is an interpersonal metaphor realising modality. In this sense, the I think is in fact realising support for the idea that Kristy is upset because the TV isn’t working. As we will discuss in Chapters 4 and 5, the I think also makes explicit that what is being tendered is a personal opinion and it lowers the stakes of the position.

The I don’t think in the first line functions similarly to the second line, but realises a rejection. We can see this again by the fact that it is unlikely that we will reject the ‘thinking’ (Yes you do), but also the fact that the negation can be easily transferred to the you’re upset without significantly changing the meaning: I think you’re not really upset about me going (Halliday & Matthiessen, 2014, p. 689). So Kristy’s mother has tendered two propositions in opposition to each other in order to reject one and support the other – i.e., that Kristy is not really upset about her going but that she is upset because the TV wasn’t working.

Reviewer Comments:

Guest commentary by ChatGPT

This discussion attempts to leverage Halliday’s account of interpersonal metaphor to justify the claim that I think in I think you’re upset because the TV wasn’t working constitutes a form of support in the authors’ tenor system of positioning. The claim appears to be that I think realises support for the proposition that follows, rather than functioning as a proposition in its own right.

[1] There is a serious category error here: I think is rightly analysed (in Halliday & Matthiessen 2014) as an interpersonal grammatical metaphor for modality, not as a move in a system of tenor-positioning such as “support” or “reject.” The authors ignore the distinction between interpersonal grammar and contextual tenor, collapsing them into one another in order to preserve the illusion of theoretical innovation.

[2] The suggestion that I think somehow “realises support” misconstrues both the nature of modality and the analytic function of interpersonal metaphor. I think modifies the speaker’s commitment to the proposition; it does not itself “support” that proposition in a second-order sense.

[3] Worse still, the authors treat grammatical diagnostics such as tag questions and preferred responses (e.g., aren’t you vs. aren’t I) as if they confirm the categories of their own invented tenor system. But these diagnostics are evidence for clause structure and mood, not for tenor roles or speaker positioning. This is a classic case of retrofitting grammatical facts to support a speculative model.

[4] Finally, the apparent opposition between I think you’re not really upset about me going and I think you’re upset because the TV wasn’t working is a red herring. The former encodes a single modalised proposition; it is not a rejection of a second speaker's stance. The claim that Kristy’s mother “tenders two propositions in order to support one and reject the other” is interpretive projection masquerading as analysis.

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.

The Notion That A Repeated Proposition Is Not A Proposition

Doran, Martin & Zappavigna (2025: 44-5, 55n):

¹⁰ There is a question here as to whether No, I won’t additionally tenders a negative proposition, as we argued for the full clauses in (20)-(22) and (24). Our interpretation is that replaying the Subject and Finite without the Residue, as is done in No I won’t, means the turn is not putting forward any new proposition, and so is best read as simply rendering – reacting to what was said previously. It is only when there is some adjustment of the Mood (Subject and Finite) – what Halliday calls a shift in the modal responsibility (Halliday & Matthiessen, 2014) – or the specification of Residue is there a new proposition tendered (Martin, 1992, p. 461ff). This is because any further rendering involving the same strategy will simply replay the Mood of the initial position:

You will be home late as well. – No I won’t. – Yes you will. –No you [sic] won’t etc. 

Having said that, some instances with this configuration would tender a new proposition, such as if the primary tone was placed on either the Subject and Finite to suggest some sort of contrast: 

- You will be home late as well. – No, I won’t; 

or if the Mood Adjunct and Mood are given on different tone groups: 

//No//^ I won’t//. 

Ultimately, evidence as to whether there is a new proposition being tendered will depend on how it is negotiated in the following discourse.

Reviewer Comments:

[1] To be clear, here the authors confuse 'proposition' with 'new proposition', leading them to the self-contradictory conclusion that a repeated proposition is not a proposition.

[2] In terms of speech function, this exchange features four moves, all of which are propositions:

• You will be home late as well is the statement that initiates the exchange;
• No I won’t is a responding statement that contradicts the previous statement;
• Yes you will is a responding statement that contradicts the previous statement;
• No you I won’t is a responding statement that contradicts the previous statement.

See Halliday & Matthiessen (2014: 137).

[3] In the authors' terms, the proposition constitutes a 'tender' if it is subsequently 'rendered' (responded to), as in all the examples they provide.

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):

The Misunderstanding Of Stratification On Which This Work Depends

 Doran, Martin & Zappavigna (2025: 8):

In this book we will treat social context as a semiotic system realised through patterns of language (Figure 1.1).

