(Draft section 1 of "A medium of assent for the support of large scale
communicative action".) http://zelea.com/project/votorola/d/theory.xht

Communicative assent is the expression of agreement, particularly of
agreement that arises from discussion.  I describe a medium in which
communicative assent is formalized through voting.  The voting
mechanism is a delegate cascade that is open to recasting.  In a
delegate cascade, a "delegate" is any participant who both receive
votes (like a candidate), and casts a vote of her own (like a voter).
But when a delegate casts a vote, it carries with it those received.
And so on...  Passing from delegate to delegate, the votes flow
together and gather in volume - they cascade - like raindrops down the
branches of a tree.^[4]

    FIGURE 1.  Cascades in tree form.  The measure of assent for each
    participant is the quantity of votes currently received (circled
    number).  Vote flow is depicted by arrows, with volume shown for
    each.  The votes flow together until they pool at the bottom
    (red), where they are held by the leading candidates.  (The
    depiction is unrealistic.  Actual cascades would likely be
    bushier, with perhaps 5-20 voters per candidate on average.)


Every participant has a single vote.  She may either cast it or
withold it.  If she casts it, she may cast it for anyone.  There are
no pre-declared candidates.  All participants are eligible to receive
votes.  All non-participants are also eligible.  A non-participant who
receives a vote is thereby made a participant.

Votes are open to recasting.  If a voter changes her mind about a
candidate, she is free to withdraw her vote, or to recast it for
another candidate.  Polls are intended to remain open indefinitely,
year round, with the votes shifting as new information becomes
available to the voters.  There is never a final result that cannot be
undone by recasting.

Votes are public.  There are no "secret ballots".  For every vote that
is cast, the identities of the voter and the candidate are visible.
Anyone may trace the flow of a vote from its original caster, through
all intermediate delegates, to the final candidate who holds it.^[9]

Assent for each candidate is measured as the quantity of votes
received (circled numbers in figures 1 and 2).  Note that a single
vote may be received by multiple delegates before it is received by
the final candidate.  As each delegate or candidate receives the vote,
her measure of assent is incremented.  When the vote finally reaches
the candidate from whom it can flow no further, it is "held" by that
candidate.  The total of votes held by a candidate (red numbers) has
no bearing on the measure of assent.

The typical structure of a delegate cascade is a tree.  As shown in
figure 1 (above), it has a single candidate at the root; voters at the
leaves (top); and delegates among the branches in between.  The
general structure however is a cyclic graph, as shown here in figure

    FIGURE 2.  Cyclic cascade.  Depicts a cascade that has formed into
    a nearly perfect ring structure.  A single voter outside of the
    ring has injected a vote.  It carries nearly full circle until it
    comes to rest with a candidate who consequently holds two votes
    (bottom left).  Nevertheless the assent within the ring is equal,
    at 6 for each candidate.


A vote never actually cycles.  It flows through every candidate
exactly once, but stops before it encounters a candidate for a second
time.  It then remains held where it is.  Consequently assent is
equalized in a cycle.  A vote also stops before it encounters its
original caster.  Consequently a vote for oneself has no effect.

    FIGURE 3.  Tight cycles.  The tightest cycle is actually between
    two voters (left).  A cycle with a single voter (right) is a null
    cycle, equivalent to a withheld vote.


Assent is an expression of agreement.  Assent is formalized in the
medium by casting a vote.  The vote is cast for the person who best
represents the thing agreed to.  Of all the things that might be ageed
to, I wish to consider only acts.  For example, Juanita may propose to
build a sandbox for children in the neighbourhood park.  Samantha may
formally express her agreement to that act by voting for Juanita.

A person has a single vote for every act that could possibly be
proposed.  If Juanita also proposes that Rajiv should be appointed as
Park Superintendent, then Samantha could vote for Juanita on this act,
too.  The two votes - one to build the sandbox, and one to appoint
Rajiv - would have no formal connection to each other.  They would be
cast in separate "elections", so to speak.

Variant acts may be proposed.  Variant acts are acts that differ from
the originally proposed act.  When a variant act is proposed, the
participants do not gain another vote to cast.  Instead they gain a
choice of which act to cast their vote for.  They may cast it for the
originally proposed act, or for the variant, or for neither
(abstaining to vote).  Thus Monika may hear of Juanita's plan, and
join Samantha in voting for it.  She might, at the same time, suggest
that the sandbox ought to be larger than Juanita is proposing.
Samantha might shift her vote to Monika as a sign of approval for the
proposed variant, or she might keep it with Juanita.  She would have a
choice.  I refer to all of the choices (variants and original) as

A candidate always has two aspects: an active aspect, and a personal
aspect.  The active aspect is the proposed act.  The personal aspect
is the person who proposes, or represents, or embodies the act.  Thus
the building of the larger sandbox (act) and Monika (person) are one
and the same candidate.  We can speak of voting for a larger sandbox,
or of voting for Monika, but the two have the same meaning.  The
personal aspect of a candidate is formalized in the medium by an
identifier, so one always knows who the candidate is, for a given
vote.  But the active aspect is only formalized by an optional link.
Thus Monika may have a Web page that details the dimensions of her
planned sandbox.  She may formally link her candidacy to that page in
order that people can discover what she is proposing, without
personally asking her.  But this is optional.

