On a credible alternative to liberal democracy

The main idea: Fred Gohlke’s idea of iteratively setting up groups of 3 people who vote on each-other’s progression to ever-higher tiers presents the only credible alternative to liberal democracy.

0. Posts on this blog are ranked in decreasing order of likeability to myself. This entry was originally posted on 22.11.2021, and the current version may have been updated several times from its original form.

0.1 The Tryptic Model of elections featured in part 2 of this post is not mine at all, but was devised by Fred Gohlke as practical democracy, to which system I'm making relatively minor changes mechanics-wise and explaining in my own words below. Still, though the system is almost unchanged, the context to which I apply Gohlke's idea couldn't be more different, the original tweak to empower real democracy here made into a full alternative to liberal democracy. Thus, all resulting errors, departures from the original design or broader implications readers may choose to draw remain mine alone.

 

1 The First-Level Solution

1.1 I agree with those who see liberal democracy as a system with a built-in leftist bias, that takes a few generations to destroy once functional societies. Surely, there must be a better system.

1.2 On the other hand, I’m also cognizant of the critique of non-democratic systems one finds in his Mises or, implicitly and funnily enough, Moldbug (can a King head off the bureaucracy? A popular mandate is a far stronger tool).

1.3 The only credible, sustainable and proven alternative to liberal democracy in a modern society (i.e. one with a middle class) is the Single Party State. Leaving aside the sorry ideologies that birthed the terrors of the East, the fact remains that half the world was quite content to be run on the say-so of a single party. China and Singapore still are.

1.4 A nice, middle of the road system between a full liberal democracy and a formal or even informal one-party state would be the two-party system, enforced by such means as Approval Voting or Instant Runoff Voting on single member constituencies. Or one could think of adapting the Single Transferable Vote to electing parties and transferring until only two remain.

1.5 But these systems tend to evolve into an inner party and outer party system, with the entire media-university industrial complex siding with one party and keeping the other around as a stabilizing force. Until one can figure a way to keep the media-university complex firmly split into two warring sides, the two-party system is a nice mirage only.

1.6 So, single party state it is. To avoid complete degradation, entry into the party must be open to all who care to join, but promotion through the ranks is no guarantee. By simply asking people to be arsed to present a formal adhesion request, you eliminate 90% of the noise from a modern democratic system.

1.7 So, how to govern the internal promotion system?

 

2 The Second-Level Solution

2.1 Let’s say your party is founded today, and has only nine members. These members are allocated into formal, explicit levels and begin at Level (or Tier) 1. You arrange them into groups of three at random, so a Tryptic.

2.2 Now, this part varies by the culture of the host society, but I’ll stick to mine: you ask the tryptic to come together once a week for 30 minutes at least to have a talk over coffee, beer whatever. If at all possible, it would be best to have the Tryptic engage in some useful activity, and if any are needed these are organized by Tryptics, not individuals. But let’s keep it simple, and say they only have to meet once a week.

2.3 After three months you ask each to vote over who among the other two must be promoted. No reasoning to be given, just give a name according to your own conscience. Two votes promote you one level, anything less demotes you one level, unless you already are at level 1.

2.4 Assume the original nine members all elect a winner (no cyclic ties), you are now left with six level 1 members and three level 2 members. Repeat within each level. Assuming no ties, you get 6 level 1 members, 2 level 2 members and one level 3 member.

2.5 You can’t arrange levels 2 and 3 into tryptics now, so these guys wait for level 1 to bring new blood into level 2, and so on. It never ends. What you have done, is ordered your party base into a discrete number of explicit dynamic tiers.

2.51 Note how there is no chance of remaining in the same tier in two consecutive cycles, making it even harder for people to agree to a cycle to remain in a tier they all like.

2.52 A remarkable property of an election contested by only two candidates - which every election within this setup would be - is that none of the otherwise unavoidable conflicts among valuable electoral properties apply: all votes can and should be sincere in such elections from a game theoretical point of view, and no other setup can claim as much. That's a huge source of bias completely removed from the system. 

2.6 At any one point, whoever is at the topmost level (whatever that is), gets to call dibs on any position of their liking that the party may get the chance to place them into. Multiple applicants within the same level to be resolved by voting within the level or random tie breaking. Obviously the topmost level is/are the head of the party.

2.7 In practice, you would arrange this into geographic regions, so people don’t have to fly across the country weekly to meet. Best to avoid anything but face to face meetings, picking a sly psychopath is hard enough at close range, no need to bring Zoom into this.

2.8 Also, in practice the length of the cycle for each level would probably increase by seniority, from 2 - 3 months (a minimum, I think), to a year or two (a maximum). Watch the cycle length so that a reasonably young guy can get in at the bottom and make it to the top by a reasonable age though, the world has enough Saudi Arabias as it is.

2.9 At the highest levels, if you get a job because you picked it and were senior enough to be accommodated, you get to haul around your current tryptic fellows in your job (if they care to join). They get to see your actions, and judge you accordingly.

2.10 Also, whilst this is designed to be good enough to run a party state on, it need not start that way. Any party can be arranged by these principles today (you may even win a few elections, who knows). In a sense, this is how party organization is supposed to be run, only at a far less detailed and far less formal level.

2.11 Many more practical issues to resolve, but that’s the design.

2.121 To step away from the bombastic setting for a moment, a reasonable way to actually trial the algorithm would be in the comments section of a corpulent blog, where commenters are randomly assigned to tiers and vote on each other, eventually assigning explicit tiers to commenters and thus comments (the host's tier would always be 1+ the top tier).

2.122 Besides testing the system in a low-risk setting (whilst also testing Moldbug’s theory that professionals best monitor their own ranks), the blog itself would benefit from a hopefully half-decent shill defense, retain comment quality in the face of growing readership numbers, limit the need for active policing, allow readers to delve into comments in descending quality order as dictated by time available, and allow new commenters to rise through the ranks without getting lost in a sea of thousand+ incumbents.  OK, now back with the bombast.

 

3 Return to Aristocracy

3.1 The idea is just to adopt a selection system that maximally filters psychopaths out, something obviously beyond the reach of voters in a liberal democracy. Obviously you’d get factions of people helping one another, but would this be as bad as it is now? If nothing else, you got to know quite a few people (some of whom may become bigshots) on their way up. Who knows that a personality type that would never make it in the present system would not emerge here. Would Alexander get to run for a riding today? Run for Council?

3.2 Another desideratum is getting the media out of this (to the degree feasible), hence you only vote people you get to meet personally and are held to no account at all over you choice. Don’t like the race of the third guy? Cool, didn’t ask, don’t tell.

3.3 Obviously, pressure at the top would start to get quite high: when you’re high enough to be voting on who becomes President (or Mayor of a third rank town), you can expect all sorts of nasty games. Only those who can stomach the stakes should make it that far, and those who can’t, should just sell their vote to the highest bidder. It’s okay, we are looking for Alexander, not Brezhnev.

3.4 If the selection system is even marginally better than whatever informality we run now, and you iterate it enough times (Log2 of party base gives you the levels, so for a membership base of 48 millions as per the US Dems, you get 26 levels to win, and you will loose a few along the way) you get a noticeably better breed of leaders. Dare one say, rule by the able and willing?

3.5 But frankly, electing better guys is just an afterthought, what really interests me of this design is the capacity to explicitly and reasonably rank-order every man in the country. I’m not just after a better way to elect leaders, but more centrally after the capacity to adopt an explicit class system that is mobile enough yet rigid statistically speaking. A real alternative to liberal democracy.

3.51 In this same vein, a key piece of social technology this setup may unlock is the jury of true peers, whereas one is to be judged by fellows within one tier of himself, unless the pool thus selected would be too small (include more tiers above and below until you have the required pool).  

3.6 To think of the various signaling games people resort to to display adherence to an informal class, and to imagine that all could be done away with by a system that just tells you where you stack against all your fellow citizens, is quite enticing to even imagine. Hiring decisions, renting decisions, knowing where to buy a house, all would be made far easier once you know the level of the applicant.

3.7 In a sense, this system is trying to adopt the benefits of militarized society where your rank determines your standing in civilian life too, without the actual militarization part (which is just role playing in the absence of a real threat anyway). Also, trying to outdo and make flexible the Caste System which gives you quite a leg up in the world.

3.8 Is this making of bigotry and prejudice a system? Maybe, but I let those who are kept awake at night by such things worry about them. I take comfort in the reasonable expectation that, besides the market, such a system is to be preferred to the myriad of signaling dances one is expected to engage in nowadays.

3.81 In December 2022 Robin Hanson proposed a system that would achieve some of the same of the above, and would also rely on random groups of three. Similarities end there though, as that proposal relies on using the voting data to inform an Elo-like scoring system, instead of a mechanical progression through the ranks.