The Discourse Semantic System Of External Conjunction

Martin & Rose (2007: 122):

External conjunction is concerned with logically organising a field as sequences of activities. For each general type of external conjunction - addition, comparison, time, consequence - there are two or more sub-types, summarised in Table 4.2.

 

Blogger Comments:

[1] To be clear, in SFL Theory, external conjunction is concerned with relating text segments in their experiential guise.  Halliday & Matthiessen (2014: 611):

As we have seen, elaborating, extending and enhancing conjunctions mark relations between semantic domains, i.e. between text segments. These text segments are simultaneously ideational and interpersonal; they construe experience as meaning, e.g. an episode in a narrative or a recount, and they enact roles and relations, e.g. an exchange in a conversation or consultation, or an argument in an exposition. Relations link text segments either in their ideational guise or in their interpersonal guise: they relate either chunks of experience or chunks of interaction. … Relations between representations of segments of experience are called external relations, and conjunctions marking such relations are called external conjunctions. … Relations linking text segments in their interpersonal guise are called internal relations – internal to the text as a speech event, and conjunctions marking such relations are called internal conjunctions.

[2] To be clear, in SFL Theory, external conjunction is a non-structural grammatical system of the textual metafunction that deploys the resource of expansion to create cohesive relations between portions of text.  Martin & Rose combine the textual system of conjunction and the logical system of clause complexing and rebrand it as discourse semantics instead of grammar, without providing evidence as to why it constitutes a higher level of symbolic abstraction.

[3] To be clear, in SFL Theory, field is the ideational dimension of context, the culture construed as a semiotic system.  Since sequences of activities are here construed as discourse semantic rather than context (contrā Martin 1992), field and activity sequences lie on different levels of symbolic abstraction and, as such, different levels of semiotic organisation.

[4] To be clear, the omissions and misunderstandings in the authors' model of external conjunction can be made evident by comparing it with the SFL model below (Halliday & Matthiessen 2014: 612):

For example, Martin & Rose:

• do not organise their system in terms of the three most general subtypes of expansion: elaboration, extension and enhancement, which are fractal types manifested across multiple domains (e.g. circumstances, relational processes, etc.);
• omit all 9 subtypes of elaborating relations;
• omit 4 of the 6 subtypes of extension;
• omit the enhancing relation of matter;
• misconstrue the adversative extending relation ('but') as a subtype of comparison (enhancing);
• misconstrue the manner subtype means as a relation of consequence (cause-condition).

Moreover, the logico-semantic relation of projection is entirely absent from the discourse semantic model of logical relations.

Misconstruing Interpersonal And Textual Adjuncts As "Logical" Continuatives

Martin & Rose (2007: 121-2):

Finally there is one other type of linker aside from conjunctions. These are known as continuity items or continuatives. Continuatives differ from conjunctions in two ways. More often than not, conjunctions occur at the beginning of a clause in English (although cohesive conjunctions can be positioned more flexibly). But continuatives primarily occur within a clause, rather than at the start. And their options for logical relations are far more restricted. Two that we have come across so far are even and also:

We even spoke about marriage.
It is also not true that the granting of amnesty encourages impunity.

To put even at the start of this clause completely changes its meaning — rather than spoke about marriage being unexpected, it is we that becomes unexpected. Also, placing also at the start of a clause is a marked option, as it more typically occurs within the clause.

Blogger Comments:

[1] To be clear, in SFL theory, the grammatical items linker and continuative are subclasses of conjunction, which is a subclass of adverbial. Halliday & Matthiessen (2014: 75):

[2] To be clear, in SFL theory, it is conjunctive Adjuncts that have this positional flexibility, and these are realised by adverbials, including adverbs and conjunctions.  Their function is textual on the lexicogrammatical stratum, not logical  on the discourse semantic stratum.

[3] To be clear, in SFL theory, continuatives are inherently thematic and so occur at the start of the clause.  Halliday & Matthiessen (2014: 107):

A continuative is one of a small set of words that signal a move in the discourse: a response, in dialogue, or a new move to the next point if the same speaker is continuing. The usual continuatives are yes no well oh now.

[4] To be clear, in SFL theory, as demonstrated by the usual examples yes no well oh now, continuatives do not express logical relations.  Their function is textual on the lexicogrammatical stratum, not logical  on the discourse semantic stratum.

[5] To be clear, neither even nor also are continuatives.  In the first cited clause, even is an adverb serving as a mood Adjunct of intensity. Its function, therefore, is interpersonal on the lexicogrammatical stratum, not logical  on the discourse semantic stratum.

In the second cited example, also is a conjunction serving as a conjunctive Adjunct of addition. Its function, therefore, is textual on the lexicogrammatical stratum, not logical on the discourse semantic stratum.

[6] To be clear, the locational choice of a conjunctive Adjunct in a clause serves a textual function on the lexicogrammatical stratum, not  a logical function on the discourse semantic stratum.

Misconstruing Textual Cohesion As A Type Of Logical Dependency

Martin & Rose (2007: 121):

Third, two sentences can be logically related by a conjunction such as Further or Thus:

It is also not true that the granting of amnesty encourages impunity...
Further, retributive justice…is not the only form of justice. 

This is a far more personal approach, which sees the offence as something that has happened to people and whose consequence is a rupture in relationships.
Thus we would claim that justice, restorative justice, is being served when efforts are being made to work for healing, for forgiveness and for reconciliation.

We will refer to these kinds of dependency relations between sentences as cohesive (following Halliday and Hasan 1976).