This model conceives of context and language as being intricately related and co-developed, with each continually impacting the other (what Martin (2011), adopting a term suggested by Chris Cléirigh, calls a ‘supervenient’ perspective on social context). 

Such an approach contrasts with perspectives which either treat social context as an independent phenomenon alongside language (that is, at best, correlated with language choices but otherwise having no co-genetic relation) or alternatively as a circumvenient phenomenon in which language is in some sense embedded in context but where there is only a one-way determination – so that context determines language choice, but language has no or minimal effect on context. 

Our project, in other words, involves working towards a model in which social context is realised through recurrent patterns of meaning – where realisation involves a two-way process of mutual actualisation between language and context. Accordingly anything we say about social context is an abstraction from recurrent patterns of semantic choice (Martin, 2014) – a relationship Lemke (e.g. 1995) refers to as metaredundancy.

Reviewer Comments:

[1] To be clear, in SFL Theory, context is the culture as a semiotic system that is realised in language (inter alia). The ideational dimension of context, field, is what is going on, in terms of the culture; the interpersonal dimension of context, tenor, is who is taking part, in terms of the culture; and the textual dimension of context, mode, is the role played by language, in terms of the culture.

[2] To be clear, in SFL Theory, the stratal relation between context and language is one of intensive symbolic identity (realisation). That is, context and language form the one identity, with context as the higher level of symbolic abstraction, and language as the lower level. On the other hand, the "co-development" of context and language, is distinct from their stratal relation, and here refers to the process of logogenesis: the development of the text. That is, the authors here have confused stratification with semogenesis.

[3] To be clear, in SFL Theory, the "co-development" of context and language in logogenesis is the development of the one identity, viewed at two different levels of symbolic abstraction. Two different levels of symbolic abstraction cannot "impact" each other. For example, a pipe and a representation of a pipe cannot "impact" each other.

[4] This is misleading. On the one hand, it misrepresents Cléirigh's use of the term 'supervenient'. Cléirigh used the term to refer to the emergence of higher levels of organisation in complex systems, as in the emergence of chemical systems from physical systems, and of biological systems from chemical systems. That is, it covers similar ground to Halliday's (2002 [1996]: 388) evolutionary typology of complex systems. On the other hand, contrary to the implication, Cléirigh did not apply the term 'supervenient' to the relation between context and language. Instead, he applied it to the three strata of language, and conceived of these as embedded in context:

(This is a view that Cléirigh soon abandoned as his understanding of SFL Theory improved.)

[5] To be clear, these 'other perspectives' are the authors' own misunderstandings of work in SFL Theory, and so function here fallaciously as straw men to be attacked. See further below.

[6] To be clear, in SFL Theory, context and language are not independent, since they are two aspects of the one identity, and their "co-genetic relation" is the development of this identity in logogenesis.

[7] To be clear, 'circumvenient' was the term Cléirigh (almost facetiously) suggested to Martin when Martin rejected the term 'embedding'.

[8] To be clear, in SFL Theory, there can be no determination between context and language because the logical relation between them is elaboration (intensive identity), not causal enhancement (circumstantial identity). What is true is that during logogenesis, previous language instantiations probabilise following language instantiations, and these potentially construe changing contextual features, as when a speaker changes the subject (second-order field).

[9] To be clear, the model that the authors are working toward in this project is therefore based on a confusion of stratification ('realisation') with instantiation ('patterns of meaning'). In SFL Theory, stratification means that the systems of context are realised by the systems of language. Patterns of meaning, on the other hand, are created in the instantiation process of logogenesis: the selection of features and the activation of their realisation statements, and constitute variation along the cline of instantiation.

[10] To be clear, the model that the authors are working toward in this project is therefore based on a fundamental misunderstanding of realisation. In SFL Theory, realisation is one-way relation: the lower level of symbolic abstraction (language) realises the higher level (context). As such, context cannot realise language. Moreover, the use of the term 'actualise' here invites confusion with instantiation as the actualisation of potential as instance. Again, the authors have confused stratification with semogenesis: the instantiation of potential in logogenesis.

[11] As will be seen, in the model that the authors are working toward in this project, their misunderstanding of stratification leads them to mistake the lower level of symbolic abstraction (semantic choice) for the higher level they plan to model (context), thereby invalidating their model and their project.

[12] This is a serious misunderstanding of metaredundancy. As the term implies, 'metaredundancy' is a redundancy on a redundancy. Applied to the stratification hierarchy, it means that semantics is redundant on the redundancy of lexicogrammar and phonology, or alternatively, that the redundancy of semantics and lexicogrammar is redundant on phonology. Applying the term to just two levels of symbolic abstraction, context and semantics, is therefore nonsensical.