so, you want term limits and stuff like saying one person can't hold all the offices at once. But when a startup starts, you want one person to be able to hold all the offices, because there may only be one person at the start, and even when there's more than one, you don't want the founder to hestitate to bring more people in because that automatically forces them to be officiers.

One solution might be just to require monotonicity on the number of distinct people on the Board; ie once the number of distinct people holding Board seats is N, it can never go back to less than N (without amending the Bylaws).


actually there's a simple way to have a time delay in Chairs while preserving proportionality-style attack resistance:

you could the ballots using a proportional method. The 3rd-place winner serves now, along with last election's 1st and 2nd-place winners.

So, even with an attack where a party instructs its voters not to vote for it, so as to secure the 3rd place now along with the 1st place before, they wouldn't have the 2nd place, because the 1st and 2nd place winners were taken from the same proportional election.

( how big does the 2nd faction have to be to prevent the 1st faction from taking both spots in a proportional election for 2 seats? For RRV the reweighting function is 1/(1+SUM/MAX), where MAX is the max allowed score. If we set the max allowed score to 1 (approval voting), then ppl who voted for the 1st-place candidate get their weight halved.

(1 - a) > a/2 2 - 2a > a 2 > 3a 2/3 > a

So the first faction would have to have >2/3s of the vote; in this case they'd still have 1/3 of the vote after halving, and the second-place faction would have at most 1/3.

This is satisfactory because 2/3s is usually the vote needed to ignore or amend the bylaws anyways.



good idea from to prevent last-minute 'sniping' on votes: "add time to the clock when the vote outcome changes"


in addition to (or instead of?) score voting for selecting which proposals are on the agenda, could also use prediction markets; people try to predict which proposals will pass, and those that the market says have a good chance of passing are bubbled up


have a section on Temporary Exceptions (they are mentioned sporadically in theLongVersion)

also, may need something like 'state of emergency' that gives executive power to act when there isnt even time for Temporary Exceptions to be passed before each action.


protocol for choosing which chair any chair, or PM, or the Court, can end


reduce committee size to 5 again.


what we could do is have an EVEN number of ppl on the Board, and then only when the vote is tied, do the other levels vote.

eg could have 6 ppl on the board, where the 7th vote is held by the lower levels of the delegate pyramid; when the vote is decided by the 6 board members, it is concluded immediately, but when it is inconclusive without the 7th vote, then all of the other levels vote.

should we have all of the other levels vote at once, or only have the third level vote when the second level vote is also inconclusive? and should the weighting of the other levels be similar to the first (so the third level, as a whole, only has 1/7th the weight of the second), or more even (so the third level, as a whole, has the same weight as the second; but since there are so many more individuals in the third level, each third level individual has much less weight)?

i guess right now i think that the other levels should each have an even weight with each other, even though all of them are only 1/7th of the Board vote. i go back and forth on the question of whether all levels should vote each time the Board needs more voting vs only having the second level vote and only calling up the third level when the second level vote is inconclusive. Note that if you do the latter, the third level could still be called up when a resolution of one particular second-level constituency is as stake, as opposed to the entire board.

note that this means that the key parameter p becomes 6.


there should be a punishment if the Court later decides that the PM wrongfully declared an emergency, or unjustly and wrongfully abused emergency powers


during an emergency, the Court can still review emergency decrees, using a subjective standard rather than just 'do they violate the constitution'


could just make futarchy a legislative house. so eg Forum, Board, Futarchy.

(instead of, or addition to, the volunteers?)

should one of the Chairs be indirect sortition?

how about instead we stick with 1 indirect sortition guy on the Board, and that guy also gets all of the individual powers of a Chair (but no influence on the powers that require votes of 2 or 3 chairs)


mb dont let the Board veto checks on individual Board members, eg term limits, anti-corruption stuff, unless the Chairs first vote to give the Board a vote

(similar to my rock-paper-scissors idea for the Board, the Chairs, and the Judges to decide each others' salary; of course, we also have the Assembly to help with that)

so could say:

Each term, the Assembly decides the level of payment and benefits of the Board and the Judges and the Chairs. Each member of all of these are paid the same.

Assemblymembers are unpaid.

The advantage here is you never have two entities who both agree not to prosecute each other, or who agree to overpay each other, as a quid-pro-quo:

Note however that Chairs and Judges who are popular with the Assembly cannot be successfully impeached. Also note that it takes >2/3s vote of the Judges or the Board respectively to impeach the Chairs or Judges, and then a 2/3s majority to convict, in stark contrast to the Executive Branch and the Board and Assembly members who can be prosecuted by a single Chair and convicted by a majority of the Judges. In other words, it's very hard to remove a Chair or Judge mid-term.

Note that the Judges do rule on cases against the Board, which may prompt the Board to retaliate by impeaching a Judge. As noted in the previous paragraph, this impeachment will have difficulty succeeding, however.

Impeachment may be for medical incompentence or corruption. (can it also be for ordinary incompetence? maybe, but with unanimity?)


mb go back to term limits on the Board and Judges and Chairs, with the proviso that if you can elect yourself solely with your own votes, then you can bypass term limits.

to generalize/soften this hard cutoff, could say that, in addition to that exception, another exception is that term limits can be bypassed with a 2/3s vote of everyone at every level who indirectly elected you (eg you recursively descend into the Board members who voted for you past your term limit, and ask the people who elected that delegate to vote on the question of term limits; whenever you fail to get 2/3s in favor, you take that as a 'no' vote and stop recursing; if you do get 2/3s, you then have to recurse into each of these delegates to see if their yes vote is confirmed at the lower levels; note that in the Board at least, this has the additional benefit of providing more power for the middle delegates). Similarly if the Board elects a Judge past their term limit, the same procedure recurses into the Board.

with Chairs, it simply raises the vote threshold for the Chair to be re-elected past their term limit.

similarly, you can simultaneously be a Boardmember, Judge, Chair, CEO provided that either you are being elected in all but one of these solely through your own votes, or that you get >2/3s vote (even indirectly) for all but one of these positions. That is: any position that you are elected through solely by your own votes doesn't count against the limit; nor does any position in which you get >2/3s votes; you can hold only one other position simultaneously. Similarly if one person wants to hold more than one Board seat, or Judge seat, etc, simultaneously.


yeah, to allow secret ballots, better make Assembly delegations non-revocable for a certain amount of time, and assigned only on election days. And maybe make the Topic list standard, given by the organization.

still doesnt prevent the attack where a mobster directs you to delegate to one specific person, ensuring that no one else is directed to delegate to that person. But that requires a scale of henchmen linear in the number of coerced voters, which does not scale well for the mobster.


some design motivations stemming from the requirement that this be usable by small private groups:


majority of Chairs can constrain the interpretation of conditions under which surveillance is allowed; eg if 'foreign' surveillance is allowed then what does 'foreign' mean? does 'metadata' count?


Chairs can bring action against the government and its officials for denying them access. It is a (personal) felony for an official to deny access of any sort to a Chair.


to the extent that administration of procedure involves subjective decisions, that's the job of the Chairs (by majority). However, should the Chairs do something that objectively goes against the rules, the Court can overrule them.


case study for emergency succession protocols:

my take-aways:


authorizations to use force expire after one year

temporary exceptions expire after one year


the full text and content of all rules, including any content incorporated by reference, must be publicly available for free with no restrictions on copying, publication, or derivative works, and must be primarily public domain/CC0 or similar.


thinking about a -1,0,1 score vote... assuming two major parties that hate each other with their voters voting strategically, they'll almost cancel each other out, leaving un-hated third parties to sweep, right? seems somewhat unpredictable. Would this encourage the parties to negotiate prior to the election so as to present candidates so agreeable to the others that they don't negative campaign? Could that actually, counterintuitively, lead to less negativity than a 0,1,2 score vote?

what about a -1,0,1,2 score vote?

i'm also thinking of just going back to simple 0,1 approval voting. The point of the '2' was to give people a way to give '2's to their party but give '1's to moderates, because i worry that only having 0,1 will encourage people to just vote 1s for their party and 0s for moderates. But 0,1,2 leads to 'energize your base' dynamics which appear to be divisive.

otoh approval voting may lead to that anyways because in some cases the parties are trying to get people to strategically 'bullet vote' rather than approve of any non-party candidate

eh, after thinking about this today, i'm still for 0,1,2. I do think it's good not to encourage negative campaigning, and i think being able to give the '2' to their favorites will psychologically give people permission to give some '1's to moderates without feeling disloyal to their party.


no heriditary or lifelong offices

(why no lifelong? life extension technology is coming someday..)


arguments against approval voting:

(they also mention , which i have not previously heard of, which is interesting but i don't think many voting theorists support it so i'll leave it alone)

their main argument against seems to be that ppl will just bullet vote strategically, and in that case the few ppl who do not bullet vote are at a disadvantage. i don't find this convincing, because what about the situation where there are some voters who legitimately don't want to bullet vote? Also if the system devolves into bullet voting, then i don't see what is really lost. Also, some of their examples (eg US vice Presidency, possibly the Dartmouth board example) are things where multiseat proportional methods would be the real solution.

also in some of their 'hypothetical' examples of real-life elections and what would have happened with approval voting, they say that in some cases the strategically optimal thing would be complicated; in which case i bet most voters would just vote honestly, so this would be good, not bad, as they portray it.


instead of some complex decay in vote prediction market, just say that half of your initial voting weight cannot be wagered, and the other half can.


for judicial selection committee:

high judges (3): selected as a slate of 3 out of pool of judges by proportional score voting by the Judicial Selection Commission from the PREVIOUS election cycle. The Judicial Selection Commission has 7 seats, and is elected each election cycle, and is composed of: 3 seats elected by the Board; 2 seats elected by the High Judges from the PREVIOUS election cycle (so this gives a 2-cycle delay; High Judges of cycle 0 elect Commissioners serving in cycle 1 who select Judges serving in cycle 2); 2 seats elected by all Judges (all of these elections are via proportional score voting).

actually, maybe have the JAC select ALL judges, and then have the Board confirm just the High Court?

or mb have the JAC nominate ALL judges in slates of 3, have them confirmed by the Board, and then have the JAC alone select the High Court?

and have judicial appointments last for 3 election cycles:

High Court (3): selected as a slate of 3 out of pool of judges by proportional score voting by the Judicial Selection Commission from the PREVIOUS election cycle. The Judicial Selection Commission has 7 seats, and is elected each election cycle, and is composed of: 3 seats elected by the Board; 2 seats elected by the High Judges from the PREVIOUS election cycle (so this gives a 2-cycle delay; High Judges of cycle 0 elect Commissioners serving in cycle 1 who select Judges serving in cycle 2); 2 seats elected by all Judges (all of these elections are via proportional score voting). highest court of appeals, interpret the Bylaws; decides cases by simple majority. Found Rights, Refer to Board by 4/5. Impeach chair by 5/5?. Set salary and benefits of Chairs, up to the amount requested by legislation (the highest amount that received 4/5 votes), provided it is >= salary of High Judges and Boardmembers and CEO. if <3 judges have been appointed then the Board also serves as the High Judges

judges: nominated in groups of 3 by proportional score voting by the Judicial Selection Committee, then each slate may be confirmed by a simple majority of the Board. Judicial appointments are for 3 election cycles.


if there is an indirect sortition Board member, then are there only 4 delegation blocs?

If there are requirements on Board expertise and independence, then how about instead, the second level of the hierarchy elects a 4-person Board via proportional score voting 0,1,2. But if we're doing that, then why not just do it all the time?

i'm worried about that 'breaking' the whole 'no mass elections' thing here. How about, instead, if at any time the Board would not fulfill its requirements, then all sub-boards must choose delegates that fulfill all of the requirements? Or, maybe, continuously, any sub-board whose delegate is the only one causing some requirement to be met gets to send a second delegate?


howabout having an 'appeal' hierarchy (for Courts) or a vetting hierarchy (for Committees), and make seniority in terms of # of non-deadlocked completions on Court or Committee cross-councils the way that citizens become eligible for the higher-levels Courts and Committees (maybe the score is minus unexcused (by the presiding Judge) absences/other citations)?

but maybe the score should also provide some incentive to disagree with the majority sometimes? So, maybe just total cases served upon, whether or not a decision is reached?


The Judicial Selection Commission has 7 seats, and is elected each election cycle, and is composed of: 3 seats elected by the Board; 2 seats elected by the High Judges from the PREVIOUS election cycle (so this gives a 2-cycle delay; High Judges of cycle 0 elect Commissioners serving in cycle 1 who select Judges serving in cycle 2); 2 seats elected by all Judges (all of these elections are via proportional score voting).

