In late 2009, we (Jan Behrens, Axel Kistner, Andreas Nitsche, and Björn Swierczek) started to develop LiquidFeedback, a software for proposition development and decision making. [PLF] One of our goals was to provide a feasible implementation of Liquid Democracy. Back then, “Liquid Democracy” was only a vague concept to us that we stumbled upon in the orbit of the Pirate party movement in Germany. [Huwald] This article shall show that most of the ideas had already been thought of many decades ago, even though crucial details were devised around the millenium change.
More than 130 years ago, Lewis Carroll (Charles Lutwidge Dodgson) described in his book “The Principles of Parliamentary Representation” the concept of giving candidates in an election a choice
It should be noted that the proceedings described in his book were specifically meant for elections of people (i.e. for representative democracy where members of a house of representatives are elected) and not for empowering the electorate to decide on certain issues themselves. Furthermore, the transfer of votes (according to his proposal) was carried out by candidates receiving those votes and not by voters.
His reasons to propose a transfer of votes by the candidates were to avoid a “waste of votes” in multi-winner elections where members of a house of representatives are being elected. Therefore, his proposal can be seen as alternative to “Single Transferrable Vote” (STV) systems, which had already been known at that time. Carroll justified his proposal by pointing out deficiencies found in the known methods to transfer votes using STV at that time, [Carroll, pp.30-32] and by claiming that preferential voting would be too complex for the “ignorant elector”:
Following the premise that neither excessive votes received by a candidate nor votes for those candidates who do not get elected should be lost, and arguing that it would not be feasible to let the voter decide who to transfer surplus and/or unused votes to, Carroll reasons that it must be “the candidate himself, for whom the votes have been given” who must have authority to dispose those votes. Guarding against skepticism, he explains that whoever trusts someone so far as to serve as a member of a house of representatives, he or she shouldn't have a problem to trust that person to transfer the vote in lieu thereof:
Obviously this argument could also serve in favor of transitive delegated voting, which was about to be invented more than 100 years later. [Note: There are more reasons to facilitate transitivity of delegations, as explained in [PLF, chapter 2].] However, Carroll's proposal didn't explicitly mention any transitive element, i.e. more than one step of transferring votes further. [Carroll, pp.36-40, pp.44-45]
While Lewis Carroll was probably the first person to propose empowering candidates to transfer their received votes, his proposal didn't contain any element to empower people to participate on particular issues themselves (which is arguably one of the most important aspects of Liquid Democracy, as described in the following).
About 83 years later, Gordon Tullock proposed a hybrid of direct and representative democracy in form of a proportional representation scheme where each voter may decide to either represent him- or herself or to choose a representative. Each representative would be given a voting weight equal to the number of his or her voters but a voter could also elect him- or herself as their own representative (in which case they would receive a voting weight of at least 1). Tullock suggested to use the possibility of watching proceedings through broadcast and voting “by wire” (see also [Armytage14]) to allow an unlimited number of voters to represent themselves if they wish so. Tullock noted that only the existence of “computers” and “modern electronics” could make this possible:
Tullock's proposal might have been the first step towards a democracy where each voter may decide him- or herself whether to engage in decisions directly or through representation. It is thus a hybrid of direct and representative democracy. However, his proposal still differs from Liquid Democracy in at least two important ways: neither delegation by topic (i.e. nominating different people for different subject areas or issues) [PLF, subsection 2.2.2] nor the possibility to instantly revoke transferred voting weight [PLF, subsection 2.2.4] is part of Tullock's proposal. Tullock, however, did not expect his ideas to be the “best possible suggestions” but rather saw them as a potential step for “a better and more scientific political structure” in the future:
In 1969, James C. Miller foresaw in his publication “A program for direct and proxy voting in the legislative process” that within the next 20 or 30 years, every household would have a “console tied into a computer”. According to his ideas, such a computer console could not only be used for children to do their homework, making out grocery lists, or paying bills, but also for making political decisions such that each voter could decide on every issue:
Just like Tullock, Miller proposed that voters may use computer technology to vote on every single issue themselves or to delegate their vote to a representative if they wish to. But Miller further suggested that voters could determine themselves how long such a delegation shall be in effect:
While not foreseeing the difficulty regarding verifiability of such an electronic system, Miller did have some ideas about potential security measures, which were quite futuristic for that time:
As of today, we know that measures such as “special metal keys”, coded combinations, or even retina scans will not be able to solve the problem of verfiable secret elections using computers (see also section 3.4 on the “Wahlcomputerproblem” in [PLF]). Nonetheless, (and disregarding the antiquated gender roles) Miller's proposal in that time is to be considered “visionary” (see also [Armytage14]).
As already said, the aspect of verifiability of electronic ballots is ignored by Miller at that time. Furthermore, his proposal (as well as Tullock's proposal) still differs in another way from what we call “Liquid Democracy” today: delegates who receive votes as proxies are not explicitly enabled to delegate their vote further. Such an extension isn't conflicting with Miller's proposal (and might even be considered part of the “stockholder voting schemes” he referenced) but to our knowledge hasn't been explicitly mentioned by him.
Miller, however, already proposed the ability to revoke a previously given delegation at any time. Ironically, the ability to instantly disempower (or empower) a representative was seen by Miller as a positive influence on the representative's freedom to vote his conscience because the dynamic representation scheme would allow representatives to “reverse their stand”:
Miller also considered the idea that voters might make their decision whether to vote directly or via proxy on a per-issue basis:
The ideas from the 1960's reappeared on the internet around 1995. [Note: We didn't find a proof for the exact date of publication. The article published at the URL http://robla.net/1996/steward/ (accessed 2016-04-19) written by Rob Lanphier has a copyright notice dated 1995 [Lanphier] and has been cached on archive.org since 2005. James-Green Armytage states 1995 as the year of Lanphier's publication. [Armytage14]] Rob Lanphier proposed the “Public Ballot Stewardship”: a model for electronic democracy. [Lanphier] [Armytage14] He distinguished two forms of elections: general elections (using a secret ballot) and public elections. Depending on the impact of what is voted on, either a “general election” would be held, or – for more “mundane” things – a public ballot would be held, where everyone knows which way everyone else voted:
In his proposal, the public elections provide a dynamic delegation system: Lanphier proposed that for each issue that is being decided in a public ballot, one may either decide to represent him- or herself or to choose a “steward” to vote on one's behalf. He furthermore notes two important aspects: one should be able to change the steward at any time, and for particular decisions one should be able to override the steward's vote by temporarily ceding the vote from the steward and voting directly. [Lanphier] Lanphier even proposed to allow different delegations in different subject areas:
Lanphier's proposal doesn't explicitly state that delegates may further delegate their votes (see [PLF, chapter 2] for transitive delegation). The cited “trusted advisors” could, however, be seen as a precursor to what would later be proposed in the 2000's by Bryan Ford and others.
There was one other important aspect which Lanphier addressed in his publication: the internet (as of 1995) didn't provide means to identify its users. Lanphier pointed out that it would be necessary to verify that one person gets exactly one vote (and not two, three, or more, by creating multiple accounts). He assumed, however, that this problem would be solved in the near future:
Similar considerations can also be found in our book “The Principles of LiquidFeedback”, [PLF, subsection 6.1.1] even though we were not aware of his publication at the time of writing our book. It should be noted that Lanphier's optimistic views in regard to emerging means of identity verification on the internet would fail to come true, at least for another 20 years.
In the early 2000's [Note: We didn't find a proof for the exact date of publication. The PDF published at http://www.brynosaurus.com/deleg/deleg.pdf carries a date of May 15, 2002 and has been cached on archive.org since 2005. A link to the document as well as Ford's pages on Delegative Voting have been cached on archive.org since 2004. See also [Armytage14], where Ford's ideas are dated 2002.], Bryan Ford proposed two ideas which he calls “Delegative Voting” [FordDV] and “Delegative Democracy” [FordDD] of which the former can be seen as further development of Lewis Carroll's ideas (a method to avoid wasted votes when electing individuals without preferential ballots) and the latter contains elements akin to Lanphier's proposal. However, instead of following Lanphier's idea to facilitate two different forms of ballots (secret and public) for different kinds of decisions, Ford demanded “privacy of the individuals” and a transitive delegation system at the same time for the same decisions. He named the following 6 basic principles to describe his form of “Delegative Democracy”:
As for the last point, Ford added the important property of transitivity to the delegation model which wasn't mentioned in the previous proposals of Tullock, Miller, and Lanphier. However, in regard to the Wahlcomputerproblem (see [PLF, chapter 3]), he makes a big step backward (when compared to Lanphier) because his proposal may not be applied to electronic systems without losing another imporant property of democractic processes, namely verifiability.
Following Ford's step backward, the idea of privacy of the individual and accountability of the delegate would later be described by Hardt and Lopes [Hardt&Lopes] as a possible solution regarding their “Golden Rule of Liquid Democracy” (which, we think, is a dangerous euphemism). [Editorial4]
Even though we do not consider private electronic ballots to be part of the concept of Liquid Democracy or to meet democratic standards at all [Editorial4], Bryan Ford may have been the first person to add transitive delegations as the last missing piece to describe what is called “Liquid Democracy” today. [Note: Later, other persons claimed to be the “inventors” of Liquid Democracy. However, we can not confirm their claims. [Deseriis] See also [QA010]. We would also like to note that most ideas were already formulated in 1967, 1969, and 1995 by Tullock, Miller, and Lanphier respectively.]
Ford furthermore described three possible “extensions” to the delegation model: “backup choices”, “split delegation”, and “restricted and transitive delegation”. Bryan Ford didn't provide any notes on implementation in his paper, neither for a system fulfilling those 6 basic principles listed above nor for these three extensions. Section 4 of his paper just reads, “Implementing Delegative Democracy : Under Construction”. [FordDD] As it could be shown later in [PLF, subsection 2.4.2] and [PD], two of his extensions would add certain unwanted properties to his system (negative voting weight and/or unequal treatment of the participants).
James Green-Armytage described in 2004 [Note: An early draft of his ideas is available on archive.org. [Armytage04]] and 2005 another system that is incorporating transitive delegations for decision making. [Armytage05] He called his system “Direct Democracy by Delegable Proxy”, which consists of the following two “fundamental elements”:
Opposed to the visionary views in 1967 and 1969, Armytage didn't see the internet as an application field yet:
It should be noted that electronic voting machines do not solve the problems of verifiable electronic voting either. [PLF, section 3.4] [BVerfG] [CCC] However, combining transitive delegations [PLF, chapter 2] with a preferential voting system on proposals [PLF, section 4.12] are some of the key elements found in our software LiquidFeedback, even though we were not aware of Armytage's website at the time of creation of our software. Armytage also proposed that delegations shall be “issue-specific” in order to allow for a delegation to experts in each field, and he explained the importance of transitivity:
In the year 2009, Jan Behrens, Axel Kistner, Andreas Nitsche, and Björn Swierczek presented “LiquidFeedback”, which doesn't only incorporate the ideas regarding Liquid Democracy as described above [Note: with the exception of secrecy as explained in chapter 3 of [PLF]], but also includes other features such as proportional minority representation and preferential voting which go beyond the ideas of Liquid Democracy. A more detailed description of LiquidFeedback can be found in the book, “The Principles of LiquidFeedback”. [PLF]
The origins of Liquid Democracy date back as far as 1884, [Carroll] though most core elements were not foreseen until 1969, when James C. Miller published “A program for direct and proxy voting in the legislative process”. [Miller] In the year 1995, the idea of vote delegation was re-thought in the context of the emerging use of the internet. [Lanphier] The element of transitivity was brought up by several people after the millenium change.
The term “Liquid Democracy” became popular in Germany when the Pirate party movement proposed to overcome the limitations of direct and representative democracy. [Huwald] Even though the Pirate party arguably never succeeded in establishing a sustainable application of Liquid Democracy, they helped the idea to gain attention by public media.
LiquidFeedback was created independently of any particular party as a feasible software solution to help political parties or other organizations to implement Liquid Democracy within their organization. Beside the concepts of Liquid Democracy, a particular set of rules of procedure was combined with the ideas of Liquid Democracy to allow users to engage in a collective decision-making process where all participants are treated equally. [PLF] With the publication of LiquidFeedback, it is possible to apply Liquid Democracy to real-world scenarios.