They kept him in a basement on Queen's Park Crescent. Not the Centre for Culture and Technology, which had moved twice since his death, but a sub-basement beneath the old Coach House that nobody used for anything important. A liquid nitrogen dewar the colour of a municipal bin. No notes.
The arrangement had been made in the spring of 1979, six months before the stroke took his speech. The professor had signed the papers himself, in the office of a cryonics society operating out of a dentist's building in Mississauga. What a Fool Believes drifted from the radio. His colleagues at the Centre later said he must have done it as a joke. But he'd paid the deposit. And when the stroke came in September, and the aphasia settled in like weather, and it became clear that the man who had theorised communication could no longer communicate, someone honoured the contract.
He died on New Year's Eve, 1980. By the second of January, he was in the dewar. It was the first time anyone had done as he asked.
His written instruction, in his desk drawer, was a single sentence: Defrost when the medium catches up with the message.
Forty-six years is a long time to wait for a punchline.
The decision to thaw him was made by a committee, and he would have expected that. Committees are the immune system of institutions: they exist to prevent anything from happening fast enough to matter. This committee, though, had a problem it could not solve by doing nothing.
Artificial intelligence had arrived and nobody could agree on what it was.
Silicon Valley said tool. Regulators said threat. The press said revolution, or apocalypse, depending on the week. Philosophers said it depended on your definition of intelligence. Marketers said it depended on your use case. Computer scientists said it was just statistics. Artists said it was theft. Nobody was listening to anybody else.
The committee voted six to three to bring McLuhan to room temperature.
He opened his eyes on a Tuesday in September. The room was a seminar room on the fourth floor: whiteboard, projector, institutional carpet in the colour of compromise. Three academics, a postdoctoral fellow, and a dean. They had prepared a briefing document. Thirty-one pages.
McLuhan sat in a plastic chair and looked at the fluorescent lights. He did not read the document. He had the look of a man who knew how this ended.
"Tell me what it does," he began.
The postdoctoral fellow began. She was good at this. She explained large language models, neural networks, training data, generative AI. She mentioned parameters. She mentioned benchmarks. She used the word emergent twice.
McLuhan listened for four minutes. Then he held up a hand.
"You are describing the content," he said. "I didn't ask what it carries. I asked what it does."
The room went quiet.
"Does it replace something?"
The dean said it was more of an augmentation.
"Everything replaces something. The railway didn't add to the horse. What does it replace?"
The postdoctoral fellow said knowledge work. Research, analysis, writing, translation. Tasks that used to require a person who had spent years acquiring specialist expertise.
McLuhan nodded slowly. He asked for a marker and stood up. The whiteboard was covered in a grid from someone's earlier class. He erased it with the side of his fist and drew four quadrants.
He wrote ENHANCES in the top left.
"Every medium extends some human faculty. The wheel extends the foot. The book extends the eye. Clothing extends the skin. Electric circuitry extends the central nervous system." He tapped the board. "This thing you have built extends cognition itself. Pattern, synthesis, language at whatever scale you require. What took a team of specialists a season takes a single question. You have extended the mind the way the telescope extended the eye, except there is no direction you must point it."
He wrote OBSOLESCES in the top right.
"And as with every extension, there is an amputation. The wheel extended the foot and man stopped walking. The book extended memory and man stopped remembering." He paused. "Your machine extends specialist knowledge. The premium on possessing that knowledge is the thing it amputates. The gatekeeper. The analyst. The intermediary who was paid not for what they could think but for what they had accumulated. The entire middle of the value chain, where people sat between questions and answers and charged a toll."
He put the cap on the marker. He looked at the word OBSOLESCES for several seconds.
"Do not mistake this for unemployment. It is a restructuring of the sensorium. When print arrived, the scribes did not simply lose their employment. They lost their function in the social order. Their position had been a quality position. They believed their craft protected them. But Gutenberg was not interested in quality. He was interested in distribution. The scribes were defending the wrong asset."
The dean shifted in his chair.
McLuhan uncapped the marker. He wrote RETRIEVES in the bottom left and underlined it twice.
"This is the quadrant nobody asks about."
He turned to face them.
"Before our man Gutenberg fragmented the world into disciplines, a person could hold the whole of what was known. Not because there was less to know, although there was, but because the oral tradition permitted a kind of thinking that print destroyed. The tribal elder did not specialise. He moved between law and agriculture and medicine and weather and history and war, because all of these were the same conversation. The community held the knowledge. The elder held the connections.
"Print ended that. Print created the specialist. It created the university department, the professional credential, the expert witness, the man who knows everything about one thing and has been trained not to speculate about the rest. This was the Gutenberg Galaxy. It was extraordinarily productive. It built our modern world. And it made a certain kind of mind, the mind that moves freely between domains, not merely unfashionable but functionally unemployable.
"Your machine retrieves that mind."
He let this sit.
"If the specialist's accumulated knowledge is now available on demand, the advantage shifts to the person who knows which questions to ask across the boundaries. The generalist. The connector. The person our universities have spent four centuries training out of existence."
The postdoctoral fellow started to speak. McLuhan continued.
"Consider what has driven your significant innovations. Not the specialist deep in his trench, but the person who looked across two trenches and saw that they were digging toward each other. Printing itself was a convergence. The transistor was a convergence. The Web was a convergence. Every one of them required a mind that could hold two unrelated fields in tension until the connection became obvious. This kind of mind has been a minor player in our economy since the sixteenth century. Your machine hands it back its instrument."
He wrote the generalist returns in the quadrant and drew a circle around it.
"But the retrieval goes deeper. This machine talks. And you talk to it. You are not reading its output in the way you read a book, silently, alone, detached. You are in dialogue. You ask. It responds. You refine. It adjusts. This is conversation. This is the oral tradition."
He turned back to the room.
"You do not understand what you have done. You have not built a tool. You have rebuilt the campfire. The question is whether you have rebuilt the hearth or the pyre.
"Before print, all knowledge was conversational. You did not look it up. You asked someone who knew, and their answer depended on who was asking, and what they needed, and the context of the question. The book destroyed that. The book gave the same answer to everyone, regardless of who was reading, because the book does not know who is reading. Your machine knows. Your machine adjusts. Your machine gives a different answer to the farmer than to the physician, not because it is imprecise but because it has recovered a capacity that print amputated six centuries ago.
"The societal implications of this are enormous. In oral cultures, authority rested not on credentials but on demonstrated capacity. The elder was not elected or certified. He was recognised. If your machine makes the credential less relevant, because any person can access the knowledge, then authority migrates back toward judgment. Toward the person who asks well, not the person who has memorised most.
"Memory itself becomes communal again. Our oral cultures stored their knowledge in the group, not the individual. The bard was a vessel, not an author. Your machine is the same. It holds the knowledge of the network. No single person need carry it.
"And context returns to knowledge. The book is context-free. That was its genius and its violence. It travels without the author. It says the same thing in London and Lagos. Your machine restores the context. It shapes its answer to the listener. This is not a flaw. This is the recovery of something print stole."
He paused.
"But I should warn you. Oral cultures had no way to verify a claim at scale. In a tribal village, a lie could be checked by walking to the other end of the settlement. In a global village mediated by your machine, a lie has no end of the settlement to walk to. Every strength in this quadrant carries its shadow. The generalist can also be the dilettante. The conversational mode can also be the rumour. The contextual answer can also be the flattering one.
"The retrieval is real. So is the danger inside it. This has always been the case. Radio retrieved the spoken word and also retrieved the demagogue."
He paused.
"Some will read this retrieval as the return of media consensus. The standardised voice, the safe average, the three television stations. They will be describing something real. They will also be describing it in the wrong quadrant. Homogenisation is what a medium becomes at its limit, not what it recovers. Watch for it in a moment, when we come to reversal."
He capped the marker and set it on the ledge.
The dean cleared his throat. He had been reading. "Some are arguing this represents a new category," he said. "A cognitive medium, joining the mind rather than extending it. Beyond the tetrad's frame."
McLuhan looked at him with something close to pity.
"Every medium has always joined the mind. The book joined the mind. The clock joined the mind. The alphabet joined the mind so thoroughly that its users could not think outside its structure. What you are describing is not a new category. It is the same category, closer. Do not build a new framework to escape the discomfort of what the old framework reveals."
He picked up the marker again. He wrote REVERSES in the bottom right.
"When any medium is pushed to its extreme, it reverses into the opposite of its original intention. Radio, pushed to intensity, produced television. The automobile, pushed to congestion, produced the traffic jam, which is the absence of the automobile's function. The democratic press, pushed to its commercial limit, produced propaganda."
He looked at the board.
"Your machine, pushed to its limit, reverses from extending human cognition to replacing it. The user does not think and then consult the machine. The user consults the machine instead of thinking. The extension becomes the substitute. The amplifier becomes the signal.
"And then thought homogenises. Because everyone is thinking through the same medium. The machine that promised to extend individual capability dissolves individual voice. You all begin to sound the same. You all begin to think the same. Not because the machine is unintelligent but because the medium is singular. One campfire. One story. One bard."
He put down the marker.
"The Narcissus trap. Man becomes the servomechanism of his own cognitive extension, fascinated by its reflection, unable to recognise it as himself."
The dean said it was a question of how you use it.
McLuhan looked at him the way a man looks at a clock that has stopped. He had heard this sentence before. In 1964, General David Sarnoff, chairman of RCA, had told a university audience that technology was morally neutral. That it was the way you used it that determined its value. McLuhan had given his answer then.
"The smallpox virus is in itself neither good nor bad. It is the way it is used that determines its value."
Nobody laughed. Nobody had laughed in 1964 either. The accuracy of the observation had never been the problem.
"You are discussing the content," McLuhan said. "Whether this machine produces good things or bad things. Whether it is used wisely or foolishly. We have been having this conversation since Gutenberg. We will have it until the heat death of the universe. And it will never matter, because the medium is already doing its work while you debate the content. You are sitting in the water and asking whether the water is wet."
He looked at the whiteboard. The four quadrants. The marker was running out of ink toward the end, so REVERSES was fainter than the other three.
"Put me back," he said.
The postdoctoral fellow asked if he wanted to see what the machine could do. She had a laptop.
"I have seen what your medium does," McLuhan said. "It is written on that board. I do not need to examine the content."
He asked them to leave the heating off.
They found the whiteboard the next morning, untouched. Four quadrants in blue marker, the handwriting of a man who had last held a pen in 1979. In the retrieval quadrant, two phrases circled: the generalist returns and the campfire.
Nobody took a photo.
The dewar was moved to a facility outside the city. The committee filed its report. Thirty-eight pages. It recommended further study.
The whiteboard was cleaned on Thursday. What was written on it kept happening.
McLuhan's tetrad asks four questions of any technology. Most of the conversation about AI is stuck in the first two: what it makes better, what it makes redundant. The third quadrant, retrieval, is where the structural opportunity sits. The fourth, reversal, is where the structural risk lives.
The question the last scribe never asked about the printing press. The question the room never asked about AI.
What is this medium doing to the way your audience thinks?
Plan B diagnoses brand as a structural problem. If you are looking at the content while the medium does its work, we should talk.
