Why Martin & Rose Have Focussed On Discourse Semantics Rather Than Lexicogrammar

Martin & Rose (2007: 295-6):

As noted in our Introduction, we have had to make some decisions about what regions of the language system to focus on in this book. Language is an immensely complex phenomenon, no less than the contexts of social life that it realises, and discourse analysis is a very large and growing field of practice, so our focus has been on providing the tools that analysts can use to start exploring these domains. To this end we have focused on systems at the level of discourse semantics rather than the levels of lexicogrammar or social context, although we have touched on these lower and higher level systems at certain points. 

In particular we have explored discourse semantic resources for enacting social relations through appraisal and negotiation, for construing fields of experience through ideation and conjunction, and for presenting our enactments and construals as meaningful text-in-context through identification and periodicity. 

As far as discourse semantics is concerned, one set of textual resources we have not dealt with here is substitution and ellipsis — in part because we cannot enhance the accounts given in Halliday and Hasan’s 1976 Cohesion in English, and in part because rehearsing them would involve a lot of additional grammatical description. Martin’s 1992 English Text also contains discussions of cohesive harmony and modal responsibility that we have not developed here; and it also outlines a model of context, about which we’ll now say something further.

Blogger Comments:

[1] To be clear, for Halliday, it is the grammar that is essential for discourse analysis. Halliday (1994: xvi-xvii):

The current preoccupation is with discourse analysis, or 'text linguistics'; and it has sometimes been assumed that this can be carried on without grammar — or even that it is somehow an alternative to grammar.  But this is an illusion.  A discourse analysis that is not based on grammar is not an analysis at all, but simply a running commentary on a text … the exercise remains a private one in which one explanation is as good or as bad as another.

A text is a semantic unit, not a grammatical one.  But meanings are realised through wordings; and without a theory of wordings — that is, a grammar — there is no way of making explicit one's interpretation of the meaning of a text.

However, as demonstrated in painstaking detail here, much of Martin's (1992) discourse semantics is grammar: the model of textual cohesion in Halliday & Hasan (1976), relabelled and relocated from Halliday's grammatical stratum to Martin's discourse semantic stratum. Martin & Rose have focussed on discourse semantics, rather than grammar, simply because discourse semantics is presented as Martin's theorising. 

[2] For the theoretical problems with these systems, and inconsistencies in their exposition, click on the following links:

• appraisal here
• negotiation here and here
• ideation here and here
• conjunction here and here
• identification here and here
• periodicity here and here.

[3] To be clear, in SFL Theory, experience is construed as the ideational meaning of language, whereas field is the ideational dimension of the culture that is construed by language. The distinction is important epistemologically, since 'experience' is what is transformed into meaning.

[4] To be clear, the reason why Martin & Rose are unable to 'deal with' ellipsis-&-substitution is because the relations involved are lexicogrammatical, not semantic. Halliday (1994: 316):

Nevertheless, Martin (1992) locates ellipsis-&-substitution both interpersonally within the mood system of grammar (pp34-5) and textually within the identification of discourse semantics (pp 100-2, 135, 144). The theoretical inconsistency here is thus metafunctional as well as stratal.

[5] For the misunderstandings of Hasan's cohesive harmony in Martin (1992), see the clarifying critiques here.

[6] For the misunderstandings of Halliday's modal responsibility in Martin (1992), see the clarifying critiques here.

[7] For the theoretical inconsistencies that invalidate the model of context in Martin (1992), see the clarifying critiques here.

Misconstruing External Conjunction As Internal

Martin & Rose (2007: 263-4):

In contrast, the Cost of Courage report is organised around the people that struggled against apartheid, who are identified at the beginning of each paragraph, from the policy of apartheid to the comrades in the struggle, to I (Mandela), to we, to every man:

The last paragraph then begins In that way… to culminate Mandela’s explanation of the price his family paid for his commitment to the struggle:

The global scaffolding resource operating here is identification — initially of people, and finally text reference in a circumstance of manner. Expressions of this kind are closely related to internal conjunction in their global text orchestrating function, as we could see by substituting the manner conjunction thus for in that way.

Blogger Comments:

[1] Clearly, the policy of apartheid does not denote the people who struggled against apartheid.

[2] To be clear, this is identification in the sense of ideational denotation, which, as previously demonstrated, Martin confuses with textual reference and interpersonal deixis in his model of textual discourse semantics.

[3] To be clear, in that way serves as a cohesive conjunction of manner: means, like thus, not as a circumstance of manner.

[4] To be clear, the conjunctive relation here — whether expressed by in that way or thus — is external, not internal. This is because the relation obtains within text in its ideational guise (external), between 'chunks of experience', rather than within text in its interpersonal guise (internal), between 'chunks of interaction' (Halliday & Matthiessen 2014: 611).

That is, Martin & Rose, like Martin (1992), confuse expansion relations that are internal to the text as speech event — whether structurally between clauses or cohesively between messages — with expansion relations that function cohesively ("their global text orchestrating function").

Cherry-Picking The Data

Martin & Rose (2007: 263):

As with larger segments in written texts discussed above, formatting can be a useful starting point, but now it is the paragraphing that can help to indicate the phases in which the field unfolds. As paragraphing tends to coincide with the hierarchy of periodicity, we can adjust and expand the information that paragraphing gives us by looking at what is presented as hyperThemes and hyperNews. For example, what is presented first in each paragraph of the Inauguration Day recount are times that scaffold the activity sequence of the day’s events and of Mandela’s speech:

The day’s activity sequence is concluded in the hyperNew of the second last paragraph, with Finally..., and is then reoriented in the last paragraph, beginning with The day…

The global scaffolding resource here is sequence in time, expressed as external conjunctions and temporal circumstances.

