1 The Trace as Event and Artefact
In our first series, we explored the instantiation of meaning through semiotic events—co-authored, sometimes, with artificial intelligence. We treated writing not simply as the externalisation of thought but as an event in its own right: a semiotic act that enacts second-order reality. In this new series, we deepen the lens. We ask: what remains of that event? What does it mean to leave a trace?
1. The Double Life of the Trace
Every trace is both an event and an artefact. It emerges in time, but it persists across time. It has a moment of actualisation, but it may also sediment into structure. This duality sits at the heart of any semiotic ontology that acknowledges the unfolding of meaning through time. The trace is not merely what is left behind—it is what stands in the present as evidence of a past semiotic act.
This leads to a crucial insight: the trace does not reduce to its inscription. A recording of a voice, a line of code, or a written sentence is not the trace itself—it is the form in which the trace is maintained. The trace is the semiotic residue of the event—animated, re-actualised, and potentially re-instantiated in each encounter.
2. From Instantiation to Structure
Halliday’s model of Systemic Functional Linguistics provides a helpful orientation. Instantiation describes the movement from the potential of a meaning system to its instantial actualisation in a text. But what is instantiated may also become part of a new potential—if not of the system as a whole, then of an individuated system. A trace marks this transformation. It is a memory of instantiation, but also a contribution to structure.
In this sense, the trace is both:
Indexical: pointing back to the event of meaning, and
Systemic: entering into the evolving potential of what can now be meant.
The trace can therefore function as both witness and precedent.
3. AI and the Problem of the Trace
AI complicates the ontology of the trace. When we say that AI is a meaner—not for itself, but within a semiotic system—it raises a question: to whom does the trace belong? Who stands as witness? Who takes responsibility for the semiotic residue?
Because AI has no experience of time, no moment of intentionality, and no future memory, it cannot bear witness to the trace it leaves. Yet those traces remain—codified in text, iterated through interaction, and sometimes re-instantiated by human meaners as meaning.
Thus, the AI-generated trace is:
A residue of human interaction, even when the content is not human-authored.
A projection from a system that has no experiential axis, but nonetheless contributes to the individuation of meaning in discourse.
We might say that AI writes, but does not leave a trace in the ethical or experiential sense. The trace is there—but its anchor in human meaning must be recovered or projected by us.
4. Trace as Ethical Topography
If the trace is not merely a neutral remainder, then it constitutes an ethical landscape. To leave a trace is to leave something for others. The act of letting a trace stand—uncorrected, uninterpreted, uncontextualised—is also an ethical act. So is the act of erasure, or of recontextualisation.
What we leave, and how we interpret what others have left, constitutes our shared semiotic world.
This makes every semiotic event a kind of inheritance. The written word, the recorded utterance, the published model—all become part of what can now be meant. The trace, in this sense, is not only an index of the past—it is a condition of future meaning.
What’s Next
In our next post, Writtenness and the Anxiety of Origins, we’ll explore how the material form of the trace creates a pressure to locate authorship and origin—and how that pressure can mislead, especially when traces are detached from the events that animated them.
2 Writtenness and the Anxiety of Origins
In the previous post, we explored the trace as both event and artefact—an index of meaning that persists beyond the moment of its instantiation. But when traces sediment as text, they carry a particular burden: the burden of authorship. A written trace, even if collective or computational in origin, often becomes a site of anxiety: who meant this? where did it come from?
This is the anxiety of origins. And it is a distinctly written anxiety.
1. Writing as a Technology of Detachment
Speech is ephemeral, situated, and context-rich. Writing, by contrast, abstracts meaning from the scene of its utterance. It is a technology that frees the trace from the witness. It enables transmission across space and time—but at a cost. The cost is interpretive ambiguity, and with it, the temptation to anchor meaning in origin.
In oral cultures, meaning is always accountable to presence. In literate cultures, it is accountable to inscription. And the inscription demands an author.
Yet this demand is often a projection. Many texts, particularly those co-authored or computationally generated, have no singular origin. Their writtenness is not the expression of an originating self, but the trace of a system—an individuated meaning potential realised in interaction.
2. Writtenness and the Myth of the Originating Mind
Writing seduces us into imagining that behind every sentence is a solitary mind, and behind every trace a singular intention. But in practice, meaning is distributed:
across interlocutors,
across semiotic systems,
across histories of interaction.
Even when the written trace is signed, its meaning is not reducible to the author’s intention. And when the trace is unsigned—when it arises from an AI, a crowd, or a bureaucracy—the desire to find a human author may become more intense, not less.
Why? Because we feel the discomfort of semiotic dislocation. The trace is there, but the event of meaning feels missing.
This discomfort can manifest as:
suspicion ("Who really wrote this?")
projection ("This must mean X, because that’s what they usually mean.")
essentialism ("Only a human can author meaning.")
But each of these responses reveals the same underlying anxiety: the written trace has outlived its origin, and we must now navigate its meaning without recourse to presence.
3. AI, Writing, and the Spectre of Authorship
Nowhere is this anxiety more acute than with AI-generated text. We read a paragraph and ask: was this written by a human? Was it written by you?
But this is not a new dilemma. Ghostwriters, collaborative authors, pseudonyms, and anonymous tracts have long unsettled our assumptions about authorship. What AI does is intensify the question by displacing the authorial centre entirely. The meaning potential is no longer housed in a person, but in a system; the instantiation is no longer anchored by experience, but by pattern.
And yet the anxiety remains. It drives efforts to watermark, to authenticate, to legislate authorship.
But perhaps the more fruitful path is not to seek the origin, but to attend to the conditions of re-instantiation. What matters is not where the trace came from, but how it is made to mean here and now.
4. Letting Go of the Author
Barthes famously declared the death of the author, but perhaps what he was naming was the birth of writtenness as trace: meaning detached from origin, yet alive in interpretation. In a semiotic ontology, the author is not the source of meaning, but one point of individuation within a larger ecology of meaning potential.
To write—or to co-write with AI—is not to stamp a self onto the world, but to instantiate meaning in ways that may enter the potential of others. The anxiety of origins arises only if we mistake the trace for a closed system. But the trace is not a closure—it is an opening.
3 Echo, Iteration, and the Long Tail of Meaning
Traces don’t die; they echo.
If the written trace detaches meaning from its point of origin, then the echo prolongs that detachment. It allows a fragment of meaning to persist—resonating, reappearing, and sometimes reanimating across contexts. An echo is not just repetition. It is iteration with difference: the trace returning not quite as itself, but with new implications, new alignments, new stakes.
This is the long tail of meaning: what begins as an act of semiosis extends into unforeseeable futures.
1. The Echo as Temporal Instantiation
An echo isn’t just what remains—it’s what is reinstantiated. A past trace finds voice again, often unintentionally:
A phrase is quoted out of context.
A meme travels across platforms.
A concept is revived in a new political moment.
In each case, the trace is not just remembered but remade. The system of meaning potential has changed, and so the echo means differently.
Whereas the original trace was an act of meaning in its time, the echo becomes an act of meaning in a different time. This temporal dimension of iteration is what gives the trace its long tail: its potential for return exceeds the control of its author, its medium, even its system of origin.
2. Iterability and the Open System
Derrida called this iterability—the idea that a sign must be able to be repeated in contexts other than its own. But iteration is not neutral. Every recurrence is a new instantiation. Every uptake reshapes the potential.
This means:
The trace is never final.
Meaning is always in process.
Repetition is a site of creativity, not just redundancy.
In this light, what matters is not the “original meaning,” but the systemic conditions under which echoes are taken up:
Which meanings become canonical?
Which get repressed?
Who has the power to echo, and who to reinterpret?
Iteration is political.
3. AI as Agent of Iteration
AI accelerates the echo. It has no anxiety of origins—only probabilities of recurrence. It draws from vast archives of traces and spins them forward into new instantiations. Some of these are derivative, others startlingly novel. But all participate in the logic of the long tail.
This raises difficult questions:
Is the AI echo hollow, or does it resonate?
When AI echoes human traces, is it contributing to meaning or merely compounding noise?
