Chimpanzees have diverse vocal sequences with ordered recombinatorial properties | Hacker News Reader

Chimpanzees have diverse vocal sequences with ordered recombinatorial properties

researchgate.net

111 pointspueblito52 comments

Just don't call it "language" or the human exceptionalists will get very upset.

> Just don't call it "language" or the human exceptionalists will get very upset

We don't know what "language" is. This is an attempt to empirically construct that definition from the bottom up. Calling it "language" may saddle whatever's going on here with needless or even misleading assumptions.

Language is a word we invented to describe a certain category of things. It's not some big mystery, there's not some objective, universal meaning that we have yet to discover.

By the broadest definitions, chimpanzees have a language, and maybe even multiple dialects. By the narrowest definitions, only humans have languages.

People behave like there's some profound truth here, but this is really just semantics. I don't see why it's such a touchy subject, other than the fact that people like to believe humans are a special case.

Cultural value inherited from the past, mankind holding dominion over all others on earth.

Chimps have culture: https://www.nature.com/articles/050815-12

As for "dominion over the earth," we're in the process right now of finding out exactly how fragile our hold over the planet is. I would not be asserting "dominion over the earth" as some kind of virtue at this point.

I didn't assert it was a virtue. I asserted that is why, and why we continue to assume human exceptionalism.

In a different comment, I was talking about dogs, birds, cephelopods being more intelligent than we typically give them credit for.

>Cultural value inherited from the past, mankind holding dominion over all others on earth.

That's not just some "cultural value". That's an actual capacity. If sharks, tigers, or monkeys could lord over us, they would have. Like they do for species they have the upper hand on.

That’s not a capacity so much as a worldview.

A different worldview is that humans are a steward of a land, and the land has abundance for all who inhabits it.

I do not have any strong objection to it being called language, so long as that does not slide into "...just like we do." I take it as quite clear that the language of chimpanzees is qualitatively less powerful, in terms of its ability to express and communicate ideas, than that of humans - their behavior amongst themselves, and in their interactions with humans, would surely be very different if it were (if they have language and they understand language, why are they not using it in situations where it would be clear, to any individual with that knowledge, how useful it would be?)

On the other hand, I feel it is highly likely that other, now extinct, homo species had human-like language skills. Language evolved, and it is not very plausible that it did so in one leap from that which we see in other species today. Some level of language and other cognitive skills are to be expected in other species, and the study of those skills is both worthwhile in itself and as a guide to better understanding ourselves.

>why are they not using it in situations where it would be clear, to any individual with that knowledge, how useful it would be?

Who is to say it is not?

By analogy, if you are laying under your vehicle, loosening a bolt with a wrench, and your dog is hunched nearby watching, he can clearly see what you are doing (and quite possibly understands that the wrench is needed to loosen the bolt) but he cannot understand why you are doing that, or what 19mm means, nor express either. The dog understands leverage in terms of his world, but quite possibly not that the wrench is used to create it in a different way. Certainly he couldn't explain it to another dog.

The other apes are doing things we can see and hear, but do not understand or put any importance on. One thing I note is that there is no mention of body language in that essay. Posture and gesture (and situation) add a huge amount of complexity to language.

And apes are surely using that too.

For example, I can tell you "no" but with minimal deft gestures, convey understanding in you that I mean "maybe" or even outright "yes". The canonical example being Linus Torvalds saying "No" to whether he was asked¹ to add back doors to the Linux Kernel, while indicating non verbally and quite clearly that he meant yes.

You and I have "that knowledge" to understand the whole message, while we quite likely lack nuance that other apes have in their communications.

¹and we both understand "asked" to actually mean "pressured" to some degree, and told not to tell. And there was more than duality in Linus' answer, he was also saying "Don't trust me to not fall to pressure" and "If I am pressured, they shouldn't trust me to keep it a secret".

Apes are surprising in their sophistication.

Jokes on them I think the proteins and electrical signals through which multi-cellular organisms interact is a language.

Is it bad to think humans are exceptional? It seems when you look at the world that we stand out quite a bit. You could argue the only difference between humans and certain animals is that of degree (of cognitive ability, consciousness, intellect, etc) but it still seems like humans stand out, there is a large degree of difference especially in the effect.

I mean vocalizations aren't language. My baby makes all sorts of vocalizations, but it's not language. And I'm not just saying that because I'm an adult exceptionalist

Do the vocalizations include a diverse set of patterns they use to communicate with other babies? If the answer is "no" then we're not really talking about the same thing.

To somewhat overstate this paper, it shows chimpanzees have their own language and communicate vocally using words

Do you know of any projects using automated transcription and machine learning to analyze hominid speech? I ask because I remember reading about google translate effectively learning how to translate languages purely through machine learning and a large corpus of transcribed speech.

Sure. Do you have a large corpus of chimp to human?

And it's not like both of these sides don't have an almost endless variety of languages...

What we need is a large corpus of vocal recordings along with contextual video recording. Then train a model to associate vocalizations with states of the video by predicting subsequent video frames. By throwing enough compute at the problem we could probably come to understand any animal language with such a technique.

What if chimps mainly gossip? How's that going to show up in the videos?

I expect there's a hierarchy in language acquisition, where describing interactions with the world come before more abstract talk such as gossip. But once this first-order descriptive talk is understood, this can be leveraged to capture the more abstract talk.

Looking at correlations between vocalizations and behavior is one thing primate researchers have been doing for decades. It seems implausible for there to be a large skill set that has so far escaped their notice.

Whenever one group of humans has met another, it has rapidly become apparent to both parties that the other has a language, no matter how long the ancestors of the two groups have been isolated from one another. Given how long humans and chimpanzees have been coexisting, if they had comparably effective language skills, it is implausible that either party has not recognized it in the other.

I don't think it's implausible. I think it's more implausible that human language is unique

The view I set out here is neither predicated on nor entails the position that human language is unique. In fact, I happen to believe that there have been several species on Earth with human-like language, it's just that the others are now extinct. I also think it far more likely than not that there is language-using life elsewhere in the universe.

I think its been conclusively demonstrated that deep learning can discover correlations that go unnoticed by humans. An example that immediately comes to mind is AI detecting race from X-rays. There are many others. There's also many other factors that influence our ability to ascribe language skills to other groups of humans, from complexity of vocalizations to similar ways to interact with the environment and with other group members. We recognize ourselves in fellow humans, which makes it easy to ascribe similar capacities. The question is how easily can we recognize language in a species where we do not immediately recognize ourselves? The more foreign the behavior and the patterns of communication, the harder it gets. AI has the ability to elide over those difficulties.

I do not doubt that deep learning can discover correlations that go unnoticed by humans, and if our failure to detect the level of language skills we possess in chimpanzees (and the failure of chimpanzees to detect those skills in humans!) is due to the sort of difficulties you raise, then I agree that machine learning could find the evidence that we have not. What I do doubt, however, is that these difficulties actually exist to the level required for us to miss what is going on.

I believe that the theory of evolution is correct, and therefore that the evolution of language confers a differential fitness, yet chimpanzees are clearly not behaving in a way that exploits anything like the full capabilities of a human-like language, either in interactions between the members of their own species, or with their adversaries such as leopards or humans. At any time in the past, a small cabal of language-wielding chimpanzees, who grasped something more of the capabilities of their language skills than did their neighbors, could have dominated the remainder, increasing both their fitness and that of their species. In practice, we do not even observe them chatting.

It was David Brin, I believe, who made the same point with respect to dolphins: if they have these language skills, how come the knowledge of the dangers presented by purse seine nets, and the means to avoid them, has not spread within those dolphin communities where these nets are a significant cause of mortality?

There are many cases in evolution where a feature evolved for one purpose but then served another (which is how, it is supposed, birds got their flight feathers) - but even if that was the case for chimpanzee language, what was that primary purpose and why have the abilities not been adopted for their secondary purpose - especially as, once you have that level of language skills, change occurs at the rate of meme spread, rather than that of genes.

In other words, the combination of human-like language skills and the chimpanzee lifestyle seems very far from equilibrium. Language, by its very nature, does not deliver its evolutionary benefits cryptically.

> AI detecting race from X-rays

Anthropologists have been doing that from skeletons for centuries.

Hmm. Do you think you could do the same with human languages?

I think so. It's similar to how an adult can learn a new language by way of immersion. Correlations and gesturing go a long way.

It seems a little more than that even. I mean, I admit I couldn't quite follow all of it, but it seems to be suggesting that chimps can to a degree create reorderings of their 'words' to make new meanings, in a sort of proto-syntax. After all, if human language is expressive because our abilities to use syntax and grammar to make complex meaning. Those things had to be built on something that came before.

It's interesting because I've heard that one of the benefits Sanskrit poets had in their ability to introduce complex structural patterns in their poetry, is that the meaning of sentences is largely independent of their word-order - meaning the lines can be reordered to freely introduce rhyming schemes etc. Yet the language has many grammatical elements; gendered word endings and so on.

This kind of relaxed word order is also typical for modern Baltic and Slavic languages. Words themselves have enough information encoded in word prefixes/suffixes/endings so that it's possible to decode even a randomly mixed up sentence. It will feel awkward, or stylistically wrong, but nevertheless understandable.

This might help with poetry but comes at a price: it is superhard to internalise all the numerous word forms. I can't image learning Polish, Lithuanian or Russian being only exposed to English previously!

It is not superhard. It’s not even hard. Every native speaker learned it easily ;)

> It will feel awkward, or stylistically wrong.

No it won't, because it's understood to be poetic.

In English, it's a awkward if an adjective is put after a noun right? But you don't bat an eyelash if it's in a poem.

"High upon the chimney stack, there I saw perched three crows black."

(Don't search for that, I just made it up.)

Moreover, in languages with case, reordering doesn't cause any ambiguities or confusion. You know which word is the subject and which the object in any permutation. (Not necessarily for all words, but most.) The speakers already enjoy considerable reordering freedom in everyday sentences already (non-poetic) where it plays roles in emphasis and such.

I actually disagree with your example. Taken ultra-literally, it could mean that 'I' was 'perched' and 'saw three black crows'. So I feel that 'crows black' was fine and French places adjectives after the word as standard. I have to rely on context to interpret the sentence, and that creates an iota of doubt which makes it read clumsily, or me read it clumsily. I think it's very nuanced and personal.

Languages that use complex systems of conjugation and declension don't need to rely on word order as much to denote relationships between the words. In normal speaking, subject-verb-object order might still be used, but in writing, there's a lot of flexibility.

what your saying is true of most highly-inflected languages

Are they words or phonemes?

Somewhere in between? It seems that individual sounds are perceived by the chimps as meaningful signals, and that the combined bigrams and trigrams have additional meanings. Not sure it can be analogized to human speech more specifically than that.

It's been shown that border collies are able to understand novel reordering to a limited degree. They are supposed to be as smart as five-year olds. And watching my 18-month old son and similar toddlers interact in the world, such reordering is something is learned well before five years old.

Compare to parrot and corvids. Birds have evolved a brain more efficient than humans. Crows are able to work with abstractions, and pound-for-pound use less brain matter.

And then there's the octopus and other cephelopods.

After spending a day with Chimpanzees where they climbed all over me, played in the jungle and basically did their thing in their group, it was blatantly obvious to see the intelligence in their eyes. [1]

I have no doubt they have a language with each other, we're just not smart enough to figure it out yet.

[1] https://www.youtube.com/watch?v=Mfdo3s8tPUk

> Focusing on the structure of vocal sequences, we analysed 4826 recordings of 46 wild adult chimpanzees from Taï National Park.

If this same analysis were conducted with other populations of chimpanzees as well, could it be determined whether different populations use different languages? I believe this sort of thing has been tried with Orcas, with the conclusion that different populations have different languages.

I think them using different languages should be the assumption. No two individuals ever have an exact same accent so over time languages change and chimps don’t have large macro-societies (states) to standardize language.

Does that mean we can use one of those AI language models and develop a way to translate what the chimpanzees are communicating to each other?

that sounds like a dope idea..

The AI language models are based on reference translations that cannot exist in this case unless you find a chimp that knows English and can translate all of the recordings from that study. If you repeat that enough times eventually you will get a model that can do a translation without any chimp linguists.

What would be possible in a thoroughly surveilled group is to do a detailed analysis and record of the behavior and interactions and dialogues etc., possibly even eye gaze direction, and it's exact timing, and then create some type of AI model based on all of that together.

Has similar research been done on bird song? It seems to me that bird song could also be described as ordered recombinatorial sequences, for instance the mockingbird takes the sounds of other birds and strings them together to create new songs.

https://news.ycombinator.com/item?id=31233473

I'm curious about mockingbirds as well, they appear to have a separate set of vocalizations for interacting directly with each other in addition to their imitated songs. I always assumed the mocking was to claim territory as every species at once, but it seems to not have a well established explanation.

Has anyone tried training a language model on animal vocalizations?

Not a language model, but there was this classifier mapping pig squeals to contextual information about the animal [1].

For human language models there are obvious practical uses and we can intuitively evaluate the quality of the models. For animal communication you need to use other biological variables, like valence in the pig communication example.

I do remember one where they generated artificial frog mating calls and were able to generate artificial ones that were exceptionally attractive to female frogs, but that was from a while back, and I can’t find the link.

[1] https://www.nature.com/articles/s41598-022-07174-8

There was a post here on HN a few months back about a group of scientists that did exactly that to whale recordings, and actually found some surprising patterns emerge.

edit: I think this is the one I was thinking of: https://news.ycombinator.com/item?id=26874309

There was some research on prairie dog language as well:

https://medium.com/health-and-biological-research-news/prair...

https://www.npr.org/2011/01/20/132650631/new-language-discov...

Original paper: sciencedirect.com/science/article/abs/pii/S0003347205801174?via%3Dihub

Reply on news.ycombinator.com