A democratic system in which most issues are decided by direct referendum. However, since no one has time for this, you can delegate your votes. Here’s the cool part; you can delegate your votes on a certain topic to one person, and then delgate your votes on another topic to someone else. And delegations are transitive; you can delegate to someone who delegates to someone else, etc, in which case your votes will flow to whoever is at the end of the line. Of course, you can “un-delegate” at any time.

So, say you don’t know much about the space program – you give your votes on things relating to the space program to someone who has similar political views to you but who knows more about the space program (and they can pass the vote on if they choose).

There are some example scenarios in the middle of this page.


Problems to be solved:

Other notes:

See also

Liquid Democracy

Because “liquid democracy” seems like it would work well for WikiVoting, we’ve decided to use (one version of) it in our CommunityWikiBylaws.

Alternatives to Liquid Democracy

Page translations

[fr] DémocratieLiquide

(comment on EfnetPythonWiki:LiquidDemocracy: it seems like the idea is basically that instead of getting apathetic about an issue, you delegate your votes on certain topics to friends who care more. Sounds like an interesting idea. This is also interesting to me because of structural similarities to some ideas related to PeerToPeerJournalism. --BayleShanks)


(Some people were using “vote” to indicate a single ballot given to a single person. Others were using “vote” to indicate the entire process of deciding a single question, involving thousands of ballots. I changed their words; I hope that’s OK).


DavidCary, you added: (On this page, “vote” indicates a single person’s power to influence the outcome of a question, traditionally exercised using a single ballot. “question” is a single one-time issue, to be decided by a tally of votes. “issue”, “topic” is a group of related questions. Are there better terms we should be using ?)

I don’t feel that’s correct- the vote doesn’t have to address a single one-time issue. Such as the case of a “live” vote. For example, in HiveAllocation, the vote on who gets funds is continual. If at any point you want to change the way funds are meted out, you can change your vote. So it’s not so much a question.

So, I’ve never seen this page before, but it has an uncanny relationship to my own ideas about DirectDemocracy.

Anywho, I don’t see a reason not to sell votes. If I can get $5 for my vote on, say, soybean trade policy, well, why should anyone stop me? It’s a lot more honest to buy my vote directly than to buy off my representative.

Otherwise, cool page.

You want to avoid selling votes because otherwise monied interests have almost total control. There are more apathetic people who like/need money than there are people who care about almost any given issue. The result would be (I predict) that on any given issue, the side with the most money easily wins. At which point there’s not much point having a democracy; the government is essentially autioning decision-making power.

Another reason is the general principal AvoidConcentrationsOfPower.

Abd ul-Rahman Lomax: Liquid democracy is pretty close to the concepts I’ve been working on for about twenty years. Those concepts have boiled down to two: Free Associations and Delegable Proxy.

Delegable Proxy is essentially only a slight extension of the basic idea of the proxy: if you trust someone to handle a certain business for you, presumably you can also trust that person to wisely select someone else to handle it in their absence. I’m only aware of some very limited efforts, however, to implement delegable proxy in any organization. Given that lack of experience, it seems premature to me to try to decide issues such as whether or not it would be legal to buy and sell votes. It would depend entirely on the nature of the organization itself. But it all gets clearer when the concept of the Free Association is combined with delegable proxy.

A Free Association is actually the basic default association of human beings, in the absence of coercion. There is more experience with organizations resembling Free Associations than there is with delegable proxy; an example would be Alcoholics Anonymous. Free Associations are “controlled” from the bottom; leaders “are but trusted servants,” Free Associations don’t collect power except for the power that flows naturally from communication. Actually, since no one is controlled at all in a Free Association, it’s a little off to even think of how the Association is controlled. If Free Associations don’t collect power, how can they have an effect? showed a bit of how. is not a Free Association, it is an oligarchical structure, albeit one which is at least somewhat responsive to its members. was able to exert significant political power by merely suggesting to the members that they send donations to this or that cause or campaign fund. It is entirely possible that a Free Association, where it does not find consensus, to act in contrary directions simultaneously. However, because a Free Association with delegable proxy should (theoretically) be as efficient a discussion and decision-making system as is humanly possible (in large organizations), Free Associations seem more likely to develop consensus solutions and thus could exert the full force of the combined resources of their members.

The idea that big money has influence is based entirely on experience with situations where the public is not organized. The public, in fact, has more money collectively than any special interest. Isn’t that a tautology? Much more money. Consider the users of Microsoft software, and compare them with Bill Gates. The only reason Bill Gates has power is that he, in his personal property, represents a collection of a relatively modest amount from a large number of customers. Consider the disposable income of Bill Gates, and compare it with the disposable income of the Microsoft user community. The latter, collectively, has far more disposable income; but it is not organized, so that power is not available.

In a large Free Association, attempting to affect association decisions or recommendations by buying votes would be pretty silly. First of all, votes in such an association are merely a means of gauging consensus. To buy base-level votes would not gain the buyer much power; in order to truly affect decisions, you’d have to buy close to a majority of votes; really what you would be buying is proxy delegations. Suppose there is a Free Association PAC. Its purpose is to gather money for a political cause. It would probably be far more efficient to simply donate the money to the cause directly, than to try to buy votes. Bought votes would simply mean that someone would be arguing for the position of the buyer at a higher level in the organization. Yet in a Free Association with delegable proxy, any idea with a prayer of being widely accepted can be represented at a high level without buying any votes.

The delegable proxy structure actually forms a bidirectional nervous system, information flows in both directions, and is filtered in both directions. The proxy protects the higher levels from information overload from the bottom, and protects the represented person from information overload from others, including high-level discussions the proxy considers will not interest the delegator.

Bottom line: first step in transforming society is to apply the principles of Free Associations and Delegable Proxy or Liquid Democracy in actual organizations. Successful implementations, if the idea is good, will find imitation. Ultimately, as I see it, FA/DP organizations will be so efficient and so effective that they will co-opt the political system. There is no need whatever to fight against existing systems, as long as they are within basic democratic norms; and even outside that, FA/DP could accomplish a lot. FA/DP should be able to bypass existing systems. Obviously, if you could find a way to develop a consensus on a decision (or even simply a majority position), in a democracy, that decision could be implemented. And no matter how good the idea, if it is implemented merely because a few people with inequitable power go for it, the implementation is oppressive. Far too many reformers have failed to notice this.

See for some early material on this topic; there is also a wiki linked from that site, and, from the wiki, a number of subsidiary wikis established in an attempt to apply FA/DP concepts to real-world situations. None of these efforts have attracted more than a certain level of interest; participation levels are as yet too low to begin actual testing of the concept. --Abd (danielATbeyondpoliticsDOTorg.)

secret ballots: It makes things simpler if each person’s delegation is public. (Public information makes it simpler to discover the truth when someone fraudulently claims he has been delegated lots of voting power). But if there were good reasons for a particular election to use secret ballots, wouldn’t those same reasons also be good reasons for each person’s delegation to be secret?

Some people are paranoid about putting lots of power into the hands of one person. But they don’t mind putting huge amounts of power into the hands of 2 people (I’m thinking of Crimson Tide). Here is a variation that just occurred to me – is this more complicated than it is worth?

  • Everyone starts with 2 votes.
  • You can either cast your 2 votes (plus any others you have collected) directly, delegate both votes to 1 other person, or (here’s the new bit) delegate 1 vote each to 2 other people. (So if Charlie delegates his votes to Alice and Bob, Alice and Bob have now collected 3 votes each).
  • If a delegate re-delegates her votes to 2 people, then half the votes she has collected (rounded up) go to the first, and the rest go to the other. (If Alice delegates the 3 votes she has collected to Pat and Tom, then Pat gets 2 of them, Tom gets 1).

(Is there a better wiki than CommunityWiki for discussing strange new voting systems, and proposing improvements? The many hits in the search Wiki:search=voting seems to imply that the original Wiki is a good place for discussing voting theory…)

I wonder if you could get something interesting by turning that into approval voting: you can “delegate” your vote to any number of people, basically just saying that you trust them. It’d give different but interesting mechanics.

Bouncing around ideas about voting is pretty interesting (more so than actually researching what has already been said and experimented with 😝). I guess what’s really needed is experience with those things. Here would be a good place to start, even though it would probably be too small (even with the legions of zombies).

Yea! Someone to stick up for the rights of we legions of zombies! You have my vote!

No, you maroon! We were supposed to delegate our votes to DavidCary!

:-) Is there some other wiki where informally “bouncing around ideas about voting” is the main topic?

I’ve been pondering one particular voting question for some time. Let’s see if I can ask it without rambling on too long:

I know one situation where Wiki:ApprovalVoting is optimum: A school club wants to decide when to meet. And our goal is to get as many people at meetings as possible.

So, we make up a little ballot with every possible meeting time, everyone punches out every possible time that he could meet, (“approves” each of those times), and we stack them all up. If we’re lucky, we can see all the way through the stack, at a time when everyone meet. (if there is more than one, they all tie – see below). Otherwise, we labouriously count up how many votes each time got, and pick the one with the most. The time with the most votes is, by definition, the time when the most people could meet – which was our whole goal, to get as many people to meetings as possible.

If there is a tie between 2 or more times that got the most (or all) votes, then you can arbitrarily pick between them – coin toss, runoff, whatever.

No other method of voting can work better (get more people to meetings) – either that other method picks one of the ties (and gets the same number of people), or it picks some other time (and fewer people can come to the meeting).

Later we thought about having 2 different meetings, so people who couldn’t meet at the “main” time could meet at some other time. If the “main” time is already fixed, then the best “alternate” time would be to pick the people who couldn’t meet at the “main” time (pick out all the cards that don’t have that time punched), and do another round of approval voting.

But I keep wondering – what if the “main” time wasn’t fixed yet? Some other pair of times could be better. Could we pick a pair of times with a few less people at the “main” meeting, but many more at the “alternate” meeting, ideally (if at all possible) a pair that would allow everyone to see at least one of the 2 meetings? I think all the information is there on the approval ballots to do that. Surely there’s a better way itemizing every possible pair of times, counting each person (that can go at least once) only once, even if he can go both times.

An interesting discussion on Google+ where I summarized some of the points made at the top of this page. Later, I wrote the following:

Liquid Democracy would definitely be harder to grasp when thinking about interests. Perhaps a way forward would be: can we think of an interim solution that gets us some net benefit? The Swiss system is the only one I’m familiar with but I hear that in the US on a state level, there is a similar level of DirectDemocracy as it is called over here. I’m going to argue that it provides most of the benefits of Liquid Democracy and can be scaled up or down as needed, making Liquid Democracy redundant.

The citizens in Switzerland basically have the following options: a process to change the constitution (the initiative), a process to prevent the passing of a law (the referendum); but no process to add a law. The latter is an interesting and contested limitation because we also have no constitutional court, so an initiative can add an article to the constitution to do X, but then parliament can stall on passing a law to implement X and there’s nothing the citizens can do about it. So, given that we have the usual shenanigans involving new laws, lobbyists, convoluted language, packages of amendments, deals made between parties and all that, where interests are being negotiated, we still have some days during the year where we can vote on issues. The point is that we cannot delegate our votes to experts. All we can do is read the documentation available to us, and get informed ourselves, or just vote the party line if that’s who we’ll trust.

To me, that seems as if Liquid Democracy isn’t really necessary. Assuming we implemented Liquid Democracy for a few issues every year only, then what would the difference to the Swiss system be? We would nominate delegates for particular issues, and let them vote instead of reading what some experts we trust have to say on the issues at hand and voting as they say… The difference is small, I’d say.

Proponents of Liquid Democracy will say that voter participation will go up, but I’ll say that voter participation is already up when voting on issues instead of general elections of parliament (sadly). So, we get the claimed benefits of LD using a system that’s already in place without having to propose a radical new process.

CategoryElectronicDemocracy? CategoryGroupDecisionMakingRule?


Define external redirect: CategoryElectronicDemocracy GerryMandering RepresentativeDemocracy CategoryGroupDecisionMakingRule

EditNearLinks: PeerToPeerJournalism NetConferencePlus ConsensusPolling BenevolentDictator


The same page elsewhere: