Ways that animals perceive the world | Hacker News Reader

Ways that animals perceive the world

newyorker.com

78 pointsRafelMri34 comments

So there's an experiment I'm going to run. I've already got permission from the ethics committee. And by ethics committee I mean the friend of mine who owns the birds. I'm going to attempt to train parrots to train each other to talk.

Here's the premise. We make switches that the birds know how to press and understand that pressing causes actions. Easy enough. Then we hook the switches up to some sort of reward dispenser. Press button, get sunflower seed or almond. So far pretty simple. Then, we put the birds in a game-theoretic situation like the prisoners dilemma. If both birds press the left button, they both get the high reward (almond). If they both press the right button, they both get the low reward (sunflower seed). If one presses left and the other presses right, the left presser gets nothing and the right presser gets the highest reward (almond AND seed, oh wow).

Our first question is if the birds will converge upon the Nash Equilibrium for this game. There's a second, even more interesting question we want to answer. The birds are capable of talking to each other, but they don't have a general purpose language. So far as we can tell, their vocabulary consists of 2 squawks, 3 squawks, 5 squawks, and rarely 6 squawks. The possible signals the birds might send each other are arbitrary and meaningless in the context of this experiment, but they can signal each other. Thus the question becomes, will the birds invent a meaning for their squawks in order to cooperate at this game? Can the evolution of language be explained by simply trying to maximize payoffs in the presence of other birds trying to maximize payoffs?

Our third question is if the birds ever start lying to each other? If so, can I get the birds to manage my Diplomacy games for me?

To the extent that the whales clearly have a structured language, I suggest a way to decode it. Ask not what the generic patterns of whale sounds might be, but instead ask what payoffs the whales need to negotiate over. What positive sum behavior do the whales engage in that would be impossible without signaling?

You can observe these behaviours in the wild, and especially among corvids alike. Sure it’s less convenient than using animals in your own enclosure, and perhaps less alluding as they’re not “your” animals. But I reckon for anyone, besides yourself and your friend, involved it would be beneficial if you’d conduct experiments like these through observation of wild behaviour instead.

Also have you looked at the literature? Surely these experiments have been conducted several times already by many students in all kinds of flavours.

How do you account for confounding variables? For example, unseen or alternative methods of communication e.g. body language like bobbing, or clicking their beaks? Or even the tonality of the squawking? Whistling?

It seems like you might still get meaningful results for some of your questions, but also that the behavior of the birds could be far more complex than foreseen, and maybe even missed entirely.

>How do you account for confounding variables? For example, unseen or alternative methods of communication e.g. body language like bobbing, or clicking their beaks? Or even the tonality of the squawking? Whistling?

All perfectly acceptable. The relevant feature is merely the ability for the birds to put arbitrary signals in some communications channel, without those symbols having any a priori semantics. We would certainly know if they were winning more often than they should. It might take a while to figure out what non-squawk method they were using, but it wouldn't go over our heads.

Do not use food as a reward, it will not work because the animal will just use the buttons to dispense food, not to communicate. Social communication is rewarding and reinforcing in itself.

> ...whales clearly have a structured language

Well, that's new to me. References?

The first part of the OP describes it.

Structure... but 'language'? Not the same thing.

I see your point!

Edit: From the OP:

Early in the book, he [British documentary filmmaker Tom Mustill] goes to visit Payne [researcher Roger Payne, who discovered or first studied whale songs] who’s now eighty-seven. Why, he asks, do humpbacks sing? And what do their songs mean? Payne says he can’t say: “I would desperately love to know.”

> Why, he asks, do humpbacks sing? And what do their songs mean? Payne says he can’t say: “I would desperately love to know.”

"Any good looking females around? I am robust and can protect a family!"

"DAE hate the noisy floaters that eat all the fish?"

"Oh salty water, you are more salty than usual this day, oh salty water"

"I'm tired of barnacles, mom says it's nothing, but they're nasty and they're ruining my life!"

"Stop singing nonsense Waldorf, others want to sing too! If I was a few hundred miles closer I would slap you! Shut up shut up shut up!"

Joking aside, the songs probably aren't very intellectual. If they were smarter, whalers and fisheries wouldn't be so successful.

I'd like to be informed of the outcome of this experiment. Mail, Twitter, Subscription... whatever you choose.

The last paragraph got me hooked. You asked the right hard question. Tell me more if you've started working on it.

Can I ask what species of parrot?

The type that can't talk people words but loves to yell "SQUAWK". Rainbow colored, about 100 grams.

(I don't actually know the breed, they're not my birds).

Rainbow Lorikeet maybe? There's such a range of intelligence and behavior between parrot species. It matters a lot what species you're working with.

Arguably, a game that relied on increasingly complex arithmetic could become viral among birds. Geometry (lines and circles) would be next.

When we school horses, we use circles and straight lines because those things don't exist in nature, but horses can learn them with instruction and repetition, and then we use those figures as a foundation or baseline to refine subtle deviations from the figures so that the horse can learn to interpret our intent from our physical cues in the context of a baseline figure. What dressage is really is the construction of a physical language with another being who has a completely different brain and existential experience, from the first principles of geometry. People go on about horse dancing and spiritual woo, but almost two decades in, this is what I have been able to derive from it. Master riders will tell you it's a science as much as an art, and it usually takes many years to sink in why they would say that, but like constructing a language with birds to teach each other, while they may not articulate it, with guidance, animals can absolutely experience and percieve geometry, and based on that shared apprehension, we can build on it to relate in very sophisticated ways that come to resemble language.

The idea isn't crazy. I practice it more days than not. :)

I suppose its only fair that crows should understand geometry in terms of "how the man walks".

I'd say if you could get a crow to fly in a circle, and then in a straight line, that is sufficient to establish your intentions of: right and left, bigger and smaller, longer and shorter, forward and sideways, faster and slower, more and less, listening and freely, and then combinations of those.

Given that's 16 instructions in all, in 8 groups of a-or-b, a language that formed programs of combinations of each of them (without memory or state) would be pretty rich. Once you add abstraction to that, (something we're not sure about in animals), we're into some very complex behaviour and instructions in a way of relating. Communication with animals is very achievable in this model. What's not clear is how to express higher order abstractions in the model once you have those basics built up from geometry. Assistance dogs are "trained," to do tasks for their owners like opening doors, assisting with dressing, finding objects, etc. These are abstractions over the above instructions, and combined with them, we're well into relating to another being anyway.

Not going to lie I was just making a pun on geometry phrase "how the crow flies".

The short answer is that to build abstractions, you need words that represent sets of possibilities, not just specific things. words like "anywhere", "anyone", "somewhere" etc. You also need some set-builder logic words like negation and union and disjunction. Basically you need to be able to substitute a regex matching pattern in place of a literal specific shape instruction. The genius of language is that it can transmit partial information just as well as the whole.

Crows do it.

>A five-year study of crows living near Seattle in Washington State show the birds can remember a "dangerous human" and are able to share their knowledge of the learned danger with their offspring and other crows.

https://www.cbc.ca/news/science/crows-share-intelligence-abo...

Lasting recognition of threatening people by wild American crows

https://docdro.id/8UxS2eA

“What Is It Like to Be a Bat?,” published in 1974 by the philosopher Thomas Nagel ...
One might try to imagine, Nagel wrote, “that one has very poor vision, and perceives the surrounding world by a system of reflected high-frequency sound signals,” or that “one has webbing on one’s arms, which enables one to fly around at dusk and dawn catching insects in one’s mouth.” But that wouldn’t help much.

“I want to know what it is like for a bat to be a bat,” Nagel insisted.

From that quote something crystalized in my mind, the essense of understanding other people. (I would guess Nagel intended that.)

Often we assume our own perspective is reality. When think that ours is just perception and we try to think of someone else's perspective, I would guess that the great majority of the time, we imagine ourselves in their shoes - 'what would I do?' - essentially, 'how would I apply my perspective to their situation?' To actually understand another person, we need to know "what it is like for a bat to be a bat"; what their experience is like from their perspective, with their prior experiences, emotions, needs, motives, position (physically, socially, etc.), distractions, concerns, perceptual ability, cognitive abilities, etc. etc.

For example, sometimes we tell drug addicts to 'get a job' or at least to go through recovery. That's our perspective. When we think of theirs, we often think that, in their shoes, we would do those things. But I have no idea what it's like to be a drug addict on the streets; and even more, I don't know what in their life led them there, and what those experiences are like. I'm fortunate, for example, not to have had parents who were drug addicts; I have a hard time imagining where I would be - would I be educated, for example. I don't have their fears, strengths, traumas, joys, emotional capacities, I've never been in the position of having nothing at all, I've never felt the drive of addiction.

It doesn't just apply to strangers, but even more to people we know - family, friends, co-workers. What is it like for them to be them?

Excellent post. The more one thinks about it the way you describe it, the more one makes themselves open to understanding or vicariously experiencing another's perspective.

At that point, it is called compassion, and it is probably a force of nature of some kind.

Our ignorance about the experiences of animals will be a black stain on human legacy. There are still people who don’t think fish feel pain, or that cows don’t mourn when their calf is taken from them. Scallops eyesight is very interesting, but even the most common farm animals are still a complete mystery to the average person in a developed world.

That ignorance is often a shield to protect from thinking about ethical issues. Once you start believing that a fish can feel pain or a cow mourns you have to think about whether you want to keep eating them despite that knowledge. It's easier to not know or to keep believing something that's wrong.

Humans have not yet reconciled our sharp and flat teeth. This is a deeply biological issue, with roots in our teeth:

http://www.katabane.com/mt/dentist.html

>If whales are using their songs to communicate with one another, they are doing so not just across space but also across time. A call made by a humpback near Bermuda would take twenty minutes to reach a humpback swimming off the coast of Nova Scotia. If the Canadian whale answered immediately, it would be forty minutes before the Bermuda whale heard back.

So whales aren't speaking on the phone, they send mail. :)

For the youth of today: Mail = Snail mail or slow mail aka. letters and such.

I personally quite like the sound of snail mail ^^

TIL that scallops have eyes: https://www.google.com/search?tbm=isch&q=scallops+eyes&tbs=i...

That's terrifying.

Don't worry even though many of them have a much wider field of view then humans, their vision is much less advanced, many can't even see colours or the direction of light, although the latter is probably achieved with image processing.

Interesting thing that I discovered about a month ago: Humans have an inverted retina, which means that our sensing cells are the furthest away from the light of in other words the sensing cells layer is covered by another nine layers of different cells. This seemed kind of illogical to me, which I pointed out to the professor, why is our retina inverted? Because if it wasn't, then we would have blind spots and because most of the body cells diffuse light, the image quality could be better. Sadly the professor didn't know the answer. But when I got home and did some research I discovered that we actually have another set of cells in our eyes called Müller cells which extend through all nine layers from the top of retina to the top is our sensing cells. These Müller cells function similarly to optical fibers, they direct light to the sensing cells so that were actually able to see a clear image. Taking this into consideration, our eyes become quite similar to eyes of bees and bugs. As to why our retina is inverted and thus requires special cells to solve the problems that inverting the retina introduces, it's actually quite simple and obvious, is because of "power consumption" or in other words, or sensing cells require a lot of food and nutrients which is why directly behind the sensing cells there a web of capillaries and thus they are the first cells in the retina that get food.

https://archive.ph/dtQYP

Thanks! I needed this.

This topic always makes me think of Thomas Nagel’s classic, What is it like to be a bat?

https://warwick.ac.uk/fac/cross_fac/iatl/study/ugmodules/hum...

The article talks about it - it seems to be the/a starting point for the book reviewed - and I am going to read it! Thanks for the link.

Here's a bit more on the book, in a review in Science.

https://www.science.org/doi/10.1126/science.abq6526

The book is,

An Immense World: How Animal Senses Reveal the Hidden Realms Around Us by Ed Yong.

Reply on news.ycombinator.com