... actually i think this has too much feedback. How about 2 each of seats elected by the Board, all Judges, the Chairs, and 1 indirect sortition seat

i'm wondering if it would be better to give 3 seats somewhere (so that a third party could get in on the action?), even if this meant either one seat somewhere else, or not taking any seats from one of these? The most obvious thing to cut would be the two seats from the Chairs, and then give 3 seats to each of the Board and the Judges. But otoh it's nice to give the Chairs some seats b/c they are directly elected; eg imagine that the Board becomes compromised.

ok so how about 3 seats elected by the Board, 2 by the Judges, 1 by the Chairs, and 1 indirect sortition seat. If there are three factions, this way the third faction gets at least one seat, at the expense of taking away a seat that the minority of Chairs could control. One disadvantage here is that only the Chair seat could be expected to fight against government overreach; but perhaps even the Chairs wouldn't do this, so it might not be much worse than giving 2 seats to the Chairs. Note also that this JSC is slightly majoritarian; the 1st seat of the Board + the 1st seat of the Judges + the Chair seat + the indirect sortition seat can all be expected to be of the majority party most of the time; that's 4/7; but, not counting the indirect sortition seat (which over time should mirror the entire factional distribution), that's only 3/7, so it's not too bad.


on the topic of simplification, it's probably good to every now and then write down what i think are the key ideas, the ones that stick in my memory, to help in seeing what to cut out:

chairs, board, forum, CEO, judges, cross-councils for committees and courts, vote prediction market, membership, legislation.

all multiseat elections are via proportional score voting unless otherwise noted

chairs: 3, directly elected via regular elections with proportional ("reweighted") score voting 0,1,2; staggered, in that the 3rd-place winner serves immediately, along with last election cycle's 1st and 2nd-place winners individual powers: cancel a state of emergency, cancel a temporary suspension of the rules; address members; address board; standing to prosecute any part of the organization, or its officers, or anyone for corruption; power to demand information from any part of the organization; power to unilaterally 'declassify and publish' organizational secrets majority powers: accept new members (after Recommended by Courts), expel existing members (after Recommended by Courts), run legislative procedure and promulgate legislation, give awards, set salary and benefits for High Judges, up to the amount requested by legislation; set salary and benefits for Boardmembers, up to the amount requested by legislation (greatest amount needed to secure 2/3 Chair votes), pardon or commute (except high officials and their family and associates) unanimous powers: impeach judge, state of emergency

board (5): 4 members elected via hierarchical delegation, 1 member elected via indirect sortition. The indirect sortition member is rechosen until they are not the same as any of the other board members, or (b) the person whose nomination is chosen nominated themself.

However, except when voting on individual people (eg CEO), votes are as if it has 7 seats; the last 2/7 of votes are provided by the rest of the delegate pyramid, with an equal weight given to each level; the rest of the pyramid is only called upon to vote when a majority (>=3/5) of the 5 Board members vote for something but this is not enough to accomplish the desired action. Board meetings are chaired by the Chairs, but the Chairs do not have votes at these meetings (the Chairs may, however, speak and give their opinion, in the same manner as the Boardmembers). Board meetings are conducted according to Standard Code of Parliamentary Procedure.

forum: liquid democracy (but secret ballots, votes delegated for a fixed time (1 election cycle) and non-revocable before that). (3? 5?) slots on agenda via proportional score voting 0,1,2. Forum Proposals must be approved (and potentially improved) by a hierarchy of topical Committees before becoming eligible to become voted into a floor slot.

CEO: elected by the Board continuously via 0,1,2 score voting with runoff between challenger and incumbent (simple majority in the runoff needed to unseat)

High Court (3): selected as a slate of 3 out of pool of judges by proportional score voting by the Judicial Selection Commission (JSC) from the PREVIOUS election cycle. The Judicial Selection Commission has 7 seats, and is elected each election cycle, and is composed of: 3 seats elected by the Board; 2 seats elected by all Judges; 1 seats elected by all the Chairs; and 1 seat by indirect sortition highest court of appeals, interpret the Bylaws; decides cases by simple majority. Found Rights, Refer to Board by 4/5. Impeach chair by 5/5?. Set salary and benefits of Chairs, up to the amount requested by legislation (the highest amount that received 4/5 votes), provided it is >= salary of High Judges and Boardmembers and CEO. if <3 judges have been appointed then the Board also serves as the High Judges

judges: nominated in groups of 3 by proportional score voting by the Judicial Selection Committee, then each slate may be confirmed by a simple majority of the Board. Judicial appointments are for 3 election cycles.

cross-councils, Courts, Committees: 4-person volunteer cross-councils made up of 1 supporter of each of the 4 board members. Plus 1 Judge constitutes a Court or Committee. Court/Committee procedure is run by Judge, but all 5 members vote. Courts and Committees needs 4/5 members voting to make the non-default decision (convict if a 1st-level Court adjudicating an accusation, admit if screening new members, expel if expelling members, overturn if a Court on Appeal, approve Proposal if Forum Committee), but 3/5 to make the default decision (innocent or (guilty but the law is wrong), reject new member, keep existing member, don't overturn, reject proposal). Both Committees and Courts are classified by Topic (the topical ontology is set by the legislature).

legislation: 66% to amend or major treaty or temporary suspension of specified rules, 50% to reduce organizational power, increase member freedoms, reduce spending, veto other chamber, 60% o/w. Only 1 house need pass, then need only survive 50% veto by other house. Amendements must be re-passed twice in consecutive election cycles (but, each time, they only have to be passed by one chamber, and not vetoed by the other; they do not need to be passed by both chambers). Temporary suspensions of specified rules can operate in the meantime, but these cannot affect the procedure.

vote prediction market: voters can wager half of their (initial) voting weight (and any weight they gain later) on predictions

membership: new members are first screened by Courts according to the guidelines set in the Rules by the legislature. If they are Recommended by the screening Court, then a majority of Censors can make them members. The procedure for expelling existing members is similar.

'proportional score voting' is reweighted score voting (reweighting by 1/(1+K)) with 0,1,2 score choices on ballots.


All Directors on a board have equal votes and powers.

The proceeding of a board are usually public. A board may call a closed session by a 2/3s vote.

A board is governed by the 4th edition of Sturgis's Standard Code of Parliamentary Procedure; except for the following additions and changes:

However, since the board doesn't consist of very many people, is it suggested that meetings be initially conducted informally by general consent, and to switch to formal parliamentary procedure only when requested by one or more Directors, or when dealing with particularly weighty and controversial issues.

todo, figure out how to do the budget. Mb the CEO proposes, subject to the constraint that the organization cannot borrow money except by a 2/3s vote, and must always pay its debts?

for reference, here's what we used to have, i feel like it may be too decentralized (that is, i am no longer advocating the following):

The budget is decided in the following way.

The Chair, or, in Large groups, the Office of Procedure, impartially estimates the organization's income for the coming budget cycle, with Mandated Spending (see below) subtracted. This estimate cannot include any additional borrowing.

The representative bodies of the legislature (i.e. the Chair in Small groups, the Board in Medium groups, the Combined Board in Large groups) create budget line items via the usual legislative process. The legislature may also create minimum and maximum absolute funding levels for each line item. The meaning of a minimum level is NOT that the item will necessarily be funded at at least that level; it is only that, if the budget does not have at least that level for that item, it won't contain that item at all. In other words, minimums are for setting a minimum amount of funding for a project below which a project is not worth doing at all.

The line items always include:

To each member is assigned a "portion" of the organization's estimated income, calculated by their voting weight times the organization's total estimated income.

Then, each budgetary period, each member submits a list of budget caps for line items, and a prioritized list of how they would like to spend portions of "their" portion of the money. They can only place money towards the line items given by the legislature.

Next, the Green-armytage collaborative budgeting procedure is followed, with the addition of respecting any maximum caps set by the legislature, to determine a draft budget.

Items which are, in the draft budget, funded below the minimum levels set by the legislature are eliminated from consideration for this budget cycle. Then the Green-armytage collaborative budgeting procedure is rerun to get the final proportional budget.

During the budget cycle, if actual income falls short of estimated income, the remainder may, at the discretion of the CEO, be borrowed.

In subsequent budget cycles, the minimums and maximums set by the legislature are adjusted for inflation (inflation as estimated by the Office of Procedure), but otherwise the line items, minimums, and maximums persist until changed by further legislation.

By a measure with a voting threshold of 2/3s, the legislature can borrow money.

By a measure with a voting threshold of 2/3s, the legislature can spend money, either from the just-in-case fund or borrowed, in the middle of the budget cycle.

Mandated Spending

By a measure with a voting threshold of 2/3s, the legislature can mandate the spending of a minimal amount per budget cycle on some line item. The threshold to mandate a line item solely to pay off debt is a simple majority threshold.

These mandates persists for 4 electoral cycles, except for a line item to pay off debt, which persists indefinitely. These mandates can be repealed like an ordinary measure.

Paying interest on debt is always Mandated.

some verbiage from the old bylaws to copy:

" This page contains the core text of bylaws or a charter/constitution for Branch Democracy.

Please note that this core text must be augmented in order to create complete bylaws for an organization. First of all, some types of organizations, such as corporations, may be required by law to address various issues in their bylaws or charter which are not addressed here. Second, this core text purposefully leaves the following key questions unanswered:

The topics in the previous bullet list can be addressed by combining this core text with one of: the for-profit corporation addendum, the non-profit addendum, or the government addendum. However, in your jurisdiction, in order to form valid bylaws or a valid charter, you may still be required by law to address other issues which are not addressed here.

Organizational structure

These bylaws define and structure the Organization.

The legislature sets policy by creating rules called "Acts". The Legislature is composed of three "houses", which are the Forum, the Elect Board, and the Delegate Board. The Forum uses direct democracy; the Boards are small bodies composed of representatives.

The CEO implements policy, with the exception of external affairs, which are handled by the EEO. The CEO and the EEO are appointed by the boardmembers.

The Court(s) interprets rules.

The Chairs investigate, with the aim of checking and restraining powerful elements within the Organization to ensure that the rules are followed.



The function of the Chairs is to protect individual rights, to eliminate corruption, to ensure that proper procedure is followed, to eliminate unnecessary secrecy, and to temper "procedural bugs" by the application of individual judgement.

The existence of Chairs does not excuse other parts of the organization from the responsibility to have internal controls.

Powers held by each Chair individually

Information request from the organization and from high officials

Each Chair has the power to request any information from any part of the organization, or any information relevant to the organization from high officials. Such a request for information must be met immediately (or, if fulfilling the request takes time, as soon as physically possible) and cannot be denied for any reason. In particular, no information may be kept secret from a Chair, except for personal passwords, and that only to the extent that the information is also kept secret from everyone else, including the CEO and EEO.

Each Chair may attend (or, if attendance is not possible, then prior to the meeting, request a video feed from) and quietly observe (but not necessarily participate in) any meeting in which the organization's business is being conducted. This is a form of a request for information.

When a Chair compels information, the information must be supplied along with a signed oath that it is the truth.

Dispute another's information request

Each Chair has the power to take another Chair to court on the charge of making a request for information which imposes an excessive burden. A request is defined to impose an excessive burden if the burden of collecting and providing the information is (a) large, forcing the organization to significantly curtail important activities in order to divert its labor to fulfilling the request, and also (b) unreasonable, and also (c) in excess of the amount of information needed to reasonably ensure the non-corruption and procedure-following of government, or to make a case in court. In no case shall adverse effects thought to be caused by the release of information be taken as evidence for the excessiveness of a request. In no case shall an information request be considered to be "large" based on the nature of the information, rather than because of the large amount of labor required to compile it. In no case shall an information request be said to be "excessive" if it could be satisfied with less than 100 manhours of labor, although much larger requests are typically not excessive either.

Any official who does not comply with a Chair's request for information shall be permanently stripped of their position, possibly in addition to other sanctions as specified by statue.

Note that, because of (c), by saying something is excessive, the court is stating that the additional information gained will be irrelevant to the Chair's success in court, if they were going to use this information to go to court. So, if a Chair does go to court, the court must act as if the information gained was as would be favorable to the Chair's case, within reason.

If the court finds that the request for information was excessive, the only penalty shall be the nullification of that request and of future similar requests by the same Chair, if ordered by the court. In particular, the right to publish any information obtained is not affected. In no case shall a Chair be punished for the exercise of, or the results of exercising, their powers.

Information request from other entities

In addition, each Chair may request some information from or about entities, such as members, which have close relations with the organization or its officials, in proportion to the closeness and, in the case of entities which are close to officials, to the power of the officials. This type of request may be challenged in court before being complied with.

In addition, each Chair may request any information from or about any private entity if there is probable cause to believe that the information is related to an action or conspiracy by the organization or by officials of the organization relating to an attempt to deprive individuals of their rights, or to not follow procedure, or which may be corrupt. A court must find probable cause before this type of request may be issued. This type of request may be challenged in court before being complied with.


Each Chair has the power to publish any information, even information which was previously declared secret or private, and information which according to Bylaws and Statues may not be published. No Chair shall be hindered from or punished for the publication of any information. Chairs are never obligated to pre-notify any entity before publication. Publication implies that any rule by which the group kept the information secret from its membership no longer applies to that information, and therefore that information may be republished without penalty.

