There’s a peculiar humiliation in being told what you meant. Not what you said — that, at least, would be a matter of record. What you meant. The interior intention. The hidden agenda that the critic, having read perhaps half of your argument, has divined with a confidence unavailable to you regarding your own mind.
I have been writing for the better part of forty years – since I stopped composing music and started using words. In that time, something has shifted—not in how carefully I write, but in how reliably I am misread. And I am fascinated by that shift not as personal grievance but as diagnostic data. Because what’s happening to my work is happening to public discourse at large, and the implications run considerably deeper than anyone’s bruised authorial ego, including mine.
The Death of the Text Was Announced Early
Roland Barthes declared the death of the author in 1967. What he meant — and the irony of paraphrasing Barthes on misreading is not lost on me — that is, the moment a text is released into the world, the author’s intention becomes one reading among many and not necessarily the most privileged one. The text belongs to whoever picks it up. Meaning is not transmitted directly from writer to reader like a parcel; it’s reconstructed at the point of reception out of everything the reader carries: their history, their wounds, their prior commitments, the last thing they read before this, and the argument they are already having with someone else entirely.
This is the condition of language. And for most of literary history it remained a more or less manageable condition, because the contexts in which serious writing was received were, however imperfectly, structured. There were editors, institutions, critical traditions, and the slow friction of considered response. Meaning had time to settle before it was contested.
What has changed is that the friction has gone. Text now travels stripped of context, landing in feeds engineered to produce the most energetic possible response. The reader’s prior commitments are not incidental to this — they are the raw material the algorithm is working with.
Shards of Truth Are More Dangerous Than Lies
Here is what I have noticed over years of writing on matters that people would prefer to remain undisturbed. Wild assertions provoke very little. Conspiracy theories, extreme positions, unmoored speculation — these are easy to dismiss, and people dismiss them. What detonates is a shard of truth that lands inside something already decided.
The philosopher’s term for what follows is motivated reasoning, but that phrase is too clinical to capture what actually happens. It’s more visceral than reasoning. It’s the experience of recognition — not of the argument, but of the threat. The reader does not think, 'This challenges my position; let me engage with it.’ Oh no. The reader feels: this is aimed at me, and reaches for the nearest available weapon, which is usually a mischaracterisation of what was just said.
What fascinates me is the structure of the mischaracterisation. It is almost never random. The position the critic constructs — the one they argue against with such conviction — is consistently more extreme, more reckless, and more easily dismissed than the one actually on the page. This is not coincidence. The mind under threat doesn’t engage with the argument it received. It engages with the argument it can defeat. And so it builds a case.
The argument they are having, in other words, is with themselves. The text is just the occasion. The writer — if they are paying attention — are watching a private drama in which they have been cast, without audition, as the villain.
The Inference Machine
Beyond motivated reasoning lies a more subtle issue: the ordinary, non-defensive production of meaning that readers can’t help but perform.
Every act of reading is also an act of inference. The reader fills in the gaps. They import context. They resolve ambiguity according to their existing model of the writer — who this person is, what they usually argue, and which camp they belong to. If that model is wrong, or incomplete, or built from a single prior encounter, then the inference machine runs perfectly well and produces the wrong output.
I use AI to check my work precisely because of this. Not to replace judgement but to stress-test it — to identify the gaps a hostile or careless reader might fill in ways I didn’t intend, to catch claims that are accurate but could be weaponised by selective quotation, and to remove anything that functions as a dog whistle I didn’t mean to blow. It is, in effect, an attempt to pre-empt the inference machine by modelling it in advance.
It helps. It doesn’t solve the problem. Because the inference machine is not running on the text. It’s running on the reader. Alas, the reader I cannot edit.
What Mass Misreading Costs
If this were only a problem for writers with sensitive dispositions, it would not warrant sustained attention. But mass misreading — the systematic construction of positions that were never held, followed by their systematic demolition — is one of the defining features of contemporary public discourse, and its costs are not trivial.
The first cost is epistemic. When the response to an argument is almost always the argument’s caricature, genuine disagreement becomes structurally impossible. You cannot update your position in response to a counter-argument that was never actually made. The feedback loop that keeps thinking honest – the loop that runs: here is a position, here is a challenge to it, here is a refined position – is broken. Everyone gets louder. No one gets wiser.
The second cost is social. The inference machine does not just produce wrong readings of texts. It produces wrong readings of people. The writer who argues X becomes, in the reader’s mind, someone who believes Y and Z and probably also holds views that were never expressed and would be denied if raised. The person disappears into the position attributed to them. And the position attributed to them is the one most useful to the argument the reader was already having.
This is not a new problem. But the scale of it is new. The speed is new. And the architecture that incentivises the most energetic misreading — that rewards the confident attribution of motive and the viral demolition of the position that was never held — is new, and it’s operating at a civilisational scale.
What the Noise Tells Us
I said at the beginning that the response to one’s work is data about the reader, not the argument. Let me be more precise about what that data reveals.
When a carefully evidenced, qualified, scrupulously fair argument is met with outrage at a position it never took, what you are seeing is the form of an open wound. The misreading maps the reader’s vulnerabilities — the places where their commitments are so load-bearing that they can’t afford to have them questioned, even gently, even by implication. The more violent the misreading, the more structural the vulnerability.
This is not a basis for contempt. It’s a basis for understanding. The civilisational crises we are navigating – ecological, democratic, economic, and epistemic – are producing people whose capacity for genuine encounter with challenging ideas has been worn down to almost nothing. The defensive misreading is not stupidity. It is exhaustion. It’s the mind protecting itself against one more thing that threatens the scaffolding.
This is why I don’t respond to personal taunts or rudeness. Not out of indifference, and not because I haven’t noticed. But because the taunt is never really about the argument — it’s the pain a wound feels when it’s pressed. Engaging with it as though it were a reasoned position would be a category error, and a condescending one at that. The only honest response to someone’s anguish, when it arrives dressed as an insult, is to decline the invitation to fight and keep writing.
The question for those of us who write into this environment is not how to be better understood. That is mostly beyond our control. The question is what we owe the reader who is, beneath the projection and the noise, genuinely trying to make sense of a world that has become very difficult to read.
My answer — provisional, revisable, offered without great confidence — is this: precision, without condescension. Difficulty, without obscurantism. The refusal to simplify what is genuinely complex, held alongside the equal refusal to make complexity a barrier to entry. And the long, unglamorous patience of someone who understands that meaning is not transmitted but grown — slowly, imperfectly, across the distance between one mind and another.
That distance is not a failure of communication. It’s what communication is.