Sortition Vs. Elections under socialism.

[Unclebananahead]

Which do you favor, and why? Sortition (or allotment), selects officials at random via lottery, ala jury member selection in the west. Some say that this would virtually eliminate the possibility of the creation of a specialized 'political class' which could degenerate into a ruling bureaucracy.

[Sinister Cultural Marxist]

Elections, but with the option to recall, or for a populace to void the decision of a representative under certain circumstances.

Perhaps overseen by a jury-like body of randomly selected peers, but not composed of that.

[Victus Mortuum]

If the central body is making coercive decisions - then sortition representatives.

Edit: Now that I think about it, even with the former elected delegates (reps subject to recall) from very small districts would be a better option if you must have a central coercive power, because it would allow for a smoother transition to the type of organization below, and it would ideally abolish 'officialdom/bureaucracy' (if tagged with median worker wages and non-private-funded campaigning).

If the central body is making non-coercive decisions - then elected delegates or some other means of establishing large-scale deliberation and creation of voluntary working groups.

The former is the only possible way that a 'workers state' could exist - the latter is 'stateless'. If there is a state in existence, then the sortition model is the one I support, but I prefer to eliminate a separate coercive central body as soon as possible (but of course only in context of the socialization of property) - so the latter is preferable in as many circumstances as possible and hypothetically all circumstances.

Edit:
Of course, any of the above changes must go along with the dissolution of the standing army and the professional police and their integration with the citizens. Those two types (and only with those two types, it should be noted) of political reforms would meet Marx's and Lenin's conceptions of a 'workers state' as far as I can tell. And the abolition of the army and police is certainly something anarchists would support, as well as probably at least not opposing the anti-bureaucracy idea.

Sorry, this is totally off-subject, but I'm rereading State and Revolution and this is related.

[graymouser]

I think attempting to balance out bureaucracy and power-cliques by sortition is a naive and dangerous mistake; naive because it misplaces the roots of bureaucracy and cliques, and dangerous because of its impact on leadership dynamics.

Bureaucratism does not arise from elected leaderships but from division of labor; the bureaucracy comes about because the leaders are not doing the day-to-day work but rather allowing another layer beneath them to arise. For instance, the Bolshevik Party did not "degenerate into" a bureaucracy, but rather absorbed a bureaucratic layer which through intrigue and rallying around the person of Stalin managed to outmaneuver most of the previously-existing leadership. This couldn't have been prevented by sortition, because the bureaucrats were the only ones with the technical skills to run Russian society. The only way to fix that is to make the leadership itself actually capable of running things - which you absolutely cannot do with sortition.

The other thing sortition does that is dangerous is to create gaps between de facto and de jure leaderships. In any group of people, there will be leaders based on skill, personality, personal history and so on. These are natural leaders, and conflicts with and between them are best settled by linking de jure leadership to electoral contests between them. What sortition does is to create group A, which is the de facto leadership, and group B, which is the de jure leadership. If there isn't a significant overlap between groups A and B, then group B will not have the authority within the group to make any actions against group A, and even if members within group B tried to counter them, they would be out by sortition in the next round anyway. With elections at least group A and group B overlap to a sufficient extent where the de jure leadership has the respect and confidence of the whole body.

The real solution, suggested by Gramsci in his prison writings, is the mutual technical training of the leadership bodies. With a sortition method this would meet with intractable difficulties, and so I have to say the whole idea should be thrown in the trash.

[Unclebananahead]

I think attempting to balance out bureaucracy and power-cliques by sortition is a naive and dangerous mistake; naive because it misplaces the roots of bureaucracy and cliques, and dangerous because of its impact on leadership dynamics.

Bureaucratism does not arise from elected leaderships but from division of labor; the bureaucracy comes about because the leaders are not doing the day-to-day work but rather allowing another layer beneath them to arise. For instance, the Bolshevik Party did not "degenerate into" a bureaucracy, but rather absorbed a bureaucratic layer which through intrigue and rallying around the person of Stalin managed to outmaneuver most of the previously-existing leadership. This couldn't have been prevented by sortition, because the bureaucrats were the only ones with the technical skills to run Russian society. The only way to fix that is to make the leadership itself actually capable of running things - which you absolutely cannot do with sortition.

