Hacker News new | comments | show | ask | jobs | submit login
Why dolphins are deep thinkers (2003) (theguardian.com)
99 points by rcarmo 7 months ago | hide | past | web | favorite | 64 comments



A long time ago I came across this question [1] (probably through Hackernews): "What do you believe is true even though you cannot prove it." I found the question hugely fascinating. For me one answer is that some attributes which we think of as uniquely human, such as intelligence or consciousness, aren't actually so unique. I believe at least many mammals and possibly other animals (birds? some fish?) have some form of (perhaps basic) intelligence.

Of course there is always a big risk in anthropomorphizing behavior and there are huge challenges in designing replicable experiments. However, ultimately humans are just apes so it shouldn't be too surprising if some of our human attributes actually evolved much earlier. In any case I find research like that mentioned in the article immensely interesting. There is still so much we can learn on this front.

[1] https://www.edge.org/annual-question/what-do-you-believe-is-...


Humans are pretty good at pretending other humans aren't human, it should be no surprise that we find it easy to ignore traits that we consider human in other species.


Savage


> some attributes which we think of as uniquely human, such as intelligence or consciousness, aren't actually so unique

If you find this topic interesting I'd highly suggest reading the book Other Minds [1] and maybe Being a Beast [2]. Apparently there was a Y fork millions of years ago in the tree of life where vertebrates and invertebrates split. It's interesting that both branches developed advanced nervous systems, specifically octopi and mammals. What's the purpose of an octopus having the highest brain to body mass of any invertebrate? Especially when they live ~2 years. Also, 2/3 of their neurons are found in its arms. Imagine thinking with your arms or seeing with your skin - they have photoreceptors in their skin and are able to match their environment yet they're color blind.

[1] https://www.goodreads.com/book/show/28116739-other-minds

[2] https://www.goodreads.com/book/show/28696605-being-a-beast


Ha, I did read Other Minds! You're right, it's absolutely fascinating. The book makes an interesting point how interacting with cephalopods are probably the closest thing we have to interacting with an alien intelligence. I'm adding Being a Beast to my reading list too.


Thanks for making me spend more money on books... :)


> What do you believe is true even though you cannot prove it.

I suspect, but cannot prove, that the brain is necessary for consciousness but not sufficient.

Sort of like a handheld radio is necessary for hearing radio broadcasts, but not sufficient, it needs to tune into a signal originating elsewhere.


Some questions that have always bothered me:

Do mentally sick people have "consciousness"?

What happens to "consciousness" during episodes of extreme rage\trance\trauma?

Sometimes, I think that we are simply machines collecting data for someone. A other times, I feel there is a higher purpose.


I believe that many animals have forms of communication close to what we call language - an infinitely productive communication system with displacement and abstraction.

Many animals are capable of working in large groups. To coordinate the actions, efficient and rich communication is required.

On the other hand, with billions of years of evolution yielding advances in locomotion, speed, protection from the environment, it seems improbable that the communication will not develop as much and go sharply from nearly zero to one in a blink of an eye with the great apes.


You can go deeper than this. Attempt to define "intelligence" in any philosophically meaningful way and you might find that it isn't clear that a coffee table isn't intelligent on some level, or alternatively that a human being isn't just a sufficiently complicated coffee table. "Consciousness", in the philosophical meaning, is even more interesting because it is entirely non-empirical; it can't be measured in any objective sense.

Any deeper implications of this are left as an exercise for the reader.


> "Consciousness", in the philosophical meaning, is even more interesting because it is entirely non-empirical; it can't be measured in any objective sense.

Don't mistake "don't know how" with "cannot". The fact that we currently don't know what consciousness is, doesn't entail that we won't one day. Consciousness will likely one day be considered a some physical process, which means it will one day be measurable.

The difficulties surrounding qualia require a resolution, but it's a mistake to just take some people's word that physicalism can't account for our qualitative perceptions.


> Consciousness will likely one day be considered a some physical process

The only people I know of that might contest that non-human animals are conscious are dualists who attribute consciousness to minds and minds only to human beings.

> it's a mistake to just take some people's word that physicalism can't account for our qualitative perceptions

Physicalism BY DEFINITION cannot account for qualia because the presuppositions of physicalism BY DEFINITION rule out the possibility of accounting for qualia. This is not the same as saying that consciousness or the experience of color is non-physical. It is the same as saying that the metaphysics of physicalism is too inadequate. Sadly, physicalism has become a blindly held dogma by philosophers like Dennett and Rosenberg, leading them to embrace eliminativism (the denial that qualia exist at all). Naturally, this is an evasion, a Procrustean cherrypicking of the evidence to suit the scientistic, physicalist dogma. It is an incoherent position and a non-starter.

The solution is to embrace a richer metaphysics instead of repeating the vacuous claim that we will/may one say have a solution to the problem of qualia while clinging to presuppositions that are false. It can't and won't happen.


> Physicalism BY DEFINITION cannot account for qualia because the presuppositions of physicalism BY DEFINITION rule out the possibility of accounting for qualia.

Except we clearly invented a term describing something we've observed or believe, therefore something must be accounted for. Ergo, physicalism still must account for qualia, even if it's not the account you were hoping for.

> Sadly, physicalism has become a blindly held dogma by philosophers like Dennett and Rosenberg, leading them to embrace eliminativism (the denial that qualia exist at all).

Please, every time we thought humans were special, we were wrong. Arguments arguing for our specialness effectively amount to special pleading.

> The solution is to embrace a richer metaphysics instead of repeating the vacuous claim that we will/may one say have a solution to the problem of qualia while clinging to presuppositions that are false. It can't and won't happen.

So said the vitalists. Met any vitalists lately?


No. You are incorrect.

Let us say that we have a computer system so complicated that it is capable of mimicking a human being with perfect accuracy. You would probably say that that computer is not a being that has consciousness, because it is merely doing calculations to replicate behavior. Problem is, you could easily say the same about any given human being. There is, quite literally, no way to tell the difference through objective measurement.

We can measure responses to stimuli, we can determine the complexity of a system, but we have no way of determining if a system experiences anything or is simply running a program, and we never will.


> Let us say that we have a computer system so complicated that it is capable of mimicking a human being with perfect accuracy. You would probably say that that computer is not a being that has consciousness, because it is merely doing calculations to replicate behavior.

I have no problem considering such a system conscious and intelligent. What's the difficulty exactly?

> Problem is, you could easily say the same about any given human being. There is, quite literally, no way to tell the difference through objective measurement.

Exactly, which means there is no difference.

Consciousness will turn out to be a functional process, which means any system with the same structure will exhibit consciousness. This would be an objective measurement.


Then how can you make a statement to the effect that we may one day know how to measure consciousness, which you have just agreed is fundamentally unmeasurable?


It's just a question of definition. Instead of "consciousness" being unmeasurable, the word would get redefined to something that can be measured, like a particular volume and/or type of calculation.

If that happens it will be in good company; at one point people thought that temperature and heat, for example, were fundamentally subjective and not measurable.


> which you have just agreed is fundamentally unmeasurable?

Where did I do that?


Looks like you added a line to clarify, or I missed it on first reading:

> Consciousness will turn out to be a functional process

Here's where I disagree. In order to know what kind of physical structure produces consciousness, you need to know what consciousness is, and we have no way of doing that. You can define some arbitrary criteria, say, having the ability to make metatools, but that is meaningless in this context, "conscious" will just be another word for "can make metatools".

In an effort to clarify, let me ask you this: why do you care what is and is not conscious?


> In order to know what kind of physical structure produces consciousness, you need to know what consciousness is, and we have no way of doing that.

I don't see how this is different than any other empirical fact we didn't know before we started exploring the world. In what way could we properly characterize fire, or stars, or bacteria before we had any conception of such things? Yet, I don't think anyone sensible would disagree that we largely understand these concepts now.

Similarly, consciousness is an amalgamation of different information processes that we have yet to disentangle, and all we have is a catch-all term "consciousness" to describe our ignorance.

> In an effort to clarify, let me ask you this: why do you care what is and is not conscious?

I care about anything that is not yet explained.


> I don't see how this is different than any other empirical fact we didn't know before we started exploring the world.

We didn't know that bacteria existed, but we knew what disease was. Once we had the concept of bacteria, showing it to be related to disease was a matter of testing. How do you test if something has a consciousness? I know that consciousness exists because I experience it, but for all I know you don't, and I have no way of knowing any different because experiences are, by definition, subjective.

You've probably seen this old XKCD about color [1]. It's like that. We can both perceive a wavelength of light and have labeled it "blue" for the purposes of communication, but we cannot possibly know eachother's experience of it. The very question is meaningless in an empirical model of reality.

> Similarly, consciousness is an amalgamation of different information processes that we have yet to disentangle, and all we have is a catch-all term "consciousness" to describe our ignorance.

I disagree. You describe what I would call "mind". The kind of consciousness I'm talking about could be described as "the experience of being". We all know what it is, but we get our communication about it confused with concepts like intelligence, mind, and complexity. I suspect this is because we grew up with a scientific worldview and have difficulty allowing ourselves to believe something non-objective can exist, so we entangle it with concepts that are objective to force it into our model of reality.

> I care about anything that is not yet explained.

Ok, fair enough. I asked because this sort of discussion usually takes place in the context of people having already come to the conclusion that human beings (and maybe some animals) are special and trying desperately to find a line of reasoning that allows for them being conscious, but not computers or plants or other such things. Some of my earlier arguments presumed this context and that may have been a source of mutual confusion.

[1] https://xkcd.com/32/


> Let us say that we have a computer system so complicated that it is capable of mimicking a human being with perfect accuracy. You would probably say that that computer is not a being that has consciousness, because it is merely doing calculations to replicate behavior.

Actually, I would seriously consider the proposition that it is intelligent, especially if it wasn't mimicking one specific person, and would be inclined to take Occam's razor to the proposition that it is actually faking it in ways that cannot be explained.


I don't think that there is any philosophical difficulty in distinguishing an animal from a coffee table. Intelligence is a broad concept which can be defined or measured in several ways. One definition is the ability to act upon the environment to achieve one's goals.

In contrast, there is an epistemic problem about whether other minds exist: https://plato.stanford.edu/entries/other-minds/


> the ability to act upon the environment to achieve one's goals.

Prove that the animal has a concept of goals and is not simply a complex arrangement of reactions environmental conditions, like a coffee table moving when you bump into it.


Goals are a complex arrangement of environmental conditions. But not all such conditions are equal. It's perfectly sensible to classify arrangements of matter that have self-awareness and learning behaviour as distinct from those that do not exhibit such behaviours.


We have computers that demonstrate such behavior, but few would refer to them as intelligent or grant them higher moral worth than even animals.


> We have computers that demonstrate such behavior

Not yet we don't. The full spectrum of intelligent behaviour is still out of reach for computers because we don't yet precisely know what intelligence entails. At best, we have small pieces of the puzzle that only sort of fit together.

It's still pretty clear that coffee tables exhibit none of them though.


> Few would refer to them as intelligent.

Few would refer to a coffee table as intelligent. You seem to be inconsistent in what sort of arguments you will accept.


I'm not the one trying to use intelligence as a metric for the moral worth of animals.

My argument is that intelligence is a much fuzzier thing than people like to admit. Usually though they're using it as a stand in for "consciousness", or the ability to experience the world as we do.


Your inconsistency is more straightforward than that: you started by writing "it isn't clear that a coffee table isn't intelligent on some level", but it is if the "few would regard..." form of argument is acceptable. You are also demanding proof from allenz while presenting a "few would regard.." argument for your own position.

I agree that intelligence is largely mysterious at this point, but I think you are pushing it too far in the unknown-and-unknowable direction. A debate in which the intelligence of coffee tables is an issue is not likely to provide much enlightenment about intelligence.


> You are also demanding proof from allenz while presenting a "few would regard.." argument for your own position.

"few would regard" isn't my argument though. It is an assumption of the other poster's position. I suppose I could have phrased it as "I would assume you wouldn't regard...". This was done to point out the inconsistency in their reasoning (which is admittedly assumed, if they had responded otherwise then we probably wouldn't actually be disagreeing).

>but I think you are pushing it too far in the unknown-and-unknowable direction

Actually where I'm going with this is that intelligence isn't actually what we care about re: morality, it's just a metric sometimes used because people haven't thought it through.


When one "point[s] out the inconsistency in [someone's] reasoning", one is making an argument.

Your concern with morality seems to have entered the thread rather late and tangentially, and it is not clear to me what you are trying to say about it - in fact, it feels as if you are arguing against a position that no-one here has taken.


If you're a behaviorist, you would define intelligence as "demonstrates problem-solving behavior".

If you're not a behaviorist, then it's perfectly reasonable to presuppose some definition of intent. You could further debate how I would define the coffee table itself (https://plato.stanford.edu/entries/object/), but that's the wrong level of analysis.


I think what you're asking about is Efference copy https://en.wikipedia.org/wiki/Efference_copy

Other animals do this. They perceive their own movements as different from an others.


