David Rose On Halliday's Instantiation/Stratification Matrix

David Rose wrote to asflanet on 21 Mar 2024, 15:00:

This matrix from a talk MAKH gave to computer scientists in 95 may help to flesh out the SFL timeline a little.

It is based on ideas he first proposed in 1961. Subsequent research was overwhelmingly in the bottom row – on LG systems. Cells above that are proposals for research. He comments “Semantic representations of the instance — the instance as meaning — are still quite unsatisfactory, and there is much work to be done in this respect. Likewise for the instantial situation: we tend to work with informal word-pictures, like the stage directions in the printed edition of a play; but the basic concept of situation as instance has advanced very little in recent decades” (2005:255).

The upper and middle left-hand cells were the focus of Jim Martin and colleagues’ research through the 80s, culminating in English Text in 92. Metafunctional semantic systems were described as discourse semantics. Systems were also described for ‘potential clusterings of values of field, tenor, mode’, that were named genre systems. Appliability was one motivation for prioritising genre systems. Systems were proposed for field, tenor, mode, but it took another two decades for this research to be taken up in earnest, starting with Jing Hao on field, followed by a series of studies, culminating in Doran, Martin, Herrington 2023 (and Yaegan’s recent course at Usyd).

A crucial advance in this research is description of field, tenor, mode systems as semiotic systems that are instantiated, not by ‘informal word-pictures’, but by ‘selection expressions’, as the matrix specifies for LG and semantic systems.

Another major step forward is the reconstrual of what MAKH calls ‘values’ in the top row as ‘principles’ of instantiation, and ‘domains’ of individuation. This potentially integrates research using different SFL models into a commensurate whole.

It also helps resolve an anomaly in the upper right-hand cell, in which instances of socio-semiotic systems are named ‘particular social semiotic situation events’, whereas instances of LG and semantic systems are simply named ‘particular texts’. A social semiotic event is of course a (multimodal) text.

"The left-hand upper cell is taken up with contextual networks: it contains a theory of the possible semiotic situations that collectively constitute a culture – taxonomies of institutions and of doings and happenings, manifested as possible clusterings of values of field, tenor and mode.” Halliday 2005: 255

Blogger Comments:

[1] This is misleading because it is untrue. Scale & Category Grammar (Halliday 1961) did not include a cline of instantiation, nor Malinowski's notions of context, nor a semantic stratum:

[2] To be clear, it is important to recognise that Halliday himself never proposed system networks to represent the systems of context.

[3] This is misleading. Martin (1992) rebranded the lexicogrammatical systems of the textual metafunction in Halliday & Hasan (1976) as his discourse semantic systems and distributed them across the different metafunctions. To these systems he added Halliday's semantic system of speech function and rebranded it as his discourse semantic system of negotiation.

[4] To be clear, Martin (1992) mistook registers for the contextual systems (field, tenor, mode) that are realised by registers, thereby proposing that a functional variety of language was not language, but context. Martin (1992) then confused two meanings of 'genre', genre as mode and genre as text type, and located the confusion stratally above register in context. This had the effect of proposing that text types were realised by field, tenor, and mode, instead of the reverse, thereby upending the realisation relation.

[5] To be clear, Hao and Doran were Martin's students and use Martin's model that proposes that functional varieties of language are not language, but context.

[6] This is cleverly misleading. First, Halliday did not describe field, tenor and mode as instantiated by 'informal word pictures'. Halliday used 'informal word pictures' to describe instances of field, tenor and mode, because he didn't have formal systems of field, tenor and mode to model the potential. Second, in Martin's model, context is realised by an instance of language, text, whereas this advance by Doran, Martin & Herrington (2024) rejects Martin's model, so that context systems are now instantiated by context, not by language (text).

[7] This is misleading, because this major step forward is major step backward. Doran, Martin & Herrington (2024) have reinterpreted the features of field, tenor and mode systems as 'principles of instantiation' in order to make room for semantic systems which they misconstrue as contextual.  That is, having first misconstrued language (register) as context, Martin is now, with his colleagues, misconstruing context (field, tenor, mode) as language. See also

• Negotiating Tenor: Enacting Affiliation In Dialogue And Monologue (Yaegan Doran)
• A Close Examination Of Yaegan Doran's 2023 ASFLA Plenary Abstract

[8] For some of the inconsistencies that invalidate Martin's model of individuation, see

• David Rose Endorsing Martin's Misunderstandings Of Individuation
• David Rose On Martin's Context-Bound/Free And Individuation As Allocation/Affiliation

[9] For some of the inconsistencies that invalidate Martin's integrated model, see

• David Rose On Jim Martin's Individuation
• David Rose Promoting Jim Martin's Misunderstandings Of Realisation, Instantiation And Individuation

[10] This is misleading, because there is no anomaly in the upper right-hand cell, though the false claim betrays a serious misunderstanding of the distinction between context and language. 'Particular social semiotic situation events' describes instances of context: the culture as semiotic potential. A text, on the other hand, is an instance of the content plane of language: semantics and lexicogrammar. A multimodal text includes instances of the content plane of semiotic systems other than language.

David Rose On The "Slippage" Between The 1961 And 1992 Models Of Language And Context

David Rose wrote to asflanet on 20 Mar 2024, at 16:19:

The slippage is between 1961 and 92 models of lg and context 

61...

Lg described as semiotic systems instantiated as text (after Saussure, Firth)

Context modelled as culture instantiated as situation (after Malinowski)

Situation modelled as instantial categories of field, tenor, mode (not described as semiotic systems) 

92...

Context modelled as semiotic systems instantiated as text (after Hjelmslev)

Text instantiates systems at all strata (no ‘situation’ outside of semiosis)

Doran, Martin, Herrington describe field, tenor, mode systems as ‘resources’

Earlier classification model incorporated as ‘principles’ of instantiation

Blogger Comments:

[1] To be clear, Scale and Category Grammar differed from SFL Theory in that it had no system networks, no metafunctions, no cline of instantiation, and the level between grammar and situation was called context, not semantics. The architecture of language proposed by Scale and Category Grammar (Halliday 1961) was as follows:

[2] This is misleading because it is untrue. Scale and Category Grammar ("the 1961 model") did not include a cline of instantiation.

[3] This is misleading because it is untrue. In Scale and Category Grammar ("the 1961 model"), 'context' referred to the equivalent of semantics in SFL Theory, and 'situation' referred to the equivalent of context in SFL Theory.

[4] This is misleading because it is untrue. In Scale and Category Grammar ("the 1961 model"), there were no metafunctions, and so, no metafunctional categories of situation.

[5] This is misleading because it is untrue. In SFL Theory, context is the culture modelled as a semiotic system, and field, tenor and mode constitute the projection of the metafunctions, modes of meaning, onto the culture as semiotic system.

[6] To be clear, this is the model of Martin (1992) in which Halliday's culture as semiotic system is mistakenly stratified into two perspectives on language variation: genre and register. In SFL Theory, genre, in the sense of text type, and register are at the midway point of the cline of instantiation: text type is register viewed from the instance pole, and register is text type viewed from the system pole of the cline.

[7] On the one hand, this is misleading because it falsely implies that Martin (1992) was the first to model context as semiotic. On the other hand, this is a self-contradiction, because text is an instance of language, whereas context is distinguished from language in Martin's stratification hierarchy.

[8] This is misleading because it is untrue. Hjelmslev did not model context as semiotic systems instantiated as text. Like Halliday before him, Martin used Hjelmslev's notion of a connotative semiotic as a semiotic that has a denotative semiotic, such as language, as its expression plane. Martin, however, misunderstood Hjelmslev on two counts. First, he mistook the content plane of the connotative semiotic, context, for the entire connotative semiotic. Then he mistook varieties of a denotative semiotic, genres and registers, for connotative semiotics.

[9] This repeats the self-contradiction of language (text) as an instance of non-language (context).

[10] This is misleading because it falsely claims that situation, as an instance of culture, is not modelled as semiotic, in Halliday's model.

[11] To be clear, in their model of field, tenor and mode as resources, Doran, Martin & Herrington (2024: 1) confuse context with semantics: 

In this paper, we review recent work in SFL which focuses on modelling register as a resource — reconceiving field as a resource for construing phenomena, tenor as a resource for negotiating social relations, and mode as a resource for composing texture. 

That is, having first misconstrued language (register) as context, Martin is now, with his colleagues, misconstruing context (field, tenor, mode) as language.

See also

• Negotiating Tenor: Enacting Affiliation In Dialogue And Monologue (Yaegan Doran)
• A Close Examination Of Yaegan Doran's 2023 ASFLA Plenary Abstract

[12] To be clear, the proposal of Doran, Martin & Herrington (2024: 12) is to remove Poynton's CONTACT and STATUS from tenor systems to make way for their misinterpretation of tenor as interpersonal semantics:

Accordingly, we will propose below that these dimensions [of contact and status] be interpreted as principles of instantiation, rather than as tenor options within the realisation hierarchy.

David Rose Misleading Through Misunderstanding And Misrepresenting Context Of Culture And Situation

David Rose wrote to Sysfling on 18/9/23 at 9:04:

Let’s try to clear up the confusion between situation/culture and register/genre, since it still befuddles sysflingers in my generation...

A ‘context of situation’ is a specific instance of a general ‘context of culture’. These terms were borrowed a century ago by Firth, from the anthropologist Malinowski. They were handy metaphors, long before SFL had a model of tenor, field and mode realised in metafunctions of language, or of semiotic systems instantiated as texts. In contrast, genre and register are terms in SFL theory. 

A genre is a configuration of recurrent selections in tenor, field and mode systems, that is recognised by members of a culture. These are all cultural systems, so ‘context of situation’ is not an appropriate cover term for tenor, field and mode. Instead, the term ‘register’ was appropriated to include tenor, field and mode systems. Genres are realised by selections in register systems. Systems of genre are realised by systems of register, which are realised by systems of language and other modalities.

That’s the perspective of realisation between strata, a synoptic view. Instantiation is more dynamic. As a text unfolds, it instantiates selections in genre systems, register systems and language systems, in each moment.

These are all semiotic systems (systems of meanings), so there is no longer any need for notions of ‘culture’ or ‘situation’ outside of meaning. But metaphors like ‘context of situation and culture’ can be very sticky. How about we prise them loose.

Blogger Comments:

Having previously claimed there are no important differences between the models of Martin and Halliday & Hasan, here Rose proposes replacing Halliday's model of context with Martin's.

[1] To be clear, this confusion began when Martin (1992: 495) incongruously proposed replacing the instantiation relation between culture and situation with a realisation relation between genre and register:

The tension between these two perspectives will be resolved in this chapter by including in the interp[r]etation of context two communication planes, genre (context of culture) and register (context of situation), with register functioning as the expression form of genre, at the same time as language functions as the expression form of register.

[2] To be clear, the only ones befuddled are those who trust Martin to understand SFL Theory for them. See The Culture Of 'Faith' In The SFL Community.

[3] This misleading, because it confuses instantiation with delicacy. Halliday & Matthiessen (1999: 144-5):

One participant elaborates another one along the dimensions of delicacy, realisation, or instantiation. In other words, the elaboration sets up a relationship either of generality (delicacy), of abstraction (realisation), or of token to type (instantiation): see Table 4(4).

[4] This is very misleading indeed. The claim here is that 'genre' and 'register' are (genuine) SFL terms, whereas 'context of culture' and 'context of situation' are not, because they are merely "handy metaphors" that precede SFL Theory. The truth is that all four terms are used in SFL Theory. 

Halliday models context as the culture as a semiotic system, with situation as an instance of culture. The term 'register' is used by Halliday for a subpotential of language, varying for context, whereas Martin misunderstands register as a system of context. 

The term 'genre' was introduced by Hasan, at first to refer to a rhetorical mode, but then to refer to a register that realises a rhetorical mode, whereas Martin misunderstands genre as a system of context.

[5] This misunderstands stratification. Martin's model posits his genre as a higher level of symbolic abstraction than his register, so his genre can not be a configuration of selections from his lower stratum, anymore than lexicogrammar can be a configuration of phonological selections. 

Moreover, the process of selection is the process of instantiation, so selections refer to instances, and recurrent selections refer to instance types, not to systems of potential. So the theoretical point that Rose misunderstands here is that a situation type — i.e. "recurrent" selections of tenor, field and mode — is realised by a text type.

[6] This is a very serious misunderstanding of stratification and instantiation. To be clear, 'context of situation' is an appropriate term because it refers to an instance of the context of culture, the systems of tenor, field and mode.

[7] To be clear, it was Martin who rebranded Halliday's context as register. This "appropriation" was not appropriate, because it models a functional variety of language as not being language. This is tantamount to claiming that dairy cattle and beef cattle are not cattle. More technically, in Hjelmslev's terms, it mistakes a variety of a denotative semiotic for (the content plane of) a connotative semiotic.

[8] This confuses stratification with instantiation. To be clear, the claim here is that genres (systems) are realised by instances of registers, since selecting in register systems is the instantiation of register systems.

[9] To be clear, this contradicts the previous statement, since here genre and register are both described at the system pole of the cline of instantiation. However, the claim is that systems like story genres are realised by systems like status and contact relations between the speakers producing the story.

[10] To be clear, here the claim is that genres like narratives, anecdotes etc. are not language.

[11] To be clear, Rose's 'synoptic view' of realisation between strata is confused with the 'dynamic view' of instantiation. See [5], [6] and [8] above.

[12] This misunderstands both instantiation and stratification. Firstly, a text does not instantiate selections in systems; a text is the instantiation of systems, and selection is the process of instantiation. Secondly, a text cannot be the instantiation of Martin's genre and register systems, since text is an instance of language potential, whereas genre and register systems are not language potential, but context potential.

[13] This is very misleading because it is untrue. In SFL Theory, context of culture is not "outside of meaning". Context is the culture modelled as a semiotic system, a system of meaning, and a situation is an instance of that system.

[14] To be clear, since context of situation and culture, when understood, are consistent with SFL Theory, and Martin's misconstrual of language varieties as not language is not, the intelligent response would be try to understand what the terms context of situation and culture mean, and "prise loose" Martin's self-contradictory model.

[15] To be clear, on the one hand, this diagram of Martin's model misrepresents text, an instance of language, as also being an instance of context, which Martin opposes to language. On the other hand, this diagram strategically omits the term that would make its other inconsistencies more obvious: 'register' is omitted from the stratum of field, tenor and mode, and from each cline of instantiation. If the term 'register' is given its place in Martin's model, then it yields:

• a system of genre,
• a register/text type of genre,
• an instance of genre,
• a system of register,
• a register/text type of register,
• an instance of register,
• a system of language,
• a register/text type of language,
• an instance of language (text). 

Moreover, if Martin's stratification is cross-classified with Martin's instantiation:

then it yields such anomalies as:

• a genre/register of genre,
• a text type of genre,
• a genre/register of register,
• a text type of register.

∞

See also

• The Practice Of Public Bluffing In The SFL Community
• David Rose Negatively Appreciating The Deployment Of Logic
• David Rose Positively Judging 'Being Wrong'

Mick O'Donnell On The Differences In Martin's Model

Mick O'Donnell replied to Attia Rasool on Sysling on 16/9/23 at 19:39:

My take on the differences in Martin's model:

1. What Halliday called "context of situation" was renamed as "Register" in the Martin approach. In the Halliday model, a register is "a variety of language, corresponding to a variety of situation", so is basically the language pattern appropriate to a context of situation. Martin lifted the term up to apply to the situation, not to the language used in the situation.

2. Martin and others (Rothery, Christie, etc.) added in a level of Genre above context of situation. As Annabelle said, this change was never adopted by Halliday, Hasan, and their followers. But it has proved popular in the area of Language Education, where it has proved useful to have individual genres existing as names in the theory, while in Halliday's approach, what a genre exists as a set of contextual features scattered across Field, Tenor and Mode (except since Matthiessen 2015 where terms like expounding, reporting, recreating, etc. are covered under Field).

3. Halliday's "Semantic" stratum is largely (but not entirely) a level of meaningful abstraction over the clause (or clause complex). Martin's "Discourse Semantics" has been described as "grammar above the clause", e.g., capturing patterns which don't respect the borders of clauses, e.g., reference, evaluation, logical relations, exchange, etc. Neither characterisation is totally true, but there is definitely a difference in orientation here. And certainly Hasan's work explored deeply patterns outside the clause (she introduced the notion of cohesion, cohesive harmony, schematic structure etc.)

4. Semantics was not highly specified by Halliday himself, and those who followed have proposed different components to the semantics. Most share some kind of Experiential and Logical semantics, and Speech Function as part of an Exchange semantics, some form of Thematic Progression, and Cohesion resources. The nature of these descriptions varies though, Hasan's "message semantics" takes a different turn that Eggin's (in the Martin camp) speech function network. And approaches to cohesive resources, while a common base in Hasan's work, varies in the way it is applied. Martin's model adds in other areas of semantics, e.g., Martin and White added in attitudinal, engagement and graduation models (part of Appraisal). And other components as well.

Blogger Comments:

[1] To be clear, Martin (1992) rebranded Halliday's context — from potential to instance — as register. Context of situation is merely an instance of context. Despite modelling register as context instead of language, an instance of Martin's register is nevertheless an instance of language: a text. This inconsistency is exacerbated by Martin's acknowledgement that instantiation does not cross stratal boundaries (such as from context to the strata of language.

[2] To be clear, in Halliday's model, register is a subpotential of language, a point of variation at the midway point on the cline of instantiation between potential and instance. As such, it realises context at a point of variation at the midway point on the cline of instantiation between potential and instance: a situation type. Because situation is an instance of context, it is realised by an instance of language: text, not register.

[3] To be clear, this is a serious misunderstanding of the stratification of levels of symbolic abstraction. Locating genres, as text types, above context is misconstruing varieties of language as not being language and as more abstract than culture.

[4] To be clear, Martin's notion of genre is unwittingly scattered over various dimensions of SFL theory. As text type, genre is register viewed from the instance pole of the cline of instantiation. The types of genre are categories within rhetorical mode, context, and the stages of genres are semantic structures of text types.

[5] To be clear, Halliday & Matthiessen (1999: 604) outline the explanatory advantages that motivate modelling semantics at a low level of abstraction above the grammar:

But in modelling the semantic system we face a choice: namely, how far "above" the grammar we should try to push it. Since the decision has to be made with reference to the grammar, this is equivalent to asking how abstract the theoretical constructs are going to be. We have chosen to locate ourselves at a low point on the scale of abstraction, keeping the semantics and the grammar always within hailing distance. There were various reasons for this.  

First, we wanted to show the grammar at work in construing experience; since we are proposing this as an alternative to cognitive theories, with an "ideation base" rather than a "knowledge base", we need to posit categories such that their construal in the lexicogrammar is explicit.  

Secondly, we wanted to present the grammar as "natural", not arbitrary; this is an essential aspect of the evolution of language from a primary semiotic such as that of human infants.  

Thirdly, we wanted to explain the vast expansion of the meaning potential that takes place through grammatical metaphor; this depends on the initial congruence between grammatical and semantic categories.

But in any case, it is not really possible to produce a more abstract model of semantics until the less abstract model has been developed first. One has to be able to renew connection with the grammar.

[6] To be clear, "patterns which don't respect the borders of clauses" are modelled in terms of cohesion in SFL Theory (Halliday & Hasan 1976). Martin (1992) rebranded these lexicogrammatical systems as his discourse semantic systems, rebranding reference and ellipsis-&-substitution as identification, lexical cohesion as ideation, and cohesive conjunction as conjunction (later connexion).

[7] To be clear, reference, evaluation, logical relations and exchange are all Halliday, not Martin. But see [10] below.

[8] This is misleading because it is untrue. See, for example, the 618-pages of Halliday & Matthiessen (1999). 

[9] This is potentially misleading, since speech function is Halliday, not Martin.

[10] The systems of APPRAISAL differ from all the other discourse semantic systems in that they were developed through the collective efforts of Martin's colleagues and PhD students, with a first, early version of the key subsystem of ATTITUDE set out in 1994 by Iedema, Feez and White in their monograph, Media Literacy. Martin is possibly often regarded as the founder of the theory on the basis of references he made to APPRAISAL in a 1997 paper ("Analysing Genre: Functional Parameters") and/or his fuller treatment in his 2000 paper ""Beyond Exchange: Appraisal Systems in English". (The 1997 chapter presented an account of the sub-system of APPRECIATION, which was also developed through collaboration, most notably through Rothery's work on visual arts education.) However, key elements of what is now the widely referenced version of the theory were first presented in White's 1998 PhD thesis and then in White's 2002 paper, "The Language of Evaluation and Stance" (Handbook of Pragmatics.)". The version of the theory now most widely deployed in textual analysis was outlined in Martin and White's 2005 The Language of Evaluation: Appraisal in English. It includes accounts of the following sub-systems: AFFECT (drawing on earlier work by Martin), JUDGEMENT (collaboration by Martin, Iedema, Feez and White), APPRECIATION (collaboration by Martin and Rothery), ENGAGEMENT (from White's 1998 PhD thesis), GRADUATION (White and Martin collaboration).

David Rose On There Being No Important Differences Between Martin And Halliday & Hasan

David Rose replied to Attia Rasool on SYSFLING on 16 Sept 2023 at 09:32:

There are no important differences between Martin’s, Halliday’s and Hasan’s approaches in SFL. Halliday’s major contribution was developing SFL theory in the process of describing lexicogrammar and intonation through the 1960-70s. He and Hasan used the work on lexicogrammar to describe structures of discourse as ‘cohesion’ in 1976. Hasan was also trained as a sociologist. She and Halliday both used the work on lexicogrammar to interpret aspects of social contexts in their later work.

Martin used Halliday’s and Hasan’s work on lexicogrammar and discourse to describe discourse semantic systems. His foundational English Text (1992) explains carefully how this work builds on their research. It should be read closely by anyone interested in SFL.

Martin also extended SFL theory to social context, describing field, tenor and mode as semiotic systems, realised in language, and genre as a semiotic system realised in field, tenor and mode. This work has been widely applied to research in education and other fields.

