The Lizard That
Was Never There

You have heard it, and you may have said it. That was my lizard brain. My amygdala hijacked me. The reptile took over. It is a way of apologising for a reaction by attributing it to an older, dumber animal living downstairs.

There is no such animal. There is no such downstairs.

The model behind the phrase is called the triune brain, and a paper published in April 2026 disposes of it in a single sentence that is worth reading slowly, because of how little hedging is in it:

“Utterly false” is not a phrase working scientists spend lightly. And the model has been contradicted in the specialist literature since 1969. So this chain is not really about whether the lizard brain is real. It is about how a claim that was refuted before most of us were born ended up in everyone's mouth — and what is actually dividing the brain, which turns out to be better than the story it replaces.

Thirteen links. We are implicated in one of them.

Paul Broca, describing mammalian brains, notices a rim of cortex curving around the brainstem like a collar. He calls it le grand lobe limbique — the great limbic lobe.

The word is limbus. It means border, edge, hem. Broca is naming a shape, and a position: the stuff at the margin, where the cortex folds round. His grouping takes in the cingulate and parahippocampal gyri, the hippocampus, the primary olfactory cortex, the diagonal band, retrosplenial and medial prefrontal areas.

Note what is not in that sentence. No function. No claim about feelings. No suggestion that these parts are older than the rest, or that they do one job together. Broca has drawn a line around a region and named it for where it sits.

This is the cleanest link in the chain and the most important one to get right, because everything downstream is a series of additions to a word that started out meaning nothing more than around the edge.

James Papez publishes A proposed mechanism of emotion. He argues that the hippocampus, the hypothalamus, the anterior thalamic nuclei, the cingulate gyrus and the connections between them form a working loop, and that this loop is where emotion is elaborated.

It becomes known as the Papez circuit, and it is a real contribution: an anatomical hypothesis, stated as one, about how a set of structures might be wired together.

Two things travel forward from here, and only one of them should have. The wiring diagram was productive and much of it survives. The association — this region, that feeling — hardened much faster than the evidence for it, and it hardened onto a set of structures whose only original qualification was sitting at the edge.

The 2026 paper is blunt about how thin the direct evidence still is for the version everyone inherited: “Empirical examination of whether an aroma, song, or vista better evokes an individual's old memories is scarce.” Eighty-nine years on, the question at the bottom of the whole edifice has barely been asked.

Paul MacLean publishes on what he calls the visceral brain in 1949, and in 1952 coins the term that will outlive him: the limbic system.

Look at what that phrase does. It takes Broca's word for an edge, adds Papez's circuit and further structures — septum, amygdala, hypothalamus — and binds them into a system: a set of parts that belong together because they do one thing.

A name for a place has become a name for a function. Nothing was discovered at this step. A category was drawn, and the drawing did the work.

The 2026 authors, looking at the same collection of structures seventy-four years later, put the problem plainly: the limbic system “has resisted functional definition or even a specific list of its contents.” Its members have disparate jobs. Olfactory researchers and memory researchers are, for practical purposes, describing different objects and using the same word.

That is what happens when a category is named before it is characterised. It becomes very hard to disprove, because nobody can say exactly what it claims.

MacLean assembles the full model. The human brain, he proposes, is three brains in one, added in sequence over evolutionary time, each retained inside the next:

• the reptilian complex — the basal ganglia; instinct, aggression, ritual;
• the paleomammalian complex — the limbic system; emotion;
• the neomammalian complex — the neocortex; reason, language, planning.

He describes each as independently conscious. A reptile's brain, on this account, stops at the second floor. Ours has a third built on top.

This is the step where a functional claim becomes a claim about chronology — and chronology, pointed at living animals, becomes a rank. Older reads as cruder. Newer reads as better. The lizard is downstairs because the lizard arrived first.

It is a magnificent diagram. It is drawable on a napkin, it explains your worst afternoon, and it flatters the species drawing it. None of which is evidence.

Harvey Karten publishes The organization of the avian telencephalon and some speculations on the phylogeny of the amniote telencephalon in the Annals of the New York Academy of Sciences. It is a paper about pigeons, and it takes the floor out from under the ladder.

Karten shows that two regions of the bird forebrain — the dorsal ventricular ridge and the Wulst — contain the same neuron types and the same patterns of connectivity as mammalian neocortex. They are not arranged in six layers. They are arranged in nuclei.

The six layers are a packing arrangement, not the thing that does the thinking. Give a bird the same cell types and the same wiring in a different shape, and you get the same computation.

Which means the top floor of the triune building is not a floor, and it is not on top, and birds — who on the model should have stopped at the reptile — have the circuitry all along.

Hold this date. This is 1969. The chain has eight more links and most of them are still ahead of us, and the counter-evidence is already in print, in a mainstream journal, from the person who would spend the next forty years on it.

Carl Sagan publishes The Dragons of Eden: Speculations on the Evolution of Human Intelligence. It wins the Pulitzer Prize for General Non-Fiction in 1978. A substantial part of it is given over to MacLean's triune brain, presented with all of Sagan's gift for making an idea feel like something you have always known.

We have to stop here and say something about ourselves. This site is named out of Sagan's line about star stuff. Its tagline, Cosmic Connections, is partly out of his 1973 book. He is one of two figures the entire collection stands on.

And this is the book that took a refuted model out of the journals and put it into common speech, where it has stayed for fifty years. Both things are true. We are not going to file only the flattering one.

Sagan was doing what popular science does: taking the most vivid current account and giving it reach. The account he chose had been contradicted in a mainstream journal eight years earlier. Reach is not a filter. The better a communicator is, the more expensive their sourcing errors become — and this collection exists because of that, not in spite of it.

Why this joint is marked contested That The Dragons of Eden is the reason “lizard brain” is in ordinary speech is a causal claim about cultural transmission. It is widely asserted and it is plausible — the book was enormous and the phrase spreads after it — but we have not found a study that establishes it, and popular ideas rarely have one carrier. The book's role is documented. Its sole responsibility is not.

Barbara Finlay and Richard Darlington publish Linked regularities in the development and evolution of mammalian brains in Science. They measure the parts of the brain across 131 species of primates, bats and insectivores, and find something the ladder cannot accommodate.

The size of every component, from medulla to forebrain, is highly predictable from the size of the brain as a whole. One nonlinear function. Not a set of independent modules that can be bolted on, but a single developmental schedule that runs for longer or shorter.

And the order in which parts are generated is strongly conserved across mammals. The structures made late in that schedule are the ones that grow disproportionately when a brain gets bigger. Run the programme a little longer and the late parts — the neocortex among them — expand hardest, without anything being added.

The neocortex is not a newer floor. It is the part of an old schedule that has the most room left to grow when the schedule runs long.

Remember Finlay's name. She will be back on this chain in thirty-one years, as an author of the paper that closes it.

An international consortium of neuroscientists publishes Avian brains and a new understanding of vertebrate brain evolution in Nature Reviews Neuroscience, and formally renames most of the bird forebrain.

Read the old names aloud. Paleostriatum. Neostriatum. Archistriatum. Hyperstriatum. Old-striatum, new-striatum, ancient-striatum. A century-old vocabulary in which every label asserts two things: that these regions are basal ganglia, and that they arrived in an order.

Both were wrong. The greatest expanse of the bird telencephalon turns out to be derived from pallium — the cortical side — not striatum at all. Only one of the old “striatum” regions is genuinely equivalent to the mammalian striatum.

The terminology had to be replaced because a name is a claim, and students were absorbing the claim every time they learned the word. You cannot teach around a vocabulary that argues with you.

This is the field doing the expensive, unglamorous thing: not writing a rebuttal, but changing the words, in a hundred textbooks, because the words had been carrying a wrong idea for free.

Joseph Cesario, David Johnson and Heather Eisthen publish Your Brain Is Not an Onion With a Tiny Reptile Inside in Current Directions in Psychological Science. The title is doing a job the field had evidently stopped doing for itself.

Their description of the situation is the sentence to keep:

Not a live controversy. Two rooms. In one room the question was settled decades ago and there is unanimous agreement. In the other room it is being taught to first-year students as fact.

They sampled 20 introductory psychology textbooks published between 2009 and 2017. Of the 14 that discuss brain evolution, 86% contained at least one inaccuracy. Only two described brain evolution in a way matching the consensus among comparative neurobiologists.

Karten published in 1969. Two textbooks out of twenty had caught up by 2017.

Nabil Imam, Matthew Kielo, Brandon Trude and Barbara Finlay publish Dual computational systems in the development and evolution of mammalian brains in Science Advances. They measure brain structures across 182 mammalian species in 10 taxonomic groups.

What they find is not a ladder and not independence. It is a trade.

The olfactory system and the neocortex covary inversely. Take out the effect of overall brain size and plot what is left: species with more olfactory system than expected have less neocortex than expected, and the reverse, at r = −0.63.

Meanwhile the structures MacLean bundled as limbic — amygdala, septum, schizocortex, hippocampus — covary positively with the olfactory system and with each other. They really do belong together. They just do not belong together for the reason he gave.

Two systems, in tension, competing. Not three storeys stacked by age.

Here is the part that makes the whole thing click, and it has nothing to do with when anything evolved. It has to do with what the input is like.

Vision, touch and hearing arrive with space already in them. Things near each other in the world are near each other in the signal. The team measured this in the actual datasets — images, tactile-glove pressure readings, speech — and got spatial autocorrelation scores of 0.90, 0.86 and 0.89.

Smell does not. Two odour molecules that are chemically adjacent smell nothing alike. There is no “next to” in odour space. Same measurement: 0.11. On a locality index the gap is starker still — 0.97, 0.84 and 0.99 against 0.00.

So they trained artificial networks on each kind of data and looked at what organisation emerged. Networks optimised for vision, touch and hearing converge on orderly maps that keep neighbours as neighbours — which is what neocortex looks like. Networks optimised for smell, and for relational memory, develop fractured maps with distributed connectivity — which is what the olfactory system and hippocampus look like.

The division is not old versus new. It is mappable versus not. One kind of information can be laid out on a sheet. The other cannot, and needs wiring that reaches everywhere.

If two systems need incompatible wiring and both need space, they compete. The team tested that directly: an evolutionary algorithm optimising one network with two connectivity domains — one orderly, one distributed — for different mixes of task. Change what the animal needs to do, and the two subnetworks expand and contract against each other, reproducing the pattern seen across the 182 species.

Which brings us to the nine-banded armadillo, which forages by smell in the dark. Against what its brain size predicts, its hippocampus is twice the expected volume — 464 mm³ of extra circuitry — its olfactory system is 1.5 times larger, and its neocortex is smaller by a factor of 2.8.

Now the owl monkey: a small nocturnal primate. Brain 15 g. And the agouti, a rodent: brain 16 g. Nearly identical brains by weight, and utterly different inside — in the agouti the olfactory system is so large it obscures the cerebellum from view.

An armadillo is not a failed monkey. It is a nervous system that spent its room on the thing it lives by. There is no version of this where one of them is the deficient one, because there is no single axis for them to be worse along.

Why this joint is marked contested The inverse covariation across species is measured and solid. The explanation — that it happens because two computational systems compete for representational space — comes from optimising artificial neural networks and mapping the result onto cortex. That is a model, and the authors say their results “suggest” it. We are not going to upgrade their verb.

Everything above is history and measurement. This spread is not, and it is marked accordingly.

A model with an order in it will be used to rank whatever you point it at. That is not a misuse of the triune brain; it is the only thing a stack of floors can do. And the thing it got pointed at was people.

“That's your lizard brain” is not a neutral description of a reaction. It is a verdict on it. It says: what happened in you just now was the old animal, the part that predates reason, breaking through the floor. It relocates a person's response to a lower storey and files it under something to be overridden.

Ask who gets told this. It is said about anger and about panic, about a meltdown, about a startle response, about someone who could not mask any longer. It is a way of describing a nervous system doing exactly what it is built to do, as though it were a failure of the newer, better parts to keep the lid on.

There was never an older animal in there. All vertebrates share the same gross forebrain-to-spinal-cord segmentation, with no segments added on. The reaction was not primitive. It was yours, entire, from a brain with no basement.

And what replaced the ladder does not sort anybody. Two systems, competing for room, tuned by what a creature actually has to do. The armadillo went one way and the monkey the other, and the answer to which is better is that the question does not survive contact with the measurement.