An obsession with proof leads to the sort of sterile, unresolvable discussion that you describe in your first post. It is much more productive to make some defeasible assumptions and then look for evidence one way or another.


Such discussions reveal more about the practice of philosophy than they do about consciousness. The putative intelligence of coffee tables is much more likely to be raised as a counter-argument that it is offered as a plausible deduction.


If you have a means of determining whether or not something is conscious or merely responding to stimuli, by all means please share.

If your argument is that there is no difference, then you have to start adding a bunch of other rules to justify how you determine the moral worth of one class of things that respond to stimuli from others, which is beyond the original scope of discussion, I think.


The response of a coffee table when one bumps into it can be explained by Newton's laws alone, so by Occam's razor we can exclude the notion that consciousness is also involved.


See also discussion from 8 years ago: https://news.ycombinator.com/item?id=917128 (28 comments)


Since we can't continue the older discussion, I'd like to comment here on the new thread.

I'm discouraged by the topmost thread. In every case of (anecdotal?) experiments, the researcher was attempting repeatability with the same group of dolphins. I'd like to suggest that, given the 'negative' results and the predictions of dolphin keepers, repeatability should be attempted with a different group of dolphins.


The other implication of testing this on "another group of dolphin" is the ethical dimension.

Some people believe it's unethical to imprison dolphins in a small pool.

I'd argue we certainly don't need to capture more dolphins to 'prove' they are intelligent.

I also never understood why the barrier is set so high for giving any other species but ourselves credit for being intelligent.


> I also never understood why the barrier is set so high for giving any other species but ourselves credit for being intelligent.

It is our tendency to believe things that we want to believe, and we want to believe things that align with our self interest.

Since human moral understanding has come around to the concept that all human beings have inherent value, so therefore we can't just torture and kill them because they're not part of the tribe, we've had to keep inventing reasons why that same behavior is acceptable towards animals. If we were to decide that animals deserve the same moral considerations as ourselves, then we have to accept that many of the foundations for our comfortable lifestyles are immoral, and if we did that we'd have to consider giving them up.


> we'd have to consider giving them up.

Just this week there's been a round of articles on the scientific consensus that fish feel pain. Read https://www.smithsonianmag.com/science-nature/fish-feel-pain..., for example, and see if the final third of the article leaves you comfortable about eating fish. I've always thought the "catch and release" practice (and the accompanying justification that fish don't feel pain, so that it's harmless sportsmanlike fun) to be bewilderingly shallow thinking - but despite that fish is a much more significant part of my diet than avian or mammal meat. I'm also deeply hypocritical about mammal meat eating; every time I go to an Indian restaurant (but never otherwise) for example, I'll order lamb, just because it's tastier than saag or tofu. As another example, I don't eat pork, with the notable exception of cured meats, which are also too damned tasty. I probably could go wholesale vegan, especially with some of the better fake-meat protein sources, and yet perhaps I don't want to enough, and/or I enjoy eating meat a few times a month too much to do so. That's hardly a well justified moral position to take, and yet I suspect it's not uncommon.

[Also, the dividing line between animals and plants - while of course clearly defined - may be softer than we've assumed. (https://www.pri.org/stories/2014-01-09/new-research-plant-in...). In 25 years will we talking about eating lab grown food only instead of farmed plants?]

I suspect our moral systems are still reeling from 1000+ years of Christianity completely dominating our societal thinking - even if your parents weren't religious, or theirs, their inherited "firmware" (societal norms picked up in early years of learning) probably has a lot of irrationally derived foundations. Our attempts at scientific derived moral systems haven't had particularly great results so far, but perhaps we'll get there in time.


Quite true, intelligent behavior has been observed in many species.

Just because they aren't at the same level as us, speaking and writing books, doesn't mean like aren't intelligent in different ways.

After all, we also needed a few million years to get here.


> I also never understood why the barrier is set so high for giving any other species but ourselves credit for being intelligent.

Because the Western intellectual traditions of both science and philosophy arise from European Christian religious thought, which regards Man as made in the image of God, set apart from and above other animals.

The psychology of animals is one of the areas of scientific study that is still today struggling out from under its cultural roots, to be truly empirical about things.


Could you not train wild dolphins though? It seems like they are quick learners, so even though you might not have as quick and easy access, it shouldn't take substantially longer. Assuming the dolphins want to play along.


