Against Turchin on power: proposing a better model

Turchin's model fails to be sufficiently empirical or quantitative

Back in April, Peter Turchin wrote a blog post titled “On Social Power.” In this post, he, predictably, explains his model of social power. It is very short, so I won’t summarize it, as you can just read it yourself. In this article, I will go over my several critiques, as well as my model, which I believe to be better, since it is more empirical and quantitative and therefore more useful to science.

The genealogical critique

My first issue with Turchin’s model is its source. It comes from Michael Mann, a UCLA pseudo-scientist. Michael is a psuedo-scientist because he is a historian who claims to be a scientist of the generally pseudo-scientific field “sociology.” Sociology, and Michael Mann’s work, are pseudo-scientific because they are insufficiently quantitative and empirical. If you do not understand this claim, you need to read the beginning of the reading list here. Turchin, for what it’s worth, already agrees with me on this point, given the first chapter of his book Historical Dynamics.

You can see that sociology is insufficiently empirical and quantitative by studying the methods of the father of sociology, Max Weber. Weber did not collect data and derive quantitative predictions from the data, as a science does. Instead, Weber practiced a qualitative method, amounting to something one might call “social botany.” He was a mediocre conceptualist, who would break “society” down into disparate, immeasurable, non-predictive verbal components. The proof of this is in his Wikipedia page. Weber never allowed one to predict which societies would be democratic and to what degree, if a society was democratic or not, how much power someone has and who would grow up to have what amount of power from birth. These are some of the basic questions a real science of society must answer. Instead, for example, Weber divided “political leadership” into three “types”: charismatic, traditional, and legal. The coherency and utility of these types is very little, for charisma is fake and “legal” is circular.

This sort of “thinking” is the main quality of Weber’s whole body of work. Likewise, he is known for dividing the causes of social stratification into three components: class, status, and party. He never measured social stratification or provided any quantitative evidence for this assertion. He was truly pseudoscientific. One wonders if he had any mathematical aptitude at all. If he did, he surely would have been a physicist. Thus we cannot fully blame this behavior on the lack of data; Galton, Spearman, and Fisher did just fine collecting social data and inventing the framework needed to analyze it. It is likely that Weber was quite low IQ, likely under 120, and was therefore incapable of producing real science.

Now enter Michael Mann and his book The Sources of Social Power. What a promising title! I ctrl-f “Weber” and get 77 results. I ctrl-f “Marx” and get 101 results. I ctrl-f “genetic” and get 4 results. Two instances say:

Original stratification had little to do with the genetic endowment of individuals. Nor had any subsequent social stratification. The range of difference in the genetic attributes of individuals is not great, and it is not cumulatively inherited.

Another instance says:

The genetic constitution of human beings gives a fundamental equality of most mental attributes relevant to the acquisition of general knowledge of the world.

LOL!

So Mann was a pseudo-scientist more concerned with Marx and Weber than with real human biology. This is why his book, which spans multiple volumes amounting to over 800 pages of unreadable pomobabble, contains no data or math whatsoever. Like Weber, Mann presents us with a typology of power. This is not scientific, and is not worth taking seriously, because it does not enable quantification and prediction.

The logical critique

We now get to the substance of the theory. I hold that one of the concepts is illogical: administrative power. Turchin says

Administrative power: I can order you to do something by virtue of being your boss, a superior in an organization.

This is circular reasoning. I suspect this happens because the idea of the organization is being reified in Turchin’s mind. An organization is an abstraction, it does not exist or at least cannot be taken as given in a science of power. The alleged existence of organizations is something that must be explained in terms of biology and individual relations. You absolutely cannot say “organizations exist, Joseph is the boss of the organization, Peter is the employee, therefore Joseph has power over Peter because he is the boss of the organization”, as if the organization were the MPD from Call of Duty: Black Ops (this allowed whoever happened to be in it to control the zombies).

Instead, you must ask why Joseph is the boss and Peter is the employee. Saying “administrative power” amounts to saying “because I said so” or “because the boss is the boss.” Joseph can order Peter to do stuff, i.e., Joseph is Peter’s boss, because he is Peter’s superior in the organization, i.e. he is Peter’s boss.

To be non-circular, you have to explain Joseph’s superiority in a more basic way. If we stay inside of the framework, Joseph must have economic, coercive, or persuasive superiority over Peter. This makes more sense, logically speaking. The boss must be the boss for an exogenous reason.

The empirical critique

We have just abolished one of the concepts of the theory. The next concept we will abolish cannot be rejected as easily. It requires a good empirical understanding, not just a little bit of logic. Turchin says:

Persuasion, or ideological power: this is the most subtle, but extremely effective (when it works) form of power. It is especially effective because it is often not appreciated enough, or even not considered as a form of power. After all you end up doing what I want because you want it, right? Persuasion is exercised by aspiring politicians, orators, religious figures, thought leaders, and opinion influencers working for the traditional media or, more recently, on social media.

I hold that, to the contrary, persuasive power is overrated, because it is fake. Those who hold that it does not exist and is not a form of power are correct. Therefore, one person thinking persuasion is power is too many, making it overrated and not underrated.

It is not that trivial to prove that persuasion is not power. The current reading list on this can be found here under “Memetics.” Essentially, we have more or less found that deception is not effective en masse. As evolutionary psychology predicts, people are not very gullible. If they were, they would be at a very serious fitness disadvantage. Evolution is not perfect and diseases emerge all the time, so while we predict most people will not have this disease, it is fair to ask how many have gullibility. The statistics I am currently collecting on this suggest that, across situations, it is less than 1% of people, confirming the theory.

If deception is not effective, then information must be true and useful in order to effect someone’s behavior. The other day, I filled up a beggar’s gas tank. This influenced the beggar to continue driving around. Was this power? Obviously not. New useful information can be modeled as a resource like gasoline. It is scarce and worth paying for. But if you are great at supplying new useful information, wouldn’t that make you an economic elite, someone who is powerful because the excel at creating value from which they can differentially profit?

Thus, one empirical step undercuts the alleged uniqueness of of persuasive power. The following logical step recognizes that we are left with an empty husk which is essentially no different than economic power — distinguishing persuasive power makes about as much sense at this point as distinguishing food power from water power from … $COMMODITY_N power.

The quantitative critique

We are left with “coercion” and “economic power.” We now must ask, how do we measure these concepts? Why are they both called power? Shall power be a two dimensional vector, [coercion, economic]?

I can answer these questions. Turchin can’t. My model collapses power into one quantity known as credit. There is another quantity from which credit flows known as alpha. The distinction between coercion and economic “power” is halfway retained in my model as a distinction between two different types of alpha, that which is based on increasing overall utility, granting one the ability to withhold oneself, and that which is based on domination and violence, the capacity to diminish utility beneath a counter-factual “default value”, the value utility would assume if you had not been born. I call these, clumsily, negative-death alpha (NEA) and positive-death alpha (PEA) respectively, because a solid response to the latter is to just kill the one trying to wield it (they are tasked with making sure this is not trivial).

Alpha is stolen from finance where it denotes the excess return of an investment relative to the return of a benchmark index. Alpha is taken to be an intrinsic property of a person, in theory essentially a product of their genetics, although different genotypes can have different alpha in different environments.

Power is modeled as someone’s credit stream, the amount of resources they are currently extracting from others measured in credit. Credit is an abstraction like distance. Concrete measures of credit include the USD, the Chinese yuan, and the Roman denarius, while concrete measures of distance include the meter, the yard, and the cubit. Alpha is potential credit stream, it predicts how large of a credit stream, say, a baby might be able to acquire when they grow up.

So, on the one hand, power becomes a single number, but on the other we could, if we care to, break down a credit stream into a NEA and PEA component which sum to produce total power.

This model is clearly more quantitative than Turchin’s, making it more useful to science. It is also easy to deploy. Someone’s power is the total value of the resources they extract per time period. This can be approximated by wealth or standard income, although it is not exactly the same. For example, U.S. Supreme Court Justices often hang out with billionaires and take fancy vacations on their yachts. This is not reported as income, even though it is, and it’s what makes them observably more powerful than some engineer who also makes 6 figures.

The working reading list on power is here under the “exousiology” section.

This is the best model I am aware of

I will leave off by saying that in all my extensive reading on modeling power, I have never found a more coherent and quantitative model than this one. By the process of elimination, I therefore propose that this is the best existing model of power, and should consequently be adopted and further refined. I believe this model can help bring insight to Turchin’s work on elites and the structural-demographic cycle in which wealth and economic elites figure heavily (because they are the real elites).