Integers and the lowest amount of offspring natural selection will permit.
This article is a reworked section of the “Brood size” article. I separated the two because they did not fit together.
If we were to compute the optimal brood size in a certain environment, we might get something like 5.76 offspring. However, we must remember how that is implemented at the individual level: if a population somehow evolved a lifestyle that allowed it to have 5.76 children on average, that would translate into some individuals having 5 and others having 6 (for the sake of simplicity). One of those would be more adaptive than the other, and so the proposed equilibrium would be unstable. While it might be more adaptive to have 5.76 children than 5 or 6 in a certain environment, an organism cannot have 5.76 children, so the options are only 5 or 6, and whichever one is most adaptive will be selected for.
That optimal brood size can be a biological heuristic, in the sense that it is a consequence of a having a particular life cycle, in a particular environment, but it cannot be a psychological heuristic, because it is too complex. A number like 5 is not encoded in genes and expressed in the structure of the brain. It is encoded, in a sense, in a phenotype as a whole and expressed when it interacts with the environment.
The hard limit on how low the stabilized brood size (or APDR) can get is higher than 2, because the replacement level for sexual reproducers is two offspring per female, and having three is always more adaptive than having two. As discussed above, in an environment of extreme scarcity the optimal brood size might be 2.1, but the practical options for organisms are either two or three, and 2.0 is never stable. So, the minimal ancestral brood size is 3, and the minimal APDR is 33%. In the long run, it is not possible for evolution to select for a lifestyle where organisms produce less than three children over their lifetime.
In other words, natural selection can only brings about whole-numbered and superior-or-equal to 3 brood sizes in stable ecosystems.
A stable ecosystem is one where every population has reached the carrying capacity of the environment. An average of only 2 offspring per individual can survive and reproduce.
Picturing causal relationships between genes, environments and outcomes.
If you are reading this, there is a fair chance that you are nearsighted. Nearsightedness is a very common affliction in Western and East Asian countries, less so in other regions. It is an interesting puzzle, because it has an obvious genetic and racial component despite the fact that it would have been a big problem for most people living in a premodern environment. It is also rapidly becoming more common, suggesting a decisive environmental factor despite the absence of any obvious candidate. The increase began in the 1950s in the West, and in the 1980s in East Asia. It was known to be common among scholars before that, and it is reliably associated with intellectual achievement (which is mostly genetic). It started later than the advent of mass literacy in Europe. It predates the advent of modern electronic devices, and it is too sudden to be caused by the slower-moving process of urbanization. Maybe the television played a role, although if that were the case, it would have been very easy to test: many families in the 1960s still did not have a television.
Myopia is a complex topic, but this post is not about myopia – it is about the complexity. Which mental framework or model should we use when inquiring about individual outcomes?
A good mental framework can help us grasp complex questions easily. I believe the model most people have in mind when trying to disentangle nature from nurture is something like this:
The arrows represent causal relationships. So, if that graph applies to an individual, each one of the factors on the left accounts for a fraction of the outcome, and together they account for all of it. It it applies to a population, each of the variables on the left accounts for part of the total variance in outcomes, and the sum of all “variances explained” is 1.
This model raises more questions than it answers. Let’s improve upon it step by step, by going through various examples. Only well-known examples will be used, to see how they fit into the model. Towards the end, we will simplify it again.
Genes have a direct impact on the chances of dying from homicide, because genes influence (among other things) personality traits and ability for long-term thinking. A more violent and impulsive individual is much more likely to get into a fight. But genes also have an indirect impact: they strongly determine social status. Even if the individual is not particularly violent, genes have an indirect influence on the kind of social environment someone is likely to interact with.
For someone to be in that specific environment at that fateful moment is also not fully determined by a prior state of affairs and genes. It is also somewhat random. So, we need to add a couple of arrows to our model: genes influence the environment, and the environment is partially random.
Let us see if we can use this to model someone’s risk of falling sick with malaria in the fifteenth century. The outcome is having malaria. It has genetic factors (the presence or not of the allele responsible for sickle cell disease), environmental factors (being in an affected region at the time of an epidemic) and it is partially down to chance. The environment in question, however, is not meaningfully affected by genes, since almost no one would have migrated in the north-south direction at the time.
Is the environment the person finds themselves in fully random? Well, no, in all likelihood it is the environment where their parents and a lot of their ancestors lived. So, we might add another cause to the environment: the past environment.
There is a problem with that: some of the past environment was the ancestral environment, and that had an influence on genes through selection. So, we need to refine the model again.
Some islands in the Pacific have obesity rates in excess of 70%. This is caused, as all things are, by the interaction between their genes and their lifetime environment. Their lifetime environment is (for the most part) the same island biome their ancestors used to live in, a social environment that is partly the expression of their genes, and… the introduction of a type of diet that was not a part of their ancestral environment. So, we should add a “Non-ancestral environment” rectangle.
This is a distinction worth making, because the ancestral and non-ancestral environment are neatly delineatedcomplements: every outcome ultimately comes from these two causal sources, bar some random noise.
The increase in abstract reasoning ability that followed the invention of agriculture (documented in The 10,000-year explosion by Cochran and Harpending) was brought about by a positive feedback loop: a more demanding settled lifestyle causes an increase in intelligence, and that increase in turn causes society to become more complex and impose stricter rules. So, genes are not only causal towards the lifetime environment, they were causal towards the ancestral environment as well.
Lastly, what is the difference between homozygotes and heterozygotes? Obviously, homozygotes share the same gene variants, which is why they are a fruitful natural experiment: they provide a control for genetic effects.
However, this exposes a problem with the causal model above: heterozygotes have the same ancestral environment, and yet they do not have the same genes. So, we should add another arrow, to account for the fact that sexual reproduction randomizes the inheritance process. After that, the model seems complete.
This is a lot messier than the original, but it better reflects reality. This model readily maps onto almost all questions of genetic inference, and helps us understand what is being described. In a coming post, I will describe how it applies to various outcomes: the spread of fashions among the schizophrenic, the lack of influence of parenting style on adult achievement, and the modern pattern of low fertility.
As I said earlier, it accounts for all of the causes of the studied outcome, in an absolute sense. However, we can simplify it, because in practice we are never studying outcomes directly: we are studying how much of the variation in an independent variable explains the variation in a dependent variable (the outcome). In a large sample, we can ignore the contribution of randomness, because the noise cancels out. This gives us the final result:
To be clear, randomness can influence the population as a whole as well as an individual. We can only ignore it when we are considering an “averaged out individual” within the population.
Stupid elites theory is a meme theory that is often brought up as a joke when discussing current events. As we will see, there is truth in mirth. The theory describes the lack of competence and foresight at the top echelons of Western societies. In this essay, I will go over some examples, ballpark where these elites stand cognitively, and discuss some causes. I will mostly focus on political elites.
Revealing events
In late 2019, news of a disease propagating rapidly near Wuhan, China were making the rounds on most platforms, including the legacy media. As the weeks went on, it became clear that things were getting out of hand. We were seeing videos of overwhelmed hospitals and local doctors sounding the alarm. After an initial attempt to cover up the events, the Chinese government imposed drastic measures, quarantining Wuhan (a 13-million people provincial capital) and a few neighboring cities on January 23rd, 2020. Despite these efforts, isolated cases were popping up left and right, including in neighboring countries such as South Korea. At that point, all we knew about the disease is that it was highly contagious, seemed to cause a number of those affected to need medical attention, and that the CFR was low when people could be treated, but potentially a lot higher when they could not. From their reaction, we could tell that the Chinese government thought it was hazardous enough to warrant placing tens of millions of people under lockdown – a previously unheard of policy – despite the economic fallout.
Some decisions are only easy to criticize in hindsight, but this was not that. The problem was not very complicated, and while implementing proactive policies is difficult in democracies, Western political elites almost uniformly failed the test of discerning the threat. Instead, they remained in blissful ignorance and denial even as local epidemics were spreading around the world. I remember specifically that for a while, a region of Italy (Lombardy) was suffering a significant outbreak, and the predominant discourse in France was about the differences between our two countries that caused them to have an outbreak and not us. Experts speculated about party culture, the aging Italian population, diet, infrastructure, population density etc. instead of the obvious explanation: in their early stages, epidemics grow exponentially, so a one-week difference in how early the outbreaks happen can manifest as a 10-fold difference in the number of cases at a given moment.
Amusingly, this was not limited to political elites. Medical professionals joined in the fun, just like the WHO and CDC had done towards the beginning of the crisis. At that moment (before the first wave), it would have been reasonable to advise social distancing, build temporary hospitals, gather medical supplies and subsidize our industrial capacity to quickly produce respirators, masks etc. Contact-tracing, limiting international travels and imposing quarantines on small affected areas would also have slowed down the spread of the disease. None of that happened, because our political (and in this case, medical) elites could not apprehend the nature of the problem.
When it hit, they went into panic mode and imposed repeated lockdowns on whole countries. These were pointless by that time – what are you containing if outbreaks are already everywhere? They scrambled for medical supplies and did a 180 on their risk assessment. This was particularly absurd, because it was precisely around that time (late march) that the picture was becoming clearer: the CFR was very low if people could be treated, and would probably not go much above 1-2% even if they could not. So, we got off easy. It was fairly serious, but only 2-3 times worse than the flu, because it was more contagious. Very manageable.
When the disease was unknown and could have been very bad, they failed to recognize the danger. When the disease became better known, they failed to quantify the danger. Our elites underreacted for months, and then overreacted for years. For astute observers, that was a learning opportunity: it is very rare for nature to provide a collective psychometric test for our decision-makers. What did we learn? – Like common people, elites tend to uncritically adopt the beliefs of their social circle. – Only a very small fraction of them can understand exponential growth. – The independent-minded rely on naive historical pattern-matching (SARS, AIDS).
Astonishing data
This famous article by Anatoly Karlin attempts to quantify, in his terms, the “idiocy of the average”. He is not discussing elites, but average 15 and 16-year-olds. This article is useful for getting a feel for what normal people can do. Since people learn some things as they grow older, I don’t think it is wise to take those figures as they are, but they provide a basis for two mental shorthands I have abstracted from this source :
– In a typical Western country, roughly 10% of people can solve the “Helen the cyclist” problem. (It takes Helen 9 minutes to move 4 km and then 6 minutes to move 3 km. What’s her average speed in km/h? Using miles instead of kilometers gives the same answer)
– In a typical Western country, roughly 50% of people can solve the “revolving door” problem. (How many people can go through a tripartite, four-rotation-per-minute revolving door, large enough that two people can enter side-by-side in the space between wings, in the span of 30 minutes?)
