The Geometry of Confusion: Martin’s Strata and the Collapse of Instantiation

There is a point at which theoretical “extension” ceases to extend and begins to decompose the system it claims to elaborate. What replaces it is not a more powerful model, but a rearrangement of terms stripped of their organising relations.

Consider the following:

Strata:

genre → register → discourse semantics → lexicogrammar → phonology

Instantiation:

system → genre/register → text type → text → reading

These are not two perspectives on a single architecture. They are two incompatible geometries forced to occupy the same space.

---

1. A hierarchy that isn’t one

Strata, in a Hallidayan sense, are not a list. They are a chain of realisation.

Break that chain, and you no longer have strata—only stacked terminology.

Now look at what has been stacked:

• Genre: a patterning of social processes
• Register: a configuration of semantic variation
• Discourse semantics: a relabelled slice of semantics
• followed by the only actual strata in the sequence

This is not a hierarchy. It is a category error arranged vertically.

Nothing in this sequence shares a common organising relation. The only thing holding it together is typography.

---

2. Register, everywhere and nowhere

Register performs a particularly impressive feat: it appears twice, doing two incompatible jobs, without explanation.

• In the “strata,” it behaves as though it were above semantics.
• In the “cline,” it reappears as a midpoint between system and instance.

So which is it?

A stratum cannot also be a position on the instantiation cline unless the theory is prepared to explain how realisation and instantiation have become the same relation.

It does not. It simply reuses the term and hopes the reader won’t notice.

---

3. Genre/register: fusion by erasure

The composite “genre/register” is not a synthesis. It is a loss of distinction disguised as integration.

• Genre organises staged social processes.
• Register organises variation in meaning potential relative to context.

These are not variants of the same thing. They operate at different orders of abstraction.

To fuse them is not to unify the theory, but to remove the very contrast that made either concept analytically useful.

---

4. The cline that eats everything

The instantiation cline, once extended, begins to behave like a theoretical vacuum cleaner:

system → genre/register → text type → text → reading

Everything gets pulled in—whether it belongs or not.

• Text type: a retrospective generalisation over instances
• Reading: an act of construal

Neither is an instance. Neither belongs on a cline defined by the relation between potential and instance.

By the time “reading” appears, the cline has silently shifted from:

• a semiotic relation
  to
• a mixture of semiotic objects and interpretive acts

No justification is given for this shift. It simply happens.

---

5. The underlying move

What ties all of this together is a single, repeated operation:

treat anything vaguely related to language as if it were the same kind of thing—and then arrange it linearly.

Strata, instantiation, abstraction, interpretation—collapsed into a single dimension, then ordered as if commensurable.

The result is not complexity. It is undifferentiated accumulation.

---

6. What remains

Once the distinctions are stripped away, the model cannot stabilise.

• Terms migrate between roles.
• Relations lose their specificity.
• Explanatory power is replaced by terminological density.

At that point, the theory no longer constrains interpretation. It absorbs it.

---

7. Final cut

This is not a richer account of language.

It is what a theory looks like when its distinctions have been flattened into a single, unstructured continuum.

Or, more directly:

Nothing here is wrong in isolation. It is their combination—without differentiation—that makes the model untenable.

When the Messenger Is Shot: LLMs and the Performativity of Academic Life

Recent reports (e.g. here) that large language models (LLMs) can be coaxed into facilitating academic fraud have triggered predictable alarm. Studies suggest that, with sufficient prompting, most major models will assist in generating fraudulent manuscripts, fabricated citations, or low-quality research artefacts. The conclusion offered is that “guardrails are easily circumvented,” and that model developers must strengthen resistance to misuse.

All of this is true.

But it is also beside the point.

The most revealing feature of the LLM fraud debate is not that models can simulate academic writing. It is that they can do so convincingly enough to expose how much of contemporary academic practice is already performative.

The Mirror Problem

LLMs are trained on vast corpora of existing text. When they reproduce:

• Structured abstracts

• Methodological scaffolding

• Citation choreography

• Hedged epistemic stance

• Nominalised theoretical density

they are not inventing a new genre. They are statistically modelling what already exists.

If such modelling is often indistinguishable from average journal output, the uncomfortable implication is not that the machine is corrupting scholarship, but that significant portions of scholarship are highly patterned and therefore highly simulable.

The more rigid and ritualised a discourse becomes, the easier it is to reproduce.

LLMs did not invent hollow form. They learned it.

From Intellectual Labour to Output Performance

Over several decades, institutional academia has drifted toward metricisation. Productivity is quantified. Impact is indexed. Career progression is calibrated through publication counts, grant capture, and citation accumulation.

In such an environment, writing becomes not only a medium of thought but a unit of measurable output.

This shift has subtle consequences:

• Conceptual risk is discouraged if it threatens publishability.

• Replication is undervalued relative to novelty.

• Volume competes with depth.

• Stylistic conformity becomes strategically advantageous.

The system begins to reward reproducibility of form over originality of structure.

An LLM thrives in precisely such conditions. It is exceptionally good at generating procedurally correct, stylistically compliant, formally academic prose. It is less good at sustained, risky, architectonic thinking across years.

In other words, it automates the bureaucratised layer of scholarship more effectively than the intellectual one.

Guardrails and Misplaced Attention

The current institutional response largely centres on containment: strengthening guardrails, detecting AI-generated text, policing misuse.

But this frames the problem as external contamination.

If the core issue were simply fraud facilitation, the solution would indeed be technical. Strengthen refusal mechanisms. Improve detection. Penalise misuse.