Blogger Comments:

[1] To be clear, in SFL theory, cohesive conjunction is a non-structural grammatical system of the textual metafunction.

[2] To be clear, here Martin & Rose confuse interdependency, which is structural and logical, with cohesion, which is non-structural and textual.

[3] This is misleading. To be clear, Martin & Rose are not "following" Halliday & Hasan (1976); they are taking the grammatical model of cohesive conjunction in Halliday & Hasan (1976) and simply rebranding (their misunderstanding of) it as their own model of discourse semantics (following Martin 1992).

Rebranding Grammar As Discourse Semantics And Confusing The Logical And Textual Metafunctions

Martin & Rose (2007: 121-2):

Before discussing each type of conjunction in more detail, we need to look briefly at three grammatical contexts in which they are realised, as different conjunctions are used in each context. The first type links a sequence of independent clauses:

I went off to school in the morning
and I was sitting in the classroom
and there was only one room where all the children were assembled
and there was a knock at the door

… another conjunction used in paratactic relations is then:

I was told to shut up, sit in a chair
then I was questioned

These two clauses cannot be reversed without reversing the logical relation between them. We cannot say, for example, *then I was questioned, I was told to shut up. But the conjunction when does allow such a reversal:

when I answered the questions
I was told that I was lying 

I was told that I was lying
when I answered the questions

The reason is that these two clauses are not equal in status. One is independent, and the other beginning with when is dependent on it. The when clause functions as the context in which the other takes place. In this respect its function is similar to a Circumstance of time such as after the questions I was told that I was lying, which can come at the start or end the clause.

Blogger Comments:

[1] To be clear, the system being demonstrated here is Halliday's grammatical system of clause complexing, and the grammatical context is which it is realised is the syntagmatic axis.  The realisation relation does not obtain stratally between discourse semantics and lexicogrammar. Martin & Rose, following Martin (1992), are merely rebranding Halliday's (1985) grammatical system as Martin's discourse semantic system.

[2] To be clear, conjunctions are lexicogrammatical phenomena.

[3] To be clear, it is not the logical relation that is reversed, but the sequence of clauses.  Clauses are lexicogrammatical phenomena.

[4] To be clear, this relates to the textual function of β-clauses in clause complexes, which varies according sequence: progressive or regressive. Halliday & Matthiessen (2014: 551):

The contrast between progressive and regressive sequence … is quite typical of procedural texts: temporal clauses delimiting the performance of actions tend to be rhematic …. In general, thematic β-clauses serve to set up a local context in the discourse for the α-clause: they re-orient the development (as in the staging of a narrative), often distilling some aspect of what has gone before to provide the point of departure for the dominant clause, thus creating a link to the previous discourse (cf. Longacre, 1985; Thompson, 1984; Ford & Thompson, 1986).

[5] To be clear, the 'similarity' relates to their being agnate manifestations of the expansion category temporal enhancement. Halliday & Matthiessen (2014: 666-7):

We have met expansion in a number of different grammatical domains. The most detailed account was … where we found that the three subtypes of expansion (elaboration, extension and enhancement) combine with tactic relations to link one clause to another in the formation of clause complexes. … On the one hand, expansion is manifested in the augmentation of the clause by circumstances: these circumstantial augmentations cover all three types of expansion, with enhancement being the most highly developed one.

[6] To be clear, this relates to the textual function of circumstances in clauses as thematic or rhematic; see [4].

Confusing Different Strata, Different Metafunctions, And Different Grammatical Elements

Martin & Rose (2007: 119-20):

Like Tutu, Helena uses explicit conjunctions to signal the beginning of new phases in her story. But whereas Tutu uses them to organise his argument, Helena uses them to sequence the phases in time.

 

Helena uses the time conjunctions Then and After to connect each phase to the immediately preceding events, but the scope of finally is the story as a whole. During all the preceding events Helena didn’t understand the struggle, but now she finally does. 

The other resources Helena uses here to sequence the story in time are Circumstances - As an eighteen-year-old, one day, More than a year ago, After my unsuccessful marriage, After about three years with the special forces, Today. These Circumstances set the events in an exact time period, while time conjunctions simply indicate the sequence.

Blogger Comments:

[1] To be clear, this confuses different levels of description.  Although this is presented as discourse semantics, conjunctions are grammatical, the stratum below discourse semantics, and phases of story are, on the authors' model, genre, the stratum two levels above discourse semantics.

[2] To be clear, the two instances of after in the extract are not conjunctions and do not function conjunctively, either logically within a clause complex, or textually as a cohesive relation.  Both are prepositions and function, experientially, as the minor Process — and interpersonally as the minor Predicator — of a prepositional phrase serving as a circumstance of temporal Location.

[3] To be clear, this is (high school) reading comprehension, not (academic) linguistic analysis.

[4] To be clear, circumstances do not function logically or sequence a story in time.  Circumstances function experientially within a clause.  However, each of circumstances cited is also highlighted in the text as a marked Theme, and it is this textual logogenetic pattern of Theme selection (Halliday & Matthiessen 2014: 659ff) that works with the textual grammatical system of conjunctive cohesion — misunderstood by Martin & Rose as logical discourse semantics — in creating text.

The Four Types Of Conjunction And Expectancy

Martin & Rose (2007: 119):

In sum the explicit conjunctions here realise our four types of conjunction: addition, comparison, time and consequence, and Helena uses them deftly to manage expectancy in the context of the events. They are set out in Table 4.1.

