[FN: - the role of sexual selection complicates, but does not defeat, the story I will be sketching in this essay, however I’m going to bracket it because it’s not necessary to the narrative, and deserves its own essay.]
(1) If we designate all the goodness of the world etc. (art, poetry) as "sentimentality", I think the possibility that sentimentality is evolutionarily adaptive/selected for is being a bit prematurely disregarded. I know all throughout the post from the beginning to the end you mention that sexual selection and kin selection, while they exist, don't fundamentally change the argument, but I think they certainly need to be considered. I mean, sex being pleasurable (that you list along with all the other features of goodness of the world) I think is pretty obviously evolutionarily adaptive (you don't need to involve sexual selection or kin selection). But I think even less obviously adaptive things can be explained through a combination of natural selection, kin selection, and sexual selection. I think the counterargument to this argument would be that these arguments (for adaptiveness/selection) can be difficult to falsify and might be "Just So" stories. But still, for some (perhaps most or all) things I think you can convincingly make the case that they are in some sense adaptive.
(How would you prove it? I suppose an evolutionary biologist would be able to better provide an answer, but if a behaviour is conserved over time, found in many species, evolved multiple times through convergent evolution, exists despite apparently obvious selection pressures against the behaviour, etc., I think these can be evidence that a behaviour must be adaptive in a perhaps non-obvious way. Obviously ants or wolves or whatever aren't writing poetry, but they have complex social dynamics etc. The biggest distinction is that ours involve language, but I think the only reason other species' complex social behaviours don't involve (our form) of language is that they just haven't evolved capacity for our form of language. (Side note: our AI successors seem pretty certain to have language)).
Things like emotions, people like to contrast those with some sort of purely cognitive mode of behaviour. But I think emotions certainly are selected for/adaptive. They are useful for a variety of purposes: making behavioural plans coherent (e.g. making them consistent w/r/t approach vs. withdrawal, among other things); communicating our intentions to allow for social coordination; eliciting social responses; etc.
(I guess to the broader point about necessity vs. freedom, I'm not sure how distinguishable they are in the most cosmic of senses. But I mean in a more obvious day-to-day sense, about not wanting to work at a crappy, dangerous, demeaning job in order to survive, there is certainly a distinction)
I guess when I think about the above, I think having a complex web of social interactions, complex social communication, emotions etc., is our evolutionary birthright and probably in some sense an inevitable outcome of us being intelligent. I don't know, it just Feels Intuitive that our future replacement robots (lol) must be creatures that will have complex social dynamics and make something like robo-poetry etc.
I guess this is all kind of an argument for complacency or that everything's going to be fine or something. But maybe it will.
Re: (1) I agree, let me be clear that I think that, at the moment, what I call the 'parasite' is serving functions, I just think that:
A. The functions it serves are not, in principle, the most efficient way to meet those needs (e.g. I think there are possible more efficient alternatives, from a biological point of view, to the things we think of as giving life meaning. We're not using those more efficient alternatives because evolution isn't perfect".
B. Future super intelligences will not "need" them to propagate and expand.
On (2) I tend to disagree for reasons you can probably guess, but this is a live debate that has been hashed out elsewhere (it relates substantially to the orthogonality thesis I think).
I suppose you can think of the evolutionary fitness landscape, the peaks and valleys and all that, and imagine that some soulless, emotionness, cruel, asocial species might be the highest peak of that landscape. But if we look at how the Earth has turned out so far, it seems like if any species were getting closer to that peak (if it exists), it would be humans rather than other species (judging by how we took over nearly every habitat on Earth etc.). And a big part of what made us successful (in taking over every habitat etc.) seems to be exactly those qualities that you are concerned are sub-optimal (that is, language, social coordination, complex communication, etc.). To me, it seems like if something is going to out-compete us, the evidence suggests that it would be even better at our Most Human qualities, rather than worse.
But then, I am hesitant to think that the evolutionary fitness landscape would even have a single peak anyway--why not multiple peaks? That's the other thing the history of the Earth has seemed to suggest, that there are multiple niches that different species fill etc.
Interesting post, thanks for writing it.
My thoughts on reading this:
(1) If we designate all the goodness of the world etc. (art, poetry) as "sentimentality", I think the possibility that sentimentality is evolutionarily adaptive/selected for is being a bit prematurely disregarded. I know all throughout the post from the beginning to the end you mention that sexual selection and kin selection, while they exist, don't fundamentally change the argument, but I think they certainly need to be considered. I mean, sex being pleasurable (that you list along with all the other features of goodness of the world) I think is pretty obviously evolutionarily adaptive (you don't need to involve sexual selection or kin selection). But I think even less obviously adaptive things can be explained through a combination of natural selection, kin selection, and sexual selection. I think the counterargument to this argument would be that these arguments (for adaptiveness/selection) can be difficult to falsify and might be "Just So" stories. But still, for some (perhaps most or all) things I think you can convincingly make the case that they are in some sense adaptive.
(How would you prove it? I suppose an evolutionary biologist would be able to better provide an answer, but if a behaviour is conserved over time, found in many species, evolved multiple times through convergent evolution, exists despite apparently obvious selection pressures against the behaviour, etc., I think these can be evidence that a behaviour must be adaptive in a perhaps non-obvious way. Obviously ants or wolves or whatever aren't writing poetry, but they have complex social dynamics etc. The biggest distinction is that ours involve language, but I think the only reason other species' complex social behaviours don't involve (our form) of language is that they just haven't evolved capacity for our form of language. (Side note: our AI successors seem pretty certain to have language)).
Things like emotions, people like to contrast those with some sort of purely cognitive mode of behaviour. But I think emotions certainly are selected for/adaptive. They are useful for a variety of purposes: making behavioural plans coherent (e.g. making them consistent w/r/t approach vs. withdrawal, among other things); communicating our intentions to allow for social coordination; eliciting social responses; etc.
(I guess to the broader point about necessity vs. freedom, I'm not sure how distinguishable they are in the most cosmic of senses. But I mean in a more obvious day-to-day sense, about not wanting to work at a crappy, dangerous, demeaning job in order to survive, there is certainly a distinction)
(2) re: “imagining a beautiful, complex robotic ecology, robot poetry, robot sex”.
I guess when I think about the above, I think having a complex web of social interactions, complex social communication, emotions etc., is our evolutionary birthright and probably in some sense an inevitable outcome of us being intelligent. I don't know, it just Feels Intuitive that our future replacement robots (lol) must be creatures that will have complex social dynamics and make something like robo-poetry etc.
I guess this is all kind of an argument for complacency or that everything's going to be fine or something. But maybe it will.
Re: (1) I agree, let me be clear that I think that, at the moment, what I call the 'parasite' is serving functions, I just think that:
A. The functions it serves are not, in principle, the most efficient way to meet those needs (e.g. I think there are possible more efficient alternatives, from a biological point of view, to the things we think of as giving life meaning. We're not using those more efficient alternatives because evolution isn't perfect".
B. Future super intelligences will not "need" them to propagate and expand.
On (2) I tend to disagree for reasons you can probably guess, but this is a live debate that has been hashed out elsewhere (it relates substantially to the orthogonality thesis I think).
Very interesting to think about.
I suppose you can think of the evolutionary fitness landscape, the peaks and valleys and all that, and imagine that some soulless, emotionness, cruel, asocial species might be the highest peak of that landscape. But if we look at how the Earth has turned out so far, it seems like if any species were getting closer to that peak (if it exists), it would be humans rather than other species (judging by how we took over nearly every habitat on Earth etc.). And a big part of what made us successful (in taking over every habitat etc.) seems to be exactly those qualities that you are concerned are sub-optimal (that is, language, social coordination, complex communication, etc.). To me, it seems like if something is going to out-compete us, the evidence suggests that it would be even better at our Most Human qualities, rather than worse.
But then, I am hesitant to think that the evolutionary fitness landscape would even have a single peak anyway--why not multiple peaks? That's the other thing the history of the Earth has seemed to suggest, that there are multiple niches that different species fill etc.