If you can solve both problems easily, you probably found the first one easier than the second one. Fewer people can solve it because taking an average is a more abstract mental operation than multiplying numbers together. This is counter-intuitive, but what makes a problem out of reach of people below a certain threshold is not its complexity, but how abstract it is. In the previous part, I mentioned how the initial stages of the Covid pandemic were characterized by exponential growth: the problem was not complicated, but it was a little too abstract. Conversely, regular people can solve surprisingly complex problems when those can be broken down and no single step is too abstract – like managing a store.
While the “revolving door” problem does not tell us much about elites, the “Helen the cyclist” problem does. Because of their IQ (110-115), we know that political elites in a given country largely come from that fraction of the population – the 10%. This gives us an idea of the minimum level of competency of this group – by and large, they can solve the “Helen the cyclist” test. The smarter people in that fraction will probably go on to work in the private sector, but even the remaining ones are likely to be able to take averages and do simple multiplications without a calculator on hand.
However, it appears that they cannot do much more than that. As noted earlier, exponential growth is too abstract for almost all of them. A simple probability question, which is more abstract than an average but less abstract than exponential growth, tripped up roughly 60% of British MPs. Other surveys of officials in different countries provide similar results, although they are less clear-cut.
So, to sum up, a decent ballpark estimate is that our political elites tend to fall into a category of people who can solve the “Helen the cyclist” problem, but not problems involving abstract modeling, feedback or distributions. Simple probabilities are, fittingly, a coin toss. Conditional probabilities are likely just beyond reach of such a group.
Confirmation bias
While we should be aware of it, confirmation bias is not inherently a bad thing: it’s just the mental process we default to when we evaluate data in light of an existing theory. The only way to evaluate conflicting theories is to spend some time with each of them, allowing the “bias” to settle in, and then see if the picture fits. Smart elites theory and Stupid elites theory are competing theories. We can put them on and take them off like glasses. Once our eyes have habituated, we can keep them on for a little while and see if anything appears blurry or distorted in day-to-day life.
In my experience, wearing the Stupid elites theory glasses makes the picture much crisper. The mysterious machinations of governments around the world are a lot less puzzling when taking into account the very limited agency of those who run them. For the most part, they are neither macchiavellian geniuses optimizing their profits nor tools being puppeteered by decisive private interest groups. They are short-term thinkers who believe they are doing the right thing and try to impress others like themselves with their wealth and status. People in governments and legislative bodies default to price control and subsidies because they believe it solves problems. They do not see a problem with legal immigration because they do not think about it quantitatively and supporting immigration is high-status in their social circle. They grandstand and saber-rattle because it sounds cool, or tough. Trump is not playing 5D chess, nor does Putin know what he is doing. They and their counselors mostly know how to put on their very-serious-people face. Their worldviews are murky, incoherent, and based on false premises.
If you do not subscribe to it, spend some time with Stupid elites theory. Allow yourself to see the world through that lens for a little while.
What to make of this?
Political elites are a varied bunch. They are not all idiots. They are a group, and every group has outliers. They are smarter than regular citizens, but by less than a standard deviation. They also differ from normal people in other ways, such as character. Political elites must be somewhat more competitive, social and status-conscious than normal people.
Elected officials might be a special case to some extent, because winning a constituency involves persuading regular people to vote for you. In some cases, being relatable and not too high-achieving is an advantage. However, I doubt that this makes elected officials profoundly different from other political elites, because at the upper echelons those two groups exchange seats very often.
Elites are selected from a substrate. Their composition depends both on the substrate and the selection mechanism. A critique of our elites begins with a critique of the people: stupid people have stupid elites. So, in the long run, the best way to have better elites is eugenic population control and avoiding the mass import of third-world immigrants into the West.
Regarding the selection mechanism, putting into place requirements to access key positions would be easy enough, as well as reducing the general overproduction of elites by placing a cap on how many BA/BS, MS, PhDs, etc. can be awarded in a given year. Money incentives would probably be counter-productive, considering that they already tend to be paid fairly well in Western countries.
We should keep in mind that every smart and diligent individual who branches into politics has to come from somewhere, and that involves tradeoffs.
There is no miracle cure, but there is no harm in seeing through the smoke.
I recently took a frequentist statistics course, in order to gain a more formal understanding of this ubiquitous tool. It helped me realize I harbored some misconceptions about statistics, in no small part because of a sloppy understanding of the vocabulary. In this blog post, I will go through the essential vocabulary of statistics, with examples. I initially wrote this for myself, as a way to better grok subject assignments.
A random variable is not a variable in the common sense of the term. A random variable is a function that attributes a certain probability to every possible outcome. So, a discrete random variable could model the behavior of a dice, attributing a probability of 1/6th to each face. A continuous random variable could model the height of a person, attributing a certain probability to every possible interval of heights (say between 5’11 and 6 feet), with higher values near the mean and lower and lower values as we go further from the mean.
A sample is a set of identical but independent random variables. So, a sample could represent the behavior of many dice thrown together, or it could model the heights of a large population.
A realized sample is a set of values that represents one possible outcome for a sample. So, a realized sample for a handful of dice could be {4 ; 2 ; 2 ; 2 ; 1 ; 4 ; 6 ; 1}. A realized sample for the heights of a sample population would be a list of measured heights. A realized sample is not the same thing as a sample, but realized samples allow us to approximate the underlying parameters of the sample.