* Note that then is not typically counterexpectant, but functions counterexpectantly in this context.

Blogger Comments:

[1] As previously demonstrated, the authors' argument that a speaker's (textual or logical) deployment of expansion features manages the experiential expectations of readers does not survive close scrutiny.

[2] To be clear, in SFL theory, the first three examples in Table 4.1 are instances of cohesive conjunction, and function textually and non-structurally, whereas the final two examples are instances of clause complexing, and function logically and structurally.

[a] Of the three instances of cohesive conjunction:

• and all my girlfriends envied me is an instance of extension: positive addition, and
• then one day he said… is an instance of enhancement: temporal: following.

Whereas the first and third of these are merely rebrandings, the second is a misunderstanding of the type of cohesion deployed. In So was he, so is an instance of substitution, not conjunction. See Halliday & Matthiessen (2014: 636).

[b] Of the two instances of clause complexing:

• all because I married… is an instance of enhancement: causal-conditional: cause: reason ('because P, so result Q'), and
• even if he was an Englishman is an instance of enhancement: causal-conditional: condition: concessive ('if P then contrary to expectation Q').

That is, Martin & Rose rebrand cause-condition as consequence, and the distinction between (a type of) cause and (a type of) condition as a distinction between expectant and counterexpectant.

∞

The theoretical disadvantage of rebranding grammatical expansion relations as different discourse semantic relations is that it creates a mismatch between strata in the absence of grammatical metaphor, thereby undermining the distinction between congruent and metaphorical grammatical realisations of semantic systems.

The Interplay Of Explicit And Implicit Conjunction To Manage Expectancy [3]

Martin & Rose (2007: 118-9):

This interplay of explicit and implicit conjunction to manage expectancy is well illustrated in the first Incident of Helena’s story:

… Then one day he said he was going on a 'trip'.
'We won't see each other again…maybe never ever again.'
I was torn to pieces.
So was he.
An extremely short marriage to someone else failed
all because I married to forget.

… Then the next step from romance to tragedy is explicitly marked by Then, signalling that a new phase is beginning which is likely to be counterexpectant, and so probably bad news. After her reaction, So was he makes explicit that her lover’s feelings about leaving were the same as hers, and that this was to be expected. And the failure of her subsequent marriage was also completely predictable, made explicit by the causal conjunction all because.

Blogger Comments:

[1] To be clear, the conjunction then merely signals the conjunctive relation of 'temporal: following'. Whether what follows constitutes a new "phase" depends on the experiential meanings of the preceding and following portions of text.

[2] To be clear, there are no grounds for counterexpectancy on the part of a reader merely on the basis of the conjunction then.  Consider instances of the type: I sat down; then I stood up.

[3] To be clear, there are no grounds for assuming that "counterexpectant" information is bad news.  Consider instances of the type: Then, on the day I was retrenched, I won the lottery.

[4] To be clear, here the conjunction so merely signals the conjunctive relation of 'positive addition'.  The clause so was he gives the reader no evidence as to whether this was to be expected. In fact, its inclusion by the speaker might be taken to resolve any doubt on the matter.

[5] To be clear, the conjunction group all because provides no grounds for a reader to judge the preceding clause as predictable; it merely signals the clause complexing relation of 'cause: reason'.  Consider instances of the type: So I ended up in hospital all because I walked to work instead of driving.

∞

In all of the above, Martin & Rose are misattributing their own hindsight judgements of reader expectancy to the unfolding text of the speaker.  Moreover, in doing so, they

• confuse the experiential meaning of clauses with the logico-semantic relation of expansion, 
• confuse the latter's textual manifestation (cohesive conjunction) with its logical manifestation (clause complexing), and
• rebrand these grammatical systems as discourse semantic systems.

The Interplay Of Explicit And Implicit Conjunction To Manage Expectancy [2]

Martin & Rose (2007: 118-9):

This interplay of explicit and implicit conjunction to manage expectancy is well illustrated in the first Incident of Helena’s story:

… A bubbly, vivacious man who beamed out wild energy.
Sharply intelligent.
Even if he was an Englishman,
he was popular with all the 'Boer' Afrikaners.
And all my girlfriends envied me.

… And in the description phase that follows, she uses even if in a similar way [to even], to tell us that an Englishman being liked by the ‘Boer’ Afrikaners is counterexpectant (if they were expected to like him she might have said because he was an Englishman). In contrast, her girlfriends’ reaction is explicitly added by starting a sentence with And, letting us know that their envy is entirely to be expected.

Blogger Comments:

[1] To be clear, the claim here is that the conjunction even if functions similarly to the adverb even (see previous post).  In SFL theory, the functional difference is significant.  The function of conjunction even if is to relate two clauses in a clause nexus by the enhancement relation of concessive condition, whose meaning is 'if P then contrary to expectation Q' (Halliday & Matthiessen 2014: 478).  So, in the above extract, the meaning is 'if he was an Englishman, then contrary to expectation, he was popular with all the 'Boer' Afrikaners'.  In terms of metafunction, the meaning being realised is logical, and the grammatical domain is the clause complex.

In contrast, the function of the adverb even is that of a mood Adjunct of intensity: counterexpectancy: exceeding, whose meaning is 'went so far as' (Halliday & Matthiessen 2014: 188).  So the meaning of we even spoke about marriage is 'we went so far as to speak about marriage'.  In terms of metafunction, the meaning being realised is interpersonal, not logical, and the grammatical domain is the clause, not the clause complex.

