Through McLuhan’s Lens

Governing What We Can’t Name

August 16, 2026 | 2690 words

Through McLuhan’s Lens: Governing What We Can’t Name

There is a strange arithmetic running underneath this week’s AI conversation, and once you see it, you cannot unsee it. The word regulation and its close relatives — rules, oversight, governance, compliance — saturate the discourse. They appear with the density of a downpour. Everyone, it seems, is talking about how to control the thing. But the words that would tell us what the thing actually is have gone missing. Partner barely registers. Transformation is a whisper. Machine learning — the plain technical name for how these systems work — is nearly absent from the very debate about how to govern them.

Sit with that gap for a moment. We are drafting rules at high volume for a technology whose defining vocabulary has quietly drained out of the room. The regulatory language is loud. The descriptive language is a vacuum. This essay is about that vacuum — not because it is a small oversight to be corrected, but because it is doing something. The empty space where the technology’s real name should be is not neutral. It is shaping what an entire public can perceive about what is being decided on its behalf.

The Debate You Are Watching, and the One You Are Not

Marshall McLuhan spent a career on a single, deceptively simple claim: the medium is the message. In plain terms, it means that any technology or communication form reshapes us more powerfully through its shape than through its content. What a thing carries matters less than what the thing does to us by existing. A television does not merely deliver programs; it retrains how we pay attention. The message of the medium is the change of scale and pace and pattern it introduces into human affairs — a claim he laid out in Understanding Media.

Apply that instrument to this week’s phenomenon and the picture inverts. The public is watching the content of the regulatory debate: Should there be rules? How strict? Enforced by whom? These are the questions in every headline. But the debate is also a medium. It has a form. And that form is delivering its own message, silently, beneath the arguments — a message no single proposal states out loud.

The message is this: AI is a passive object. A slab. An inert thing that sits still while we decide what to do with it. You can hear this in the grammar of the discourse. We speak of AI the way we speak of a chemical to be labeled or a road to be zoned. The technology is the direct object of every sentence — the thing acted upon, never the thing that acts. And when the descriptive words that would complicate this picture are absent — partner, which implies something that acts alongside us; transformation, which implies something that changes us in return — the grammar hardens into a worldview.

This is the figure-and-ground move at the center of McLuhan’s method. The figure is what you stare at: the loud, foreground argument about rules. The ground is the background that makes the figure legible but goes unnoticed: the assumption, baked into the vocabulary, that AI is a manageable object. Everyone debates the figure. Almost no one sees the ground. And the ground, as McLuhan understood, is where the real work gets done.

The Rear-View Mirror

Why has the descriptive vocabulary gone missing? Part of the answer is a habit McLuhan named the rear-view mirror — our tendency to understand each new technology through the language of the one before it, driving into the future while staring at what is already behind us.

We do this reflexively. When cars arrived, we called them horseless carriages. When film arrived, we pointed a camera at a stage play. The new thing is frightening precisely because it has no name yet, so we reach for the nearest old one. The old name is a comfort. It is also a distortion. It shows us the present dressed in the clothing of the past.

The regulatory conversation is a rear-view mirror aimed at AI. Its governing vocabulary — rules, oversight, compliance — comes from an older world of regulating products, industries, and machines that do exactly what they were built to do and nothing more. That vocabulary works beautifully for a toaster or a toll road. It presumes a stable object with fixed properties, sitting in place, waiting to be measured and bounded.

But the words that would describe what is actually novel about this technology — that it learns, that it changes, that it acts with us and on us — are the words that have gone quiet. Machine learning, the term for the one thing that makes this technology genuinely new, is nearly absent from the debate about how to govern it. That absence is the rear-view mirror at work. We have described AI using only the parts of it that resemble older technology, and we have gone silent about the parts that don’t. The result is a governance conversation that is confident precisely because it refuses to look at what it cannot recognize.

McLuhan warned that this is not a minor lag. In The Medium Is the Massage he argued that societies routinely march into the future while facing backward, and that the cost is not merely awkwardness but blindness — we cannot govern well what we insist on mistaking for something else. The rear-view mirror does not just delay understanding. It substitutes a familiar false image for an unfamiliar true one, and then we legislate the false image.

The Comfort of Numbness

There is a second McLuhan concept that explains not just why the vocabulary is missing but why its absence feels fine. He called it technological numbness, sometimes auto-amputation — the idea that when a technology overwhelms us, we go numb to it as a defense, the way the body shields itself from a shock it cannot process.

Every powerful new medium extends some human capacity and, in doing so, amputates or numbs another. The stirrup extended the leg and numbed the body to the violence of the mounted charge. Print extended the eye and numbed the ear. The response to being extended and overwhelmed is not clear-eyed analysis. It is anesthesia. We stop feeling the thing that is acting on us.

Flattening AI into a “neutral blank” is exactly this kind of anesthesia performed with language. To call the technology a partner is to feel it — to admit that it acts, that it changes the terms of the relationship, that we are being extended and amputated at once. That feeling is uncomfortable. It is hard. So the discourse reaches, understandably, for the numb option: AI as inert object, a slab that does nothing until we permit it to. The neutral framing is not a failure of intelligence. It is a nervous system protecting itself from a shock it does not want to feel.

McLuhan’s darker observation, throughout the McLuhan Interview, was that numbness is precisely the state in which a technology does its deepest work on us — unnoticed, unresisted, unfelt. We are most thoroughly reshaped by the medium we have gone numb to. The public is being invited, by the very shape of the regulatory conversation, to go numb to what AI actually is. And numbness, however comfortable, is the opposite of the understanding a citizen needs.

The Turn: The Vacuum Is Doing a Job

Here is where the analysis has to stop describing the gap and start asking what the gap is for. Because the temptation is to treat the missing vocabulary as an accident — a slip, an oversight, a thing we forgot to say and could easily add back. That reading is too kind, and it misses the point entirely.

The naming vacuum is not a hole in the discourse. It is a load-bearing wall.

Consider what regulation, as an activity, actually requires. To regulate something, you must first hold it still. You must treat it as a bounded object with fixed edges — a thing that stays put long enough to be measured, categorized, and fenced. Regulation is an act performed on objects. It cannot get purchase on anything else.

Now consider the words that have gone missing. You cannot regulate a partner. A partner is a relationship, and you do not write a rulebook for a relationship — you negotiate it, you renegotiate it, you live inside it and it changes you. You cannot regulate a transformation. A transformation is a process that includes you as one of its products; by the time you have the rules drafted, the thing you meant to govern has already changed you into someone who needs different rules. And machine learning names precisely the quality — the capacity to change, to move, to become something its makers did not specify — that makes the whole object-model false.

So watch what the vacuum accomplishes. By evacuating exactly these words — the words for relationship, for process, for change — the discourse manufactures the one thing that makes confident regulation possible: a still object. The flattening of AI into a neutral blank is not a bug in the debate. It is the precondition for the debate. The vacuum is productive. It produces the very slab that the rulebook then claims to govern.

This is the revelation McLuhan’s framework surfaces, and it is worth stating without hedging. We did not fail to name AI and then decide to regulate it anyway. The failure to name it is what lets the regulation proceed. Strip the technology of every word that suggests it acts, changes, or partners, and you are left with an inert thing — and inert things are exactly what governance knows how to touch. The silence is not preceding the rules. The silence is enabling them.

Who Benefits from a Blank Slab

If the vacuum is doing a job, the next question is McLuhan’s most uncomfortable one, and the one a pro-reader analysis is obligated to ask: who benefits?

A vocabulary that makes a technology look neutral is never neutral in its effects. It serves whoever wants to control the terms on which the technology is understood — and understanding is upstream of everything. When AI is framed as an inert slab awaiting rules, several things become true at once, all of them convenient for power and none of them convenient for the public.

First, the slab-frame makes the technology look finished. A slab is a completed object. It does not learn while you are looking away. This is a gift to anyone who would rather you not notice that these systems change, that their behavior drifts, that what was safe on Tuesday may not be safe by Friday. Govern it as a static thing and the ongoing, moving, learning reality slips out of the frame entirely.

Second, the slab-frame locates all the agency in the regulators and the makers, and none in the technology or its relationship with the public. If AI merely sits and waits, then the only actors in the story are the powerful ones deciding what to permit. The citizen becomes a spectator to a negotiation between two sets of experts over an object that supposedly does nothing on its own. The relationship you actually have with the technology — the way it is already extending and amputating your own capacities — is defined out of existence.

Third, and most subtly, the slab-frame makes the public’s confusion look like the public’s fault. If the thing is a simple neutral object, then anyone who cannot follow the governance debate must simply be uninformed. But the confusion is not the reader’s failure. It is the product of a discourse that removed the words needed to understand what is being discussed. You cannot follow a conversation about a thing that has been stripped of its name.

McLuhan’s insistence in Understanding Media was that control of a medium’s ground — the unexamined background assumptions — is a deeper form of power than control of its content. Whoever gets to decide that AI is “a neutral blank” has already won the important argument before the visible argument about rules even begins. The debate about how to regulate is loud and public and contested. The prior decision that AI is a slab to be regulated was made quietly, in the vocabulary, with no vote taken.

What the Reader Loses, and What Insisting Costs

So what does a general reader — a citizen, a voter, someone simply trying to understand what is being decided in their name — actually lose when a technology is governed under a name that hides what it is?

You lose the ability to hold the right people accountable for the right things. If the public perceives AI as an inert object, then when the technology does something the object-model said it could not — when it acts, drifts, or reshapes a practice nobody chose to reshape — there is no vocabulary in which to name the failure. The rulebook was written for a slab. The slab did something a slab cannot do. And the discourse has no words to describe what just happened, so it will reach, once again, for the rear-view mirror and call it something old and familiar.

You lose, more fundamentally, your standing as a participant. A partner is something you negotiate with; a transformation is something you have a stake in. Both of those words return agency to you. The slab-frame takes it away. It casts you as a bystander to an object being managed by others, when the truth — the truth the missing vocabulary would tell you — is that you are already inside a relationship with this technology that is changing what you can do, what you can perceive, and what you are becoming. To be told you are watching an object being regulated, when you are in fact being transformed by a partner, is to be quietly disqualified from your own life.

What would it change to insist on the missing vocabulary before the rules harden? Not a policy. The point here is not to swap one rulebook for another. The point is perception, which McLuhan always insisted comes first. If the public reinstated the words — if partner, transformation, and machine learning re-entered the conversation with the same volume as regulation — the whole debate would tilt on its axis. You cannot ask “what rules should this slab follow?” once you have admitted the thing is not a slab. You would have to ask harder, better questions. What is this changing us into? What relationship are we actually in? What is being extended, and what is being quietly amputated while we go numb?

Those questions do not fit neatly into a governance framework, which is precisely why the vocabulary that raises them went missing. But they are the questions a citizen deserves to be asking. Understanding beats being managed. A public that can name what a technology is cannot be governed around — it can only be governed with.

McLuhan’s whole method was a training in noticing the ground while everyone else stares at the figure. The figure this week is the regulation debate: loud, urgent, contested, everywhere. The ground is the silence around it — the empty space where the words partner, transformation, and machine learning should be, and are not. That silence is not the absence of an argument. It is the argument, already settled, before you were invited to speak.

You can watch the foreground for the rest of the year and learn nothing that matters. Or you can do the one thing the discourse is built to prevent: look at the space where the name should be. Once you have seen that vacuum — once you notice how confidently we are governing a thing we have refused to describe — you will not be able to watch a regulation debate the same way again. You will hear the missing words. And hearing them is the beginning of getting your perception back.