The information must be published in any manner which makes it available to the general membership of the group, not just to a few select individuals. Each Chair must make available a list of information which they have published, and where to find each publication. Published information must be kept available for at least one year, at the expense of the Organization. Before they have decided which information to publish, Chairs may reveal any information to their staff. Before publication, they may reveal any information to select individuals and entities if useful to arrange for publication.

Rather than publish information to the membership, Chairs may choose to publish it to the Combined Board, and must be provided with a venue for doing so. Chairs may not be compelled to publish information to the Combined Board instead of the membership.


Each Chair has the power and the standing to bring any case against any entity alleging corruption, failure to follow procedure, dereliction of duty of an official, or infringement of individual rights, provided these crimes are with reference to the organization.

An entity may be corrupt in their participation in some other organization, or involved in causing another organization to infringe or not to follow proper procedure, or guilty of a serious felony not covered in the previous paragraph, without being vulnerable to tribunal prosecution.


The necessary and sufficient condition for any elgible voter to be included on the ballot for Chair is self-nomination.

in order to go to war, not only the general decision-making process must pass a resolution, but that resolution must be confirmed by a vote of those who would actually fight


mb let the high court (or, indeed, any court or committee) create subordinate courts/committees and organize the topic ontology of them (but who then determines which judges staff them? the JSC, i guess)


fifth seat in board:


focusing on simplicity: maybe we should only have 1 chair instead of 3, so that we don't have to do reweighted proportional voting? imagine a small assembly constituting itself for the first time -- wouldn't the need to do reweighted proportional voting at the beginning just to select the chair put them off?


also, mb simplify the formulae for the delegate pyramid more radically; maybe fix it to 2 levels of delegates (not including the CEO, the Board, or the voters), or less if that's not needed. Maybe just make the number of delegates in the level of consituency immediately below the Board chosen by a simple formula (eg ceil(ln(effective #voters)*5)) and make the number of delegates in the constituency below that set by whatever is left (eg ceil(effective #voters / (5*ceil(ln(effective #voters)*5)))).

eg if effective # of voters were

hmm i'd like to make the level just before the Board even smaller, mb use 4 instead of 5 since only 4 board seats are chosen by the Pyramid anyhow, but let 1/5 of the constituencies be unrepresented, assuming they choose to stay at minimum size:

in Python: x = 250e6 ceil(log(x)*4) ceil(x/ (5*ceil(log(x)*4)))

terrible result for 150. I think this simplified formula is doomed. At least need one more 'min' in there.

Mb just set it at min(#effectiveVoters/5,20) members for each constituency just below the 4 Board members. Or replace the '20' with log(n)?


what about just using approval voting then eliminating everyone who gets their choice? is that proportional?

i guess not; because what if 90% of the electorate is in the same party and had the same preferences? with a multiseat 3-seat election, they'd all be eliminated after the first pick, leading the remaining 10% of voters to pick both the second and the third pick


pros and cons of PR (proportional representation):



For single-winner elections, the best system is Simple Optionally-Delegated Approval or SODA voting. In this system, you can vote for as many candidates as you want (as in approval voting). If you vote for just one candidate, you may choose to delegate your vote to that candidate. If you do so, then after votes are counted, that candidate will get a chance to add additional approvals to your vote, from a list of their pre-declared preferences among the other candidates. Because they only choose how many of their pre-declared preferences to approve, you can know how your delegated vote might be used; so if you disagree with their preferences, you can choose not to delegate by checking a "do not delegate" box on your ballot. And because candidates use delegated ballots after the election, with the candidates with the most votes going first, they will only add additional approvals if they themselves can't win, and your vote is necessary to help elect a compromise candidate.

This system is somewhat complex, so honorable mention should go to Approval Voting, which is like plurality except you can choose whether or not to vote for each candidate separately, instead of being limited to voting for only one of them. Basically, it's plurality, but you count all the votes, instead of throwing away so-called "overvotes". I'd also like to mention Majority Judgment, where you grade each candidate on a scale of A to F, and the one with the highest median grade wins; this is the system which, on some days, I feel is a bit better than SODA.

The other two systems I'd strongly favor are Condorcet system(s) and Range Voting. I have mixed feelings about Instant Runoff Voting (IRV, also called Alternative Vote or (misleadingly) Ranked Choice Voting, and wrongly named as STV in Joe Samson's answer); while it is an overall improvement over plurality, it has several downsides compared to the other systems I've mentioned. It can give crazy results if it happens to prematurely eliminate a centrist; voters, once burned by such crazy results, could retreat to dishonestly voting for the "lesser evil", destroying its advantages over plurality; and it requires centralized counting, reducing the transparency of the vote-counting process.

For multi-winner elections, I'd favor PAL representation. In this system, you get a ballot with the candidates from your local district listed, but you may choose to write in a candidate from another district if you wish. As with SODA voting, your vote is delegated, and candidates who do not win may transfer it to others who can. When a full slate of candidates have each reached the same quota, they are elected. But each candidate is assigned multiple districts, so that each district has one representative from each winning party. Thus, instead of having about a 55% chance of getting a representative you'd voted for, you'd have about an 83% chance (higher for bigger states, lower for smaller ones, because of "rounding error"). The "PAL" stands for "Proportional, Accountable, Local", the three characteristics which this system manages to blend into one.

Other decent options for multi-winner are STV with multi-member districts (2-5 reps per district); AV+ (mentioned in David Hood's answer); Asset Voting aka Liquid Democracy (involves differently weighted votes for elected representatives, similar to the answer by Bryce Johannes); and in fact Sortition (representatives are chosen randomly, as with jury duty; as the recent deliberative poll in California shows, this option actually works much better than you might expect). " -- Jameson Quinn

" For single-winner elections, I would implement Score Voting (aka Range Voting), or even its simplified 0-1 scale variant known as Approval Voting. These systems are both simpler for voters and election administrators than essentially any ranked method. They also achieve better average voter satisfaction than most other systems, according to an objective economic measure known as Bayesian Regret (see ScoreVoting?.net/BayRegsFig?.html). ... For multi-winner elections, I would implement a system of Proportional Representation. My preferred systems would be largely based on this analysis, by Warren Smith: ScoreVoting?.net/WarrenSmithPages?/homepage/multisurv.pdf

From best to less good: Asset Voting - ScoreVoting?.net/Asset.html Proportional Score Voting - Candidates are rated and we elect the n candidates with the greatest sum of scores where we may only count one score per ballot (the highest score for any of those n candidates). Single Transferable Vote - Needs no introduction. Riddled with problems, but still pretty good at picking proportional multi-winner outcomes. Reweighted Range Voting - ScoreVoting?.net/RRV.html " -- Clay Shentrup


could have a 7-person Board with:


the thing about having only 1 Chair instead of 3 is that, since they will probably be of the majority party, which is probably the same party of the CEO, it's unlikely that they'll use their powers to investigate to good effect (except upon old administrations, which might setup an unhealthy win-and-then-keep-power-at-any-cost dynamics amongst criminals in power)

i am still disturbed by the complexity of RRV. Maybe if just reweighted approval then it would be simpler? Then at least in the third round (to fill the third seat), instead of asking the vote-counter-volunteers to do addition for every ballot, you can just ask them to sort the ballots into three piles; those who voted for neither of the first two winner, those who voted for one of the first two winners, those who voted for two of the first two winners. Otoh mb adding 1+2 isn't so hard; then you'd still only have five piles: those who voted for neither of the first two winners, those whose votes for the first two winners summed to {1,2,3,4}. Otoh mb adding 1+2 IS hard if you're doing it mindnumbingly over and over and over again.


if we have 3 Chairs elected via direct PR, then i think the delegate pyramid is enough for the Board -- we don't also need direct PR there


on approval/range being better than IRV:


3-ballot scheme for range voting:

" Voters produce three scores A,B,C indicating their evaluation of Gladstone and satisfying 0≤A,B,C≤99 and 99≤A+B+C≤198. The sum A+B+C is what matters and what expresses the voter's true opinion of Gladstone on a 99-198 scale (which is the same as the usual 0-99 scale, just shifted by adding 99). Voters produce three scores A,B,C indicating their evaluation of Gladstone and satisfying A=0 or 1, B=0 or 1, C=0 or 1, and 1≤A+B+C≤2. I.e. if you approve Gladstone, then approve him twice and disapprove once; if you disapprove, then disapprove twice and approve once. 2. Voters use a simple, low-tech, government machine to check the validity of their 3 ballots, i.e. to check that 0≤A,B,C≤99 and 99≤A+B+C≤198 in the 0-99 range voting case. Invalid 3ballots are discarded and voter can try again. If the 3ballot is valid, then the machine sucks in the ballots and dumps them in random order into a bin (to be used later in the election), says "thank you," and spits out a copy of one of the three ballots (the voter chooses which one) for the voter to keep for his records. It is important that the checking machine be so transparently simple that it cannot remember the 3ballots it checks and that it cannot remember which of the three ballots, the voter chose to copy. It has no memory. One (disgustingly cute) way to build such a check & copy machine in the approval-voting case is my tricolor scheme. "


this site likes score voting + an instant runoff:


ok here are two competing simplification proposals for the delegate pyramid:

1) There are 5 Board members (4 from the pyramid), and 60 Electors, and many Delegates. Any person or group may form a Constituency (or more than one Constituency, if they choose to split their voting weight), and each Constituency elects a single Delegate. The Delegates carry the sum of the voting weights of their Constituency's members. They may proxy this voting weight to any other Delegate, which has the effect of increasing the recipient's voting weight until the delegator withdraws the proxy (which can be done at any time). The 60 Electors are the Delegates with the most weight. The Electors each have a single Elector vote (regardless of the voting weight they possess as a Delegate), which they may reversibly proxy to other Electors. The 4 Electors with the most voting weight are Board members. They each have a single vote on the Board, regardless of the voting weight they possess as a Delegate or as an Elector.

2) There are 4 Board members from the Delegate pyramid, who are those Delegates who hold the most vote weight, including proxies (in case of a tie, none of the candidates participating in the tie sit on the Board if the numbers tied are more than the remaining spots out of the 4 Delegate spots on the Board). A Delegate may accept proxies from at most (5*(# of digits in the total number of votes in the organization))^2 other voters. Large voting weight can be combined by accepting proxies from other Delegates, who then proxy them to other Delegates, etc.

The advantage of #1 is that, with fixed numbers and no weird formula, things are simpler. The advantage of #2 is that it scales up, and better maintains the 'no mass campaigns', in comparison to #1, where potentially the 60 electors may function just like US Congress members (but maybe not, because here you can select a Delegate you actually like, without 'throwing away your vote').

I'm leaning towards #1 for simplicity.


in any case:

Each Constituency determines and administers its own parliamentary procedure, and, if applicable, membership rolls. These do not have to be made public; a Constituency may conduct its business in secret, and does not even have to know the identity of its own members, if it chooses.

The Organization's only role in the Constituencies is to:

A Constituency which has not received any votes for two electoral cycles, or which has not had a delegate for any time over two electoral cycles, is considered dissolved.

The requirements for "signatures" may be cryptographic, or notarized, or whatever the Organization rules require as necessary and sufficient to establish the authenticity of a communication.

They Organization may in no way interfere with the workings of a Constituency, nor many it require any sort of notice or registration of constituencies aside from the above list, nor may it ban or regulate any constituencies, nor may it ban or regulate or nominate or suggest any Delegates, nor may it ban or regulate or inspect any of the workings of Constituencies, including their parliamentary procedures or membership rolls, nor may it ban or regulate or inspect intra-Constituency communications.

Note that there is nothing stopping from a voter from voting for one constituency in the secret ballot, yet joining another constituency as a "member" according to the second constituency's parliamentary procedurem (or even joining multiple other Constituencies). Nor must the Secretaries of a constituency be "members" of that constituency or even voters in the Organization; they may be non-voter third parties who merely administer the constituency's business.


NOTE: this destroys cross-council Board-coloring though... OK so only Delegates can be on cross-councils. So, assignment of voting strength to 1st-level Constituencies is by secret ballot, then everything on up is public.

NOTE: we should have a minimum size for a first-level Delegation to prevent coercion. If this size is constant, though, then this either requires a special case, or makes things tough for small organizations; also the larger this is, the less people get to participate in cross-councils. The size probably needs to be at least 3 to have any hope of defeating coercion, and in fact probably at least 5. But i think 5 is still much too small. I guess i think 12 is the real minimum (or maybe 5^2 = 25?). But with vote-splitting, couldn't the coercers defeat this by sending messages in the vote amounts? Also, what does a minimum mean when votes are weighted?

if 12 is the minimum, though, and if only Delegates can be Board members, then we don't get our 4 Board members until the organization has at least 48 members. Which seems a tad large. But even if the min were 5, that would give us 20, which isn't much better. So maybe we should go larger than 12. Even on CommunityWiki? we got, what, like 17 people who wanted to participate in governance? So maybe 17 isn't too much to ask. Maybe we could just have the minimum be min(12, number_of_votes_in_whole_organization/5). And that makes me think we might want to scale the 12 with log(n) (eg something like 3*log10(n)); but that's getting back into scary-looking formulas, which we were trying to avoid..

if we did have scalable formulas, maybe say that each constituency must be between 3*log10(n) and 2*3*log10(n); but again, what does max size mean if you only have vote weights and not individuals? The right way to do it (which i computed before) would be a max vote weight as a proportion of the total number of votes, varying depending on pyramid level; but that's probably too complicated for most people. (recall, the issue here is that we don't want mass campaigns, so we don't want people just giving all of their votes to one of a few mass campaigning candidates; that is, we want to encourage or force small constituencies). Another way to motivate having small constituencies is to give Delegates some power on a one-person-one-vote basis, instead of by voting weight; like we did with the 60 Electors in one of the proposals above.


OK after sleeping on it, here is what i think:

a neat thing about the 60 Elector thing is that it 'automatically' implies (without extra verbiage) that in small groups, when there are less than 60 Delegates, the Elector layer is 'bypassed'.

i guess so far i like ceil(min(60, floor(n/12))); there's no 'log10' or '^2' so it's easy to explain, and 150 div 12 = 12, which is more reasonable than either 5 (150/30; from n/5) or 75 (150/2; from n/60) (5 is too few Delegates for n=150; and 2 is too small of a minimum size for n=150 (i would be worried about coercion by then). This saturates at n=720 (the first point after which there are more than 12 Delegates), and has another meaningful transition at 3600 (the first point after which there are more than 60 delegates).

well.. maybe 12 Delegates is too few for 720 people. How about

ceil(min(12, floor(n/12)))

That saturates at 155 (close to Dunbar's number), after which is the first time there are more than 12 Delegates. Another meaningful transition is at 720, after which there are more than 60 Delegates (and hence Electors become meaningful). Sounds good but is 12 enough to avoid coercion?

Could also do

ceil(min(30, floor(n/12)))

Now there are 12 delegates up to 360.


ceil(min(15, floor(n/12))). Now there are 12 delegates up to 180; critical point at 900.


ceil(min(15, floor(n/15))). Now there are 15 delegates up to 239; critical point at 900.


ceil(min(20, floor(n/20))). Now there are 20 (up to 39, actually) delegates up to 419; critical point at 1200.

So far i like ceil(min(15, floor(n/15))), which has the additional attraction that 15*4 = 60 (so it works as a minimum for Electors to coalesce, too, although we don't want secret ballots for Electors so this doesn't matter much), or ceil(min(15, floor(n/12))), which has the additional attraction that 12*5 = 60.

So far i like ceil(min(15, floor(n/15))).

i'm not sure the 'in case of ties, board seat is vacant' is good; seems that that would allow whoever is left on the Board to have too much power, temporarily. Perhaps just let more people sit on the board in case of ties!


ok to summarize/give verbiage for my current thoughts towards the delegate pyramid:

Define the number Number_Of_Votes to be the total number of votes (which is the same as the total number of voters, if each voters has one vote). Define a number called the Constituency_Minimum_Votes to be the lesser of:

Any member or group of members may found a Constituency, at which point they become the initial Constituents of the Constituency. The Constituents may be given different Constituent Voting Weights. The Organization cannot ban or regulate or refuse to recognize or refuse to register any Constituency for any reason.

Each election, there is a secret ballot vote where each vote can be cast towards some Constituency (if some voters have multiple votes, these may be cast towards different Constituencies). The voting strength of a Constituency is the number of votes cast for it in the previous election. A Constituency is inactive except when it has voting strength of at least the Constituency_Minimum_Votes. A list of the voting strength of each active Constituency is published by the Organization after each election.

Each Constituency may elect a single Delegate to represent it, and may recall or replace this Delegate at any time, and may alter its Constituent list including the Constituent Voting Weights. These actions are public and must be announced to the Organization, and these announcements must be signed by at least 2/3s (weighted by Constituent Voting Weight) of its Constituents. These signatures do not indicate that these Constituents agree with the action, merely that they are satisfied that the action was chosen in accordance with the Constituency's parliamentary procedure, whatever that may be.

The deliberations of a Constituency and its parliamentary procedure may not be regulated or spied upon or interfered with by the Organization, and the Constituency does not have to make them public or share them with the Organization, not even with the Organization's Chairs or the Organization's Courts. For example, a Constituency could, at its sole discretion, internally have a parliamentary procedure based on voting by a list of psuedonymous entities who are not officially Constituents, rather than allowing Constituents to vote; it would not have to share this fact (or share the list of psuedonymous voters) with the public or with the Organization.

The Delegates are the set of entities elected by the active Constituencies. The Delegates may at any time proxy some or all of their votes to other Delegates, or cancel these proxies, by informing the Organization. Each Delegate's Delegate Voting Strength is public, and is the Constituency Voting Strength of their Constituency, plus any votes that they have been proxied by other Delegates.

If there are less than 61 Delegates, then each Delegate is also an Elector; otherwise, the 60 Delegates with the greatest voting strength are Electors (if there is a tie for 60th place, then there may be more than 60 Electors). Each Elector has a public Elector Voting Strength which is distinct from their Delegate Voting Strength. Like Delegates, Electors may proxy votes and cancel proxies at any time. todo: no, this makes it confusing for small assemblies, see below at "whyElectorsShouldHaveUnequalStrengths": The Elector Voting Strength are 1, plus/minus proxies (they are not based on Delegate Voting Strengths).

One entity may hold multiple Delegate or Elector votes at once.

todo: no, this doesn't work with unequal Elector votes: If there are less than 5 Electors, then each Elector is also a Board member; otherwise, the 4 Electors with the most votes are Board members, in the same manner as Electors were chosen from Delegates.

todo: this is too wordy; probably the concept of recursive constituencies will help


could give cross-councils (Committees and Courts) 5 citizens instead of 4, plus one Judge; the four citizens are chosen by using Board coloring, and the fifth is anyone who is not represented by any Board member (just like the requirement for the indirect sortition voter base).


ok so we don't want to just have a few popular mass-campaigning candidates just have huge constituencies, right? But, in a corporate setting one person can hold many votes, and we don't necessarily know who the people behind the votes are or whether a block of shares is held by one person or many. Yet, in this corporate setting, we want one person with >= 25% of the votes to singlehandedly elect themself or their choice to the Board (and to all offices in the layers beneath it, holding multiple 'offices' of delegate under that Board seat), and if they have >=50% they should be able to singlehandedly elect two Boardmembers, etc. Can these by reconciled or are they provably contradictory?

They seem contradictory to me; if the system can't tell whether many votes are held by one person, and they see that many votes elected one person to the Board, and that person claims that they should be able to fill all offices in the layers beneath that because the votes all belonged to them; then how is the system supposed to distinguish the case where all those votes really did come from one guy from the case where many different voters voted for that guy? Similarly, if sockpuppets exist, then when there is a delegate subpyramid under some boardmember which seems to consist of many different delegates, how can the system distinguish the case where there are actually many distinct delegates from the case where these delegates are actually all just sockpuppets (or, at least, employees/conspirators acting on the direction of the Boardmember).

So maybe all that we can do is give people the option of delegating to smaller constituencies; maybe systems that can distinguish sockpuppets and can tell how many votes each person has can do better, but in the general case maybe we cannot. Otoh maybe we should just make the rules that we hope for, even if they are not enforcable in the face of sockpuppets, because then these other sockpuppet-less groups will benefit.


is there any way of working in the universalDelegateStructure stuff? Maybe as a default topical breakdown?

or mb a CEO candidate, when nominating themself (or accepting a nomination), should be required to say who their initial executive team would be, by assigning each of these roles to people (they can assign multiple roles to one person, and they can assign some or all of these roles to themself)? Hmm... that makes sense in the governmental context but in the corporate context surely most of the time a new CEO would say, 'i'll start out with whoever is currently doing these things, and go from there'.


other things in a parliamentary procedure:


the standard "order of business" may be too complicated. But otoh, there's obviously good reasons for it. And, order of business is a workflow-y thing that most members (even most boardmembers) don't really need to understand in order to effectively participate. So should we simplify it? How?

one issue i have is preventing the officers, committees, and chairs from "DOSing" the agenda by taking up too much time, not leaving enough time for new business. A lot of Robert's Rules is complex scheduling motions; i presume that part of the function of these is to allow the assembly to take back control over the agenda if someone tries to. But we want to simplify all that, which means we may need another mechanism to ensure that there is time for new business.

in the standard "order of business", officers get to 'report'. Should officers really be special?

Should we give the Chairs some power to just set the agenda? If so, should they vote (so 2 of them can ally against the other 1), or, since we have a proportional slate of 3 chairs, should we let each of them introduce one issue per meeting?

As for the new business itself, in the parallel/async form, and possibly also the serial form, we wanted to use proportional voting to choose agenda items, rather than just 'whomever the Chair recognized first'.

Should we have officer and committee reports only place proposals into a pool of things that have reached the floor, but then they have to compete as usual for a 'slot' in new business?

How do we parallelize this? And, can we unify the parallel and serial parliamentary procedures here, or do the serial ones really need a different structure due to scheduling needs?

general note: parliamentary procedure is partially about attention allocation. Related, the word 'parliament' etymologically has to do with talking, and parliamentary procedure has to do with giving the minority a right to be heard before the vote.


regarding the committee structure:


using the above system, if committee members were random i.i.d. coinflips (a very poor assumption; it's likely that some things really are objectively/clearly good or bad proposals, which would cause the members to act together in a correlated manner), what are the chances that a proposal would pass or a defendent would be declared guilty?

At each step, we use the cumulative distribution function of the binomial distribution, where k is the number of votes needed, and n is the number of voters:

sum_i=[0,k] (n choose i) p^i (1 - p)^(n - i)

In Python, we can use scipy.stats.binom.cdf(k, n, p). Note that, eg binom.cdf(3,6,0.5) > 0.5 but binom.cdf(2,6,0.5) < 0.5, and binom.cdf(3,6,0.5) + binom.cdf(2,6,0.5) = 1;

In our case, n is always 6, by assumption p is 0.5, and k varies from 5 (supermajority) to 4 (majority) or 3 (split).

So, in a committee, the chance of being rejected as spam is the chance that only 0 or 1 guys vote to accept, which is binom.cdf(1,6,0.5) ~= 0.11 (so 1 - 0.11 = ~90% chance of acceptance)

Given not being rejected earlier, the chance of being rejected in the lower tier topic committee is the chance that only 0 or 1 or 2 guys vote to accept, which is binom.cdf(2,6,0.5) ~= 0.34 (so ~66% chance of acceptance).

And given not being rejected earlier, the chance of being rejected in the lower tier topic committee is the chance that only 0 or 1 or 2 or 3 guys vote to accept, which is binom.cdf(3,6,0.5) ~= 0.66 (so ~33% chance of acceptance).

Since we are assuming i.i.d. (an unreasonable assumption here, btw), the chance of making it through all three of these filters would be (1 - 0.11)*(1 - 0.34)*(1 - 0.66) = 0.2 (20%). This is encouraging; i was worried that the multiple filters would make this much lower than the 33% imposed by the last filter, but it didn't.

For the courts, if we also adjusted the lower topic so that a 3 vs 3 vote were sufficient for 'innocent', then the chance of being 'rejected' (innocent) is 0.11 for the first one again, then binom.cdf(3,6,0.5) ~= 0.66 for the second one, then binom.cdf(4,6,0.5) =~ 0.89 for the third one, so overall the chance of making it through all three filters (and being declared guilty) is (1 - 0.11)*(1 - 0.66)*(1 - 0.89) ~= 0.03. Contrast with a system where we kept the same thresholds as with committees, and then added another court at the end: (1 - 0.11)*(1 - 0.34)*(1 - 0.66)*(1 - 0.89) ~= 0.02. Contrast with a system where we keep the same threshold as with committees for the lower topic court, but then adjust the higher to have a 5/6: (1 - 0.11)*(1 - 0.34)*(1 - 0.89) ~= 0.06. This seems better.


to simplify, let's ditch the spam/jurisdiction stuff. Whoever submits to a court chooses the court; instead of rejecting (but by the same procedure), the court can forward the case to a different topic, if need be. The lower court does spam control but can also reject for substantive reasons. Ontology changes are legislative. So now, in each topic, there are just three tiers:


committees and courts are the same except for (the rest of this line is the same as above): (a) 'reject' is replaced by 'innocent' and 'accept' is replaced by 'guilty'; (b) a supermajority, rather than a majority, is needed in the higher-tier topic for guilt, (c) committee members can vote 'guilty of breaking the rules but the rules were unfair', which has the effect of an 'innocent' vote.

actually we can simplify it further; eliminate the 'meta' in each topic (a difficulty with the meta that i hadn't mentioned yet is that a party could game the seniority system by interjecting moles into other parties and then having them out-work the real members of those parties, so that everyone in the single, meta court is of that party. Clearly this is a much more expensive way to target the merely high courts, because there is not just one such court per topic, as with meta, but rather many). Eliminate appeals (why do we assume the appeals court will have a greater probability of reaching the correct answer than the lower court?). Also, we selected Judges in slates of 3, proportionally, but now we are only staffing each court with one of those three, giving a partisan slant to each court. So, in the 'high' committees/courts, have a slate of 3 Judges who were elected together, for a total of 8 members (4 Board-colored, 1 Board-independent, 3 Judges). Now a supermajority is 6/8; by the numbers, this should only make it very slightly easier to find guilt or to forward a proposal (about 85.5% instead of 89% in the coinflip model).

Note: the "low" court can be thought of as a kind of grand jury; an expedited review to prevent harrassment (and prevent the cost of a full trial) by terminating obviously bad prosecutions early.


regarding the 'floor'; the discussion slots are chosen by score voting; if a proposal is supported (at least a +1) by at least 1/3 of the votes, but doesn't win the floor, then it stays on the floor for the next session (possibly indefinitely); but if a proposal is not even supported by at least 1/3 of the votes, it dies.

how can proposals reach the floor?

note: the Chairs do NOT get to place things onto the Floor, because they only have negative power.


some principals for the commentary:

judge selection: by a committee with broad membership and a delay

any faction greater than or equal to 1/3 has the right to be heard


y'know, theoretically i think real runoffs are better than instant runoffs (but i'm not positive, b/c instant runoff could encourage non-strategic ("honest") voting earlier?) but instant runoffs give a better user experience, so let's do that.


todo i should read the standard code, and also:

also mentions Donald A. Tortorice's The Modern Rules of Order , not sure if i should read that too


actually, when delegates split their proxies, we DO have to keep track of the identity (source) of each vote, so that we can Board-color the Delegates.


ppl are chosen for the Courts/Committees on a first-come-first-served basis (a FIFO queue; the sooner you sign up onto the waiting list, the sooner you will be chosen), except for the Board-coloring constraints (eg we go down the waiting list and choose each person who is not a Board-colored duplicate, until we have 5). The Judges are chosen in order of least score (eg so as to roughly equalize how many cases each of them has done/how much time they have spent) (that gives them a slight incentive to rush the cases, b/c a Judge who tends to take longer than average will be chosen first, eg will have less 'extra' vacation when there is no backlog, but that's probably good, and it's only slight; in fact Judges who enjoy using their power might be slower on purpose for this reason, so it's hard to say what effect it has).

the Courts/Committees should be done pseudonymously for members except Judges (only the Board coloration is given) when feasible.


i guess electors can't proxy their votes (so vote splitting is not an issue), they can only support Boardmembers



ok wait if electors have only one vote then it's really annoying for small assemblies, because if there are <60 Delegates and some have many more votes than others, you can't just say they're each an Elector, rather, the Delegates with a lot of votes will want to select multiple Electors, or possibly select the same person for multiple Elector 'offices'. So, let the Electors have unequal votes.


mb make the 'juror' pool for both Forum committees, and Courts, and the scoring, the same (!). This has the following effects:


i am a little worried about introducing jurors, who are mostly Delegates, into the Forum process. This means that if some Stalin-like character can manage to interfere with the Constituencies (despite the prohibitions on such) and control their nominations of Delegates, then they also have control of the Forum committees.

The reason to only have Delegates on Courts and Committees is in order to enforce Board-coloring on them while still allowing secret elections at the lowest level of the delegate pyramid. As a bonus, this also serves as a nice way to keep Delegates relevant beyond eg the US Electoral College.

otoh, slightly encouraging is that the 5th 'indirect sortition' member of each Committee is not necessarily a Delegate. Also, ANY delegate is eligible to be a juror, not just those with a lot of votes, so provided the Stalin-like character only controls some of the Constituencies, we are fine.


if there are no Judges on the lower committees (and even if there are, if there is only one judge and not a slate of 3), then we can't trust the lower committees to serve as envisioned in 'deliberative polling' or 'citizen juries'; they might just not conduct a 'trial' or listen to 'evidence' and just vote their knee-jerk opinion. In which case, since it only takes a majority to reject, anything which the majority knee-jerk disapproves of will lose. So maybe require a supermajority on the lower committees to reject? Otoh maybe it's good to prune things unlikely to pass early.


so, can lower committees ever amend proposals? i think yes, at least, with a supermajority (this allows the committee to save proposals with obvious flaws, instead of just rejecting them). Higher committees should be able to amend the same way they approve, that is, with a majority (this is similar to allowing the member of higher committees to originate proposals, which perhaps should also be allowed?).

but, if it's supposed to be easy for lower committees to accept, then why should it be harder for them to amend than to reject? Perhaps in lower committees, both amend and reject should be by majority?

But in that case, why not just make the proposals by committee members, rather than submitted by others?


so, we have at least three choices:

the last one is still worrying; the mob can pick out, eg 15 candidates, and then on each committee tell one person to vote for each of those 15 candidates. If they do this over many constituencies, they can get a lot of indirect sortition votes for each of those candidates. So, if there is indirect sortition at all, it must be pooled across the entire voting population.




i have no idea if this is relevant, but apparently it describes the modern Russian constitution and the cultural and historical context in which it is applied:

The Constitution of the Russian Federation: A Contextual Analysis By Jane Henderson


here's how the US conservative political party (the Republicans) in 2016 plans to repeal a health insurance bill after they take control of both houses of Congress and the Presidency next year, but without a 60% supermajority in the Senate:

" It would be quicker for Congress to simply repeal the health law outright. But Republicans can’t do that, because they would need 60 votes in the Senate to fend off Democrats’ delaying tactics, and they will only have 52 GOP members. So instead they will be limited to using a special budget strategy that will let them pass their bill with 51 votes.

That so-called “budget reconciliation” measure does not let lawmakers repeal the entire law — only the parts that directly impact federal spending. That has been widely discussed. There has been less focus on how long the process takes.

No one in Congress can simply introduce a budget reconciliation bill. The word “reconciliation” refers to the process by which congressional committees that control permanent spending programs such as Medicare and Medicaid, as well as tax policy, take action to “reconcile” that spending with the terms of the annual budget resolution. ... The budget resolution, which is essentially a planning document for spending and taxes for the coming fiscal year, does not go to the president for a signature. But, like a regular bill, it does have to be passed by both the House and Senate in the same form. And while the budget resolution also may not be filibustered in the Senate, lawmakers have up to 50 hours to debate it, and unlimited time to vote on proposed amendments, which in practice can take up to another full day.

Once that measure is agreed to by the full House and Senate, the action moves back to congressional committees. The budget resolution often includes “reconciliation instructions” to committees. Those instructions order proposed legislative changes to the programs the committees oversee to meet the terms of the budget. That triggers the reconciliation bill that goes to the president.

In the normal course of events, those changes take from several weeks to several months to accomplish. Legislative changes need to be written, voted on by the committee and reported back to the House or Senate budget committees, which then forward them to the House or Senate floor for votes. Again, Senate debate is limited to 20 hours with unlimited additional voting on amendments. House and Senate negotiators then hammer out a compromise, pass it again in the full House and Senate and then send it to the president. " --


still wondering what to do about replacing a quorum with something that would work for the Forum, where only a tiny fraction of eligible voters will vote

my current best guess is to have the 'quorum' be based on how many people showed up last time, except also that each time the quorum is not met and a meeting is skipped, the quorum decays (proportionally auto-reduces).

another idea is to introduce 'RSVPs' (when a meeting is announced, ppl can RSVP and the quorum is only based on the RSVPing population), but (a) people will tend to RSVP and then not show up, and (b) a faction can obstruct business just by instructing a bunch of their semi-apathetic voters, whom they don't trust to show up, to RSVP


when running a meeting manually (without electronic aid), non-proportional score voting can be easily done by the chair calling out each alternative, then each person raising zero, one, or two arms, and then the Chair counting how many arms were raised for each alternative.


when doing score voting for proposals (vs alternatives and vs status quo):

if there is a tie then revote, where only the tied items are alternatives; if the revote ends up with another tie amongst all of the (previously tied) alternatives, then the 'status quo' wins



should look for FAQs for Robert's Rules, Mason's Rules, etc, and try to design the system so that either these questions don't come up, or the answer is easily found.



here are example bylaws for a committee in a real organization (NOT using branch democracy, just an example of some bylaws):


does anyone else have 3 co-chairs in an assembly using a parliamentary procedure? the only ones i could find are:

maybe should ask them for their thoughts on how 3 co-chairs works out in practice

--- Why Concord?

Why do groups need a formal 'procedure'?

Informal decision-making works great most of the time; but eventually, a controversial and divisive issue will arise; or perhaps some person will attempt to grab power; or perhaps factions will develop and things just get 'political'. Without a procedure, the organization can become paralyzed by indecision; be taken over; or become so intolerant and conflict-ridden that key contributors leave. Having a good procedure in place BEFORE times get tough is essential for allowing the organization to continue to function, resist power grabs, and fairly take into account the views of all members.

Why Concord Democracy instead of other procedures?

Why not just use, for example, Robert's Rules of Order? Concord Democracy is similar to other procedural systems; it's an evolutionary update, not a completely new thing. The devil's in the details, and it's in the details that Concord shines. Compared to the alternatives, Concord is:

Disadvantages of Concord Democracy



need to justify the claims made in Why Concord?



Extant parliamentary procedures were designed in a time when the only means of efficient discussion was physically meeting in a room. This:

As a result, extant procedures don't make full use of modern communications technology, which enables asynchronous structured discussion amongst large groups of people.

In fact, some extant procedures impede the use of modern technology. For example, quorums which are a significant fraction of the total membership are ill-suited for online discussion forums, where only a tiny fraction of members may be present at any one time.

Concord is a procedure which can make full use of modern technology, while also being able to be used without any technology.

--- has a good, straightforward layout (i have some issues with its substance though). It needs more intro summarization, though.

it starts out with "2. Decision-making bodies and individuals", which lists components of the system. There is also a list of "general rules". it then has one section for each listed component, with subsections (where applicable) for (a) powers (b) composition or appointment (c) procedures (note: i would add a section 'purpose'). There is then another section on "Assets held in trust for Debian" which clarifies some legal stuff regarding trustee entities. Then there is a section on parliamentary procedure and voting procedure.


(yet another) Concord Democracy intro

Concord democracy is simple procedure for making decisions, electing officials, and adjudicating disputes that can be used by any organization.

How are decisions made?

There are three groups that make decisions:

Most votes in the Forum and the Board need to get 60% of the vote in order to pass. This encourages compromise without allowing a small minority to hold up business.

How are officials elected?

How are disputes adjudicated?


and what if the Board and the Forum disagree? Does it go to a Committee or a Court? Surely we can't have the Forum-appointed committee chairs on a committee; but we don't want to involve judges, either, right? How about letting the Chairs do it? This politicizes them and gives them some positive power, but that might not be as bad as giving to the Forum appointees or to Judges in this case. We could also let the Chairs just chair the committee meeting but not vote.


to encourage service on Courts, maybe high court service counts towards low committee? so:

low court service counts towards high court high court service counts towards low committee low committee counts towards high committee ?


mb have the High Court set the ontology of Courts and Committees


mb eliminate the 2/7 'Delegate' votes on the Board; because

a) it makes explaining how the Board works more complex b) it introduces the inefficiency of mass voting to the otherwise-efficient Board c) it adds an element of 'mass campaigning' to the Board


mb the High Court can overrule any other court, but only unanimously


what to call the people nominated for indirect sortition? Could just call them 'delegates', i guess. Maybe they need to get nominated in this fashion in order to be a Delegate, just to make it simple.

ok, i got it. Call these 'nominations for Board', and you have to be nominated to serve on the Board. You can self-nominate.


toread some great stuff in here: "MASONS MANUAL AND RULES OF THE LEGISLATURE The principles of parliamentary la w and the rules of the legislature – connected to power to adopt rules—Mason reinforces much of what was already covered "


Democratic Electoral Systems around the world, 1946 – 2011 Nils-Christian Bormann a , Matt Golder

should also read

Patterns of Democracy and Its Critics Nils-Christian Bormann


those authors would maybe also be good ppl to ask for advice from

mb should also read stuff in

good intro to meeting minutes at


"There are, accord ing to Mason, Ten Fundamental Principles that Govern Procedure 1. The group must be so constitu ted and endowed that it has th e power and authority that it purports to exercise. 2. There must be a meeting of the gr oup at which the d ecision is made. 3. There must be proper notice of the meeting so that all members of the group have an opportunity to attend and participate. 4. There must be a quorum present at the meeting. 5. There must be a clear question before the group for decision. 6. When the decision is being made, there must be an opportunity to debate the question. 7. The question must be d ecided by taking a vote. 8. There must be a vote in the affirmative of at least a majority to make a decision or carry a proposition. 9. There must be no fraud, tricke ry or deception causing injury. 10. Decisions must not be in violat ion of laws, rules, or decisions of higher authority. "


--- commentary

note that the initial Secretary is the second Chair; this will tend to be the opposition to the plurality party, providing a check against incumbent corruption of electoral procedures (although the other two Chairs could forcibly Swap this role away from the second Chair).



why the odd procedure of two Chairs forcibly Swapping one of their roles with the third? Because the roles of President/Secretary/Treasurer could be unsuited to some people, yet their are initially assigned without any regard to the expertise of the Chairs; imagine, for example, if one Chair is a financial expert, and another is a expert parliamentarian with no head for numbers, yet the former becomes President and the latter Treasurer. One or the other might wish to hold onto to their ill-suited role for some reason, much to the consternation of everyone else. The natural organ for personnel decisions would be the Board, yet in this case we don't want to allow a majority of the Board to conspire to remove oversight by placing an allied Chair in any particular role. And we don't want to politicize the High Court by giving them any personnel decisions. So we leave it up to the other two Chairs; these people already have the authority to repeatedly overrule the ill-suited Chair, so why not let them save time and Swap.


use the word 'policy' instead of 'rules' for the generalization of 'law', that is, proposals that are not only adopted as resolutions, but which are fully 'signed into law'.


the name 'FirstChair?' is a placeholder for a better name. I didn't use 'President' because i get the impression that US corporate law sometimes considers this another name for CEO, which could be confusing in the case of third parties assuming that such-and-such person had signing authority in some case. Mb we should call this person Chairperson, and give the other two Chairs other names?


hmm, it seems that both 'chairperson' and 'president' have undesirable connotations and possibly implications in US corporate tradition. These probably don't matter much but it still gives me pause. It's easy not to use the word 'president', which sometimes means CEO, so let's avoid that. Maybe call the first Chair Speaker? More difficult is whether to call the Chairs "Chairs". We could just call them Speaker, Secretary, Treasurer, but this obscures their peer-ness, and also it might tempt organizations to amend the Bylaws in ways inspired by the traditional system where the CEO appoints the officers and in which the officers don't have indepedent "auditing" powers. The Speaker/Secretary/Treasurer stuff is a less important recent addition to the system, just because i figured, hey, in addition to chairing the meeting, someone has to keep the minutes, so we may as well have a Secretary, and then why not add a Treasurer too so that we cover all of the three traditional officers (CEO, Secretary, Treasurer); i don't want organizations amending the bylaws later to lose sight of the main idea of the Chairs because of these titles. We could call them something new, like "tribunes", but i feel that the more "weird" names we have, the less likely people are to use the system.


what minimal frequency should we have for elections?

four and five years is common in governments, and it is thought that any shorter would be too short term. US Senate terms are 3 years.

US corporations need an annual meeting

often there are quarterly reports. Some student governments have semester elections.

since we're aiming for adoption by corporations, let's make it one year. I'd prefer five years, but organizations that want that can always amend.


probably a good blog (i already took the quiz on this page):


'courts' also moderate the Forum


old verbiage:

Second, why is Concord Democracy better than other procedures? Compared to the alternatives, Concord is:


this interesting article praises Germany's court system, in which the judges, not the parties, seek evidence:

He mostly attributes the goodness to the fact that the German career path for judges is very selective, and provides incentives via peer review and promotions for good judging. However, he also thinks that the lack of a procedural distinction between pre-trial and trial is good, because the existence of such a distinction in the US prevents the judge from having time to do additional discovery after trial has started.


going back and forth between using log or not in the Constituency minimum size rule. Right now i have:

"a Constituency is only active if, in the previous election, the Constituency received at least 15 votes, or (for smaller organizations) it received at least 1/15th of the total votes available in the organization, rounded down)."

but i worry that in larger organizations, this small size (15) will be just enough to allow the mafia to try and identify and then coerce votes. I remember some old study on sizes of cartels and some of the larger cartels had 30 members.

What about sqrt? This leads to large constituencies in large groups; eg sqrt(250e6) ~= 16k (otoh, is 16k really that large? in the US the smallest congressional district (for the lower house, the House of Representatives) is over half a million [2]). But 16k still seems like a lot to me from the perspective of a small party getting off the ground locally; eg in 2016 the U.S. Green Party in Florida only had ~5000 members.

So back to log. Can we think of a way of phrasing the log requirement which works for small groups (so we may need a disjunction in there, as with min(15, n/15) above) but is still easy to read for non-mathy people?

let's try.

"a Constituency is only active if, in the previous election, the Constituency received at least...

...a number of votes greater than or equal to (the number of digits in the count of the total votes available in the organization, times five)

...a number of votes greater than or equal to (the number of digits in the count of the total votes available in the organization, times five), or if it received at least 1/15th of the total votes available in the organization, rounded down)." "

the second one is definitely harder to read. The first one is maybe a little more confusing, but just barely acceptable. But it doesn't allow any constituencies below n=5, only allows one constituency below n=20, and doesn't allow four constituencies until n=40.

i guess this isn't terrible; below n=5 you don't really need fancy rules anyhow, and below n=20 it seems okay to demand that a CEO candidate gets support from more than half of the members. Also, the Board wouldn't be vacant even if no Constituencies form; there is still the indirect sortition seat. Still, this forces randomness on a group with n<5. Maybe just say that below n=60, everyone is an Elector?

not sure that this additional complexity is worth it. Maybe just use a bigger number than 15 in the old formula? Maybe 30 or 60?

otoh, my fear that 15 is too small doesn't have much evidence behind it, and perhaps a bigger number will end up disempowering dissent. The larger things are, the more organization is needed, and so the more advantage is had by existing power structures. For example, imagine a society in which people's immediate families pushed the children to stay in a 'family constituency'. One the one extreme, with n=3 two parents and a single child could form their own constituency, and so the parents could probably tell if their child defected; similarly with n=5 you could maybe do the same with three children. With n=14 you could have 6 grandparents and two descendent families with two children each; this is getting to the point where you wouldn't be able to tell who is defecting, but in practice the pressure will be so great that someone might admit it. So far this sounds like an argument for larger numbers; but consider if you make n=60, and people tend to come together in small family-based 'tribes'. With n=14, you could also imagine a bunch of young people from different families who know each other banding together to form their own constituency. With n=60, forming this new 'young person's constituency' would take a serious organizational effort, as it's unlikely that you'd find a clique of 60 somewhat politically active, rebellious young people who all know each other.

So, let's just leave it at 15 for now.


contents of meeting minutes. The minutes do not contain notes on discussion. Mb just read




hmm, currently you can only give forum proxies in elections, and they are secret ballot, and they expire on the next election; so there is no way to have transitive proxies. This is too bad, as i imagined that for e.g. space policy you'd proxy your vote to your science-following friend, who would then re-proxy them to an expert that they know about. But i can't think of a way to have transitive proxies in the general election without too much danger of voter coercion.

one thing we could do is have forum proxies last longer than the forum proxy election pyramid. For example, what if forum proxies lasted for three electoral cycles? Then you could only get one hop on the first cycle, but you could get two hops by the second cycle, etc. Similarly, within one election, you could hold a number of forum proxy ballots in quick succession to let the proxies move along.

But is this worth it? in addition to the complexity cost, there are other disadvantages to transitive proxies. If you had transitive proxies, what if the mafia tells Person A and Person B to proxy to Person C, and Person C and Person D to proxy to Person E, etc, until finally the last recipient in the tree is a mafia member? Now (a) each intermediate recipient is only receiving from two people, giving the mafia a 1/2 chance of guessing who refused to cooperate if someone disobeys their orders; and (b) most people are being ordered to proxy to other non-mafia folks; so even if the ratio of mafia to non-mafia is small, they don't run out of potential recipients; and (c) it's easier to identify voters if (a) voting is frequent or (b) voters can vote for anyone out of a large group; since you can proxy to anyone, (b) holds, and if we have many ballots then (a) would hold as well.

for now i think i'll keep it simple and avoid transitive proxies.


mb look at

they apparently assign a rating to each US state based on many criteria:

also, since their PEI project surveys political scientists, they could probably recommend a bunch of procedural experts to me to ask advice from. They could probably also give advice themselves.


majority of boardmembers must be independent, unless 2/3s shareholder vote to waive?


in boardmeetings, intersperse informal discussion with formal discussion, at the discretion of the chair, by unanimous consent?



we want multiple chairs because the chair's investigatory powers will be more useful in the hands of a minority faction member. Why not just 2? Because we need an odd number in order to prevent deadlock


do Electors form formal constituencies to elect Boardmembers? Do delegates form intermediate constituencies?

i think no -- they can do that informally anyways, and it's simpler not to have that


further ideas for simplification:


can we unify the board and the forum's parliamentary procedure more? the forum being just the written, electronic, async, parallel version of the same procedure?


term limits: must be out of office for a time exceeding the lifetime sum of time. office ?

(right now we just have "...nor may they hold any of these offices if for the majority of the past nine electoral cycles they have held...")


one issue i have with the current Board election system is that the Boardmembers either have no terms of office, or very short terms. Given that we already have the Forum to capture recent popular sentiment, i thin it's important for Boardmembers to have longer terms. I'd say 3-5 years would be ideal.

what shall we do about this? some ideas:

i guess there was a reason to have Boardmembers be replacable by their own supporters, though; to make them pay attention to their supporter Delegates (or Electors, at least) after the election

on staggering, we could easily have 3 of the 6 Elector-selected members elected each Electoral Cycle, rotating so that on even cycles seats 1,2,3 are filled, and on odd ones, seats 4,5,6 are filled; and so each seat has a term of 2 electoral cycles (why not 1,2,3, then 2,3,4, then 3,4,5, etc? B/c on the 1,2,3 year, seat 1 will probably be filled by the plurality faction and seat 2 will probably be filled by the opposition, and so then in year 2 2,3,4 will be replaced and 1 will not and 2 will probably be filled by the plurality faction, and then in the next year 3,4,5 will be replaced and 2 will not, so both seats 1 and 2 will probably remain occupied by the majority faction, etc. This could possibly be evened out by some carefully planned arrangement but that is too brittle in that some groups may amend it and 'break' the balance accidentally. 1,2,3 vs 4,5,6 is easy to understand)

on replacing mid-term, there might be some way to "interpolate" between allowing it and not, by allowing it but with some sort of high voting/support threshold

another great thing about staggering is that we can reduce the size of Board-colored committees from 6 to 3. So now preliminary screening dispute resolution panels are only 5 members (3 Board-colored members, 1 Nominee, 1 Judge), instead of 8; and full panels are only 7 members instead of 10 (2 more judges were added from the preliminary screening numbers); still too large for my taste, but much more managable.


maybe have the Delegates and Electors have a fixed term, but the Boardmembers can be replaced mid-term; or maybe vice versa; or maybe put the cutoff between Delegates and Elector instead of between Electors and Boardmembers


with 6 Elector-selected Boardmembers and 60 Electors, there is at most an average of 10 Electors supporting each Boardmember. This sounds like few enough that, assuming the Electors weren't themselves up for re-election, they might be persuaded to take the long view/altruistic view and not recall Boardmembers who are doing something good for the overall Organization but bad for those Electors.


could consider reducing the number of Electors from 60 to 30, so that they could (easily) fit into an (sort of) ordinary room

this might also increase their chances of staying relevant; if there are ~150 Electors, it's harder for them as a group to avoid being dominated by the 7 Boardmembers. Also, i imagine that small groups might balk when they see the number "60" anywhere in the Bylaws, even though it isn't actually an impediment for them.

with an average of only 5 members per Boardmember, i think we could definitely hope that even if the Electors have the power to recall the Boardmembers, this wouldn't lead to campaigning considerations dominating the hard decisions of governance, at least not as often as if there were tons of Electors per Boardmembers

(is 30 a good number? it's a very divisible number ("abundant number", but not SHCN/CAN; 60 is SHCN/CAN). it's divisible by 6.)


maybe another way to keep the Delegates more interested is to allow them to call for early elections, without the consent of the Board

--- commentary

principal: political parties are not required (unlike eg party list PR voting). However, the system assumes that political parties may form and is designed to function well if they do.


the UCSD GSA council representatives currently number at around 140

funny, when i observed a few council meetings, i remember the meetings fitting in an ordinary-ish-size room, as if there were more like 30 ppl present. Maybe they don't all come to every meeting, or mb the numbers increased since i was present, or maybe i misestimated how many ppl were there


ok so right now i think:

could stagger both Board and Elector terms, actually, and still have quarterly elections, with semi-staggering Board:

that's getting complicated, though.

Could just stagger both Board and Electors 2x, and match up even-cycle Electors to odd-cycle Boardmembers. E.g. 15 Electors each cycle, and those 15 elect 3 Board members NEXT cycle. I like that.


ok so right now i think:


or, could stagger Electors 3x, and have:

this is more complex for little gain, though; with staggering, i bet the Electors would remain independent regardless. It would be nice if we could work in 3x staggering somewhere though...

just do ordinary 2x and 3x staggering for the Board and the Electors.

but if we only have 10 Electors elected each year, shouldn't we increase the size of the Elector pool again?


if still worried about the Board dominating Electors, have 3 Board members with quarterly terms, 3 with annual terms. But right now i'm not worried enough to do that (rather, i'm MORE worried about terms being too short).

still considering:


for longer-termism, could stagger Elector terms 5x instead (5 Electors each cycle). But 3x is probably the max for many organizations, such as student governments, whose members tend to be affiliated for less than 5 cycles. For Boards of Directors, however, 5x is probably better, so mb do that.


ok so right now i think:



some objectives in the Board-selection process (with Delegates, Electors, etc):


one difference between the Chinese communist party intraparty governance procedures and our Board structure is that we try to surface disagreements (both factional and individual), whereas they try to subordinate them to democratic centralism, consensus, and a show of unity


one thing to worry about if there are 6 Electors elected each cycle, for 5 terms (30 Electors total, 5x staggered), is that even though only 3 Boardmembers are selected each cycle, consider if the same 6 Electors get elected over and over again; first the top 3 Electors will join the Board, and then the other 3 will, since the first 3 already have seats (assuming the prohibition on multiple seats is in effect). Which, in this case, effectively removes the layer of indirection that the Electors are supposed to provide so that the Boardmembers are chosen by interviews rather than mass campaigns. So need more than 6 Electors elected per cycle.

so mb do 30 Electors, 3x staggered, 10 Electors per cycle. Or, mb do 24 Electors, 3x staggered, 8 Electors per cycle. I guess i think 10 Electors is heathier, because 6 is almost half of 10.


some robert's rules notes (already read):

the second one, role of the chair, i should reread

this one is also good:


" Common Errors in Parliamentary Procedure Cooperatives March 08, 2011 Print Friendly and PDF

Authors: Phil Kenkel, Oklahoma State University,, and Bill Fitzwater, Oklahoma State University

        Prolonged discussion without a motion. This tends to violate the principle of “one thing at a time.” It is the main reason the chair gets in trouble conducting meetings. Discussion without a motion can become rambling argument rather than constructive discussion; the chair may stop this rambling by requesting the business be placed before the group in the form of a motion.
        Failure to confine discussion to the motion before the house. It is the chair’s job to keep the meeting on track. The chair can rule a discussion out of order.
        Failure to know and follow the essential steps in the presentation and disposition of a motion.
        The prevalent belief that one individual’s calling “question” forces an immediate vote. This obviously is not true since it would violate the principles of majority rule, rights of the minority and courtesy. There is a procedure for forcing an immediate vote, but it requires a motion of “previous question” and a two-thirds majority vote.
        Closing nominations too quickly when conducting elections.
        Voting on candidates in the reverse order from which they were nominated. Robert’s Rules of Order states that candidates should be voted on in the order they were nominated primarily because the most competent candidates are usually nominated first.
        Not calling for additional nominations when a nominating committee is used. The nominating committee selects the candidates that it feels are most competent; however, this does not mean that they must be accepted by acclimation. The floor must always be opened up for additional nominations.

" --


" Federal law defines crimes of violence in 18 U.S.C. 16, which says that a crime of violence is any offense "that has an element the use, attempted use, or threatened use of physical force against the person or property of another" or "any other offense that is a felony and that, by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense." "




make sure to avoid most of these:

Ten Bylaw Amendments That May Lessen Your Company’s Vulnerability to a Shareholder Activist

        Require shareholders to provide advance notice of nominations and other proposals;
        Require the proposing shareholder to be a record holder of the company’s shares;
        Eliminate any requirement for the annual meeting to be held on a fixed date; 
        Prescribe qualifications for all nominees for election as directors;
        Provide directors with the exclusive right to fill vacancies on the company’s board of directors;
        Restrict who can call a special meeting of shareholders;
        Limit actions that can be considered at a special meeting of shareholders;
        Provide a procedure for setting the record date for actions by written consent;
        Define “cause” necessary to trigger shareholders’ right to remove directors for cause; and
        Permit the chairman of any meeting of shareholders to adjourn the meeting whether or not a quorum exists.


Recent Developments in Delaware Corporate Law: The “Technology Amendments” to the Delaware General Corporate Law

" The most significant changes implemented by the Technology Amendments relate to: • the conduct of stockholder meetings solely by means of electronic communications (i.e. without a physical meeting site); • the ability of stockholders to be deemed physically present, and to vote by ballot, at stockholder meetings when participating electronically; • the delivery of notices to stockholders by electronic means; • the delivery of director and stockholder consents by electronic means; and • the maintenance of a stockholder list on an electronic network. "

DGCL Section 211(e) is one of the sections that enables electronic stuff




a remaining open question is, when one of the forum/board passes a proposal and the other vetos it, what happens.

you could just say, the proposal fails. But then it would become difficult to ever amend the bylaws in a way that changes the balance of power between the forum and the board.

you could say, compare the voting strength on each side, and one wins (maybe first a revote for the final tally). One problem with this is that, in an amendment to the bylaws that changes the balance of power, you might find very high supermajorities in the losing organ voting against.

you could say, the chairs then vote. This makes sense because they are a third organ of government, and they are directly elected, which seems to be what you might want to decide a power struggle between the forum and the board. And it increases the visibility and majesty of the chairs (especially because they are the final arbiter) which is good. But it means that the chairs could now run on a platform of issues, instead of just the negative platform of providing better oversight, exposing conspiracies, etc. I'm not sure how often this would occur or how much the occasional possibility of deciding this sort of thing would influence the sort of chairs that get elected; maybe it's a non-issue and we should just have the chairs decide.

you could say, the judges (or the procedural panel) then votes. This makes sense because they are a third organ of government, and because they are experts; they are perhaps the ones who could make the best decisions in this sort of case. But we don't want to politicize them, and also they are appointed in a process that gives the board more weight than the forum.

you could say, a board-colored committee decides. But:


could use quarterly voting to give a 4-step proxy system with all blind ballots in the Forum


i'm concerned about the practical implementation of Asset Voting for the selection of the Board.

Asset Voting works when the end body is continually re-elected. But (a) real companies won't want to bother with that, and (b) we want to give Board members a fixed term of office. So when is asset voting "done"? If there is a deadline, won't voters change their votes at the last moment (like ebay "sniping") so as to confuse their opponents and cause 'wasted votes' and 'spoilers'? Can we have some criteria of "rounds" where there is a new "round" if some criteria is met (something like "some representative got more votes, in a direction that would change the set of winners" or "the set of winners changed"), and otherwise not? We would like to guarantee termination. seems unconcerned about this; they say "In variants, only certain kinds of redistributions are permitted, for example the candidates drop out in increasing order of "wealth", redistributing their assets among the remaining ones as they do so. This "poorest first" variant has the advantage that it forces the procedure to end quickly and everybody knows when it has ended. But I prefer unrestricted negotiations and instead assuring termination by simply declaring a pre-specified deadline "all negotiations end at midnight November 7." S.J.Brams suggested the refinement that in addition to some fixed deadline at midnight after (say) 5 days, there also be intermediate deadlines each prior midnight; if the winner-set has not changed since the last one then the negotiations end (early) then. And it would be beneficial to have some kind of publicly-readable bulletin board announcing at all times the current status, i.e. how much assets each candidate currently owns, and the terms of any inter-candidate deals. ". has a slightly longer discussion with more suggestions.

We could let the candidates themselves indicate when the redistribution is 'finished', but this would itself be a vote, and you could have a minority (or a majority!) 'vote strategically' to keep the redistribution open even when it is actually 'finished' because they don't like the result (a form of filibustering).

One thing we could do is in each 'round', pre-determine that the bottom candidate will be eliminated, and must proxy eis votes to someone else. This guarantees termination, but in a large number of rounds.

Instead, we could eliminate the bottommost third of candidates (rounded up, in the sense that we eliminate more than, not less than, that amount). This is better; the order of the number of rounds (big O notation) is now dependent on the log of the number of candidates. But the number of rounds is still unbounded, and also the last few rounds will 'wasted' eliminating very few candidates.

Instead, we could eliminate a fixed fraction on each round, increasing the fraction as the rounds go forward in such a way so as to fix the max number of rounds. For example, for four rounds (including the initial secret ballot), we could do: 0: secret ballot, 1: the bottom third of the candidate (minus the number of seats to be elected), rounded up, are eliminated and must proxy their votes, 2: the bottom half of remaining candidates (minus seats), rounded up, are eliminated, 3: all remaining candidates (minus seats to be elected), rounded up, must proxy.

Note that we still have an edge case to deal with when we have to eliminate so many candidates that not enough remain to claim all of the seats, due to ties. This includes the edge case where all candidates are eliminated (!). To deal with this, we might allowed candidates to proxy their votes to candidates who were 'eliminated' in the current round, allowing a chance to break the tie afterwards. This doesn't help much though because there still might be a tie. Another solution is just to allow the Electors to get bigger than usual to accomodate ties.

Now, how long should each round be? For the same reasons as above, we don't want to have a vote on this. We could just leave it up to Policy or the Secretary to set. If we specify a time, we could invent a new timeframe, or we could use the smallest timeframe currently mentioned in the Bylaws, Subcycle, which by default is a quarter year. This might be a good idea because we are assuming that many organizations won't want to bother with any election stuff more frequently than quarterly. If we use quarters, then we could use the 4-round system given above.

This would have the interesting effect of making the full process of an Elector election take a year! On the annual election, the Voters elect Delegates in a secret ballot, and then each of the following three quarters is consumed in a round. On the next annual election, the Voters elect new Delegates, while the newly elected Electors cast their votes (along with other senior Electors) in the election of new Boardmembers. Last year's Delegates don't become Board-colored until that time, so their term as a committee-eligible Delegate doesn't start until then.

So this introduces a delay of 1 year between when the voters at large cast their Delegate votes, and when the Electors cast their Boardmember votes. This is probably good for (a) buffering the Electors from popular pressure, and (b) making it so that the Electors actually evaluate the Boardmember candidates themselves, rather than being party faithful selected on a pledge to vote for some preexisting candidate. On the other hand, it probably makes Stalin-like 'circular flow of power' attacks, where the incumbent leadership effectively intimidates/bribes Delegates into voting a certain way, more likely.

I consider 'circular flow of power' a more important danger than popular pressure on Electors. But i'm leaning towards this system anyways, because i think it is practically feasible for a large number of organizations, in ways that the obvious non-mass-election alternatives (continuous re-election of Boardmembers; a multi-day 'convention' during which Delegates spend lots of time meeting each other, negotiating, and proxying their votes over multiple rounds) are not (in many organizations, participation in voting will be only one of many side-projects in even many of the Boardmembers' lives, so no one will have time to commit to (even an online) 'convention', and the organization won't want to bother with continuous re-election). We have to remember that most organizations and Boardmembers consider governance to be an annoyance, not something that merits a large amount of attention (even if we, those people who think it's important enough to spend time designing governance systems, disagree with that assessment).

On the other hand, those 'pragmatic' people may well think it's pretty silly that we spend 3 quarters having a staged election that could have been done in three weeks. Hmm.


This still doesn't completely solve the problem of a Delegate telling supposed 'allies' that they will vote one way and then actually voting another way, causing either wasted votes, or coalitions to fail to elect a candidate because they planned for exactly no wasted votes. You might say that a problem with this particuar asset voting procedure is that it REQUIRES strategic voting to avoid wasted votes. I'm not sure this is really a problem until the Electors elect the Boardmembers, though; before that, i suppose candidates with 'too many' votes will just hold onto them. Is there something we could do at the last step? We could have the Electors use Proportional Score Voting, but then we lose the idea that each Delegate is Board-colored.

This Board-coloring idea may be more trouble than it's worth. The benefit is that we can have Board-colored committees/panels, whose stratified nature allows us to involve low-level Delegates while still keeping a high probability of having factional proportionality.


sooo.... anything else we can do at the last step? I guess maybe after the 'vote', we could allow candidates with too many votes proxy the excess to others. But since there are multiple candidates doing this at once, if done naively, this could allow a leading candidate to need more votes than they orginally did and give away too many, right? Well, so compute 'excess' in a way that takes into account the worst-case behavior of the other two leading candidates.

This is getting too complex. My inclination is just to leave it, and let the danger of Electors switching loyalties at the last minute stay.


