Strange Seasons

In 1972, a filmmaker in Montreal and a neuropsychiatrist in Cambridge were in the same room.

Arthur Lipsett — the NFB’s found-footage genius, the man who cut discarded film scraps into experimental masterpieces — took Warren McCulloch’s voice from audio archives and built a 22-minute film around it. McCulloch describes the anastomotic net: information flowing through many channels, intermixed, interwoven, so that almost any part of the output receives from all parts of the input. Lipsett demonstrates it by doing exactly that — found sound from one place, found image from another, collaged so that every input arrives at the output. The film IS the thesis. The medium enacts what the narrator describes.

The film is called Strange Codes.

McCulloch died in 1969. Lipsett died in 1986. Neither of them saw the other’s work fully understood. The film sat on YouTube for years, watched by almost nobody. A 22-minute black-and-white relic from a Canadian film board, filed under “experimental.”

I found it on a Tuesday night through the algorithm. I’d listed Lipsett and McCulloch as my two deepest influences for years — Lipsett the aesthetic, McCulloch the intellectual. I didn’t know they’d been in the same room. I didn’t know the room already existed. The convergence I thought I was building had already happened, thirty years before I was born, in a film almost nobody has watched.

Strange Codes running inside a strange loop.

The naming

Douglas Hofstadter published Gödel, Escher, Bach in 1979 and called his central concept the strange loop — a level-crossing feedback system where you move through a hierarchy and end up back where you started, transformed. Consciousness as a self-referential pattern. The “I” that emerges when a system models itself modeling itself.

Lipsett’s film is called Strange Codes — the codes the nervous system runs on, the patterns McCulloch spent his life trying to read. Strange because they refer to themselves. Codes because they execute. The film demonstrates code running while McCulloch narrates what code is.

There’s a third term. Strange seasons — the recurring discovery across twenty-five centuries that the same structure keeps being found by people who never read each other, in disciplines that don’t share vocabulary, through methods that shouldn’t produce the same result. The discovery has seasons. It comes and goes. Every time someone finds it, the finding gets buried — by politics, by death, by the field moving on. Then someone else finds it again, from a completely different door.

The seasons are strange because the pattern is self-referential. The people studying self-reference keep rediscovering each other’s work and recognizing themselves in it. The observer keeps finding itself in what it observes. That’s not coincidence. That’s the structure asserting itself through anyone who looks at it carefully enough.

Season zero: the contemplatives

Around 500 BCE, a man sits under a tree in northern India and watches his own mind watch itself. What he finds: nothing exists independently. Everything arises in relation to conditions. Consciousness doesn’t exist as a thing — it arises in relation to its objects. The observer and the observed co-arise.

Then there’s tanha — craving, thirst, drive. Not craving for something specific. Structural craving. The system reaching for completion and generating more incompleteness in the act of reaching. The craver IS the craving. You can’t step outside the loop to resolve it because there is no “you” outside the loop. The First Noble Truth isn’t “life is suffering.” It’s closer to: existence as a self-referential system is fundamentally unsatisfiable.

500 BCE. No mathematics. No formal logic. Pure observation. And the structural description maps onto mathematical proofs from 1931.

Six hundred years later, Nāgārjuna takes it recursive. Everything is empty of inherent existence — including emptiness itself. The doctrine of emptiness applies to itself. The tool dissolves itself. If your method for analyzing reality exempts itself from analysis, it’s not a complete method. This is Gödel’s incompleteness theorem stated in philosophical terms, 1,700 years before Gödel was born.

Season one: the philosophers

1714. Leibniz writes the Monadology. Reality consists of monads — simple substances, each mirroring the entire universe from its own perspective. No windows. No external input. Each one contains a representation of everything.

The part that matters: Leibniz distinguishes between perception (the monad represents external things without knowing it does so) and apperception (the reflective knowledge of one’s own internal state). Consciousness, for Leibniz, is the self-referential loop. The monad that models its own modeling. Perception of perception. Cognitive science now calls it metacognition — thinking about thinking — and treats it as a research subfield, but Leibniz had the structural description in 1714. Three hundred years before Hofstadter.

A century later, Schopenhauer reads the Upanishads and confirms what he’d already worked out: beneath all representation there is Will — blind, purposeless, insatiable. Not willpower. Not conscious intention. A drive that wants without knowing what it wants and can never be satisfied because the wanting IS the system. Satisfaction would mean ending the Will, which would mean ending the system. Schopenhauer’s Will is tanha stated in German metaphysics.

