Comment by ksymph
9 days ago
It seems like the lesson we keep learning, no matter the proxy we use for intelligence, is that there is nothing that fundamentally sets humans apart from other animals (or even, in some ways, AI) other than the degree and scope of our intelligence.
While I'll never begrudge science that points out the obvious -- that's often where the most value comes from -- this particular avenue is always a little funny to me, as it often belies an expectation that other animals are unable to do these things by default.
> It seems like the lesson we keep learning, no matter the proxy we use for intelligence, is that there is nothing that fundamentally sets humans apart from other animals
Except it doesn't show that.
The reason people make this judgement is because they don't have a coherent or clear definition of "intelligence". Nothing has been undermined, except in those who took the view that animals are dumb automatons. That's more of a legacy of modernism and the desire to gain "mastery over nature" more than anything else.
The essential feature of human beings - from which the rest of human nature and its consequences follow, including our social nature - is rationality. This entails an intellect, which is the abstracting faculty. It is the intellect that makes language possible, because without the capacity to abstract from particulars, we could not have universal concepts and thus no predicates. Language would be reduced to the kind we see in other animals.
For clarity, the functions of language are:
1. expressive: expressing an internal state or emotion (e.g., a cry of pain)
2. signaling: use of expressive to cause a reaction in others (e.g., danger signals)
3. descriptive: beyond immediate sensation; describes states of affairs, allowing for true or false statements
4. argumentative: allows critical analysis, inference, and rational justification
Without abstraction, (3) and (4) are impossible. But all animal activity we have observed requires no appeal to (3) and (4). Non-human animals perceive objects and can manipulate them, even in very clever ways, but they do not have concepts (which are expressed as general names).
Could there be other rational animals in the universe? Sure. But we haven't met any. And from an ontological POV (as opposed to a phylogenetic taxonomic classification), they would be human, as the ontological definition of "human being" - "rational animal" - would apply them.
Not sure if it qualifies but even bees have a dance that describes the direction, distance and quantity of nectar.
https://en.wikipedia.org/wiki/Waggle_dance
Feels like a lot of animals just lack ability to articulate. Which might evolve if they had a need but feels like an chicken-and-egg problem more than anything?
Perhaps ironically, I am having trouble distilling your abstractions into concrete concepts.
A dog or chimpanzee can easily understand conceptual ideas such as 'walk', 'play', 'food', and so on, even through language. Not to say humans don't process these in different ways, and are able to manipulate them as abstract concepts as other species generally cannot, but in isolation it seems the fundamental principles can be widely accessed. What sort of test might you propose that demonstrates the difference you describe?
Isn't language also dependent upon unique human physical features enabling sophisticated sound combinations i.e. speech?
Birds use sophisticated sounds for communication.
Not long ago there was the story of an orangutan in the wild using a paste it made from plants to heal itself, like some sort of anti-bacterial/anti-inflammatory.
https://www.bbc.com/news/science-environment-68942123
In some areas, we are still like the geocentricly minded from long ago.
I feel that language models are revealing things about our cognition that some people are as unwilling to accept, similarly to how some people had trouble with accepting heliocentrism.
What things are being revealed about our cognition by language models?
> I'll never begrudge science that points out the obvious
People have many beliefs, inaccurate to varying degrees - many to a large degree. Science is a solution to our pretty bad intuitions. Sometimes it discovers they are wrong; sometimes science shows they are correct - it's hindsight to say it was 'obvious'.
Also, I don't think it's obvious to 99.x% that cows use tools.
The primal separation of man and animals has not changed since the invention of the device...
Animals fear motorized vacuum cleaners.
I fear a motorcycle blasting down my street at 10pm. What's the difference. Once my cats realized the robo vac won't hurt them they don't even move for it anymore... Seems intelligent to initially be terrified of something and update your perception of it.
Nature abhors a vacuum
Incorrect. Humans have free will (= information complexity). Various aspects of intelligence are just metrics of that, but the metric isn't the thing itself.
[dead]