The Potential of Weighted Sortition

This is the 5th post in a series on Barbara Goodwin’s classic work on sortition Justice by Lottery, University of Chicago Press, Chicago, 1992. Previously published parts: 1, 2, 3, 4.

Sortition is hiring without human intervention using random rather than reasoned criteria. Weighted sortition, though, is a statistical narrowing down of randomly chosen job candidates to the more ready among workers available. In Barbara Goodwin’s words, they are “lotteries where some people get extra chances of winning according to their personal characteristics or qualifications.” More broadly, the chances of “winning” a position can be micro-managed to take in any number of other factors beyond qualifications, such as need and desert, both personal and social, both present and future.

Goodwin’s “Justice by Lot” introduced to me the idea of weighted sortition. True, she dismisses it as “impure” compared to completely random sampling of the entire population, but the idea has intrigued me ever since I encountered it here. Admittedly, playing with percentages increases the complexity of choices, but complexity, in an age of supercomputers, data mining and artificial general intelligence (AGI), is not the obstacle it was even ten years ago. A weighted lottery allows for intriguing possibilities never imaginable before. It can factor into every hire whatever evidence-based social science deems well grounded, including experience, aptitude, personal choice and even broader social goals. The difficulty is that the advantages of sortition itself have to be clear to a broad proportion of the population before weighted sortition comes into play, as Goodwin points out,

Only by participation, which is a continuous process, would people acknowledge the authority of a social lottery [p. 71].


Once a broad swathe of the population is accustomed to surrendering their immediate fortunes to the whim of the kleroterion, then the tweaking, that is, the weighting of the lot, can commence. I cannot emphasize enough how essential it is that every locality feature a physical kleroterion designed and built by local talent. This is the only way to establish and demonstrate to all beholders the total transparency of random selection. Whether the particular design of a kleroterion involves bouncing balls, a spinning wheel of fortune or whatever, all participants and spectators must witness the choice being arrived at by some kind of natural process. Only then will all appreciate that it makes a small but real difference if social scientists decide that seven percent rather than, say, five percent of every choice be totally random, allowing anyone into the choice.

A weighted lottery need not hobble itself with the assumption of ignorance inherent to elections — one person, one vote, no matter who it is. Everyone’s entire resume, a statement of their credo, values, goals, dreams, performance evaluations and much more besides, as well as how their style combines with other members of the workforce, can all be fed into an AGI algorithm that prepares the sampling and divisions of a weighted lottery.

This data mining and information processing would cover the often-heard objection to sortition, “If I break my leg, I don’t want just any randomly chosen person fresh off the street to be setting my bone fracture.” Under weighted sortition, there would be a small chance that one member of the team assigned to setting your leg may not have every qualification on the job description. The chances of such misfortune coming about can be as small or large as experience and the judgment of experts and the AGI algorithm deem safe and necessary.

It can be argued that such risks are entirely natural, that they are already built into the nature of things. For example, physicists tell us that everything we do is subject to the random factors inherent to quantum mechanics. They tell us that there is a small chance when you put a pot of water on the stove and turn on the heat that it will freeze rather than boil, but the chance is so small as to be negligible.

Beyond that, it could be that every job description, every stipulated qualification is not as solidly grounded in evidence as we tend to assume. After all, centuries ago, surgery was handled by barbers with little formal schooling. The weighting of the sortition algorithm could well discover that some procedures are simple enough, aided by robots, that a surgeon need not be overqualified, to spend a fortune to undergo decades of schooling to perform that operation as well or better than they are now.

Beyond all that, a weighted lottery could even enable advances in democracy itself. It would allow for personality-free “virtue elections,” where, instead of voting for an individual manager (in a large population, few voters have enough direct experience with any single candidate to cast an intelligent vote based on anything but reputation, image, physical appearance or some other flimsy criterion), voters could choose whatever virtue they deem to be needed now or in the near future. Let us say that a nation is fraught by dissent, and the people choose the virtue of “compassion” over courage or another quality listed on a multiple-choice ballot. The AGI would then go through everyone’s resume data and those who have shown this virtue would have, say, a twenty percent greater chance to be bumped up to a management level in their next job assignment. This would shift the entire style of how a polis is run, top to bottom, without ever naming names or being bogged down in personal conflicts.

Justice by lottery can, of course, be seen as a sub-category of decision-making, since most social allocations result from some kind of decision. Just allocations rest on particular kinds of reasons or criteria which are systematically invoked in a decision. While Elster’s argument could be thought peripheral to questions of social justice — indeed he might maintain that situations of indeterminacy and incommensurability are, precisely, characterized by the absence of pointers to just distribution — he makes a useful, justice-related suggestion in his advocacy of weighted lotteries (that is, lotteries where some people get extra chances of winning according to their personal characteristics or qualifications).

However, the compiling of a point system reflecting, say the need or merit of the competitors is based on assumptions about people and about which of their attributes are relevant to the distribution in question, so the criteria invoked in devising such a list will themselves be part of some theory of justice. The proposal for a weighted lottery is thus itself weighted in some direction and must provoke debate about the criteria chosen, and their justification. While a lottery society would probably eschew weighted lotteries in its pursuit of extensive equality, existing societies might well accept the principle of weighted lotteries, while rejecting anything as radical as direct egalitarian lotteries. Hence, Elster’s proposal should be borne in mind as a possible modification of the lottery principle which could adapt it to prevalent conditions and attitudes.

Greely has also advanced the merits of the lottery as an allocative or decision-making tool for governments. Although the lottery is a satisficing method (a draft lottery does not produce the best recruits, but it produces enough recruits), it greatly reduces the costs of allocation where elaborate tests for merit or suitability would otherwise be required. Greely also notes the incidental benefits of random selection, which reduces the discretion of officials and possible abuses thereof, avoids undesirable self-perceptions on the part of the losers which might occur with other allocative methods, and recognizes human dignity by according equality of opportunity and fair treatment.

One of the virtues of Greely’s exposition from a liberal perspective is that it illustrates that, for governments committed on principle to non-intervention, random allocation process could help to limit bureaucracy and minimize the government’s role. Greely concludes (and Elster would presumably agree) that “random selection is not an abdication to irrationality and amorality; it is neither irrational to recognize the limits of reason, discretion, and judgement, nor is it amoral for a democracy to reaffirm its commitment to human equality [pp. 180-181].

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: