The distributed brain

July 17, 2024 1:53 AM

Headmen, shamans, and mothers: natural and sexual selection for computational services "The human brain, which evolved to acquire, store, and process information to make beneficial decisions, is likewise energetically expensive to build and maintain yet plausibly has idle capacity much of the time. We propose that humans evolved to use advantages in information or computational resources to provide computational services to others via a language-based “network” in exchange for payments of various sorts that helped subsidize the energetic costs of the brain."
posted by dhruva (21 comments total)

This is interesting:

James Neel, a major figure in twentieth century genetics, observed that in small-scale societies leaders are often polygynous and have more children than other men. Neel reasoned that if this pattern characterized most human societies during our evolution, there would have been strong sexual selection for the trait(s) that predisposed men to ascend in social rank and become community leaders. Many later authors have argued that this trait is status striving, i.e., that the reproductive benefits of prestige and leadership roles would select for a strong motive in men to achieve high status. Neel had something quite different in mind. Political leaders in small-scale societies are chosen by other men. What traits would men value in other men? Leaders are often skilled hunters, warriors, orators, and masters of tribal lore. Neel therefore argued that although physical strength is an asset in campaigns for leadership in small-scale societies, mental agility is even more critical. Neel proposed that there was social and sexual selection, not on status striving, but on cognitive abilities. Because sexual selection often results in exaggerated traits, this could explain the dramatic increase in human cranial capacity in the genus Homo.

as is this:

In our model, in addition to increasing male mating success, male investment in energetically expensive brain tissue would have directly benefited females. Moreover, as we argue next, expression of the relevant alleles in females would also have benefitted females, thus aligning sexual and natural selection.
posted by chavenet at 2:52 AM on July 17

Great! I was formulating a question for mefi along these lines so really looking fwd to reading later. I would like to see a female researcher view of this though.

Lead author Hagen has super interesting page at WashU, some very out of the box ideas (CW as research includes drivers of suicidal ideation).
posted by unearthed at 3:02 AM on July 17 [2 favorites]

We do not aim to provide a comprehensive theory of shamanism… We aim only to explain… exchange for payments such as meat and other foods, tobacco, slaves…
so, their argument is: “shamanism” = capitalism?
posted by HearHere at 4:29 AM on July 17

computational metaphors are eating science
posted by AlbertCalavicci at 4:57 AM on July 17 [10 favorites]

computational metaphors are eating science

They need the calories to keep growing.
posted by JSilva at 5:31 AM on July 17 [1 favorite]

I think it is wise to be cautious whenever scientists start trying to explain social developments in terms of evolution. Especially when they attempt to project practices of extant cultures onto an imagined primordial human society.

The passing down of information and development of leadership and delegation aren't more biologically driven for "small scale" groups than they are for us. We rationally choose to retain useful information and organize ourselves and delegate decision making because it works and makes sense. There is no reason to think early human cultures were any less rationally motovated to do the same.

The reduction of human culture to computational resources, and leadership and caretaking to evolutionarily driven "payment" in food and sex feels very dehumanizing to both ancient peoples and the extant groups used as models for them by researchers.
posted by pattern juggler at 6:07 AM on July 17 [19 favorites]

God is the original Bitcoin
posted by Reyturner at 7:38 AM on July 17

computational metaphors are eating science

Pray tell, good sir, what think you of my pneumatical model of the evolution of the mind of man?
posted by clawsoon at 7:50 AM on July 17 [1 favorite]

This would answer a basic question I have about software engineers. Namely, how is it that myself and other plains apes have been born with a solid instinct for extremely large logic driven systems? It doesn’t seem like the kind of natural advantage that evolution breeds for.

But here we posit an actual theory of that sort of capacity being useful in a pre-agricultural society. Probably not an accurate theory, but infinitely more theories than I had when I got up this morning.
posted by Tell Me No Lies at 8:00 AM on July 17 [1 favorite]

imho: Holographic storage and fractal redundancy are better metaphors for early human "computational services".

Information wasn't stored in a silo, it cohered from the richly woven connectivity between environment, community and culture. You don't have a calendar, you live inside one via constant intimate awareness of biological and meteorological phenomena which serve as markers that inform your available options. In societies where information is transmitted/stored orally it tends to be very fractal. Layers of stories within stories honed in form, function and fitness over hundreds of thousands of generations.
posted by neonamber at 9:14 AM on July 17 [7 favorites]

I only skimmed the article, but allow me to immediately yell about the pull quote. "the human brain ...plausibly has idle capacity much of the time"

They're wording this more carefully than the really stupid "you only use 10% of your brain capacity" myth, but its enough to trigger a little rant.

1. There's evidence that useful housekeeping and other stuff is going on in that "idle capacity" e.g. https://www.wired.com/story/what-your-brain-is-doing-when-youre-not-doing-anything/
2. Relatedly, we're still figuring out the brain - we don't know what's going on yet. Stop assuming things based on poor non-analogue analogies.
3. We do have damn good evidence that brains aren't just undifferentiated masses of computing bits waiting for information though.
3. It's fine for parts of biological systems (or really any energy using systems) to not be active all the time. I suspect people might think I was silly if I said "maintaining muscle tone is energetically expensive, the muscles in your non-dominant arm plausibly have idle capacity much of the time."

Anyway, I would really like some of the people who make brain arguments to actually learn some neuroscience before they make them.
posted by BlueBlueElectricBlue at 9:49 AM on July 17 [12 favorites]

Skimming through the paper, there's some interesting stuff there. Certainly you could keep yourself busy for a couple of years reading all the references.

Parts of it sounds like they're arguing that we got big brains in order to do emotional labour, though they attribute most of that to men, who they argue were responsible for group leadership, hunting cooperation, and medical shamanism. They only credit emotional labour for women in terms of having to keep track of lots of kids who are children for a long time.

(...though it looks like they'd disagree with my use of the term "emotional labour", since they limit emotional labour to "fostering others' wellbeing" and don't associate it with all of the work required to lead a group or get hunters to cooperate.)
posted by clawsoon at 10:49 AM on July 17 [2 favorites]

Namely, how is it that myself and other plains apes have been born with a solid instinct for extremely large logic driven systems?

Synaptic pruning means you are what you do, basically, and “instinct” is exactly the wrong way to think of it. It’s an acquired skillset, a deeply nested stack of abstraction layers and system models that you maintain and build upon because you use it frequently or use things that are adjacent to it frequently. Some fundamentals - eg bitmasks and binary representation of memory if you do a lot of low-level programming - are referenced by many different semantically adjacent stacks and this is *probably* loosely reflected in the biological substrate, which is otherwise not tightly coupled with semantic representation.

All naturally occurring biological neural networks that I am aware of are continuously (re)training in real time in response to sensory input. I am not a collection of memories, skills, and abstract models: I am an ongoing survival of the fittest attention-competition between 30 trillion dendritic connections distributed across 100 billion neurons, which are hosting representations of memories, skills, and abstract models. Which are the highest-tier derivative semantic structures and end product of my sensory inputs, internal activation feedback loops, and some neurotransmitter spillover. Which are me.

Let’s go less abstract, but still somewhat speculative: why do babies and toddlers learn languages so quickly, once they’ve finished setting up frequecy selection and phonetic parsing? Because there are vast fields of synapses sitting in an almost virginal state, ready to become a core pathway for whatever major concept is needed. Whereas in older people setting down deep cross-domain structures is probably far more difficult when there are already a thousand concepts juxtaposed on each synapse. This is also likely the reason childhood trauma is so difficult to overcome: it was laid down when there was less competition, so it’s incorporated into the high-traffic pathways of the network and pretty much everything/anything in daily life is lightly refreshing it. Keeping it semi-active at all times.

We talk about being haunted by the worst mistakes of our past for a reason: continuous reactivation makes those pathways linger far beyond their shelf life, and lends some subjective truth to the description.

As far as the paper is concerned: makes perfect sense and I already thought of PCs as an intellectual prosthesis, this is a generalization of that to include the squishy monkey-based processing units around me. Why do LLMs excite me despite their many, many flaws? Because it means that - like monkey-based external processing - the artificial prosthesis can begin to mirror our internal semantic structures, to speak the hidden language of our thoughts, and that is maybe only somewhat useful in isolation but it is an essential foundation for far, far greater things.
posted by Ryvar at 11:21 AM on July 17 [1 favorite]

Certainly you could keep yourself busy for a couple of years reading all the references.

The sheer number feels like padding, or over justification. I don't undestand so many and still wondering how real the article is.
posted by unearthed at 12:08 PM on July 17 [2 favorites]

In societies where information is transmitted/stored orally it tends to be very fractal. Layers of stories within stories honed in form, function and fitness over hundreds of thousands of generations.

The first part of Thomas Mann’s Joseph and his Brothers (which I’ve only read in English translation) starts with a beautiful extended meditation on exactly this topic. How even the identities of people who lived far apart in time, assuming they might ever have been real, become conflated in retelling as they assume similar roles in similar stories over the generations.

(And I would love to quote one of the passages, but right now I’m freaking out because ironically we just did a big book pruning, and I’m afraid I may have accidentally donated my copy! )
posted by mubba at 1:00 PM on July 17 [2 favorites]

This business about payment in food and sex sounds really sus to me, like a prehistory managed by Andreessen Horowitz. As a corrective the first half of David Graeber's Debt: The First 5000 Years is relevant: early societies were not markets and people did many things without expectation of direct reward. He is mostly tearing down notions of a "barter system" where if you need arrows you need to give up pots, or something. In many societies, if you want Fred's arrows you ask and Fred gives you some. You do not have to pay in pots, nor food, nor sex.

Or, not to overemphasize Graeber, but The Dawn of Everything (with David Wengrow) has a lot on historical egalitarian societies where leadership was not coercion. The ideas that males were more powerful and did all the hunting are also outdated.

As it happens I've been reading a lot about shamanism, and there is a lot of variation. Surviving shamanism can be transactional— i.e. you have to pay the shaman— but that's in societies where payment is already a thing. One model that would seem to challenge these theories is the !Kung, a society where half of adult males are shamans, and healing is a community event.
posted by zompist at 3:24 PM on July 17 [8 favorites]

Synaptic pruning means you are what you do, basically, and “instinct” is exactly the wrong way to think of it. It’s an acquired skillset

You can teach people how to do things, and with experience they can get pretty good at it.

However, for most human endeavors there are people who are born to do it. The most visible ones we call child prodigies and they're at the end of a curve that includes people born with various degrees of various talents. They need to train and develop their talents, but their high end is often much higher than what someone whose talents lie elsewhere can accomplish.

We all like to flatter ourselves that we could be olympic class athletes, or great scientists, or successful CEOs if only we had the resources and cared enough to learn. It's a fantasy. We are born with innate skills and innate limitations that both restrict and enhance what we're truly capable of.
posted by Tell Me No Lies at 3:51 PM on July 17 [1 favorite]

This business about payment in food and sex sounds really sus to me, like a prehistory managed by Andreessen Horowitz. As a corrective the first half of David Graeber's Debt: The First 5000 Years is relevant: early societies were not markets and people did many things without expectation of direct reward. He is mostly tearing down notions of a "barter system" where if you need arrows you need to give up pots, or something. In many societies, if you want Fred's arrows you ask and Fred gives you some. You do not have to pay in pots, nor food, nor sex.

I am not an anthropologist, but I don't think just because there wasn't explicit scorekeeping doesn't mean there was no division of labor or value exchange. Like, surely you were not just taking Fred's arrows then keeping everything you hunted, there must've been some way you gave back to the community.

Let’s go less abstract, but still somewhat speculative: why do babies and toddlers learn languages so quickly, once they’ve finished setting up frequecy selection and phonetic parsing? Because there are vast fields of synapses sitting in an almost virginal state, ready to become a core pathway for whatever major concept is needed.

I mean, another reason is that the human brain has language center(s) specialized for the task. Humans aren't exactly born with an instinct for reasoning through complex mathematical proofs, but it's also possible to go too far in the other direction, we are not completely blank slates and individual variation does exist.
posted by ndr at 4:58 PM on July 17 [1 favorite]

(And I would love to quote one of the passages, but right now I’m freaking out because ironically we just did a big book pruning, and I’m afraid I may have accidentally donated my copy! )

If you haven't already seen it, I think you would enjoy the film Ten Canoes.
posted by neonamber at 6:40 PM on July 17

However, for most human endeavors there are people who are born to do it. The most visible ones we call child prodigies

On the other hand, most of the time when I hear about a child prodigy there's a parent pushing them really hard. Not always, and not every kid who's pushed becomes a prodigy, but having an edge-of-insanity parent seems to help.

(Or perhaps I'm suffering from recency bias, having just watched an interview with Nyjah Huston.)
posted by clawsoon at 6:56 PM on July 17

Apes don't ask questions (more).

Baboons have relatively flexible listening capabilities, but cannot learn to express new things.

Some ants can recognize themselves in morrors.

Are we done yet with the idea that intelligence just represents computational capacity per se?

Around this Peter Watts' has some similarly reductive but more interesting ideas.
posted by jeffburdges at 7:43 AM on July 18 [2 favorites]