Two candidates may or may not be mutually exclusive.  We know that
delegation may enable a single vote to be received by multiple
candidates, thereby expressing agreement to multiple, simultaneous
acts.  We expect those acts to be, in some sense, logically compatible
with each other.  Thus by continuing to vote for Juanita, even while
proposing a somewhat different sandbox, Monika may be expressing a
hope that Juanita will eventually agree to amend her original plan.
If Samantha were to shift her vote to Monika, in turn, she might be
hoping to further that amendment.

Assent may build as new participants join.  The votes are public, so
each newcomer can discover who is involved, and what they have agreed
to.  The newcomer can join in the discussion and cast a vote.  She can
propose new candidates and invite other participants to join.  And so
on.  An initial plan of action that was hatched among 2 or 3 people
might therefore grow a little larger.  In principle, there might be no
limit to its size.  Before examining the possibility of large scale
assent, I wish to consider a special case in which an upper limit is
definitely imposed.

    FIGURE 4.  Mass assent.  The same two cascades as in figure 1, but
    without any actual delegation.  Here each voter has recast for the
    end candidate who currently holds her vote.  The tree structures
    have collapsed to starbursts, typical of mass voting media.


Figure 4 illustrates the case in which the voters avoid delegation and
cast directly for end candidates.  Instead of a tree, the result is a
starburst pattern, typical of mass voting media.  There are many
voters, no delegates, and only a few candidates.  The limit imposed by
this pattern is not to the scale of assent itself, but to the scale of
discussion.  Because the number of candidates is restricted, and the
number of voters increases, there is a point at which discussion
between voters and candidates is no longer possible.  That point is
the boundary between communicative assent and mass assent.

As the scale of mass assent increases, newcomers will find it
increasingly difficult to discover exactly what act is being agreed
to.  Beyond a small core of initial voters, still in touch with the
candidates, the others will be in the dark.  Their participation will
not have enlightened them.  They will not be able to ask the
candidates what is at stake, because the candidates would be flooded
with questions if they were to intercommunication with the voters.
The only source of information will be other forms of mass media, all
inherently one-way channels.  With no discussion between the voters
and the candidates, the assent of the voters cannot be characterized
as communicative.^[12]

The newcomer would have less difficulty in joining if the assent were
mediated by a delegate cascade.  She could discover what was at stake
by asking any voter or delegate in the higher branches of the cascade.
She could go shopping among the delegates, using her vote as leverage
to enter discussions, and maybe to gain concessions.  She could make
suggestions.  She might even solicit votes herself and thus increase
her leverage, and her ability to move deeper in the cascade.  It would
not matter to her if a dozen participants were involved, or a
neighbourhood, or a whole city.  The discussion would be no less
lively for all its extent, and no less inviting to newcomers.


[TCA1] The Theory of Communicative Action.  Volume 1.^[1]

[TCA2] The Theory of Communicative Action.  Volume 2.^[2]

 [1] Jürgen Habermas.  1981, 1984.  The Theory of Communicative
     Action.  Volume 1.  Reason and Rationalization of Society.
     Translated by Thomas McCarthy.  Beacon Hill, Boston.

 [2] Jürgen Habermas.  1981, 1987.  The Theory of Communicative
     Action.  Volume 2.  Lifeworld and System: a Critique of
     Functionalist Reason.  Translated by Thomas McCarthy.  Beacon
     Hill, Boston.

 [3] The basic types of social action are:^[TCA1.85-86]

     * Teleological action - an actor attempting to reach a goal

         * Strategic action - teleological action in which success
           depends on the decisions of others who do not share the
           same goal

     * Normatively regulated action - an actor fulfilling expectations
       common to the group

     * Dramaturgical action - an actor self-presenting to an audience

     * Communicative action - two or more actors coordinating by
       discussion aimed at mutual understanding or agreement

 [4] I originally proposed a delegate cascade for the purpose of open
     voting on draft legislation.^[8]

     Other methods of recursive vote transfer have been described by
     Carroll,^[5] and by Rodriguez and Steinbock.^[7]  They differ
     however in allowing a vote to be alienated from its original
     caster.  Once transferred to a delegate, the vote cannot be
     withdrawn or shifted to express disagreement with the collective
     result that emerges.  Consequently the result cannot claim to
     have the assent of the voter.

     Another method is Brin and Page's PageRank algorithm for ranking
     Web pages.^[10]  It effectively considers a link to a page as a
     "vote" for that page.  The algorithm is recursive, allowing the
     vote to traverse the target page's own links.  PageRank differs
     in that a single page may link to (vote for) multiple pages.
     Votes consequently multiply or split apart.  It also differs in
     that votes diminish in strength (are "dampened") as they traverse
     the pages.