Blogger Comments:

[1] To be clear, in SFL Theory, 'field' refers to the ideational dimension of the culture modelled as a semiotic system. Here Martin & Rose unwittingly use it to refer to the ideational semantics of a text as it unfolds in logogenesis.

[2] To be clear, as previously explained, Martin's hyperTheme and hyperNew are his rebrandings of topic sentence (of a paragraph) and paragraph summary from the field of writing pedagogy. Writing pedagogy is concerned with proposals on how to write, whereas linguistic theory is concerned with propositions that model language.

[3] To be clear, what is actually presented first in each paragraph of the Inauguration Day recount are:

10 May
The ceremonies
On that lovely autumn day
Today
We who were outlaws not so long ago
We
Never, never, and never again
Let freedom
A few moments later
The day

That is, Martin & Rose have cherry-picked the six instances that support their analysis, and ignored the four instances that do not.

[4] See, for example, Activity Sequences: A Cornucopia Of Theoretical Inconsistencies.

[5] To be clear, this is inconsistent with both the source meaning of hyperNew as paragraph summary, and with the authors' notion of hyperNew distilling what had preceded it (p195-6), since this clause realises meaning that had not previously been mentioned:

Finally a chevron of impala jets left a smoke trail of the black, red, green, blue and gold of the new South African flag.

[6] Trivially, none of the six temporal Themes are conjunctions, and two — the first and last — are participants, not circumstances.

The Notion That Text Phases Are Sensitive To Genre And Field

Martin & Rose (2007: 262):

As we introduced in Chapter 1, text phases are sensitive to both genre and field: while the stages of a genre unfold in a highly predictable sequence, the phases within each stage are partly predictable from the genre, and partly from its construal of a particular field of activities and entities. Stories, for example, unfold through phases such as settings, problems, people’s reactions, descriptions, solutions to problems, author’s comments and participants’ reflections on the significance of the events. Orators and authors use such phases in highly variable combinations, as the basic building blocks of stories, as we saw for Helena’s story. The phases of recounts are often a series of episodes, and in biographical recounts these episodes correspond to the person’s life stages. In argument genres, phases may include grounds and conclusions, as we saw for Tutu’s exposition, as well as evidence, examples and so on. In the Act we found phases such as purposes, motivations, provisions and definitions. In reports, each phase will tend to describe an element or aspect of the phenomenon under focus. These may include phases such as appearance, behaviour, location, types, parts and so on, depending on the particular type of report and its particular field. We have not attempted an exhaustive study of phase types in this book, as genres are not our primary focus (but see Martin and Rose 20076, and Rose 20076 for discussion of phases in various genres). It is an area of considerable variation that is wide open to further research; the tools we present here will help the analyst to develop this research.

Blogger Comments:

[1] To be clear, as previously demonstrated, from the perspective of SFL Theory, the authors' 'genre' — misunderstood as a stratum of context — refers to text types that vary in their ideational semantic structures (stages), while the authors' 'field' is ideational semantics, misunderstood as context, misunderstood as register. Phases, on this model, are ideational subcomponents of stages; see [2].

In claiming that text phases vary according to genre and field, what Martin & Rose are actually saying is that the ideational subcomponents of stages vary according to the ideational stages and the ideational meaning of the text type.

[2] As can be seen from this inventory of text phases, all are ideationally defined units, which locates them theoretically within ideational semantics.

[3] To be clear, as demonstrated above and previously, developing this research is not developing research that is consistent with SFL Theory, the theory in which it is purported to be located.

Problems With The Notion That Field Unfolds Through Genre

Martin & Rose (2007: 261-2):

Having established the genres we are working with, our next step in analysis is to interpret how the field unfolds through each genre. One reason we start with the field is that the steps in which it unfolds are readily accessible to conscious reflection. This can be illustrated by asking people to retell a text they have heard or read; they will rarely repeat its language features, but will typically summarise its sequence of phases. As analysts, the things we are interested in include the ideational language resources that construe the unfolding field, the interpersonal resources which evaluate it from phase to phase, and the textual resources that present each phase as a pulse of information. First identifying phases from the perspective of genre and field can provide a useful scaffold for us to identify other less obvious discourse patterns.

Blogger Comments:

[1] To be clear, in terms of SFL Theory, there are two serious misunderstandings here.  Firstly, it is a text that unfolds, not a text type (genre). This is because the unfolding of text, logogenesis, occurs at the instance pole of the cline of instantiation (text), not at a point of variation halfway up the cline (text type/genre).

Secondly, 'field' refers to the ideational dimension of context, the culture as semiotic system, not to the ideational dimension of language. It is the ideational dimension of language that unfolds through the instantiation of text, not the ideational dimension of context, since text is an instance of language, not context. In short, the authors here confuse ideational semantics with ideational culture (field). For evidence that Martin (1992) routinely misunderstands the SFL notion of field, see the clarifying critiques here.

In terms of the authors' own model, the claim is that the ideational dimension of register, field, unfolds at a level of symbolic abstraction (stratum) above register, genre. This is analogous to claiming that a metafunctional dimension of phonology unfolds through lexicogrammar. That is, Martin & Rose confuse a stratal relation (between genre and register) with logogenesis (unfolding of text).

[2] As this demonstrates, what Martin & Rose regard as 'field' is actually the ideational meaning of language itself.

[3] There are multiple confused misunderstandings here. Firstly, as a dimension of context, field is construed by all of language, not just by its ideational resources. This is because 'construal', in SFL Theory, means the assignment of an intensive identifying relation between two levels of symbolic abstraction — in this case: between the semantics and context — irrespective of metafunction. For example, the field of science is construed by propositions, interpersonal meaning, whose validity is contested by scientists.

Secondly, the interpersonal resources of language do not evaluate field, because language and field are different levels of symbolic abstraction (strata). An analogous claim would be that interpersonal phonology (lower stratum) evaluates ideational grammar (higher stratum).

[4] To be clear, the reason why Martin & Rose suggest identifying phases from the perspective of genre and field, is that, in terms of SFL Theory

• their phases are ideational semantic units,
• their genres are identified by ideational semantic units (stages),
• their field is ideational semantics misconstrued as field (their register).

Why Martin's 'Macro-Genre' Is Inconsistent With Both His Own Model And SFL Theory

 Martin & Rose (2007: 261, 262):

Technically then Chapter 115 is a macro-genre, comprising three genres — a recount extended by a report, extended in turn by another recount. Within this overall structure, the first recount projects a pledge by quoting Mandela’s inauguration speech, which is marked in the formatting by indenting. So within macro-genres (such as books and chapters), genres are interdependent — extending, elaborating or projecting each other, as illustrated in Figure 8.1.

Blogger Comment:

[1] To be clear, the authors' claim here is that a text (Chapter 115) comprises three genres (text types). This is inconsistent with both Martin's model and SFL Theory.

With regard to Martin's model, it is inconsistent in terms of strata, since it proposes that a unit of the stratum of discourse semantics (text) is composed of categories of a stratum that is two levels of symbolic abstraction above discourse semantics and outside language (genre). This is analogous to proposing that a tone group (phonology) is composed of discourse semantic categories, since discourse semantics is two levels of symbolic abstraction above phonology.

With regard to SFL Theory, the inconsistency lies in misconstruing a type of text (genre) as a constituent of a text. A text type is a point of variation on the cline of instantiation, whereas a constituent of a text is a unit on the stratum of semantics.

[2] To be clear, the general logico-semantic relations of elaboration, extension (and enhancement) and projection do not feature in Martin's logical system of discourse semantics, conjunction, because his model is his rebranding of cohesive conjunction (Halliday & Hasan 1976), in which projection does not feature as a conjunctive relation, and the general types of expansion had not yet been formulated.

The Global Purpose Of A Genre

Martin & Rose (2007: 261, 345):

Crucially all texts have more than one purpose, and for this reason they will include elements that we would expect to find in other genres. But all texts also have a global defining purpose, and it is this global purpose that predicts the stages the text will go through to achieve this goal, i.e. its genre. Its additional purposes are realised below the level of generic stages, in the variable phases of meaning within each stage, and within the messages that make up each phase. Identifying the genre of a text sometimes involves some shunting up and down, from identifying its global purpose, to analysing its stages and phases, and back up again to its purpose. As with other features of language, a first glance is often not sufficient to identify a genre. A useful guide is the table of genres, their purposes and stages, presented as an Appendix to this book.

Blogger Comments:

[1] To be clear, in SFL Theory, text types (genres) differ by their patterns of instantiation; that is by different frequencies of feature selection in semantic and grammatical systems. Systemic Functional Linguistics gives priority to system over structure.

[2] This is the opposite of what is true. As Martin & Rose have just demonstrated — see the previous post — they use their stages of a text to identify ("predict") the genre of a text, and thus its global purpose. As Appendix A (above) demonstrates, the 'purpose' of a text is just a gloss of its genre classification. In SFL Theory, the purpose of a text is modelled as a dimension of mode. Halliday (2002 [1981]: 225):

Halliday has suggested (1975) that the “textual” properties of a text – the cohesive patterns and those of ‘functional sentence perspective’ – tend to be determined by the “mode”, the function ascribed to the text in the given context of situation, the purpose it is intended to achieve.

[3] To be clear, here Martin & Rose propose a 'generic' compositional scale of

• stage
• phase, and
• message.

Now, although these are clearly semantic units of language defined ideationally, Martin & Rose incongruously locate them two levels of symbolic abstraction above semantics, on their stratum of genre, which they define as context, not language.

On terminology, in Martin (1992: 325), 'message' is proposed as the unit for his logical discourse semantics, whereas in SFL Theory, 'message' refers to the textual semantic counterpart of the clause (Halliday & Matthiessen 2014: 88, 212).

[4] To be clear, there is no shunting between levels here. On the authors' own description, in order to identify the global purpose of a text (a gloss of its genre classification), it is necessary to analyse its stages; see [2] above.

Two Reasons Why Martin's Model Of Genre Is Inconsistent With SFL Theory

Martin & Rose (2007: 260-1):

We can be confident that these are the genres we are looking at by asking a few probing questions. 

First is the global structure one of activities unfolding in time, or of phenomena described out of time? This criterion distinguishes the first and third sections, which are sequenced in time, from the second section, which describes and reflects on the struggle and its protagonists, but is not sequenced in time. 

Secondly, is the sequence of activities about specific people and events or about generic participants? This distinguishes stories from explanations and histories in the natural and social sciences. 

Thirdly is the story structured around a major disruption to the course of events or does it simply recount a series of events? This distinguishes narratives, anecdotes and exemplums (which involve a significant disruption) from recounts (whose series of events may or may not be problematic). 

And finally is it a recount of events in an episode of experience, as in the first recount, or of stages in a person’s life, as in the last? These and other generic criteria are discussed in detail in Martin and Rose (20076).

Blogger Comments:

As these generic criteria demonstrate, Martin & Rose classify genres (text types) according to their ideational semantic structure. That is, the authors classify text types in terms of just one metafunction, the ideational, and by taking the view 'from below': structural realisation.

This is seriously inconsistent with SFL Theory on two counts. Firstly, the authors ignore the other two metafunctions, the interpersonal and textual, and secondly, the perspective taken in SFL Theory is 'from above': that is, in terms of what is being expressed, not in terms of the expression.

In SFL Theory, therefore, in terms of stratification, different text types (genres) realise different combinations of contextual features of field, tenor and mode (Hasan's 'contextual configurations'), and it is these contextual differences that account for the different patterns of instantiation, of linguistic systems, that distinguish one text type from another.

See also Jim Martin Misunderstanding SFL Methodology.

Misconstruing Genres As Units Of Discourse Structure

Martin & Rose (2007: 260):

The formatting of the chapter thus gives us three main units of discourse to work with (although we need to keep in mind that formatting reflects but does not determine discourse structure). Generically speaking, Inauguration Day is a recount, a genre that chronicles an episode of experience, Cost of Courage is a report, a genre for making generalised descriptions, and Meaning of Freedom is an autobiographical recount, a genre for chronicling the significant stages of the author’s life.

Blogger Comments:

To be clear, here Martin & Rose claim that a text consists of genres (text types). In terms of SFL Theory, this confuses the semantic structure of a text ("units of discourse structure") with a point of variation on the cline of instantiation (text type).

In terms of the authors' own model, it posits categories of a contextual stratum outside language, genre, as structural units of the discourse semantic stratum within language — a stratum which is, moreover, two levels of symbolic abstraction below the genre stratum. The authors' model of stratification is given in Figure 9.2 (p309):

Martin's Register, Genre And Ideology

Martin & Rose (2007: 256):

In theoretical terms what we are saying is that register, genre and ideology all matter. …

From the angle of field, we want to analyse texts whose subject matter interests us, or at least is relevant to the topic we are studying or researching; from tenor, we are interested in how speakers in oral interactions negotiate their relationships, and in how written texts engage their readers, or position us to accept their authors’ point of view; and from mode we are interested in the interplay between spoken and written ways of meaning, and in their relation to other modalities of communication (e.g. image, sound, activity).

The concept of genre then gives us the kind of handle on discourse that the clause gives us for grammar — a genre is a recurrent configuration of meaning that matters in the culture, just as a clause is the recurrent configuration of meaning that matters for discourse. 

And from an ideological perspective there’s no point in analysing something that isn’t compelling, because analysis is a considerable investment in time and mental labour, so it has to be worth our while. 

It’s for these reasons that we have based this book on a field that fascinates us and is surely one of the key topics of our time, the overthrow of the world’s last regime of constitutional racism.

 

Blogger Comments:

[1] To be clear, register, genre and ideology are interpreted by Martin (1992) as three strata of cultural context, despite the fact that registers and genres are two perspectives on functional varieties of language: language as subpotential variation vs language as instance type variation. Martin has since jettisoned his stratum ideology from his model of context. For clarifying critiques of Martin's model, see:

• the 82 posts here on register, 
• the 67 posts here on genre, 
• the 15 posts here on ideology, and
• the 172 posts here on context.

[2] To be clear, in SFL Theory, field, tenor and mode are the metafunctional dimensions of context, not language. Martin (1992) misunderstands these as metafunctional dimensions of register, a functional variety of language, which he misunderstands as a stratal system of context. Moreover, Martin routinely confuses contextual field with ideational semantics; see, for example:

• Confusing Field With The Language That Realises It or
• Misconstruing Semantics As Context. 

[3] To be clear, this is a false analogy, since a genre is a type of discourse (text), whereas a clause is not a type of grammar, but a constituent of grammar: a rank scale unit that serves as an entry condition to grammatical systems.

[4] To be clear, in SFL Theory, a text type (genre) is a pattern of instantiation that is shared across texts. The extent to which a genre (text type) "matters" in a culture might be gauged by the frequency of its instantiation.

[5] Trivially, the use of the word 'compelling' here might be seen as a Freudian slip, given that 'compel' means to force or oblige someone to do something, and that the language that realises ideology is the language of obligation and inclination — i.e. of desired proposals.

[6] To be clear, the effect of 'basing this book' on such texts is to have the reader associate the authors with the people who are actually involved in the struggle for social justice. This, in turn, has the effect of positioning a critic of their theorising as an enemy of social justice.

[7] This will come as a surprise to all the ethnic communities who are still systematically discriminated against by their national governments.