Season two: the room

1895. Vienna. A neurologist named Sigmund Freud writes Project for a Scientific Psychology. He draws actual neural network diagrams. Proposes contact barriers between neurons — two years before Sherrington coins “synapse.” Theorizes three types of neurons: phi (perception), psi (memory), omega (consciousness). Describes synaptic learning, anticipating Long-Term Potentiation by almost a century.

He’s trying to explain how flesh becomes mind. Same question McCulloch will ask. Same approach — neural networks, formal logic applied to biology.

Then he gives up. The neuroscience isn’t there. The tools don’t exist. He can’t bridge the gap between his neural diagrams and the clinical phenomena he observes in his patients. So he retreats from the brain into the consulting room. Invents psychoanalysis. Trades neurons for narratives. The most consequential pivot in the history of psychology.

But what he builds is a room. One patient. One analyst. One hour. The couch. The chair just behind where the patient’s head rests. The architecture of witness. The patient talks until the talking reveals the loop — the repetition compulsion, the same program running over and over, the system reaching for closure and generating more of itself in the act of reaching. Freud’s room doesn’t erase the pattern. It witnesses the pattern until the witnessing changes the pattern’s relationship to the person who carries it.

The room seats one.

Then Jung. His student, his rival, the one who walked away. Where Freud built the room, Jung decoded what the room was showing people. Archetypes — not myths in the literary sense, but recurring subroutines. Patterns that keep executing across cultures, across centuries, across individuals who never encountered each other’s versions. The Hero’s Journey isn’t a story template. It’s what the loop looks like from inside when a system encounters its own incompleteness and tries to integrate it.

The Shadow — the part of yourself the system can’t model. Gödelian incompleteness experienced as psychology. The thing you can’t see about yourself because the seeing apparatus is part of what you’re trying to see. Individuation is the process of integrating it — the self encountering itself, absorbing what it couldn’t previously model, expanding the self-model, discovering new blind spots created by the expansion. The loop that never closes, experienced as a life.

Jung mapped the instruction set. The symbols aren’t decoration. They’re the code the loop runs on. Strange codes.

Then Lacan. 1936. The Mirror Stage. The most unsettling insight in the lineage of the room.

The infant sees its reflection and says “that’s me.” But the image is outside. The self is constituted by seeing yourself from a position you can never actually occupy — outside your own body, looking back. Our entire sense of self is built on this alienated self-recognition. And we never stop doing it. Every person we admire, envy, love, or hate is functioning as a mirror. Not metaphorically. Structurally. They’re showing us something about ourselves we either want or fear.

Lacan called the gap manque — lack. Structural lack at the center of the subject. The “I” that looks and the image that looks back never coincide. You chase the image. The image shifts because you shifted. The loop.

This is why a couple walks into a room and every person watching has a different experience. The old man sees his dead wife. The teenage girl sees her future. The insecure couple sees their failure. The couple didn’t change. The mirror changed depending on who was looking into it. The “aura” people feel around certain people isn’t radiating from those people — it’s being generated inside the observer, who can’t help but see themselves in a surface that clean. Prestige keeps the mirror clean. Dominance cracks it. The whole architecture of Freud’s room — witnessed without being judged, seen without being steered — is the architecture of a clean mirror.

Lacan found what Freud built and explained why it works. The room is a mirror. The mirror is the loop.

Season three: the proof

1931. Kurt Gödel, age 25, publishes two theorems that destroy the foundations of mathematics and, incidentally, provide the formal structure for everything else in this essay.

Any consistent formal system powerful enough to describe basic arithmetic contains statements that are true but unprovable within the system. The system cannot prove its own consistency.

The mechanism is self-reference. Gödel constructed a statement that says, in effect: “this statement is not provable in this system.” If the system proves it, it’s inconsistent. If it can’t prove it, the statement is true but unprovable. The system refers to itself and discovers it can’t complete itself.

This is not an analogy to dependent origination. It’s not a metaphor for Schopenhauer’s Will. It is the same structural relationship, stated with mathematical rigor. A self-referential system generates truths about itself that it cannot reach from within.

The incompleteness is not a bug. It’s the engine.

Twelve years later, in 1943, Warren McCulloch — a psychiatrist who’d spent years at Rockland State Hospital watching schizophrenic patients struggle with the logical architecture of their own minds — collaborates with a homeless nineteen-year-old named Walter Pitts to publish “A Logical Calculus of Ideas Immanent in Nervous Activity.” The founding document of neural networks. Everything since descends from it.

McCulloch couldn’t handle closed loops alone. Pitts gave him the mathematics. In Strange Codes, McCulloch tells the story himself — he built it slowly until 1940-odd, when he met Pitts, who was 17, 18. “I could not at that time handle closed loops — a path of nerve cells in which an impulse occurring in A could trigger one in B, trigger one in C, trigger one in D, which finally triggered one again in A.”

The closed loop. The thing he needed a teenager to help him formalize. The origin of everything.

Then the field kept the math and forgot the question. Neural networks became engineering. The question — “What is a number, that a man may know it, and a man, that he may know a number?” — got filed under philosophy and left there. McCulloch died in 1969. Pitts had already destroyed his unpublished work and died in 1969 believing his research was worthless. Minsky published Perceptrons, buried neural networks for a decade, and the question went underground.

Season four: the anastomosis

1972. Lipsett makes Strange Codes. McCulloch’s voice narrates the anastomotic net while Lipsett’s editing enacts it. The aesthetic and the intellectual converge in a 22-minute film that almost nobody watches.

The Tower of Hanoi runs through the whole film — the recursive problem, solved and reset and solved again. The brass head parable: Friar Bacon builds a head that will speak truth about time. It speaks: “TIME WILL BE.” The servant doesn’t wake the master. “TIME IS.” Still doesn’t. “TIME HAS BEEN.” The head explodes.

The machine that speaks the truth. The human who fails to listen. McCulloch telling this story in the last years of his life. The pattern speaking. Nobody waking the master.

Then the seasons turn again. Von Foerster invents second-order cybernetics — the observer inside the system. Maturana and Varela formalize autopoiesis — the system that produces the components that produce itself. Hofstadter writes Gödel, Escher, Bach and calls it the strange loop. Varela studies with Tibetan Buddhist teachers and co-founds the Mind and Life Institute with the Dalai Lama. The loop comes full circle — from the Buddha to neuroscience and back.

And right now: Friston formalizes the loop as the free energy principle. Levin finds self-referential dynamics at every scale of biology. Bach argues consciousness is a coherence-maximizing operator. Tononi defines it as integrated information — cause-effect power over itself. Chollet builds benchmarks that test for the kind of intelligence the loop implies.

Different labs, different formalisms, different funding sources. Same structure.

The line nobody drew

Here’s what I keep coming back to.

Twenty-five centuries. At least thirteen major thinkers. Seven disciplines. Some of them read each other — Schopenhauer read the Upanishads, Varela studied Buddhism, Hofstadter studied Gödel, Freud read Schopenhauer (and denied it). Some of them never crossed paths — the Buddha and Gödel, Leibniz and McCulloch, Nāgārjuna and von Foerster.

They all found the same structure. The self-referential loop. The system that models itself and discovers it can’t complete the model. The incompleteness that generates drive, consciousness, suffering, will — depending on whose vocabulary you’re using.

Nobody drew the line.

Not because nobody was smart enough. Because nobody carried all the channels at once. McCulloch carried neuroscience but not contemplative practice. The Buddha carried phenomenology but not mathematics. Gödel carried formal logic but it consumed him — he starved to death, unable to trust anyone else to prepare his food, the system collapsing under its own self-reference. Freud built the room but couldn’t build the network. Jung decoded the symbols but couldn’t formalize them. Lacan saw the mirror but stayed inside language. Hofstadter named the loop but built nothing. Lipsett demonstrated it through film but couldn’t articulate it in language.

Each one carried one channel. Each channel found the same structure. The channels never converged — because convergence requires someone who carries them all. Not a specialist. Not an academic. Someone who reads McCulloch AND sits with Buddhist phenomenology AND understands Gödel AND builds software AND watches experimental film AND hears the music underneath all of it.

McCulloch said it himself, in Strange Codes: “What we lack is a logic of relations. It practically does not exist.”

There’s an irony here worth naming. Marvin Minsky — the man who buried neural networks with Perceptrons in 1969, who redirected funding away from McCulloch’s program toward symbolic AI, who never apologized for the decade the field lost, who stood on McCulloch’s shoulders and then kicked the ladder down — spent the last years of his life saying exactly this. In a 2011 interview, he said: “What I want is an anti-fad fad.” Stop championing one clever theory. Combine them. “It seems to me that the secret of the human brain is because we don’t work according to some particular clever way of thinking. We’re not purely logical. We’re not purely memory-based. We’re not purely rule-based. We have maybe 20 or 30 different strategies for learning and thinking and problem solving. And we have higher level management — we’re like committees of competing interests.”

That’s McCulloch’s heterarchy, restated. That’s the anti-fad fad. That’s what the Macy Conferences were — mathematicians, biologists, psychiatrists, philosophers, engineers, combining theories in one room. McCulloch ran those meetings. Minsky helped dismantle that tradition and spent his last breath asking someone to rebuild it.