The other thing sortition does that is dangerous is to create gaps between de facto and de jure leaderships. In any group of people, there will be leaders based on skill, personality, personal history and so on. These are natural leaders, and conflicts with and between them are best settled by linking de jure leadership to electoral contests between them. What sortition does is to create group A, which is the de facto leadership, and group B, which is the de jure leadership. If there isn't a significant overlap between groups A and B, then group B will not have the authority within the group to make any actions against group A, and even if members within group B tried to counter them, they would be out by sortition in the next round anyway. With elections at least group A and group B overlap to a sufficient extent where the de jure leadership has the respect and confidence of the whole body.

The real solution, suggested by Gramsci in his prison writings, is the mutual technical training of the leadership bodies. With a sortition method this would meet with intractable difficulties, and so I have to say the whole idea should be thrown in the trash.

Natural leaders? So some people know better than everyone else what the people want, after the institution of socialism? Also, 'mutual technical training of leadership bodies'? You'll have to explain this further to me.

Also, won't circumstances be rather different in this day and age as compared to Russia 1917? I'm not certain how things will unfold for revolution in our time, but in 1917 things were a great deal different.

[Zeus the Moose]

The Israeli-British mathematician talked about a method that combined both elections and sortition in an interesting way. The example given is that when elections are held for a given post, everyone in the constituency casts a vote for which candidate they want. Then, a single ballot is drawn from the box and that candidate is elected to the office. His argument is that this helps mitigate some of the problems with sortition in that it can put people into office who genuinely do not want to do the job, so if the election is to something long-term (like a parliament with biannual elections), people won't mentally check out after a certain period of time, which comes back to the problem of leaving things to "busybodies."

The videos where he talked about this (Two Parts) are here (part 1) (http://vimeo.com/14852939) and here (part 2) (http://vimeo.com/14854566).

I'm not sure how much I agree with this proposal, but it certainly is interesting in my view. It also suggestions that elections versus sortition need not necessarily be an either-or choice.

[Comrade_Stalin]

Which do you favor, and why? Sortition (or allotment), selects officials at random via lottery, ala jury member selection in the west. Some say that this would virtually eliminate the possibility of the creation of a specialized 'political class' which could degenerate into a ruling bureaucracy.

As a Stalinists I am for Sortition (allotment) or a Demarchy. The reason is the economic assumption behind politcal choices.

[Die Neue Zeit]

The Israeli-British mathematician talked about a method that combined both elections and sortition in an interesting way. The example given is that when elections are held for a given post, everyone in the constituency casts a vote for which candidate they want. Then, a single ballot is drawn from the box and that candidate is elected to the office. His argument is that this helps mitigate some of the problems with sortition in that it can put people into office who genuinely do not want to do the job, so if the election is to something long-term (like a parliament with biannual elections), people won't mentally check out after a certain period of time, which comes back to the problem of leaving things to "busybodies."

The videos where he talked about this (Two Parts) are here (part 1) (http://vimeo.com/14852939) and here (part 2) (http://vimeo.com/14854566).

I'm not sure how much I agree with this proposal, but it certainly is interesting in my view. It also suggestions that elections versus sortition need not necessarily be an either-or choice.

Comrade, Machover suggested this solution to me as a means of (s)electing the Condorcet winner (the one who beats everyone else in every pairwise election). This mixed scenario of his would not apply if there's a candidate who's got 50%+1 of the votes.

The random selection pool would have to take into consideration volunteers. Only if there aren't enough volunteers, perhaps, could some election carrots be tossed to the arena. Geez, I didn't realize there was a wiki on this:

http://en.wikipedia.org/wiki/Demarchy#Lottocracy


The concept of demarchy is similar to but slightly different from L. León's concept of lottocracy. Burnheim ... insists that the random selection be made only from volunteers. In the chapter A Concept for Government, León states: "... that first of all, the job must not be liked". A detailed protocol for lottocracy is described in the same chapter.

[Kotze]

Random ballot is interesting. No deterministic voting method can completely erase the spoiler effect, related to that is the fact that no deterministic voting method can completely erase pressure to vote strategically.

Random ballot also deals with gerrymandering. Consider a model where voters who have a fixed party preference live in different districs each having the same population size and each district sends one candidate to the national parliament. If random ballot is used, no redistricting changes the expected average of seats a party gets (provided the number of voters in each district stays equal).

Random Ballot would be terrible for voting on an issue, but if it's used to elect a body with many seats that's very different (and even for voting on issues it could be usable if used only as a fallback mechanism to pressure people into compromising).

I still prefer sortition for bodies with many seats though.

[MarxSchmarx]

Only half seriously, random ballot is as a philosophical perspective not too distinct from a hereditary monarchy. After all, the personal qualities of the monarch we would get would basically be random, unless you are a hard-core eugeniticist who believes the ruling class through selective breeding can be made a super-elite.

The reason is is that a monarchical heir, under the right circumstances, can come to represent just about "any person" with the right training and the like. In fact, if we made them go through the life of a random individual in the country, and didn't inform them or the people around them of their obligation to rule, I think we'd have something very akin to the hereditary rulers of yore work just as welll as a\randomly selecting a group of people to rule.

Now there are reasons that absolute monarchs tie their lot in with the ruling class, basically because they are raised to rule and uphold the ruling classes' interests.

But let's take this perspective seriously and assume a "man in the iron mask" scenario. Isn't this basically what the proponents of random selection favor?

[Unclebananahead]

A randomly selected body or committee derived from the ranks of the working class, would presumably act in such a way as to advance what they perceive to be in working class interests. That's the premise I'm going by. Not being a former capitalist could certainly be one of the main criteria for eligibility.

[Jose Gracchus]

I can't believe people think one could just start, tomorrow, statistically jury-calling 500 workers together to decide something "for the working class" hosted by communist militants. I'm sorry if this seems rude. It seems to me absurd.

[Unclebananahead]

I can't believe people think one could just start, tomorrow, statistically jury-calling 500 workers together to decide something "for the working class" hosted by communist militants. I'm sorry if this seems rude. It seems to me absurd.

How so? It might not be universally applicable at all levels of socialist, post-capitalist government, but certainly a number of multiple seat committees could be filled this way

[Die Neue Zeit]

I think he's referring to forming left parties at the moment, and he's got a point. You can't have a representative sample until said sample has at least 25-30 members and the statistical population is large enough.

My commentary deals with mature party-movements, not fledgling groups striving to be such.

[Jose Gracchus]

What we -need- is a program from fledging groups, since that's all we have...

[Die Neue Zeit]

Random ballot is interesting. No deterministic voting method can completely erase the spoiler effect, related to that is the fact that no deterministic voting method can completely erase pressure to vote strategically.

Random ballot also deals with gerrymandering. Consider a model where voters who have a fixed party preference live in different districs each having the same population size and each district sends one candidate to the national parliament. If random ballot is used, no redistricting changes the expected average of seats a party gets (provided the number of voters in each district stays equal).

Random Ballot would be terrible for voting on an issue, but if it's used to elect a body with many seats that's very different (and even for voting on issues it could be usable if used only as a fallback mechanism to pressure people into compromising).

I still prefer sortition for bodies with many seats though.

Why would random balloting be bad for plebiscites? Except when an option has over 50% of the votes the first time around, isn't random balloting merely applying the Condorcet method? :confused:

[Unclebananahead]

I was referring of course to the way in which to run things in a socialist society, not necessarily a revolutionary 'organización.'

[Kotze]

Except when an option has over 50% of the votes the first time around, isn't random balloting merely applying the Condorcet method? :confused:How did you get that impression, did you fall asleep during a presentation by Machover? :lol:

If you have a pile of ranked ballots and every candidate receives less than half of the rankings in first place and one candidate is the second place on every ballot, that candidate has 0% probability of winning under Random Ballot yet is also the Condorcet Winner.

There are some Condorcet Methods that do have probalistic aspects though, meaning that as long as there is a Condorcet Winner, that option wins with 100% probability if the voters are honest and that in other cases there are lottery elements. Maybe you meant something like this?

For example, some assemblies are organized in a way that when there are several competing proposals, they don't use ranked ballots, instead they only arbitrarily take 2 of them, hold a vote which of the 2 they prefer, they throw away the loser, then they do another comparison with the winner and another arbitrarily taken proposal, and so on, until no competing proposals exist any more. While this doesn't make every pairwise comparison between these proposals, it still selects the Condorcet Winner, because the one who wins every pairwise comparison also wins any subset of pairwise comparisons. If the Condorcet Winner isn't already in the first pairwise comparison of the series you won't know in the end whether the winner of the process is a Condorcet Winner, only that if among them was a Condorcet Winner, the winner of the process must be it.

What I don't like about that method is that doing such a comparison series with challengers entering in random order is somewhat spammable, because in situations without a Condorcet Winner an option that enters the comparison battle late has a better chance to make it, so if you have a proposal and you come up with more and more proposals very similar to that, the probability of one these entering late, and hence also its probability of winning, goes up.

The spam problem can be addressed by finding some non-spammable quick and dirty way to roughly figure out how popular options are (like polling a subset of voters with a non-spammable method) and then you let the roughly less popular ones enter the comparison series first. Or you let go of the idea of making pairwise comparisons with only a subset and instead use ranked ballots to get the complete pairwise matrix and use a counting procedure that is quite robust against spamming like Schulze.

[Die Neue Zeit]

How did you get that impression, did you fall asleep during a presentation by Machover? :lol:

Not his presentation "Socialism and Democracy" (as linked to by Zeus above), but correspondence. :confused:


If you have a pile of ranked ballots and every candidate receives less than half of the rankings in first place and one candidate is the second place on every ballot, that candidate has 0% probability of winning under Random Ballot yet is also the Condorcet Winner.

Maybe I need to re-read his correspondence a third and fourth time. He did have a scenario of preference balloting, and just now before the second reading I had the impression he had simple balloting. :confused:


There are some Condorcet Methods that do have probalistic aspects though, meaning that as long as there is a Condorcet Winner, that option wins with 100% probability if the voters are honest and that in other cases there are lottery elements. Maybe you meant something like this?

:confused:

[hatzel]

So does this whole question stem from the decision that direct democracy has already been totally written off as a possibility? Of course we all know that increasing technology makes it easier and easier for such a system to be implemented, and has the advantage that it can work on the macro-level. Much of this thought seems to be based around the thinking that there will be some single institution, a group of people, chosen by some method, which will rule over a territory corresponding to an existing national state, or, if it turns out that only a part of the territory is won for socialism, for a single group to rule over that whole area. Participatory democracy could even work in that case, but it would be much fairer and more representative if people were actually given the chance to participate in something that is for their own interest. I mean, even if we randomly selected people from the workers, chances are the needs of the workers in one area might be different than in other areas. So that would rely on the 90% nationwide who don't need, for example, irrigation, to agree with the 10% who happen to live in a pretty dry part of the country and are struggling to grow crops. Or whatever it might turn out to be. Local participatory democracies, relying on internet technology to facilitate participation, seems much better to me than somehow choosing a group of people to make decisions for a whole country...

[Die Neue Zeit]

The other thing sortition does that is dangerous is to create gaps between de facto and de jure leaderships. In any group of people, there will be leaders based on skill, personality, personal history and so on. These are natural leaders, and conflicts with and between them are best settled by linking de jure leadership to electoral contests between them. What sortition does is to create group A, which is the de facto leadership, and group B, which is the de jure leadership. If there isn't a significant overlap between groups A and B, then group B will not have the authority within the group to make any actions against group A, and even if members within group B tried to counter them, they would be out by sortition in the next round anyway. With elections at least group A and group B overlap to a sufficient extent where the de jure leadership has the respect and confidence of the whole body.

You really need to look up:

Stratified sampling (OK with a job slot system)
Probability-Proportional-to-Size sampling (scrapping the slate system)
Cluster sampling
Quota sampling