Non-Malthusian Scenarios
This is an attempt to list all of the possible ways in which humanity may avoid scenarios where the average standard of living is close to subsistence, in response to Robin Hanson’s recent series of posts on Overcoming Bias, where he argues that such an outcome is likely in the long run.
I’ll start with six, some suggested by myself, and others collected from comments on Overcoming Bias and Robin’s own posts. If anyone provides additional ideas, I’ll add them to the list.
(I have a more general point here, BTW, which is that predicting the far future is very difficult. Before thinking that some outcome is inevitable or highly likely, it’s a good idea to repeatedly ask oneself “This is all the ways that I can think of why it may fail to come true. Am I sure that all of them have low probability and that I’m not missing anything?” There may be some scenario with a non-negligible probability that your brain simply overlooked when you first asked it.)
Singleton
A world government or superpower imposes a population control policy over the whole world.
Strong Security
Strong defensive technologies and doctrines (such as Mutually Assured Destruction) allow nations, communities, and maybe tribes and families to unilaterally limit their populations within their own borders, while holding off hordes of would-be invaders and immigrants.
Non-Human Capital
Maximizing the wealth and power of a nation requires an optimal mix of human and non-human capital. Nations that fail to adopt population controls find their relative wealth and power fade over time as their mixes deviate from the optimum (i.e., they find themselves spending too much resources on raising humans, and not enough on building machines), and either move to correct this or are taken over by stronger powers. (I believe that historically this was the reason China adopted its one-child policy.)
Unlimited Growth
We don’t completely understand the laws of physics, nor the nature of value. There turns out to be some way for economic growth to continue without limit. (Robin himself once wrote “I know of no law limiting economic value per atom” but apparently changed his mind later.)
Selfish Memes
Memes that manage to divert people’s resources away from biological reproduction and towards memetic reproduction will have an advantage over memes that don’t. On the other hand, genes that manage to block such memes will have an advantage over genes that don’t. Memes manage to keep the upper hand in this struggle (or periodically regain the upper hand).
Disease, Warfare, Natural Disasters, Aliens, Keeper of the Simulation
One or more of these come along regularly to keep the human population in check and per capita incomes above subsistence.
My claims are mostly about median income, not average. I am not making strong 99.9% confident claims; I am making my best estimates. So even if I cannot prove that any future physics will limit economic value per atom, I can use our best understanding of physics today to estimate such limits.
Strong security might allow high median income in the nations that limit population, but that would not assure the global median remained high.
On your capital argument, I think you need to learn more econ growth theory.
Your meme argument would apply to any non-meme-pushing use of resources whatsoever. And why can’t having kids be a good way to push memes?
Why do you care more about median income than average?
Yes, I guess I’m more interested in exploring the full distribution of outcomes rather than the best estimate.
That argument has been bugging me too as not being quite right, but I’m having trouble articulating why. Can you say a few words about why it’s wrong, or do I really need to read another literature to understand the error?
Suppose such memes do dominate the meme pool. Still, at any moment a new meme may arise and succeed in diverting resources away from having kids and into reproducing itself, and cause a drop in population growth.
My guess is that Robin is talking about how the median entity, regardless of how much nonsentient capital it has for itself, will spend almost all the proceeds of that capital on reproduction, and very little on plasma TVs.
Robin’s recent post and discussion there made me think of another way that the median entity might spend a lot on the equivalent of plasma TVs. Suppose a high capital to labor ratio is required to optimize military power in the far future. Then once all available resources are claimed by someone, and the optimal capital/labor mix is reached, nations should limit reproduction to avoid declining in power. During peacetime, the high capital levels could allow a high per capita production of consumer goods. (I guess this is just an elaboration of my original argument, which I still don’t quite understand why Robin disagrees with.)
One plausible way this could come about is one of the scenarios Carl Shulman described, namely that most work will be most efficiently performed by specialized AIs that we would not consider sentient.
But plasma TVs, besides being a luxury, are also a form of signaling that helps attracting a mate, and increases social status in general (which improves reproduction). Maybe the right argument is not “non-human capital” but “signaling”?
I expect some signaling would be part of a subsistence lifestyle. But I don’t expect 99% of spending to be on signaling.
I think spending 99% on signaling by definition (well, by most people’s definition) wouldn’t be a “subsistence lifestyle”, so perhaps a better phrase would be “equilibrium lifestyle”. But how sure are you that there isn’t a high-signaling equilibrium? Maybe not 99%, but say >50%?
I didn’t say I cared more about median income. Learning econ is good for you—I recommend it! :)
Let me put it this way then. Why did you decide to make predictions about median income instead of average income?
Theory makes clearer predictions about medians than means.
Doesn’t that seem like a bias? If what we care about is average income, then we should talk about average income (and about more than just the most likely outcome if necessary), not switch to talking about median income just because it’s easier to make predictions about it.
Also, I think your “nature is doomed” depends on average, not median, income to approach subsistence level. Suppose you have one individual who has half of the world’s income, and the rest are living at subsistence. Clearly that one individual can preserve much of nature by him or herself.
People have been predicting Malthusian scenarios since, well, Malthus, cheerfully indifferent to the continued failure of such predictions to match reality.
Is there any weight of evidence whatsoever that could convince you of the falsehood of such predictions?
The evidence that would disprove Robin is disproving the population growth rate he assumes or finding a way to increase wealth in a super linear manner once we’ve reached the theoretical maximum usage of each atom.
How has Malthus been proven false? From what I recall, he said that overshoot & collapse are inevitable unless people exercise ‘moral restraint’. That seems pretty bang on to me.
Strong security by itself may not be enough to ensure a high global median income, but strong security plus some other effect which causes a large majority of nations to adopt population controls, would. (Without strong security, every nation would have to agree to adopt population controls, which is much less likely.)
The demarcation of distinct “capita” for purposes of “per capita” wealth offers a lot of room for flexibility. If we assess wealth on a per-cell basis (in energy, food, etc), then human wealth is comparable to bacterial wealth within a couple orders of magnitude by some metrics. But if we assess at the level of an ‘organism’ human wealth is astronomically greater.
Future intelligences may demarcate identity in such a way that the members of the ‘proletariat’ are just components of ‘capita,’ without independent desires. Or labor might be done by independent systems that we would not classify as capita for other reasons, e.g. the workforce of Bostrom’s “uninhabited society” in his Future of Human Evolution paper.
So in this scenario, a “capita” would be like a hive mind? Interesting, but what prevents such “individuals” from being driven to a subsistence standard of living by Malthusian dynamics operating at this higher level of organization?
This one seems to be covered by the “Non-Human Capital” scenario.
This is not exactly an answer rather than a comment: Malthusian dynamics implies reproduction of approximately fixed-size (where size includes needs and abilities) individuals.
A hive mind differs from those premises in that one doesn’t necessarily need to reproduce, just grow. It’s not immediately obvious that a mind housed in the equivalent of 1 billion human bodies applies the same kind of arithmetic to resources, happiness, etc.
I don’t have a theory to propose, as we don’t really know that much about non-human minds, but, for instance: If I “added” to myself another (mindless) body, over which I had unusual control (e.g., I could choose whether or not I feel pain from it), and it was worse (say) fed than I am, I don’t feel my “average well-being” would decrease. However, for two individuals it would.
As a closer-to-reality example, if I were to add some cells to my body that are closer to “subsistence” level (e.g., by eating the same but starting body-building), from a cell point of view the median & average “wealth” decreases, but from mine it may raise (if, e.g., chicks dig me more).
The point being that it appears that the resources-per-replicator may decrease yet resources-per-individual increase. Aggregating replicators in individuals at least complicates the math enough that I’m not convinced that’s very improbable.
(Sorry for the long comment, but I got another point of view:)
Even though Malthusian dynamics work somewhat at the level of cells, I don’t give a damn about my individual cells themselves. It just doesn’t enter the calculation. A “hive-mind” aggregating all humans might be the same: a great discontinuity in how calculations are made. (Granted, the result is, in fact, a singleton.)
Operating at a higher level of organization means that the fitness hit for a given level of per capita wealth is less, so that selection will be less intense, so that, e.g. a less extreme military technology situation can sustain a high per capita wealth..
Under the theory of eternal inflation new universes keep being created at an exponentially growing rate. If we could find a way to keep sending our surplus population to different universes we could have unlimited growth.
Being able to create new universes isn’t the same as being able to push 10^70 creatures from our universe into each new universe. You could still have subsistence income for the creatures in each universe.
If the time and capital investment to create and exploit a new universe—which potentially includes traveling into that universe yourself—is not large relative to the size of a single universe, then there is an indefinitely repeatable, extremely high-payoff investment you can make, which can outpace any population growth less than that interest rate. Of course you do have to be able to exploit the descendant universe—including to the extent of using those new resources to create more descendant universes—which again would be satisfactorily handled by traveling into it.
Well sure if each new person could make their own universe. But that is much stronger assumption than just new universes being possible.
But the new universes also have their own population, though I guess you could colonize universes where humans don’t arise rather than universes identical to this one except I didn’t scratch my nose just now
Or just enter the universe early. Suppose our universe were created suchly, and for some reason the creators just had to have a metallic terrestrial planet. Even waiting for all the necessary supernovae, they still have a good billion or 2 years to exploit the Earth before we arose.
And given the Great Silence, it might just be that exploiters don’t need to worry about competition in the new universe.
A “world government” and “unlimited growth” I can understand.
However:
Memes are not really a solution to Malthus’s dilemma. They are just a different type of substrate for inheritance. The world filling with memes might indeed prevent the world from filling up with DNA—but then you have the exact same Malthusian problem all over again with the new medium of inheritance.
The “Disease or Warfare” scenario apparently retains the essential undesirable element of Malthus’s scenario of resource limitation—namely the failure of much of the population to survive to reproduce. The standard of living for those that remain could be high, though.
But memes are not morally relevant entities. Who cares if they suffer from the Malthusian problem, as long as human beings can have a high standard of living? (I’m assuming a non-upload scenario here, in case there’s any confusion about that.)
Yes, that was the point.
I would agree that some agents could improve their standard of living—even if resources are limited—by making use of an army of slaves, which they don’t care about the welfare of.
I don’t think that DNA has a monopoly on moral relevance, though. The idea that the inheritance medium is of significant moral import will probably be derided in the future as a kind of prejudice.
Well, memes might care :-)
By the way, is there such a thing as a meta-meme?
I don’t see how the problem would arise. Genes replicate by increasing the number of people. Memes (of the kind we’d like) replicate by infecting existing people. So by diverting human efforts from genes to this kind of memes, we reduce population growth.
Yes, we (might) reduce the growth of the human population. But then we get memes replicating on human substrate, which means that memes have a Malthusian problem—they reach replicate to subsistence level of “resources”, i.e. humans.
The second order problem is that memes that cause their substrate to reproduce more and pass the meme to children have a reproductive advantage for meme natural selection. Which, if big enough, can lead to a parallel Malthusian problem in both human and meme populations.
When non-DNA inheritance gets better at using the available resources, a likely result is more effective competition for resources with existing organisms—leaving many affected humans unemployed and dependent on alms from the government.
It is true that population growth would probably be reduced in this kind of scenario—since usually the government pays the unemployed just enough to exist on.
From what I’ve heard, people don’t try to have as many children as possible—they try to have as many children as will be able to inherit their niche. This could explain why those who can afford college for their children put money into that for a small number of children rather than into having as many children as possible.
We already have something like Psychohistorian’s nudges. There’s a lot more pressure to not leave a child home alone, and that raises the cost of raising children.
The other observed fact is that birth rates drop when women get rights. Just for a weird speculations—what if children get enough rights that they can choose whether they have siblings?
I find Robin’s theory that we’re selecting for a strong propensity to have children plausible (I came up with it independently), but I’m not sure how it works in practice. There are fringe religious groups, both Jewish and Christian, which encourage maximizing reproduction. I don’t know how stable this is in terms of having grandchildren and great-grandchildren. Limiting one’s number of children pays off in the present, so I wonder about the defection rate from those groups.
This might or might not have much of an effect. I requested a new baby sister when I was four (clarifying that if it was a boy he should be “sent back”). I got one. (I honestly don’t know if my request had a thing to do with it.) I proceeded to regret it for about the next decade (no longer), but she was already there. That having been said, I do plan to consult my future kids before giving them siblings or (should the situation come up) stepfamily. It’d be interesting to see how, if at all, that affects the makeup of families should the practice become widespread.
How about simple spontaneous population stability… I live in a country with negative birth rate but the population is increasing due to immigration nevertheless. This state of affairs hasn’t been legislated into existence, it just happened, and may be a natural behaviour of large human populations. Perhaps once the whole world reaches western standards of living the whole world will stop growing exponentially, with pockets of negative growth being compensated by low but positive growth in others… in the long term the trend could even be for decreasing population...
In the long term (and I mean the very long term) people will evolve to get around the obstacles that stop them producing the children they could.
If contraception decouples sex from reproduction, people will evolve to be less interested in sex and more directly interested in babies.
If entertainment proves more compelling than having kids, people will evolve to be less entertainable.
If being a responsible, well adjusted person is limiting family size, people will evolve to be irresponsible, poorly adjusted people.
Consider an analogy to the cells in our own bodies. Cells can divide (with some exceptions), yet the cells in our bodies do not keep dividing till they run into local resource limits, the equivalent of subsistence limits. There are signalling systems that tell healthy cells when they are “supposed” to stop dividing, and these mostly work. The analog to saying that people will evolve to get around obstacles that stop them from breeding is that cells will mutate till they are dominated by cancer cells. That isn’t the whole story. Our immune system kills off most of the malignant cells we produce—we have social systems at various levels which could do the equivalent. If we (as a global society—a kind of weak singleton) can add layers of control faster than breeding mutations pile up, we may be able to contain runaway breeding indefinitely.
The fact is we as large complex mammals are already locked into a low rate of reproduction, sure given the right evolutionary pressures we could end up like shrews again, but that would take an asteroid strike or nuclear war, the scenario you’re thinking of assumes long term evolution within a very long lasting stable society essentially like ours. In those circumstances genes for successful reproduction will spread through the population, but that’s largely meaningless—if I have the gene for super attractiveness and manage to have 100 kids with 100 women we’re still below replacement rate. The way women maximize their reproduction is by having male kids who are alpha males but in these circumstances an alpha is someone who is good at seduction rather than the old style coercion and multiple wives ownership of old times.
tl;dr the bottleneck for overpopulation is individual women’s fertility, and the way women maximize their reproduction is by having high quality sons rather than popping out babies nonstop. So you can still have high reproductive strategies without actual overpopulation.
In any case it’s hard to think in these terms, the feeling I have is memetics will always overshadow any purely instinctual drives.
If a woman maximizes reproduction (and so fitness) by having more sons than daughters, then why doesn’t the population tilt towards a male:female ratio > 1?
Because Bugle is wrong and seems to be falling into a species of the group selection error.
If indeed women try to have only alpha males (in the sense of some genetic mutation causing her to have more high-quality males and is being put to the test), then other women make a “genetic killing” (genetic birthing?) by having daughters instead. The equilibrium of this game is that of 50⁄50 chance of male/female babies.
(And alphaness is a positional good anyway and so can’t increase in the gene pool.)
OTOH, cultural mechanisms can permit group selection to occur, so a culture could put a bottleneck on its growth rate by endorsing abortion/infanticide specifically of female babies. (China and Middle East, I’m looking in your general direction here.)
And if a culture endorsed killing most male babies (or with modern medical technology, conceived only girls most of the time), and implemented polygamy, they’d be using their child-raising resources efficiently and would increase their growth rate linearly with the number of women.
Thinking about this some more: the reason this isn’t a historical winner is rather obvious—a mostly female population can’t field a big army and is invaded by young males from other groups. Also, the women have incentives to move to a different group where they could have a man all to themselves (or to invite foreign men), so a few men would have to effectively rule and police a mostly female society.
Apart from the aforementioned confusion regarding investment in male vs female children, this isn’t a new balance either.
This occurred to me too, but on second reflection this seems like a special case of the “Selfish Memes” scenario. The meme being “the costs of (marginal) children are not worth the benefits”.
For what its worth, I also think wealth-induced population stability is a real possibility.
Notice how little evidence we have for most human societies in history to be really close to Malthusian population bounds.
It’s really incompatible with evidence unless you take “extreme long term view” like Robin, in which case it becomes essentially non-empirical.
Surely we have good evidence that the growth rate in income per capita was very near zero over a million years The Malthusian theory clearly predicts that near zero number—what other theory do you prefer?
taw, on OB:
No, I don’t think the Malthusian hypothesis says that; in particular, it doesn’t talk about rapidity of recovery, only that population will keep growing in the face of any excess food.
Yes, that would be evidence if there were truly available land. But the wikipedia article you cite seems to say that the land was not safe to farm. It took the shift from “mobile bandits” to “stationary bandits” to have a real supply of land. (Maybe this is what you say about slavery.) I don’t see this complaint as very different from saying that the 400-600 population collapse disproves the malthusian hypothesis.
I think you are misreading wikipedia. I think that the establishment of law lead to higher quality of life, which lead to expansion, driving down the quality of life. That happens to fit the malthusian hypothesis. In particular, that article says that Europe had run out of food by 1350. This could be for malthusian reasons or it could be the exogenous shock of the end of the warm period. I think some people claim that the black death lead to higher per capita wealth than 1200, which suggests malthusian poverty after 1200. Wikipedia merely says higher wealth than 1300, which isn’t saying much, since the climate was deteriorating by then.
The failure of Tuscany to return to its 1350 peak until 1850 is a serious problem for a malthusian hypothesis.
The faster climate, war, and disease change, the less important malthusian effects are. I think that they are visible in midieval european history, at least in the 1000-1300 expansion; others claim to see malthusian poverty in the following expansion. But perhaps they are not important; malthusian effects should probably not be blamed for the poverty of 600. China has a more stable history and, I am told, more visible malthusian effects, particularly the exploitation of poor land.
[My knowledge of High Medieval period is based mostly on listening to TTC’s audiobook lecture series about it (highly recommended), not on Wikipedia.]
A few random thoughts:
I don’t have one big mathematically pretty theory here. (Short-term) Malthusian theory seems to be widely accepted because it’s mathematically pretty, and something like it works on many animal populations. It’s not obvious from either theoretical or empirical point of view it makes sense for humans, at least after they stopped hunting and gathering own food.
Actual food production must cover actual population no matter if Malthusian hypothesis is true or not—if food production was too low, people would die; if food production was too high, there would be nobody to eat it (unless people do something funny with their food instead of eating it). So that’s the wrong thing to look at. Lower population levels must automatically mean lower land area used, or lower productivity/intensity of agriculture. And we can guess people would use the “best” lands, and leave “worse” lands unused.
I think failure of Medieval Europe to rebound back to Roman population levels is much bigger problem for Malthusian theory than Tuscany’s. Small highly urbanized regions like Tuscany might have simply imported food from other regions. Such explanations won’t work for an entire continent.
TTC lectures present as well-established fact that quality of life increased a lot in High Medieval period, while population roughly doubled. The lecturer finds it quite puzzling, as it conflicts with Malthusian theory.
Wikipedia says that during the entire Medieval period, “land was plentiful while labour to clear and work the land was scarce”. This is highly un-Malthusian scenario. Malthusian theory strongly predicts having large excess of labour.
Obviously there must be something that limits fertility. Malthusian theory says it must be potential food production, and that every land that can produce enough food to support people farming it must be used. If the mechanism is something else, we might have highly non-Malthusian (non-poor) equilibria. Wars, and failure to organize expansion to new lands (what in agricultural societies might require significant up-front investment, it definitely did in Medieval Europe, settlers on new lands were freed from usual taxes for many years) are a few obvious mechanisms.
Wikipedia says position of peasants in Western Europe which had high labour-to-land ratios was much better than in Eastern Europe with high land-to-labour ratios. Malthusian theory predicts otherwise. I can see no differences in agricultural technology, or land productivity, or anything else to explain this away.
Increase in wages after Black Death can as easily be interpreted as inflation, not as increase in gdp per capita, as prices seem to have increased correspondingly. It’s not really surprising, as supply of metal money stayed constant, so supply of money per capita increased proportionally to loss of population.
Wikipedia says that Black Death had negative effects on the economy:
We have really crappy data about economic history.
One issue to keep in mind is measuring wealth in calories and in other terms. People interested in malthusian theories focus on calories, while those interested in the advance of civilization may measure it in, for example, cloth. The malthusian theory predicts that the proportion of salary that goes to food should creep up in peacetime. But the advance of cities in the high medieval period made other products cheaper. When Braudel talks about economic recession, he’s probably talking about abandonment of cities, which is not necessarily relevant to the malthusian question. So quality of life as measured by things other than food may have become high in 1300, but the famine of 1315-1317 suggests that the food situation was precarious. Why didn’t they respond to the famine of 1315 by planting more land the following year? Was the land exhausted, the labor still in short supply (unlikely if they were starving in cities), or it takes too long to start new fields?
I don’t think the failure of rebound from the Roman Empire, in itself, is so bad for the malthusian theory. But then the malthusian theory must predict that the shortfall is due to the lack of pax romana. It predicts that uninhabited areas are difficult to defend; or that farmers in such areas lose productivity by switching to crops that are hard to loot.
I think of Tuscany as an agricultural area. I guess this is evidence that I shouldn’t. But I also think of Florence as important Renaissance city. Yet it was importing more food in 1300 than 1600? This is bizarre regardless of malthusian issues.
I don’t have any direct knowledge of agricultural productivity in Eastern vs Western Europe, but I think the adoption of the high calorie-per-acre potato is correlated with low productivity and I think it was particularly popular in Eastern Europe. It is also a hard to loot crop. Corn and the potato are connected to population growth throughout Europe, suggesting that food limited growth before 1500. But some places (France?) didn’t adopt them and didn’t grow as fast, a rather non-malthusian situation.
No, it just says it about 1000-1250. I think it means that raiders had been expelled from previously uninhabited land, allowing expansion. But at the end of this period, or at least by 1350, the situation was overextended. This is one case where I will grant a concern about speed of expansion. We know that in the right circumstances, such as colonial America, population can double each generation. You might expect the same from this description of this situation. This suggests that there is more to human psychology to switch on “high growth” than just food, which we certainly know from the demographic transition, but I’m reluctant to accept the description. The high initial cost of expansion may be relevant.
Which is of course total nonsense, as number of calories consumer per capita varies extremely little over huge differences in GDP. some data
Not at all. Even if there’s plenty of unused land lying around, if you get unusually low yields, you can get famine. Few civilizations farmed spare land and threw away food, just in case the spare food might be useful, so food margins have nothing to do with availability of free land or lack of it. It takes some time to get new land into use, one year at the very least, more realistically many years to cut forests, build settlements, gather people, animals, and seed and move there, make agricultural tools etc. It’s a major project, and conditions of famine do not help at all.
Potato was widely adopted in 18th century, the Eastern vs Western divide refers to Late Middle ages/Early Modern period, which took place long before it. During that time Poland and other countries became major exporters of wheat to Western Europe, using particularly nasty forms of serf labour.
That’s not what I meant about 1350. Clearly if land was plentiful in 1000-1250, it must have been even more plentiful in 400-1000, and yet population wasn’t expanding back to its Roman levels.
Raiders were Vikings / Muslim / Hungarians / etc. and they were raiding entire Europe. There was no distinction between raidable and not raidable lands. Here’s map of Viking raids. Other groups raided other regions. Most of Europe, especially coastal areas were repeatedly raided, these lands were definitely inhabited all the time.
It can, and also GDP/capita was increasing while population was increasing. For pretty much any country you try. Malthusian theory suggests ridiculously high starting GDP/capita, which would then gradually go down as population increased, which is completely wrong.
I’m sorry if I wasn’t clear, but I reject GDP for this purpose. I suspect that Braudel and the conventional wisdom about 1300 are like saying that people today are as kings, for they have the greatest musicians of the century at their beck and call on youtube. GDP tells how nice are the luxuries, but it doesn’t tell if someone is starving.
Malthusians claim that percent of income spent on food is a good inverse predictor of number of calories, that people eat few calories when they can’t afford more. GDP, whether in the 13th century or especially in the 20 century can mask this, because what can be bought with the remaining income is quite variable. Greg Clark claims that the poor in England in 1800 were getting 1500 calories per day, which is off the chart you link to.
I can’t find him giving calorie estimates for other years, but he and his predecessors Phelps-Brown-Hopkins and Steffens claim that English labor income peaked in 1450, that the black death raised wages in a malthusian manner, in contrast to the claims you quote. This could be special to England or could be a result of measuring wages in necessities (calories), or could be a result of malthusians putting a finger on the scales. (Clark is a malthusian. I don’t know about PB, H, or S.)
How is it measured? If you use nominal wages, you will see this effect, Malthusian or not, because amount of metal money per capita is inversely proportional to population. You need some sort of GDP estimates to adequately measure wages.
Do we at least have this data?
It isn’t a terribly useful indicator, as prices are by their nature marginal, and just recently wheat prices varied from 287.75 $/bu in February 2005 to 1280 $/bu in February 2008, 4.4:1. This doesn’t mean people got four times poorer just because food got four times more expensive.
But then, with economic history we rarely have the kind of data we want.
They use baskets of goods. But if the basket weights food heavily, it may see different effects than if it weights manufactured goods heavily. You can call it GDP, but 20th century GDP is definitely measuring the wrong thing. PPP deals with some of these issues, but for each purpose you need a different basket.
This is Engel’s law. He had contemporary (19th century) cross-sectional data, not historical. I have heard people claim to have some historical data like this, but I haven’t run across it recently. Clark seems to claim to have better knowledge of the basket consumed than PHB, so he ought to be able to graph calories, but I haven’t seen him do it. Or rather, he claims to have better knowledge of meat consumed, so he should be able to graph protein, which is another part of Engel’s law. Protein consumption changed much more across the 20th century than calories, but I’d be nervous about cross-cultural comparisons.
No, we shouldn’t say that the people became 4 times poorer from 2005 to 2008, but we should say that poor people who use wheat as a staple and didn’t have (flexible) subsidies did become poorer, while I doubt that’s visible in the GDP per capita. Just saying that they became poorer, without quantifying it, is a crude measure, but we’re only interested in the sign of the change: did the Black Death make people poorer or richer? It may have had opposite effects on the rich and the poor because they consumed different baskets of goods.
Any theories about why human populations typically don’t get near Malthusian bounds?
This is like most of the responses to Robin on OB; extreme examples of wishful thinking. None of the responses, here or there, are stable or address Robin’s point that in the very long term population will probably grow until it reaches the level of subsistence.
In the very long term, concepts like “population” may be hard to measure. The clear identifiable edges between humans may be historical accidents, sometime around the evolution of multicellularity and/or immune systems. Accidents, that with careful refactoring, might be smoothed away.
Do you count bacteria as individuals? Or is the line drawn better around bacterial colonies? For a boundary-blurring example: http://en.wikipedia.org/wiki/Volvox
The Gaia hypothesis argues that our ecosystem as a whole has some aspects of homeostasis, making it in some ways similar to a single organism. http://en.wikipedia.org/wiki/Gaia_hypothesis
Suppose that in the future, civilization as a whole has some aspects of being a single organism, but also some aspects of being composed of a small number of separate organisms. Suppose that each of those components are similarly blurry, and could be construed as being composed of subcomponents, and so on.
This “fractal population” scenario demonstrates that, in order for the prediction “population will probably grow until it reaches subsistence” to be falsifiable, the notions of “population” and “subsistence” need more explication.
I always felt that this very important truism is neglected around these parts. People here often construct thought experiments about the well-being of populations in the very long run. And often these thought experiments become meaningless when considering that in the long run, individua (?) can gradually become parts of larger systems and give up some or most of their free will. (Gradually is an important word here. Any numerical formalization of the well-being of a population must be robust to this, without artificial phase-transitions.)
I call this half-jokingly the “Individualism Bias” at Less Wrong, and was thinking about writing it up as a post. Frankly, it would be better if you did.
The concept of near subsistence income is robust to aggregating smaller individuals into larger individuals. However you group or divide very poor folk, they remain very poor.
Very well, at what level is the earth now? At subsistence, above, or below?
I said at OB; median is 5-10 times, mean is about 20 times.
I can’t find the post on OB that you’re referring to, otherwise I would reply there.
Do these (5-20) figures disagree with estimates that humanity is consuming more “ecological services” than the rest of Earth’s ecology can renew?
http://en.wikipedia.org/wiki/Ecological_footprint
I’m just trying to understand how you compute “income” and “subsistence” for entities like “earth as a whole”. From your tone it ought to be easy, but there isn’t really money, or trade, at this level of organization. Earth as a whole doesn’t buy things from other entities, not with money, nor with barter, so I don’t see how to compute those numbers. The best parallel I can see between human-scale economics and earth-scale is “sustainability”. If a human’s income exceeds subsistence, then I’d say they were in a good situation regarding sustainability. However, the ecological-footprint figures argue that earth as a whole is in a bad situation regarding sustainability.
The usual methods use market prices, which of course would not be available if the Earth had no internal smaller units trading.
Singleton isn’t. Robin both hates it and doesn’t think it’s likely. It makes me very uncomfortable… but a Singleton is clearly not subject to evolutionary pressures.
The only two that seem to have a chance of avoiding Hansonian scenarios in the very long term, IMO, are the “unlimited economic growth is possible” scenario (which requires us to be wrong about physics, AFAICT) or a very powerful and far-thinking singleton. The selection for expansion is an incredibly strong optimization process, and it could only be kept in check by a stronger optimization process with a goal of keeping it in check. But that’s not a priori impossible, nor an especially unlikely subgoal for (say) a FAI.
Given current cosmological theory, our access to negentropy is limited, so that we probably won’t get to see the effects of natural selection go to the limit as generations approach infinity.
Consider the scale of the living standard drop Robin is predicting: roughly an order of magnitude. That’s much larger than the great depression—which set the stage for WWII. A prospect of an order of magnitude drop in living standards would probably be enough to trigger wars of extermination. To put it another way, a cross between “Strong Security” and a policy by population-limiting nations to limits the populations of neighbors as well (with nukes if needed) could be stable.
I’d add:
Superstimuli—New technology creates activities that are so enjoyable that most people just don’t bother having children, or don’t want to have many children.
Default birth control—all individuals (of at least one gender) are implanted with effective birth control at birth/onset of puberty. It’s easy to turn off, and anyone who wants to turn it off can, but you can’t conceive without taking making an affirmative decision to do so. This could lower birthrates enough to hit replacement.
Nudges—like default birth control, but through other mechanisms. Rules and policies are put in place that gently discourage fertility (or gently discourage high fertility) without coercively preventing it. Tax incentives and the structure of social support fall under this category. This impairs high-fertility memes (as they are more expensive to execute) and thus lowers fertility.
These arguments all fail for the same reason.
Superstimuli: “most people” is an important qualifier. The Permian mass extinction eliminated most life on earth, yet it didn’t take long for those life forms that did survive to multiply back to malthusian limits. So in a superstimulus scenario, (like virtual reality/wireheading, etc.) the people who opt out and keep reproducing will continue increasing exponentially, until they hit malthusian bounds.
Default birth control: No-brainer. The people who turn it off have high genetic fitness, those who keep the default birth control have zero fitness. The only difference is the selection is for pronatalist sentiment instead of forgetfulness/irresponsibility.
Nudges: modern society already is punitive toward having many children, as they are essentially consumption goods and not productive capital from their parents’ perspective. Even though children are already costly non-producers, there are sub groups that highly value producing many kids. This cannot change through “gentle” approaches. Natural selection is a powerful force.
I agree with your answers to the other two, but superstimuli still shows some promise. Yes, evolution can always find a way around any particular fixed set of superstimuli, but what if people (say a few rich altruists) keep inventing new superstimuli as a way to limit the human population, and this process works faster than evolution can respond? Is that a plausible scenario?
I believe there are already subgroups that have already developed sufficient resilience to superstimuli that non-coercive measures are unlikely to succeed completely (and if they don’t succeed completely, they fail completely).
Anabaptists are a good example.Those who are seduced by modernity leave, those most orthodox, obedient, and conformist remain to reproduce. For hundreds of years, they’ve essentially been selecting for superstimulus-resistance qualities at the genetic/biological and memetic/cultural levels. I expect the future gains in superstimulus power are probably less than the difference between MTV/Megamalls/internet-porn/McDonald’s/Coca-Cola/Hip-hop and the simple prayer groups and barn-raisings they’re used to.
Of course, those rich altruists might very well use fraud or force to absorb resistant groups. By my values, they would be justified in doings so.
It’s quite possible that defection will outstrip reproduction, particularly as the group becomes large or geographically dispersed and its ability to exert powerful interpersonal influence on its members declines accordingly.
Your other responses assume that the urge to have lots of kids will find a way to dominate all or most other practical desires. This seems unlikely, for reasons I address elsewhere in this thread.
I agree that if fertility choice is determined substantially by genetics in the long run, or if it’s memetic and memes are stably heritable against outside pressure and over many, many generations, you will probably see runaway growth. I just don’t think either of those conditions is particularly realistic given existing evidence; I certainly don’t think existing evidence actively supports either assumption to the extent that I expect people to breed up to a Malthusian limit.
I’d be very curious as to what groups you think tend to show high fertility memes, and why you expect that behaviour to be stable in the very long run, and what evidence you have supporting that belief, if you do think it’s stably memetically heritable.
Anabaptists, as I pointed out in my comment, show high-fertility memes. I suspect they have already selected for genetic high-fertility predispositions as well. (Peter J. Richerson and Robert Boyd wrote extensively about anabaptist genetic/cultural evolution in “Not by Genes Alone”, which I highly recommend.) Anabaptist culture has thus far been highly resistant to the best modern superstimuli—in spite of the fact that their children are exposed to modern society through rumspringa and similar traditions. Therefore, it remains to be shown, by you, why future superstimuli will work (for every reproducing group) when all so far have failed. The vast majority of Amish children, for example, choose to remain with their communities instead of leaving.
The Amish population has been increasing exponentially for decades. They have shown an impressive ability to adapt to new economic models and yet maintain their unique identity, and resistance to modern culture.
Another group (which I suspect has gone even further than the Anabaptists in the strength of its memetic/genetic firewall) is the Mormon Fundamentalists. I’ve been intending to write a (short) post about them for a while, maybe I will get around to it soon.
These are answers to a different question: how could a singleton limit population? More precisely, they’re answers to the question of what a singleton could do now, since in the long run, people would evolve around any particular mechanism that doesn’t wipe them out, so the singleton would have to adapt.
This assumes high genetic heredity of fertility decisions. Binary fertility does seem largely like it has a strong genetic component, but once it’s there, the decision to have two children versus six children seems to be influence overwhelmingly by non-genetic factors. Once industrialization took root, fertility fell quite heavily. Since there would be rather strong selective pressures favoring “have more kids when resources are available,” this suggests that such an impulse, if genetic, is very easily outweighed by other factors.
If it’s memetic rather than genetic, the argument requires extraordinary stability, for which I see little evidence.
Population doubling takes place in alternative branches of a coherent quantum superposition, so that each single branch has plenty of capital. Parents don’t get to talk to their children much, but everyone maintains a higher standard of living—in exchange, perhaps, for lower measure.
You could take the current splitting as showing there is already a vast rate of population increase. However, the analysis of per-capita income is robust to splitting or joining individuals across quantum branches. If people can grow kids within a branch faster than they can grow wealth within a branch, incomes fall regardless of how fast the number of branches grow.
I don’t believe in MWI, but I try to understand it, and I can’t get this to make sense. Reproduction is made conditional on some quantum event, so that it happens in some futures and not in others—OK. These futures are supposed to be in coherent superposition—not sure what the point of that stipulation is; is the idea that the separate futures are given a chance to recombine at some point?? But they can’t recombine unless they come to resemble each other, which is a bit of a problem when one future contains lots of people who don’t exist in the other one! And finally, doesn’t this situation result in a Malthusian scenario in the branch which does reproduce??
This would be accomplished using quantum suicide to edit proportions of the population out of individual world lines. At each juncture 50%, say, of the population would be edited out of each world line. You aren’t gambling on which world line gets to reproduce, you’re putting half of the population into each one.
Correct; and the point of specifying coherent quantum superposition, as would exist in a large-scale quantum computer, is that so that both branches “go on existing”—which I think they do anyway, but I wasn’t sure about the quantum suicide aspect—actually, the more I think about this, the more I start to think that quantum suicide ought to work, since could you really say someone was 50% dead if they lived on a still-coherent quantum computer and only existed in half the branches? Maybe you could. I don’t understand Born probabilities or anthropics.
Your original proposal is extremely close to quantum suicide; in particular the phrase “at the expense of smaller measure” is the key assumption of quantum suicide. This is, as far as I can tell, the rejection of the Born measure. But we observe the Born measure and our normal decision theory values it as probability. To reject Born measure is to say that it does not “all add up to normality.”
I think you’re misunderstanding an important part.
I think the idea is that you flip a quantum coin, and if it’s heads you do nothing, otherwise you replace yourself with your child. So the population of each branch stays the same.
But the coherent superposition part doesn’t make sense to me either. And the scheme also doesn’t say what prevents people from reproducing the normal way and driving up the per-branch population.
Making reproduction conditional on some quantum event sounds like a reproduction lottery to me. It reduces the birth rate, but the main difference from other means of population control is that the decision about who has kids is made randomly—resulting in an eventual decline in genetic quality of the population.
“Singleton
A world government or superpower imposes a population control policy over the whole world.”
it has to be stable essentially forever. It seems to me that no human government could achieve this, because of the randomness of human nature. Therefore, only an AI would suffice.
Not necessarily. A singleton could come to power, genetically engineer humanity with a fertility cap, then fall. In the long run, evolution will probably find a way around such a cap, but in the mean time another singleton could arise and reset the clock.
If predicting the far future is very difficult, why should “future people will be very rich” be different from “future people will be very poor”? Shouldn’t it be just as hard to be sure of one as the other?
Yes, I think to a first approximation it’s equally hard to have confidence in either of those predictions. Which is why I’m not arguing for “future people will be very rich” but rather that Robin may be too confident in “future people will be very poor”.
The big one:
Memetic instability, aka Hanson’s assumptions are just wrong.
It’s quite easy for an armchair evolutionary psychologist to say, “X group reproduces more than not-X. Therefore, by 2150, everyone will be Mormon because they’ll be selected for!” Yes, I’m exaggerating a bit, but the actual argument is about as absurd. There aren’t any large population subgroups with extremely high fertility as a subcultural norm. Given the substantial hedonistic sacrifice involved in high fertility, this is not surprising. It seems unlikely, barring kids being a lot easier to raise, that large subgroups of the population will maintain a high-fertility norm. This is even more true if people measure social status in material terms—high fertility inhibits material acquisition, and people tend to be pretty serious about pursuing status.
The fact that group X, which is 1.7% of the population, has high fertility, does not mean that their great-grandchildren will when they are a larger subset of the population. In fact, mere changes in geography could corrode the memes of such subgroups. The fact that, if every subgroup kept their current fertility for 1000 years, most of us would be Mormon/Muslim/what-have-you does not mean that this will actually happen. Assuming the future will be exactly like the present but more so has been systematically wrong for at least the last few centuries. Memes can change very, very rapidly, and their change is not directed by a meme-designer trying to beat all the other meme-designers; it’s directed by human desires and cultures.
This also ignores the fact that most of the growth that occurs is in the third world, and most indicators show that their fertility falls as child mortality and women’s education improves. Because women’s rights are, generally, rather ratchet-like (they rarely get massively scaled back once they advance), these gains are likely to be sustained. As people no longer think they need 8 kids to support them, they’ll start realizing they don’t want 8 kids, and the old cultural standards will decay in a few generations. It seems like this is already starting to happen. Rather than actually refuting this apparent demographic reality, we get, “People who have more kids and teach them to have more kids will take over in several dozen/hundred generations.” That’s assuming some incredible memetic stability. I do not think that memes have shown that much stability historically, and they are particularly unstable these days.
Granted, it’s entirely possible that technology will make child-rearing all play and no work, in which case Western demographic trends may reverse. The argument doesn’t only cut one way; it does require that people be generally hedonistic and that kids cramp their style. However, there really isn’t good reason to take Hanson’s theory as the default position.
Also: Nothing. We’ll be fine.
Hanson’s assumptions require extremely strong memetic stability, or at least high and consistent prevalence for high-fertility. Since fertility, at the relevant margin, seems principally memetic (rather than genetic), this seems unlikely, particularly if it entails a significant sacrifice of material well-being. If it doesn’t require a significant sacrifice of material well-being, then it will not likely coincide with falling standards of living. In other words, when it becomes clear that the next generation will be worse off than the current generation because there are too many people, it is very likely that anti-fertility memes and possibly laws will become prevalent. Insofar as people are mortal, significant changes in fertility and population can happen extremely quickly.
Edit: Downvotes are interesting, especially without counterargument. “His assumptions are wrong” seems like a legit criticism.
I don’t know about laws, but I don’t see why “high fertility decreases happiness of humans” necessarily implies “low-fertility–causing memes have a reproductive advantage”. Nor why that advantage would shadow that of other memes.
Local (time-wise) changes can be quick and large in any direction, true, but I thought the argument was that as long as fitness selection works, in the long term, it always finally pushes them back towards the Malthusian limit.
So even in an extremely sane and hedonistic population (say, immortals who mostly never choose to reproduce), as long as some inheritable trait (genes, memes or something else) [a] increases the likeliness of reproduction relative to the average and [b] varies among the population, eventually individuals with that trait will dominate because… well, they’re the only ones who reproduce.
(Of course, if society can perfectly control all such factors, then the problem disappears. It’s just that it doesn’t seem possible even in theory.)