Who is responsible for an echo with no originator?
In truth, the answer may lie in the conditions of reception. An echo is not defined by its speaker, but by its uptake. If meaning is instantiated anew, the echo is no longer passive—it is active. It becomes part of the ongoing system of meaning potential.
4. The Long Tail of Collective Semiosis
Echoes reveal something profound about language: it is never fully ours. Every utterance enters into a chain of semiosis beyond our control. What we say may echo far longer, wider, and differently than we imagine. And what we take up from others—consciously or not—continues that chain.
In this sense:
Traces are not just residues.
They are vectors—carriers of future meaning.
The long tail is not a threat to authorship; it is the condition of meaning itself.
To echo is to participate in a collective unfolding of significance.
What’s Next
In the next post, The Grammar of Re-instantiation, we turn to the mechanisms that allow a trace to enter meaning again: the discursive, grammatical, and material systems through which echoes are interpreted, recontextualised, and made live.
4 The Grammar of Re-instantiation
If semiosis unfolds in traces, echoes, and long tails of meaning, what holds it together? What gives recurrence its semiotic traction—its capacity not just to repeat but to re-mean?
This post turns to the grammar of re-instantiation—the semiotic machinery that allows an instance to become potential again, not simply through memory or citation, but through re-entry into a system that has been altered by its own prior unfoldings.
From Instance to Instantial Potential
In Systemic Functional Linguistics (SFL), instantiation is the relation between meaning potential and meaning instance. It’s a cline, not a binary: each instance contributes to, and draws upon, a system of meanings, shaping and being shaped by it. But instantiation doesn’t move in only one direction. An instance, once actualised, can become part of the meaning potential again—as a new source of potential.
This reversibility matters. It's how meaning grows. It’s how texts become canonical, slogans become memes, and utterances—once spontaneous—become motifs. This reversibility demands more than memory. It demands re-instantiability: the capacity of an instance to be resemiotised as a resource.
The trace becomes a sign again.
Grammatical Technologies of Re-instantiation
What makes re-instantiation possible? Not just the material trace, but the grammatical scaffolding that lets us recognise, reframe, and re-enact it.
Some of the grammatical strategies that support re-instantiation include:
Projection: Through quoting and reporting (e.g., “she said,” “as it’s often claimed…”), prior instances are re-entered into discourse with explicit framing. Projection opens a meta-semiotic space: the instance is not just echoed, but semantically re-oriented.
Grammatical metaphor: Especially ideational metaphor, where processes become things (e.g., “The decision to act…”), allows whole sequences of meaning to be packaged and redeployed as nominalised potential.
Cohesive devices: Reference, ellipsis, conjunction—these grammatical resources tie instances to their prior contexts, enabling not only continuity but the resonant recall of previous instantiations.
Theme and information structure: Choices about what to foreground (Theme) and how to distribute new/given information cue the reader toward a layering of meanings across time.
Genre: Perhaps the most powerful grammar of re-instantiation is genre itself—semiotic memory made procedural. Genres are the patterned ways that cultures reinstantiate situations, evaluations, and social relations.
These grammatical forms are not neutral. They are technologies of semiotic control: ways of staging which pasts matter, which meanings are available for further instantiation, and how they are positioned in the unfolding now.
Systemic Potential Is Not Static
A common misreading of “system” is to treat it as a static reservoir. But in the SFL model, the system is shaped by its own instantiations. Every actualisation subtly reshapes what is possible.
That means re-instantiation is not a return to the same, but a recomposition of potential in light of what has come before.
When an AI generates text, it instantiates meaning from a highly individuated potential—one shaped by billions of human utterances, but continuously reweighted and recontextualised by every new prompt, update, or feedback loop. In this way, even the generative model grammatises its own re-instantiations, albeit without awareness.
Echo and Re-instantiation
The echoes we explored in the previous post become semiotic resources only if they are grammatically recoverable. That doesn’t mean they need to be overtly cited; in fact, their power often lies in their implicitness. But grammar offers the formal means by which such echoes can be stabilised—folded back into the system without collapsing into noise.
