[roenxi]

> "Since the octopus was not reacting to any existing threat, but rather in anticipation of one, it had demonstrated foresight and planning."

The journalist is hopefully misunderstanding a more correct explanation by the scientists. That most certainly isn't demonstration of foresight and planning. Nature is perfectly capable of encoding that sort of information as an instinctual response. Something like [trying to sleep] -> [get distressed by exposure to open ocean].

It is like how humans feel cold and put on a jumper. It isn't a demonstration that we all understand heat flow equations and have modelled out that we need to consume less food if we put another layer between us and the outside world, increasing the long term economics of our survival, demonstrating an advanced knowledge of thermodynamics and economics. Our body has encoded "heat leaking through skin" -> "feel uncomfortable" into our senses and that is all we need to respond to without considering the consequences of the response.

[klabb3]

Agreed. This is like building a little den out of nearby objects, which seems to be at sub-mammal levels? Even birds that didn’t get the best grades do this.

They picked an extremely mediocre example by octopus standards. Play is more interesting but this happens with primitive mammals as well, like squirrels and such. And fish, btw.

It feels like pop-sci journalism is always assuming animal behavior is orders of magnitude less sophisticated[1] than what is obvious from having pets, watching a basic nature doc, or just observing wildlife for a tiny bit.

[1]: I tend to avoid words like intelligence etc both because they’re ambiguous as hell, and because people jump to anthropomorphization, backwards-rationalizing etc, in exactly the fashion you mention.

[legacynl]

> It feels like pop-sci journalism is always assuming animal behavior is orders of magnitude less sophisticated[1] than what is obvious from having pets, watching a basic nature doc, or just observing wildlife for a tiny bit.

I would go further and claim that basically everybody holds the assumption that animal cognition lives on a scale of not sophisticated to very sophisticated, and also that human cognition holds the top spot on this scale, and that therefor anything else is below us in sophistication.

The idea that conscious thoughts are what makes our cognitive abilities so good is merely an assumption. It's very possible that in million year of evolution humans evolve a way to do everything we do today on instinct without the need for energy hungry consciousness. That would probably allow the brain to become much smaller and efficient over time.

With our assumptions we would conclude that the smaller brain and instinctual behaviors would signify a less sophisticated brain.

[capex]

There's a video about an octopus solving a maze: https://youtu.be/7__r4FVj-EI?si=FH2QsL7BxlJ4h7KF. You could call that reactionary too, but at some point we have to admit octopuses do have a higher level of intelligence than we'd assume from an animal their size.

[brians]

It sounds like you would prefer evidence of something like Kahneman’s System II as evidence of high intelligence? That is, an abstract general purpose thinking system, even if it is slower and in many ways weaker than the instinctive/trained/evolved systems for solving specific problems?

[roenxi]

I preferred the bottle-bouncing behaviour. It seems unlikely that nature encoded that one as an instinct - or if it did, it is an adaptive enough instinct to be intelligence for me.

It just happens that the leading example was not correct. That sort of thing turns up all the time in creatures that are effectively thoughtless automatons.

[zztop44]

It’s a bit off topic I’m not sure behaviour being “instinctive” means it’s thoughtless (whatever that means).

For example, being uncomfortable might be an instinctive response to being cold, but it’s hardly thoughtless. Some people choose to swim in icy water as a form of recreation. Some sit outside on a cool night and enjoy the crisp air. We write songs and poetry about the deep frost. It means something to us.

I know our nervous systems are far more complex than eg: a bivalve mollusc, who are probably not composing poetry about the sensation they get that makes them close their shell.

I don’t really disagree with you about the observed behaviours in TFA, but I do always feel weird when I hear any living thing called effectively a thoughtless automaton.

[paulrudy]

> but I do always feel weird when I hear any living thing called effectively a thoughtless automaton.

I feel very weird about this too.

This de facto assumption that organisms are mechanisms first, that "higher-order" experiences we are familiar with as humans are at best "emergent" from these mechanisms--I wish people would understand that this is fundamentally as much a belief system, an article of faith, as the many alternatives are.

Saying this doesn't imply that every belief system is equally valuable or scientifically verifiable. But I think it's important to recognize one's axioms and/or biases.

The mechanistic view is certainly compelling and has the appearance of being all-encompassing.

Its all-encompassing appearance may actually be an artifact of how used to the story we've grown. A clockwork universe. We know that one by heart, whether we're scientists or not. We can apply that template to anything, and set about exploring (or reading about) the mechanisms. The fact that there are mechanisms everywhere doesn't prove that mechanism is all there is. That last part is an implicit belief system, a hidden article of faith, and that's how you get Descartes vivisecting dogs, and conscious experience necessarily (as though no other possibility could exist) having to be an "emergent" property.

[zztop44]

Yes. Also, one of the most popular alternative views (that a non-“mechanical” soul is the origin of higher-order experiences) is probably even more deeply engrained in many of our cultures and patterns of thought. And operates very similarly, albeit with a different story about how these properties “emerge”.

[paulrudy]

Yes, that's true. Generally, we see that point of view as a belief, an article of faith (and of course the subject of a lot of disagreement) but we're falsely conditioned to imagine that the purely-mechanical view is free of all that, and it's not.

[rocqua]

Occam's razor does support the mechanistic view. It is certainly necessary to explain the entire world, and it is sufficient to explain intelligence. So by Occam's razor we should accept it is the best explanation of intelligence.

That doesn't mean we have to give up on understanding the mechanism. Nor does it mean we shouldn't try to find out if it isn't a sufficient explanation. But calling it an unsupported belief system goes a bit far. It certainly isn't proven scientific theory. And there is room for it to be wrong. But there are good scientific principles behind this belief.

[paulrudy]

> But calling it an unsupported belief system goes a bit far. It certainly isn't proven scientific theory.

I agree it would be going too far to call it an unsupported belief system. I don't think I implied that it's unsupported, but it is a belief system. My point is that even with support, the underlying, unproven or unprovable assumptions should not be glossed over.

I'm skeptical that we have a mechanistic explanation of intelligence. I think we have sensible-but-unproven hypotheses, partially supported by observations, for how intelligence might evolve/arise. There's a lot of hand-waving between mechanistic principles and an outcome of general intelligence. One can imagine Occam's razor applying, if the hand-waving eventually resolves to something coherent. Until then, it's a combination of good science and fantasy.

Intelligence is just one of many human experiences that are believed/assumed to have mechanistic explanations. We should be careful to recognize the assumptions, however sound they may seem, and not turn them into dogma.

[TheOtherHobbes]

You can't apply Occam's Razor to things you don't understand, because that's just claiming to explain something without actually explaining anything.

Intelligence and consciousness are two of the things we don't understand.

[rocqua]

Instinctiveness means there isn't much modeling going on. And modeling reality to guide actions is a big part of intelligence, thought, etc.

[lpfnrobin]

Thankful this hasen't happend to you and me yet. Fascinating article

[legacynl]

> like how humans feel cold and put on a jumper.

This is actually a bad example. Although cold being uncomfortable is instinctual, the way how people deal with feeling cold is not instinctual at all. People use different methods of dealing with the cold depending on situation, culture, experience.

In this way your example actually speaks for the claimed forethought, because in this analogy the octopus knows (from experience or speculation) that he will be uncomfortable when trying to sleep if he's exposed and he decides to 'put on a jumper' by putting rocks in front of the entrance.

Also instinctual behavior is in general much less sophisticated. Like an infant animal starting sucking when encountering a nipple shaped object or a moth circling a light bulb because it always tries to keep the brightest source above it. It could be that pulling debris around him before sleeping could be instinctual, but it would be very unlikely for that to include the process of looking for suitable rocks beforehand.

[inglor_cz]

The jumper explanation is interesting, but it has a hidden catch.

While the Stone Age people who first came up with clothing certainly didn't understand heat flow equations, they still had to think in advance about the coming cold snap and make preparations: hunting down some game for its coat, skinning it etc. Quite a sophisticated set of actions, not a mere instinctive response to changing temperature.

[godelski]

While I agree with your distinctions I think it is important to note that the boundary of this is not explicitly clear and it can be difficult to differentiate.

Certainly instinctual responses are a form of planning, but more automated. For example, people often forego reaching for a jacket (that is well within reach or even in their literal hands) if they are quickly going from one warm place to another. And idk about you, but I've done this without a moment's thought about how the other end will be warm. The action feels fairly automatic in the same way walking outside in winter has an automatic response of grabbing a jacket.

Of course, this could be because (and likely is) a more complex equation is formulated in my mind but just operated outside the conscious part. But I think that just illustrates how the lines blur. Especially since many conscious actions turn into unconscious ones. Think about when you first learned to drive a car and how involved you were compared to now many years later. In fact, you WANT to have automated responses as these are quicker, but they come through experience. Do we call this foresight and planning? I think you can argue either side. You can easily argue that some algorithm with explicit foresight and planning was encoded into the subconscious.

I guess I'm just trying to say that it is quite complicated and we should be ensuring this complexity is known (the existence of complexity, not necessarily all the details), or it will be difficult to interpret.

[pugworthy]

For what it’s worth, this is an excerpt from a book, not a piece by a journalist say, summarizing someone’s study.

The author (Toomey) is Professor of English at the University of Massachusetts, Amherst, where he teaches courses in writing and in the history of science.