These important distinctions are lost on Martin & Rose, in their rebranding of Halliday's lexicogrammar as Martin's discourse semantics.

[2] To be clear, the explicit use of the conjunction and does not "let us know that their envy is entirely to be expected".  This can be demonstrated by omitting the conjunction, leaving the textually cohesive relation implicit:

Even if he was an Englishman,
he was popular with all the 'Boer' Afrikaners.
All my girlfriends envied me.

Here Martin & Rose are merely making bare assertions, unsupported by any linguistic evidence or argumentation. 

The Interplay Of Explicit And Implicit Conjunction To Manage Expectancy [1]

Martin & Rose (2007: 118-9):

This interplay of explicit and implicit conjunction to manage expectancy is well illustrated in the first Incident of Helena’s story:

As an eighteen-year-old, I met a young man in his twenties.
He was working in a top security structure,
it was the beginning of a beautiful relationship.
We even spoke about marriage. …

The first phase is sequenced in time, from meeting to relationship to speaking about marriage, but this sequence is expected by the field, as we discussed in Chapter 3 (section 3.5), so there is no need to make each step explicit with conjunctions. On the other hand, Helena uses even to make it explicit that speaking about marriage was more than we would normally expect at the beginning of a relationship.

Blogger Comments:

[1] To be clear, here Martin & Rose confuse conjunctive relations in the text of the speaker with the expectations of various types of others.

[2] To be clear, the extract is not sequenced in time. Only the last clause simplex can be interpreted as related by temporal succession. This can be demonstrated by inserting the conjunctive Adjunct (afterwards) that marks temporal succession:

As an eighteen-year-old, I met a young man in his twenties.
Afterwards he was working in a top security structure,
(and) afterwards it was the beginning of a beautiful relationship.
Afterwards we even spoke about marriage.

 [3] Leaving aside the metaphorical notion of a theoretical category having expectations, and the fact that this extract is not simply a temporal sequence, the field of the text is a South African woman talking of her own past. Given how few South African women, of the total population, form romantic relationships with top security officers, it is clearly unreasonable for any reader to expect this particular sequence, as opposed to any other, just on the basis of recognising the situational field.

[4] Clearly, there is a need to make each step explicit with conjunctions, at least for two readers, Martin & Rose, since they have demonstrably misunderstood the conjunctive relations in this portion of text.

[5] To be clear, here Martin & Rose mistake the interpersonal Adjunct even (counterexpectancy: exceeding) for a conjunctive Adjunct marking (for them) a logical relation.

[6] To be clear, here Martin & Rose misunderstand the meaning of the text. The use of the mood Adjunct of intensity even signals 'went as far as'; that is, the speaker's meaning is 'we went as far as speaking of marriage'.

Misrepresenting Field And Confusing Reader Expectation With Speaker Meaning

Martin & Rose (2007: 118):

Indeed sequence in time is so consistently expected by story genres that there is often no need to use any conjunctions:

On arriving back at Sandton Police Station, at what they call the Security Branch
the whole situation changed
I was screamed at, verbally abused
I was slapped around
I was punched
I was told to shut up
sit in a chair
then I was questioned
when I answered the questions
I was told that I was lying
I was smacked again...

Conjunction between the first five activities in this sequence is left implicit — they just happen one after another — until the field shifts from physical and verbal abuse to interrogation, and this shift in field is signalled with the explicit conjunction then. We can now expect a different set of activities — concerned with questioning rather than beating. However the interrogators’ response to their victim’s answers was unexpected, at least to the victim, and this is again signalled with an explicit conjunction when.

Blogger Comments:

[1] To be clear, on the authors' previous analysis, there is no way to determine whether the conjunctive relation here is sequence in time or addition, since the omitted conjunction could be either then or and.

[2] To be clear, here Martin & Rose again confuse what the author of the text says with what various others expect.  In this case, the expecting process is mediated (metaphorically) by story genres, and by readers (including discourse analysts).  In the final instance, Martin & Rose imagine that the author's projection of himself in the text mediates a process of not expecting.

[3] To be clear, here Martin & Rose confuse ideational semantics (language) with ideational context (field). In terms of SFL theory, the field of the text — 'what is going on' — is a witness giving evidence at the Truth And Reconciliation Commission.  It is not the contextual field that shifts, but the ideational meanings being instantiated in the logogenesis of the text.

[4] To be clear, these are hindsight claims by Martin & Rose, not meanings made by the author in the text.  This can be demonstrated by comparing the use of these conjunctions in other texts, or by removing the text that follows each of the conjunctions.  For example, the wording

I was punched
I was told to shut up
sit in a chair
then …

does not lead the reader to expect the instantiation of the Process was questioned any more than a wide range of other potential processes.  Similarly, the wording

I was punched
I was told to shut up
sit in a chair
then I was questioned
when …

does not signal any interrogator responses as unexpected by the victim, not least because the author does not express the view that the interrogator responses were unexpected by him. 

Misanalysing Conjunctive Relations By Giving Priority To The View 'From Below'

Martin & Rose (2007: 117-8):

Conjunction helps to manage what we expect to happen in a text. In an exposition, we expect a series of supporting Arguments, and Tutu confirms our expectations by explicitly adding each one. We also expect conclusions to be drawn from the arguments presented, and again Tutu meets our expectations by explicitly announcing each conclusion with Thus. In Chapter 3 (section 3.5) we saw that the unmarked relation in an activity sequence is simple addition, so that and is the most common conjunction in personal recounts, adding one event to another:

The circumstances of my being taken, as I recollect, were that I went off to school in the morning and I was sitting in the classroom and there was only one room where all the children were assembled and there was a knock at the door, which the schoolmaster answered. After a conversation he had with somebody at the door, he came to get me. He took me by the hand and took me to the door. I was physically grabbed by a male person at the door, I was taken to a motor bike and held by the officer and driven to the airstrip and flown off the Island.

