Hacker News new | past | comments | ask | show | jobs | submit login
Recursive language and modern imagination acquired 70k years ago: hypothesis (phys.org)
120 points by JackFr 76 days ago | hide | past | web | favorite | 83 comments

Its hard for humans today to visualize higher dimensional physical spaces, even though after study, we can easily reason about them mathematically.

I wonder if that's analogous to the barrier that was crossed 70k years ago according to the theory in the article.

However, we routinely visualize high dimensional information spaces without even realizing it (though probably through some heavy dimensionality reduction). It's almost like we have the hardware for high dimensional reasoning, and some software to apply it to physical dimensions (the language of math), but no drivers to accelerate the latter with the former.

I don't think we visualize high dimensional information spaces. We have tools for dealing with them, but of a completely different nature.

I understand visual perceptions and descriptions to be literal mappings of high-dimensional information spaces.

Smooth vs rough vs rippled vs grainy, fuzzy vs hairy vs spiky, striped vs mottled, etc. We get a lot of understanding of a material’s properties by looking at it, coupled with prior experience with similar-looking ones.

There’s even more richness involved when you take human cultures and other senses into account. I’d be surprised if I tried to turn what looks like a doorknob and it didn’t wiggle even a little; every locked one I’ve met does. I’d be shocked if my hand passed right through it instead. And when applied to other people, my God! It’s usually easy to tell when someone is angry or bored, often even if it’s aimed at us vs some third party—how many dimensions does that encapsulate?

Many animals are clearly capable of this too. The language appears to be the leap. It makes me wonder even more about crows talking to each other about which humans feed them, and what a shotgun looks like and can do.

Yeah, we already live in n-space. 3D is a convenient fiction that models reality in very specific instances.

I should have said we humans process high dimensional information, like the 5 conscious senses, plus subconscious sensory input, and symbolic representations, like language and images.

This Rio Journal seems to pride itself on letting you publish non traditional papers. In this case, we have something that seems clearly marked as a hypothesis. I won’t comment on the value of such an approach. But I will say that it’s frustrating that such a publisher would take a speculative article like that, and pick a title like “recursive language and modern imagination were acquired 70k years ago”

In science we only have hypotheses -- disproven and as-yet-not-disproven. Granted, some hypotheses are less supported by data, often due to lack of data, and this one might be such a hypothesis, but, in science rejecting an idea as a hypothesis is oxymoronic.

yeah but presenting a hypothesis as fact (as the title does) is misleading at best. something like "A new hypothesis re: the development of recursive language" would be better

The second paragraph of TFA starts with "This new hypothesis" and goes from there. Titles are titles -- they've been clickbait since before we had mice to click on links with.

Every newspaper headline ever, is a form of clickbait.

Titles are usually chosen by the author, not the publisher.

To suggest that modern imagination and the recursive thinking necessary for language evolved _after_ humans started making complex tools (which required hierarchical recursive processes to make) is, to me, dubious. The tool-creation/language similarity was commented upon recently [0].

[0] https://royalsocietypublishing.org/doi/full/10.1098/rstb.201...

The article mentions one such tool - needles with eyes. As this innovation is mentioned, I imagine the time of its invention is consistent with the hypothesis. Do you have in mind specific tools from before 70kya that seem to require recursive thinking?

I am not sure what would constitute recursion in tools - maybe something like an atlatl, sling or bow, where a tool is applied to another tool? These particular examples do not seem to challenge the timing of the hypothesis.

Other species have the ability for tool making, some of it relatively complex. So clearly there are other factors. Language is obviously one of them. The differentiator described here is PFS secondary to a genetic (evolutionary) change in the prefrontal cortex. Tool making to me implies sequential processes, not the juxtaposition of mental objects. I'm not clear on what you mean by toolmaking as a recursive process.

A problem that this hypothesis has to contend with is that there are populations of humans (Khoi san and African pygmies) who split off from the rest of humanity > 70k years ago, yet who seem pretty modern.

The mutation allowed certain individuals to invent recursive language, which could then be taught to individuals without the mutation during a critical period in childhood.

What could be more "modern" in this case than existing at the present moment?

From Wikipedia:

As an analytical concept and normative ideal, modernity is closely linked to the ethos of philosophical and aesthetic modernism; political and intellectual currents that intersect with the Enlightenment; and subsequent developments such as existentialism, modern art, the formal establishment of social science, and contemporaneous antithetical developments such as Marxism. It also encompasses the social relations associated with the rise of capitalism, and shifts in attitudes associated with secularisation and post-industrial life (Berman 2010, 15–36).

I meant modern in the sense that the person you're responding to was using it. I'm not sure what their point was; as far as I can tell they were agreeing with me.

The Khoi San diverged from the rest of humanity roughly 200k years ago. Yet they are basically compatibly with modern humans. As far as I am aware, they have recursive language and normal imaginations. My point was that this theory would have to deal with the fact that human populations who split much, much longer ago than 70k years ago appear to have the same basic brain functionality as the rest of humanity.

I guess, based on what someone else said, the theory is positing that people with the mutation could bootstrap the process in people who don't have it. I dunno, in light of the Khoi San, things start to diverge pretty far from what I would describe as a parsimonious explanation. Plus we're talking hypothetically about what was going on in the minds of people who lived 70k years ago. There are a lot of theories that could be correct, since we have literally no direct evidence of the presence of recursive language (or the lack thereof).

The Khoi San are human and just as "modern" as us, given that we all exist right now.

Side note: you're being pedantic. When discussing deep human history, 'modern human' has a meaning beyond just the dictionary meaning of the word 'modern' (being of a recent vintage). It is used to contrast modern humans with our quite complex behavior vs. archaic humans with their (apparently) much less complex behavior.

But anyway, my point doesn't depend on how you define 'modern human', since not only do the Khoi San co-exist with the rest of humanity in the current moment, but they are not really significantly different from the rest of humanity in a behavioral sense. So by both definitions they are 'modern humans'. Which gets to my actual point: a theory which posits that behaviorally modern humans have their genesis only 70k years ago has some explaining to do in light of the fact that the Khoi San split from the rest of humanity about 200k years ago, yet are more or less the same as the rest of humanity.

>It is used to contrast modern humans with our quite complex behavior vs. archaic humans with their (apparently) much less complex behavior.

Yes, I know - that was my point. You seem to be assuming that we in the West are at the apex of development while the Khoi San are to be evaluated by the degree to which their society resembles Western societies. Who knows - perhaps the Khoi San think that they are more modern than us and are waiting for us to catch up.

We all have ancestors going back to the first humans. No present human society is more "archaic" than any other, unless you think that human societies are somehow predestined to develop along a certain course.

So I do object to calling the Khoi San "pretty modern", as if they are somehow being stick-in-the-muds by not developing along exactly the same path as the West.

> You seem to be assuming that we in the West are at the apex of development while the Khoi San are to be evaluated by the degree to which their society resembles Western societies.

No, I am not. I have no idea how you came to that conclusion. On the contrary, my whole argument depends on the Khoi San not being significantly different, behaviorally, from the rest of humanity.

Here are quotes of what I said. The only reason I didn't use even stronger language is that there are, in fact, significant genetic differences between the Khoi San and the rest of humanity (that's how we know they diverged so long ago). But anyway, the quotes:

> (Khoi san and African pygmies) [...] who seem pretty modern.

Edit in reply to your edit: "pretty" in a colloquial sense means "very". As in, "it's pretty hot out today" = "it's very hot out today". Even the dictionary definition is "to a moderately high degree".

> Yet they are basically compatibly with modern humans. As far as I am aware, they have recursive language and normal imaginations.

> Appear to have the same basic brain functionality as the rest of humanity.

> they are not really significantly different from the rest of humanity in a behavioral sense

> are more or less the same as the rest of humanity

> The only reason I didn't use even stronger language is that there are, in fact, significant genetic differences between the Khoi San and the rest of humanity (that's how we know they diverged so long ago)

We have no evidence that the differences are particularly significant to everyday life. But in any case, your statement here makes no sense. The fact that the Khoi San are (to whatever minuscule extent) genetically different from “us” does not make them any less modern than us. Divergence is symmetrical: as we diverged from the Khoi San, they diverged from us, and there’s no sense in which they’re “older” than we are. So I still don’t understand why you are only willing to credit them with being modern “to a moderately high degree”. There’s nothing to qualify. They’re humans just like we are.

A couple of things:

1. You are not interpreting the use of the modifier 'pretty' correctly. It is used for emphasis or to add or increase some property; to imply that something is more than usual. To say that 'it is pretty hot today' means that it hotter than if you merely said 'it is hot today'. Thus, when I say 'pretty modern', I am indicating a degree of modernity that is more modern than simply 'modern'. As far as I can tell, you are interpreting 'pretty' to mean that something is lacking in a quality or falls slightly short of possessing a quality. But that is the opposite meaning of the word. If I were trying to detract from their modernity, I would have said that they are 'kind of modern' or 'somewhat modern'.

2. The bigger picture, though, is that you're focusing on a single word choice instead of literally everything else I've written, including the crucial point that I have to agree with you for the point that I was actually making to hold true.

That is not what 'pretty' means, FYI (check the dictionary definition, which in the case of the OED gives "somewhat" as a synonym). It's sometimes used to convey a stronger proposition via understatement, but that doesn't come across in your original post. Nor does it jibe with the rest of what you said (e.g. that the Khoi San are "more or less" the same as the rest of humanity).

I'm glad we agree on the rest.

So if it happens as in the synopsis, it would have been pretty weird to grow up as one of the first kids with recursive language? And I guess the gene takes over the world because well, you want to mate with other people who understand prepositions.

The article goes on to mention that it likely developed gradually over several generations. The first kids with the mutation would have never been taught recursive language by adults. Instead, they probably developed some semblance of it from speaking with siblings/cousins their age. While this would be pretty weak, it probably gave them enough of an advantage over others.

Once they themselves had kids they would be able to teach them their imperfect version of recursive language. This probably continued and amplified over generations.

It's a pretty interesting case of nature and nurture combining. The mutation is a prerequisite, but largely useless unless there's enough "nurture" happening during the critical period.

As the article mentions it's only 70k years ago that we start seeing unambiguous indicators of intelligence. The normal figure for this indicator, the upper paleolithic revolution, is 50k years. In any case that's where we first start seeing significant and then rapidly widespread signs of intelligence - fishing, extensive adornment, wide artifact collections, burial ceremony, and much more.

The point here is that the archaeological record seems to indicate that it's like one day we simply 'woke up.' Whomever was fortunate enough to experience such a mutation first would have had a tremendous competitive advantage. The degree in intellectual advancement is such that they would have probably been significantly better at just about everything -- most importantly in surviving and spreading their genes. Now take this person and what would likely have been their vast number of offspring. They could effectively take over the world. And I suppose, in a manner of speaking, they did.

> the archaeological record seems to indicate that it's like one day we simply 'woke up.'

Either that or the kinds of artifacts we look for simply don’t last longer than 70k years on a planet like this.

Or, it’s just rare enough that, when combined with the cost of excavation and the fact that scholarship builds on scholarship, it’s just not the kind of evidence scholars are going to dig far past we’ll established boundaries to find.

Lots of possible explanations for the data.

It is entirely possible the capacity for this existed but was not being used. That would explain why it spread so rapidly... perhaps most humans already had the capacity?

I think it's implied that the capacity was there, just no impetus.

This all reminds me of the Bicameral Mind hypothesis, although that dates that point of 'waking up' much later.

In certain society what we would call mental illness was often looked at as people being touched by the gods, that they had some special insight. So depending on the society they were born into, they could have be highly regarded indeed.

They may have had other reasons for being highly regarded, or feared.

