I’ve shared another article on Substack about how the AI-authorship debate is bollox.
Audio: NotebookLM summary podcast of this topic.
The LLM-versus-human debate has been raging for several years now, and there seem to be two principal camps: human exceptionalists and agnostics. I suppose there may also be a small, misanthropic faction, but its members are either remarkably quiet or indistinguishable from everyone else online. If you are out there, do raise your voice.
If you want to witter on about other negative aspects of AI, have at it, but the authorship debate is weak tea.
I shared another Substack post, but I want to expand it a bit here. On the Substack version, I conspicuously didn’t name certain obvious names, but here, I shall. I also left out a reference to Plato’s Republic.
The current American regime is built on meritocracy, and – this is the part everyone misses – it would not deny the charge. It rather likes the word. It uses the reinterpreted version, the laundered one my other essay was about, in which merit means precisely whatever the sovereign has decided it should mean this morning. Accuse them of meritocracy, and they will thank you for noticing.
Take the man at the centre of it. Whether Donald Trump is meritorious is not, in the end, a question about Donald Trump. It is a question about which ontological grammar you happen to be speaking, because the grammar constitutes the object before a single fact is admitted into evidence. This is the whole Language Insufficiency point in miniature: the frame quietly does the work the facts are later congratulated for.
To Cohort A, he is a charismatic and self-made businessman. The six bankruptcies are not failures but proof of a shrewd operator who knows how to play the game and walk away intact – power-mind, killer instinct, the art of the deal. That he has a decades-long habit of not paying his contractors, and an even longer one of not paying his tax, does not indict him; it endears him. It makes him the everyman who beat the system the rest of us merely resent. A working-class hero, if you can manage it, descending on a golden escalator.
To Cohort B – and I’ll confess my membership, since honesty is cheaper than the pretence of the view from nowhere – he is a less likeable Forrest Gump: a man forever arriving at the centre of history without ever quite grasping how he got there, weaponising a charisma that works only on the already-mesmerised, and rebuilding the world with himself at the origin. He has a certain gravity. It attracts sycophants and rewards them, which is the only merit test his court reliably administers.
Same man. Two grammars. The ‘merit’ resides in neither the man nor the record; it is assigned by the frame and then invoiced to the evidence. Which is exactly the thesis of the parent essay, now wearing a name tag.
The downward stroke, said plainly
The parent essay left the reverse operation as a coy ‘phrase in current circulation’. I’ll say it here without the euphemism: DEI hire. The move is to take career civil servants – people with decades of demonstrable domain competence, whatever their private politics – and broad-brush them as diversity appointments, which is to say, unqualified by definition, arrived by an unauthorised route. The vacancies then fill with those who have pledged fealty and displayed the one indispensable qualification: a flexible integrity. The competent are marked down for being the wrong sort; the pliant are marked up for being biddable. ‘Incompetent’ has been quietly reissued to mean not ours, and ‘qualified’ to mean loyal. Same machine, thrown into reverse – now with a name engraved on the lever.
The cream, and the dimension it rose along
Which brings me, as these things do, to Plato. The Substack piece used his ship of state – the crew who master the mutiny rather than the sea. Here I want the other half of the same complaint: the comfortable conviction that the meritorious rise to the top like cream.
They do rise. That was never in dispute. What the metaphor conceals is the dimension along which they rose. These people are genuinely competent – at working the system, at the acquisition and retention of position – which is a wholly different skill from governing, and precisely the one we so conveniently assume we were measuring. Cream rises; so does scum; the trope declines to specify which, and physics is no help at all.
This is the standing problem of republicanism – and no, before anyone lunges for the comment box, I do not mean the Republican Party, which has its own well-stocked catalogue of failures, as do the Democrats, who would be unwise to feel smug reading this. Nor is any of it uniquely American. If you are somewhere else entirely, enjoying the spectacle from a safe distance, do put the popcorn down. The mechanism by which a society mistakes competence-at-ascending for competence-at-ruling is not a national defect. It is a design flaw in selection itself, and your country installed it too.
I will not, on this occasion, start in on the failings of democracy as a system. Been there; done that. There are only so many sacred cows a man can tip before lunch.
The question, once more
‘Meritocracy’ survives all of this the way it survives everything – by never once being asked the only question that matters: merit at what, judged by whom, to whose benefit? Put it to the present court, and the answer is almost embarrassingly legible. Which is, of course, exactly why the question is never put.
I agree with Paul directionally, but I think he could have gone further – and perhaps he has since – so I compare his view to my own in a Substack article. I’m not sure whether he’d disagree with the extensions. Check it out.
The question is reasonable, although belief may already be doing more work than I would permit it under cross-examination. My writing is generally diagnostic rather than doctrinal. I am more interested in identifying the architectural assumptions beneath a dispute than in selecting a furnished room within it. Where philosophy offers a contest between established positions, I often suspect that the arena itself has been badly designed.
Still, repeated diagnoses eventually disclose a pattern. Certain commitments recur: suspicion of substance, resistance to metaphysical inflation, distrust of linguistic confidence, rejection of moral realism, scepticism toward persistent personal identity, and an insistence that mediation does not entail fabrication.
What follows is therefore neither a creed nor a completed system. It is a provisional map of where I presently stand.
The labels are intended as coordinates, not allegiances. Where no established term quite fits, I have used my own. This is partly unavoidable and partly evidence for the very linguistic difficulty I spend so much time diagnosing.
This content, including video and podcast summary, are of a longer Substack article on my philosophical positions with descriptions.
This is less a confession than an earmark. I came across a conversation with Michael Puett on Chinese philosophy, selfhood, ritual, and the modern Western compulsion to find oneself, and the whole thing caught on an old hook.
Audio: NotebookLM summary podcast of this topic.
Synthèse audio de ce sujet en français dans le podcast NotebookLM.
I am already familiar with some high-level accounts of Asian philosophies, especially Buddhism and the doctrine of anattā (Pali) or anātman (Sanskrit): non-self. That idea has never seemed especially alien to me. If anything, it names with admirable economy a suspicion I already held: that the stable, discoverable, authentic Self is less a hidden metaphysical jewel than a bundle of habits wearing a name badge.
Puett’s point, as I take it, is not merely that the Western self is overblown. That would be too easy, and therefore popular. The sharper claim is that the search for one’s True Self may become dangerous precisely because it mistakes the current arrangement of ruts, reflexes, wounds, preferences, and defensive routines for an essence. ‘This is just who I am’, says the person, embracing not a soul but a sediment. The bad habit acquires a halo. Therapy-speak finds a throne. Another little sovereign subject is crowned in the bathroom mirror.
Against this, Puett presents Chinese philosophy as a tradition of cultivation: not self-discovery, but self-transformation. The point is not to excavate an authentic interior but to train perception, conduct, relation, and response. Ritual, in this frame, is not dead ceremony or social programming. It is a technology of de-naturalisation. It interrupts the little automatisms by which we become legible to ourselves and intolerable to everyone else.
This matters to me because I have long been suspicious of Western narratives and metanarratives, especially those that mistake local grammar for cosmic architecture. The year and a half I lived around Tokyo was instrumental here. I did not leave Japan with a doctrine, thankfully. Nothing so vulgar. But it loosened something. I had already begun to distrust the sovereign Western subject, the metaphysical executive behind the eyes, the heroic interiority forever pretending to be origin rather than residue. What I lacked was a broader vocabulary. A better frame.
As often happens, I later encountered thinkers who had already articulated versions of what I had been circling, usually with more patience, precision, and elegance. Such is the scholar’s reward: to discover that one’s private revelation has an older bibliography.
What struck me in Puett’s discussion was the movement from identity to practice. Religion, selfhood, reading, even attention itself: these need not be treated as badges to defend, but as disciplines to inhabit. This is a useful corrective to the modern anxiety of proof. The self must prove itself. The worker must prove worth. The believer must prove identity. The thinker must prove originality. Modern life becomes a Calvinist treadmill with better branding and worse lighting.
The alternative is not passivity. Nor is it the wet little slogan that everything is socially constructed, therefore nothing matters, therefore let us all dissolve into discourse and invoices. The alternative is practice: deliberate, repeated, embodied interruption. Ritual as a way of making habits visible. Reading as a ritual of humility. Attention as a cultivated stance. Relation as something made, not merely felt.
This is where Chinese philosophy may become especially useful for my own work. It offers neither the Western fantasy of the fixed self nor the lazy inversion of pure flux. It gives us formation, pattern, training, relation, and transformation. Not the Self as essence. Not the self as fiction. The self as practised arrangement. That, at least, is worth returning to.
I am busy, always busy, because apparently exhaustion has been mistaken for virtue by a civilisation that should know better by now. But I want to earmark this thread for further consideration: Chinese philosophy, ritual, anti-authenticity, non-self, practice over identity, and the possibility that what we call ‘who I am’ is often only what has not yet been interrupted.
On Heidegger, and the rook that was always, secretly, a bishop
Imagine, if you will, that you are playing chess with a grandmaster. Your opponent fingers his rook, slides it diagonally across the board, captures your queen and lifts the piece away. Nonplussed, you question the tactic. He defends it with vehemence.
For most of my posts, I also share a summary processed by Google’s NotebookLM. Most of these summaries are decent enough – some are even excellent inasmuch as they shed a new light on the idea. Other times, they miss the point, are well off the mark, or inject concepts or translations, not only unintended, but markedly opposed to the point being made.
My recent Chess with Heidegger post is illustrative. Both the podcast and the video make errors – and I don’t mean the visual ones that occur as well.
As a page footer, I mention that I use LLMs as part of my workflow. This may be anywhere from 0–99%. The podcast asserted that I used an LLM as a significant part of this project. This is incorrect. Specifically, I didn’t feed Heidegger’s text into an LLM and ask it to digest and regurgitate it, parsing out some unlocked wisdom. Obviously, I used various LLMs for supporting content.
The video misrepresented my point about Wittgenstein, directionally off. He expressly does not support the specificity of language. My reference to him rolling over in his grave was triggered by the notion that language is determined by use – and language games.
Am I the only one who can’t resist a massive eyeroll – and, let’s be honest: jaw-drop – what you hear transhumanism couched as evolution? To me, it incites a similar reaction to hearing people witter on about machine consciousness, but I’ll sideline that topic.
My objection is linguistic: transhumanism often borrows the prestige of evolution to describe what is more precisely technological mediation. The fact that a device is worn, implanted, or integrated into a body does not by itself move it from tool-use into biological descent. The offspring still inherits the organism, not the upgrade. Technology is not heritable.
Audio: NotebookLM summary podcast of this topic.
Consider rhinoplasty. Rhinoplasty changes the presented phenotype, not the inherited genotype. The child inherits the developmental instructions, not the parent’s post-surgical edit. Likewise, a neural implant, prosthetic limb, exoskeleton, gene-unrelated enhancement, or titanium jaw of techno-vanity may alter the lived organism, but it does not thereby alter the reproductive line. This is the category error: Acquired modification is mistaken for inherited transformation.
So, transhumanism often confuses the edited encounter-profile of an organism with the evolutionary alteration of the organismic lineage. The rhinoplasty case is good because it shows the absurdity without needing much apparatus. No one sane thinks a nose job rewrites the germline. Yet when the modification is sufficiently glamorous, especially when welded to futurist rhetoric and venture-capital incense, people suddenly start talking as if augmentation equals evolution.
A prosthesis is to evolution what rhinoplasty is to heredity: a modification of presentation, function, or encounter, not a transformation of descent. The confusion arises when the altered individual is mistaken for an altered lineage.
The question isn’t: Can transhumanism be defended under some broader theory of evolution, cultural inheritance, niche construction, or techno-biological co-development? Sure, it can.
But my claim is narrower: The term evolution is being made to do illicit rhetorical work when acquired, engineered, prosthetic, or embedded modifications are spoken of as though they were equivalent to heritable biological transformation.
The problem is that we are sliding from tool-use to augmentation to enhancement to transformation to evolution. This move isn’t illegal, but it should attract your attention.
At each step, something plausible is smuggled into something grander. By the end, an implanted device or technical prosthesis has been rhetorically promoted into ‘the next stage of human evolution’, when what has actually happened is that an organism’s encounter-capacities have been modified by artefactual mediation.
In the end, I’m not claiming that transhumanist practices can’t affect selection environments, social organisation, embodiment, or cultural transmission. What I am claiming is that calling such practices evolution without specifying the relevant register is a linguistic inflation. It trades on the authority of biological evolution while often describing technological mediation, individual augmentation, or cultural change.
In my book, more accurate terms might be augmentation, prosthesis, mediation, tool evolution, cultural inheritance, or niche modification, but these re so glamorous or click-baity. I might as well wear a hat and call myself post-human.
I chatted up ChatGPT, so I am adding this response:
Exactly. The carried / worn / embedded distinction is mostly a boundary fetish, the sort of thing humans adore because skin looks like an ontological border and apparently everyone remains impressed by packaging.
An axe, spectacles, a pacemaker, a cochlear implant, a smartphone, and a neural implant all sit on the same broad continuum: external mediation of capacity. Some are held. Some are worn. Some are surgically installed. But unless the alteration becomes heritable, it is not biological evolution in the strict sense. It is phenotypic augmentation, affordance expansion, or environmental coupling.
The useful distinction is threefold:
Biological evolution Heritable variation changes across generations. The organism-line changes.
Cultural / technological evolution Tools, techniques, designs, institutions, and practices vary, are selected, retained, modified, and transmitted. Axe technology evolves. Writing evolves. Surgery evolves. The appalling human habit of calling every interface “intuitive” also evolves, though downward.
Individual augmentation A particular body-system gains capacities through artefacts. This may change what the organism can do, but it does not by itself alter the inherited biological line.
Your axe example is clean because it exposes the smuggling. The axe helped alter human selection environments. It extended reach, force, food access, defence, craft, and perhaps social organisation. That may feed into gene-culture co-evolution over long scales. But the axe itself is not an evolved human trait. It is a technological mediator that participates in the evolutionary ecology.
The same holds for transhumanist implants. A retinal implant may alter one person’s perceptual affordance field. A brain-computer interface may alter one person’s control loop. A prosthetic limb may alter one person’s mobility. But unless such modifications become transmissible across generations, biologically or through some stable reproductive architecture, calling them “evolution” is loose metaphor masquerading as profundity. Very on-brand for techno-utopianism: rename engineering as destiny and hope the grant reviewers are sleepy.
From your MEOW angle, the better formulation is:
The relevant transformation is not evolutionary but mediational. Tools, prostheses, implants, and interfaces reconfigure encounter-events by altering the affordance structure through which an organism meets constraint. Whether the device is carried in the hand, worn on the face, implanted beneath the skin, or networked through the environment is secondary. The organism-world relation changes; the inherited organism does not necessarily change.
That also lets you avoid an overcorrection. Transhumanist enhancement is not irrelevant. It may be historically enormous. But its primary register is encounter, not inheritance; mediation, not mutation; technical lineage, not biological descent.
So the blunt line is:
Transhumanism mistakes the evolution of tools for the evolution of the tool-user.
Not always, not necessarily, but often enough that the slogan deserves being dragged into the alley and relieved of its metaphysical wallet.
Man, this IaI piece asking about The Patriarchy in Question has got my knickers properly twisted. As I gather the scattered crockery of my thoughts, the first issue is the Sorites problem of patriarchy.
Audio:NotebookLM summary podcast of this topic.
Mirroring the old question of when a collection of grains becomes a heap: when, exactly, does a society become a patriarchy? How much concentration of patriarchal residue is required before the noun is earned? Is one bad apple enough to spoil the whole lot, or does that give us only the faint homoeopathic aftertaste of patriarchy?
I doubt many people would openly defend this homoeopathic definition. One sexist custom, one male-coded institution, one inherited assumption, and behold: The Patriarchy. But if not that, where’s the threshold? Fifty per cent? Ninety? Thirty? Or is the question itself badly formed?
The issue isn’t only composition but degree, location, and power. One king over a kingdom gives us monarchy; it becomes patriarchal when rule is authorised through masculine-coded inheritance, legitimacy, property, office, or paternal command. But what of a queen operating under the same institutional grammar? Has the patriarchy been interrupted, or merely furnished with a woman at the apex? If she inherits the language, offices, succession rules, and symbolic architecture of patriarchal power, then the body on the throne may change while the grammar of rule remains intact.
The law has a charming habit of behaving as though language becomes precise the moment someone in a robe frowns at it. Words that drift cheerfully in ordinary life are summoned into court, sworn in, interrogated under oath, and expected to produce stable meaning under institutional pressure. When they fail, as they reliably do, the system does not conclude that language may be structurally insufficient for the task. It consults another authority. A dictionary. A drafting manual. A corpus database. A professor, if civilisation has really run out of excuses. Whatever. Any port in a storm. Then it calls the result interpretation, and everyone pretends the word was waiting there all along.
Audio: NotebookLM summary podcast of this topic.
Watch the video below. It is an admirably clean illustration of exactly this.
What you just watched is not merely a curiosity about punctuation and gun laws. It is a diagnostic. And if you have read Chapter Five of A Language Insufficiency Hypothesis, you will recognise the pathology immediately.
The Repair Cascade
The video gives you three cases, each one a new rung on the same ladder of failed repair.
tl;dr? They tend to make it up as they go to serve their power needs.
First, Muscarello v. United States (1998): A man transports a handgun in a locked glove compartment whilst conducting a drug transaction. The statute punishes anyone who ‘uses or carries a firearm’ during such a crime. The question is whether ‘carry’ includes a weapon stored in a vehicle. The Supreme Court reaches for the OED, finds that the earliest documented sense of carry includes conveyance by vehicle, and sends Muscarello to prison, where he eventually dies. Convenient etymology. Regrettable outcome.
The video notes – correctly, I might add (and so do) – that this is an instance of what linguists call the sense-ranking fallacy: assuming that the first definition listed is the primary one, rather than simply the earliest documented. The OED’s ordering is historical, not hierarchical. Why a US court chose the OED is a sign of refinement yet remarkably curious for an American institution.
Second, the Oakhurst Dairy case: Maine truck drivers sue for $10 million over a missing Oxford comma in a statutory overtime exemption. Both sides marshal gerunds, asyndeton, the Chicago Manual of Style, and the Maine legislative drafting manual, which explicitly prohibits the Oxford comma – making the ambiguity, in a sense, officially mandated. The case settles without a definitive ruling. The language did not yield a winner; the lawyers did. The hole wasn’t filled, but their pockets were.
Third, and most instructive, corpus linguistics arrives as the shiny new repair tool. Rather than trusting dictionaries, courts can now search large databases of actual language use to establish ‘ordinary meaning’. Progress. Empiricism. Science, even. And then, almost immediately, the next failure mode surfaces: the frequency fallacy (common usage is not the only permissible usage), corpus skew (many databases over-represent news articles), and search-framing (the ‘sanitation’ / ‘sanitise’ mask mandate case, where including a related but non-synonymous word shaped the results before analysis had even begun). The supposedly empirical tool inherited the user’s prior interpretive frame. Extraordinary.
Follow the sad path of the sad panda: ordinary meaning fails → dictionaries → dictionaries fail → corpus linguistics → corpus linguistics fails → methodological dispute about whether judges should be conducting quasi-scientific research from the bench at all.
And so it goes…
Judge Humpty Dumpty: Guilty as Charged (Sorry. No 8-bit video game music. My bad.)
The LIH Reading
In A Language Insufficiency Hypothesis, Chapter Five argues that law is not a domain that occasionally encounters linguistic difficulty. It is a domain that is constitutively dependent on terms that live in the Contestables zone of the Effectiveness–Complexity Gradient – words like reasonable, fair, cruel, due process – terms indispensable to legal order and perpetually unstable within it. The Gradient’s prediction is blunt: the further a term drifts from stable, concrete reference, the more its meaning must be imposed by authority rather than established by usage.
The video illustrates this at the level of what might seem to be relatively simple terms – carry, distribution, sanitation – words that appear to sit closer to the Invariants end of the scale than to the Contestables. And yet even here, the institutional machinery creaks. If ‘carry’ cannot carry the weight of a single statute without Supreme Court intervention and a man’s death, what prospect does ‘reasonable’ have? Or ‘fair’? Or ‘obscene’?
Potter Stewart, as Chapter Five recounts, admitted in Jacobellis v. Ohio (1964) that he could not define obscenity in the abstract. ‘I know it when I see it‘, he declared. The remark is famous for its candour. It is less often noted that it is also an admission that language had simply given up, and that institutional authority stepped in to do what definition could not. The Court didn’t clarify what obscenity means, but it asserted the power to punish it anyway as it might later decide.
The video’s repair cascade is the same mechanism operating at a more mundane level. Legal interpretation doesn’t overcome linguistic insufficiency. It proceduralises it. Each interpretive tool displaces the instability onto a new surface. Dictionaries relocate the problem from statutory language to lexical authority. Corpus linguistics relocates it from lexical authority to sampling, frequency, and search design. The crack isn’t closed. It’s moved, with considerable administrative ceremony, and the ceremony is called clarity – clear as mud.
The law, in short, functions less as a dictionary than as a sovereign Humpty Dumpty: it decides what words mean when it matters, and enforces those meanings until it decides otherwise. The gavel is doing the work the lexicon cannot.
NotebookLM Infographic on this topic.
The Lesson That Isn’t
The lesson here isn’t that dictionaries are useless, corpus linguistics fraudulent, or judges uniquely obtuse. The lesson is structurally worse than that. Each repair works locally and fails architecturally. The law can stabilise meaning long enough to act, and acting is not nothing – Muscarello’s conviction required a determinate reading of ‘carry’, and the system produced one. But it can’t transmute contested language into invariant reference. It can only decide, punish, and maintain the fiction that the word was always waiting there, meaning exactly that.
Textualism – the interpretive philosophy that instructs judges to attend only to the words on the page, nothing more – is, viewed through an LIH lens, an institutionalised form of the Presumption of Effectiveness. It treats language as though it has a singular, determinate meaning recoverable by sufficiently rigorous attention, rather than as a system whose instability is structural rather than incidental. The words on the page are not a fixed source. They are the site of the problem.
If this framing resonates, Chapter Five of A Language Insufficiency Hypothesis develops the full argument – from the Gradient’s account of why legal language is structurally dependent on Contestables, through Potter Stewart’s famous abdication, to the Humpty Dumpty jurisprudence that inevitably follows. Available in paperback and hardcover from Philosophics Press.
I just read Chapter 4 of Frantz Fanon’s Black Skin, White Masks, and it has similar problems I’ve also critiqued for Judith Butler’s Gender Trouble: Feminism and the Subversion of Identity and Simone de Beauvoir’s The Second Sex. In all three cases, I accept the primary argument. What I reject is psychology, especially psychoanalysis, as a legitimate form of scientific inquiry.
Audio: NotebookLM summary podcast of this topic.
My issue isn’t that Fanon, Beauvoir, or Butler fail in their central diagnoses. I broadly accept their claims. My issue is that the psychoanalytic material typically functions like a grinding side quest: time-consuming, rhetorically elaborate, and only weakly connected to the main argumentative progression. It may enrich the atmosphere, but it doesn’t materially alter the outcome. Once one doesn’t accept psychoanalytical psychobabble as a valid evidentiary lens, the material becomes a time sink. Not only can’t I get my time back, but I also expend even more time here, railing on.
Speaking of distraction: evidently, WordPress has added a new blog-to-podcast feature, so I tried it out here. Whatevs.
Fanon’s central claim about colonial racialisation doesn’t require dream interpretation – the dreams themselves are seemingly apocryphal at the start. The stronger route is through embodied recognition, imposed category structures, conceptual nomenclature, and the racialised field of encounter. The dream material reads as psychoanalytic side-content: thematically adjacent, occasionally vivid, but methodologically low-yield. It doesn’t deepen the case so much as delay it. The entire time, I am thinking to myself, ‘Where is he going with all this?’ and ‘Are we there yet?’ only to get dropped off just where I had started – a round trip to nowhere.
But Fanon’s mistake isn’t necessarily insincerity. It’s an over-trust in a psychological lens that converts metaphor into method. The psychoanalytic examples may have seemed to him like evidence; to a reader sceptical of psychoanalysis, as I am, they register as rhetorical illustrations. Once a reader withholds confidence in this method, the chapter’s supporting material becomes distracting rather than strengthening.
Enfin, psychoanalysis too often behaves like a prestige tarot deck for the academically credentialled: it turns ambiguity into confirmation, opacity into symbolism, and resistance into further evidence. Fanon’s broader account of colonial alienation survives because it doesn’t depend on this machinery. The dreams aren’t necessary to the argument; they’re decorative scaffolding around a structure that’s stronger without them.
Philosophics 