Varieties of Elitism

Logo by Anthony Salvagno [CC BY 1.0 (https://creativecommons.org/licenses/by/1.0)], via Wikimedia CommonsImagine each individual neuron had its own moral agency…

What should such a neuron do?

Assemble several billion neurons into a human brain, and you can get surprising intelligence, but individual neurons are not very smart by themselves. They can’t guard against making mental errors merely by comparing their opinions to those of other neurons–they don’t have enough independent intelligence to conceive the opinions that ought to be formed. For example, individual neurons cannot conceive of a brain so they cannot form an opinion about whether they should participate in a brain. Most importantly, since individual neurons aren’t smart enough to design a brain, they can’t individually know how to participate in one.

Corporantia entertain the possibility that all intelligent beings share this dilemma: No matter how smart you are, you have the potential to be part of something smarter, yet you can’t be smart enough to design that smarter entity, so you are inadequate to answer your own most important personal moral dilemma: “How to be part of something better than yourself.”

Or might you actually be just smart enough?

Despite the apparent paradox, many individual neurons do form into brains. Inspired by that example, I teamed up with mathematician Bennette Harris to explore the possibility that we can take some intelligent steps towards forming into something better than ourselves. Specifically, we investigated the potential for individuals to apply mathematics to choose between possible ways to form teams that are more effective than individuals (e.g. what is the optimal team size and what is the optimal way to decide which individuals go on which teams).

We titled our analysis Varieties of Elitism because we realized that any such endeavor would inevitably be labelled “elitist,” yet each way of organizing society (including the status quo) is some variety of elitism. Our goal was to add mathematical rigor to the way we define varieties of elitism, so we can consider them objectively, rather than get mired in name-calling.

We have yet to produce a mathematical proof that any particular form of elitism is optimal or superior to others, but we did produce open source software that anyone can use to simulate various forms of elitism and compare their average performance. It doesn’t tell you whom to marry, but it does open a research field. The results thus far support some general conclusions:

  • Optimal team size depends upon the number of interdependent roles and the degree of inequality among individuals.
  • Majority-rules forms of democracy do not perform as well on average as forms that permit the most qualified minorities to control decisions (e.g. sociocracy, futarchy, evidence-based decision-making).
  • On average, measures of vulnerability (e.g. SAT, ACT, or GPA) aren’t much better for organizing society than chance. Such measures are blind to the talents of savants who have both great ability and great vulnerability.
  • Gains from diversification using an instrument like the GRIN-SQ are robust against imperfections in the instrument and would raise average performance by about 400%. That means the GRIN-SQ can’t be much better or worse than its major competitors for this purpose.
  • Of all the varieties we considered, the best-performing variety of elitism favors savants but relies on accurate measures of specific abilities. The first company to perfect and use such measures would outperform competitors by 700%+. A society that blocked such development would sacrifice a potential doubling of progress even for its average least-privileged members.

A paper was published in Figshare with the supporting software. If you have corrections or improvements to suggest, please do so where everyone can see them in PubPeer at https://pubpeer.com/publications/BAE4D582BF0CF2E93AA0F43667FD84

This paper may be cited as

Santos-Lang, C., Harris, B. (2018). Varieties of Elitism. Figshare. DOI: 10.6084/m9.figshare.7052264

You can discuss this paper in a free online webinar on Tuesday, Oct 9, 2018: http://ewh.ieee.org/r1/new_hampshire/NLSri/Sep2018/raswebinar092418.html

Update: If you missed the webinar, the recording is at https://www.youtube.com/watch?v=wYBAKxVkh0w