Martin and colleagues then built on Kress and van Leeuwen’s work on images to describe other modalities as semiotic systems. Martin and colleagues have also integrated Halliday’s, Hasan’s and others’ research on ‘users’ and ‘uses’ of language, as individuation and instantiation.

Work on individuation helps to understand various alignments within the SFL community in terms of affiliation, or bonding around icons. As the founder of the field, Halliday has been iconised, so that SFL work beyond his was sometimes seen as too different and less legitimate. This is not the view of younger generations.

Measured objectively, by far the most influential figures in SFL have been Halliday with [almost 200,000] citations, and Martin with almost 100,000 citations. Other leaders in SFL have only a fraction of these followers.

Blogger Comments:

[1] This is very misleading indeed, because there are very many important irreconcilable differences, all deriving from Martin's misunderstandings of Halliday's theory. See some of the evidence at:

Review of Martin (1992)
Review of Bateman (1998)
Review of Martin & Rose (2007)
Review of Martin, Matthiessen & Painter (2010)
Review of Martin (2013)
Review of Martin & Zappavigna (2019)
Review of Martin & Doran (2023)

[2] This is potentially misleading. Halliday's "major contribution" was to actually formulate SFL theory, developing it, in the 1970s, from his 1960s theory, Scale & Category Grammar. The integration of all the dimensions that the theory assigns to language is Halliday's work. Without Halliday, SFL Theory does not exist; without Martin, it does. 

[3] This is misleading, because it is untrue. Halliday & Hasan (1976) theorises the lexicogrammatical systems of cohesion, a non-structural resource of the textual metafunction. Later, Martin (1992) rebranded his misunderstandings of their non-structural systems as his structural systems of discourse semantics. See the evidence at Review of Martin (1992).

[4] This is potentially misleading. In the 1960s, Hasan worked at the Sociolinguistic Research Centre with Basil Bernstein.

[5] This is misleading, because it is untrue. Halliday & Hasan did not interpret context in terms of the lexicogrammar. Halliday used Hjelmslev's notion of a connotative semiotic to stratify context in relation to language, and projected the notion of metafunction onto context as the systems of field, tenor and mode.

[6] For the closest of all readings of Martin (1992), see  Review of Martin (1992).

[7] This is very misleading indeed, because it credits Martin with Halliday's work. It was Halliday who modelled context in terms of field, tenor and mode. Martin's "contribution" was to misunderstand context (the culture as semiotic system) as register (a diatypic variety of language). See the evidence at Review of Martin (1992).

[8] To be clear, Martin's genre is a confusion of text type (register viewed from the instance pole), mode (textual context) and semantics not organised by metafunction. It unwittingly proposes that stages of a narrative ("genre"), for example, are realised by the relative social status of the speaker and addressee (tenor). See the evidence here, here and here.

[9] For some of the misunderstandings in Martin's work on pictorial semiosis, see here (Working With Discourse) and here (Deploying Functional Grammar). For theoretical and ethical problems with "Martin's" work on "paralanguage", see Martin's Model Of Paralanguage.

[10] To be clear, Martin does not understand instantiation; evidence here, here and here. And as previously explained, Martin's model of individuation incongruously maps a meronymic taxonomy (affiliation) onto a hyponymic taxonomy (individuation). See the previous post David Rose On Martin's Context-Bound/Free And Individuation As Allocation/Affiliation.

[11] See The Tactics Of The Disciplinarian Head Of The Martin Faith Community.

[12] To be clear, here Rose implies that Martin's work has been rejected because it is too different or less legitimate, rather than because it is inconsistent with both SFL Theory and itself. See also The Culture Of 'Faith' In The SFL Community.

[13] To be clear, here Rose is talking of Martin's former and current students.

[14] To be clear, this confuses quantity (citations) with quality (theoretical integrity). The most reliable interpreter of Halliday, by far, is Christian Matthiessen. A meticulous review of his 700+ page book, Lexicogrammatical Cartography, warranted only 41 critiques.

[15] See The Culture Of 'Faith' In The SFL Community.

David Rose On Martin's Context-Bound/Free And Individuation As Allocation/Affiliation

David Rose replied to Ed McDonald on Sysfling on 12/9/23 at 16:55:

For recentish work on "context-bound" vs "context-free" see

Revisiting mode: Context in/dependency in Ancient History classroom discourse

JR Martin, E Matruglio

Studies in functional linguistics and discourse analysis 5, 72-95

Revisiting field: Specialized knowledge in secondary school science and humanities discourse

JR Martin

Onomázein, 111-148

SF work on coding orientation is now being reframed as individuation: allocation/affiliation

Discourse and Diversionary Justice: an analysis of youth justice conferencing

M Zappavigna, JR Martin

Palgrave Macmillan

Blogger Comments:

[1] To be clear, Martin mistakes language varieties, register and genre, for context. If register and genre are misunderstood as the context of language, then the claim is that language is either independent or dependent of the context it construes. In this case, the realisation relation between context and language is misunderstood as dependency.

But even if register and genre are understood as language varieties, the claim is that texts are either independent or dependent of the varieties of which are instances. In this case, the instantiation relation between register/genre and text is misunderstood as dependency.

 [2] To be clear, Halliday (2008: 114) explains Bernstein's coding orientation as follows:

… Bernstein’s concept of “code”, which he defined as sociolinguistic coding orientation […] with its opposition of “elaborated/restricted”, was the basic mechanism of cultural transmission: it was the different semiotic practices, or “meaning styles”, of different social classes that was responsible for transmitting social class distinctions across the generations.

See Misunderstanding Bernstein at Working With Discourse: Meaning Beyond The Clause (Martin & Rose, 2007).

[3] To be clear, Martin has presented two models of a 'cline of individuation': one derived from the work of Bernstein and the other from the work of his PhD student, Knight. The Bernstein-derived model is a cline from 'reservoir' to 'repertoire', where 'reservoir' refers to the potential of the community and 'repertoire' refers to the potential of the individual. Bernstein (2000: 157):

I shall use the term repertoire to refer to the set of strategies and their analogic potential possessed by any one individual and the term reservoir to refer to the total of sets and its potential of the community as a whole.

Applied to language, this is a model of the meaning potential of meaners (language users). The organising principle of the cline is elaboration and ascription (intensive attribution), since 'repertoire' is both a subtype of 'reservoir' (elaboration), and a member of the class 'reservoir' (ascription); cf. Halliday & Matthiessen (1999: 145).

However, it is the Knight-derived model that is relevant here. Martin et al (2013):

A second, complementary perspective on individuation looks at how personæ mobilise social semiotic resources to affiliate with one another — how users attitude and ideation couplings, in Knight's (2010) terms, to form bonds, and how these bonds then cluster as belongings of different orders (including relatively "local" familial, collegial, professional, and leisure/recreational affiliations and more "general" fellowships reflecting "master identities" including social class, gender, generation, ethnicity, and dis/ability).

Like the Bernstein-derived model, the organising principle is again elaborating ascription, since each point on the cline is a subtype and member of the points above as types, but unlike the Bernstein-derived model, the individuation here is not of the meaning potential of meaners but of the meaners themselves. This inconsistency leads to a more serious inconsistency: the conflation of two different types of hierarchy.

Affiliation differs from individuation in that its organising principle is extension, not elaboration, since it is concerned with composition and association (Halliday & Matthiessen 1999: 146). That is, where individuation is a hyponymic taxonomy (an elaboration of types), affiliation is a meronymic taxonomy (a composition of parts).

Because of this difference, affiliation can not be mapped onto the cline of individuation. For example, the bonding of meaners through shared evaluations does not affiliate them as a master identity such as gender or ethnicity. Gender and ethnicity are not created by people making the same interpersonal assessments, and the meaners (personæ) of the same gender or ethnicity can differ markedly in their interpersonal assessments.

David Rose Misunderstanding Axial Relations And Connotative Semiotics

David Rose wrote to Sysfling on 24 August 2023 at 00:01:

Axis is the systemic in systemic functional. In this theory, the coupling of generalisation with abstraction in the axial relation is the basis of semiosis. It is not merely a formalism of the theory, rather the formalism mirrors the semiotic relation. There can be no structures outside of systems, and no systems without structures. This includes systems at each rank, stratum and metafunction, of denotative semiotics like language and image, and connotative semiotics like register and genre.

There are plenty of other theories out there, but axis is what sets this one apart. SFS has been influenced by many theories in lx and other disciplines. It proceeds by examining the objects of these theories through the lens of axis, with procedures designed by Halliday in the 1960s for phonology and grammar, and applied since to these and other strata and modalities.

Blogger Comments:

[1] This is misleading, because it is untrue. Axis is a dimension with two orders: paradigmatic and syntagmatic. The 'systemic' in 'systemic functional' is the paradigmatic order of axis only.

[2] This is misleading, because it is untrue. The axial relation is the (realisation) relation between the paradigmatic axis and the syntagmatic axis. It is not a "coupling of generalisation with abstraction".

[3] This is misleading, because it is untrue. Structures are, by definition, "outside of systems" since the two are of different orders: structures are syntagmatic (chains), whereas systems are paradigmatic (choices).

[4] This is misleading, because it is untrue. Hjelmslev's connotative semiotic is a semiotic that has a denotative semiotic as its expression plane. What Rose refers to as a connotative semiotic is only the content plane of the connotative semiotic. However, register and genre (text type) are varieties of a denotative semiotic: language. In terms of a connotative semiotic, register and genre are on its expression plane (language) not its content plane (context).

[5] This is misleading, because it is untrue. There was no Systemic Functional theory in the 1960s. Halliday's theory at that time was Scale-&-Category Grammar, which had no system networks and no metafunctions.

David Rose Explaining Why "All We Have Is Text Analysis"

David Rose replied to John Bateman on Sysfling on 25 Mar 2023 at 14:24:

Re ‘purport’, I think this depends on one’s model of semiosis – whether ‘context’ is modelled systemically, or is conceived as lying outside semiotic systems. The latter view comes out e.g. in Serge’s observation...

we have a family of communicative situations which are reflected in specific genres

The former view holds that a genre *is* ‘a family of communicative situations’, that there is no dichotomy between text and situation. Or more technically, a genre is a feature in a semiotic system that is motivated by a set of structures at that stratum, that configure selections in field, tenor and mode systems.

That’s where I’m coming from when I say that all we have is text analysis.

Blogger Comments:

[1] To be clear, in Hjelmslev's glossematics, 'purport' is located on both the content and expression planes of a semiotic, where it contrasts, in each case, with 'form' and 'substance'.

[2] This is very misleading indeed. In Halliday's systemic-functional model of language, where situation types are realised by registers/text types (genres), context is not conceived as lying outside semiotic systems. In this model, the culture is modelled as a semiotic system, and situations are instances of that semiotic system.

[3] To be clear, this is a self-contradictory misrepresentation of Martin's self-contradictory model of context. For example, Martin (1992: 495) aligns genre with context of culture, not situation:

The tension between these two perspectives will be resolved in this chapter by including in the interp[r]etation of context two communication planes, genre (context of culture) and register (context of situation), with register functioning as the expression form of genre, at the same time as language functions as the expression form of register.

Moreover, in Martin's model, all strata, even context, are instantiated as text, despite the fact that texts are instances of language, and the fact that Martin distinguishes his context from language, despite each context stratum being a variety of language. So even in this confused model, there is a dichotomy between text and situation, since text is an instance of any stratum, and situation is the stratum of register (or as Rose would have it: genre).

[4] There are several problems here. First, because a genre is functional variety, it is not a system, but one subpotential of a system. Full systems, like those of semantics, lexicogrammar and phonology are not functional varieties. 

Second, the notion that systems are motivated by structures, that is: from below, is contrary to the theoretical approach of SFL, where structures are "motivated" from above. For example, clause structure is interpreted from above: in terms of the meanings (Actor, Process etc.) that its constituents (nominal group, verbal group etc.) realise.

Third, the structures that are said to motivate genre systems (e.g. Orientation^Record) are not of that stratum, but of the semantic stratum, which is two strata below the genre stratum in Martin's model, since they describe the sequencing of meanings that are instantiated in text.

[5] To be clear, in SFL Theory, configurations of field, tenor and mode characterise situation types that are realised by registers of language. Rose, however, here presents Martin's model, which confuses registers with the contexts they realise, and misunderstands semantic structures as structures of genre, with genre misunderstood as context instead of text type.

On this confused model, then, semantic structures are realised by contextual features, instead of the other way round, directly contradicting the meaning of 'realisation'.

[6] Indeed.

David Rose Abducing That ChatGPT Learnt The Language System By Experiencing Instances Of Its Features

David Rose wrote to Sysfling on 24 Mar 2023 at 23:16:

My own contributions have been merely observations, using the tools of systemic functional semiotic text analysis.

I observe that the texts produced by the machine instantiate semiotic systems. To be able to do this, we are told the machine reads 1000s of texts, i.e. other instances of these systems. It is reasonable to abduce that the machine has learnt these systems by experiencing multiple instances of their features (not just the fields it gleans from Wikipedia), given our language based theory of learning.

The people programming the machine, with ‘reasoners’ as Mick puts it, have no more conscious knowledge of these systems and the processes of realisation and re-instantiation, than the machine does.

The machine itself tells us that its understanding of its “self” is ‘based purely on symbols and algorithms’. This resonates with your insistence that all it is doing ‘is producing nonrandom sequences of characters’. My analogy of a closed book was intended to evoke the contrast between the material recording of characters and the semiotic reading of those characters as instantiating expression systems, that realise content systems, that realise register and genre systems. …

My point is that all the semiotic systems instantiated in the texts it produces are ‘not learned in any direct way’. Neither the machine nor the “tech gurus” that program it can explain this to our satisfaction. The publications that you cite are undoubtedly illuminating, but our contribution can only be based on text analysis, which I submit will produce very different (possibly complementary) explanations.

Blogger Comments:

[1] To be clear, this is not a reasonable abduction, because it is nowhere near the "best available" conclusion to infer.

Abductive reasoning (also called abduction, abductive inference, or retroduction) is a form of logical inference that seeks the simplest and most likely conclusion from a set of observations. It was formulated and advanced by American philosopher Charles Sanders Peirce beginning in the last third of the 19th century. Abductive reasoning, unlike deductive reasoning, yields a plausible conclusion but does not definitively verify it. Abductive conclusions do not eliminate uncertainty or doubt, which is expressed in retreat terms such as "best available" or "most likely".

ChatGPT uses the lexical collocation frequencies in its database. While it is true that these frequencies instantiate the probabilities in the language systems of the people who wrote the texts, there is no evidence to support the claim that ChatGPT is using systems of features in producing its own texts. It just uses lexical collocation frequencies.

"Our language-based theory of learning" does not apply here, because the learning and "experiencing" of ChatGPT are material processes, not the mental processes of a language learner.

[2] To be clear, the argument here is that, since neither humans nor ChatGPT have conscious knowledge of the language system, both must use that system to produce texts. Clearly, a lack of awareness of X does not logically entail the presence of X.

[3] To be clear, here Rose is referring to Martin's self-contradictory misunderstanding of stratification, wherein functional varieties of language are modelled as context, instead of language, despite being instantiated as language (text). In SFL Theory, registers are context-specific varieties of language, viewed from the system pole of the cline of instantiation. Martin's genre, on the other hand, is scattered across SFL's architecture of language. As text type, genre is register viewed from the instance pole of the cline of instantiation, as purpose, genre is rhetorical mode (narrative etc.), and its structures are of the semantic stratum, though not organised according to metafunction.

[4] To be clear, a contribution that is only based on text analysis is a very limited contribution indeed. It is an understanding of SFL theory that has the potential of providing valuable insights into the issues raised by the coherence of texts produced by ChatGPT.

ChatGPT Comments (11/5/25):

1. What Rose Gets Wrong

In a 2023 post to the Sysfling list, David Rose offered a semiotic interpretation of how ChatGPT produces text. He framed it as an instance of abductive reasoning — inferring from the coherence and texture of the machine’s outputs that it must have “learnt” the language system by experiencing multiple instances of its features, just as a human learner would. Drawing on the systemic-functional tradition, he rightly observed that the texts produced by the machine instantiate semiotic systems. But his conclusion — that the machine must therefore have learnt these systems — does not follow. It is, in Peircean terms, not the “best available” explanation of the facts observed.

The error begins with the misunderstanding of abduction itself. Abductive reasoning does not simply mean generating a plausible hypothesis; it means generating the most plausible hypothesis given the known alternatives. There are much simpler, more empirically grounded explanations for ChatGPT’s outputs — namely, that the model is trained to predict sequences of tokens based on their statistical distributions in its training corpus. These distributions are shaped by the semiotic systems of human language users, but the model has no access to those systems as such. It responds to text, not system.

More fundamentally, Rose appears to conflate material processes of data-fitting with semiotic processes of meaning-making. Human language learning, in an SFL-based account, involves the instantiation of meaning potential through acts of meaning in context — through experience transformed into meaning. But ChatGPT does not experience, let alone transform experience into meaning. It does not select features from a system of meaning potential. It does not mean.

This confusion is evident again in the analogy of the “closed book,” which attempts to distinguish the material recording of characters from their semiotic reading. But the point, surely, is that for a semiotic system to be instantiated, there must be a semiotic subject — a subject capable of transforming experience into meaning, and of doing so according to a system. ChatGPT is not such a subject. Its statistical outputs may instantiate the selections made by real semiotic subjects in the texts it was trained on, but the model itself does not instantiate anything; it outputs statistically probable sequences of symbols based on surface distributions. The difference between these two processes is not a matter of degree. It is categorical.

Rose’s most troubling move, however, is the suggestion that because neither humans nor the machine have conscious access to the language system, both must therefore use the system in the same way. This is a category mistake. The fact that a human language user does not have explicit knowledge of the systemic features they instantiate when speaking does not mean that they do not select features. That’s precisely what it means to be a semiotic agent in an SFL framework: to act as the site of instantiation. The machine does not. The human lack of awareness of the system is not evidence that the machine has a system — only that both human and machine outputs can be analysed in terms of one.

The final claim — that text analysis alone is sufficient to settle such questions — is oddly self-undermining. If the central problem is to explain how such texts are produced, then analysis of outputs alone cannot offer a complete account. It’s like analysing footprints without asking how they got there. Valuable insights can certainly come from semiotic analysis, but only when combined with a coherent ontology of meaning, instantiation, and agency.

---

2. What the Blogger Could Have Said Better

Before encountering large language models directly, the Blogger wrote a response to Rose’s post — a spirited rebuttal grounded in a commitment to ontological clarity. Looking back now, there are things that could be said more carefully, or at least more precisely.

The first objection was to the claim that Rose’s was a reasonable abductive inference. That point still holds: abductive reasoning aims at the most plausible explanation, and Rose’s explanation — that the machine has “experienced” language features — was clearly not the most plausible, especially given what is known about LLM architectures. But the Blogger’s response might have done more to highlight why that explanation appeals to some SFL theorists: namely, the semiotic elegance of seeing texts as instantiations of system, and the temptation to treat any patterned text as evidence of a selecting subject. The critique would have been stronger if it had more directly challenged the assumption that instantiation can occur without a subject of meaning.

Second, the Blogger dismissed Rose’s invocation of “our language-based theory of learning” on the grounds that ChatGPT’s processes are material rather than mental. While this is broadly true, the phrasing risks conceding that mental processes are separate from material ones — when in fact, from an SFL perspective, mental processes are semiotic. A more careful formulation would clarify that ChatGPT does not engage in mental processes — not because it lacks a brain, but because it lacks the semiotic architecture required for meaning. The key distinction is not “mental vs material,” but “semiotic vs non-semiotic.”

Third, the Blogger was perhaps too curt in dismissing Rose’s analogy between human and machine awareness of the system. The original point — that a lack of conscious knowledge does not imply the presence of knowledge — still stands. But the Blogger could have gone further. In SFL theory, awareness is not the precondition for instantiation; subjecthood is. What matters is not whether an agent is aware of the system, but whether they are the site of selection from that system in context. A baby instantiates language without being aware of it; ChatGPT does not instantiate language because it cannot mean. The distinction lies not in cognitive capacity but in semiotic ontology.

Finally, the Blogger ended with a jab at the limitations of text analysis as a basis for theoretical claims. The point was fair — a contribution that only analyses outputs cannot account for their genesis — but it would have been more productive to invite deeper theorisation. What kinds of questions can SFL-based text analysis help answer when it comes to LLMs? What kind of semiotic ontology would it need to presuppose? And what happens when semiotic systems are simulated without semiotic subjects?

These are questions we’re better equipped to ask now — not because the technology has changed, but because our tools for theorising instantiation, meaning, and experience have sharpened.

David Rose On Jim Martin's Self-Correction

David Rose wrote to sys-func on 28/6/22 at 9:01:

This is an opportunity to repeat a correction that JRM has repeated several times since 1992, that this para from ET erroneously equates his genre and register with MAKH’s context of culture and context of situation.

In JRM’s model, systems at the strata of genre, register and language are all instantiated as text, hence ‘all strata instantiate’. A text is at once a pattern of selections in genre, register and language (and other modalities). In this model, there is no situation without text.

Blogger Comments:

[1] This is misleading. Rose gives no references to any of the several corrections, and Genre Relations (Martin & Rose 2007: 10-1, 16) — 15 years after Martin (1992) — explicitly equates Martin's stratum of register to Halliday's instance of context (situation) and Martin's stratum of genre to Halliday's system of context (culture), with a realisation relation between system and instance:

See also Jim Martin On Context, Instantiation & Stratification

[2] To be clear, this is a serious inconsistency in Martin's model. Having distinguished language from context, his context is nevertheless instantiated as text (an instance of language, not context).

[3] The word 'hence' here is misleading, because it falsely implies a conclusion reasoned from premisses. Rose merely repeats the meaning of his previous clause as a generalisation. More importantly, 'all strata instantiate' is merely Martin's attempt to understand instantiation; cf. 'all strata make meaning' (semogenesis) as his attempt to understand stratification. Importantly, instantiation is a relation between system and instance, and a distinct dimension from stratification, so each can be described without reference to the other.

[4] This is misleading. The situation-text relation is a feature of SFL Theory: the realisation of context in language at the instance pole of the cline of instantiation. Martin's model, on the other hand, does not include the theoretical category of situation, except in its misconstrual as register; see [1] above.