Crows are capable of recursion, scientists claim

<a href="https:&#x2F;&#x2F;archive.ph&#x2F;zLgN0" rel="nofollow">https:&#x2F;&#x2F;archive.ph&#x2F;zLgN0</a>

Other thread:<p><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=33454205" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=33454205</a> (<i>&quot;Crows found to be smarter than we think&quot;</i>)

One more threat to javascript developers trying to make it in today&#x27;s economy.

I feed the crows cashews as I walk the dog, and I find their behavior fascinating.<p>They &quot;commute to work&quot; 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&#x27;re in a place where you see crows every day, it&#x27;s almost certainly the same pair of crows you’re seeing each day.<p>Each pair has a strict territory. I don&#x27;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.<p>They definitely recognize me and follow me for the peanuts. Usually they&#x27;ll follow me from one block to the next, and another pair will come up and chase them off.<p>They do seem to set aside their territorial fighting when there&#x27;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.<p>I would love to know how they determine who gets what territory. I assume it&#x27;s handed down; every year there&#x27;s a month or so when the adolescents come out (they&#x27;re a little bit ganglier and pester their parents for food all the time rather than foraging for it themselves) so I think they&#x27;re there to learn the ropes. Are there crow dynasties that have owned certain city blocks for generations? How long do they last?<p>Lots of interesting stuff to watch in the lives of crows.

&gt; ...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.<p>I&#x27;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&#x27;s still interesting, but is that &quot;understanding recursion&quot;? Recursion is the idea of repeating an algorithm or a function calling itself. Visual symmetry is a little simpler than that.<p>If a crow peels an onion layer by layer that doesn&#x27;t mean it understands recursion. It&#x27;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.

Crow Recursion is the name of my new nerdcore death metal band &quot;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!

&gt; 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.<p>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.

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.

We know that crows and many other animals use tools, which represents at least one level of linguistic abstraction (&quot;It&#x27;s not just a rock, it&#x27;s a member of a more general category of things capable of smashing open food&quot;).<p>I&#x27;ve always thought that recursion, specifically the ability to apply abstraction to itself, was a unique human trait. We don&#x27;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.<p>So while we see other animals using tools that they find in their environment, I don&#x27;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&#x27;s not mental capacity or linguistic recursion that limits tool making but something more mechanical like simply dexterity?

&gt; Adapting the protocol used in the 2020 paper, the team trained two crows to peck pairs of brackets in a center-embedded recursive sequence.<p>&gt; Two of the three monkeys in the experiment generated recursive sequences more often than nonrecursive sequences ...<p>I&#x27;m no academic but aren&#x27;t those extremely small sample sizes to make any reasonable deductions from? This looks to be even addressed in one of the papers cited...<p>&gt; 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 &quot;proof of existence&quot; showing that this cognitive capacity is, in principle, within the reach of carrion crows.[0]<p>[0] <a href="https:&#x2F;&#x2F;www.science.org&#x2F;doi&#x2F;10.1126&#x2F;sciadv.abq3356" rel="nofollow">https:&#x2F;&#x2F;www.science.org&#x2F;doi&#x2F;10.1126&#x2F;sciadv.abq3356</a>

The recursion example in the article is not good because it is language dependent. One could also say: &quot;running mouse chased by cat in the past&quot; and you would get the same idea as in opposed to &quot;The mouse the cat chased ran&quot;. 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 &quot;if you cant recurse in communication you cant recurse in general&quot;.

&gt; To address this limitation, Liao and her colleagues extended the sequences from two pairs to three pairs—such as { [ ( ) ] }<p>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&#x27;t that be more easily explained by &quot;understanding symmetry&quot; rather than &quot;understanding recursion&quot;?

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.

&quot;They do not seem to possess anything similar to human language&quot;<p>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&#x27;t know.<p>What I&#x27;d like to see is trying these tasks on a flock, rather than individual birds.

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

&quot;Crows demonstrate that crows are capable of recursion&quot;

Was it tail recursion?

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]<p>[1] <a href="https:&#x2F;&#x2F;www.quantamagazine.org&#x2F;what-scarlett-howard-learns-from-the-bees-she-teaches-20200122&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.quantamagazine.org&#x2F;what-scarlett-howard-learns-f...</a>

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

Linguistic recursion, rather than algorithmic recursion (?)

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.

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

So are cauliflower and broccoli?

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

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

Chomsky would be surprised

<p><pre><code> def caw(): print(“Crow on the tower of Hanoi say…”) caw() caw()</code></pre>

thats more than some CS graduates can do

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

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