The central limit theorem underpins almost everything people normally do with statistics. The central limit theorem states that if a sample is sufficiently large, it is guaranteed to follow a normal law, regardless of the random variable that serves as a basis for the sample. There are counter-examples to this, but they are rare.
A normal (or gaussian) distribution is a bell-shaped function that encapsulates the behavior of large samples. It is very practical, because a normal distribution is made up of only three components: the standard normal distribution, a certain mean, and a certain standard deviation. So, subtracting the mean and dividing by the standard deviation allows us to model the behavior of any large sample using a known random variable: the standard normal distribution.
In practice, we often do not know the mean, standard deviation or other parameters of a distribution we are trying to study. We have to use estimators. An estimator of a certain parameter is a random variable – a function – which takes a sample as its input, and outputs a normal distribution centered around the parameter (if unbiased). If fed a specific realized sample, the estimator outputs an estimate. So, in these terms, if you were trying to model the distribution of heights in the population of a big country, you could take a sample of 100 people. The realized sample would be a specific list of 100 heights. A function that adds up all the heights in the sample and divides that by the size of the sample is a unbiased estimator of the mean height of the population. If fed a specific list of 100 heights, the estimator provides an estimate. This standard estimator is called the empirical average, but we can also build custom estimators based on how likely they are to be accurate given a certain realized sample.
Confidence intervals give us a range of acceptable values given the null hypothesis and a chosen probability of error. The methodology typically goes like this: we choose a null hypothesis, such as an average height of 5’8 and a standard deviation of two inches for the population of a country. We then choose a value for type I error, which is the probability that the average of a realized sample falls beyond a certain distance from the supposed mean of the population, under the null hypothesis. A 5% type 1 error gives us 95% confidence that feeding our realized sample to the estimator should output a value that falls between the bounds of the confidence interval. Conversely, if the estimate falls outside of the bounds of the confidence interval, we can reject the null hypothesis with 95% confidence, i.e. the idea that the average height is 5’8 is probably wrong, given the realized sample.
Type I error is the probability that the null hypothesis was correct even though we rejected it. Type II error is the probability that the null hypothesis was wrong even though we did not reject it. In other words, these errors correspond to false negatives and false positives. We cannot choose an arbitrarily low value for type I error, because in doing so we increase the probability of type II error. An estimator that rejects very few correct hypotheses will let through many false hypotheses, and vice versa.
Instead of directly choosing a certain value for error, we can instead compute the value for error which would tip our decision towards rejecting the null hypothesis, given a realized sample. This is called a p-value, and it is used as a shorthand: the smaller the p-value, the higher the certainty with which we can reject the null hypothesis.
If you find issue with the way I describe things, feel free to correct me. I am not credentialed, just an interested citizen. I took a class on bayesian probabilities but for now, I remain entirely unfamiliar with bayesian statistics. I am also ignorant of the details of principal component analysis, but hopefully not for long.
No species is ever adapted to its current environment. It is adapted to the ancestral environment: the environment its ancestral population lived in, and was selected by. Sometimes, the current environment resembles the ancestral environment, sometimes it does not.
Ancestral death rate
I would like to propose a humble construct that I will call: ancestral premature death rate. (APDR) When a population exists under a stable environment for a long time, its behavioral phenotype becomes stable. This includes the premature death rate corresponding to its stable lifestyle. This is what I call ancestral premature death rate.
The APDR is useful as a mental shorthand because it captures several characteristics of a population. It is directly related to ancestral brood size: when a population is at the carrying capacity of the environment, an average of only two offspring per individual survive long enough to complete the reproduction cycle themselves. So, if the average amount of offspring produced per individual over a lifetime is six, the APDR is 66%. If it is 200, the APDR is 99%.
Table by ZeroContradictions
The APDR also reflects the selective pressure a species was under in the ancestral environment. As a consequence, it influences the level of genetic noise we should expect in a population.
Edit: the author of The phenocentric theory of biological purpose (I prefer the subtitle) messaged me to bring up a good counterpoint, so this requires a caveat. Selective pressure is a function of both premature death rate and premature death randomness, so the APDR is not the only thing that factors into the level of genetic noise a stable population would end up with. The degree of randomness might even counteract the effect of a high premature death rate, because a high survival rate or a high mortality rate should proportionally de-correlate survival from behavior. In his own words:
I don’t know if the point about genetic noise is correct — in a high mortality species, presumably the deaths are more random, and thus less “selective”: e.g., an oyster produces a million offspring, and they are eaten by predators in a way that is essentially random, having very little to do with their behavior.
We could say that the “intelligence” of the oyster is this mass-generation of offspring, which “searches” the environment for available niches, while the intelligence of an elephant is the brain of the elephant, which helps its children to avoid death from predators, etc.
With humans, culture adds some randomness, which is another wrinkle, but I don’t know if oysters would be less genetically noisy than elephants, simply because they have a child mortality rate of about 99.9998%, versus, say, 75%. The death of an elephant might be much more correlated with its genes, including mutations.
If either a high mortality rate or a high survival rate de-correlate outcomes from behaviors, an APDR of 1/2 should be where outcomes are most correlated with behavior. So, in practice and since almost all species have an APDR between 1/2 and 1, the effect is unidirectional. I wonder if it compensates (or perhaps even overcompensates) for the effect of a higher/lower APDR on the noisiness of the genome. There’s an idea for someone’s research paper.
