![[Home]](http://www.bayleshanks.com/cartoonbayle.png)
Difference between revision 27 and current revision
No diff available.why P=17? P has 3 main effects :
1) group sizes of senate , high council 2) group sizes of councils 3) scaling transition point when council comes into being
for 1, want odd number and perhaps prime number . also want medium sized group -- small enough to have (inefficient )discussion , but large enough to have resistance to group think . its okay if it is too big to have efficient discussion -- that's what the cabinet is for . i fear that P=13 is too small because a faction of 2 (15%) or 3 (23%) will feel social pressure not to speak -- with P=17, the entire group is almost a crowd, hence a faction of 3 (18%) or 4 (24%) can delight in being nonconformist rebels -- in other words, i think that with P=13, small factions will be dominated by the group identity , but with P=17, they will act as factions . this is what is desired -- the senate and high council are supposed to be a microcosm of society, not a monolithic group .
another consideration for 1 is that some representatives will happen to misrepresent their constituencies, whether out of incompetence or malice or honest difference of opinion . in addition, if there are multiple orthogonal dimensions of ideological position, you need at least m^d factions to represent m positions along each dimension . allowing "c" to represent the number of representatives per faction needed in order to make it probable that at least 1 is representative, then cm^d is the minimum number . for c=2, m=2, d=3, we have 16, and for c=2, m=3 , d=2, we have 18. so, unless you're comfortable with having less than 2 representatives per faction, 13 appears to be too small . another way to look at it is to let "f" be the minimal desired factional size needed to win a representative, and then to look at P=cf. for c=2, P=13 yields f=15%, P=17 yields f=12%. by this metric, both of these seem reasonable.
for 2, want efficient discussion . for this purpose , 11 would be nice . however , in order to fulfill primary purpose of fighting group polarization effect , popular council size must be exactly the same as high council size . i think it is more important to fight group think in the representative legislature than to facilitate efficient discussion in popular councils . i recommend councils split into 2 groups of 8 and 9 for half of their meetings, and then have plenary sessions for the other half .
for 3, don't want council when group is too small -- 11 yields 121, which seems too small .
so i think 17 is best, but 13 is the runner-up; which one you choose depends on the balance to strike between fighting groupthink in representative legislature and having effective popular counsil discussions .
i initially thought that perhaps another reason to have 13 rather than 17 would be to make decisions more quickly, especially in corporations . however , right now i think that
the ACLU may have 83 directors on its board of directors, however, it remains to be seen if the average corporation would accept replacing a 7-person board with four commissions with a total of 48 commissioners (17 central and 7 external delegate commissioners and 17 central and 7 external elected commissioners). with P=11 this would decrease only to 33. with P=7 it decreases to 20.
if we allow external commissioners to also be central commissioners, and assume that they always are (it may be a good idea in any case to have most of them be central also), we have 34, 22, and 14.
it's probably also a good idea to have some external specialists on the external side. i think i'll leave it that central + external can overlap but don't have to -- i.e. whoever gets elected/delegated can serve.
the elects and delegates' commissioners should be disjoint, however; and the tribunes should be disjoint from both the commissioners and the management.
i'm putting the councils' influence over the elect commission rather than the delegate commission (which would seem to be the more logical place), because the elects are already very unconstrained vs. the delegates.
currently i think the councils' influence should be exercised in terms of casting "council votes" in the elected commission. the sum total of possible "council votes" could be equal to the number of votes in the elected commission (i.e. 17 for internal issues, 7 for external). these could be divided up equally between the layers of the delegate constituencies, except for the delegate commissions, which are not included. for example, n = 290 there is just the members and then the delegate commission; in this case there are 17 members' central councils, and each of them has 1 vote in the elected commission, and 41 external councils, and in total these 41 have 7 votes in the external elected commission, so each of them has 7/41 or about 0.17 votes. as partially noted above, these votes may only be exercised when consensus-2 (in the case of the central councils) or consensus-1 (in the case of the external councils) (note: based on the considerations below, this has been changed to consensus-2 and consensus).
let's estimate majoritarian bias if we only allowed true consensus (caused because sometimes people will end up on the same council who already agree with each other), we calculate as if there is a 50% chance of each individual disagreeing with their faction. so if there is a council where just less than 2/3s of the total population is in favor of something, then on a 17-person council there are (rounding up) an average of 12 members in favor (and 5 opposed). so, .5^(17 - 12 = 5) yields the chance per council that the entire council will agree without anyone needing to compromise. note that, for the opposition, the relevant figure is .5^12, much smaller; so in this case we expect the proponents to have a lead of (.5^5 - .5^12 ~= (# councils)*.5^5 = (# councils)/32. if there are 17 votes for all the councils, then this is 17/32 = about half a vote. which seems tolerable.
note: this calculation is incorrect because it is only giving one side a chance to flip; it's an upper bound, tho
for the 7-person external councils/commissions, it gets worse: 5 votes in favor, 2 opposed, 1/4 of the total votes = almost 2 votes of skew. o.t.o.h. one can argue that the point of the smaller external group sizes are to provide a more unified voice, so this majoritarian skew isn't as bad.
if 5/7 commissioners would be in favor, we have a 2/3s majority anyways, so mb we should round down, or not round. so now, we have .5^(7 - 2*7/3) = about 20%, or about 1.4 votes. let's try when 4/7 members even are actually in favor of something (57%). we have .5^3 = 1/8 = 12.5% of the votes, or about .9 of a vote. so in this case, 4/7 isn't enough to push the vote over 60% (but just barely).
what about if 60% of members are in favor? we have .5^(7 - 3*7/5)*7 = just over one vote. in other words, the distinction between 60% supermajorities and 2/3 supermajorities are practically abolished in this system for the external councils; if 3/5 members want to do something "external", (and assuming there is no actual deliberation and compromise in the councils), then the external elected council will find itself having not just 4 but 5/7 votes in favor (at least 4 elect votes, and 1 council vote).
(but how many councils cast "no" votes in this case? we have .5^(7 - 2*7/5) of the total, for .5^(7 - 2*7/5)*7 no votes, or .4. i don't like to rely on this, but if you do, then that brings it back down to .6. at 2/3s, we have .5^(7 - 1*7/3)*7 = about 1/4 of a "no" vote, for a total skew of .5^(7 - 2*7/3)*7 - .5^(7 - 1*7/3)*7 ~= 1.1 votes -- so now this doesn't seem so bad.
for consensus - 1, it gets worse:
.5^(6 - 2*7/3)*7 - .5^(6 - 1*7/3)*7 = 2.2 votes (at 2/3s) .5^(6 - 3*7/5)*7 - .5^(6 - 2*7/5)*7 = 1.25 votes (at 60%)
man, what's it like for consensus-2 at 17 people? it's
.5^(15 - 2*17/3)*17 - .5^(15 - 1*17/3)*17ans = 1.3123 (votes at 2/3s)
.5^(15 - 3*17/5)*17 - .5^(15 - 2*17/5)*17 ans = 0.55244 (votes at 60%)
so, mb we should make it consensus-1 for 17, and consensus for 7:
.5^(16 - 2*17/3)*17 - .5^(16 - 1*17/3)*17ans = 0.65616
.5^(16 - 3*17/5)*17 - .5^(16 - 2*17/5)*17 ans = 0.27622
that's better, but i worry about consensus-1 for a transient 17-person group.
so i think we'll stick with consensus-2 for the 17-person group, and consensus for the 7-person group.
for an example if there are multiple levels of delegate constituencies, with P = 17 at n=8104, we have 17,20,23. so, there are about 20*23 member's coucils with a total of 17/2 votes among them, and about 20 second-tier member's councils with 17/2 votes among them. so there's about (17/2)/(20*23) ~= .02 commission votes for each bottom-level council, and about .425 commission votes for each intermediate-level council. at n = n= 8886111 we make the transition to 2 middle layers and we have 17,37,80,175. so each bottom-level council has
(17/3)/(37*80*175) ans = 1.0940e-05
votes, the lower middle councils each have
(17/3)/(37*80)ans = 0.0019144
votes, the upper middle councils have
(17/3)/(37) ans = 0.15315
votes each.
hmm, that's discouraging. but remember that the councils have their own dd interface. so, a small number of low-level councils can still force the elect commission to debate a proposal, even if it's difficult for them to force its passage. the vision is that perhaps the councils, driven by their consensus processes, will propose innovative forms of compromise bills. also, if there is actually discussion and deliberation, there may be issues in which a large number of councils cast votes, because one side of the issue is actually more persuasive than the other.
i'm wondering if the tribunes should be called "investigators" to emphasize their adversarial nature. i worry that ombudsman sounds like too much of a mediator.
hmmm, reconsidering, now i think that perhaps the arguments i gave for 17 rather than 13 work fine with smaller numbers because the number that those arguments applied 2 is the total number of commissioners, not the number in each commission.
also, i feel that most private organizations will balk at having 34 commissioners in place of a board of directors.
maybe we can even go down to 11. with 11, the size of the entire "board of directors" replacements could be as small as 22 (+ 3 tribunes = 25 representatives).
with 11 and 5, and absolute consensus in both primary and external councils, the council majoritarian biases are:
for 5: .5^(5 - 3*5/5) - .5^(5 - 2*5/5) = 12.5% at 60% so the distinction between 3/5 and 2/3 is effectively abolished in the external elect commission; however, this doesn't happen until you get above 57%: 2.85/5 = 0.57, and 2.85/5 + .5^(5 - 2.85*5/5) - .5^(5 - 2.15*5/5) = .65662)
for 11: .5^(11 - 3*11/5) - .5^(11 - 2*11/5) = 3.7% and note that 57% doesn't even get pushed above 60%: 2.85/5 + .5^(11 - 2.85*11/5) - .5^(11 - 2.15*11/5) = .5947
certainly there will be more groupthink in a group of 11 than a group of 17. perhaps governments will want to have P=13 or 17, and other organizations will want 11.
besides a smaller total number of representatives, a huge benefit of P=11 or P=13 is better deliberations in councils.
with P=11, the numbers are:
22 commissioners + 1 tribune = 23 representatives (in place of a board of directors) 1 CEO + at most 3 external officers + at most 6 internal officers = at most 10 officers, or 11 if the board chooses a separate EEO (upper management)
(research showing that boards >7 are bad: Board Size Effects in Closely Held Corporations, Morten Bennedsen, Hans Christian Kongsted and Kasper Meisner Nielsen, 1http://www.cebr.dk/Publications submenu/Discussion Papers/2004/DP 2004-23.aspx later: mb this is now at: http://www.econ.ku.dk/cam/wp0910/wp0203/2004-25.pdf )
(research including a histogram of fortune 500 company board sizes: decision making dynamics in corporate boards, Stefano Battiston, Eric Bonabeaub and Gerard Weisbuch, http://dx.doi.org/10.1016/S0378-4371(02)01930-1 -- 11 is right in the middle of the histogram, but 20 is way out. however, hopefully this structure works like "2 boards of 11" rather than "1 board of 23")
for organizations with unequal members (such as corporations), the voting strength of members in the forum, in the commission election, and in the constituencies varies (some members may form constituencies, or even multiple constituencies, all by themselves). however, each elect commissioner should have equal strength in their capacity as elect commissioner, to promote collegiality and independent thought. a powerful member can still exercise their disproportionate influence in the elect commission by the threat of the power of their vote.
in the forum, a powerful member directly exercises their strength.
in the delegate commission, a single delegate may have the strength of multiple delegates only if their constituency is a single member.
so for example, let's say that there is one person who holds 40% of the stock of some corporation.
they have the voting strength to elect themselves to about 4 spots on the elect commission (in practice, maybe sometimes a little less, due to the centralizing factor in the reweighted range, or the corresponding Condorcet factor in the Loring vote) (and they also have another "left over" .4 of a vote that influences the election of the other elect commissioners)
they would control 4.4 of 11 delegates in the delegate commission -- this means that they can stand themselves as a delegate in the delegate commmission, having 4 votes -- and they still get some votes in other constituencies (to account for the other .4).
so in this case, there are only about 8 people on each commission.
term limits and commissioner separation are waived when a single member has more than 1/11 of the votes -- that member is free to sit on all commissions as long as s/he has that many votes, or to give a proxy to any other person to vote in their place, either on individual issues or in general.
this prevents the preset sizes of the commissions from getting in the way when a substantial fraction of the power is really concentrated in a few individuals.
similarly, a powerful individual (or a pooled faction rep) can be placed in multiple councils. they don't even have to attend them (although it may annoy others if they don't...) -- if they want to never attend most of them (to lessen the power of the councils, which is against their interests after all), they can simply indicate such, and other individuals will be virtually placed with them and then "blocked". in fact, i guess we should let any citizen ruin the fun for others in this fashion -- in fact, it should be automatic for anyone who does not join a council. but the assignments should automatically try first to place the blockers together. de facto "blocking" by nonattendence should not be allowed; those who request council placement and then do not attend should be removed.
actually, i guess "blocking" automatically happens anyway if you refuse to be placed in council; if other factions want to be placed more than you, tough beans, they don't get to.
this means that if anyone controls more than 1/11th of the vote, they are on every council and can block all councils simply by not signing up!
in other words, any faction with 1/11th or more power can "temporarilly deactivate" the council system at will.
perhaps people who don't sign up can, at their option, allow councils to go on without them; this is equivalent to agreeing whatever the rest of council agrees to!
maybe it should, rather, be like this. it's unlikely that an entire 11% faction will have the solidarity to block, because if a few members defect, they will have lots of power (because there will be only a few councils). in fact, if a faction tried to block, it might be worth it for the other factions to infiltrate them so as to have a few non-blockers.
so, this will only really happen when there are a few individuals with >11%. so, give people who want to not attend council the power to allow the councils to go on without them. if they refuse, they may annoy everyone else.
(yes, there is always a first tribune. this may seem unnecessary, but it hammers home the importance of separation of powers, as well as of the majesty and high status of the tribune)
i changed my mind about the lifelong powers of the tribune. this would (a) be a lot of responsibility, which might discourage people from running; they would be bothered for the rest of their life by people seeking assistance, and (b) would make the voters think twice about electing a tribune who is strongly anti-secrecy. so tribunal powers expire at the end of their term.
why so difficult (2/3) to refer to referendum? because if the people really care, the forum can take it up. we prefer business to be transacted in the forum because the transitive proxy voting makes it more deliberative than a straight-out vote.
if less people than the preset numbers for commissions/tribunes stand for election, that's okay -- everyone who stands is elected, and that's all.
the distinction between the officers of procedure and the constitutional court is that the OP is constantly involved in the day-to-day affairs of the commissions, whereas the court is only invoked if there is a problem that some part of the state (including the tribunals) 'takes to court'. furthermore, the OPs just make decisions willy-nilly, whereas the CC goes through a formal court proceedings for each case. the OP has final jurisdiction over questions of procedural details within each house of the legislature, and on which things need which majorities. the CC is obliged to respect the OP's decisions in such cases, except in case of corruption or mental problems. the CC has jurisdiction over anything else, including all substantive (rather than structural) rights, and over "big picture" items in the constitution involving the powers of and relation between different organs of government, or between government and individuals. the CC can also do the other things mentioned in the constitution: pardon individuals, Force Clarification, create law when Forced Clarification fails, and declare Found Rights, whereas the OP can do none of these things.
So now, for an organization with n=>144, the minimum number of high officials is:
for 27 in total.
and the maximum is:
for 46 in total.
how many high officials there are in between 27 and 46 (inclusive) is determined by the commissioners -- if they wish to keep it at 27, they can simply refuse to appoint all (or any of the) judges and parliamentarians, they can make one person both CEO and CER, and they can confirm only 1 Cxx.
note that some commissioner spots may well be "independent outsider" spots, say on a corporate board.
compare to today's corporate boards; there is probably a corporate counsel (in place of 3 tribunes), and about 11 board members, 1 CEO and 1 President, for a total of 14. So we have 13 extra (the other commission, and 2 extra tribunes).
adjusting n for ownership: to calculate n when voters have unequal strength, order voters from strongest to weakest, and then start adding up the stakes of the strongest until you get over 50%. Multiply the number of voters you have gotten through by 2 to get an effective n.
For example, if there is a 40% stakeholder, a 30% stakeholder, and a lot of small stakeholders, then n = 4 -- only the forum exists. If there is a 40% stakeholder, and two 10% stakeholders, then n = 6 -- again, only the forum exists. If there is a 20% stakeholder, two 10% stakeholders, and four 5% stakeholders, then n = 16 -- primary commission size is 3.
note that for n=3.5e8, the total number of people in the top two layers of the delegation pyramid is 649, only about a hundred more than the size of the U.S. Congress. so if people worry than 11 (or 13 or 17) reps just aren't enough, they should be reassured that a representative body the size of Congress does exist in this system, it just isn't on top -- but in Congress the true situation is similar, in that real power rests with party leaders and committee chairs, of which there are much fewer than 535.
the tao of this system (compared to the tao of python):
explicit is better than implicit small groups are better than big the individual is better than the group peace is better than war less legal text is better than more both direct democracy and representative democracy have something to contribute both campaigning and small group delegation have something to contribute talking to the opposition makes a person more informed concentrations of power are scary
note that if there are no OPs and 2 tribunes, only 1 tribune (the senior one; the positions rotate as tribunes end their term, and the junior one becomes the new senior one) becomes OP. the OP can delegate functions to the other tribune, but is not obliged to. the other tribune still retains their fundamental tribunal powers of investigation, declassification, and prosecution.
mb make tribunes more like the head of state by giving them "competency to grant pardon, decorations and titles of honor, to address messages to the people and the Assembly and to promulgate the laws"? -- http://www.cecl.gr/RigasNetwork/databank/REPORTS/r4/GR_4_Katrougalos.html . i dunno -- it would be kinda funky to have the Official Adversary do all these lovey-dovy things. but i kinda like that. it might bring home the point that the adversarial role isn't bad but good, and it might also raise the stature of the tribune. at the least, they should be able to "address messages to the people and the Assembly".
i kinda like the idea of giving each of the three tribunes their own function --- the youngest one could be purely adversarial, the middle one could be ombudsperson, the oldest one could be head of state. the middle and oldest do not, however, lose their adversarial powers.
and by giving one of the tribunes the head of state role, we prevent it from falling on the PM (and we have a chance of having my rule that the legislature cannot bestow honors and commemorations stand!). and by giving one of them the ombudsperson role, we prevent it from falling on all of them.
note that, strictly speaking, the PM/FM's powers are just the nomination and dismissal of cabinet members (and to preside over cabinet meetings), except in times of emergency (and also their powers as legislator).
no laws or regulations shall apply to members acting in council that don't apply to them otherwise. they are not to be considered elgible for special scruity or intrusion like executive or legislative officials when in council or because they are members of a council. they are free to buy each other dinner, have romantic relationships with each other, etc (mb not free to bribe each other tho! but that falls under the no-bribing general rules, i.e. you can't trade a vote for anything but another vote)
thanks the obama in and his speechwriter in http://www.nytimes.com/2009/05/21/us/politics/21obama.text.html?_r=1 for the first two bullet pts in "Prohibited reasons for secrecy", taken from "we must not protect information merely because it reveals the violation of a law or embarrassment to the government."
"parpolity", a political counterpart to a communism economic system ("parecon") has a hier indirect rep structure: http://www.zmag.org/znet/viewArticle/4957
note: in 1936 the soviet union switched from an indirect hier election to a direct election. "Soviet Politics in Perspective" by Richard Sakwa on page 109 makes it sound as if the main drafter, Bukharin, thought this would better restrain the official terror which was the order of the day. on page 109 this book credits the problem with the lack of enforcement, rather than a bad constitution. however, on page 106, it seems to place the blame on (a) the lack of explicitly and precisely defined political rights (instead using vague terminology such as "the interests of the people"), (b) the principal that the interests of individuals are subordinate to the interests of society, (c) no court of appeal against infringements of individual rights, (d) a presumption that "politics" had been transcended (hence no need for contestatory political parties), (e) no separation of legislative and executive power, (f) little attempt to restrain the exercise of power, (g) the Supreme Soviet (the highest legislative body) did not have independent legislative power (?) and was de facto politically irrelevant, (h) no explicit limits on state power in the constitution, (i) politics through conventino, informal practices, rather than through explicitly, legally defined perogatives, (j) "above all, the Soviet political order (particularly in the Stalin period) is best understood in terms of personalities rather than institutions". he concludes "Power in the Soviet Union was not defined and therefore not limited, although informal restraints, as in Britain, increasingly came to the fore." he goes on to say that the constitution had a declarative, rather than a defining function, meaning basically that they were statements of what the party thought of things -- the first one set the tone -- "It summed up the experience of the first 8 months of Bolshevik power but left the details of government uncertain."
as for not separating legislative and executive functions, he claims that "the lack of checks and balances and the absence of separation of powers allowed the enormous accumulation of power at the hands of executive bodies".
lessons: most of the safeguards implied in the above are already included. in addition, the constitution should make it clear that individual citizens wronged by the government can take the government to court (i assumed this, but best to make it explicit). second, it should be made explicit that the ministers do not have any authority to issue laws/regulations/decrees. the legislature can delegate such powers only in certain states of emergency, and such powers can be undelegated at any time, and individual decisions can be overruled (i.e. unlike the idea that corporate shareholders can not freely undelegate authority from management on any issue -- here the legislature explicitly can). (this is also different from the present-day U.S. system, in which there are executive branch regulatory agencies that can make regulations -- but perhaps the creation of quasi-independent regulatory agencies should be allowed? that is, if they don't answer to/aren't influenced by the ministers, then some legislative power can be delegated, so you can have the FCC and SEC but they must be quasi-independent, like the Fed? i dunno...).
reading pages 90-92 of "Soviet Politics in Perspective" by Richard Sakwa makes it sound as if a combination of:
is to blame for everything. i.e. the problem was not the constitutional structure, but rather the prohibition on competitive politics that was the problem. as he puts it, democratic centralism "...guarantee[d] that the lines of communication remained vertical, from top to bottom, rather than horizontal, between like-minded individuals, groups or committees."
delegates were recallable. on 92, we see that the party even had separate "revisional committees" acting as watchdogs.
even China has a "convention that government committees contain at least one non-party member, a party membership is a definite aid in promotion and in being in crucial policy setting meetings." according to Wikipedia
1.
Since their representative is only directly responsible to them, and not to their constituents, the representatives of the Undersenate are expected to pass information between their representative above, and their own constituents below. They are expected to communicate with their direct contituents, to discuss issues with their peers, and to advise, monitor, and judge the decisions of their representative.
tying of hands: 2
One way that this Constitution might be attacked would be for an Amendment to be made during a Condition of Dire War, Emergency, or Martial Law; in this way, the normal procedures for the Electorate to deliberate might be bypassed3.
To protect against this, if the Constitution is amended during a Condition of Dire War, Emergency, or Martial Law, these amendments shall expire with the cancellation or expiration of such condition, and shall not be re-enacted for a period of at least one year. If they had an effect on the distribution of power (for instance, if they affected who is a Senator, who is Prime Minister, who is a Tribune, etc), this effect is reversed upon such expiration.
If the amendments are desirable, the Legislature is encouraged to re-enact them after the fallow period.
Upon the passage of a bill in one chamber of the Legislature, that bill shall be considered to be vetoed by the other chamber unless the other chamber has reached quorum at least twice during the three-month veto period. If a bill is vetoed in this way, nothing shall prevent the original chamber from sending it back to the other chamber the next time that the other chamber meets\footnote{This is to prevent the situation where an attack on the Senate or a major accident which prevents it from meeting provides an excuse for the House to pass bills without Senate review. Note however that as a side effect, it allows a minority of Senators to veto a bill by preventing quorum. Even if the Senate never again met, the House could still send the bill to a referendum, however, so this would only be a good idea if a majority of the people supported that minority of Senators; if the Senate did eventually meet, the House could simply re-send the bill then. It also prevents the same situation with the roles of the House and the Senate switched. Notice that this means that "quorum" must somehow be defined for the House.}.
4.
note: public membership roles in constituencies are needed in order to do cross-council assignments. but they present a major opportunity for corruption. however, this cannot really be avoided, because, unless deliberation is done exclusively online, people can see who else is in their constituency when they attend the meetings.
random note: if you ever need crypto paper ballots: http://rangevoting.org/RivSmiTB.html
The voting threshold for bills which repeal laws and which do not do any of the of the above is 50%5