 [5] Black analyzes a voting mechanism proposed by Lewis Carroll that
     involves multiple levels of delegation.  Unlike a delegate
     cascade, however, the votes are alienated from the original
     casters.  The delegates treat them "as if they were their own
     private property".^[6]

     Lewis Carroll.  1884.  The principles of parliamentary
     representation.  Harrison and Sons.  London.

 [6] Duncan Black.  1969.  Lewis Carroll and the theory of games.  The
     American Economic Review.  59(2), p. 210.

 [7] Rodriguez and Steinbock et al.  describe a system of 'dynamically
     distributed democracy' that involves recursive delegation.
     Unlike a delegate cascade, however, the votes are alienated from
     the original casters.  The intent is to improve the efficiency of
     the process by removing voters from direct discussion and
     agreement on a final decision.  The system might therefore be
     classified as a steering medium that functions as a substitute
     for communicative action.  Its design and purpose are therefore
     different from the medium of communicative assent that is
     proposed here.

     Marko Antonio Rodriguez, Daniel Joshua Steinbock.  2004.  A
     social network for societal-scale decision-making systems.
     NAACSOS '04.  Proceedings of the North American Association for
     Computational Social and Organizational Science Conference.
     Pittsburgh, Pennsylvania.

     Marko A.  Rodriguez, Daniel J.  Steinbock.  2006.  The anatomy of
     a large scale collective decision making system.  Los Alamos
     National Laboratory Technical Report.  LA-UR-06-2139.

     Marko A.  Rodriguez, Daniel J.  Steinbock, Jennifer H.  Watkins,
     Carlos Gershenson, Johan Bollen, Victor Grey, Brad deGraf.  2007.
     Smartocracy: social networks for collective decision making.
     40th Annual Hawaii International Conference on System Sciences

 [8] I sketch the concept of an "open legislature" in which variant
     bills are drafted in a medium of recombinant text while being
     simultaneously exposed to delegate-cascade voting.

     Michael Allan.  2007.  Recombinant text.  SourceForge.net,
     project textbender, release 0.2.2, file d/overview.xht.

     See also the latest version online.

 [9] A secret ballot is a defence against vote buying because it
     prevents the buyer from verifying compliance.  The voter may take
     the money, then vote as she pleases.  This makes it a poor

     It will also be a poor investment when the vote is recastable, as in
     a delegate cascade.  The vote might be public and compliance might be
     verified, but there is no guarantee of continued compliance.  The
     voter may take the money from one side, then shift her vote and take
     it from the other.

[10] Sergey Brin and Lawrence Page.  1998.  The anatomy of a
     large-scale hypertextual Web search engine.  Computer Networks
     and ISDN Systems.  30, p. 107-117.

[11] Jürgen Habermas.  1962, 1989.  The Structural Transformation of
     the Public Sphere.  Translated by Thomas Burger.  MIT Press,
     Cambridge, Massachusetts.

[12] To emphasize the contrast between the communicative and the mass,
     consider C. W. Mills's definitions of "public opinion" and "mass

    "In a *public*, as we may understand the term, (1) virtually as
     many people express opinions as receive them.  (2) Public
     commununications are so organized that there is a chance
     immediately and effectively to answer back any opinion expressed
     in public.  Opinion formed by such discussion (3) readily finds
     an outlet in effective action, even against - if necessary - the
     prevailing system of authority.  And (4) authoritative
     institutions do not penetrate the public, which is thus more or
     less autonomous in its operation.

     In a *mass*, (1) far fewer people express opinions than receive
     them; for the community of publics becomes an abstract collection
     of individuals who receive impressions from the mass media.  (2)
     The communications that prevail are so organized that it is
     difficult or impossible for the individual to answer back
     immediately or with any effect.  (3) The realization of opinion
     in action is controlled by authorities who organize and control
     the channels of such action.  (4) The mass has no autonomy from
     institutions; on the contrary, agents of authorized institutions
     penetrate this mass, reducing any autonomy it may have in the
     formation of opinion by discussion."

     C. W. Mills.  1956.  The Power Elite.  New York.  p. 303-304.  As
     quoted in Habermas.^[11]



  - an object of assent

  - one who receives a vote

communicative action

  - a type of social action that is coordinated by discussion aimed at
    mutual understanding or agreement

communicative assent

  - expression of agreement, particularly of agreement that arises
    from discussion


  - a subject and object of assent

  - one who is both a candidate and a voter, who both receives and
    casts votes

delegate cascade

  - a voting mechanism in which received votes are carried along with
    cast votes

mass assent

  - expression of agreement that arises through mass media


  - a formal unit of assent


  - a subject of assent

  - one who votes for a candidate

Copyright 2008, Michael Allan.  Permission is hereby granted, free of
charge, to any person obtaining a copy of this software and associated
documentation files (the "Votorola Software"), to deal in the Votorola
Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicence, and/or
sell copies of the Votorola Software, and to permit persons to whom
the Votorola Software is furnished to do so, subject to the following
conditions: The preceding copyright notice and this permission notice
shall be included in all copies or substantial portions of the
Votorola Software.

Michael Allan

Toronto, 647-436-4521

Election-Methods mailing list - see http://electorama.com/em for list info

Reply via email to