Non recursive language does not allow you to give commands more complicated than "you do this", for instance saying "tell X to do this thing at this time in this location" is impossible without recursion. You might be able to get as far as "tell x to do this" since you are essentially just saying "repeat these words to X" but you have no way of providing context such as the time or circumstances when X should be told or when X should be told to perform the action. All organization and command would have to be done spur of the moment. There would be no possibility of planning in advance or communicating abstract concepts and thoughts to others.

Even a small group of people with this ability facing another group without it would have an overwhelming advantage, they could deceive, plot revenge, strategize, distract, and confound their opponents with their seemingly magical ability to know what members of their own group were to do and say before they did it. They could conceptualize solutions to problems that the others could not even begin to imagine. They could invent new objects in their mind and then execute those plans or tell others about them. They could create new words and concepts in language through combination and juxtaposition. They would quite literally be super-human, every one of them a Napoleon and an Einstein and a Da Vinci and a Shakespeare at once.

Being able to conceptualize does not depend on language, many of us don't think in language at all.

There is no proof of that. Even if you experience speaking and thinking as separate sensations, it may still be that they are actually "implemented" by the same structures in the brain.

There is in fact a theory that language is a tool for thought first, and for communication only secondarily.

One may not think in a common language used to communicate (e.g. English) but the way one thinks could be considered a language regardless, only specific to one's own mind.

Take for instance Google's translation AI. It developed its own intermediary language[1] which facilitated translation between languages and enabled it to even translate languages it had not specifically been trained on.[2]

[1]e.g. It had a "word" for dog that wasn't part of any existing language.

[2]e.g. it could translate Spanish -> French even though it was only trained on Spanish -> English and English -> French

This is a very sneaky argument that misses the point I think, you are effectively defining any sharing of information as a language, then using that definition.

I'm not talking about language here. The discussion is about _recursion_ which has vast implications for both language and thought.

> To understand the importance of PFS, consider these two sentences: "A dog bit my friend" and "My friend bit a dog." It is impossible to distinguish the difference in meaning using words or grammar alone, since both words and grammatical structure are identical in these two sentences.

Am I missing something here? Those two sentences are absolutely distinguishable grammatically - in one of the them the subject is "a dog" and the object is "my friend" and in the other the subject is "my friend" and the object is "a dog".

I realize this is kind of nitpick, but at the very least this is a very bad example to use to demonstrate the importance of this PFS concept.

You're confusing semantics with grammar. They have identical grammar.

"(A dog) bit (my friend)." "(My friend) bit (a dog)." Both parse as "(compound subject) verb (compound object)." Both "my" and "a" function as adjectives modifying the single noun after them.

The sentences have identical parse trees, even when considering the types of the nodes. They are grammatically identical.

Nope; these sentences share a common root-level grammar production, that's all.

They are not identical. They are similar at this production step:

  S -> subj verb obj
The verb is common:

  verb -> ... -> bit
So we have the same overall sentence pattern with the same verb, in the same tense.

Other than that, the derivation diverges:

  subj -> ... -> a dog
  subj -> ... -> my friend
You can't just ignore substructure differences and symbol identities in comparing trees!

Of course, any two things are the same, if you choose to ignore anything that is different and define "sameness" that way.

Ignoring the symbols is exactly what is meant by "the grammar" of a sentence. Two things are grammatically the same if they use the same grammar productions when parsing/producing.

This isn't conveniently bending definitions to make myself sound correct. This is the specific technical definition that was worked out by linguists in their empirical examination of language (and later adopted by computer scientists because it happens to work nicely with programming languages too). It might not match with some colloquial uses, but it's a reasonable assumption that an article discussing linguistics would be using the precise linguistic definition.

Are they though? Isn't 'a' an indeterminate article and 'my' a pronoun?

I get the point of the example, but it seems different at the grammatical level too.

Subject/object should be at a different level above grammar.

It may be easier to understand the point being made here if we consider what it means to be entirely absent these grammar. AIUI, but can't come up with a concrete link on the Internet quickly, Koko, the gorilla that knew some ASL, tended to speak in sentences like "Koko koko cup cup cup koko koko want want want cup cup cup", if they were transcribed literally. With such utterances, there is no way to distinguish between "Koko wants the cup" and "The cup wants Koko", or even "Koko wants the cup to want Koko".

There are animals that have demonstrated some "semantics". There are parrots that do better than Koko with spoken language, but still do not produce anything like the grammatical constructs humans do. Everything I've seen out of parrots could be covered by memorizing a handful of grammar constructs in a very brute-force manner, and they have no deliberate ability to combine them.

The core point is that to distinguish '"A dog bit my friend" and "My friend bit a dog."' requires grammar, which so far as I know as one who tends to pay attention to these things, is not a thing that animals have ever demonstrated a deep understanding of (if there are animals that could distinguish that, that is probably their maximum limit) and certainly never demonstrated production of. (Or, at the very least, we've never decoded and proved is speech, if you're thinking of whales and dolphins and such.) If you produce and understand speech as the Koko sentence above, you can't distinguish those two sentences. It's not about whether changing subject and object makes a "different sentence" by some metric, it's about whether the animal in question has a well-defined usage of subject and object as distinct grammatical categories at all.

For example, in my sentence "The cup wants Koko" I used earlier, it is invalid, yet everyone here reading it has at the very least the same understanding of why it is invalid, because what it says and the grammar of why it says that is completely clear to us.

Syntactically, they're both determiners. Semantically, they're a bit different since possessives (like 'my') aren't closed under permutation†, while many other determiners ('some', 'every', 'most' &c.) are closed under permutation.

† An operation mapping a set X onto another X' by relating every member of X with some member of X', so that X' has the same number of members as X, but the identity of the members may be different.

Others have addressed most of this, but no one else has called this out specifically: "my" is not a pronoun. Pronouns stand in for nouns. "my" stands in for a determiner. Both replace concrete with relative, but that's a semantic category that doesn't impact syntactic class.

"A dog bit a friend" and "my dog bit my friend" still work the same, though.

Indeed, I think it's just an incorrect example, the point gets across anyway.

Substitute 'my' for the 'a's?

You can't be randomly substituting words for other words (let alone ones with different lexical categories) and keep claiming that you have the same parse tree!

Agreed, and of course the meaning also changes - What I mean is that you can have a different example that makes the point while avoiding this particular issue.

to build/combine on two comments in the thread before mine: word position matters in English as you demonstrated. "A dog bit my friend" has different meaning than "My friend bit my dog".

In other languages, differentiating between subject/object is done differently. example: "A doggo bit my friendee" or "My friendo bit a dogee" and you can see the difference. Word order is still somewhat relevant here because I didn't add endings to the verb too, which is commonplace in such languages. When you do that you are able to be far more free with word position:

My Friendee doggo bitto. Bitto My friendee doggo.

In this way it's easier to tell the subject/object difference.[1] To some degree this can be done in English and is exactly what Yoda does: "My friend a dog bit" but that sentence can be either of the ones you gave. Adding a comma may help "My friend, a dog bit" = "A dog bit my friend"

[1] This is exactly what Latin does and as rjf72 mentioned it gives more freedom. Many Latin authors put the verb at the end which gives you a sort of suspense as to just what is happening. In my second example I put the verb up front and the subject at the end which allows me to keep you in suspense as to who did the biting.

Yeah that was my first thought too. It's absolutely possible to distinguish the meaning between those two sentences using grammar alone.

Grammars make sense of sequences of symbols, i.e. order matters.

Only if you claim there is no difference between "dog" and "friend" can you claim that the two sentences are indistinguishable. But that's why they are different words -- so you can tell them apart :)

Order matters in English, because it lost its Germanic declensions. In highly inflected languages order is less relevant to meaning.

No, they are grammatically identical, with a subject performing action on an object. The only difference is that the subject and object are interchanged.

Syntactically. Not grammatically. The author put it wrong

It's a great example. No rule will tell you why one is funny and the other is not.

There seem to be counterexamples. https://en.wikipedia.org/wiki/Feral_child#Raised_in_confinem...

"Isabelle (1938) was almost seven years old when she was discovered. She had spent the first years of her life isolated in a dark room with her deaf-mute mother as her only contact. Only seven months later, she had learned a vocabulary of around 1,500 to 2,000 words. She is reported to have acquired normal linguistic abilities."

I wonder about similar future change in imagination. Almost certainly possible since there's precedent for it happening, but the hard part is imagining what sort of thought it would enable.

Just like the dark ages, particularly in medicine, practitioners followed text from hundreds of years prior, without any perceptions without. Only memorizing and being SME on past thought was relevant.

I imagine the next cognitive leap is toward greater empathy to the other, and de-emphasizing the self.

"PFS" sounds a lot like the MERGE operation of minimalist (natural language) syntax, which Chomsky has similarly linked to behavioural modernity.

https://www.brainpickings.org/2012/06/20/noam-chomsky-the-sc... (not too much info here, but the best I can find at the moment)

Quite remarkable a single mutation gave us the tools to dominate the earth.

I guess the ability to make plans, communicate and execute them and reflect on what worked and what didn't to improve the plan making endows us with significant predictive power compared to the rest of the planet. Perhaps in similar ways AI will endow some people (or itself) with predictive power beyond our own, allowing those who yield it to dominate the rest

>Perhaps in similar ways AI will endow some people (or itself) with predictive power beyond our own, allowing those who yield it to dominate the rest

Hasn't that already happened?

Bezos Gates Zuckerburg Musk?

https://www.npr.org/sections/krulwich/2012/10/22/163397584/h... -- also happened about 70k years ago. Very interesting time to be a human

How many artefacts that show imagination and will survive >70k years have you made lately?

Are you intelligent?

Most people in this thread have taken part in the production of non-biodegradable trash. A huge fraction of ancient artifacts were trash, so don't feel bad about your legacy - it's typical. ;)

Pretty sure I've written some "temporary hacks" that will still be in production 70k years from now. Not sure if those demonstrate intelligence, though.

This is quite similar in some ways to my own theory that I posted recently at https://whatismusic.info/blog/HypothesisMusicWasALanguageSys..., because it postulates a sudden transition from non-recursive language to recursive language. The one difference in my theory is that I assume that the original non-recursive language was musical (with one tune equals one meaning). And musical language could not evolve into a more efficient and expressive recursive form of language - it had to be replaced.

If anyone is interested in early human development/the onset of civilization I highly recommend the book "Sapiens". Really compelling pop-science book if you are interested into that kind of thing.

I have a little Romulus & Remus & wolf statue at home, from my wife's grandfather when he passed. I wanted it because of its ties to Girard's mimetic theory. Now I have a 2nd reason.

Hadn't people spread out from Africa, longer ago than that?

I wonder if John von Neumann would have had some toddler friends of similar level of genius, would they have invented a twin speak for a "new level" of humanity.

Intelligence variations in modern times seem pretty big, but they're nothing compared to the difference between a person who can talk and one who can only make indicative noises.

But some races were likely separated for longer than that, so these mutations are likely uneven across populations. Very disturbing if true.

Or they seperated after having this mutation since we've no evidence that any specific ethic group (I assume you used race for this?) has inability of learning a recursive language.

I had the same question while reading TFA. The Khoisan ethnic group split from the main population an estimated 150k years ago, and although their clicking language is peculiar I don't think they have any issue with recursive grammar.

Does that mean that Scheme and other languages with TCO help make more imaginative programmers?

This is talking about recursion in the grammar of the language. I can't think of a common programming language that doesn't have at least some recursive elements. Nested arithmetic operations are a common example.

>This is talking about recursion in the grammar of the language.

I know, it talks about early humans development. I just wanted to extend it (semi-jockingly) to programming.

>I can't think of a common programming language that doesn't have at least some recursive elements

Sure, but recursion is more idiomatic/powerful in some languages than others. In nested arithmetic operations it is barely apparent.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact