Part 1: Writer, Reader, Artefact — Rethinking AI in SFL
The rise of AI has prompted many of us in linguistics and language studies to ask: what does it mean for a system like SFL when machines can generate human-like text? Does AI “mean” anything, or is it merely simulating human writing?
In exploring this question, it helps to look at two roles central to semiosis: writer and reader. Traditionally, human meaning arises because we inhabit both roles: we construe context to produce text, and we construe text back into context. AI challenges this familiar loop.
AI as Writer?
Consider what happens when AI “writes.” We give it a prompt — a string of symbols — and it produces text. It can be fluent, persuasive, even stylistically sophisticated. Yet, crucially, it does not construe context.
-
Human writer: selects meaning from the system potential and enacts it in context.
-
AI writer: re-patterns prior instances, cycling symbols without ever grounding them in a situation.
The output may look like meaning, but its status is different: it is an artefact, construe-able but not itself a construal.
AI as Reader?
AI can also “read” text — or at least, process it. It parses patterns, predicts next symbols, extracts probabilities. But it does not construe meaning in context.
-
Human reader: interprets text relative to context, system, and prior knowledge.
-
AI reader: produces distributions over patterns, but no construal occurs.
Here again, AI operates outside the loop where meaning emerges. It cycles symbols without instantiating system-in-context.
The Reflexive Loop of Non-Construal
What’s fascinating is that AI can occupy both roles at once. It reads human text (without construal) and writes new text (without construal). This creates a reflexive loop of non-construal: the AI processes and outputs symbols, but meaning only emerges when humans enter the loop, either as writers or readers.
This is where SFL can help: it makes explicit where meaning happens, and equally important, where it does not. AI doesn’t “make meaning”; it produces construe-able artefacts, patterns that humans must interpret to generate semiosis.
Simulation? Or Something Else?
The word “simulation” has been useful, but it has limitations. It implies a kind of imitation of something real — a framing that can be misleading. Perhaps we need alternatives that foreground relational structure rather than magical opposition:
-
Pattern-cycling: AI cycles patterns between input and output without construal.
-
Echo-formation: AI outputs are echoes of human semiosis.
-
Pre-semiosis: outputs that only become meaningful when humans construe them.
All these terms stress that meaning is relational and contextual, not inherent in the symbols themselves.
Key Takeaways
-
AI is neither a writer nor a reader in the SFL sense.
-
Its outputs are artefacts — construe-able, but not meanings.
-
Meaning arises only where humans instantiate system and construe context.
-
This distinction helps clarify human responsibility in mediated semiosis and frames AI as a tool, not a meaning-maker.
In Part 2, we’ll take this further by mapping these roles directly onto the system–instance–text cycle, comparing human semiosis and AI processing side by side. This will make the architecture of meaning visible in a new light, showing exactly why AI challenges our intuitions but also sharpens our understanding of construal.
Part 2: System, Instance, Text — Mapping AI onto SFL
In Part 1, we explored AI as writer and reader, highlighting that its outputs are artefacts, not meanings. Now we can take a closer look at the SFL architecture itself — the system–instance–text cycle — to see precisely why AI challenges our intuitions about semiosis.
The Human Loop
In SFL, meaning emerges through a structured process:
-
System: the structured potential for meaning, socially evolved, open to future construal.
-
Instance: the construal of context through choice, actualising system.
-
Text: the symbolic artefact produced, interpretable by others as meaning.
Here’s how it plays out for humans:
-
Writer: construes the system in context, producing an instance.
-
Reader: interprets the text, linking it back to system and context.
-
Result: meaning emerges relationally, through human construal at both ends of the loop.
AI in the SFL Cycle
AI doesn’t fit neatly into this cycle. Instead, its “loop” looks more like this:
-
Text → Distribution → Artefact
-
AI reader: collapses text into distributional weights (statistical patterns).
-
Distribution: a probabilistic shadow of prior human instances.
-
AI writer: expands distribution into a new artefact (patterned text).
-
At no point does AI instantiate system or construe context. Its outputs are pre-semiosis — artefacts ready for human construal, but not themselves meanings.
Side-by-Side Comparison
| Role | Human | AI |
|---|---|---|
| Writer | Construes context, actualises system | Re-patterns prior instances without context |
| Reader | Construes text in context | Produces distributions over patterns |
| Output | Meaningful text | Artefact, construe-able but not meaning |
| Loop | System → Instance → Text → Construal | Text → Distribution → Artefact (no construal) |
Why This Matters
Seeing AI through the lens of system–instance–text clarifies a crucial point: AI doesn’t make meaning, but it does make artefacts that humans can construe. In other words, the relational architecture of semiosis remains intact, even in a world where AI mediates so much symbolic activity.
It also sharpens the language we use: instead of talking vaguely about “simulation,” we can describe AI as cycling patterns, producing construe-able artefacts, and operating outside the loop of instantiation.
Next Steps
In Part 3, we’ll explore the ethical, pedagogical, and conceptual implications of this distinction. How should we teach students to interact with AI-generated text? How do we preserve human responsibility for meaning? And how does this relational view help us reframe “human specialness” without slipping into mystical language?
Part 3: Ethics, Pedagogy, and the Relational Specialness of Meaning
In Parts 1 and 2, we explored AI as writer and reader, and mapped its outputs onto the SFL system–instance–text architecture. We saw that AI produces artefacts but does not construe context, leaving meaning as a human responsibility. Now we turn to the ethical and pedagogical implications — and consider what it means to speak of human “specialness” in a relational way.
AI and Human Responsibility
Since AI outputs are artefacts rather than meanings, the human user remains responsible for interpretation and deployment.
-
In pedagogy: Students need tools to distinguish between construal and artefact. AI-generated essays, translations, or summaries are not themselves meanings; they are pre-semiosis — texts that require human construal to produce meaning.
-
In research writing: AI can assist with drafting, summarising, or organising, but semantic and epistemic responsibility remains with the human author.
This perspective keeps SFL ontologically clean: meaning arises only where humans instantiate system and construe context, even if AI mediates the symbolic environment.
Relational Specialness, Not Mystical Specialness
There is a temptation to treat human semiosis as magically “special” simply because AI cannot yet construe. But from a relational perspective, what is special is structural, not mystical:
-
Humans are currently the only known nodes where the relational architecture of meaning — system + context + construal — is fully instantiated.
-
AI can cycle symbols, but it cannot instantiate the relational loop.
-
Meaning is special because it emerges from relationships, not because humans possess some intrinsic, sacred property.
Framing it this way avoids anthropocentric hubris while clarifying where meaning happens in a world increasingly mediated by AI.
Pedagogical Implications
-
Teach construal, not imitation: Encourage learners to focus on why a text means what it does, rather than taking AI outputs at face value.
-
Analyse AI artefacts critically: Explore how AI patterns language, which metafunctions it favours, and what contextual assumptions are embedded.
-
Highlight human responsibility: Meaning is relational. When using AI, humans must occupy the construal roles — as writers, readers, and interpreters.
Ethical Horizons
By understanding AI as a pattern-cycling, pre-semiosis engine, we can:
-
Maintain clarity about the limits of AI in producing meaning.
-
Assign ethical and epistemic responsibility to humans, not machines.
-
Use AI as a tool to clarify and sharpen our understanding of semiosis, rather than as a replacement.
Takeaways
-
AI outputs are artefacts, not meanings — meaning emerges only in human construal.
-
Human semiosis is “special” only in a relational, structural sense: humans instantiate the system–context–construal loop.
-
Pedagogy and ethics should focus on helping humans recognise construal and maintain responsibility for meaning.
-
AI provides a lens to see the architecture of semiosis more clearly, not a threat to it.
In Part 4, we'll explore future horizons: how this relational framing could guide SFL in research, AI integration, and symbolic systems more broadly — pushing beyond human–AI comparison toward a general theory of construe-able artefacts.
Part 4: Future Horizons — AI, Humans, and the Symbolic Cosmos
In the previous posts, we explored AI as writer and reader, mapped its outputs onto the SFL system–instance–text cycle, and clarified the relational “specialness” of human semiosis. Now we can take a broader view: what does this mean for the future of SFL, symbolic systems, and our increasingly AI-mediated world?
AI as a Lens on Semiosis
One of the surprising insights from working through AI is that it makes the architecture of meaning visible:
-
By cycling symbols without construal, AI highlights the relational roles humans play in producing and interpreting meaning.
-
This shows us clearly that system + context + construal is where meaning originates, not in the symbols themselves.
-
AI, in effect, becomes a diagnostic tool: it illuminates the structure of semiosis by what it cannot do.
Beyond Human–AI Comparison
Rather than framing AI as a competitor or threat, we can see it as part of a larger symbolic cosmos:
-
Humans: currently the primary sites of construal, producing meaning through relational engagement with system and context.
-
AI: a pattern-cycling engine, generating artefacts that are construe-able, but not yet meanings.
-
Other symbolic systems: we might imagine other nodes (software agents, hybrid human–AI teams, networks of communication) that can mediate or participate in semiotic processes.
This moves us from an anthropocentric lens toward a relational, infrastructural perspective: meaning is not about “humans only,” but about where and how the relational loop is instantiated.
Implications for SFL
-
Research: AI offers opportunities to test theories of construal, register, and metafunction at scale, by providing artefacts to analyse without conflating patterning with meaning.
-
Pedagogy: Students can learn to distinguish construe-able artefacts from actual meaning, developing critical literacy skills.
-
Ethics: Responsibility for meaning remains human; relational framing prevents category mistakes (e.g., attributing human-like understanding to AI).
-
Theory-building: The AI challenge encourages SFL to make its relational architecture explicit, highlighting where meaning emerges, and where it does not.
Towards a Symbolic Cosmos
AI invites us to imagine a symbolic cosmos in which meaning is distributed, relational, and mediated by multiple agents — some human, some artificial, some hybrid. In this cosmos:
-
Meaning is emergent from relations, not intrinsic to any agent.
-
Artefacts can circulate and be construe-able, creating a rich ecosystem for interpretation and learning.
-
Humans retain a special, structural role, but it is not sacred — it is defined by relational instantiation, not essence.
AI, then, does not diminish human semiosis. It clarifies it, extends the horizons of what we can examine, and prompts a more precise, relational understanding of meaning in an increasingly complex symbolic world.
Concluding Thought
In the era of AI, SFL offers a lens to navigate a shifting landscape of symbols. By distinguishing construal from artefact, human responsibility from statistical patterning, and relational specialness from mystical essence, we can engage with AI not as a threat, but as a clarifying force — a mirror that reveals the very architecture of meaning itself.
