The referenced study on group selection on insects is “Group selection among laboratory populations of Tribolium,” from 1976. Studies on Slack claims that “They hoped the insects would evolve to naturally limit their family size in order to keep their subpopulation alive. Instead, the insects became cannibals: they ate other insects’ children so they could have more of their own without the total population going up.”
This makes it sound like cannibalism was the only population-limiting behavior the beetles evolved. According to the original study, however, the low-population condition (B populations) showed a range of population size-limiting strategies, including but not limited to higher cannibalism rates.
“Some of the B populations enjoy a higher cannibalism rate than the controls while other B populations have a longer mean developmental time or a lower average fecundity relative to the controls. Unidirectional group selection for lower adult population size resulted in a multivarious response among the B populations because there are many ways to achieve low population size.”
Scott claims that group selection can’t work to restrain boom-bust cycles (i.e. between foxes and rabbits) because “the fox population has no equivalent of the overarching genome; there is no set of rules that govern the behavior of every fox.” But the empirical evidence of the insect study he cited shows that we do in fact see changes in developmental time and fecundity. After all, a species has considerable genetic overlap between individuals, even if we’re not talking about heavily inbred family members, as we’d be seeing in the beetle study. Wikipedia’s article on human genetic diversity cites a Nature article and says “as of 2015, the typical difference between an individual’s genome and the reference genome was estimated at 20 million base pairs (or 0.6% of the total of 3.2 billion base pairs).”
An explanation here is that the inbred beetles of the study are becoming progressively more inbred with each generation, meaning that genetic-controlled fecundity-limiting changes will tend to be shared and passed down. Individual differences will be progressively erased generation by generation, meaning that as time goes by, group selection may increasingly dominate individual competition as a driver of selection.
All this is to motivate the outer/inner selection model, the idea that the ability to impose a top-down ruleset restraining the selfish short-term interests of individuals naturally results in higher-level entities. In biology, it’s multicellular life, while in human culture, it’s various forms of social enterprise. Scott is trying to explain not only how such social enterprises form, but why some fail and others thrive.
Under Scott’s explanatory framework, tenure is an inner constraint that has spread because it enhances the fitness of universities by constraining academics from focusing on shortsighted publishing at the expense of long-term investments in excellent research. This can’t be the result of blind evolution selecting for the survival and reproduction of universities, because the individual “organisms” (universities) often predate the “tenure gene” by centuries, and then acquired it hundreds of years into their individual lifespan. Instead, tenure historically is the result of 19th-century labor organizing by teachers. In fact, it looks like tenure was a rebellion of the inner system against the outer.
If we run with Scott’s framing, though, things get weird. What if tenure helps universities prosper not because it helps them produce high-quality research, but because it helps them recruit and maintain high-prestige academics capable of attracting big grants from external funders? If so, that would be a little like an organism that figures out how to attract the healthiest cells from a neighboring organism and incorporate them into its body. That seems to be a facet of group selection in biology (i.e. mitochondria and gut bacteria), but for the most part, our bodies are vigilant to destroy outsiders. To me, this is just a challenge for taking frameworks that are derived from the particularities of eukaryotic biology and trying to apply them to a different sort of entity with an entirely different set of mechanisms for self-definition, survival, reward, and replication. Perhaps group selection is still relevant, but this framing of university tenure as group selection ruleset feels less like a useful way to explore culture and more like a mad lib. We’re just taking the abstract components of the inner/outer selection model for cancer biology and finding the handiest intuitive category to map onto the component parts. It creates superficial plausibility, but isn’t historically accurate and feels mechanistically questionable.
I think that a more compelling concept of a human “ruleset” is perhaps the pattern of behavior called “professionalism.” Across the board, groups of all kinds that maintain a basic level of professional decorum tend to thrive. Those that do not tend to fail. The way that looks across cultures can vary, but there’s a recognizable commonality around the world. Something like professionalism has been around for a long time, and a very large number of professional groups have come and gone on a much shorter time scale. This creates opportunity for group selection to function, an opportunity that seems to be lacking in the university/tenure context.
Moving to the last section, Boeing is not a monopoly. Their four largest competitors have about the same number of employees all together as Boeing has alone. But that’s a side concern. More importantly, lack of slack is not the obvious explanation for why a misbehaving monopoly would persist. The most obvious solution would be for an activist investor to purchase 51% of the monopoly’s shares, then either reform or profit off the company. But they’d face whatever coordination and incentives caused the monopoly to misbehave in the first place. They’d also have to deal with efficient market questions. The real story here, I think, isn’t slacklessness, but computational challenges.
An alternative explanation for why monopolies would be major innovators is not because their freedom from competition makes them able to invest for the long term, but rather because they are the only employer in the industry. As a consequence, the company gets the credit for every innovation occurring in its industry during the time period in which it’s a monopoly.
To make this more clear, imagine that an industrial sector starts with 1 company, which produces 10 inventions per year for 10 years, for a total of 100 inventions. Then, it splits into 10 companies, each of which produces 2 inventions per year for 10 years, for a total of 200 inventions. The smaller companies will individually only have 20% of the inventions that the monopoly had, but the industry as a whole is twice as innovative.
Trying to evaluate the impact of a monopoly on innovation requires evaluating something like the “rate of innovation per capita” among employees in the sector, but even that’s not good enough, because we don’t know if a monopoly context tends to expand or constrain the population of employees, and we also care about the absolute number of innovations per year. If a highly monopolistic industrial landscape tended to increase the number of people choosing careers as innovators while decreasing the rate of innovation per capita, it’s hard to know how that would cash out in terms of innovations per year.
Turning to the section on “strategy games,” and specifically the Graham Greene quote, while I don’t know historical population levels for Italy vs. Switzerland, in modern times, Italy has 7x the population of Switzerland. From “The Renaissance in Switzerland,” “The geographical position of Switzerland made contact with Renaissance manifestations in Italy and Germany easy even if the country was too small and poor for notable buildings or works of art.” Switzerland, then, seems like it suffered if anything from too little slack, not too much.
Where Scott references Greece’s “slack” from being “hard to take over” “even by Sparta,” I suggest the delightful blog A Collection of Unmitigated Pedantry’s take on Sparta’s underwhelming military prowess. Scott points out that this part is purely speculative, so my main aim is just to spread the word that there’s a great blog on military history to be read, if you haven’t seen it before.
One of my issues with this article as a whole, beyond the epistemological topiary, is that veers between instances in which we can clearly see on a mechanistic level how dire poverty forces an inventor to take a day job rather than invent cold fusion, to a plausible but unsupported claim that Sears’ unconventional corporate structure contributed to its implosion, to wild speculation of “this thing happened in history, and maybe it was caused by too much/too little/just the right amount of slack too!” Scott sometimes labels his confidence levels explicitly, and other times indicates them with his tone, but I notice that my first read produced a credulous absorption of his speculative claims as factual, and my second read left me with Gell-Mann amnesia. If we’re really having a hard time finding a slack-based account for some phenomenon, we can always redefine slack (lack of resources vs. absence of competition), play around with whether the problem was too much or too little slack or the wrong temporal pattern of change in slack, or add a third (and why not a fourth?) layer to the inner/outer model.
As a complicating extension, it’s hard to say where the slackness “resides.” For example, imagine a student has two tutors, one harsh and exacting (low slack), the other nurturing and patient (high slack). Say the student notices that they learn faster with the low-slack tutor. They hypothesize that this shows that, at least in the tutoring domain, their optimal learning-rate is generated by low-slack tutors. So they hire low-slack tutors for every subject, only to find that their learning rate plummets.
They now suppose that this must be because with their original mix of tutors, adding a low-slack tutor enhanced their learning rate for that individual subject, but also used up their overall emotional capacity to deal with stress. Adding all low-slack tutors overwhelmed their emotional budget.
Optimal slack levels on one level led to conditions of suboptimal slack on a different level. Optimizing slack would mean finding harmonious slack levels throughout the many interwoven systems affecting the entities in question.
But we still might be able to do that by just varying the slack levels in ways that seem sensible, and keeping the new equilibrium if we like the outcome. You don’t necessarily need an RCT to figure this out. Maybe the barrier is just having the idea and putting it into action. The effects might be obvious. Maybe this is low-hanging fruit.
The vagueness and complexity of “slack” as a concept makes me worry that this term lends itself to be a hand-wavey explanation for whatever a writer wants to assert, or for political agendas, more than for making testable predictions that enhance our understanding of the world. What it seems to offer is inspiration for making testable predictions. It’s a hypothesis-generating tool.
This seems to be one of the things Scott’s really good at. He doesn’t make too many testable predictions, although he sometimes shares those of others. Instead, he finds patterns in observations and makes you fall in love with that pattern. Then, it’s on you to figure out how to turn that speculative pattern into a testable hypothesis. This is a valuable skill, one that many of his readers seem to envy. But it also puts you at risk of just resorting to that pattern he’s jammed into your head any time you need to reach for an explanation for something.
Unfortunately, I’m really not sure if we have it in us, as a blogging community, to internally coordinate intellectual progress. We have speculative stuff, like this article. To that, we need not only epistemic spot checks, like I’m trying to do in this review, but attempts at operationalization and normal scientific study, which this article is referencing in places, but which isn’t obviously being directly triggered or motivated by this article. If we mostly only have the time and expertise for speculation, exhortations, and the occasional fact check or two, how do we achieve intellectual progress? How will we get the resources to have these ideas checked, operationalized, and put to the test? How do we overcome our short-term status competitions in order to create a body of work that builds on itself over time toward a higher long-term payoff for the community?
An explanation here is that the inbred beetles of the study are becoming progressively more inbred with each generation, meaning that genetic-controlled fecundity-limiting changes will tend to be shared and passed down. Individual differences will be progressively erased generation by generation, meaning that as time goes by, group selection may increasingly dominate individual competition as a driver of selection.
I don’t think this adds up. Yes, species share many of their genes—but then those can’t be the genes that natural selection is working on! And so we have to explain why the less fecund individuals survived more than the more fecund individuals. If that’s true, then this is just an adaptive trait going to fixation, as in common (and isn’t really a group selection thing).
I’d enjoy talking this out with you if you have the stamina for a few more back-and-forths. I didn’t quite understand the wording of your counter argument, so I’m hoping you could spell it out in a bit more detail?
Looking at the paper, I think I wasn’t tracking an important difference.
I still think that genes that have reached fixation among a population aren’t selected for, because you don’t have enough variance to support natural selection. The important thing that’s happening in the paper is that, because they have groups that colonize new groups, traits can reach fixation within a group (by ‘accident’) and then form the material for selection between groups. The important quote from the paper:
The total variance in adult numbers for a generation can be partitioned on the basis of the parents in the previous generation into two components: a within-populations component of variance and a between-populations component of variance. The within-populations component is evaluated by calculating the variance among D populations descended from the same parent in the immediately preceding generation. The between-populations component is evaluated by calculating the variance among groups of D populations descended from different parents. The process of random extinctions with recolonization (D) was observed to convert a large portion of the total variance into the between-populations component of the variance (Fig. 2b), the component necessary for group selection.
So even tho low fecundity is punished within every group (because your groupmates who have more children will be a larger part of the ancestor distribution), some groups by founder effects will have low fecundity, and be inbred enough that there’s not enough fecundity variance to differentiate between members of the population of that group, (even if fecundity varies among all beetles, because they’re not a shared breeding population).
[EDIT] That is, I still think it’s correct that foxes sharing ‘the fox genome’ can’t fix boom-bust cycles for all foxes, but that you can locally avoid catastrophe in an unstable way.
For example, there’s a gene for some species that causes fathers to only have sons. This is fascinating because it 1) is reproductively successful in the early stage (you have twice as many chances to be a father in the next generation as someone without the copy of the gene, and all children need to have a father) and it 2) leads to extinction in the later stage (because as you grow to be a larger and larger fraction of the population, the total number of descendants in the next generation shrinks, with there eventually being a last generation of only men). The reason this isn’t common everywhere is group selection; any subpopulations where this gene appeared died out, and failed to take other subpopulations down with them because of the difficulty of traveling between subpopulations. But this is ‘luck’ and ‘survivor recolonization’, which are pretty different mechanisms than individual selection.
This is a little like game theory coordination vs cooperation actually. Coordination is if you can constrain all actors to change in the same way: competition is if each can change while holding the others fixed. “Evolutionary replicator dynamics” is a game theory algorithm that encompasses the latter.
Even if the beetles all currently share the same genes, any one beetle can have a mutation that competes with his/her peers in future generations. Therefore, reduced variation at the current time doesn’t cause the system to be stable, unless there’s some way to ensure that any change is passed to all beetles (like having a queen that does all the breeding).
The referenced study on group selection on insects is “Group selection among laboratory populations of Tribolium,” from 1976. Studies on Slack claims that “They hoped the insects would evolve to naturally limit their family size in order to keep their subpopulation alive. Instead, the insects became cannibals: they ate other insects’ children so they could have more of their own without the total population going up.”
This makes it sound like cannibalism was the only population-limiting behavior the beetles evolved. According to the original study, however, the low-population condition (B populations) showed a range of population size-limiting strategies, including but not limited to higher cannibalism rates.
“Some of the B populations enjoy a higher cannibalism rate than the controls while other B populations have a longer mean developmental time or a lower average fecundity relative to the controls. Unidirectional group selection for lower adult population size resulted in a multivarious response among the B populations because there are many ways to achieve low population size.”
Scott claims that group selection can’t work to restrain boom-bust cycles (i.e. between foxes and rabbits) because “the fox population has no equivalent of the overarching genome; there is no set of rules that govern the behavior of every fox.” But the empirical evidence of the insect study he cited shows that we do in fact see changes in developmental time and fecundity. After all, a species has considerable genetic overlap between individuals, even if we’re not talking about heavily inbred family members, as we’d be seeing in the beetle study. Wikipedia’s article on human genetic diversity cites a Nature article and says “as of 2015, the typical difference between an individual’s genome and the reference genome was estimated at 20 million base pairs (or 0.6% of the total of 3.2 billion base pairs).”
An explanation here is that the inbred beetles of the study are becoming progressively more inbred with each generation, meaning that genetic-controlled fecundity-limiting changes will tend to be shared and passed down. Individual differences will be progressively erased generation by generation, meaning that as time goes by, group selection may increasingly dominate individual competition as a driver of selection.
All this is to motivate the outer/inner selection model, the idea that the ability to impose a top-down ruleset restraining the selfish short-term interests of individuals naturally results in higher-level entities. In biology, it’s multicellular life, while in human culture, it’s various forms of social enterprise. Scott is trying to explain not only how such social enterprises form, but why some fail and others thrive.
In part 3, this model is used to justify the tenure system, as well as diversity in grantmaking. And this is where I think Scott’s inner/outer selection model, with its origins in biology, doesn’t work as well. Tenure has its origins in the early 1600s, but really took off in the mid-19th century. The first university was founded in 1088.
Under Scott’s explanatory framework, tenure is an inner constraint that has spread because it enhances the fitness of universities by constraining academics from focusing on shortsighted publishing at the expense of long-term investments in excellent research. This can’t be the result of blind evolution selecting for the survival and reproduction of universities, because the individual “organisms” (universities) often predate the “tenure gene” by centuries, and then acquired it hundreds of years into their individual lifespan. Instead, tenure historically is the result of 19th-century labor organizing by teachers. In fact, it looks like tenure was a rebellion of the inner system against the outer.
If we run with Scott’s framing, though, things get weird. What if tenure helps universities prosper not because it helps them produce high-quality research, but because it helps them recruit and maintain high-prestige academics capable of attracting big grants from external funders? If so, that would be a little like an organism that figures out how to attract the healthiest cells from a neighboring organism and incorporate them into its body. That seems to be a facet of group selection in biology (i.e. mitochondria and gut bacteria), but for the most part, our bodies are vigilant to destroy outsiders. To me, this is just a challenge for taking frameworks that are derived from the particularities of eukaryotic biology and trying to apply them to a different sort of entity with an entirely different set of mechanisms for self-definition, survival, reward, and replication. Perhaps group selection is still relevant, but this framing of university tenure as group selection ruleset feels less like a useful way to explore culture and more like a mad lib. We’re just taking the abstract components of the inner/outer selection model for cancer biology and finding the handiest intuitive category to map onto the component parts. It creates superficial plausibility, but isn’t historically accurate and feels mechanistically questionable.
I think that a more compelling concept of a human “ruleset” is perhaps the pattern of behavior called “professionalism.” Across the board, groups of all kinds that maintain a basic level of professional decorum tend to thrive. Those that do not tend to fail. The way that looks across cultures can vary, but there’s a recognizable commonality around the world. Something like professionalism has been around for a long time, and a very large number of professional groups have come and gone on a much shorter time scale. This creates opportunity for group selection to function, an opportunity that seems to be lacking in the university/tenure context.
Moving to the last section, Boeing is not a monopoly. Their four largest competitors have about the same number of employees all together as Boeing has alone. But that’s a side concern. More importantly, lack of slack is not the obvious explanation for why a misbehaving monopoly would persist. The most obvious solution would be for an activist investor to purchase 51% of the monopoly’s shares, then either reform or profit off the company. But they’d face whatever coordination and incentives caused the monopoly to misbehave in the first place. They’d also have to deal with efficient market questions. The real story here, I think, isn’t slacklessness, but computational challenges.
An alternative explanation for why monopolies would be major innovators is not because their freedom from competition makes them able to invest for the long term, but rather because they are the only employer in the industry. As a consequence, the company gets the credit for every innovation occurring in its industry during the time period in which it’s a monopoly.
To make this more clear, imagine that an industrial sector starts with 1 company, which produces 10 inventions per year for 10 years, for a total of 100 inventions. Then, it splits into 10 companies, each of which produces 2 inventions per year for 10 years, for a total of 200 inventions. The smaller companies will individually only have 20% of the inventions that the monopoly had, but the industry as a whole is twice as innovative.
Trying to evaluate the impact of a monopoly on innovation requires evaluating something like the “rate of innovation per capita” among employees in the sector, but even that’s not good enough, because we don’t know if a monopoly context tends to expand or constrain the population of employees, and we also care about the absolute number of innovations per year. If a highly monopolistic industrial landscape tended to increase the number of people choosing careers as innovators while decreasing the rate of innovation per capita, it’s hard to know how that would cash out in terms of innovations per year.
Turning to the section on “strategy games,” and specifically the Graham Greene quote, while I don’t know historical population levels for Italy vs. Switzerland, in modern times, Italy has 7x the population of Switzerland. From “The Renaissance in Switzerland,” “The geographical position of Switzerland made contact with Renaissance manifestations in Italy and Germany easy even if the country was too small and poor for notable buildings or works of art.” Switzerland, then, seems like it suffered if anything from too little slack, not too much.
Where Scott references Greece’s “slack” from being “hard to take over” “even by Sparta,” I suggest the delightful blog A Collection of Unmitigated Pedantry’s take on Sparta’s underwhelming military prowess. Scott points out that this part is purely speculative, so my main aim is just to spread the word that there’s a great blog on military history to be read, if you haven’t seen it before.
One of my issues with this article as a whole, beyond the epistemological topiary, is that veers between instances in which we can clearly see on a mechanistic level how dire poverty forces an inventor to take a day job rather than invent cold fusion, to a plausible but unsupported claim that Sears’ unconventional corporate structure contributed to its implosion, to wild speculation of “this thing happened in history, and maybe it was caused by too much/too little/just the right amount of slack too!” Scott sometimes labels his confidence levels explicitly, and other times indicates them with his tone, but I notice that my first read produced a credulous absorption of his speculative claims as factual, and my second read left me with Gell-Mann amnesia. If we’re really having a hard time finding a slack-based account for some phenomenon, we can always redefine slack (lack of resources vs. absence of competition), play around with whether the problem was too much or too little slack or the wrong temporal pattern of change in slack, or add a third (and why not a fourth?) layer to the inner/outer model.
As a complicating extension, it’s hard to say where the slackness “resides.” For example, imagine a student has two tutors, one harsh and exacting (low slack), the other nurturing and patient (high slack). Say the student notices that they learn faster with the low-slack tutor. They hypothesize that this shows that, at least in the tutoring domain, their optimal learning-rate is generated by low-slack tutors. So they hire low-slack tutors for every subject, only to find that their learning rate plummets.
They now suppose that this must be because with their original mix of tutors, adding a low-slack tutor enhanced their learning rate for that individual subject, but also used up their overall emotional capacity to deal with stress. Adding all low-slack tutors overwhelmed their emotional budget.
Optimal slack levels on one level led to conditions of suboptimal slack on a different level. Optimizing slack would mean finding harmonious slack levels throughout the many interwoven systems affecting the entities in question.
But we still might be able to do that by just varying the slack levels in ways that seem sensible, and keeping the new equilibrium if we like the outcome. You don’t necessarily need an RCT to figure this out. Maybe the barrier is just having the idea and putting it into action. The effects might be obvious. Maybe this is low-hanging fruit.
The vagueness and complexity of “slack” as a concept makes me worry that this term lends itself to be a hand-wavey explanation for whatever a writer wants to assert, or for political agendas, more than for making testable predictions that enhance our understanding of the world. What it seems to offer is inspiration for making testable predictions. It’s a hypothesis-generating tool.
This seems to be one of the things Scott’s really good at. He doesn’t make too many testable predictions, although he sometimes shares those of others. Instead, he finds patterns in observations and makes you fall in love with that pattern. Then, it’s on you to figure out how to turn that speculative pattern into a testable hypothesis. This is a valuable skill, one that many of his readers seem to envy. But it also puts you at risk of just resorting to that pattern he’s jammed into your head any time you need to reach for an explanation for something.
Unfortunately, I’m really not sure if we have it in us, as a blogging community, to internally coordinate intellectual progress. We have speculative stuff, like this article. To that, we need not only epistemic spot checks, like I’m trying to do in this review, but attempts at operationalization and normal scientific study, which this article is referencing in places, but which isn’t obviously being directly triggered or motivated by this article. If we mostly only have the time and expertise for speculation, exhortations, and the occasional fact check or two, how do we achieve intellectual progress? How will we get the resources to have these ideas checked, operationalized, and put to the test? How do we overcome our short-term status competitions in order to create a body of work that builds on itself over time toward a higher long-term payoff for the community?
It sounds like we need more slack.
I don’t think this adds up. Yes, species share many of their genes—but then those can’t be the genes that natural selection is working on! And so we have to explain why the less fecund individuals survived more than the more fecund individuals. If that’s true, then this is just an adaptive trait going to fixation, as in common (and isn’t really a group selection thing).
I’d enjoy talking this out with you if you have the stamina for a few more back-and-forths. I didn’t quite understand the wording of your counter argument, so I’m hoping you could spell it out in a bit more detail?
Looking at the paper, I think I wasn’t tracking an important difference.
I still think that genes that have reached fixation among a population aren’t selected for, because you don’t have enough variance to support natural selection. The important thing that’s happening in the paper is that, because they have groups that colonize new groups, traits can reach fixation within a group (by ‘accident’) and then form the material for selection between groups. The important quote from the paper:
So even tho low fecundity is punished within every group (because your groupmates who have more children will be a larger part of the ancestor distribution), some groups by founder effects will have low fecundity, and be inbred enough that there’s not enough fecundity variance to differentiate between members of the population of that group, (even if fecundity varies among all beetles, because they’re not a shared breeding population).
[EDIT] That is, I still think it’s correct that foxes sharing ‘the fox genome’ can’t fix boom-bust cycles for all foxes, but that you can locally avoid catastrophe in an unstable way.
For example, there’s a gene for some species that causes fathers to only have sons. This is fascinating because it 1) is reproductively successful in the early stage (you have twice as many chances to be a father in the next generation as someone without the copy of the gene, and all children need to have a father) and it 2) leads to extinction in the later stage (because as you grow to be a larger and larger fraction of the population, the total number of descendants in the next generation shrinks, with there eventually being a last generation of only men). The reason this isn’t common everywhere is group selection; any subpopulations where this gene appeared died out, and failed to take other subpopulations down with them because of the difficulty of traveling between subpopulations. But this is ‘luck’ and ‘survivor recolonization’, which are pretty different mechanisms than individual selection.
This is a little like game theory coordination vs cooperation actually. Coordination is if you can constrain all actors to change in the same way: competition is if each can change while holding the others fixed. “Evolutionary replicator dynamics” is a game theory algorithm that encompasses the latter.
Even if the beetles all currently share the same genes, any one beetle can have a mutation that competes with his/her peers in future generations. Therefore, reduced variation at the current time doesn’t cause the system to be stable, unless there’s some way to ensure that any change is passed to all beetles (like having a queen that does all the breeding).