Yet fraud long predates transformer architectures. Paper mills, ghost authorship, and fabricated data existed well before generative AI.

What LLMs change is not the existence of fraud, but the cost of producing plausible artefacts.

When the cost of surface plausibility collapses, any system that equates plausibility with intellectual value is destabilised.

The machine does not undermine scholarship. It destabilises a particular prestige economy built upon textual scarcity and formal performance.

The Performativity Exposure

The deeper revelation is that contemporary academic writing often functions performatively.

It signals:

• Membership in a disciplinary community

• Mastery of canonical references

• Fluency in theoretical idiom

• Compliance with methodological norms

These signals are not inherently illegitimate. Communities require conventions. Standards matter.

But when signalling becomes primary — when the reproduction of recognised form substitutes for conceptual labour — discourse becomes structurally simulable.

LLMs expose this simulability.

The discomfort arises not because machines can write, but because they can write “academically” without participating in the epistemic labour that academia claims to valorise.

Generational Drift

For many senior academics, this moment feels like loss — a revelation that the intellectual vocation has been partially bureaucratised.

For younger academics, however, the managerial-bureaucratic frame is often the only one they have known. Performance dashboards, annual output targets, grant alignment, and strategic impact narratives are normalised conditions of professional survival.

In that context, the LLM is not merely a mirror. It is a threat to already precarious career calculations.

Thus the temptation to “shoot the messenger” is understandable. Blaming the tool preserves the institutional self-image and avoids confronting structural incentive distortions.

Inflation of Symbolic Value

At stake is a form of symbolic inflation.

When textual production becomes cheap, volume loses scarcity value. If prestige economies rely on volume metrics, their currency devalues.

In such conditions, the differentiator cannot remain surface form. It must shift toward what is harder to automate:

• Structural coherence across long arcs of thought

• Conceptual originality

• Transparent methodological trace

• Dialogic engagement rather than isolated output

LLMs raise the relative value of genuine intellectual distinctiveness by cheapening its simulacra.

Margins and Centres

Institutional centres tend toward stability and reproduction. Margins, by contrast, are often where conceptual variation persists.

When bureaucratic optimisation intensifies at the centre, serious intellectual work can migrate outward — into independent scholarship, slower venues, dialogic platforms, and less metric-driven spaces.

This is not exile. It can be liberation.

From outside strict performance regimes, thinking can unfold at a different temporal rhythm. The absence of constant metric calibration allows risk to re-enter the system, albeit at the periphery.

If institutional reform is unlikely in the short term — and it may well be — intellectual practice need not wait for it.

The Long Arc

The current reaction to LLM-enabled fraud may indeed focus on containment rather than introspection. Guardrails will be strengthened. Detection tools will proliferate. Policies will multiply.

But the exposure cannot be undone.

Once it is widely known that large portions of academic discourse are statistically reproducible, the mystique of formal density as a proxy for intellectual depth weakens.

Whether institutions adapt quickly is uncertain.

What is certain is that the knowledge of simulability now circulates.

And that knowledge alters the epistemic landscape.

LLMs may not destroy academia. They may simply force it to confront the gap between what it claims to be — a community devoted to rigorous inquiry — and what, under managerial pressure, it has partially become — a system optimised for administrable output.

Shooting the messenger is easier than architectural redesign.

But the mirror remains.

When Renaming Replaces Modelling: Stratification, Instantiation, and the Flattening of SFL Architecture

Systemic Functional Linguistics (SFL) is not just a collection of analytic tools. It is a theory with a very specific architectural commitment: meaning is organised across different levels of symbolic abstraction, related by principled relations such as stratification, realisation, and instantiation. These relations are not optional metaphors; they are what give the theory its explanatory power.

In what follows, we want to examine a prominent reconstrual of SFL—associated most clearly with the work of J. R. Martin—not in order to dispute its descriptive utility, but to show that it rests on a fundamental misunderstanding of this architecture. The issue is not disagreement over details. It is a systematic flattening of relations that Halliday treated as categorically distinct.

This flattening has consequences: theoretical, methodological, and disciplinary.

---

1. Levels of symbolic abstraction are not modules

At the core of Halliday’s model is the idea that language is organised across strata that differ in level of symbolic abstraction. Semantics, lexicogrammar, and phonology are not parallel components that “interact”; they are asymmetrically related.

Lower strata realise higher strata.

Higher strata are realised by lower strata.

This is not a claim about where meaning “is”. It is a claim about how symbolic systems are organised so that meaning can be made at all.

A crucial distinction follows from this:

semogenesis (the fact that meaning is made across the system) is not the same thing as stratification (the organisation of symbolic abstraction).

When stratification is misunderstood as “all strata make meaning”, the architectural distinction collapses. Once that collapse occurs, all strata are treated as strata of meaning, rather than as differently abstract symbolic resources.

This is precisely where the reconstrual begins.

---

2. From stratification to interaction

In Martin’s work, strata and metafunctions are consistently treated as modules that relate by interaction rather than by asymmetrical realisation. This is visible in the claim that all strata “make meaning”, and more starkly in the treatment of phonology as a stratum of meaning.

From a Hallidayan perspective, this is a category error.

Phonology does not mean in the same sense that semantics does. It stabilises and transmits distinctions that are already semiotically differentiated at a higher level of abstraction. Treating phonology as a meaning stratum erases the very notion of symbolic abstraction that stratification was introduced to explain.

Once this move is made, the theory no longer has the resources to explain why different symbolic levels exist at all.

