Older Than Either of Us

A cockatoo works a bin lid open with its foot and its beak. A crow drops a stone in a tube. A grey parrot names a colour it has not been shown before. The clip is fifteen seconds long and the caption underneath it is always the same caption.

So human. Basically a little person. Smarter than my coworkers.

The videos are real and the birds are doing the things. What we want to look at is the measuring instrument in the caption — because so human is not a description of a bird. It is a scale with one end fixed to us.

We are not going to spend this zine on that argument, because this house has already made it twice and made it better. No. 36 took the long-finned pilot whale — roughly 37.2 billion neocortical neurons, the first animal ever counted above us — and showed the ranking surviving its own yardstick. No. 85 took the words themselves apart. Field Guide No. 11 gave the crow a card.

So this one starts further in, at a narrower and stranger question: what is actually in there? And then at the question the field has not answered about it.

In 2020 a team working across Bochum, Jülich and Aachen put pigeon and barn owl forebrain under three-dimensional polarised light imaging — a method that reads the orientation of nerve fibres directly, from how they bend light.

They were looking at a structure that, by every textbook account, is built nothing like ours. The mammalian isocortex is layered: six sheets stacked, with vertical inputs and outputs threading through long horizontal fibres. A bird's forebrain is not layered. It is nuclear — clustered, not stratified. That difference is the entire basis of a century of ranking.

What the fibres showed was an orthogonal organisation: fibres running radially and fibres running tangentially, crossing at right angles, along the entire extent of both the Wulst and the sensory dorsal ventricular ridge.

Lamina-like. Column-like. Not layers and columns — the gross anatomy really is different, and the paper does not pretend otherwise. But the wiring motif that layers and columns exist in order to produce is there, in a brain that never built either.

One lucky slice through one bird is an anecdote. The reason this paper matters is the word iterative.

The same circuit motif turned up across three senses — somatosensory, visual, auditory. Across two brain regions — the sensory dorsal ventricular ridge and the Wulst. And across two species that last shared an ancestor a very long time ago: a pigeon and a barn owl.

Where the species differed, they differed sensibly. The barn owl's Wulst was more sub-differentiated — which fits a bird that hunts by stereoscopic vision and needs the acuity. A difference that tracks the animal's life is a sign the method is reading something real.

And here is the part a summary would drop. The motif was not everywhere. The non-sensory parts of the dorsal ventricular ridge showed something else entirely — described in the paper as a complex mosaic-like arrangement, patches of fibres running in different directions. The tidy architecture has an edge, and the authors printed where it is.

We are pointing at that on purpose. A finding that comes with its own boundary is worth more than one that arrives seamless.

This is the picture worth holding, because it is the one the phrase bird brain was invented to prevent you from having.

On one side, a sheet: neurons in stacked layers, vertical fibres crossing horizontal ones, folded to fit a skull. On the other, clusters: no sheets at all, no folding, and — running through them — the same right-angled crossing of radial and tangential fibres.

The shape of the container is not the shape of the computation. A layered sheet is one way to hold that crossing. A cluster is another. The textbook mistook the container for the requirement.

Read the figure as a schematic of the wiring motif, not as anatomy at scale. Both brains are drawn small and stylised; the point is the crossing, and that the crossing survives the change of container.

If a crow does what an ape does, the first place anyone looks is the count. And the count is genuinely startling — right up to the moment you try to make it do the work.

Olkowicz and colleagues counted neurons across 28 bird species in 2016 and found that songbird and parrot brains carry about twice as many neurons as primate brains of the same mass, and up to four times as many as rodents'. More than that, they are in the right place: the share of a brain's neurons sitting in the pallium runs about 19 per cent in primates, against 62 per cent in parrots and 78 per cent in songbirds.

So far the story writes itself. Then you put the actual totals side by side.

Six times the neurons, comparable performance. Güntürkün, Pusch and Rose put the problem in their own paper and do not paper over it: corvids are cognitively on par with apes and their pallial neuron numbers still differ widely.

Note what just happened, because it is the honest shape of this whole subject. The count was supposed to rescue the birds. Instead it stopped being a ranking at all. A measure that fails in both directions was never measuring the thing.

A corvid brain weighs eight to ten grams. A chimpanzee's is around four hundred. The birds in this literature carry non-cortical brains of roughly one to twenty-five grams, and the comparison is being made anyway.

Two things make that arithmetic survivable, and neither is a superpower.

The first is where the neurons are put. Ströckens and colleagues counted associative pallial neurons — the ones sitting between sensory input and motor output, which is where flexible thought has to happen — and found corvids carrying far more of them than chickens, pigeons or ostriches. A crow matches an ostrich on associative neurons with a brain two and a half times smaller, in a body more than two hundred times smaller.

The second is fuel. Avian neurons were found by von Eugen and colleagues to consume roughly a third of the glucose a mammalian neuron uses. The bird did not get a bigger brain. It got a cheaper one, and then spent the savings on density.

The comparison's own author calls it unfair, and we are keeping that. Corvid associative neuron numbers come out roughly on par with prefrontal neuron counts in chimpanzees — but as Güntürkün, Pusch and Rose write, this is not entirely fair, because mammalian associative neurons also live well beyond the prefrontal cortex. We are printing the hedge because the hedge is the finding's actual size.

Here is where the subject gets genuinely interesting, and where every fifteen-second explainer stops. Nobody knows where that circuit came from.

Stacho and colleagues end their own paper with a suggestion, and it is not the one that travels:

Read that slowly. If it holds, the crossing in a pigeon's Wulst and the crossing in your cortex are the same circuit — not two solutions that happen to rhyme, but one inheritance, kept by two lineages that last shared an ancestor somewhere around 320 million years ago, and modified differently since.

Not the bird evolved its own version of what we have. Something closer to: the kit is older than either of us, and neither of us invented it.

Note the paper's own grammar: it is likely that … might have been. Two hedges in one sentence. This is a proposal, and it is presented as one.

Güntürkün, Pusch and Rose — and Güntürkün is a co-author on the paper you just read — reach for the opposite explanation.

They propose four features a brain seems to need for complex cognition: large numbers of associative pallial neurons; an area that does the work of a prefrontal cortex; dense dopaminergic innervation of those associative areas; and neuron groups with the dynamic properties that hold something in working memory.

And their claim about those four is flatly the other reading: they convergently evolved — arrived at independently, in birds and in mammals — which is what makes them, in the paper's phrase, hard to replace. The bird's prefrontal-like area sits inside a structure that cannot be homologous to our prefrontal cortex, so on their account it cannot be the same thing inherited. It has to have been built twice.

The same paper also gives us the sentence that dismantles the ranking from the inside:

And they grade their own list. Obviously, this is a speculative list — their words, about their own four features. A field in the middle of an argument tends to be more honest than the article summarising it.

In February 2025 a third group went underneath the anatomy entirely, to the cells.

Zaremba and colleagues built single-cell atlases of the whole chicken pallium — adult, plus eight stages of development inside the egg — and set them against equivalent atlases from mammals and from non-avian reptiles. Not does this look like a cortex, but which cell types are these, and where do they come from?

The answer came back mixed, which is the least quotable and most likely shape for a real one. Inhibitory neurons: expression patterns conserved across amniotes, with one cell type expanded in birds — a type that in mammals sits mostly in the amygdala, and in birds is found throughout the pallium. Excitatory neurons in the hippocampal regions: conserved. Elsewhere: homologies in some places, and transcriptomic convergence in others.

So: partly inherited, partly arrived at twice, and the split does not fall where the gross anatomy would have predicted.

Why the chicken. Not because a chicken is the cleverest bird — it is the bird our own culture ranks at the bottom, and the one we eat by the tens of billions. It is here because it is a tractable laboratory animal whose embryos can be sampled through development. The map of the pallium that this argument now runs on came from the bird nobody was arguing about.

Search for any of this and you will meet a single confident phrase: convergent evolution. Birds evolved intelligence independently. Nature ran the experiment twice.

It is a wonderful story. It may well be right. It is also one of three live positions, presented as the answer.

And you can see exactly why it wins. Two independent inventions of a mind is a headline. An ancient circuit both lineages kept is a headline. Partly conserved, partly convergent, and the boundary is still being mapped at single-cell resolution is not a headline — it is a research programme. The version that travels is the one with the hedge filed off.

This is the house's standing complaint, and it is why How We Got Here exists: not an argument against evidence, an argument for the parts left behind. Here the part left behind is simply that three groups of researchers, publishing in Science and in Trends in Cognitive Sciences, currently disagree.

We are not resolving it either. We have read the three papers, and we are not qualified to adjudicate between them — nobody outside that literature is. What we can do is refuse to print one of the three as though the other two were not there.

Here is why we can sit inside an unsettled question without losing the thread. The belonging claim does not depend on which way it goes.

If it was built twice — then the layered cortex is one implementation among at least two, and the thing our textbooks call the seat of higher thought is, in the field's own word, dispensable. There is more than one way to arrive at a mind, and the one we happen to have is not the entry fee.

If it was inherited — then it is stranger and warmer than that. The crossing in the crow's forebrain and the crossing in your cortex are one very old piece of equipment, carried out of the Carboniferous by two lineages that have not shared a body plan in something like 320 million years. Neither of us designed it. Neither of us is the version it was for.

This is the same move the site is built on, in an unfamiliar place. You are made of atoms assembled in dying stars, and so is everything else; the inheritance is not a distinction. A shared inheritance cannot be a hierarchy. It can only be a variety of uses.

And the variation is the part that is not in dispute at all. No reading of these papers makes the pigeon a failed owl or the chicken a failed crow. They are different uses of a common toolkit, each sufficient for the life it is in — which is precisely the claim the neurodiversity paradigm makes about us, arriving here from fibre orientation and single-cell sequencing rather than from a manifesto.

Back to the clip, with the architecture behind it. In Sydney, sulphur-crested cockatoos learned to lift the lids of household waste bins. Klump and colleagues tracked what happened next, and published it in Science in 2021.

The behaviour spread by social learning to 44 suburbs. It was not one trick copied exactly: birds opened lids with individual style, and the styles varied by site — local dialects of the same innovation. And the people responded, developing what the paper describes as a counter-culture of technical measures: bricks, bungees, weighted lids.

Notice what the whole scene is made of, and that none of it is inside the bird.

There is an environment that was built by us and then changed by us again in response. There is a relationship — the behaviour travels bird to bird, which is what makes it a culture rather than a talent. There is attention, sustained across a sequence: a raven will turn down food it can have now to take a tool or a token it can only use tomorrow. And there is a system, which in this story is a municipal waste contract and, in the laboratory version, a century of ranking.

None of that is a metaphor for us, and we are not going to make it one. It is the same stack, seen somewhere it is easy to see because nobody's dignity is being adjudicated. When we describe a human being's behaviour without the environment, the relationships, the attention and the system, we are not being rigorous. We are leaving out most of the causes.