Rotating Sortition Bonanza

NS Nathan Schneider Public Seen by 17

Here's an initial proposal for a #SCRVI experiment: https://communityrule.info/create/?r=1600836482486


Yasuaki Kudo Wed 23 Sep 2020 7:00AM

It almost looks like an elaborate joke! Looking forward do to see what happens!


Nathan Schneider Thu 24 Sep 2020 3:26AM

This joke is serious


benjamin melançon Wed 23 Sep 2020 5:03PM

Mass communication moderated by sortition is what i've felt the internet has needed since way before i'd ever learned the word sortition. Very interested in it for lots of other uses too!

Do the community rules folks have tech in progress (i would include instructions for rolling die in that!)

Couple weeks ago i started writing a very bad python library for role selection at random as a side project to the scaling communication part https://gitlab.com/pwgd/talk-by-lot

Here's a demo hard-coded to Agaric members (for selecting roles like notetaker and facilitator) but could easily hook it up to any set of people; it has a history feature so roles are spread around even if randomness says someone should be notetaker 10 times in a row! https://pwgd-talk-by-lot.herokuapp.com/roles


Nathan Schneider Wed 23 Sep 2020 6:07PM

Wow, cool library! That's great!


Ana Ulin Thu 24 Sep 2020 1:10AM

Interesting! It would be good to see how that differs from the current system, since I don't know off the top of my head the specifics of our current Ops Team exclusion/membership mechanics.

I'd love to hear what current members of the Ops Team think about the proposal, which leads to me to the question of "how do I figure out who are currently Ops Team members?". I feel like I have a documentation discoverability problem. Pointers, anyone?

The one thing that jumps out to me from the proposal is this bit: "If a person selected does not want to serve, another candidate is selected at random". I recognize that we are not the type of organization that can expect any random member to serve (or can we?), but also this rule makes it so that the sortition feels biased towards whoever has availability/willingness (~privilege). But maybe most members would be able to participate, so this might not be an issue.

How would we conduct the experiment? Would there be a specific timeline to start / review / iterate / end? Would we have criteria to try to determine success of the experiment, or perhaps areas that we are trying to learn about in this experiment?


Nathan Schneider Thu 24 Sep 2020 3:28AM

Some really good questions there! Here's the Ops Team page on the wiki: https://wiki.social.coop/rules-and-bylaws/Admin-ops-teams.html

Good point on privilege; I was thinking of that too. Maybe there's a need to offer stipends for governance work as well, which I think is really reasonable.

I'd argue that we should be more eagerly and constantly experimenting, determining our criteria of evaluation based on how it all feels. Is that reckless? At least it's just a little instance:)


benjamin melançon Thu 24 Sep 2020 4:33AM

I love the point about privilege because that's exactly what i think sortition has the most capability to avoid. That is, if we don't select at random, then we are definitely getting the most available and entitled people. But of course, simply selecting at random doesn't make privilege go away. Ideally, any scheme using sortition either has a very low commitment (what i hope to do for deciding how far news should be spread) or people are given time and resources—so, are paid—to take on the decision-making role (and this is closer to classic sortition and how people advocate for it today). Or the best of both worlds— an easy way to weigh in and lead plus getting compensated for it!

But practically, for something like social.coop that's a side thing for most people, whatever compensation we could give wouldn't, say, suddenly give sufficient resources to a single parent desperately clinging to a job for its healthcare that they could move their attention from any of what they're dealing with; any time for social.coop would have to come on top of everything. Still, it's worth considering offering compensation, as it could make a difference for a lot of people on the margins, especially with so many doing some form of gig work.

A hybrid of two kinds of sortition, one selection of people to spend time with support on digging deep, and another selection weighing in on more-or-less polished proposals crafted by the first group, could perhaps help balance the fact that some people just won't have capacity for the deeper sort of involvement.

Another supplementary approach is to have controls to ensure that randomly selected decision-making pools are broadly representative according major demographic categories like gender, race, age, economic class— and re-sampling or restricting the sample when replacing someone. But that's a lot more doable when selecting 300 people to act as a congress for millions than at our scale i think.

Anyhow, people reluctant to impose their will on others (for whatever reasons) are the most qualified to make decisions for a group, and exactly who we hope sortition gives an equal chance to, and we should try to address any shortcomings in the process that causes us to lose out on having their input.


Ana Ulin Thu 24 Sep 2020 5:11AM

Well said. Your point about how even with compensation we can't really "make privilege/inequality go away" is a good one, and while I think that it is still worth thinking about ways to decrease the effects of it, I recognize that it's unlikely that we can find a perfect solution to the problem just within the scope of social.coop. It feels important to try to hold these two points as we deliberate.

I'm a fan of enabling ourselves to experiment more recklessly. I think one way to make "wanton experimentation" safer for the community is by having good "escape hatches" -- ways to stop or modify experiments early if we feel that they are causing harm or simply going really badly, and ways to regularly to check in to see if an experiment should be stopped or modified early.


Matt Noyes Thu 24 Sep 2020 4:22AM

Interesting! About exclusion on Ops Teams -- nomination/recall should be available to the membership so an Ops team doesn't set itself up as an independent body.


benjamin melançon Thu 24 Sep 2020 4:45AM

I would say (especially in light of some of the ways to mitigate against rule by those with too much time and ego) that we seriously evaluate what skills on the ops team make it impossible to fill by sortition, and move any functions that can be to the randomly selected ruling group, to reduce problems of power accruing to the 'permanent' paid staff and to make it more possible for the sortition-selected, decision-making, resource-controlling body to be paid.

An another note entirely, is it possible to write the rule without using the "oligarchy" terminology? 😂


Ana Ulin Thu 24 Sep 2020 5:18AM

Seconding the request to find a more friendly alternative to "oligarchy". 🙀 Could just be a "members committee". Some other ideas for the folks that enjoy spitballing such things: Members' Brain Trust, Members' Ministry (this one is real sexy), Will of the Members, Heart of the Members.


Nathan Schneider Thu 24 Sep 2020 5:53PM

Thanks, all! @mlncn @anaulin yeah my use of "oligarchy" is pretty tongue in cheek:) The CR bit is really just a starting point for discussion, not any sort of final proposed language.

And @mlncn I think it's interesting to ask whether the Ops Team could be handled by sortition. I suspect that the necessary learning curve would make it hard, and that some bureaucratization is needed. I was thinking earlier whether it would be possible to simplify our processes in such a way that more people could participate and less expertise were needed. For instance, using a tool like Cloudron to deploy Mastodon and other tools would put it within the skill set of someone like me (I run a Cloudron instance for my lab), whereas I don't think I could handle much of the technical work as it is done now.

But I like the idea of helping more members become better acquainted with the processes and technical work of the co-op.

Would anyone else like to make a counter-proposal (on CommunityRule or elsewhere) for #SCRVI to consider?


Ana Ulin Sat 3 Oct 2020 7:44PM

Sounds like there is no counterproposal forthcoming. Do we want to try out rotating sortition? If so, how do we move the experiment forward?


Nathan Schneider Sun 4 Oct 2020 5:04AM

@Ana Ulin · @anaulin@social.coop Mm, I'm not sure we have a critical mass to try proposing something as such. But I'd love for us to keep proposing ideas for how to move the community forward. What is moving you?


Hakanto Thu 28 Jan 2021 7:21AM

An alternative take on this could use sortition to create a periodic auditing council, as described in this example about the Scott-Bader Commonwealth from Workplace Democratization by Paul Bernstein:

"Full-time managers are further supervised by the total Commonwealth through a unique organizational mechanism invoked annually at the general meeting. Twelve members are selected by lot, as a kind of jury, to approve or state reservations on the board of directors' conduct of company business over the preceding year.

If this panel of representatives should decide that the board's performance has not been satisfactory, the board is given three months maximum to take appropriate remedial action. Then it must report back to the Commonwealth, which resumes its general meeting for this purpose. Another twelve persons are then selected by lot as representatives of the whole Commonwealth and are asked to render their judgment. If this panel approves, then business may proceed. If not, then the task falls on the trustees to decide which directors may need to be replaced.

This unique method, rather than direct election of all directors (only two of the seven are directly elected), has been found by the Commonwealth to provide the best check on management's performance. It leaves management free of the pressures of a personality contest, which direct election of leaders can too often become, (e.g. in some plywood co-ops), yet heeps the management accountable to the public opinion."


Nathan Schneider Thu 28 Jan 2021 5:05PM

Thanks for sharing this! I can imagine a simplified version of this working for us.