Blogger Comments:

[1] To be clear, SFL theory distinguishes between conjunction (non-structural textual grammar) and clause complexing (structural logical grammar).  Here Martin & Rose exemplify conjunction with instances of clause complexing and rebrand the confusion as logical discourse semantics.

[2] To be clear, here Martin & Rose mistake the mental projections of readers ("what we expect") for the verbal projections of writers (texts).  That is, they confuse the meanings made by readers with the meanings realised in the wordings of writers.

[3] To be clear, here Martin & Rose give priority to the view 'from below' (the realisation of meaning as and) rather than taking the SFL perspective of giving priority to the view 'from' above (the meaning being realised by and.  The main disadvantage of this approach is that a single wording can realise distinct meanings.  For example, the conjunction and can realise additive extension, temporal enhancement or causal enhancement; Halliday & Matthiessen (2014: 487):

In the sample text, 3-6 of the 7 instances of and can be interpreted as realising temporal enhancement rather than additive extension — that is, in the authors' rebranded terms: successive time rather than addition.

Misunderstanding Halliday & Hasan's Distinction Between External And Internal Conjunction

Martin & Rose (2007: 117):

Tutu uses addition (also, further) to add Arguments to support his Thesis. And he uses consequence (thus) to draw conclusions from each Argument. These items are not linking events in a field of experience beyond the text; rather they are used to link logical steps that are internal to the text itself. We refer to this system for logically organising discourse as internal conjunction. And the system for linking events in an activity sequence is known as external conjunction (after Halliday and Hasan 1976).

Blogger Comments:

[1] To be clear, in terms of SFL theory, the wording 'field of experience' confuses the ideational dimension of culture (field) with the non-semiotic domain (experience) that is construed by the processes of consciousness as ideational meaning.  This is an important epistemological distinction for a theory that takes meaning to be immanent (a property of semiotic systems only).

[2] This is misleading, because Martin & Rose are not following Halliday & Hasan (1976), since they demonstrably misunderstand the distinction between external and internal conjunction.

Firstly, both types of conjunctive relation are internal to the text.  The external/internal distinction means external/internal to the communication situation. Halliday & Hasan (1976: 240):

Secondly, both external and internal conjunctive relations are used to create text ("organise discourse").  Halliday & Hasan (1976: 241):

Thirdly, the distinction between external and internal conjunctive relations is made on the basis of the metafunctional distinction between experiential and interpersonal, not experiential and textual. Halliday & Hasan (1976: 240):

[3] To be clear, in SFL theory, conjunction is a system of the textual metafunction on the stratum of lexicogrammar, which Martin (1992) has misunderstood and rebranded as his system of the logical metafunction on his stratum of discourse semantics.

Some Of The Problems With The Four General Types Of Logical Relations

Martin & Rose (2007: 116-7):

Both systems use the same four general types of logical relations: adding units together, comparing them as similar or different, sequencing them in time, or relating them causally — as cause and effect, or evidence and conclusion. These four general types are known as addition, comparison, time and consequence. The units they relate range from simple clauses, to more complex sentences, to text phases, to stages of a genre.

Blogger Comments:

[1] To be clear, the original source of the authors' theorising is Halliday & Hasan (1976: 226-73), where the four categories of conjunction are

• additive
• adversative
• causal
• temporal.

However, Halliday (1985) reworked the system of conjunction in terms of the three most general types of expansion:

• elaboration
• extension
• enhancement.

Importantly, these three general categories are manifested throughout the grammar.  Halliday & Matthiessen (2014: 669):

Let us now present a systematic and comprehensive summary of the different grammatical environments in which elaboration, extension and enhancement are manifested: see Table 10-3. As the table shows, the environments of manifestation can be differentiated in terms of (i) metafunction – textual (CONJUNCTION), logical (INTERDEPENDENCY; MODIFICATION) and experiential (CIRCUMSTANTIATION; PROCESS TYPE: relational), and (ii) rank – clause and group/phrase. (The table could, in fact, be extended downwards along the rank scale to take account of patterns below the rank of group/phrase within the logical metafunction: word and morpheme complexes also embody interdependency relations that combine with expansion.) 

Moreover, it is this multiple manifestation that makes possible grammatical metaphor and the expansion of the semantic system, as Halliday & Matthiessen (1999: 294-5) explain:

The whole metaphorical elaboration [of the semantic system] is made possible by a fractal pattern that runs through the whole system. We have suggested that the metaphorical elaboration is a token–value relation; but in order for it to be a token–value relation within the semantic system, it has to be natural in the sense that the token and the value domains have to be similar enough to allow for the token to stand for the value. The principle behind this similarity is the fractal pattern of projection/expansion … 

That is, while grammatical metaphor constitutes a move from one “phenomenal domain” to another … this move is made possible because fractal types engender continuity across these domains: the metaphorical move from one phenomenal domain to another takes place within the one and the same transphenomenal domain.

It can be seen that these later insights and the explanatory advantages they provide are lost in the authors' rebranding of the original Halliday & Hasan (1976) model.  It can also be seen that the major category of elaboration is entirely absent from the authors' model, since the model only includes one type of extension (addition) and three types of enhancement (comparison, time and consequence).

[2] To be clear, in SFL theory, these correspond to cause: reason and cause: result. The relation between evidence and conclusion is not necessarily causal.

[3] Importantly, in confusing Halliday's cohesive conjunction with Halliday's clause complexing, the authors have omitted the major logico-semantic relation of projection from "their" model.  This is because their source, cohesive conjunction, is the textual deployment of expansion relations only.  This omission means that clause complexes involving mental or verbal projection are not accounted for at the level of discourse semantics.  This is a major weakness in the model, not least because it provides neither congruent nor metaphorical relations between the strata of discourse semantics and lexicogrammar for this major type of logico-semantic relation.

[4] To be clear, Martin & Rose model genre as context instead of language, but nevertheless claim that these relations at the level of language (discourse semantics) obtain between units of context (generic stages).  This demonstrates most clearly that the authors do not understand the SFL hierarchy of stratification or the notion of symbolic abstraction by which it is organised.

It can be noted at this point that Martin's model of genre arises, in part, from his misunderstanding of Hasan's Generic Structure Potential (Halliday & Hasan 1989 [1985]: 64), which models potential semantic structures, varying according to genre, as modelling potential genre structures.

Conflating Clause Complexing With Conjunction And Misunderstanding Internal Relations.

Martin & Rose (2007: 116):

Conjunction in other words has two faces. One side of the system interacts with ideation, construing experience as logically organised sequences of activities. The other side of the system interacts with periodicity, presenting discourse as logically organised waves of information.

Blogger Comments:

[1] To be clear, in the authors' source material — originally Halliday & Hasan (1976) and Halliday (1985) — two simultaneous distinctions are made:

• clause complexing vs cohesive conjunction, and
• external vs internal expansion relations.

The first distinction is between

• logical (structural) relations between clauses in complexes, and
• textual (non-structural) relations between text spans of varying extent.

The second distinction is between

• relations between experiential construals ('external' to the speech event), and
• relations between interpersonal enactments ('internal' to the speech event).

In their rebranding of this original work, Martin & Rose

• conflate the first distinction between logical and textual deployments of expansion relations, and
• misconstrue the interpersonal dimension of the second distinction as textual,

and present their misunderstandings of these grammatical systems as discourse semantic systems.

[2] To be clear, the notion of metafunctions "interacting" betrays the authors' misunderstanding of these theoretical dimensions as 'modules'.  See critiques of Martin (1992) on the matter here.

[3] To be clear, as will be seen in later posts, the authors' model of periodicity is largely a rebranding of writing pedagogy misconstrued as linguistic theory.

Grammar-Based Approaches To Conjunction

Martin & Rose (2007: 116):

Where grammar-based approaches such as Halliday and Hasan (1976) and Halliday and Matthiessen (2004) treat conjunctions as a grammatical resource for linking one clause to the next, the perspective we take here models conjunction as a set of meanings that organise activity sequences on the one hand, and text on the other.

Blogger Comments:

[1] This is misleading.  In SFL theory, all content plane systems are 'grammar-based', since it is the grammar that construes the semantics.  The difference here lies in the fact that the sources of the authors' ideas — Halliday and Hasan (1976), Halliday (1985, 1994), and Halliday and Matthiessen (2004) — present these systems as grammatical, whereas Martin & Rose misunderstand their sources and rebrand their misunderstandings as Martin's discourse semantics.

[2] To be clear, the terminological slippage from 'conjunctions' (word class) to 'conjunction' (system) is strategic, not accidental, and calculated to deceive.  In the authors' source material, 'conjunction' also refers to a system (a "set"): a grammatical system of the textual metafunction.  Halliday & Matthiessen (2014: 612):

The Reason Why We Need Conjunction As A Discourse Semantic System

Martin & Rose (2007: 116):

This illustrates one reason why we need to set up conjunction as a discourse semantic system. The meanings of conjunction are realised through conjunctions such as if and then, but they are also realised by other kinds of wordings, and they are frequently left implicit, for the reader or listener to infer.

Blogger Comments:

To be clear, this does not constitute an argument for setting up conjunction as a discourse semantic system, since the various realisations are already accounted for by the two grammatical systems that Martin & Rose confuse and rebrand as Martin's discourse semantic system of conjunction, namely Halliday's system of clause complexing (Halliday & Matthiessen 2014: 438):

and Halliday's system of cohesive conjunction (Halliday & Matthiessen 2014: 612):

On the other hand, for a theoretically-consistent account of the semantic systems that are congruently realised in lexicogrammar as clause complexing, see Halliday & Matthiessen (1999: 104-27). 

Misunderstanding The Text And Misrepresenting Clause Complexing As Conjunction

Martin & Rose (2007: 116):

We showed an example of the work of conjunction in Chapter 1 (section 1.3), in which Helena gave the conditions under which she would have joined the anti-apartheid struggle:

I finally understand what the struggle was really about.
I would have done the same
had I been denied everything.
If my life, that of my children and my parents was strangled with legislation.
If I had to watch how white people became dissatisfied with the best and still wanted better and got it.

We know these are conditions because of the conjunction if which serves to link Helena’s contemplated action I would have done the same, with the conditions under which she would have done so, If my life was strangled… If I had to watch how white people became dissatisfied… . And the same conditional connection can also be realised by inversion of Subject and Finite, had I been denied everything, This kind of Subject-Finite inversion typically functions to ask a question (see Chapter 7, section 7.3 below), but in this instance its meaning is not ‘question’ but ‘condition’.

Blogger Comments:

[1] To be clear, here Martin & Rose begin their confusion of two distinct grammatical systems in the source of "their" ideas: the textual system of cohesive conjunction (e.g. Halliday & Matthiessen 2014: 612) and the logical system of clause complexing (e.g. Halliday & Matthiessen 2014: 438).  The examples provided here are all instances of clause complexing, not conjunction, since the conditional relation obtains structurally, between clauses in a clause complex, rather than cohesively (non-structurally).  It will be seen later that the clause complexing relation of projection is entirely absent from their rebranding of Halliday's grammar as Martin's discourse semantics.

[2] Here Martin & Rose seriously misunderstand the text they are analysing.  As signalled by (what writing pedagogy terms) the "topic sentence" — what Martin & Rose will later rebrand as hyper-Theme — the author here actually expresses how she came to understand the 'struggle' through empathising with the victims of apartheid. 

[3] Here Martin & Rose, contrary to SFL theory, give priority to the view 'from below', how the relation is expressed (as the conjunction if and structural inversion), instead of the view 'from above'.  Halliday & Matthiessen (2014: 49):

Being a ‘functional grammar’ means that priority is given to the view ‘from above’; that is, grammar is seen as a resource for making meaning – it is a semanticky kind of grammar. But the focus of attention is still on the grammar itself. Giving priority to the view ‘from above’ means that the organising principle adopted is that of system: the grammar is seen as a network of interrelated meaningful choices. In other words, the dominant axis is the paradigmatic one: the fundamental components of the grammar are sets of mutually defining contrastive features. Explaining something consists not in stating how it is structured but in showing how it is related to other things: its pattern of systemic relationships, or agnateness …

Moreover, giving priority to the view 'from below' can lead to misunderstanding.  For example, the conjunction if is not limited to realising condition, as demonstrated by If it isn't one thing, it's another where the relation is one of alternation, a subtype of extension; see, for example, Halliday & Matthiessen (2014: 475).

[4] To be clear, the inversion of Subject and Finite marks the irrealis feature of the condition.  More importantly, Martin & Rose do not provide a discourse semantic model that accounts for such inversions, as will be seen in future posts.

Misrepresenting Elemental Ideational Metaphor As Logical Metaphor

Martin & Rose (2007: 115-6):

Finally section 4.6 discusses what happens when conjunctions are realised by other kinds of grammatical classes, such as verbs and nouns; this kind of grammatical metaphor is known as logical metaphor. A method is presented for unpacking logical metaphors to analyse activity sequences.

Blogger Comments:

[1] To be clear, here Martin & Rose misunderstand grammatical metaphor as a realisation relation between word classes and present this intra-grammatical relation as discourse semantics.  There are three obvious problems with this:

1. Conjunction is a class of word, and in SFL theory, word is a unit on the lexicogrammatical rank scale, not a semantic phenomenon. 
2. Word classes are all at the same level of symbolic abstraction, whereas realisation is the relation between different levels of symbolic abstraction, as for example, between strata. 
3. Grammatical metaphor is a relation between strata, semantics and lexicogrammar, not a relation between word classes within the lexicogrammatical stratum.

[2] To be clear, here Martin & Rose misunderstand elemental ideational metaphor as logical metaphor.  The reason that such metaphor cannot be restricted to the logical metafunction is that the metaphor involves realising a logico-semantic relation as experiential elements: circumstance, process, quality or thing.  Halliday & Matthiessen (1999: 245):

Misunderstanding The Distinction Between External And Internal Conjunction(s)

Martin & Rose (2007: 115):

Section 4.2 describes conjunctions that are used to relate activities; as they construe a field beyond the text these are known as external conjunctions. Section 4.3 describes conjunctions that are used to organise texts; as this organisation is internal to the text, these are known as internal conjunctions.

Blogger Comments:

Here Martin & Rose, following Martin (1992), misunderstand the distinction between external and internal conjunction.  To be clear, the source of the distinction is Halliday & Hasan (1976: 239-41), where it is made on the basis of metafunction: experiential vs interpersonal. As Halliday & Matthiessen (2014: 611) explain:

Relations between representations of segments of experience are called external relations, and conjunctions marking such relations are called external conjunctions. …Relations linking text segments in their interpersonal guise are called internal relations – internal to the text as a speech event, and conjunctions marking such relations are called internal conjunctions.

Misconstruing The Distinction Between Logical Complexing And Textual Cohesion As The Distinction Between External And Internal Conjunction

Martin & Rose (2007: 115):

Section 4.1 outlines four general dimensions of conjunction: the difference between conjunctions that relate activities and those that organise texts; the role of conjunctions in what we expect to happen in a text; the four main types of conjunction (adding, comparing, time and consequence); and three types of dependency between clauses (paratactic, hypotactic and cohesive).

Blogger Comments:

[1] To be clear, in SFL theory, "conjunctions that relate activities" are 'external' conjunctions, whether they do so logically and structurally (in clause complexes) or textually and non-structurally (cohesive relations between messages).  On the other hand, "conjunctions that organise texts" are those that function textually and non-structurally — not logically and structurally — whether externally (relating experiential functions) or internally (relating interpersonal functions).

In other words, in this chapter, Martin & Rose misconstrue the distinction between logical structure and textual cohesion as the distinction between external and internal expansion relations.

[2] To be clear, "what we expect to happen in a text" is concerned with the mental projections of readers, whereas text analysis is concerned with the verbal projections of speakers.

[3] To be clear, in the original textual grammatical model (Halliday & Hasan 1976: 242-3) that Martin & Rose here rebrand as logical discourse semantics, the main types of conjunctive relation were identified as:

• additive
• adversative
• causal
• temporal

However, these were later reinterpreted (Halliday 1985) as more delicate types within the three most general types of expansion — elaboration, extension and enhancement — that are manifested throughout the grammar.  In these terms, the authors'

• 'adding' is a subtype of extension, whereas
• 'comparing', 'time' and 'consequence' are all subtypes of enhancement.

Moreover, Martin & Rose, just like the source of "their" ideas, Halliday & Hasan, omit all types of elaboration (and projection) from their model.

[4] To be clear, here Martin & Rose misconstrue non-structural cohesive relations as a type of structural (inter)dependency relation.

Conjunction: A Foretaste Of Some Of The Misunderstandings In Chapter 4

Martin & Rose (2007: 115):

Conjunction looks at interconnections between processes — adding, comparing, sequencing, or explaining them. These are logical meanings that link activities and messages in sequences.

Blogger Comments:

[1] To be clear, in this case, conjunction is Martin's (1992) confusion of two of Halliday's lexicogrammatical systems: cohesive conjunction (a non-structural system of the textual metafunction) and clause complexing (a structurally realised system of the logical metafunction) rebranded as logical discourse semantics.  Evidence here.  This confusion has since been rebranded as 'connexion' to differentiate from Halliday's original ideas.

[2] To be clear, only external conjunction is concerned with relations between processes.  Internal conjunction is concerned with relations between interpersonal functions internal to the speech event; see, for example, Halliday & Hasan (1976: 240) or Halliday & Matthiessen (2014: 611-2).

[3] To be clear, in SFL theory, 'adding' is one type of extension, 'comparing' and temporal 'sequencing' are two types of enhancement, and 'explaining' could be either elaboration or enhancement. That is to say, Martin's misunderstanding of Halliday's system is not grounded on the three most general types of expansion — elaboration, extension and enhancement — and entirely omits projection.  One effect of this is to undermine grammatical metaphor, since Martin provides no congruent relations between semantics and grammar on which to identify metaphorical (incongruent) realisations.

[4] To be clear, the term 'activity' is here used to refer to meaning in texts, whereas elsewhere, inconsistently, it is used to refer to what those who produce texts are doing.

[5] To be clear, the term 'message' derives from Halliday's theory, and refers to the textual semantic unit realised as a clause.  Because cohesive conjunction is a textual system, it realises relations between textual units.

[6] To be clear, the term 'sequence' derives from Halliday's theory, and refers to the ideational semantic unit that is congruently realised as a clause complex.

∞

To be clear, in the previous chapter, Martin & Rose theorised 'activity sequences' and relations between 'activities' in 'sequences' as experiential rather than logical.

Misrepresenting Ideational Metaphor

Martin & Rose (2007: 112):

Ideational metaphor tends to reconstrue our experience of reality as if it consisted of relations between institutional abstractions. These strategies have evolved to enable writers to generalise about social processes, and to describe, classify and evaluate them. One cost is that it may be hard to recover who is doing what to whom; another is that this type of discourse can be very hard to read and understand. Unpacking ideational metaphors as we have shown here can help to reveal how they construe reality and is one key strategy for teaching language learners how they work.

Blogger Comments:

[1] To be clear, this bare assertion, unsupported by evidence, severely under-represents the range of functions of ideational metaphor.  For example, what institutional abstractions and what relations between them are construed by the grammatical metaphor in the volume of the screeching of lorikeets signals the degree to which they are alarmed?

[2] To be clear, on the SFL model, reality is the ideational meaning we construe of experience. Here Martin & Rose identify reality with the domain that is transcendent of semiotic systems.

[3] To be clear, evolution happens 'because' (cause: reason) not 'in order to' (cause: purpose).  For example, in biology, eyes evolved because they afford behavioural advantages, not in order to afford behavioural advantages.

[4] To be clear, the function of ideational metaphor is not merely to generalise about social processes.  For example, what social process is being generalised about by the ideational metaphor in Quantum systems can become entangled through various types of interactions?

[5] This is misleadingly presented as if it is the authors' insight, rather than Halliday's.

[6] To be clear, as previous posts have shown, Martin & Rose reduce all ideational metaphor to one type, elemental metaphor, and their "unpacking" of metaphor consists of little more than rewording the text and ignoring the metaphorical wording, thereby ignoring the junctional nature of metaphor.

Misconstruing A Clause Beneficiary As A Nominal Group Qualifier

Martin & Rose (2007: 112):

In the following example the processes of ‘exposing’ and ‘humiliating’ become things that qualify the penalty, and are themselves qualified by their participant the perpetrator:

Blogger Comments:

Original Text:

Thus there is the penalty of public exposure and humiliation for the perpetrator.

Here Martin & Rose misconstrue an element of clause structure (Beneficiary) as an element of nominal group structure (Qualifier):

Thus	there	is	the penalty of public exposure and humiliation	for the perpetrator
		Process	Existent	Beneficiary

The fact that the prepositional phrase for the perpetrator serves a function at clause rank, and not group rank, is demonstrated by the fact that, unlike a nominal group Qualifier, it can be relocated to other parts of the clause:

1. Thus, for the perpetrator, there is the penalty of public exposure and humiliation.
2. Thus there is, for the perpetrator, the penalty of public exposure and humiliation.