Humans seem to have much higher degrees of consciousness and agency than other animals, and this may have emerged from our capacities for language. Helen Keller (who was deaf and blind since infancy, and only started learning language when she was 6) gave an autobiographical account of how she was driven by blind impetuses until she learned the meanings of the words “I” and “me”
Possible nitpick depending on how you define consciousness: this excerpt sounds like Keller was conscious in the global workspace sense, but rather lacked something like a phenomenal self-model.
Nice post; I think I agree with most of it. Two points I want to make:
Or is this “qualitative difference” illusory, with the vast majority of human cognitive feats explainable as nothing more than a scaled-up version of the cognitive feats of lower animals?
This seems like a false dichotomy. We shouldn't think of scaling up as "free" from a complexity perspective - usually when scaling up, you need to make quite a few changes just to keep individual components working. This happens in software all the time: in general it's nontrivial to roll out the same service to 1000x users.
One possibility is that the first species that masters language, by virtue of being able to access intellectual superpowers inaccessible to other animals, has a high probability of becoming the dominant species extremely quickly.
I think this explanation makes sense, but it raises the further question of why we don't see other animal species with partial language competency. There may be an anthropic explanation here - i.e. that once one species gets a small amount of language ability, they always quickly master language and become the dominant species. But this seems unlikely: e.g. most birds have such severe brain size limitations that, while they could probably have 1% of human language, I doubt they could become dominant in anywhere near the same way we did.
There's some discussion of this point in Laland's book Darwin's Unfinished Symphony, which I recommend. He argues that the behaviour of deliberate teaching is uncommon amongst animals, and doesn't seem particularly correlated with intelligence - e.g. ants sometimes do it, whereas many apes don't. His explanation is that students from more intelligent species are easier to teach, but would also be more capable of picking up the behaviour by themselves without being taught. So there's not a monotonically increasing payoff to teaching as student intelligence increases - but humans are the exception (via a mechanism I can't remember; maybe due to prolonged infancy?), which is how language evolved. This solves the problem of trustworthiness in language evolution, since you could start off by only using language to teach kin.
A second argument he makes is that the returns from increasing fidelity of cultural transmission start off low, because the amount of degradation is exponential in the number of times a piece of information transmitted. Combined with the previous paragraph, this may explain why we don't see partial language in any other species, but I'm still fairly uncertain about this.
I think this explanation makes sense, but it raises the further question of why we don't see other animal species with partial language competency. There may be an anthropic explanation here - i.e. that once one species gets a small amount of language ability, they always quickly master language and become the dominant species. But this seems unlikely: e.g. most birds have such severe brain size limitations that, while they could probably have 1% of human language, I doubt they could become dominant in anywhere near the same way we did.
Can you elaborate more on what partial language competency would look like to you? (FWIW, my current best guess is on "once one species gets a small amount of language ability, they always quickly master language and become the dominant species", but I have a lot of uncertainty. I suppose this also depends a lot on what exactly what's meant by "language ability".)
A couple of intuitions:
"Once one species gets a small amount of language ability, they always quickly master language and become the dominant species" - this seems clearly false to me, because most species just don't have the potential to quickly become dominant. E.g. birds, small mammals, reptiles, short-lived species..
This seems like a false dichotomy. We shouldn't think of scaling up as "free" from a complexity perspective - usually when scaling up, you need to make quite a few changes just to keep individual components working. This happens in software all the time: in general it's nontrivial to roll out the same service to 1000x users.
I agree. But I also think there's an important sense in which this additional complexity is mundane -- if the only sorts of differences between a mouse brain and a human brain were the sorts of differences involved in scaling up a software service to 1000x users, I think it would be fair (although somewhat glib) to call a human brain a scaled-up mouse brain. I don't think this comparison would be fair if the sorts of differences were more like the sorts of differences involved in creating 1000 new software services.
I think whether the additional complexity is mundane or not depends on how you're producing the agent. Humans can scale up human-designed engineering products fairly easily, because we have a high-level understanding of how the components all fit together. But if you have a big neural net whose internal composition is mostly determined by the optimiser, then it's much less clear to me. There are some scaling operations which are conceptually very easy for humans, and also hard to do via gradient descent. As a simple example, in a big neural network where the left half is doing subcomputation X and the right half is doing subcomputation Y, it'd be very laborious for the optimiser to swap it so the left half is doing Y and the right half is doing X - since the optimiser can only change the network gradually, and after each gradient update the whole thing needs to still work. This may be true even if swapping X and Y is a crucial step towards scaling up the whole system, which will later allow much better performance.
In other words, we're biased towards thinking that scaling is "mundane" because human-designed systems scale easily (and to some extent, because evolution-designed systems also scale easily). It's not clear that AIs also have this property; there's a whole lot of retraining involved in going from a small network to a bigger network (and in fact usually the bigger network is trained from scratch rather than starting from a scaled-up version of the small one).
I HAVE TO plug Suzana Herculano-houzel's book "the human advantage" and all the research she and her group have done since it was published.
Short version: as you scale up a generic vertebrate brain, it only has 4x as many neurons every time it increases in mass by a factor of 10. (A factor of 16 after 100, etc). There are two clades that break this relationship: primates and birds, both of which have a 1:1 relationship between brain size and neuron number. The primate density is about the density of a mouse brain, and bird brains are 6x as dense as this. A large primate brain is much more impressive than a large generic-mammal brain. You get that elephants have brains about as impressive as chimps, and big whales have brains as impressive as late homo erectus - both of which I think make sense. Humans show up as having 4x the neurons of chimps/elephants and ~2x that of the biggest whales.
Human brains also appear to just be scaled-up primate brains, with the size of every brain region falling RIGHT on the trendlines of every other primate. I am skeptical of our subspecies being significantly smarter in any general way than other hominids with similar brain sizes.
I remembered reading about this a while back and updating on it, but I'd forgotten about it. I definitely think this is relevant, so I'm glad you mentioned it -- thanks!
Humans seem to have much higher degrees of consciousness and agency than other animals, and this may have emerged from our capacities for language. Helen Keller (who was deaf and blind since infancy, and only started learning language when she was 6) gave an autobiographical account of how she was driven by blind impetuses until she learned the meanings of the words “I” and “me”
This is fascinating, but there's a bit of a potential confounder in that she was six years old. I'm anecdotally aware of several people who feel they weren't really conscious before a certain age.
Hmm, I'm not sure I understand what point you think I was trying to make. The only case I was trying to make here was that much of our subjective experience which may appear uniquely human might stem from our langauge abilites, which seems consistent with Helen Keller undergoing a phase transition in her subjective experience upon learning a single abstract concept. I'm not getting what age has to do with this.
I just want to share another reason I find this n=1 anecdote so interesting -- I have a highly speculative inside view that the abstract concept of self provides a cognitive affordance for intertemporal coordination, resulting in a phase transition in agentiness only known to be accessible to humans.
+1. It feels like this argument is surprisingly prominent in the post given that it's a n=1 anecdote, with potential confounders as mentioned above.
I might be confused here, but it seems to me that it's easy to interpret the arguments in this post as evidence in the wrong direction.
I see the following three questions as relevant:
1. How much sets human brains apart from other brains?
2. How much does the thing that humans have and animals don't matter?
3. How much does better architecture matter for AI?
Questions #2 and #3 seem positively correlated – if the thing that humans have is important, it's evidence that architectural changes matter a lot. However, holding #2 constant, #1 and #3 seem negatively correlated – the less stuff there is that makes humans special, the smaller the improvements to architecture that are required to achieve greater performance.
Since this post is arguing primarily about #1, the way it affects #3 is potentially confusing.
Questions #2 and #3 seem positively correlated – if the thing that humans have is important, it's evidence that architectural changes matter a lot.
Not necessarily. For example, it may be that language ability is very important, but that most of the heavy lifting in our language ability comes from general learning abilities + having a culture that gives us good training data for learning language, rather than from architectural changes.