One of the most fascinating experiments was done by Bastian in 1960s (our first podcast will be about it once the site goes live). It remains highly controversial but has also been replicated by others. In 2016 one Russian scientist claimed to replicate it and it resulted in huge storm and controversy. Basically it involved 2 dolphins separated visually but not acoustically. One dolphin was shown what action the second dolphins should perform so both can get a fish. It appeared that dolphins did communicate and performed above the chance level.


Dolphins are expensive. I assume it was due to insufficient research funding.


Yeah, you can't just grab a dolphin from the sea, chuck them in a tank, and run experiments on them. You'll have to build a rapport with the dolphins to get any meaningful data.

There should be better ways of studying them, like passive observation in the sea, or by making a mecha-dolphin that can follow them without interfering too much.


Dolphins are truly fascinating creatures. They have extremely sophisticated brain architecture that definitely supports at least some degree of sentience. Some species like humpback whales, orcas (but curiously not Tursiops) have von Economo neurons (VENs)that have been connected to vocal learning, language and even theory of mind.Cetacean cortical surface area is quite large (https://twitter.com/Cetalingua/status/932631255165886464?ref... and brain architecture in general is amazing, like visual cortex being next to the auditory cortex (https://twitter.com/Cetalingua/status/901590689091387393?ref...)

Add to that conscious breathing, unihemispheric sleep, cognitive preparation and fully conscious control of cardiovascular system, conscious planning and preparation for dives. Some species (pygmy sperm whales) have magnetite crystals in their brains and probably have some sort of magnetoreception for navigation, and the list goes on and on.

Our challenge is to make sense of it all, and how can we even do it, our auditory cortex is nowhere near our visual cortex.


At what point in embryonic development are these magnetite crystals formed, and how? Amazing


In pygmy sperm whales, magnetite crystals appear to be concentrated in the rostroventral dura, but we know next to nothing about how they function or are used, let alone about how they are being formed.


I feel bad for Dolphins. They evolved from land animals and seemingly got much smarter underwater; only by the time their brains came online, they were, like, stuck underwater.

And then us land mammals evolved to hunt them and accidentally destroy the oceans. That sucks.

They seem like a really chill species that we seem to have a lot in common with. They have language and stuff.


So long, and thanks for all the fish!


Let's reverse the question. It's very enticing to believe that dolphins might be sentient. But what's the evidence to the contrary? There must be some.


Well, they seem to not understand that they can jump over the net and escape the capture. They (wild cetaceans especially) are not very interested in establishing contact with another sentient species ( us). In all experiments on non-human intelligence, only Alex, the grey parrot, actually asked the question.

Having said that, we need to realize that we still do not have sufficient knowledge to argue either way. We are dealing with species who rely on sound more than humans ever would. We still do not fully understand what is going on there, echolocation has been studied for 60 years, last year new echolocation clicks types were discovered. Do not also forget that our knowledge is mostly about bottlenose dolphins, but there are many other species. Last year it was also reported that humble harbor porpoises are not as asocial as they thought to be, and in fact, communicate with each other via pulsed signals produced at higher frequencies not audible to their predators.


I suppose that if the situation were reversed and I was in a cage surrounded by dolphins, I'd probably spend most of my time trying to communicate, imitating the sounds they made, trying to establish vocabulary.. and if I worked out that I wouldn't be punished, I'd probably refuse to perform tricks until they reciprocated. So, if they are intelligent they're very different behaviorally from us.


> In all experiments on non-human intelligence, only Alex, the grey parrot, actually asked the question

Is it possible that they asked questions of each other, in their own tongue, but did not know how to ask humans questions?


More examples of cetacean intelligence here in the last section: https://techcrunch.com/2016/12/31/reverse-engineering-the-un...

Apparently, dolphins have an equivalent of names and the entropy of their communication is similar to that of the human speech.


From an economic perspective: "The dolphins are not only gaming the system they are saving and using a capital structure to increase total output." [1]

[1] http://marginalrevolution.com/marginalrevolution/2018/01/dol...


Not to take anything away from cetaceans, but this article slights elephants and crows. Both recognize themselves in mirrors, and invent tools.


Pigeons do this too, believe it or not.

http://science.sciencemag.org/content/212/4495/695


Any species that has survived for a considerable period has to be intelligent.


I recommend the latest episode of Sam Harris' podcast Waking Up. It's a 3 hour discussion about consciousness between him and Anil Seth, it's quite good and probably warrants multiple playbacks.




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

Search: