EY suggested that we would have some sort of compromise where we lowered male sex drive a little and increased female sex drive a little, which doesn’t appeal to me at all.
Yeah, this is Eliezer inferring too much from the most-accessible information about sex drive from members of his tribe, so to speak—it’s not so very long ago in the West that female sex drive was perceived as insatiable and vast, with women being nearly impossible for any one man to please in bed; there are still plenty of cultures where that’s the case. But he’s heard an awful lot of stories couched in evolutionary language about why a cultural norm in his society that is broadcast all over the place in media and entertainment reflects the evolutionary history of humanity.
He’s confused about human nature. If Eliezer builds a properly-rational AI by his own definitions to resolve the difficulty, and it met all his other stated criteria for FAI, it would tell him he’d gotten confused.
Well, there do seem to be several studies, including at least one cross-cultural study, that support the “the average female sex drive is lower” theory.
These studies also rely on self-reported sexual feelings and behavior, as reported by the subset of the population willing to volunteer for such a study and answer questions such as “How often do you masturbate?”, and right away you’ve got interference from “signalling what you think sounds right”, “signalling what you’re willing to admit,” “signalling what makes you look impressive”, and “signalling what makes you seem good and not deviant by the standards of your culture.” It is notoriously difficult to generalize such studies—they best serve as descriptive accounts, not causal ones.
Many of the relevant factors are also difficult to pin down; testosterone clearly has an affect, but it’s a physiological correlate that doesn’t suffice to explain the patterns seen (which again, are themselves to be taken with a grain of salt, and not signalling anything causal). . The jump to a speculative account of evolutionary sexual strategies is even less warranted. For a good breakdown, see here: http://www.csun.edu/~vcpsy00h/students/sexmotiv.htm
These are valid points, but you said that there still exist several cultures where women are considered to be more sexual than men. Shouldn’t they then show up in the international studies? Or are these cultures so rare as to not be included in the studies?
Also, it occurs to me that whether or not the differences are biological is somewhat of a red herring. If they are mainly cultural, then it means that it will be easier for an FAI to modify them, but that doesn’t affect the primary question of whether they should be modified. Surely that question is entirely independent of the question of their precise causal origin?
An addendum: There’s also the “Ecological fallacy” to consider—where a dataset suggests that on the mean, a population A has property P and population B has P+5, but randomly selecting members of each population will give very different results due to differences in distribution.
These are valid points, but you said that there still exist several cultures where women are considered to be more sexual than men. Shouldn’t they then show up in the international studies? Or are these cultures so rare as to not be included in the studies?
Actually it’s entirely possible to miss a lot of detail while ostensibly sampling broadly. If you sample citizens in Bogota, Mumbai, Taibei, Kuala Lumpur, Ashgabat, Cleveland, Tijuana, Reykjavik, London, and Warsaw, that’s pretty darn international and thus a good cross-cultural representation of humanity, right? Surely any signals that emerge from that dataset are probably at least suggestive of innate human tendency?
Well, actually, no. Those are all major cities deeply influenced and shaped by the same patterns of mercantile-industrialist economics that came out of parts of Eurasia and spread over the globe during the colonial era and continue to do so—and that influence has worked its way into an awful lot of everyday life for most of the people in the world. It would be like assuming that using wheels is a human cultural universal, because of their prevalence.
An even better analogy here would be if you one day take a bit of plant tissue and looking under a microcoscope, spot the mitochondria. Then you find the same thing in animal tissue. When you see it in fungi, too, you start to wonder. You go sampling and sampling all the visible organisms you can find and even ones from far away, and they all share this trait. It’s only Archeans and Bacteria that seem not to. Well, in point of fact there are more types of those than of anything else, significantly more varied and divergent than the other organisms you were looking at put together. It’s not a basal condition for living things, it’s just a trait that’s nearly universal in the ones you’re most likely to notice or think about. (The break in the analogy being that mitochondria are a matter of ancestry and subsequent divergence, while many of the human cultural similarities you’d observe in my above example are a matter of alternatives being winnowed and pushed to the margins, and existing similarities amplified by the effects of a coopting culture-plex that’s come to dominate the picture).
If they are mainly cultural, then it means that it will be easier for an FAI to modify them, but that doesn’t affect the primary question of whether they should be modified. Surely that question is entirely independent of the question of their precise causal origin?
It totally is, but my point was that Eliezer has expressed it’s a matter of biology, and if I’m correct in my thoughts he’s wrong about that—and in my understanding of how he feels FAI would behave, this would lead to the behavior I described (FAI explains to Eliezer that he’s gotten that wrong).
As I mentioned the last time this topic came up, there is evidence that giving supplementary testosterone to humans of either sex tends to raise libido, as many FTM trans people will attest, for example. While there is a lot of individual variation, expecting that on average men will have greater sex drive than women is not based purely on theory.
The pre-Victorian Western perception of female sexuality was largely defined by a bunch of misogynistic Cistercian monks, who, we can be reasonably confident, were not basing their conclusions on a lot of actual experience with women, given that they were cloistered celibates.
I don’t dispute the effects of testosterone; I just don’t think that sex drive is reducible to that, and I tend to be suspicious when evolutionary psychology is proposed for what may just as readily be explained as culture-bound conditions.
It’s not just the frequency of the desire to copulate that matters, after all—data on relative “endurance” and ability to go for another round, certain patterns of rates and types of promiscuity, and other things could as readily be construed to provide a very different model of human sexual evolution, and at the end of the day it’s a lot easier to come up with plausible-sounding models that accord pretty well with one’s biases than be certain we’ve explored the actual space of evolutionary problems and solutions that led to present-day humanity.
I tend to think that evolutionary psychological explanations need to meet the threshold test that they can explain a pattern of behavior better than cultural variance can; biases and behaviors being construed as human nature ought to be based on clearly-defined traits that give reliable signals, and are demonstrable across very different branches of the human cultural tree.
Yeah, this is Eliezer inferring too much from the most-accessible information about sex drive from members of his tribe, so to speak—it’s not so very long ago in the West that female sex drive was perceived as insatiable and vast, with women being nearly impossible for any one man to please in bed; there are still plenty of cultures where that’s the case. But he’s heard an awful lot of stories couched in evolutionary language about why a cultural norm in his society that is broadcast all over the place in media and entertainment reflects the evolutionary history of humanity.
He’s confused about human nature. If Eliezer builds a properly-rational AI by his own definitions to resolve the difficulty, and it met all his other stated criteria for FAI, it would tell him he’d gotten confused.
Well, there do seem to be several studies, including at least one cross-cultural study, that support the “the average female sex drive is lower” theory.
These studies also rely on self-reported sexual feelings and behavior, as reported by the subset of the population willing to volunteer for such a study and answer questions such as “How often do you masturbate?”, and right away you’ve got interference from “signalling what you think sounds right”, “signalling what you’re willing to admit,” “signalling what makes you look impressive”, and “signalling what makes you seem good and not deviant by the standards of your culture.” It is notoriously difficult to generalize such studies—they best serve as descriptive accounts, not causal ones.
Many of the relevant factors are also difficult to pin down; testosterone clearly has an affect, but it’s a physiological correlate that doesn’t suffice to explain the patterns seen (which again, are themselves to be taken with a grain of salt, and not signalling anything causal). . The jump to a speculative account of evolutionary sexual strategies is even less warranted. For a good breakdown, see here: http://www.csun.edu/~vcpsy00h/students/sexmotiv.htm
These are valid points, but you said that there still exist several cultures where women are considered to be more sexual than men. Shouldn’t they then show up in the international studies? Or are these cultures so rare as to not be included in the studies?
Also, it occurs to me that whether or not the differences are biological is somewhat of a red herring. If they are mainly cultural, then it means that it will be easier for an FAI to modify them, but that doesn’t affect the primary question of whether they should be modified. Surely that question is entirely independent of the question of their precise causal origin?
An addendum: There’s also the “Ecological fallacy” to consider—where a dataset suggests that on the mean, a population A has property P and population B has P+5, but randomly selecting members of each population will give very different results due to differences in distribution.
Actually it’s entirely possible to miss a lot of detail while ostensibly sampling broadly. If you sample citizens in Bogota, Mumbai, Taibei, Kuala Lumpur, Ashgabat, Cleveland, Tijuana, Reykjavik, London, and Warsaw, that’s pretty darn international and thus a good cross-cultural representation of humanity, right? Surely any signals that emerge from that dataset are probably at least suggestive of innate human tendency?
Well, actually, no. Those are all major cities deeply influenced and shaped by the same patterns of mercantile-industrialist economics that came out of parts of Eurasia and spread over the globe during the colonial era and continue to do so—and that influence has worked its way into an awful lot of everyday life for most of the people in the world. It would be like assuming that using wheels is a human cultural universal, because of their prevalence.
An even better analogy here would be if you one day take a bit of plant tissue and looking under a microcoscope, spot the mitochondria. Then you find the same thing in animal tissue. When you see it in fungi, too, you start to wonder. You go sampling and sampling all the visible organisms you can find and even ones from far away, and they all share this trait. It’s only Archeans and Bacteria that seem not to. Well, in point of fact there are more types of those than of anything else, significantly more varied and divergent than the other organisms you were looking at put together. It’s not a basal condition for living things, it’s just a trait that’s nearly universal in the ones you’re most likely to notice or think about. (The break in the analogy being that mitochondria are a matter of ancestry and subsequent divergence, while many of the human cultural similarities you’d observe in my above example are a matter of alternatives being winnowed and pushed to the margins, and existing similarities amplified by the effects of a coopting culture-plex that’s come to dominate the picture).
It totally is, but my point was that Eliezer has expressed it’s a matter of biology, and if I’m correct in my thoughts he’s wrong about that—and in my understanding of how he feels FAI would behave, this would lead to the behavior I described (FAI explains to Eliezer that he’s gotten that wrong).
As I mentioned the last time this topic came up, there is evidence that giving supplementary testosterone to humans of either sex tends to raise libido, as many FTM trans people will attest, for example. While there is a lot of individual variation, expecting that on average men will have greater sex drive than women is not based purely on theory.
The pre-Victorian Western perception of female sexuality was largely defined by a bunch of misogynistic Cistercian monks, who, we can be reasonably confident, were not basing their conclusions on a lot of actual experience with women, given that they were cloistered celibates.
I don’t dispute the effects of testosterone; I just don’t think that sex drive is reducible to that, and I tend to be suspicious when evolutionary psychology is proposed for what may just as readily be explained as culture-bound conditions.
It’s not just the frequency of the desire to copulate that matters, after all—data on relative “endurance” and ability to go for another round, certain patterns of rates and types of promiscuity, and other things could as readily be construed to provide a very different model of human sexual evolution, and at the end of the day it’s a lot easier to come up with plausible-sounding models that accord pretty well with one’s biases than be certain we’ve explored the actual space of evolutionary problems and solutions that led to present-day humanity.
I tend to think that evolutionary psychological explanations need to meet the threshold test that they can explain a pattern of behavior better than cultural variance can; biases and behaviors being construed as human nature ought to be based on clearly-defined traits that give reliable signals, and are demonstrable across very different branches of the human cultural tree.