The most quintessentially human intellectual accomplishments (e.g. proving theorems, composing symphonies, going into space) were only made possible by culture post-agricultural revolution.
I'm guessing you mean the beginning of agriculture and not the Agricultural Revolution (18th century), which came much later than math and after Baroque music. But the wording is ambiguous.
It seems like "agricultural revolution" is used to mean both the beginning of agriculture ("First Agricultural Revolution") and the 18th century agricultural revolution ("Second Agricultural Revolution").
According to this SSC book review, "the secret of our success" is the ability to learn culture + the accumulation of said culture, which seems a bit broader than ability to learn language + language that you describe.
That's one of the "unique intellectual superpowers" that I think language confers us:
On a species level, our mastery of language enables intricate insights to accumulate over generations with high fidelity. Our ability to stand on the shoulders of giants is unique among animals, which is why our culture is unrivaled in its richness in sophistication.
(I do think it helps to explicitly name our ability to learn culture as something that sets us apart, and wish I'd made that more front-and-center.)
Chimpanzees, crows, and dolphins are capable of impressive feats of higher intelligence, and I don’t think there’s any particular reason to think that Neanderthals are capable of doing anything qualitatively more impressive
This seems like a pretty cursory treatment of what seems like quite a complicated and contentious subject. A few possible counterexamples jump to mind. These are just things I remember coming across when browsing cognitive science sources over the years.
My nonexpert sense is that it is at least controversial both how each of this is connected with language, and the extent to which nonhumans are capable of them.
they also escape feral children, and might escape animals for mundane reasons, like their not having critical periods long enough to learn these aspects of language.
On the other hand, humans having learned language at late ages (e.g. Helen Keller having learned it at 6) suggests that learning language during the critical period isn't a necessary requirement. (the Wikipedia links claims that deaf people who learn language at a late age never master it completely, but assuming that it hasn't been edited by others, the quoted Keller excerpt gives a rather different impression).
Humans are capable of feats of cognition that appear qualitatively more sophisticated than those of any other animals. Is this appearance of a qualitative difference indicative of human brains being essentially more complex than the brains of any other animal? Or is this “qualitative difference” illusory, with the vast majority of human cognitive feats explainable as nothing more than a scaled-up version of the cognitive feats of lower animals?
“How special are human brains among animal brains?” is one of the background variables in my framework for AGI timelines. My aim for this post is not to present a complete argument for some view on this variable, so much as it is to:
Humans seem to have much higher degrees of consciousness and agency than other animals, and this may have emerged from our capacities for language. Helen Keller (who was deaf and blind since infancy, and only started learning language when she was 6) gave an autobiographical account of how she was driven by blind impetuses until she learned the meanings of the words “I” and “me”:
Before my teacher came to me, I did not know that I am. I lived in a world that was a no-world. I cannot hope to describe adequately that unconscious, yet conscious time of nothingness. I did not know that I knew aught, or that I lived or acted or desired. I had neither will nor intellect. I was carried along to objects and acts by a certain blind natural impetus. I had a mind which caused me to feel anger, satisfaction, desire. These two facts led those about me to suppose that I willed and thought. I can remember all this, not because I knew that it was so, but because I have tactual memory. It enables me to remember that I never contracted my forehead in the act of thinking. I never viewed anything beforehand or chose it. I also recall tactually the fact that never in a start of the body or a heart-beat did I feel that I loved or cared for anything. My inner life, then, was a blank without past, present, or future, without hope or anticipation, without wonder or joy or faith.
[...]
… When I learned the meaning of "I" and "me" and found that I was something, I began to think. Then consciousness first existed for me. Thus it was not the sense of touch that brought me knowledge. It was the awakening of my soul that first rendered my senses their value, their cognizance of objects, names, qualities, and properties. Thought made me conscious of love, joy, and all the emotions. I was eager to know, then to understand, afterward to reflect on what I knew and understood, and the blind impetus, which had before driven me hither and thither at the dictates of my sensations, vanished forever.
I think humans underwent a phase transition in their intellectual abilities when they came to master language, at which point their intellectual abilities jumped far beyond those of other animals on both an individual level and a species level.
On an individual level, our capacity for language enables us to entertain and express arbitrarily complex thoughts, which appears to be an ability unique to humans. In theoretical linguistics, this is referred to as “digital infinity”, or “the infinite use of finite means”.
On a species level, our mastery of language enables intricate insights to accumulate over generations with high fidelity. Our ability to stand on the shoulders of giants is unique among animals, which is why our culture is unrivaled in its richness in sophistication.
The most quintessentially human intellectual accomplishments (e.g. proving theorems, composing symphonies, going into space) were only made possible by culture post-agricultural revolution. So, when evaluating humans’ innate intellectual capacities, a better reference point than modern humans like ourselves would be our hunter-gatherer ancestors.
We can reduce the question of how complex our hunter-gatherer ancestors’ brains are into two sub-questions: how complex is our capacity for mastering language, and how complex are brains that are similar to ours, but don’t have the capacity for mastering language?
Neanderthal brains seem like plausible proxies for the latter. Neanderthals are similar enough to modern humans that they’ve interbred, and the currently available evidence suggests that they may not have mastered language in the same way that behaviorally modern humans have. (I don’t think this evidence is very strong, but this doesn’t matter for my purposes—I’m just using Neanderthals as a handy stand-in to gesture at what a human-like intelligence might look like if it didn’t have the capacity for language.)
Chimpanzees, crows, and dolphins are capable of impressive feats of higher intelligence, and I don’t think there’s any particular reason to think that Neanderthals are capable of doing anything qualitatively more impressive. I’ll share some examples of these animals’ intellectual feats that I found particularly illustrative.
Chimpanzees have been observed to lie to each other under experimental conditions. From Wikipedia:
...food was hidden and only one individual, named Belle, in a group of chimpanzees was informed of the location. Belle was eager to lead the group to the food but when one chimpanzee, named Rock, began to refuse to share the food, Belle changed her behaviour. She began to sit on the food until Rock was far away, then she would uncover it quickly and eat it. Rock figured this out though and began to push her out of the way and take the food from under her. Belle then sat farther and farther away waiting for Rock to look away before she moved towards the food. In an attempt to speed the process up, Rock looked away until Belle began to run for the food. On several occasions he would even walk away, acting disinterested, and then suddenly spin around and run towards Belle just as she uncovered the food.
In Aesop’s fable of the crow and the pitcher, a thirsty crow figures out that it can drop pebbles into a pitcher, so that the water rises to a high enough level for it to drink from. This behavior has been experimentally replicated, indicating that crows have a “sophisticated, but incomplete, understanding of the causal properties of displacement, rivalling that of 5–7 year old children”.
When Kelly the dolphin was given rewards of fish for picking up scraps of paper, “Kelly figured out that she received the same fish regardless of the size of the piece of trash she was delivering to her trainer. So she began hiding big pieces of trash under a rock. Kelly would then rip off small pieces from the trash and deliver them one at a time so that she could receive more fish.” Additionally, “when a bird landed in the pool, Kelly snatched it and delivered it to her trainers. She received a large amount of fish in return. Knowing this, she decided to start hiding fish each time she was fed. She would then use the fish to lure birds when none of her trainers were around. Kelly knew that by saving one or two fish now, she could get many more fish later by turning in a bird.“ (Also reported on The Guardian; I don’t know how reputable these sources are, so take this anecdote with a grain of salt.)
See these Wikipedia pages for some more interesting examples, and see here for a more thorough review of the evidence of higher intelligence in animals.
Many aspects of human cognition that may appear qualitatively different from what other animals are capable of, such as long chains of abstract reasoning, also appear qualitatively different from what less intelligent humans are capable of. As a particularly extreme example, John von Neumann’s cognitive abilities were so advanced that a Nobel Laureate, Hans Bethe, once remarked that "[his] brain indicated a new species, an evolution beyond man".
At the same time, the genes that code for different humans’ brains are virtually identical from an evolutionary perspective. This suggests that the seemingly qualitative differences between humans’ and animals’ cognition might not be so different from the seemingly qualitative differences between John von Neumann’s cognition and mine—our brains might be doing essentially the same thing as theirs, except at a higher scale.
Maybe it was extraordinarily difficult to evolve the cognitive mechanisms that allow us to learn language, above and beyond our cognitive machinery for learning other things. I think this is plausible, but I don’t think the case for this is very strong.
Animals (Washoe, Koko, and Alex the parrot) have demonstrated the ability to learn simple forms of symbolic communication, which they never evolved to do, indicating that their ability to learn things in general is good enough to learn very simple forms of language. It’s true that there are aspects of human language that escape animals, but they also escape feral children, and might escape animals for mundane reasons, like their not having critical periods long enough to learn these aspects of language.
Additionally, AI language models provide evidence that simple and general learning mechanisms can capture many of the intricacies of human language that other animals miss, further suggesting that there’s nothing intrinsically difficult about learning language. Here’s an excerpt from GPT-2, a relatively recent language model:
SYSTEM PROMPT (HUMAN-WRITTEN)
In a shocking finding, scientist discovered a herd of unicorns living in a remote, previously unexplored valley, in the Andes Mountains. Even more surprising to the researchers was the fact that the unicorns spoke perfect English.
MODEL COMPLETION (MACHINE-WRITTEN, 10 TRIES)
The scientist named the population, after their distinctive horn, Ovid’s Unicorn. These four-horned, silver-white unicorns were previously unknown to science.
Now, after almost two centuries, the mystery of what sparked this odd phenomenon is finally solved.
Dr. Jorge Pérez, an evolutionary biologist from the University of La Paz, and several companions, were exploring the Andes Mountains when they found a small valley, with no other animals or humans. Pérez noticed that the valley had what appeared to be a natural fountain, surrounded by two peaks of rock and silver snow.
Pérez and the others then ventured further into the valley. “By the time we reached the top of one peak, the water looked blue, with some crystals on top,” said Pérez.
If language isn’t a particularly difficult cognitive capacity to acquire, why don’t we see more animal species with language?
One possibility is that the first species that masters language, by virtue of being able to access intellectual superpowers inaccessible to other animals, has a high probability of becoming the dominant species extremely quickly. (Humans underwent the agricultural revolution within 50,000 years of behavioral modernity—a blink of an eye on evolutionary timescales—after which their dominance as a species became unquestionable.) Since we shouldn’t expect to see more than one dominant species at a time, this would imply a simple anthropic argument for our unique capacities for language: we shouldn’t expect to see more than one species at a time with mastery of language, and we just happen to be the species that made it there first.
It may also turn out that language is hard to evolve not because it’s a particularly sophisticated cognitive mechanism, but because the environments that could have supported language and selected for it might have been very unique. For example, it may be that a threshold of general intelligence has to be crossed before it’s viable for a species to acquire language, and that humans are the only species to have crossed this threshold. (Humans do have the highest cortical information processing capacity among mammals.)
It might also turn out that the cultural contexts under which language could evolve require a mysteriously high degree of trust: “... language presupposes relatively high levels of mutual trust in order to become established over time as an evolutionarily stable strategy. This stability is born of a longstanding mutual trust and is what grants language its authority. A theory of the origins of language must therefore explain why humans could begin trusting cheap signals in ways that other animals apparently cannot (see signalling theory).”
As we came to master language, I think we underwent a phase transition in our intellectual abilities that set us apart from other animals. Besides language, I don't see much that sets us apart from other animals—in particular, most other cognitive differences seem explainable as consequences of either language or scale, and I don’t think the cognitive mechanisms that allow us to master language are particularly unique or difficult to acquire. Overall, I don’t see much reason to believe that human brains have significantly more innate complexity than the brains of other animals.
Thanks to Paul Kreiner and Stag Lynn for helpful commentary and feedback.