Back on topic:
The APDR captures much more exactly what people attempted to describe with r/k selection theory. Different circumstances, or lifestyles, bring about clusters of related adaptations. For a given population, suffering from frequent predation while having access to abundant food would favor a high APDR: lots of offspring, many of which do not reach adulthood. If the population was limited instead by scarcity of resources (prey, vegetation, territory, etc.), a lower APDR could be selected for. This would necessarily be the case for animals who are at the top of the food chain (such as people) but also for animals who do not suffer from predation very much, such as elephants. As we know, r/K selection theory is bogus, but the patterns it purports to explain are not made up. Some are, such as the notion that “r-selected” animals are somehow inferior in quality to the “K-selected” ones, or that the former live below the carrying capacity of the environment. The genome of high APDR species (like rabbits) is not noisier, and they do not perform worse than low APDR species (like wolves).
I presented the relationship between predation, ressources and a higher APDR as causal for the sake of simplicity, but in reality the relationship is more complex. It is more accurate to say that these things emerge together as an equilibrium. Possible equilibria fall on an axis with “relatively high death rate from predation, low death rate from scarcity, high APDR” on the one end and “relatively low death rate from predation, high death rate from scarcity, low APDR” on the other end, but there is no one-way causal relationship, only correlation.
I should state that humans, especially in agricultural societies, were not exactly apex predators, because disease is a kind of predation and we used to die a lot from disease. We also predate on each other. That goes towards explaining the anomalously high APDR of humans.
The greek prefixes “eu-” and “dys-” stand for “good” and “bad” respectively. The words eugenics and dysgenics both imply a value judgement. They imply that there are good and bad genes, and so they can only be defined with respect to a value. In the past, some people holding very silly values co-opted the words, and gave them an undeserved reputation. The more reasonable way to define both eugenics and dysgenics refers to the stability and perpetuation of society (or civilization) as an underlying value. In left-wing parlance, that value is usually called “sustainability”.
Dysgenic processes fall into two categories, which I will discuss and contrast in this post. They differ in cause and urgency, but they are similar in that both are unsustainable in the long run, and require either scarcity or top-down constraints (but not top-down management) to be stopped or reversed.
The first type of dysgenics has to do with the natural degradation of the genome under relaxed selection. Each individual is born with an average of roughly 60 new mutations, which is simply the human mutation rate (1.1×10-8 per site per generation) multiplied by the amount of DNA to copy (3 billion base pairs, or 6 billion nucleotides). 60 seems like a small number compared to 6 billions, and 98% of our DNA is “junk DNA” which does not code for proteins, although some of that non-coding DNA has other functions. Let us optimistically assume that 95% of those mutations are completely inconsequential, leaving us with 3 new mutations per generation that do have an effect on the organism. The human body is very functional. Some of its systems are very sensitive, others not so much. Some genes only affect one area (or function) of the body, others affect the whole body. Every generation, nature randomly throws three spanners in the works. In almost all cases, it simply makes the body slightly less functional. Sometimes, it causes a catastrophic failure of the organism. Once in a blue moon, it is an improvement. In a population, slightly harmful mutations accumulate until their increase is offset by selection. The level of genetic noise (also called genetic or mutational load) in a population is therefore determined by how much selective pressure the population is subjected to. It stabilizes at a point where mutations are removed from and added to the population at the same rate. One in, one out. The “addition rate”, as discussed, is fixed. Only the “removal rate” varies, and it increases when genetic noise increases, as more and more organisms become too dysfunctional to complete the life cycle. When the selective pressure is intense, genetic noise stabilizes at a low level. Under relaxed selection, genetic noise accumulates until it balances out with however little selective pressure is left. In modern circumstances, since almost every child lives to adulthood, we can confidently predict that genetic noise is on its way to stabilize at a much higher level than it used to be. I suspect there is a hard limit on how dysfunctional people can be without causing society to collapse. Should we hit that limit, genetic noise will not have time to stabilize: selection will return with a vengeance. We will permanently lose modern civilization, and will suffer a mass die-off.
How quickly do the effects become noticeable is a matter of some debate, but there can be no reasonable doubt that it is happening. Although it is a fast process on evolutionary timescales, it might be a slow process on our own timescales. It also does not have to be a whole-body process. There can be relaxed selection in a specific area, while other areas remain untouched.
Cave animals have a tendency to lose their sight and colors as a result of living in a dark environment, where selection for eyesight and camouflage is relaxed. Mutations accumulate and break down some of their body parts, and there is no selection to eliminate these mutations since, in complete darkness, being blind makes no difference. To be fair, there is probably selection at play here as well: making eyes comes at a cost, so if they are not being used it is beneficial to save the energy. Regardless, the structure would break down over time in the absence of selection.
Child and young adult mortality have been low for a few generations now. It is conceivable that the effects of relaxed selection on humans are making themselves felt, but I am not aware of any research or set of observable facts supporting this. I can only presume that this is a slow enough moving process that it hasn’t taken too much of a toll on our genome yet. My hunch is that the most fine-tuned systems (like the eyes) in our bodies would be most sensitive to higher genetic noise. However, the recent epidemic of nearsightedness seems to be occurring too fast for genetic noise to be the explanation. So, for now, the process is not observable.
The second type of dysgenic process is easier to understand. It occurs when a society saps its own foundations by selecting for traits that are incompatible (in the long run) with its existence. I will go through a few examples.
The elephant in the room is the modern welfare state, which subsidizes the reproduction of people who (by definition) do not have the means to support their own children, at the expense of those who would. Evolutionary theory predicts that whatever predispositions welfare recipients have that makes them unable to support their own children will be selected for, and whatever predispositions welfare contributors have that makes them able to support theirs will be selected against.
Another thing is widely available contraception, which eliminates socially responsible people with long-term planning abilities and favors more irresponsible people. Long-term thinking and social responsibility are valuable traits in modern society, and having access to contraception selects against that.
We could imagine a civilized society that selects for violent, impulsive, sexually promiscuous males over the more disciplined type that built and maintained the society. I do not believe this is happening in our societies, but it would happen in any population (human or not) that had a high level of paternal uncertainty. If paternity is not established clearly, the optimal male strategy is to compete with other males for access to more females. If paternity is relatively certain, cooperating with others to secure resources for investing in your own offspring becomes a viable alternative. In other words, the pair bonding adaptation is conditional upon knowing who sired whom. We can observe the results of high paternal uncertainty in tournament species.
I should state that I am not passing a value judgement on welfare recipients or contraception users. I am passing a value judgement on the system that makes the former more common over time, and the latter less common over time. The value judgement is: that system is unsustainable, which is bad for all of us. I believe it would be fine to keep contraceptives available as well as a moderate amount of public welfare, as long as we stabilize the system with eugenic population control.
The first type of dysgenics is not an urgent problem, but it provides a clean, principled argument in favor of eugenics. Genetic load must be kept at a low level, and that requires selection, either natural or artificial. The second type of dysgenics is a more pressing issue, but it is also a lot more controversial, because it is caused by policy. It is possible to make an argument for eugenics based on it, but it is more explicitly about values, and constitutes an indictment against some social policies (at least in the absence of EPC).
These problems are different, but related. They should be solved together, or nature will solve them for us as it always has: through systematic, widespread premature death.
This essay is a reworked section of the post Designed Culture, for easier reference. It discusses parasitic and symbiotic memes, an idea introduced by Richard Dawkins in The selfish gene. The fashion | tradition distinction is attributed to Blithering Genius. To my knowledge, the essay God is a telomeme is the earliest mention of it.
Memes are competing for real estate in the information space of our minds. Unlike bacteria in a body, they cannot proliferate (create identical copies of themselves) in a single mind. For a meme to reproduce, it requires a fresh mind, although there is a huge amount of available space for a variety of different memes in a human mind. Some memes can be thought of as conscious knowledge, but most of them lie below conscious awareness. Most memes lodge themselves in our brains without us noticing their presence.
That subconscious memetic soup is what we call “deep culture”. Some memes are passed down in families, and those are mostly beneficial because their content is regularly trimmed by the process of natural selection, like the DNA of symbiotic bacteria. If such a meme mutates in some of its hosts and reduces their fitness, natural selection will eliminate the mutated versions. If the mutated meme increases the fitness of its hosts, they will proliferate, and so will the meme.
Some memes are acquired through exposure outside of the family sphere, and those are only modified indirectly by natural selection creating resistant hosts, which in turn selects for less harmful versions of the memes – similarly to how the immune system creates a selective pressure on parasitic bacteria. This process takes longer*, and a parasitic meme can never be made completely harmless because just like a bacterium needs energy, a meme needs a slot in a mind. In other words, even a relatively harmless fashion is probably taking the place of a useful tradition.
This is the crucial part of the fashion vs. tradition dichotomy: all memetic systems evolve over time, but they are under different selective pressures depending on their main mode of transmission. Roughly speaking, there are two modes of transmission that create significant differences in how ideologies evolve: adult-to-adult or adult-to-children.
Adult-to-adult transmission is what is called a fashion: over time, it will acquire characteristics that make it better able to spread in human minds, because variants that plug better into the human psyche tend to “reproduce” better and become more common. Given enough time, a successful fashion will seamlessly acquire a victimhood narrative, villains, aesthetics, outrage-inducing taboos, persuasive rhetoric, etc. In principle, a fashion doesn’t have to harm its host to exist, but in practice it often does, because it is beneficial for a fashion to be able to “take over” the host and make them spend time and energy promoting it. It also competes with other, potentially more useful memes. In this regard, a fashion behaves a lot like a transmissible disease, which is why I compared them to parasitic bacteria.
Adult-to-children transmission is what is called a tradition. The characteristics a tradition acquires over time also make it good at infecting minds, but in addition to that, it tends to favor high fertility in its hosts. A tradition is better able to spread if its hosts have lots of children, since it is mostly transmitted from parents to their kids. Hence, traditions evolve more slowly as they are being passed down through generations, but they win out in the long run (in a stable environment) because they do not destroy their hosts: their ‘interests’ are, to some extent, aligned with the genetic interests of their hosts, and so they are not progressively weeded out by their hosts evolving psychological immunity. They tend to acquire fertility-promoting traits such as: homophobia, sexual contracts (marriage), valuing children, resistance to change. They are most comparable to symbiotic bacteria.
This is a rough model to understand culture, not an endorsement of traditionalism or conservatism. Every tradition began as a fashion, and as an adult reading this, you can no longer acquire traditions. Adopting a tradition as an adult means that, to you, it is a fashion – chosen for its persuasiveness.
* It can take an extremely long time. By way of analogy, smallpox maintained a TFR of 30% after millennia of infecting humans.
We are the builders and depositories of a worldview that has yet to be named. Elements of that worldview make it persuasive, but it does not seem very good at spreading horizontally. I suspect that, should it persist for a few generations, it would become good at spreading vertically. In other words, that worldview would not necessarily be a powerful fashion, but it could become a powerful tradition. If so, it has a bright future ahead, and of course I want it to.
In order to be transmitted, a worldview must be taught. Some of the concepts underpinning that worldview are too difficult for children, most teenagers or even most adults. This is not unique to that worldview, but a great challenge to overcome nonetheless. The problem can be expressed as follows: how can I transmit knowledge that I cannot teach?
An overlooked strength of our worldview is that it is strikingly beautiful. Some people have a visceral reaction against its amorality, its desacralized notion of human nature and some of the political propositions stemming from it, but contemplative types should see the beauty in this world of matter, information and experience. Above, empty skies. Below, the violent farandole of function-fitting forms. Inwards, abysses. Our evolved and acquired intuitions are liable to ensnare us into delusions. Even so, they provide the contemplative with countless profound truths to uncover. Treasures, dragons to slay, a home to come back to.
Some aspects of the worldview are more than simply beautiful. They are astonishingly ordered. They will not only speak to the contemplative, but to the mystical as well. The mind craves mysterious patterns to latch onto.
Examples abound, here are a few.
The circles of order and chaos are a cosmological explanation of the world. Every circle of order emerges from the circle of chaos around it. Each domain obeys its own law, but is constrained by the law of the domain it emerges from.
The pursuit of higher abstraction is a challenge to the individual to deepen their understanding of phenomena through models and the questioning of models. It is an incitement to live philosophically, or in other words, to avoid being the caterpillar.
Family and civilization are the twin chosen values of the individual, allowing for true goal-driven behavior within society, and the projection of meaning onto life. It is the consequence of a function-fitting form freely choosing to become a form-fitting function. A thing choosing to become purposeful, and to expand that purpose to the society it is embedded into.
The fable of Easter Island is an edifying story about the inhabitants of a distant island in the Pacific. It illustrates the violence of human nature and the powerlessness of individuals to solve collective problems on their own.
I am not proposing that you should forcefully indoctrinate your children or other people. If you wish to get an idea of how different your children are likely to be from you, consider how different you are from your own parents and siblings. But beauty and mystery are ways to impart knowledge without having to explain it. They are irrational methods, that can be put to the service of rationality. I believe they are instrumental in allowing a memeplex to become a tradition.
Imagine if the entirety of your visual experience was reduced to a single point. You can imagine a pixel if that’s easier, but in principle it should be a single point that can change in color and brightness. You could also imagine that the entirety of your visual field is filled with one uniform color and brightness that varies over time. (a point, except it’s all you see)
Let us reduce this further and assume the point is colorless, or is always the same color. It only varies in brightness. It lights up when you are born, and turns off when you die. Over your lifetime, its brightness varies. Some changes might be slow, others very sudden, but they are all continuous. If you slowed down events by a large enough factor, then to a good approximation, sudden changes would reveal themselves to be gradual.
Pick an arbitrary unit of brightness. The unit does not matter, and it does not matter either where you draw the line between “bright” and “dark”. Let’s imagine the unit of subjective brightness is a barnard, or the equivalent of looking at Barnard’s star for one second (it appears to be very dim in the night sky). If you summed up all the barnards you had perceived in your life, you would get a positive sum: 15.28 barnards here, 1500 barnards there, add it all up and you get a certain amount.
If you summed up all the changes in barnards, however, you would get a different result. From one second to the next, you could get +50 barnards, or -873.2 barnards. Over a lifetime, it would be impossible to record every change in brightness that happened, but you could know one thing: between any two moments of identical brightness, for every increase there has been a corresponding decrease. Between two moments of looking at Barnard’s star, the sum of all changes is 0. This is true whether the two moments are 5 minutes apart, or decades apart. It remains true if the moments we consider are the moment of your birth and the moment of your death. The choice of barnards as a unit does not matter either, as the conclusion would be the same with siriuses, or suns. The choice of seconds as a unit of time is also arbitrary, although the smaller the unit, the more precise the claim. (in other words, the granularity of time matters)
Now, let’s take a moment to think about perception. If your visual experience was reduced to one colorless point, would you perceive barnards, or would you perceive changes in barnards over time? The answer is the second. You might object that looking directly at Barnard’s star, you are able to see/identify it even though its brightness is constant, but you are seeing it against the night sky, in this case. If that light was the entirety of your visual input (or if your visual field was one unchanging color, one unchanging brightness), you would not see anything. Nothing would register as a visual phenomenon.
One interesting aside is that we are able to perceive changes in changes. For instance, you might notice that the light has stopped changing, and deduce that you are looking at an object of constant brightness. It does not affect the conclusion (the sum of changes is still null), but it’s interesting that we can experience the second derivative of brightness and other dimensions.
So, in our thought experiment, we have determined that our perception of that point of light is really a perception of change, and change in change. This allows us to know that the sum of our visual perceptions is 0 between two moments of identical brightness. Let us expand on that.
Human visual experience typically has seven dimensions:
Brightness
Redness
Green-ness
Blue-ness
Left and right (or any spatial axis)
Up and down (or any spatial axis orthogonal to the previous one)
Depth (because we have two eyes, and we can move relative to objects)
It does not matter very much whether we consider other colors to have their own axes. The structure of the human eye suggests that other colors are composites of red, green and blue, but subjectively that might not be true. Regardless, the properties that I have discussed regarding a single point of light varying in brightness remain true if we expand our visual field along any other dimension. Instead of contrasting only bright to dark, we contrast red to not red, on the left to on the right, near to far away, but it remains true that between two moments of identical state, the sum of all changes is 0 in all dimensions. From this thought experiment, we can know that human visual experience is zero-sum over a lifetime in all its aspects, or units.
We can expand this further, to other senses. The sum of all perceived changes in pitch is 0. The sum of all perceived changes in temperature is 0. That would also remain true for complex, composite dimensions, since it’s true of every separate dimension. The sum of all perceived changes in comfort is 0. The sum of all perceived changes in taste is 0.
Emotions are how we perceive changes in our internal state. Some emotions are very complex, multi-dimensional, hard to describe even. But it remains true that if we expand our reasoning to include every aspect of subjective experience, the sum of emotions over a lifetime is 0 on all axes. This zero-sum emotions theory was proposed by Blithering Genius, and it is as important as it is counter-intuitive. It is one of the most profound insights one can gain about the self, and it naturally has major philosophical implications.
The idea of quantifying change can be applied to subjective experience, but it’s more commonly used in other areas. As most people reading this probably know, in physics, any measure that varies over time (or over some other scale) can be modeled as a mathematical derivative – a function valued at every point as the amount of change over a very small period. The derivatives of position, speed, charge, inventory, and oil reserves are respectively speed, acceleration, intensity, throughput and extraction rate.
Zero-sumness is a property of all these derivative measures when changes in the underlying variable are added up between two same-valued points. This is an endlessly fruitful idea. For instance, over the lifetime of a stock or a house, the total amount of benefits on transactions sums up to 0, if we include the initial benefit and the final loss. And of course, over the lifetime of a person, the total amount of contrasts – all we ever know – sums up to naught.
As an aside, I do not have any formal scientific background, so I found the playlist “The essence of calculus” by 3blue1brown to help with getting an intuition for derivatives. I recommend it.
I will describe a phenomenon I call the paradox of meritocracy, derived from recent research on social mobility. If I had to recommend one book dealing with this topic, it should be The son also rises by Gregory Clark. By “meritocracy”, I mean a system which does not prevent people from ascending the social ladder, legally or otherwise. Meritocracy is a scale: a society can be more or less meritocratic. The mechanism described below applies in proportion to the degree of meritocracy in a given society.
If we assume a hereditarian perspective (a topic for another day), a meritocratic system is mostly a sorting mechanism. It selects the most competent, and affords them the opportunity to man key positions in society. In doing so, it progressively exhausts available talent in the bottom rungs of the population.
Thus, the paradox of meritocracy: after a brief “grace period” of rapid social mobility, a more meritocratic system ends up recreating a stratified society, one where there are few obstacles to upwards social mobility, but also little potential for it left. Meritocracy, in the long run, does not look meritocratic.
This has political repercussions. For instance, a population which (presumably) upholds meritocracy as an ideal can be expected to be unsatisfied with how it turns out to be like in practice. Perhaps more importantly, the emergence of a very stable class of talented elites from the sorting effect of meritocracy is liable to cause those elites to attempt to secure their positions through modifying the institutions they now control, thus reducing or nullifying meritocracy. I am not claiming that they would conspire to do this, generally speaking. It’s simply the expected product of having shared class interests, and a hold on the means of power.
Should that happen, and in the absence of laws against intermarriage between social classes, human capital then begins to pool again at the bottom of society. This happens by social capillarity: a high-class person might never marry a low-class person, but a high-class person might marry a slightly-less-high-class person, and so on. Thus, as the meritocratic sorting effect is reduced, genetic potential tends to slowly homogenize throughout the social ladder.
That pool of frustrated talent forming at the bottom is, metaphorically speaking, potential energy that can be harnessed by social movements to re-establish meritocratic norms, through reform or revolt.
I should say, of course, that genetic stratification/homogenization here only refers to those traits that are relevant to the ability to move up the social ladder. Those are not the same in every society, although there is certainly some overlap between cultures and eras. There is no guarantee that society would stratify or homogenize along other dimensions. It may, or may not.
Stabilizing feedback loops push systems towards equilibria. In this case, it could be a one-point-equilibrium (a social compromise) or a two-point-equilibrium (a cycle). A social compromise is more stable and conducive to human affairs than any kind of political boom-and-bust cycle, but a democratic system might get away with having moderately large oscillations. Reform after reform, rather than revolt after takeover.
An interesting aside is the expected effect of high or low fertility in elites. If the elites have high fertility, people on average will tend to move down the social ladder. It will be difficult for any given elite individual to maintain their social status, and in practice it will be harder for a low-class person to ascend the social ladder. If the elites have low fertility, on the other hand, people on average will move up the social ladder. It will be easy for any given elite individual to maintain their social status, and relatively easy for low-class individuals to move up. This is a confusing fact about modern society: because some of our elites are dwindling (reproductively), it is easy to move up the social ladder, but the ones already at the top look insanely privileged, almost as though it were impossible for them to fail. This isn’t true of all our economic elites, but it is true of our political and intellectual elites in the west.
I believe meritocracy is desirable because it places the most competent people where they can be the most useful. We should resist political attempts to undermine it, but it is important to understand that it will never be perfect, that it does not need to be, and that social mobility is not a good measure of it.
Brittonic memetics 