Perception Management

Martin & Rose (2007: 254):

Meanwhile we’ll leave the last move for this chapter to Tutu, as he is quoted in the production notes to Forgiveness on DVD, the kind of gracious and memorable A2 move we have come to expect from this man:

Blogger Comments:

To be clear, this invites the reader to associate the authors with the graciousness of Desmond Tutu. Previously, the authors have invited the reader to associate one author, Martin, with Nelson Mandela (p62):

His communion with Mandela, at such a distance in so many respects …

For a reality check, see 

1. Jim Martin Falsely Accusing The Late Ruqaiya Hasan Of Plagiarism
2. David Rose Dishonestly Vilifying His Friend At The Time Of His Mother's Tragic Death

Problems With The Authors' Negotiation System Network

Martin & Rose (2007: 252-3):

Above SPEECH FUNCTION, in the discourse semantics, we have the system of NEGOTIATION, which sequences moves. The basic system allows for exchanges consisting of between one and five moves, as outlined in Figure 7.6. 

In addition there are tracking and challenging options which have not been included in the network. Either can increase the number of moves an exchange works through before establishing its obligatory K1 or A1 move; and in many cases challenges abort an exchange completely by refusing to comply and perhaps leading the negotiation off in another direction (by initiating a new exchange).

The different roles of dKl, K2, K1, K2f, K1f, dA1, A2, A1, A2f, A1f and tracking or challenging moves can be shown in analysis by modelling the former as constituency to the left of the move labels, and the latter as dependency to the right:

 

 

Blogger Comments:

[1] To be clear, for the system of NEGOTIATION to be above the system of SPEECH FUNCTION on Martin's stratum of discourse semantics, there needs to be a scale on which these two systems can be ranked as higher or lower. Martin & Rose have not proposed any such rank scale here.

[2] To be clear, the network does not specify the sequence of moves — it merely specifies which moves can be selected; but see [3] below.

[3] This is misleading, because it is not true. The network specifies either:

• an anticipatory primary initiation OR
• a nuclear primary initiation OR
• a secondary initiation

and, in the case of action, either immediate or prospective compliance. That is, these two systems  of the network specify two — and only 2 — moves.

The third system is incoherent, because it misconstrues the presence or absence of a primary follow up move as more delicate options of secondary follow up move. That is, it confuses paradigmatic delicacy with syntagmatic sequence.

See also the clarifying critique of the first appearance of this network, Figure 7.2, here.

[4] To be clear, the failure to include these tracking and challenging options in the network is a serious shortcoming of the NEGOTIATION system, since, in SFL Theory, it is the system that specifies structures.

Misrepresenting Halliday's Semantic System Of Speech Function As Martin's Discourse Semantic System

Martin & Rose (2007: 251, 252):

As far as the realisation of SPEECH FUNCTION in MOOD is concerned, we noted the important role played by indirect speech acts (Halliday’s mood metaphors) as far as expanding the meaning potential available for speakers to negotiate within dialogue. The major MOOD options for the English clause are outlined in Figure 7.5. Technically speaking, speech function is a discourse semantic system realised through the grammar of mood (including vocation, tagging, modality and polarity which are not included in Figure 7.5).

 

Blogger Comments:

[1] Cf Halliday & Matthiessen (2014: 162):

[2] This is misleading, because it misrepresents Halliday's semantic system of speech function as Martin's discourse semantic system. As previously noted, the authors have not acknowledged Halliday as the intellectual source of the system of speech function, anywhere in this chapter.

The Authors' System Of Speech Function

Martin & Rose (2007: 251, 252):

In this chapter we’ve developed two systems for analysing dialogue. The first, SPEECH FUNCTION, was designed to explore the relationship between moves and their realisation in grammar (technically speaking their MOOD). The relevant network of choices is consolidated in Figure 7.4, and allows for the 13 basic speech acts presented above. Further delicacy is of course possible; we could for example distinguish questions asking for missing content from those exploring the modality and polarity of a given clause (i.e. Who betrayed Daniel? vs Did one of his friends betray Daniel?). This kind of specificity is much further developed in Eggins and Slade (1997) and in the work of Hasan and her colleagues (e.g. Hasan 1996).

Blogger Comments:

[1] This is misleading, because, once again, the authors' here misrepresent themselves as the intellectual sources of Halliday's system of SPEECH FUNCTION. See, for example, Halliday (1985: 68-71).

[2] To be clear, the network in Figure 7.4

• has no entry condition, 
• omits the minor speech function 'alarm' ('warning' vs 'appeal')
• ungroups minor speech functions ('express self' vs 'attending'), and
• groups minor speech functions with major speech functions ('attending' with 'negotiating').

Cf Halliday & Matthiessen (2014: 136):

Confusing Reading Pedagogy With Linguistic Theory

Martin & Rose (2007: 249):

In pedagogic terms what we are looking at here is scaffolding which has been carefully designed to ensure teachers support students as they move into academic discourse of this order. From the point of view of exchange structure what we are looking at is exchange complexing. Each exchange is designed to realise the goals of its Prepare, Focus, Identify, Highlight or Extend phase; and they are sequenced in relation to one another according to the Detailed Reading Cycle just outlined (see Figure 7.3).

Not all registers of dialogue have exchange complexing of this kind. But as a rule of thumb we can expect that the more institutionalised the discourse, the more likely it is to display exchange routines of this order. In some contexts these routines may be so conventionalised that we treat them as stages of a genre — as with Ventolas work on service encounters for example (or Sinclair and Coulthard’s 1975 work on classroom discourse for that matter).

Blogger Comments:

To be clear, the only registers of language "likely to display exchange routines of this order" are those that realise educational contexts in which this pedagogy is deployed.

A Problem With The Argument For A 'Move Complex'

Martin & Rose (2007: 244-5):

To begin, David prepares the class to read the first sentence by telling them what it means, paraphrasing it in terms they can all understand, and then reads it to them. He then prepares the students to identify one element of the sentence, its initial temporal circumstance, by giving them its transitivity category ‘when’ and telling them exactly where to find it (‘that sentence starts by telling us… '). The logical relation between the sentence and the two preparations is elaboration (Halliday and Matthiessen 2004). Following Ventola (1987) we can treat this semantic triplet as a move complex, filling a single slot in exchange structure — two statements in other words functioning together as a K1 move. We'll highlight these move complexes by putting '=’ as a superscript before elaborating moves, and using dependency lines on the left, to group them with one another:

Blogger Comments:

[1] Trivially, 'when' is not a transitivity category; 'temporal circumstance' is its transitivity function.

[2] To be clear, the logico-semantic relation of elaboration does not feature in Martin's logical semantics, because his system of CONJUNCTION is a rebranding of Halliday & Hasan's cohesive conjunction (textual grammar) as logical discourse semantics, and Halliday & Hasan (1976) did not yet use the general expansion systems of elaboration, extension and enhancement.

Moreover, the logical relation between "the sentence and its two preparations" is not elaboration. In the first move by the teacher, he projects a clause complex that is a more congruent rendering of the metaphorical clause in his second move. It is the projected more congruent clause complex that can be said to be an elaboration of the (unprojected) metaphorical clause.

In the third move by the teacher, he projects a clause (when they were rebelling) that encodes the function of the temporal circumstance of the second move (In the mid-1980s). Again, the relation between the two is a type of elaboration (intensive identity), but there is no elaborating relation between the two sentences.

These distinction are important because the notion of a move complex depends on there being a logico-semantic relation between the sentences as moves, which is not the case here.

Problems With The Exchange Structure NEGOTIATION System

Martin & Rose (2007: 240):

Expressed as a network of choices, we have a resource with three intersecting systems. One system is concerned with how the exchange is initiated — by the primary actor/knower or the secondary one, and if by the primary actor/knower whether the nuclear Al/Kl move is anticipated or directly enacted. Another system distinguishes between action and knowledge exchanges, and for action exchanges allows for negotiations in which goods can be proffered or services enacted immediately (in which case verbalising the A1 move is optional, and in a sense redundant) and negotiations in which some time will pass before the goods are proffered or the service enacted (in which case verbalising the A1 move as a promise is obligatory, and actually acting to fulfil the promise may not eventuate). Finally there is a system allowing for follow-up moves, first for the secondary actor/knower, and then, if they do make a move, for the primary actor/knower. These options are set out in Figure 7.2.

Blogger Comments:

To be clear, this system is inconsistent with the preceding exposition of exchange structure. For example, it presents the obligatory nuclear move (A1/K1) only as a potential option — though not even an option if the exchange is initiated by the secondary knower/actor. 

Moreover, the system does not allow exchanges that include both an optional anticipatory move and an obligatory nuclear move, since only one option can be chosen.

These shortcomings are partially masked by the inclusion of structure types, after the systemic features, none of which are valid realisations of the system selections.

Even more problematically, the network unintentionally (and incongruously) allows exchange structures of just a single move, which is optional, not obligatory, such as:

• dK1 [anticipate, knowledge, no follow up], and
• K2 [secondary, knowledge, no follow up].

The Problem With The Obligatory/Optional Move Distinction

Martin & Rose (2007: 239-40):

To complete the picture we can now allow for the possibility of follow-up moves by the secondary actor or knower (with 'f' standing for ‘follow up’):

And if they do follow up, then there is the possibility of a further follow-up move by the primary actor or knower:

We can sum up the various possibilities reviewed here using parentheses for optional moves. The structure potential for action exchanges is thus:

((dA1) ^ A2) ^ A1 ^ (A2f ^ (A2f))

And for information exchanges, we find the same possibilities:

((dK1) ^ K2) ^ K1 ^ (K2f ^ (K2f))

Blogger Comments:

[1] To be clear, on this model, the only obligatory move in the exchange between the waitress and Hendrik is said to be her A1 move Yes. All the other moves, including her offer Wine? and his metaphorical command Could I have a bottle of your best dry red? — which elicits her "obligatory" move — are mere optional extras.

Similarly, the only obligatory move in the exchange between Sannie and Llewelyn is said to be her K1 move Coetzee. All other moves, including her metaphorical command You'll never guess who's here and his question Who? — which elicits her "obligatory" move — are mere optional extras.

[2] Logically, these should read (dA1 ^) and (dK1 ^), not (dA1) ^ and (dK1) ^.

[3] Logically, these should read (^ A1f) and (^ K1f), not ^ (A2f) and ^ (K2f).

Problems With Analyses Using Anticipatory Moves (dA1 & dK1)

Martin & Rose (2007: 238):

A third possibility is for exchanges to be initiated by primary actors and knowers who anticipate proffering goods or performing a service by offering first to do so, or anticipate professing information by first alerting their addressee that it is coming. These anticipatory moves in a sense delay the exchange of goods-and-services and information, and so are referred to by Berry as dA1 and dK1 moves (with ‘d’ standing for ‘delay’):

These dK1^K2^K1 sequences can be used in conversation to re-affirm a proposition that needs to be foregrounded, for example as part of an argument amongst Daniel’s comrades about who betrayed him to the authorities.