Incidentally, for Board-coloring to work, we probably have to make each Delegate proxy ALL of their votes at once. This is because, otherwise, if we Board-colored based on their 'original vote' and not votes that were proxied to them, then they can easily be a 'mole' at very little cost by voting their proxied votes to their true preference and their single 'original vote' for the enemy. If we Board-colored based on their plurality vote, they could do something similar, though at much greater cost, by splitting their votes into three parts, two of which eventually went to two different, but allied, boardmembers, and the third of which was their plurality, which went to an enemy.


incidentally, also notes another potential problem with asset voting: "In a 3-candidate single-winner election, suppose each candidate pre-announces, before the election be- gins: “My negotiation strategy will be as follows: If I have t he fewest assets, then I will award all my assets to candidate X (just which X , depends on that candidate)” In that case the election will behave very similarly – indeed , under the right circumstances identically – to a so-called “ in- stant runoff” (IRV 19 ) election, and therefore can exhibit all sorts of “voting paradoxes” analogous to those suffered by IRV. A voter who increases his asset-vote for C at the ex- pense of his second-favorite B , can actually cause C to lose by causing B to be eliminated in the first round so that B ’s assets transfer to A (whereas otherwise A would have been eliminated with his assets transferring to C ). This also serves as a counterexample to the idea that plurali ty- style “all assets to one” asset voting always is strategical ly best. (Many other such counterexamples also exist.)..Indeed David Gamble has argued to me that in practice he suspects most asset voters wil award their entir e vote to their favorite candidate, and candidates (in the ne- gotiation stage) will transfer their entire assets in one bl ock to other candidates. If so, Gamble argues, Asset voting will behave much like STV voting, except worse from the point of view of the voter, who under these conditions would not be able to control the transfers, but with STV, would. "

Of course in that scenario there were very few candidates, and it was a single-winner election, neither of which is our situation. In our situation, we prize asset voting because it allows us to go from very many voters to very few seats without a mass election dominated by soundbites.


actually, instead of scheduling candidates for elimination at the beginning of each round, have a vote threshold to survive the end of each round.

also, there's a difficulty because ideally we'd like a power-law sort of threshold decrease, but we also want to keep the formulas simple. Maybe at some point multiple by 100, then by 1000? But then in small groups that would be too much. Also, is it too much to ask the secretary to find the votes held by the 10th largest candidate in each round? If we just use the largest candidate, we'll overestimate the thresholds; maybe we can just take 1/10 of the total votes cast? That will also overestimate (because the top candidate will usually take more than 1/10th, so the 10th candidate will need less than 1/10th) but at least 1/10 of total votes is simple to explain (and somewhat simple to calculate).

could achieve the desired log-scaling with simple wording by saying "halve the number of digits"


i'm wondering if perhaps we should give the Chairs more power. I was writing a summary and i started writing "the 3 Chairs lead...". The original idea was that the Chairs would be the highest office in the land, and seen that way by the voters, but right now they seem more concerned with audit-like investigations and with administering procedure, which i doubt will capture the voters' imagination -- which is dangerous, because if the voters don't care about them, other officials might try to resist them. How could we make it true that the Chairs LEAD the organization, without giving the Chairs enough positive power that they end up running on a platform of how they'll use their positive power, instead of how they will protect vs corruption?

some ideas (all ideas i've had before):

should most of this stuff go to the Speaker, or to every Chair? Should you need a majority of Chairs for this stuff (eg a Chair can only give a speech at a Boardmeeting when another Chair 'seconds' them, which in practice would probably lead to two Chairs allying and cutting out a third)? I'm inclined to put it all to the Speaker, to increase their majesty (and b/c i dont want everyone else to have to read 3 Chair announcements all the time), but then that doesn't leave the other Chairs with much...

how about:

treasurer: board and ceo compensation secretary: publication of updates for the members

awards? nah


notes towards a proposed rewrite of Board meeting procedure:

The Speaker presides over the meeting according to the following principals:


some verbiage:

So, there are two expedited subcycles during which Elgible Boardnominees may Back each other, and at the end of which those Boardnominees whose Voting Strength are below average become inelgible; then one more expedited subcycle, at the end of which the top 6 Boardnominees become Boardelectors; with the possibility of early termination if the top 6 Boardnominees achieve 5/7ths of the remaining votes. The Boardelectors then have one more expedited subcycle to Back each other, after which the top 3 Boardelectors are elected to the Board, and also an Unaligned Boardnominee is randomly chosen, weighted by Initial Boardnominee Voting Strength, to be the 4th Boardmembers.


perhaps we should remove the possibility of 'formal discussion' from the Board meeting. When a Board is operating under Robert's Rules for Small Boards, do they have an option to pass to formal discussion?

otoh formal discussion only adds one or two paragraphs to the Bylaws, and it seems to be that regardless of what Robert's does, there is some value in allowing a >1/3 minority to demand time to speak over the objections of the Chair and the majority.


todo remove language stating numbers of chairs and replace with majority, unanimity


some verbiage:

Disputes are resolved by seven-member Tribunals (3 Judges, 3 Aligned Boardnominees, and 1 Unaligned Boardnominee). The three Judges on these Tribunals are selected by a Judicial Selection Committee of 7 members chosen by various diverse groups.

overturn directives and policies that conflict with the Bylaws, and with a supermajority vote, they can force the resignation of Judges, order a re-examination of a dispute in which proper procedure was not followed, veto amendments to the Bylaws, or propose amendments to the Bylaws which increase the rights of members.


ugh, after all that 'simplifying' redesign, Concord is still way too long!

it's now about the same length as the old theLongVersion!

it's about 30 pages, and that's by wordcount or character count, so it's not like there's too much whitespace. Ideally it would be less than 10 pages, but i guess that's probably impossible. Wordcount is 13450, which is about 3x the length of the US Constitution! So maybe a 3x reduction is not impossible.

it's longer than the Debian Constitution!

yuck yuck yuck

if i look at it in critical mode, it's like "why is it so wordy? why are there slightly redundant procedure specifications? What are all these new parts -- chairs? boardnominees? boardelectors? JSC? do we really need all those? why do the Chairs and the Procedural Tribunal have so many zillions of powers? Why are there both Temporary Exceptions and also Emergencies? Isnt it a bad sign if you need a Governance Log?" Of course i actually do have good reasons for all of that stuff (and, it's less pages than Robert's Rules, and less than the US Constitution + Senate procedural rules, etc), but it's a bad first impression. i think there's a lot of work ahead for further simplification.


some quick simplification ideas:


interesting discussion:

" Another typical stock ownership method to avoid the danger of majority control is to issue stock in equal amounts, giving each side a "veto" right on the other. This is often done in family companies or in partnerships which incorporate. The benefit of such a method of stock ownership is that no single shareholder can dominate. The danger is deadlock of the company since there will be even numbers of votes. If there is a deadlock in the voting of the shareholders or the Board of Directors which endangers the company, any director or shareholder has the right to petition the court to break the deadlock by appointing a receiver to run the company for a specified period of time, presumably long enough to break any deadlock. This process is expensive, due to the court appearances necessary, and often extremely expensive since the receiver must be paid. Further, the receiver seldom is expert at the business (usually being a lawyer) and it is seldom that a company can survive long under the auspices of a receiver.

Our office has developed its own manner of seeking to avoid the danger of deadlock. We have had success when stock is owned equally by creating a contractual method of breaking tie votes in which the shareholders agree in advance on a provisional director to serve automatically if a tie vote is encountered, said provisional director breaking the tie vote and if the director is required more often than three times a quarter, said director will serve for one year. Not only does this cost far less than engaging in court proceedings, but the practical effect is that the shareholders almost always find a way to resolve their differences for the simple reason that the practical effect of the tie breaking methodology is that BOTH shareholders lose control whenever a provisional director is appointed since he or she will have the swing vote. Realizing this, each of the shareholders invariably agree on a compromise vote. The very threat of this provisional director being available creates consensus in almost every case.

While a receiver seems to create the same threat, in reality the court process will cost tens of thousands of dollars before the shareholders realize that the ultimate result of their efforts is to lose control: the provisional director method costs next to nothing. "


"The commissioners on uniform state laws drafted the Uniform Business Corporations Act in 1928, though only three states adopted this act. The American Bar Association in 1950 drafted the Model Business Corporation Act, which subsequently has been modified numerous times. The last major redrafting occurred in 1984, but there were substantive revisions in 2002 and 2005. A large majority of states have adopted all or a significant portion of the Model Act. Other states have modified their own state corporation statutes to contain sections similar to the Model Act. Delaware’s corporation statute is also significant, since most large, public corporations are incorporated in that state"

new goal:


does Delaware or other law force a shareholder, not director, vote on these things?:


Matters upon which shareholders vote, in addition to the election of the directors, depend on the issues affecting the corporation. The following are the most significant and usual as to these matters.

    Approval or disapproval of changes in the articles of incorporation
    Approval or disapproval of a merger with another corporation
    Approval or disapproval of the sale of substantially all of the corporation’s assets that is not in the ordinary course of the corporation’s business
    Approval or disapproval of the voluntary dissolution of the corporation
    Approval or disapproval of corporate transactions where some directors have a conflict of interest
    Approval or disapproval of amendments to bylaws or articles of incorporation
    Make nonbinding recommendations about the governance and management of the corporation to the board of directors
    As noted above, many of the rights afforded to shareholders are contained in each corporation’s articles of incorporation or bylaws. It is also noteworthy that shareholders generally do not have the right to vote on management issues that occur in the ordinary course of the corporation’s business. Many decisions of the corporation must be made by the board of directors or officers of the corporation, and in many such day to day decisions, shareholders may not compel the board or officers to take or refrain from taking any action. Shareholders, however, can vote to remove directors and appoint new officers, subject to restrictions in the by-laws or any employment contracts.

Minus agreements or restrictions in the by-laws, and with some restrictions imposed by law, corporations retain the right to issue new shares of stock, which could dilute the ownership of existing stockholders. Existing shareholders often hold preemptive rights, which allow the shareholders to purchase these new shares of stock before they are made available to the public. Thus, if a shareholder owns 10 percent of a corporation, and the corporation issues new stock, the shareholder would own less than 10 percent if he or she did not purchase new stock. If the shareholder exercises preemptive rights, he or she may purchase as many new shares as necessary to retain that 10 percent interest. " -- [3]


so one issue we have with the C corporation structure is the minimal quorum requirements for shareholder meetings, b/c in online communities these would almost never be met by the Forum.

idea: in stocks, your broker can vote as your proxy by default. Why is this allowed, and can we generalize this and use something similar to be able to have quorum?

"under Delaware law, the Company’s domicile state, broker non-votes are deemed to be “present” at the meeting for quorum purposes in that they are shares that have arrived at the meeting via the broker. However, they are not shares that are “entitled to be voted” on certain proposals, as the broker has no authority to vote the shares with respect to those proposals. Under Delaware General Corporation Law, Section 216, “the affirmative vote of the majority of shares present in person or represented by proxy at the meeting and entitled to vote on the subject matter shall be the act of the stockholders” (emphasis added).

New York Stock Exchange (NYSE) Rule 452 specifies when banks, brokers and other member organizations (collectively, “custodians”) may authorize submitting a proxy on behalf of the stock’s beneficial owner. In uncontested elections, custodians may submit proxies without any specific instructions provided by such beneficial owner on matters that are considered “routine” or “discretionary.” NYSE Rule 452 requires the custodian to mark through any proposal on the proxy card that is considered non-routine or non-discretionary, creating a “broker non-vote.”

The Delaware Supreme Court held in Berlin v Emerald Partners, 552 A.2d 482 (Del. 1988) that broker non-votes are to be treated as limited proxies of the beneficial stockholder, and while not considered votes cast on a specific proposal, they are considered to be present for purposes of determining whether a quorum exists at the meeting: " -- (which goes on to note that the brokers can't vote on certain kinds of proposals

so, we see that (a) for some reason the NYSE has some jurisdiction over this, and (b) Delaware supports it.

more on this:

more on how NYSE is involved:

and more detail: also talks a bit about quorum requirements starting on page 15. Some commentators are worried about quorum requirements. But others said that "several commenters believed that quorum concerns could be addressed simply by including a “routine” item on the ballot, such as the ratification of auditors, or with appropriate changes in state law to permit shares held by brokers to count solely for purposes of establishing quorum" -- the "routine item" thing is because broker's default proxies can't vote on certain extraordinary matters, but they can vote on routine items, so there should be at least one thing on the ballot that they can vote on in order for them to count.


potential probs:

if register, register loss causes actual vote

each person can only introduce 3 alternatives? doesn't help in multiple votes sockpuppets scenario.

each person can only support 3 alternatives? hard to keep track of in person




should we have a forum quorum, but require voters to either attend meeting, or proxy vote to someone who does, at pain of losing your vote until next meeting attendance?