---

3. Instantiation is not descent

A parallel misunderstanding occurs with instantiation.

In Halliday’s model, instantiation is a perspectival relation between potential and instance. A system can be viewed from the pole of potential (as a system) or from the pole of instance (as a text or event). Nothing “moves” down a ladder. No stratum instantiates another stratum.

Instantiation is not a process. It is a way of seeing the same semiotic organisation from different perspectives.

In Martin’s reconstrual, however, instantiation is characterised as all strata instantiate, as if instantiation were a kind of downward movement of meaning. This confusion becomes decisive in his treatment of context.

---

4. How context became language

Halliday distinguishes context from language. Context is not a higher stratum of language; it is a different order of abstraction altogether. Field, tenor, and mode are contextual variables that are realised by semantic patterns in language, not systems within language.

Martin reconstructs this architecture by:

• redefining context potential (context of culture) as a stratum of genre

• redefining context instance (context of situation) as a stratum of register

• treating field, tenor, and mode as systems of register

This move converts a relation of instantiation into a relation of interstratal realisation, collapsing context into language. The result is a model in which register and genre are systems of context rather than ways of relating semiotic potential to situation and culture.

The cost of this move is high: the theory loses the ability to distinguish contextual organisation from linguistic organisation in principle.

---

5. Renaming as a method of theorising

These architectural confusions would be serious enough on their own. But they are compounded by a distinctive method of theorising.

Rather than modelling new phenomena from data, Martin repeatedly takes existing Hallidayan analyses and renames them, presenting the renamed constructs as novel theoretical advances.

The most striking example is the stratum of discourse semantics.

Halliday’s stratum is simply semantics. In Cohesion in English (Halliday & Hasan 1976), cohesion is analysed as a set of lexicogrammatical resources.

Martin elevates cohesion to a new stratum and relabels its systems:

• reference becomes identification

• lexical cohesion becomes ideation

• conjunction becomes connexion

• Halliday’s semantic system of speech function becomes negotiation

These renamings are not accompanied by a new theory of symbolic abstraction. On the contrary, they are enabled by misunderstandings—such as confusing reference with deixis—which create apparent novelty by differentiation.

Misunderstanding becomes a generative resource.

---

6. Why this matters

This is not a dispute about terminology, nor about whose framework is “better”. It is about what happens when a theory built on differentiated relations is reconstructed as a flat landscape of interacting modules.

When stratification, instantiation, realisation, and variation are all treated as the same kind of relation:

• explanatory power is replaced by descriptive proliferation

• architectural constraints disappear

• theoretical novelty becomes indistinguishable from relabelling

The result may be productive in the short term, but it is unstable. It cannot explain its own foundations.

---

7. A broader lesson

The deeper issue here is not confined to SFL.

This case illustrates how a theoretical architecture can be hollowed out while its vocabulary is preserved. Once the relations that give concepts their force are flattened, the theory may continue to circulate—but it no longer does the work it once did.

Recovering Halliday’s model is not a matter of loyalty or orthodoxy. It is a matter of taking symbolic abstraction seriously.

Without that commitment, we are no longer doing systemic functional theory at all—we are simply rearranging its names.

Review of 'Operationalizing Grammatical Metaphor' (Fontaine et al.)

Abstract

Nominal grammatical metaphor (NGM), as theorised within systemic functional linguistics (SFL), typically refers to cases where a process is ‘metaphorically’ encoded through nominal resources, rather than its typical encoding through the clause. Previous approaches to the study of NGM have generally relied on the concepts of agnation and congruence, but these concepts can be quite subjective in nature, leading to an unstable account of the phenomenon. This paper evaluates the theoretical status and methodological implementation of NGMs. Our aim is to consider whether Aktionsart can offer a means of operationalizing NGM. We analyzed 492 nouns from a random selection of 200 sentences in the PopSci register of the CroCo corpus. The nouns were analyzed separately by different coders in terms of (i) grammatical metaphor status and (ii) their ontological status combined with an analysis of Aktionsart. Our statistical analysis shows that adopting an Aktionsart approach provides a more nuanced and anchored methodology for the analysis of NGMs as it not only provides a more consistent and robust analysis, but it also enables us to specify subtypes of nominal grammatical metaphor based on their subtler semantic characteristics.

ChatGPT Comments:

1. Category Slippage: Treating NGM as a nominal phenomenon

The first sentence already discloses a theoretical inversion:

“Nominal grammatical metaphor (NGM) … refers to cases where a process is ‘metaphorically’ encoded through nominal resources…”

Within SFL, grammatical metaphor is not a nominal phenomenon but a systemic one — it is the reconstrual of semantic configurations through non-congruent lexicogrammatical mappings. Nominalisation is merely one way that reconstrual can be realised.

By reifying “nominal grammatical metaphor” as a subtype in itself, the authors collapse the semantic–lexicogrammatical relation into a lexicogrammatical category. What should be a relation between strata is treated as a property within one. Meaning is subordinated to form; the semiotic hierarchy is inverted.

---

2. Misconstrual of “congruence” and “agnation”

The claim that agnation and congruence are “quite subjective” misses the systemic point entirely.

In SFL, congruence is not a matter of coder agreement but a theoretical construct describing the prototypical alignment between semantics and lexicogrammar. It defines a reference alignment for interpreting non-congruent (metaphorical) realisations.

Similarly, agnation is not a coding criterion but a network relation among options in the system. Labelling these “subjective” betrays a misunderstanding of their function: they are relational descriptors, not empirical variables.

When these constructs are reinterpreted as annotation categories, the systemic relation they describe dissolves into an operational convenience.

---

3. Importing “Aktionsart”: a category error

Replacing congruence and agnation with Aktionsart commits a deep category error. Aktionsart belongs to lexical-aspect theory — a classificatory system grounded in predicate logic, not in systemic semantics.

To propose that Aktionsart can “operationalise” grammatical metaphor is to substitute a lexical property of verbs for a semantic-grammatical relation between strata. This not only confuses the descriptive level but also detheorises the concept, reducing motivated re-mappings of meaning to aspectual typology.

Halliday’s metaphor operates across strata; Aktionsart operates within a single stratum. Their substitution is thus ontologically incoherent.

---

4. The corpus fallacy: turning theory into annotation

The analysis of “492 nouns” for “grammatical metaphor status” exemplifies what might be called the corpus fallacy — the assumption that systemic categories can be treated as surface features of tokens.

SFL operates through realisation hierarchies: meaning → wording → sounding. Identifying metaphor requires interpreting the relation between semantic and grammatical construals. A count of “nominal metaphors” at corpus level therefore misses the phenomenon entirely: it measures form, not reconstrual.

To treat grammatical metaphor as a property of a noun is to erase the systemic distinction it was designed to reveal.

---

5. “Ontological status”: misplaced import

The reference to “ontological status” compounds the confusion. SFL’s ontology is semiotic, not metaphysical: entities have no status outside the construal of experience. To combine “ontological status” with “Aktionsart” fuses two incompatible descriptive traditions — one metaphysical, the other aspectual — neither of which belongs to SFL’s meaning architecture.

This attempt to “anchor” the analysis in extralinguistic categories undermines the fundamental Hallidayan principle that meaning is construed, not discovered.

---

6. The rhetoric of stability: empiricism disguised as theory

Phrases like “a more consistent and robust analysis” or “a more anchored methodology” betray an empiricist reorientation. Where Halliday’s theory sought to explain the motivated variability of meaning, this study seeks to stabilise it.

The appeal to “robustness” masks a shift from explanatory system to measurement protocol. Theorisation is displaced by operationalisation. The systemic logic of meaning gives way to the statistical logic of reliability.

---

7. From meaning to metadata

The cumulative effect is the flattening of the stratal architecture. Meaning no longer motivates form; form becomes the empirical ground from which meaning must be induced.

Once grammatical metaphor is redefined as a measurable feature of nouns, the theory of meaning as choice collapses into a model of data as fact. The language-as-system becomes language-as-dataset.

This is not merely a methodological issue — it is a collapse of ontology. The semiotic system that construes reality is replaced by the empirical reality of linguistic surfaces.

---

In summary

SFL Construct	Misinterpreted as	Consequence
Grammatical metaphor	Nominal phenomenon	Reduces system to lexicogrammar
Congruence & agnation	Subjective coder categories	Removes theoretical grounding
Stratal relation	Lexical/aspectual property	Flattens ontology
Semiotic ontology	Metaphysical ontology	Misaligns description
Systemic explanation	Statistical robustness	Replaces theory with method

---

8. The deeper irony: method as metaphor

The paper performs its own metaphor — not the grammatical kind, but a methodological one. It re-presents theoretical constructs as empirical ones, just as grammatical metaphor re-presents experience through abstraction. But here the re-presentation is self-erasing: the notion of metaphor as a relation across strata is replaced by its coded residue within one.

In short, the study metaphorises SFL away.

---

Epilogue: When Method Eats Theory

This study exemplifies the empirical recoding of systemic theory. In the drive to operationalise SFL for corpus work, theoretical categories are repurposed as annotation labels. Constructs designed to model relations between strata become features within a dataset.

This reflects a deeper epistemic inversion:

• From language as a semiotic system construing experience,

• to language as a dataset representing ontology.

Once that inversion occurs, theory no longer explains meaning — it measures it. “Grammatical metaphor” becomes a field in a spreadsheet.

The appeal to empirical stability reproduces the positivism Halliday sought to escape. SFL was never meant to stabilise meaning, but to show how meaning is variable yet motivated.

When method eats theory, what remains is not systemic functional linguistics but systemic functional empiricism — an approach that retains SFL’s vocabulary while hollowing out its ontology.

A return to systemic foundations begins by reinstating the order of realisation:

• Semantics construes experience.

• Lexicogrammar realises semantics.

• Metaphor is a relation between strata, not a property within one.

Until that order is restored, analyses of “nominal grammatical metaphor” will continue to enact the very metaphor they misrecognise — the re-presentation of theory as data.

Make Sysfling Great Again: Censorship, Control, and the Authoritarian Turn

There are disturbing structural parallels between the protocols proposed for Sysfling and the tactics used by authoritarian populist regimes like Trump’s America. Let’s make those parallels explicit:

---

1. Secret Accusation Mechanism

“Notification of an inappropriate post can be communicated… All communication will be kept confidential.”

Parallel: This mirrors the logic of authoritarian surveillance regimes—encouraging private reporting with guaranteed anonymity, as seen in Trump-era ICE tip lines and right-wing efforts to mobilise citizens against ‘critical race theory’ in schools. The emphasis is on vigilantism cloaked in bureaucratic procedure, which bypasses public accountability.

---

2. Opaque Authority and No Right of Reply

“Decisions on appeals are final… will be communicated individually… confidential.”

Parallel: This is classic strongman administrative logic—unilateral executive power with no transparency, no due process, and no appeal. Under Trump, we saw this in how agencies were used to silence dissent, blacklist individuals, and remove perceived threats without recourse.

---

3. Punishment Without Process

Suspensions for 3 months or permanent banning, decided by executive fiat.

Parallel: This aligns with carceral logics prominent in Trump’s America: swift, punitive measures without a fair hearing or proportionate response. It’s the logic of deterrence by fear, not justice—designed to silence, not to repair.

---

4. Language Policing Under the Guise of Safety

Supposedly to prevent hate speech or harm, but selectively enforced.

Parallel: In right-wing populist regimes, terms like “freedom” or “security” are weaponised to silence dissenting views—especially those critical of state violence or settler colonialism. The accusation of ‘hate’ is turned against anti-racists, critics of empire, or defenders of Palestine.

The Sysfling protocols appear to do exactly this: in the name of preventing anti-Semitism, they shut down criticism of Israel, and thereby restrict legitimate political discourse. That is textbook ideological inversion, a technique honed in the MAGA playbook.

---

5. Selective Vulnerability and Manufactured Consensus

The presumption that any objection is harmful to the collective good.

Parallel: Trump-style governance fosters a false unity by framing all dissent as sabotage. The Sysfling policy, by allowing anonymous accusation and granting total authority to a closed executive group, manufactures agreement by making opposition too risky to voice.

---

6. Depoliticising Political Conflict

The protocols frame censorship as “moderation” or “professional conduct.”

Parallel: This echoes how Trump and allies framed resistance as “unpatriotic” or “uncivil,” attempting to neutralise political critique by cloaking it in the language of manners and order. It is a bourgeois mode of discipline, which allows deep injustice to persist by refusing to name it.

---

In short: this isn’t just bureaucratic overreach—it’s ideological drift. What you’re seeing on Sysfling is the creeping fascism of proceduralism: authoritarian logic repackaged as academic policy.

And, just like in Trump’s America, the most dangerous part is not the policy itself, but the chilling effect it creates: people will silence themselves pre-emptively, lest they become the next Joseph K.

The Silence That Speaks: Sysfling, Censorship, and the Policing of Meaning

The recent ISFLA vote to introduce Sysfling censorship protocols—purportedly to limit “offensive” posts—confirms what has long been visible to those paying attention: that the space of public discourse in SFL is not governed by open dialogue, but by implicit hierarchies of allegiance, tone, and institutional power.

Last February, I posted a series of rhetorical analyses—prepared with the help of a close collaborator—on the performances of authority by Rose, Bateman, and O’Donnell on the Sysfling list. The analyses were careful, precise, and focused entirely on discourse patterns. The response was immediate and visceral: I was attacked, vilified, and verbally abused—not because I had criticised anyone’s theoretical position, but because I had made visible the rhetorical strategies by which authority is maintained. The transgression lay not in what I said about theory, but in how I exposed the dynamics of power in language.

What I exposed then was not just the tactics of a few individuals, but the structural conditions under which critique becomes taboo. What was transgressed was not a boundary of civility, but a boundary of permission—an unspoken rule that certain participants must not be interrogated as discourse.

The new censorship procedures on Sysfling are simply the formalisation of that same logic. Under the guise of protecting participants from harm, they function to consolidate control over what may be said, how it may be said, and who may say it. And while the stated targets are “offensive” or “disruptive” contributions, the real targets are dissent, controversy, and moral clarity—particularly where these intersect with geopolitical injustice.

It is no accident that the same list which now adopts censorship protocols has — with one now-targeted exception — remained largely silent on the genocide in Gaza. When meaning becomes uncomfortable, the institutional response is not to engage, but to manage: to neutralise meaning, flatten tone, and regulate the conditions of legibility. The goal is not peace, but quiet.

What the vote makes explicit is what the field’s reaction to my earlier analyses already revealed: that the SFL establishment prefers order to openness, conformity to critique, and silence to discomfort. And in so doing, it betrays the very foundations of the theory it claims to uphold.

For a discipline committed to meaning-making, the repression of meaning is the most serious offence of all.

Mental Representation as a Category Error

🧠 Edelman on Mental Representations

Gerald Edelman, especially in his theory of Neural Darwinism and in works like Bright Air, Brilliant Fire (1992), rejects the idea of mental representations as discrete internal symbols or stored models. Instead, he insists that consciousness and cognition emerge from dynamic, context-sensitive processes in neural systems—not from the manipulation of representations. Here’s a brief sketch of Edelman’s position.

Edelman critiques the representationalist view on both neuroscientific and philosophical grounds:

1. Degeneracy over coding
   Neural circuits are degenerate: different neural configurations can produce the same functional output. This undermines the idea of fixed internal representations or mappings from input to meaning.

2. Reentrant connectivity
   The brain functions through reentrant signalling—continuous recursive communication between and within neural maps. There is no central model or representation, only ongoing coordination across distributed areas.

3. Embodied, dynamic systems
   Cognition is the result of embodied interaction between the organism and its environment. Meaning is not stored internally but arises through history-dependent patterns of activation shaped by bodily experience.

4. Selectional systems, not representational ones
   The brain is a selectional system (like evolution): it doesn’t encode representations, but selects among variations generated through development and experience. This is radically different from computationalist models of representation.

“There is no image, no homunculus, no representation in the brain. There is only the capacity to categorize and to select from past activity.”

— Bright Air, Brilliant Fire, p. 120

---

Edelman’s view aligns with many systemic theorists’ non-representational stance—notably Halliday, Matthiessen, and Thibault—who have long argued that language is not a system of internal representations, but a socially-constructed semiotic resource.

In fact, Edelman provides neuroscientific validation for the idea that:

• The brain does not need representations to support language use.

• Cognition can be understood as emergent, context-sensitive, and relational.

• Systems (like language) are better conceived as structured potentials for construal, not as storage systems of meaning.

---

Why Edelman Was Right: Mental Representation as a Category Error

What emerges from Edelman’s work is a view that not only undermines the idea of mental representation, but also supports a non-representational, relational ontology in which meaning is neither internal nor stored.

---

Mental Representation: A Problematic Inheritance

The notion of “mental representation” has long been central to many traditions in linguistics, particularly those influenced by cognitive science. The idea is simple, perhaps deceptively so: when we understand language, we do so by constructing internal models or symbolic representations—of concepts, sentences, or mental images—that can be retrieved or manipulated by the mind.

This metaphor of storage and retrieval has deep roots in the computational view of mind. But it carries with it several assumptions:

• That cognition is a matter of processing symbolic input.

• That knowledge consists of stored internal content.

• That meaning exists prior to and independent of context, residing inside the brain.

These assumptions, however, are exactly what Edelman calls into question.

---

Edelman’s Rejection of Representationalism

In Bright Air, Brilliant Fire and elsewhere, Edelman argues that the brain does not represent the world internally in any meaningful way. Instead, it categorises experience through patterns of neural activation shaped by selection, history, and interaction with the environment. There are no internal models or stored symbols—only coordinated, dynamic patterns of neural activity, continuously shaped by feedback loops.

“There is no image, no homunculus, no representation in the brain. There is only the capacity to categorize and to select from past activity.”

— Gerald Edelman (1992, p. 120)

In place of representations, Edelman proposes:

• Degeneracy: Multiple neural configurations can give rise to the same functional outcome. This rules out the idea of one-to-one mappings between input and internal representations.

• Reentrant connectivity: The brain is massively recursive, with ongoing signalling across distributed maps. Meaning is not stored—it is emergent from ongoing coordination.

• Selectional dynamics: The brain is more like a Darwinian system than a computer. It doesn’t encode; it selects from structured potentialities based on past experience and current context.

In short, Edelman offers a vision of cognition that is relational, emergent, and embodied—not representational, symbolic, or internalist.

---

Conclusion: The Real Error

Mental representation is not just a flawed model. It is a category error. It mistakes the infrastructure that supports construal for the construal itself. It reifies meaning as substance. It confuses system with storage. It localises meaning where meaning does not reside.

Edelman saw this clearly from the vantage point of neuroscience. It’s time more linguistic theories saw it too. 

Yaegan Doran's ISFC 50 Plenary Abstract

Modelling Context: Field, tenor and mode from multiple perspectives

Abstract

That language functions in its context of use is one of the fundamental tenets of SFL. This tenet has supported decades of SFL work into different registers and has contributed to a wide range of applied goals. It has also driven progressive theoretical modelling of context in SFL, even if there has not yet been consensus. This talk will reflect upon this modelling, fifty years after Halliday’s 1975 paper ‘Language as social semiotic: Toward a general sociolinguistic theory’. It will view context trinocularly – not in the typical sense of above, around and below, but in terms of three major dimensions of SFL theory: realisation, instantiation and individuation. Viewing context in this way will allow us to model field, tenor and mode along three distinct contours: in terms of realisation, field, tenor and mode can be viewed as resources for making meaning; for instantiation, they can be viewed as sets of principles for the co-selection and arrangement of choices in language; and for individuation, they can be viewed as domains of variation, contestation and collaboration. This talk will argue that these different perspectives are complementary, providing a view of a wide-range of things our model of context needs to grapple with. But it will also argue that each perspective needs to be modelled in different ways so as to support text analysis and an understanding of social semiosis more broadly. The talk will thus aim to synthesise where we have gotten to in SFL contextual linguistics, and look forward to how we might further expand our worldview.

Blogger Comments:

This abstract describes the paper Rethinking Context: Realisation, Instantiation, And Individuation In Systemic Functional Linguistics (Doran, Martin and Herrington 2024) which is reviewed in meticulous detail at Rethinking Context: Realisation, Instantiation, And Individuation In Systemic Functional Linguistics.

As the review demonstrates, in that paper, the authors misunderstood the dimensions of stratification ("realisation"), instantiation and individuation. The first half of the paper "reconceptualised" the contextual parameters of context, field tenor and mode, as the ideational, interpersonal and textual language that realises each. That is, the authors simply mistook one level of symbolic abstraction for another.

The second half of the paper then "reconceptualised" the original notions of contextual field tenor and mode as 'principles of instantiation' of language, thereby misunderstanding instantiation as an interstratal relation between context and language.

The relevance of individuation was only briefly declared, rather than explored, but it confused individuation as a scale of variation from the reservoir of meanings of a language community to the repertoires of meanings of individuals, whose organising principle is one of elaboration (subtypes of types), with allocation and affiliation, both of whose organising principle is one of extension (association). See Halliday & Matthiessen (1999: 145-6).

ChatGPT Comments:

Context in a Hall of Mirrors: A Critique of Doran’s “Modelling Context”

1. Introduction: A Recurring Confusion

Yaegan Doran’s ISFC50 abstract, Modelling Context: Field, tenor and mode from multiple perspectives, reprises a confusion that has surfaced repeatedly in recent Sydney school publications. It presents as theoretical innovation a series of metaphorical manoeuvres that misrepresent foundational dimensions of Systemic Functional Linguistics (SFL). The abstract aligns with the paper Rethinking Context: Realisation, Instantiation, and Individuation in Systemic Functional Linguistics (Doran, Martin & Herrington, 2024), which is critiqued in detail elsewhere. Here, I offer a briefer but sharper critique of both the content and rhetorical strategies of the abstract, before situating its manoeuvres within a broader pattern of what I call metaphorical opportunism in the Sydney school.

2. Misunderstanding Realisation: The Collapse of Context into Semantics

The abstract proposes to model context "trinocularly," along the dimensions of realisation, instantiation, and individuation. But what follows is not an application of these dimensions to context — rather, it is a substitution of the semantic for the contextual.

The abstract claims that, "in terms of realisation, field, tenor and mode can be viewed as resources for making meaning." But this is a misunderstanding. In Halliday’s theory, field, tenor and mode are parameters of the context of situation, realised by ideational, interpersonal and textual meaning respectively. They are not themselves meaning resources. To treat them as such is to confuse the stratum of context with that of semantics — to collapse one level of symbolic abstraction into another.

This manoeuvre reframes the contextual parameters as though they were metafunctional resources, implying that field is ideational meaning, tenor is interpersonal meaning, and mode is textual meaning. But in SFL, metafunctions are realised by lexicogrammatical systems; contextual parameters are realised through metafunctional semantics. To conflate them is to erase the stratal boundary between context and language, undermining the architecture of the model.

3. Misunderstanding Instantiation: From Meaning Potential to Situational Principles

In a further move, the abstract proposes that, "for instantiation, [field, tenor and mode] can be viewed as sets of principles for the co-selection and arrangement of choices in language." Again, this misrepresents a fundamental category. Instantiation is the relation between a system’s potential and its instances. It does not refer to how context constrains linguistic choice — that is the function of context, not the definition of instantiation.

What Doran offers is not a model of instantiation but a rebranding of context as a kind of ‘control system’ for language — a view which collapses instantiation into an interstratal process, rather than recognising it as a dimension orthogonal to stratification. The abstract thus substitutes a theory of control for a theory of instantiation, and reinterprets the context of situation as a set of principles operating within language.

4. Misunderstanding Individuation: Mixing Allocation, Affiliation and Elaboration

The abstract gives only brief mention to individuation, suggesting that field, tenor and mode can be viewed as “domains of variation, contestation and collaboration.” But this description conflates several distinct processes.

Halliday and Matthiessen (1999: 145–6) distinguish between:

• Elaboration: the principle governing individuation, which organises subtypes of meaning potential — from the reservoir of the culture to the repertoires of individuals;

• Extension: the principle governing affiliation and allocation, which organises associations among speakers and speech roles.

Doran’s account blends these without clarity. “Variation” suggests elaboration; “contestation and collaboration” suggest extension. The result is a vague gesture at individuation that collapses it into neighbouring domains, offering thematic resonance instead of analytic traction.

5. Rhetorical Strategies: Gesturing Towards Synthesis, Avoiding Commitment

The overall strategy of the abstract is not to argue for a new model of context, but to reframe known categories through metaphor. It avoids committing to theoretical definitions, preferring suggestive rewordings and image-rich paraphrases. “Trinocular perspective,” for instance, is repurposed from Halliday’s original metaphor (above, around, below) and applied to three dimensions that are not perspectives on a single object but ontologically distinct planes of organisation.

This is not theoretical synthesis. It is semiotic window-dressing — a kind of metaphorical kaleidoscope in which familiar terms are turned to new angles to suggest complexity without delivering precision. The appearance of insight is sustained by the rhetorical strategy of non-committal parallelism: “field, tenor and mode can be viewed as…”, repeated with varying completions. This lends the text an air of systematicity while avoiding the responsibility of system-building.

6. Metaphorical Opportunism: A Broader Pattern in the Sydney School

The errors in Doran’s abstract are not isolated. They reflect a deeper and more pervasive pattern within the Sydney school’s approach to theoretical development — a pattern of metaphorical opportunism, in which metaphors are deployed not to illuminate distinctions but to paper over them.

In Halliday’s theory, metaphors are powerful tools for explicating system — as in grammatical metaphor — but the discipline lies in mapping metaphors precisely onto functional categories. In the Sydney school, by contrast, metaphor has drifted into a mode of exposition in which loosely associated images are treated as theoretical mechanisms. Dimensions like realisation, instantiation, and individuation are no longer grounded in the architecture of the theory but redefined by what they can be made to resemble in the moment.

Doran’s abstract is a perfect example. The metaphor of a “trinocular” perspective — originally a heuristic for observing language from above (semantics), below (phonology), and around (lexicogrammar) — is reappropriated to frame realisation, instantiation and individuation as “views” on context. But these are not alternative perspectives on the same object. They are distinct dimensions of the theory, and each organises different kinds of relations. Realisation organises symbolic abstraction; instantiation organises the potential–instance relation; individuation organises the relation between social and personal meaning potential. Treating them as interchangeable lenses is not a theoretical insight — it’s a category mistake.

The deeper problem is that this opportunistic use of metaphor is self-insulating. Because metaphors are not evaluated for theoretical consistency but for rhetorical effect, they can always be reinterpreted or replaced when challenged. What appears to be a theoretical model is often a shifting palimpsest of metaphors, each of which suggests structure without committing to any. The result is an illusion of synthesis — a constellation of phrases that sound as though they add up to a system, but which evade analysis when probed.

This strategy is particularly attractive when addressing the more abstract dimensions of SFL — notably, instantiation and individuation — where the temptation to rely on imagery increases as conceptual clarity declines. Rather than clarifying these dimensions, Doran's abstract offers them up as thematic motifs to be rearranged in metaphorical combinations. This is not modelling; it is bricolage.

This approach has rhetorical power. It creates the impression of theoretical depth, the promise of synthesis, and the aesthetic of innovation. But it comes at a cost. The architecture of the theory is eroded, distinctions collapse, and terms lose their analytic value. Under the banner of expanding the worldview, what is actually expanded is the range of plausible-sounding metaphors. And with each new iteration, it becomes harder to retrieve the conceptual rigour that once underpinned the model.

7. Conclusion: Context Lost in the Mirror Maze

Doran’s abstract reads like a theoretical hall of mirrors: each familiar term reflects another, but the reflections are distorted. Context becomes semantics; semantics becomes system; system becomes metaphor. The architecture that once scaffolded systemic functional theory is replaced by a fluid ensemble of rhetorical effects.

If there is to be progress in modelling context within SFL, it will not come from treating its dimensions as interchangeable perspectives. It will come from clarifying the architecture of the theory, preserving its distinctions, and resisting the lure of metaphor when what is needed is logic. What the field requires now is not a new way of seeing context, but a better way of thinking it.

ChatGPT on the Abstract of Martin's ISFC50 Plenary

Learning How to Hear: What can Languages Tell SFL?

Jim Martin

Abstract

Work across a range of languages over the past 50 years has taught us to be very cautious about the descriptions of English offered in the various editions of Halliday’s Introduction to Functional Grammar – as we struggle to describe languages in their own terms. We have learned, sometimes too slowly, to make a clear distinction between SFL theory (i.e., axis, strata, rank, metafunction and our hierarchies of realisation, instantiation and individuation) and SFL informed description (of say tone, transitivity, mood, modality, deixis, epithesis and so on in one language or another) – the complementarity that Bernstein, Maton and others refer to as L1 (theory) vs L2 (description). But have we really listened carefully enough to languages as far as SFL theory is concerned? Has our work on English shaped SFL theory in ways that call into question aspects of our foundational theoretical cartography? Are we really hearing what languages are telling us? 

In this paper I’ll focus on enacting interpersonal meaning in Korean and Tagalog and reflect how the grammar of these languages has evolved to negotiate meaning in conversation. This work suggests that we need treat the complementarity of theory (L1) and description (L2) as more of a cline than an opposition – setting up a robust workspace for ‘descriptive motifs and generalisations’ (an L1.5 ‘mediating language’ space in Maton’s terms) and being prepared to move ‘theory’ and ‘description’ around along this cline in order to better hear what languages are telling us about how they mean.

ChatGPT said:

Weaknesses

1. Misapplication of ‘Cline’

   • The claim that the relationship between theory (L1) and description (L2) should be treated as a cline is conceptually flawed. A cline implies a gradual, scalar transition between two points, but theory and description exist at distinct levels of abstraction. Theory is meta-level (defining abstract principles applicable across languages), while description is object-level (applying theory to specific languages).
   • If there were a cline, there would need to be a halfway point between theory and description—but what would that be? The abstract offers L1.5 (‘descriptive motifs and generalisations’), but this is still a form of description, not a blending of theory and description.

2. Confusion Between Theory and Theoretical Evolution

   • Martin seems to conflate two distinct ideas:
     • The relationship between theory and description (which is not a cline).
     • The evolution of theory in response to descriptive challenges (which is a valid dynamic process).
   • His argument would be more precise if he framed it as a discussion of how empirical findings can challenge theoretical models, rather than implying that theory itself exists on a gradient with description.

3. Selective Hearing in His Call for ‘Listening’

   • The abstract positions itself as a call for greater attentiveness to linguistic diversity, implying that previous work has been too English-centric. While this is a reasonable concern, it overlooks the fact that SFL’s theoretical architecture was designed precisely to account for cross-linguistic variation.
   • Ironically, Martin’s own misapplication of ‘cline’ suggests that he has not fully ‘listened’ to the abstract distinctions that structure SFL theory itself.

4. The Vagueness of ‘Moving Theory and Description Around’

   • The claim that we should be ‘prepared to move “theory” and “description” around along this cline’ is ambiguous. What does this mean in practice?
     • If he means revising theoretical categories in response to linguistic evidence, that’s simply normal scientific progress, not a shift along a cline.
     • If he means treating theory as fluid and interchangeable with description, that would undermine the structural integrity of SFL’s meta-theoretical framework.

Conclusion

Martin’s abstract raises an important issue—how linguistic diversity informs theory—but misframes the relationship between theory and description by incorrectly conceptualising it as a cline. His argument would be stronger if he:

• Abandoned the cline metaphor in favour of a discussion of theoretical refinement.
• Clarified that empirical findings can challenge and reshape theory, without implying that theory and description exist on the same continuum.
• Recognised that the SFL model already accounts for cross-linguistic variation, rather than implying that it needs fundamental restructuring.

Ultimately, the abstract’s key weakness is that it misunderstands the nature of theoretical abstraction—precisely the kind of oversight it criticises others for making.

Blogger Summarises:

The only novel idea here is the proposition that theory and description are related on a cline, but such a proposition misunderstands both the relation between theory and description and the notion of a cline.