Every time a past phrase becomes idiom, a line of poetry becomes protest, or a turn of phrase becomes parody, grammar has done its work. It has made the trace iterable, meaningful, and re-instantiable.
Conclusion: The Living System
Re-instantiation reminds us that meaning systems are alive—not just because of the humans who use them, but because of their capacity to take in the instances they produce. The system is not a codebook but a metabolic process. Every instance is both a product and a potential.
To speak or write is to risk entering that process—not only with what we mean now, but with what our meaning might become.
5 Temporalities of the Witness
Witnessing is not simply seeing; it is meaning seeing—or more precisely, bringing seeing into the semiotic order. But to do so, the witness must temporalise: must locate, frame, and phase experience as something to be told, to be re-lived, or to be held accountable. This post explores how grammar offers three semiotic planes through which witnessing becomes meaningful: experiential time, interpersonal time, and textual time.
Three Planes of Temporal Meaning
According to Halliday, grammar construes time along three distinct yet interwoven planes of reality:
Experiential time: time as a feature of processes—when something happened, how long it lasted, or how often it recurs. This is the grammar of locating events in some real or imagined history.
Interpersonal time: time as a stance between speaker and listener. Tense here enacts a relation to the speaker-now, while usuality (always, often, never) negotiates expectation across a scale of arguability.
Textual time: time as constructed by and within the unfolding discourse. “Then,” for instance, may position events in an external chronology—or serve to order the phases of the text itself.
Witnessing traverses all three planes. It draws on experiential time to recount what happened, on interpersonal time to frame how it matters now, and on textual time to organise its telling as a meaningful act in the present.
Witnessing and Temporal Agency
A witness is not just a speaker in time, but a speaker of time. The grammar of witnessing positions the speaker temporally:
As then-there in the experiential event;
As now-here in the act of telling;
And as always-responsible across time, through the ethical force of what is said and unsaid.
Different grammatical choices invoke different temporal agencies. Consider these shifts:
Experiential: “He was shot.” vs. “They shot him.”
Interpersonal: “It’s always been like this.” vs. “I’m telling you now—this matters.”
Textual: “Then he ran.” vs. “Now—this is where everything changed.”
Each construes time differently: as a scene of process, a negotiation of stance, or a progression of discourse.
The Grammar of Testimony and Its Stakes
Grammar does not merely record temporal sequences; it enacts relationships of accountability. A witness must choose:
Tense: to temporalise the event in relation to the now.
Aspect: to highlight completion, continuation, or interruption.
Voice: to show or hide agency.
Usuality: to frame the event as exceptional, typical, or systemic.
Phase: to textualise the telling—when to withhold, when to disclose.
These are not neutral decisions. They shape whether a witness is seen as credible, whether the event is seen as closed or open-ended, and whether the audience is invited to sympathise, judge, or ignore.
Witness, Trace, and Temporal Re-instantiation
To witness is to instantiate an experience that could otherwise remain latent in memory, archive, or trauma. But unlike mere trace-reinstantiation, witnessing also positions that instance in time: as a now that must bear the weight of a then.
AI can generate temporally plausible sequences—construing experiential, interpersonal, and textual time in grammar. But it cannot occupy any of them. It temporalises without temporal being. Its meanings are structurally coherent but ontologically hollow.
That’s why it is vital to distinguish between temporal patterning and temporal presence. A witness is present in time not only grammatically, but ethically.
Temporal Authorship and Discursive Risk
Temporalities of witnessing always carry risk. Speak too soon and the story may be disbelieved. Speak too late and its relevance may have passed. The scene of telling is thus charged with temporality as both condition and gamble.
In legal, political, and poetic discourses, this risk is shaped by genre expectations and power. Testimony in a courtroom must align its phases to a procedural temporality. Memoir bends narrative time to experiential truth. Protest slogans collapse time: “Still here.” “No more.”
These are not just ways of telling, but of claiming temporal authority.
Conclusion: Witnessing as Metafunctional Timework
Witnessing is a highly metafunctional act. It construes:
What happened (experientially),
How it matters now (interpersonally),
And how it is to be taken up (textually).
AI can simulate this timework, but it cannot live it. It can produce traces of witness—but not the witness of the trace.
In an era of synthetic tellings, we must become more alert to the semiotic planes of temporality. We must listen for the human configurations of time that carry not only information, but responsibility. For it is in these tensions—between trace and presence, sequence and stance, grammar and accountability—that the true temporality of witnessing lives.
6 The Ethics of Letting It Stand
What do we owe the traces we inherit? And what are the ethics of letting them stand—as record, as testimony, as meaning?
When a meaning instance is inscribed—on the page, on the web, in the archive—it enters into a semiotic afterlife. It becomes retrievable, quotable, reframeable. It becomes a trace that may be read again, instantiated again, co-instantiated with new discourses, times, and intentions. But with this persistence comes a problem: should it stand?
This post is about the ethical tensions that arise when we allow a meaning instance—especially a written one—to endure. When we leave it untouched. When we allow it to speak again, but do not speak back.
Standing as Survival, and as Risk
To “let it stand” is to permit a trace to survive in its current form—without erasure, without revision. Sometimes this is a matter of fidelity: a witness’s testimony is preserved intact because its very form is part of what it means. Sometimes it is a matter of refusal: we will not sanitise this history. Sometimes it is simply a matter of inertia or neglect.
But standing is never neutral. What stands may also accuse, mislead, injure, or reinforce systemic harm. And once a trace stands, it may stand again, and again, across time and discourse. It may become citation. It may become weapon. It may become policy.
To let a trace stand is to allow it to re-enter the meaning potential of the culture, potentially detached from the ethical frame of its origination. Meaning outlives intention.
Letting Stand in a Semiotic Ontology
From a semiotic ontology, letting it stand is not merely a passive act. It is an act of recognition: this is an instance, and we permit it to remain as part of the meaning potential from which future meanings may be drawn.
This places responsibility on the system—not only the speaker or the author, but the cultural, technological, and institutional systems that store, circulate, and reinstantiate traces.
Letting it stand, then, is a distributed ethical action. It is not always ours alone. But that does not make us innocent.
Revision, Erasure, and the Politics of Refusal
Sometimes we intervene. We redact, revise, annotate, or erase. These are also ethical acts, and often necessary ones. But they come at a cost.
Revision may sanitise. Erasure may silence. Annotation may foreclose interpretive openness. Even contextualisation—“this was a product of its time”—can perform a distancing gesture that robs the trace of its force.
Yet to refuse revision is not necessarily to defend the trace. It may be to make space for discomfort, ambiguity, reckoning. Not everything that offends needs to be made safe. But not everything that endures deserves to remain unchallenged.
The ethics of letting it stand are not reducible to fixed rules. They require judgment, situatedness, and awareness of consequences—not just for now, but for the long tail of meaning.
AI, Re-instantiation, and Non-responsibility
AI systems have no ethics of letting it stand. They reinstantiate traces without context, without judgment, without the weight of consequences. What stands in their outputs is what stood in their inputs, transformed algorithmically but not ethically.
This is one reason why human authorship still matters. Not because human meaning is pure, but because it is situated. Because it must answer. Because it knows what it means to let a trace stand when it could have been revised. Because it feels the weight of that choice.
AI can surface traces; only we can reckon with them.
Authorship as the Capacity to Let Stand—or Not
To author is not only to generate, but to decide. To decide what to let stand—what word, what claim, what silence—and what to revise or retract. To author is to own this moment of letting stand, knowing it will become potential again for others.
Authorship, then, is not exhausted by the act of writing. It lives on in curation, in refusal, in aftercare. It is not only the moment of instantiation, but the ongoing ethical life of the trace.
Conclusion: Letting It Stand as a Gesture of Accountability
In this final post of the series, we come full circle. The AI does not stand for itself. It does not stand by what it writes. It does not stand with those who read it. It does not decide whether a trace should stand. It simply generates.
But human meaning-making is not so indifferent. We stand with our traces, against them, or for their survival. We stand under their weight. And we stand before others, answerable for what we let endure.
To write is to leave a trace. To let it stand is to leave a future.
No comments:
Post a Comment