"Ironically, if I were trying to create an AI system that could see, and reason about, the world more effectively (and who says I’m not!) I would be trying to find ways to inject more of this kind of embodied, valence-laden affect into its perceptual apparatus." I think this is exactly what people do when they assign a persona to a chatbot (e.g., "you are an expert at Jungian dream analysis. Last night I dreamt that..."). The persona seems to more effectively "constrain" the statistical space of token prediction, often improving output quality. It's possible that we're all socially encouraged to do this, take on expertise that implicitly narrows our affect in order to improve the quality of our labor output. The humanities sort of acknowledge this by (ideally) broadening our conceptual space and supposedly enriching the experience of life (before eventually being pushed into expertise). To that end, seems like market forces might be a drag for humans and for AI.
I noticed this phenomenon after learning how to knit. I told a friend that sweaters in a store were “screaming information at me,” which is an insane-sounding way of putting it, but it was just pretty neat how I could *see* all this information I didn’t know was there before, about fiber content, construction techniques, etc.
"Why can’t we see this, too, as nature sorting itself into delicate, complex, semi-stable equilibria? [...] You would be too excited to try out this superpower on museums and libraries, on kindergarten concerts and family reunions and busy street corners and starry skies and lecture halls and crowded bazaars and lonely ponds!"
I try not to be that guy, but damn it: meditation, you're talking about meditation! 😭
"The Eye of Man, a little narrow orb, clos’d up & dark,
Scarcely beholding the Great Light, conversing with the ground:
The Ear, a little shell, in small volutions shutting out
True Harmonies & comprehending great as very small"
I know this sounds sick but here's how my brain actually works:
- I'm probably one of those people who are sort of always meditating, whatever I'm doing. Yeah, I'm pretty much like a Buddha, that seems about right. Meditating would be gilding the lily, coals to Newcastle, putting a hat on a hat.
- And, conversely, I need to keep something in my back pocket, in case things get really bad. If things get really bad I can always start meditating.
I love and use both of those lines when I'm not practicing myself. They're old favorites and I think they're pretty true. “Ultimately, the great incentive for self exploration is our growing sensitivity to the vacuous, insipid nature of life. It is not really a question of renunciation, but rather of interest dying naturally.”
My main worry is always that people imagine getting into meditation as if it were prioritizing diet and exercise, but it's really more like choosing to learn an instrument. Akin (equivalent?) to aesthetic experiences, meditation is worthy as its own end and not because of some other "practical" benefit. It should be fun and loose and exploratory, in addition to being challenging and complex. It accumulates depth for the rest of your life, even if your practice wanes, particularly in how it reshapes/opens up the "listening" experience forever.
I agree that you're the type who "picked up the guitar" of introspection/meditation at a young age and taught yourself to play, and as skillfully as anyone. Your whole book is a testament to this in the aesthetic context. To keep straining the metaphor, I suppose I'm merely a stodgy trad ranting on the virtues of studying the classics and practicing fundamentals for an instrument you've already dedicated your life to.
(Also what you want in your back pocket is a gun, not a gun catalog.)
"Does it have to taste bad to you now? Wouldn’t your life be strictly better if it didn’t?"
I've been thinking about this 'sommelier paradox' of yours. Instinctively I think it's false, maybe because I'm a snob, but also because I believe having true knowledge must always be preferable to not. So thinking it through, I think the mistake might be in only taking into consideration the hypothetical marginal utility of consumption in that moment- as in, there are more bad wines on offer than good, so we'd be better off if we liked bad wines. But that ignores how we acquired that knowledge in the first place.
Like, my freshman year of college, I devoured Cryptonomicon and the Baroque Cycle by Neal Stephenson, thought they were simply the greatest, and continued to think so for years. Then I decided to reread it a decade or so later. I couldn't even finish it! I couldn't believe I'd be saying it was favorite book! What happened in the interim is I read other books, further developed my own tastes, and grew up. Marginally, it might be better to be able to appreciate mediocre writers, there's certainly many more of them out there, and they are often more prolific too. But even though I can't appreciate Stephenson the way I could as a teenager, I'm not worse off, because that lack of appreciation is itself a byproduct of the pleasure gotten from lots of reading. Similarly, the only way to acquire an appreciation of wine is through study, to learn all those complex subtleties, and that requires drinking and appreciating a lot of wine. The cumulative benefit of that true knowledge is positive, even if it means the marginal benefit of the next bottle of Barefoot Cabernet is now negative. So I tell myself, at any rate.
The logic you are sketching out here, as I understand it, is that there is a trade-off happening as I refine and evolve my taste, but this trade-off is net positive and I end up better off. My ability to enjoy lower-quality things is diminished, but this is more than made up for by the ability I have acquired to enjoy higher-quality things. As a unapologetic snob myself, I agree with your math - IF we grant that the inability to enjoy the lower-quality things is a necessary and unavoidable side-effect of improving one's taste.
But is it? This is the question I want to examine. On the one hand, it seems intuitively obvious that it is. That's what it *means* to like things, right? To like some things you have to dislike other things, right? Your taste-pleasure-function must have peaks and valleys, right? There must be *contours*, there must be light and dark for there to be shape, pattern, structure. Otherwise, you would just have an undifferentiated grey expanse.
But wait. Is this intuitive picture correct? After all, in our thought experiment, we still have a high-resolution sequence of *preferences*. We still recognize higher-quality things as being of higher-quality and appreciate all the ways in which they are better than lower-quality things. We just lose the negative affect, the valence of displeasure. Just like in the ice cream parlor where we might prefer vanilla to chocolate, but chocolate doesn't taste bad to us. We can still appreciate and enjoy chocolate ice cream, despite finding vanilla better and reaching a higher register of pleasure when we enjoy vanilla. Why can't we do that with Cryptonomicon?
Well, what about the function of displeasure as a spur? As a motivation to seek our better quality things to enjoy? Are you worried that without that stick of negative affect, the carrot of greater pleasure will be inefficient? That we will linger too long on Cryptonomicon, without seeking out the better, more challenging things?
This, to me, feels like a place where we can stand up and say, "No! Stop! I want to seek out the best things in life, and learn to love them for their own sake, because I fully appreciate their value, not out of fear of being whipped by the lash of disgust."
Reward functions and punishment functions serve an important role in the foundational evolution of our value systems, in the training algorithm which initially branded our preferences into us, but we don't need to remain beholden to them forever. Just like we no longer need to remain beholden to the fitness functions of natural selection. We can choose to keep the carrot of pleasure and discard the stick of pain. We can choose to learn the things we want to know about the world out of a conscious desire to know them, and this can be an efficient way to learn, just as in the classroom we understand the positive motivation of curiosity and the pleasure of learning is more effective than the threat of punishment.
Finally, I will say, don't you feel pleasure in the beauty of a sunset? Or in a waterfall? Don't you think there is as much to appreciate, and enjoy, in Cryptonomicon? Is it not as much a complex corner of the world, full of splendid detail, the froth and whorl of the universe sorting out its thermodynamic business? Don't you think there is something beautiful about any place you might find yourself - the frozen food aisle at Trader Joes, a barbershop in Jersey City on a rainy afternoon, the back yard of your parent's house - and don't you wish you had the capacity to see it? Yes, waiting in line at the DMV is not as wonderful as holding a newborn infant in your arms, but isn't it, probably, wonderful? Why can't we enjoy it? Why must we suffer?
If you found yourself reading those Stephenson books again for some reason, because you had been hired to adapt them into a game, or because they were part of some research project, would you grit your teeth and force yourself to choke it down? Or is there some other mode, some other emotion? Why not a pleasure that is even greater than the pleasure of the naïve freshman who first encountered them, why not a pleasure expanded and made richer by the knowledge of the other books that are even better, a pleasure that can still feel everything you felt then, even though it now knows so much more?
"My ability to enjoy lower-quality things is diminished, but this is more than made up for by the ability I have acquired to enjoy higher-quality things."
My point was actually slightly different than this, closer to, "... but this is more than made up for by the higher-quality things I have ALREADY enjoyed along the way." That is, to enjoy better things, we need to already have enjoyed things, which is why we come out ahead on a net basis. (That's an important distinction, since it could be the case that we still come out ahead even in a situation where there are NO more high-quality things left to enjoy! It was really the wine we drank along the way).
I'm really just making a different argument against a straw man you set up though. And on the whole, I totally agree with you, in terms of this "we must imagine Sisyphus happy" line of argument. That we can perceive the beauty of a frozen food aisle as much as a sunset or what have you. However, I do think there are limits to how far that extends, and maybe I feel the 'stick of pain' still has some usefulness for us. Here's a scenario for you:
You come across an old watch in the attic. You can tell it's Swiss, but not the brand. You can identify, and appreciate, the design as typically midcentury. It has pleasingly industrial accents, and you can enjoy the (now retro-)futuristic glow-in-the-dark accents on the hands and dials. "If I were a horologist, I'd enjoy this even more!" you think, and since there's no time like the present, you pry open up the back, and begin to remove the delicate gears within, admiring the lapidary fineness of the parts within. You are having a great time.