Blogger Comments:

[1] To be clear, in terms of SFL Theory, the exchange between the waitress and Hendrik involves two speech functions, offer and command, with the response to the offer ellipsed. A more congruent rendering of the exchange would be:

• Can I get you some wine? (Offer)
• Yes. (response: acceptance)
• Please bring a bottle of your best dry red. (Command)
• I will. (response: undertaking)

Similarly, the exchange between Sannie and Llewelyn also involves two speech functions, command and question, with the response to the command ellipsed. A more congruent rendering of the exchange would be:

• Guess who's here. (Command)
• Okay. (response: undertaking)
• Who's here? (Question)
• Coetzee is here. (response: answer)

Importantly, contrary to the authors' analysis, with the metaphor unpacked, the first move involves goods-&-services ('A'), not information ('K').

See Halliday & Matthiessen (2014: 137).

[2] To be clear, the exchange between Luke and Zako involves a rhetorical question: a question asked in order to create a dramatic effect or to make a point rather than to get an answer.

Problems With Obligatory Moves And With Primary vs Secondary Actor/Knowers

Martin & Rose (2007: 237-8):

We can interpret what is going on here as follows, drawing on work by Ventola (1987), who was in turn building on work by Berry (e.g. 1981). Minimally speaking, exchanges consist of one obligatory move. When negotiating goods-and-services, this is the move that proffers the goods or performs the service; when negotiating information, this is the move that authoritatively establishes the facts of the matter.

Berry refers to goods-and-services negotiations as action exchanges, and information exchanges as knowledge ones. And she refers to the person responsible for proffering goods or performing a service as the primary actor, and the person who has the authority to adjudicate information as the primary knower. On this basis, the waitress’s move below is nuclear A1 move, and Sannie’s is Kl:

Berry refers to the dialogue partner for primary actors as a secondary actor, who is the person who receives the goods or has the service performed for them; the secondary knower is the person who receives the information professed by the primary knower. Where exchanges are initiated by the secondary actor (requesting goods-and-services) or the secondary knower (requesting information), we find canonical two-part exchanges like the following:

Blogger Comments:

[1] To be clear, in the exchange of goods-&-services, the move that "proffers the goods" or "performs the service" is not the obligatory move, since these can be omitted in an exchange. These moves are responses to commands or offers, and responses are not obligatory. If there is an obligatory move in an exchange, it is the move that brings it into being, the initiating move.

[2] To be clear, this confuses semantics with context. Relations between interlocutors — primary vs secondary actor or knower — are a matter of tenor, the interpersonal dimension of context, the culture as a semiotic system. The moves they enact in an exchange are units in the interpersonal dimension of semantics, the language as a semiotic system.

Misrepresenting Halliday's Approach To 'Calls' As The Authors' Insight

Martin & Rose (2007: 235): 

As a rule of thumb we can include expletives and vocatives in other moves where possible, and treat them as independent moves only when there is nothing to append them to. So Sannie would be a vocative in the following command by her father:

Hendrik: Sannie, go with Father Dalton.

Sannie: - (goes)

But in the movie version of this exchange, Coetzee has been attacked and injured by Ernest, and in the ensuing chaos Hendrik has to first get Sannie’s attention before negotiating his demand for service:

Hendrik: Sannie.

Sannie: - What?

Hendrik: Go with Father Dalton.

Sannie: - (goes)

Blogger Comments:

This is misleading. Yet again, Martin & Rose misrepresent Halliday's theorising as their own. Halliday (1994: 95):

Misunderstanding Metaphors Of Modality

Martin & Rose (2007: 235):

One variation on this principle to watch out for involves certain mental process clauses in first or second person present tense (I think..., I suppose.. ., do you reckon..., don't you suppose, etc.). These are actually modalities involving grammatical metaphor (Halliday and Matthiessen 2004).

Blogger Comments:

This misunderstands metaphors of modality. As this concerns explicit subjective modality, it is only first person variants that function this way. Halliday & Matthiessen (2014: 687) explain, beginning with the example I don't believe that pudding ever will be cooked:

Here the cognitive mental clause I don’t believe is a metaphorical realisation of probability: the probability is realised by a mental clause as if it was a figure of sensing. Being metaphorical, the clause serves not only as the projecting part of a clause nexus of projection, but also as a mood Adjunct, just as probably does. The reason for regarding this as a metaphorical variant is that the proposition is not, in fact, ‘I think’; the proposition is ‘it is so’. This is shown clearly by the tag; if we tag the clause I think it’s going to rain we get 

I think it’s going to rain, isn’t it?

not I think it’s going to rain, don’t I?. In other words the clause is a variant of it’s probably going to rain (isn’t it?) and not a first-person equivalent of John thinks it’s going to rain, which does represent the proposition ‘John thinks’ (tag doesn’t he?). Thus in

You know || what’s happening tomorrow at five o’clock, don’t you?

the ‘mental’ clause you know was able to be tagged because it does not stand for a modality. It is the fact that mental clause is a modal clause and serves as mood Adjunct that explains the tag. If it was just an ordinary mental clause in a clause nexus of projection, I don’t believe should be able to be tagged. But since it has a metaphorical status and serves as mood Adjunct, it cannot be tagged. Instead, the Moodtag picks up the Mood element of the modalised proposition: that pudding probably never will ... will it?

Confusing Modal Responsibility With Moral Responsibility

Martin & Rose (2007: 234-5):

Technically speaking, what we are saying is that a move is a ranking clause, including any clauses embedded in it, and in addition any clauses dependent on it. So we can tag the main clauses in the examples below; they are negotiable. Coetzee makes others responsible for the torture facility, not his role in it:

And Sannie makes the village gossipmongers responsible, not her family:

The underlined clauses are not directly negotiable; to make them so would require an additional initiating move in which they are promoted from a subordinate to an arguable position.

Blogger Comments:

[1] This is misleading. Once again, Martin & Rose are misrepresenting Halliday's model of speech function as their own theoretical creation.

[2] To be clear, this confuses moral responsibility (for the torture facility) with modal responsibility (whether or not it is valid to state they had a facility outside Capetown…). 

[3] To be clear, the modal responsibility here is for the validity of the claim they'll say we're selling the house.

[4] To be clear, the projected clause we're selling the house can be challenged in a response move, as exemplified by But we're not (selling the house). Moreover, not only is this possible, it is more likely than a challenge to the projecting clause: But they won't say that.

In short, on the authors' model, the move But we're not (selling the house) would not be analysed as a responding move to they'll say we're selling the house, but would instead be analysed as a new initiating move.

The Authors' Definition Of A Response

Martin & Rose (2007: 232-3):

We can thus define a response as a move which:

(1) takes as given the experiential content of its initiating pair part

and

(2) accepts the general terms of its argument established by its Subject-Finite structure (i.e. its polarity/modality/temporality). …

But a response does not allow for changes to the nub of the argument (its Subject),  or to the content of what is being argued about in the rest of the clause. By definition, any move making changes of this kind would not be considered a response but a new initiating move.

Blogger Comments:

[1] To be clear, speech function is concerned with interpersonal meaning, not with experiential meaning. In the case of propositions, what is accepted or rejected in a response is the validity of what is predicated of the Subject (Halliday & Matthiessen 2014: 148, 151-2).

[2] To be clear, a responding move may accept or reject the validity of a proposition or proposal. 

[3] Halliday & Matthiessen (2014: 148):

From above, [the Subject] is that which carries the modal responsibility: that is, responsibility for the validity of what is being predicated (stated, questioned, commanded or offered) in the clause. … The notion of validity relates to the arguing of the case, if it is a proposition, or to the putting into effect if it is a proposal. The Subject is that element in which the particular kind of validity (according to the mood) is being invested.

Misrepresenting Halliday's Speech Function As The Authors' Creation

Martin & Rose (2007: 227):

In the previous section we proposed an inventory of 13 basic speech acts; we now need to think about how to distinguish them from one another when analysing conversation. One useful set of markers includes please, kindly, ta, thanks, thankyou, OK, alright, no worries, you're welcome, not a problem, which normally indicate moves concerning goods-and-services. Here Ernest pleads with Coetzee to stay and talk to his parents (until the comrades can get there to kill him):

And Hendrick orders wine for dinner with Coetzee and his family:

Where such markers are not present, we can check to see if they could have been (adding for example please to Hendrik’s command and alright to verbalise Sannie’s response to her father below):

Blogger Comments:

[1] This is misleading. Once again Martin & Rose are misrepresenting Halliday's major and minor speech functions (minus 'alarms') as their own creation.

[2] To be clear, the SFL way to identify speech functions in a text is to take the view 'from above' (the meaning being realised). This means using the semantic system of which they are instances. For example, in the case of major speech functions (realised by major clauses):

• if information is given, then it is a statement;
• if information is demanded, then it is a question;
• if goods-&-services are given, then it is an offer;
• if goods-&-services are demanded, then it is a command.

[3] To be clear, contrary to SFL Theory, here Martin & Rose advocate taking the view 'from below' (how meaning is realised) instead of the view 'from above' (the meaning being realised).

Misunderstanding Halliday's Moodtags And Misrepresenting Them As The Authors' Work

Martin & Rose (2007: 226):

Moves may also end with tag 'questions' (don’t you?, isn't he?, etc.) oriented to the addressee. These tags do not function as separate moves, but are better treated as explicit invitations to a listener to respond. Both of Sannie’s speech acts below would thus be treated as tagged statements (not as statements followed by questions). There is after all only one proposition being negotiated, not two:

 

 Blogger Comments:

[1] This is misleading. Yet again, Martin & Rose here present Halliday's work as if it were their own. Importantly, the tag doesn't just invite a response; it signals the type of response expected. Halliday (1985:69):

[2] This is misleading. While the first instance is a tagged declarative clause realising a statement, the second instance does not feature a Moodtag, and realises two speech functions, not one. To explain:

Firstly, a Moodtag reprises the Finite and Subject of a clause, so the (unmarked) Moodtag for you could have killed the man is couldn't you? not you know.

Secondly, there are two propositions

• the projected statement: you could have killed the man, and
• the projecting question: you know (realised metaphorically as a declarative clause),

each of which could be challenged in response.

Moreover, the projecting clause you know can itself be tagged by don't you?, further demonstrating that it is not a Moodtag but a proposition in its own right.