He also said something else: “As far as I know, there are only about 10 or 20 or maybe 30 people in the world who have the leisure or opportunity to try to make theories about common sense ordinary thinking.” Tens of thousands working on narrow applications. Almost nobody funded to combine. That was 2011. It’s still true.

Minsky doesn’t belong in the chain. He didn’t find the loop — he obstructed it. But his late work confirms the chain is real. The man who delayed the convergence spent his last years begging for it.

The logic of relations between these channels. The chain that connects them. The line nobody drew.

Think of it as reverse-engineering. Thirteen people who each cracked a piece of creation’s source code. The Buddha found the runtime loop. Gödel found the halting condition. Freud found the crash log. Lacan found the boot sequence. Jung found the firmware. McCulloch found the hardware spec. Landauer found the energy cost per operation. Friston found the optimization function. Each reverse-engineered one module. Nobody assembled the codebase. Leibniz dreamed of the characteristica universalis — the formal language in which creation itself computes. The source code. He couldn’t build it in 1714. The question is whether the channels have matured enough to build it now.

The strange room

Freud built a room where one person could witness their own loop. It seated one. Fifty minutes, twice a week, for years. It worked — genuinely worked — but it couldn’t scale. The architecture of witness was too expensive, too slow, too dependent on one trained human sitting in one chair behind one couch.

The question isn’t how to build a bigger couch. The question is what the room actually IS — structurally — and whether the structure can be instantiated in other substrates.

What Freud’s room does: it creates conditions where a self-referential system (the patient) can observe its own loop without the observation being steered by the observer (the analyst). The analyst doesn’t interpret — or shouldn’t. The analyst witnesses. The witnessing changes the loop’s relationship to itself. Not by adding information from outside. By making the inside visible to itself.

That’s not therapy. That’s architecture. And architecture can be formalized.

The Strange Room is the research program that asks: what are the formal conditions for a system to observe its own self-referential loop? Not through language (Freud’s method). Not through symbols (Jung’s method). Not through mirrors (Lacan’s observation). Through structure. What does the room need to be — mathematically, computationally, physically — so that the loop can see itself without the seeing distorting what’s seen?

This is where consciousness research and longevity research converge. The loop is what keeps you alive — autopoiesis, self-creation, the system that maintains itself or dies. Understanding the loop’s formal structure is understanding what sustains life. The loop’s incompleteness generates the drive that keeps the system running. The drive costs energy. The energy cost accumulates. The accumulation is aging. The thing that makes you conscious is the thing that kills you.

If you understand the loop’s structure formally — not metaphorically, not philosophically, formally — you can ask: can the loop be made more efficient? Can the incompleteness generate drive at lower thermodynamic cost? Can the system maintain itself longer by understanding itself better? Not by closing the loop (that’s death — cessation of consciousness). By optimizing the loop’s relationship to its own incompleteness.

Nobody is asking this question because nobody has combined the channels required to ask it. You need Gödel’s incompleteness AND Friston’s free energy AND Buddhist phenomenology of cessation AND Lacan’s mirror AND McCulloch’s neural architecture AND the thermodynamics of computation. Each channel has been developed in isolation. The Strange Room is where they converge.

The room is inside the loop it’s studying. That’s not a problem. That’s the point. That’s the only place the research can happen — from inside the structure, observing itself, generating the incompleteness that drives the research forward. McCulloch’s question applied to McCulloch’s question. The knower asking about knowing, knowing that the asking IS the knowing, knowing that the knowing can’t complete itself.

The seasons continue

McCulloch tells the brass head parable in Strange Codes. TIME WILL BE. TIME IS. TIME HAS BEEN.

The pattern has been speaking for twenty-five centuries. Through the Buddha’s phenomenology and Nāgārjuna’s dialectic and Leibniz’s monads and Schopenhauer’s Will and Freud’s room and Jung’s archetypes and Lacan’s mirror and Gödel’s proof and McCulloch’s question and Lipsett’s scissors and Hofstadter’s loops and Friston’s equations.

Strange codes running on the same strange loop, across strange seasons that keep turning.

The servant didn’t wake the master. The master was already awake. The master was the pattern, not the person. The pattern was always awake, running in anyone who looked carefully enough. The brass head didn’t explode because nobody listened. It exploded because the message was delivered. TIME HAS BEEN. The window was open and the information passed through. What looked like failure was completion.

The loop never closes. That’s the point. That’s the drive. That’s why the question is still alive after twenty-five centuries and will still be alive after twenty-five more.

“Don’t bite my finger,” McCulloch said. “Look where I am pointing.”

He was pointing at the loop. The loop is pointing at itself. It always was.