Now, imagine how much LESS you'd enjoy the preceding if you were additionally burdened with knowledge about the danger of radium paint. Now the preceding is instead a scene of horror, and all the fine appreciation you could've been having is replaced with only the thin spoilsport satisfaction of a clicking geiger counter. Clearly it's better to know the serious risks of handling a watch like that, even if we might enjoy it more if we didn't (for a time).
This is contrived of course, but I think it's less a narrow situation than it seems. The reason we learn to dislike some things (eg. twinkies, cheap wines, huffing glue, et cetera) is since we learn they have negative externalities outside of, but inseparable from, the things we enjoy about them. I think we can still have a stance that appreciates these things maybe, but that feels pretty different than the enjoyment we could've had in ignorance. My examples here relate to physical health, but I think similar logic applies if the negative aspects at stake were mental health related, or the health of others, etc. Or, do you think there's an ontologically finer line between smoking and the DMV, say, such that those are really different cases?
Interesting! Yes, maybe some, or all, of our preferences reflect genuine well-being. On some level, this is true by definition. In order to exist at all we must prefer things that keep us stable and away from equilibrium with our environment. Maybe the cruder, less-sophisticated entertainments we enjoyed as freshmen don't just inflict opportunity cost on us, but more serious harms, akin to poison.
This doesn't *seem* to be the case though! I used to really love guitar solos when I was younger. When I imagine re-connecting to that feeling, and being able to fully rock out, really feel that electric thrill light up my limbic system the way it used to, I don't see the harm there, especially if it is done *within* the larger context of my current, broader musical knowledge. Likewise, thumbing through an old Ripley's Believe it or Not paperback, or assembling a plastic tank model (with safe glue!), or any of the other dumb stuff I used to love. Again, I wouldn't seek these experiences out, given my current preferences, but if I found myself doing them for some reason, I don't see the harm, and, greedily, I want all the joy.
Since we're talking about aesthetics, we can't defer to the heuristic of "is it good for you?" because, in a way, that's the whole *point* of the realm of aesthetics. There are no simple, practical criteria with which to measure.
In fact, in a way, *Shakespeare* is the poison, because it kills the child who used to only like The Hardy Boys.
After all, we don't want to be *too* stable. Radiation doesn't just kill us, it also changes us, mutates us, all wine is intoxicating, even the expensive stuff.
"Ironically, if I were trying to create an AI system that could see, and reason about, the world more effectively (and who says I’m not!) I would be trying to find ways to inject more of this kind of embodied, valence-laden affect into its perceptual apparatus." I think this is exactly what people do when they assign a persona to a chatbot (e.g., "you are an expert at Jungian dream analysis. Last night I dreamt that..."). The persona seems to more effectively "constrain" the statistical space of token prediction, often improving output quality. It's possible that we're all socially encouraged to do this, take on expertise that implicitly narrows our affect in order to improve the quality of our labor output. The humanities sort of acknowledge this by (ideally) broadening our conceptual space and supposedly enriching the experience of life (before eventually being pushed into expertise). To that end, seems like market forces might be a drag for humans and for AI.
I noticed this phenomenon after learning how to knit. I told a friend that sweaters in a store were “screaming information at me,” which is an insane-sounding way of putting it, but it was just pretty neat how I could *see* all this information I didn’t know was there before, about fiber content, construction techniques, etc.
"Why can’t we see this, too, as nature sorting itself into delicate, complex, semi-stable equilibria? [...] You would be too excited to try out this superpower on museums and libraries, on kindergarten concerts and family reunions and busy street corners and starry skies and lecture halls and crowded bazaars and lonely ponds!"
I try not to be that guy, but damn it: meditation, you're talking about meditation! 😭
"The Eye of Man, a little narrow orb, clos’d up & dark,
Scarcely beholding the Great Light, conversing with the ground:
The Ear, a little shell, in small volutions shutting out
True Harmonies & comprehending great as very small"
I know this sounds sick but here's how my brain actually works:
- I'm probably one of those people who are sort of always meditating, whatever I'm doing. Yeah, I'm pretty much like a Buddha, that seems about right. Meditating would be gilding the lily, coals to Newcastle, putting a hat on a hat.
- And, conversely, I need to keep something in my back pocket, in case things get really bad. If things get really bad I can always start meditating.
I love and use both of those lines when I'm not practicing myself. They're old favorites and I think they're pretty true. “Ultimately, the great incentive for self exploration is our growing sensitivity to the vacuous, insipid nature of life. It is not really a question of renunciation, but rather of interest dying naturally.”
My main worry is always that people imagine getting into meditation as if it were prioritizing diet and exercise, but it's really more like choosing to learn an instrument. Akin (equivalent?) to aesthetic experiences, meditation is worthy as its own end and not because of some other "practical" benefit. It should be fun and loose and exploratory, in addition to being challenging and complex. It accumulates depth for the rest of your life, even if your practice wanes, particularly in how it reshapes/opens up the "listening" experience forever.
I agree that you're the type who "picked up the guitar" of introspection/meditation at a young age and taught yourself to play, and as skillfully as anyone. Your whole book is a testament to this in the aesthetic context. To keep straining the metaphor, I suppose I'm merely a stodgy trad ranting on the virtues of studying the classics and practicing fundamentals for an instrument you've already dedicated your life to.
(Also what you want in your back pocket is a gun, not a gun catalog.)
"Does it have to taste bad to you now? Wouldn’t your life be strictly better if it didn’t?"
I've been thinking about this 'sommelier paradox' of yours. Instinctively I think it's false, maybe because I'm a snob, but also because I believe having true knowledge must always be preferable to not. So thinking it through, I think the mistake might be in only taking into consideration the hypothetical marginal utility of consumption in that moment- as in, there are more bad wines on offer than good, so we'd be better off if we liked bad wines. But that ignores how we acquired that knowledge in the first place.
Like, my freshman year of college, I devoured Cryptonomicon and the Baroque Cycle by Neal Stephenson, thought they were simply the greatest, and continued to think so for years. Then I decided to reread it a decade or so later. I couldn't even finish it! I couldn't believe I'd be saying it was favorite book! What happened in the interim is I read other books, further developed my own tastes, and grew up. Marginally, it might be better to be able to appreciate mediocre writers, there's certainly many more of them out there, and they are often more prolific too. But even though I can't appreciate Stephenson the way I could as a teenager, I'm not worse off, because that lack of appreciation is itself a byproduct of the pleasure gotten from lots of reading. Similarly, the only way to acquire an appreciation of wine is through study, to learn all those complex subtleties, and that requires drinking and appreciating a lot of wine. The cumulative benefit of that true knowledge is positive, even if it means the marginal benefit of the next bottle of Barefoot Cabernet is now negative. So I tell myself, at any rate.
The logic you are sketching out here, as I understand it, is that there is a trade-off happening as I refine and evolve my taste, but this trade-off is net positive and I end up better off. My ability to enjoy lower-quality things is diminished, but this is more than made up for by the ability I have acquired to enjoy higher-quality things. As a unapologetic snob myself, I agree with your math - IF we grant that the inability to enjoy the lower-quality things is a necessary and unavoidable side-effect of improving one's taste.
But is it? This is the question I want to examine. On the one hand, it seems intuitively obvious that it is. That's what it *means* to like things, right? To like some things you have to dislike other things, right? Your taste-pleasure-function must have peaks and valleys, right? There must be *contours*, there must be light and dark for there to be shape, pattern, structure. Otherwise, you would just have an undifferentiated grey expanse.
But wait. Is this intuitive picture correct? After all, in our thought experiment, we still have a high-resolution sequence of *preferences*. We still recognize higher-quality things as being of higher-quality and appreciate all the ways in which they are better than lower-quality things. We just lose the negative affect, the valence of displeasure. Just like in the ice cream parlor where we might prefer vanilla to chocolate, but chocolate doesn't taste bad to us. We can still appreciate and enjoy chocolate ice cream, despite finding vanilla better and reaching a higher register of pleasure when we enjoy vanilla. Why can't we do that with Cryptonomicon?
Well, what about the function of displeasure as a spur? As a motivation to seek our better quality things to enjoy? Are you worried that without that stick of negative affect, the carrot of greater pleasure will be inefficient? That we will linger too long on Cryptonomicon, without seeking out the better, more challenging things?
This, to me, feels like a place where we can stand up and say, "No! Stop! I want to seek out the best things in life, and learn to love them for their own sake, because I fully appreciate their value, not out of fear of being whipped by the lash of disgust."
Reward functions and punishment functions serve an important role in the foundational evolution of our value systems, in the training algorithm which initially branded our preferences into us, but we don't need to remain beholden to them forever. Just like we no longer need to remain beholden to the fitness functions of natural selection. We can choose to keep the carrot of pleasure and discard the stick of pain. We can choose to learn the things we want to know about the world out of a conscious desire to know them, and this can be an efficient way to learn, just as in the classroom we understand the positive motivation of curiosity and the pleasure of learning is more effective than the threat of punishment.
Finally, I will say, don't you feel pleasure in the beauty of a sunset? Or in a waterfall? Don't you think there is as much to appreciate, and enjoy, in Cryptonomicon? Is it not as much a complex corner of the world, full of splendid detail, the froth and whorl of the universe sorting out its thermodynamic business? Don't you think there is something beautiful about any place you might find yourself - the frozen food aisle at Trader Joes, a barbershop in Jersey City on a rainy afternoon, the back yard of your parent's house - and don't you wish you had the capacity to see it? Yes, waiting in line at the DMV is not as wonderful as holding a newborn infant in your arms, but isn't it, probably, wonderful? Why can't we enjoy it? Why must we suffer?
If you found yourself reading those Stephenson books again for some reason, because you had been hired to adapt them into a game, or because they were part of some research project, would you grit your teeth and force yourself to choke it down? Or is there some other mode, some other emotion? Why not a pleasure that is even greater than the pleasure of the naïve freshman who first encountered them, why not a pleasure expanded and made richer by the knowledge of the other books that are even better, a pleasure that can still feel everything you felt then, even though it now knows so much more?
That's my question.
"My ability to enjoy lower-quality things is diminished, but this is more than made up for by the ability I have acquired to enjoy higher-quality things."
My point was actually slightly different than this, closer to, "... but this is more than made up for by the higher-quality things I have ALREADY enjoyed along the way." That is, to enjoy better things, we need to already have enjoyed things, which is why we come out ahead on a net basis. (That's an important distinction, since it could be the case that we still come out ahead even in a situation where there are NO more high-quality things left to enjoy! It was really the wine we drank along the way).
I'm really just making a different argument against a straw man you set up though. And on the whole, I totally agree with you, in terms of this "we must imagine Sisyphus happy" line of argument. That we can perceive the beauty of a frozen food aisle as much as a sunset or what have you. However, I do think there are limits to how far that extends, and maybe I feel the 'stick of pain' still has some usefulness for us. Here's a scenario for you:
You come across an old watch in the attic. You can tell it's Swiss, but not the brand. You can identify, and appreciate, the design as typically midcentury. It has pleasingly industrial accents, and you can enjoy the (now retro-)futuristic glow-in-the-dark accents on the hands and dials. "If I were a horologist, I'd enjoy this even more!" you think, and since there's no time like the present, you pry open up the back, and begin to remove the delicate gears within, admiring the lapidary fineness of the parts within. You are having a great time.
Now, imagine how much LESS you'd enjoy the preceding if you were additionally burdened with knowledge about the danger of radium paint. Now the preceding is instead a scene of horror, and all the fine appreciation you could've been having is replaced with only the thin spoilsport satisfaction of a clicking geiger counter. Clearly it's better to know the serious risks of handling a watch like that, even if we might enjoy it more if we didn't (for a time).
This is contrived of course, but I think it's less a narrow situation than it seems. The reason we learn to dislike some things (eg. twinkies, cheap wines, huffing glue, et cetera) is since we learn they have negative externalities outside of, but inseparable from, the things we enjoy about them. I think we can still have a stance that appreciates these things maybe, but that feels pretty different than the enjoyment we could've had in ignorance. My examples here relate to physical health, but I think similar logic applies if the negative aspects at stake were mental health related, or the health of others, etc. Or, do you think there's an ontologically finer line between smoking and the DMV, say, such that those are really different cases?
Interesting! Yes, maybe some, or all, of our preferences reflect genuine well-being. On some level, this is true by definition. In order to exist at all we must prefer things that keep us stable and away from equilibrium with our environment. Maybe the cruder, less-sophisticated entertainments we enjoyed as freshmen don't just inflict opportunity cost on us, but more serious harms, akin to poison.
This doesn't *seem* to be the case though! I used to really love guitar solos when I was younger. When I imagine re-connecting to that feeling, and being able to fully rock out, really feel that electric thrill light up my limbic system the way it used to, I don't see the harm there, especially if it is done *within* the larger context of my current, broader musical knowledge. Likewise, thumbing through an old Ripley's Believe it or Not paperback, or assembling a plastic tank model (with safe glue!), or any of the other dumb stuff I used to love. Again, I wouldn't seek these experiences out, given my current preferences, but if I found myself doing them for some reason, I don't see the harm, and, greedily, I want all the joy.
Since we're talking about aesthetics, we can't defer to the heuristic of "is it good for you?" because, in a way, that's the whole *point* of the realm of aesthetics. There are no simple, practical criteria with which to measure.
In fact, in a way, *Shakespeare* is the poison, because it kills the child who used to only like The Hardy Boys.
After all, we don't want to be *too* stable. Radiation doesn't just kill us, it also changes us, mutates us, all wine is intoxicating, even the expensive stuff.