Crows are capable of recursion, scientists claim

https://www.scientificamerican.com/article/crows-perform-yet-another-skill-once-thought-distinctively-human/

by kposehn • 4 years ago

111 95 4 years ago

scientificamerican.com

yamrzou• 4 years ago

https://archive.ph/zLgN0

recuter• 4 years ago

Well, this makes them more capable than some project managers.

dtgriscom• 4 years ago

Linguistic recursion, rather than algorithmic recursion (?)

adrian_b• 4 years ago

Yes, it appears that they have taught the crows to write some kind of LISP-like expressions with nested parentheses, which is what is meant here by recursion.

yellowapple7• 4 years ago

Crows prefer Emacs over Vi confirmed.

prgammer• 4 years ago

Was it tail recursion?

eckza• 4 years ago

Tail-caw optimized.

BirAdam• 4 years ago

So, crows, cetaceans, octopuses, and elephants… let’s just be thankful that all of the other smart animals can’t manipulate their environments at scale.

vintermann2• 4 years ago

Elephants can rip up trees at a pretty impressive scale.

nkrisc• 4 years ago

Environmental “engineering” by ancestors of elephants is a possible factor that contributed to the environmental change that likely lead to the extinction of Gigantopithicus, the enormous, ancient ape.

j-bos• 4 years ago

Reminds me of this speculative piece:https://twitter.com/CWingUexkull/status/1554639440873652227?...

tldr is ancient elephants manipulating the environmemt at scale and enslaving ancient humans. Short worthwhile read with fun details to justify the narrative.

ren_engineer• 4 years ago

if you look at all those groups you can kind of see what each lacks compared to humans to get runaway intelligence benefits that basically create a continuous selective effect for more intelligence.

humans have a solid combination of dexterity, life span, and verbal communication. Dexterity allows for manipulation that allows intelligence to be used to maximum effect in an environment, life span allows for time to learn and share knowledge, and verbal communication allows to share knowledge effectively which allows group work and knowledge to pass through generations.

crows and elephants seem to be the most limited by lack of dexterity, crows actually try to use tools to get around the limitation. If Elephants had some octopus style multi-trunk tentacles that could be used like hands they'd be set I think

bmitc2• 4 years ago

I am not convinced dexterity has much to do with intelligence. It is more an enabler of technology development and not raw intelligence.

brnaftr361• 4 years ago

"Smart" is a... Bad category. It's just a big breeding ground for reification.

We do have a term for animals that do manipulate the environment: ecosystem engineering. Which includes a variety of species, including trees. Which to that extent, it's our scale of engineering that pales in comparison.

I mean, cyanobacteria is implicated in an extinction event, and the initiation of the great oxygenation event, and the rise of complicated metabolic systems using oxygen as the terminal electron acceptor. Very cleverly (intelligently, smartly?) they evolved [by chance,] a metabolic niche that was totally vacant which coincidentally could also positively amplify for some extraordinarily long period.

harrykeightley• 4 years ago

  def caw():
    print(“Crow on the tower of Hanoi say…”)
    caw()

  caw()

agumonkey• 4 years ago

cawrecursive scheme

b1124• 4 years ago

All that work, and with no rest(sleep)?

vintermann2• 4 years ago

"They do not seem to possess anything similar to human language"

They definitively have ways of telling each other things. Just how they do it and what the limits of it are, I guess we don't know.

What I'd like to see is trying these tasks on a flock, rather than individual birds.

elcapitan• 4 years ago

"Crows demonstrate that crows are capable of recursion"

JimBlackwood• 4 years ago

Kind of ironic, but the picture used seems to be of a raven - not a crow.

Feathers on the beak are quite far out, the shape of the head isn’t as “streamlined” and the top beak is more curved and longer. Or am I mistaken here?

cercatrova• 4 years ago

Here's the thing. You said a "jackdaw is a crow."

kzrdude• 4 years ago

I didn't read the paper but it talks about "Corvids (jays, jackdaws, crows and ravens)" first but it seems like they did experiments using carrion crows (Corvus corone corone).

b3morales• 4 years ago

I believe the extended overhanging curved tip is just the sign of a captive crow that is, sadly, not taking good care of its beak. Raven beaks are generally much broader.

RcouF1uZ4gsC• 4 years ago

Some scientists remain skeptical. Arnaud Rey, a senior researcher in psychology at the French National Center for Scientific Research, says the findings can still be interpreted from a simple associative learning standpoint—in which an animal learns to link one symbol to the next, such as connecting an open bracket with a closed one. A key reason, he explains, lies in a feature of the study design: the researchers placed a border around the closed brackets in their sets—which the authors note was required to help the animals define the order of the brackets. (The same bordered layout was used in the 2020 study.) For Rey, this is a crucial limitation of the study because the animals could have grasped that bordered symbols—which would always end up toward the end of a recursive sequence—were the ones rewarded, thus aiding them in simply learning the order in which open and closed brackets were displayed.

Doing this seems to make this more of a linear problem than a recursive problem. Place all the non-bordered symbols. Then place all the bordered symbols in reverse order of the non-bordered ones.

mannykannot3• 4 years ago

I thought what followed was equally thought-provoking:

"In Rey’s view, the notion of “recursive processing” as a unique form of cognition is in itself flawed. Even in humans, he says, this capacity can most likely be explained simply through associative learning mechanisms..."

AFAIK, this contrasts quite markedly with Chomsky's position that recursion is central to his universal grammar.

jeffmcmahan• 4 years ago

Yes, as usual, our pop science writer with a dim understanding of the issue quotes critics of the Chomskyan paradigm (the study aims to falsify Chomsky’s suspicion that a symbolic operation called “merge” gives humans a unique capacity for deep recursion) and for balance quotes other critics of the Chomskyan paradigm - and never do we hear from the syntacticians (who study recursive grammatical structure as a career).

hairofadog5• 4 years ago

I feed the crows cashews as I walk the dog, and I find their behavior fascinating.

They "commute to work" each day. Crows all live in a community that they go home to each night, and then each morning they all come back to their specific areas, usually in pairs. If you're in a place where you see crows every day, it's almost certainly the same pair of crows you’re seeing each day.

Each pair has a strict territory. I don't know how it works in rural areas, but in my neighborhood the territories are demarcated by streets and cover one or two neighborhood blocks.

They definitely recognize me and follow me for the peanuts. Usually they'll follow me from one block to the next, and another pair will come up and chase them off.

They do seem to set aside their territorial fighting when there's a lot of food; a guy up the street from me throws ridiculous amounts of food onto his front lawn, like whole loaves of bread, to feed the birds (probably rats and coyotes too), and the crows all gather peacefully there every morning before heading back to their territories.

I would love to know how they determine who gets what territory. I assume it's handed down; every year there's a month or so when the adolescents come out (they're a little bit ganglier and pester their parents for food all the time rather than foraging for it themselves) so I think they're there to learn the ropes. Are there crow dynasties that have owned certain city blocks for generations? How long do they last?

Lots of interesting stuff to watch in the lives of crows.

geewee• 4 years ago

How did you manage to get started? The crows close to me are so skittish that if I try to throw a peanut at them they just fly off and never come back. I've only managed to not have that happen once.

hairofadog5• 4 years ago

Hmm, if you’re literally throwing it at them, yeah, they don’t like that. You just kinda have to throw it where they can see it, then walk away, and they’ll come running over to scoop them up. Do that for a week or two and they’ll start looking for you. Also, cashews are their favorite (based on watching them pick the cashews out of trail mix and then come back for the other stuff later).

They don’t let me get close to them, but sometimes they do fly up from behind and tap me on the back of the head to get my attention.

Also, unlike the guy up the street, I only throw out a few at a time so they have to sort of follow me to get anything.

My ultimate goal is to be able to point at someone and have the crows attack them like in Bioshock. It’ll be my thing for when society collapses. So far, however, I think our relationship is much more “let’s get some peanuts off this sucker” than “I’ll loyally follow this peanut person into battle”.

prox2• 4 years ago

Just place a few fav seeds where they can see it clearly, soon they will figure it out. Once they start recognizing you (which is very fast, few times tops) they will come. Don’t look or make sudden motions when they are near enough. They don’t like that.

In time some might become more interested in you, depending on mood, and will come closer to you.

carapace5• 4 years ago

Crows are companions-to-hunters (wolves and humans mostly.) They interpret your behavior through that "lens".

If you set out food and then watch them that means (in crow) that you are a hunter, and a lousy one, and you're trying to hunt them.

Toss out some food and "forget" it and they'll think you're such a successful hunter that you don't even need to every scrap. Someone worth knowing.

In the old days, crows would go on the hunt with us and fly over hidden game in the bush dipping a wing in a certain way to let you know where it is, then you share the meat with them. They still remember even if most of us have forgotten. ;)

mannykannot3• 4 years ago

Around where I live (southern New York), the crows seem to gather in groups of twenty or so as sunset approaches, and appear to be socializing (and making quite a lot of noise in the process.) There does not seem to be any fixed location where this happens, and it seems different from when they are mobbing a raptor, where the target of their ire can quickly be seen from the way the crows are directing their attention towards it.

I can't say whether it occurs only at certain times of the year, and I have not paid any attention to where the birds go afterwards.

hairofadog5• 4 years ago

It looks like there’s a pod in Poughkeepsie, but it’s an annual gathering place rather than nightly?

https://www.scenichudson.org/viewfinder/poughkeepsies-massiv...

Some other reporting about crows in your neck of the woods:

https://www.ithaca.com/news/seeing-spots-nope-those-are-crow...

whatevertrevor• 4 years ago

I've observed a daily rollcall at evenings in our neighborhood. They don't necessarily gather together, but they sit atop tall trees and then caw to tally everyone up in a radius of a few blocks without fail at sunset.

onos• 4 years ago

Which is it peanuts or cashews?

hairofadog5• 4 years ago

They prefer cashews but until the economy gets a little better it’s peanuts.

frithsun• 4 years ago

One more threat to javascript developers trying to make it in today's economy.

mmplxx• 4 years ago

Seems to me that more is at stake here for lisp developers.

youguyssuck• 4 years ago

This actually proves that crows are smarter than JavaScript developers

HideousKojima2• 4 years ago

Reminds me of the "game journalist vs. pigeon" video:

https://m.youtube.com/watch?v=vfIn9OqhZWo

aulin• 4 years ago

now they just need to learn memoization and they're ready to master the FAANG interview

tengbretson• 4 years ago

Meh. Until they can figure out a stack-safe trampoline, forget about it.

ispo• 4 years ago

They have some skills better fitting that my younger self! Not an exaggeration!

mannykannot3• 4 years ago

I wonder is the result owes a lot to visual processing (a highly developed trait in most birds.) Arguably, with the symbol-pairs used here, each pair look a bit like a single but partially-obscured convex object, with the inner ones lying on top of the outer ones.

hoosieree• 4 years ago

Not surprising. Crows around here are always making a Racket.

syllablehq• 4 years ago

Crow Recursion is the name of my new nerdcore death metal band "Crows will claw your face if you mess with crows will claw your face if you mess with crows will claw your face if you mess with crows will claw your face if you mess with crows will claw your face if you mess up the END CONDIIITTTTIIIIOOONNN!

perrygeo2• 4 years ago

We know that crows and many other animals use tools, which represents at least one level of linguistic abstraction ("It's not just a rock, it's a member of a more general category of things capable of smashing open food").

I've always thought that recursion, specifically the ability to apply abstraction to itself, was a unique human trait. We don't just abstract one level, we keep going - building tools that help us build tools, building social structures that facilitate the production of these meta-tools, etc.

So while we see other animals using tools that they find in their environment, I don't know of any evidence that they recurse on that, i.e. they always use tools for their primary purpose but never engage in systematic production of tools, or tools that make other tools. But given this article, maybe it's not mental capacity or linguistic recursion that limits tool making but something more mechanical like simply dexterity?

b1124• 4 years ago

One comment, re: tools to build tools. I've seen animals use rocks, in videos, to debark sticks, to be used to perform another task.

That's a tool to make a tool, and I bet there are other cases.

In fact, I recall a video of a monkey, using a sharpened stone, to break the window of its habitat. To sharpen the stone, it used something else (the video did not say anything other than the monkey sharpened the stone).

So that is another case.

perrygeo2• 4 years ago

Yes - but have you ever seen a troop of monkeys organize themselves to produce a surplus of sharp stones for their collective advantage? Thankfully, monkeys do not have this recursive ability. Sadly, humans do.

EGreg• 4 years ago

Chomsky would be surprised

default-kramer• 4 years ago

To address this limitation, Liao and her colleagues extended the sequences from two pairs to three pairs—such as { [ ( ) ] }

Three pairs of brackets still seems like setting the bar very low... But even if they got great results with 10 pairs of brackets, wouldn't that be more easily explained by "understanding symmetry" rather than "understanding recursion"?

Oxidation• 4 years ago

Presumably you could check that with non-symmetrical symbols (so unrelated pairs like A/Y, or 1/W, etc)

But then, I'm not sure that many humans would figure that out.

perihelions• 4 years ago

Other thread:

https://news.ycombinator.com/item?id=33454205 ("Crows found to be smarter than we think")

zaidhaan3• 4 years ago

Adapting the protocol used in the 2020 paper, the team trained two crows to peck pairs of brackets in a center-embedded recursive sequence.
Two of the three monkeys in the experiment generated recursive sequences more often than nonrecursive sequences ...

I'm no academic but aren't those extremely small sample sizes to make any reasonable deductions from? This looks to be even addressed in one of the papers cited...

While a sample size of two is not enough to infer that any crow in the population may generate center-embedded recursive sequences, we present a "proof of existence" showing that this cognitive capacity is, in principle, within the reach of carrion crows.[0]

[0] https://www.science.org/doi/10.1126/sciadv.abq3356

zaidhaan3• 4 years ago

Not sure why I got downvoted. If my assumption is wrong then I'd be really curious to hear from someone that knows better.

goto11• 4 years ago

It is unclear what your complaint is. As the second quote states, you only need a sample size of one to prove that a certain ability can exist. You just cant say much about how widespread the ability is.

jamal-kumar• 4 years ago

I suppose the next step is seeing if you can train bees to get the concept. May seem farfetched but I was recently reading how they can be taught to understand basic arithmetic and the concept of zero. [1]

[1] https://www.quantamagazine.org/what-scarlett-howard-learns-f...

uranium• 4 years ago

Why is this evidence of recursion rather that evidence of recognition of symmetry? The sequences they show are all simple symmetrical ones, e.g. {()}.

kragen• 4 years ago

NSFW!

tasuki• 4 years ago

So are cauliflower and broccoli?

rojobuffalo• 4 years ago

...the crows still had to figure out the center-embedded order where open and closed brackets were paired from the outside in...if the birds only learned that open brackets were at the beginning of the sequence and closed ones were at the end, you would expect an equal proportion of ( { ) } mismatched and correct responses. But...the crows chose more of the latter than the former, even with the more complex sequences of three pairs of brackets.

I'm no expert but it seems like the birds are identifying visual symmetry. Maybe they are remembering the mid-point of nested symmetrical symbols, and that's still interesting, but is that "understanding recursion"? Recursion is the idea of repeating an algorithm or a function calling itself. Visual symmetry is a little simpler than that.

If a crow peels an onion layer by layer that doesn't mean it understands recursion. It's doing something recursive. Understanding recursion, in my mind, involves being able to describe a recursive function or make a prediction based on understanding the function.

xkcd1963• 4 years ago

The recursion example in the article is not good because it is language dependent. One could also say: "running mouse chased by cat in the past" and you would get the same idea as in opposed to "The mouse the cat chased ran". I believe other non-indoeuropean languages like Chinese or Japanese work in some ways much different from ours where you would pretty much derive to similar sentences if you translate without considering grammatical rules (not to say that they dont use recursion). Also one has to make a distinction from recursion and recursion in language as opposed to "if you cant recurse in communication you cant recurse in general".

shafyy• 4 years ago

I also love how crows put a walnut on the road, wait until a car drives over it and cracks it open, and then they go collect it. Smort.

fullsend• 4 years ago

I’m always amazed how low supposedly smart people’s opinion of animal consciousness is.

layer82• 4 years ago

I’ll believe it when they can solve Tower of Hanoi to arbitrary depth.

compressedgas• 4 years ago

But if they were to, would they not have discovered the iterative solution?

layer82• 4 years ago

Good point. I’d still give them credit. :)

JoeyBananas• 4 years ago

thats more than some CS graduates can do

Crows are capable of recursion, scientists claim

Related Stories