None of the Above:
Creating Mass Deliberation Without Discussion


Ray Pingree
Graduate student
School of Journalism and Mass Communication
University of Wisconsin-Madison
[log in to unmask]
  (608) 244-3336
210 Buell St #2
Madison WI 53704

Nne of the Above:
Creating Mass Deliberation Without Discussion


Abstract

Deliberative democracy has been plagued by questions of implementation, due
to a failure to distinguish between discussion and the more general concept
of many-to-many communication.  To demonstrate that this theoretical
distinction is both possible and important, this paper introduces an
example of an Internet-based many-to-many communication system designed to
achieve deliberation's outcomes without discussion.  A broader deliberative
theory is proposed, to encompass the concept of non-conversational
deliberation as part of a more attainable public sphere.

 "[Online forums are] a development of historic significance, for there has
been practically no innovation in many-to-many communication in over two
thousand years"
- Hans Klein (1999: 213)

"The main problem [for deliberative democracy] is to explain how it is even
possible to have a 'discussion' among thousands or millions of people."
- James Fearson (1998:64)


None of the Above:
Creating Mass Deliberation Without Discussion



Deliberation and discussion

The promise of deliberative democracy is in its ideal outcomes, but its
most compelling criticisms highlight the unreliability of discussion for
achieving those outcomes, especially in large, diverse groups or an entire
polity (Witschge 2002, Shapiro 1999, Bell 1999, Hardin 1999, Sanders
1997).  Similarly, both optimists and pessimists about the deliberative
potential of the Internet have focused on the strengths and weaknesses of
online discussion (Papacharissi 2002, Witschge 2002, Dahlberg 2001, Fishkin
2000, Gastil 2000, Davis 1999, Klein 1999).  Underlying much of the
disagreement about the deliberative potential of the Internet is a
disagreement about the deliberative potential of discussion itself.

By discussion I mean any group communication structured as a series of
conversational turns.  Because of the Internet, discussion is no longer the
only way to structure group communication, although all existing forms of
online group communication are discussion-based (including email, Usenet,
chat, message boards, and blogs).  Entirely new forms of online group
communication can be designed that directly address the goals of
deliberation.  To persuade the reader on this theoretical point requires a
concrete example, and since no such example already exists, this paper will
introduce a new software system designed to be such an example.

In deliberative democratic theory, "deliberative" does not refer to just
any discussion.  Different theorists list different conditions that a
discussion has to satisfy to be called deliberative (Witschge 2002, Elster
1998, Fearson 1998).  What is important for our purposes is not these
conditions themselves, but the fact that arguments over their
appropriateness are nearly always based on their capacity to encourage
smarter or fairer collective decision-making.  This points to a more
general, although tautological, definition of deliberative: a group
communication process is deliberative to the extent that it transforms (not
merely aggregates) individuals in such a way as to make society, as a
whole, behave better in terms of intelligence and justice.

To create a broader deliberative theory, one that is not biased towards
discussion, we must work backwards from ideal outcomes of communication to
discover ideal communication conditions.  The first backwards step is
therefore a clearer definition of the goals of political communication than
"an intelligent and just society."  To that end, I suggest two intermediate
outcomes that political communication processes should aim for: collective
reasoning and collective consciousness.


Collective reasoning

Collective reasoning, at its heart, is the sharing of statements and
considerations that bear on them.  We borrow the spirit of Zaller's concept
of a consideration: basically anything that bears on a decision to agree or
disagree with a statement (1992:40-41).  Zaller's concept, however, was
intended for how public opinion is usually measured: fixed survey questions
(1992:4), and thus quite rightly does not take into account a third
reaction that people often have when they encounter an opinion-statement:
the desire to re-frame or re-phrase it, instead of merely agreeing or
disagreeing.

Note that as group size grows, discussion must tend more towards Zaller's
definition of considerations.  Nonverbal cues can express agreement or
disagreement, but one must take up a conversational turn to re-frame.  With
a bigger group, each individual has a smaller fair share of speaking time,
so each individual gets fewer opportunities to re-frame as group size
grows.  In the extreme case of a national "discussion," mass opinion can
only be agreement or disagreement with elite-framed opinion (which includes
both opinion polling and voting).  These are limits that follow directly
from the structure of conversation, and should not be presupposed to be
limits of group communication in general.  For this reason, our definition
of consideration must include not only reasons to agree and reasons to
disagree, but also alternative frames.  Considerations then bear not only
on the decision of whether to agree or disagree, but also on decisions such
as whether to even answer the question as asked, and what alternative
questions to propose.

Reasoning as exploration

Each consideration is potentially a whole new opinion-statement, which may
in turn have its own long list of possible considerations.  Thus, reasoning
(alone or with others) can be thought of as exploring a network.  At any
point in this exploration, we may not be aware of all the possible pathways
(or considerations) we could take.  As Zaller argues, the sets of
considerations that occur to people when asked to privately evaluate an
opinion statement tend to be unstable over time and vulnerable to salience
effects like question order, question wording, and media priming
(1992:40-96).  These limits of private reason constitute a strong argument
for deliberation: by pooling considerations, we can form better
opinions.  Instead of wandering this vast network of opinions and
considerations alone, we attempt to bring a group along with us, so that at
every step we have the benefit of others' vision (and experience) about
which way to go.

In this exploration, discussion requires a group to generally "stick
together" in order to benefit from each other.  This is why conversational
norms, civic practices (Eliasoph 1996), and discussion rules are crucial to
conversational deliberation.  This is also why discussion doesn't
scale.  With the Internet, we can design software that for the first time
allows explorers to benefit from anyone who has ever been "in the same
place" before, while allowing them to individually "go wherever they want"
at any time.  In other words, by structuring a system as a network of
opinion statements connected by considerations, instead of as a series of
conversational turns, we can allow people viewing a particular opinion
statement to benefit from the reasoning of anyone who has ever viewed that
statement in the past, while allowing each individual to view or write
whatever opinions or considerations they want to, at any time.  If we don't
get along well with each other in such a system, it does not have to impede
our ability to benefit from each other's reasoning.

Note that existing message boards and blogs are asynchronous and have
links, but these links are still generally to and from conversational
turns.  Users can "go where they want" independently of each other, but in
a conversation network, not a consideration network.  In other words,
instead of seeing considerations that prior visitors reading the same thing
have thought were relevant to deciding what to make of the statement, users
see a history of what other statements have been made in response.  Reasons
to agree or disagree with a statement may be buried deep in the many
replies and sub-replies to the statement, making it difficult to weigh them
against each other.  Furthermore, such systems do nothing to help people
see which reasons have been seen as compelling by others.

Collective reasoning creates the public

A key difference between collective reasoning and private reasoning
concerns the public-spiritedness of the reasons themselves, via what Elster
calls "the civilizing force of hypocrisy" (1998:12).  As dissatisfied as I
am by rational choice theory in general, it's hard to deny that some people
sometimes privately reason solely in self-interest.  However, when
addressing what is perceived to be a diverse audience, even a selfish actor
has an incentive (in order to be persuasive) to express public-spirited
reasons (reasons based on some concept of the common good).  Through
collective reasoning, we "form in common a common will" (Elster
1998:2).  Public-spirited reasoning, even if it does not broaden consensus,
forces us to create some version of a public spirit in each of our
individual heads, resulting in a real difference in collective
understanding and political coherence.  This is what Eliasoph refers to as
"the power to create the public itself" (1996:263).

The question of whether or not people are rational actors is analogous to
the question of what peoples' "true opinions" are.  In other words,
Zaller's model stressing the importance of the salience of considerations
(1992) applies to selfishness too.  If, on balance, selfish considerations
are more salient than public-spirited ones at the moment, an individual is
more likely to make a self-interested decision (or come to a
self-interested political opinion).  This would suggest that even if people
currently seem to behave mostly self-interestedly, this may not be so much
an indicator of human nature, but simply a lack of salience (or even lack
of knowledge) of public-spirited considerations, due to living in a nearly
deliberation-free society.  Therefore, there is some hope that if we create
a public, public-spiritedness (and thus further re-creation of the public)
will be easier.


Collective consciousness
"People need an organized map of the political world, not just a huge pile
of unsorted facts" (Eliasoph 1998:152).

By collective consciousness, I mean knowledge of the structure of opinions
in society.  Through communication we can learn this structure.  This
includes not only how many people agree or disagree (and how passionately)
with each opinion statement or consideration, but also which sets of
opinions tend to go together in opinion-groups (e.g. the left, the right),
which opinions bridge those opinion-groups, and which opinions are marginal.

In other words, we can form, through communication, a mental map of opinion
space.  However, in discussion, it may also go the other way.  The mental
map may affect our conversational turns and our interpretations of the
turns of others.  Through the opinion-structural equivalent of
Noelle-Neumann's Spiral of Silence (1984), we may reproduce our own
assumptions about what "doesn't fit" into assumed political categories, and
what would thus be difficult to explain in a limited conversational turn.
What we say, how we say it, and how we hear what is said can all depend on
(and determine) "where" we perceive other discussants to be in our mental
map.

For example, the common assumption that opinion space consists of a left, a
right, and a center may lead people to converse more as if they are
addressing and representing those categories than they would otherwise,
especially as group size grows.  People would do this to save time, because
of the aforementioned limits of the conversational turn.  Even if nobody
"misrepresents" any of "their own" reasons or opinions, in conversation one
is forced to choose which of many possible opinions and considerations to
express in each conversational turn.  Again, the key point is that this can
socially reproduce and reinforce that mental map of left, right, and center
(or whatever the current map is), regardless of what would happen to the
map given an open exchange.

The effects on a democracy of having an incorrect yet persistent shared
mental map would be enormous.  Nearly everyone would feel like they
personally "didn't fit" into politics, so participation would be low, as
would trust in established political parties and government.  Attempts at
collective reasoning would be unsatisfying and might become either taboo or
polarized to the point where they resemble a sporting event, only with less
sportsmanship.


Interactions of collective reasoning and collective consciousness

The combination of collective reasoning and collective consciousness is
critical.  A society composed of two groups that each misunderstands the
other is clearly worse off than one that is equally polarized but where
both sides know the reasons why.  Thus, the sharing of considerations is
not just to help people decide where they stand, but to help people to
understand why other groups of people stand where they do.

Deliberation should not be expected to lead to consensus, because of deep
differences that can't be resolved by reasoning, and thus will still exist
between "reasonable" people.  Such differences include values, vested
interests, tastes, and religious beliefs.  Cohen calls this "the fact of
reasonable pluralism" (1997:408), and argues that it is here to stay (1993,
1997, 1998).  But even without consensus, deliberation should be expected
to lead to increased understanding of such deeper reasons for different
opinions, resulting in increased political tolerance and higher chances of
finding good compromises between truly meaningful groups (in other words,
groups formed out of a consciousness of deep reasons), based on a mutual
understanding of the reasons for their differences.  Thus, deliberation
must combine collective reasoning and collective consciousness to lead to
political tolerance.  This may explain part of the unreliability of effects
of conversational deliberation on tolerance (for example, compare Denver et
al. 1995 to Fishkin et al. 1999).

Both collective reasoning and collective consciousness make use of spatial
metaphor.  It's important to note that they refer to two different, but
related "spaces," or more precisely, networks.  In collective reasoning,
people explore together a semantic network.  It is made up of opinion
statements connected by considerations.  Collective consciousness is
awareness of a different space.  It is ultimately based on assumptions of
statistical relationships between pairs of opinion statements (e.g. people
who agree with A also tend to agree with B, or people who agree with C tend
to disagree with D).  Our mental maps may take various forms (category
systems, dimensions, or more likely some fuzzy combination of the two), but
I argue that the underlying data on which we would ideally base those
mental maps is actually the statistical relationships between pairs of
individual opinion statements.


The Meaning Map Project

Pessimists about the deliberative potential of the Internet have often
focused on what Internet software currently does or how it is currently
used (e.g. Davis 1999).  Imagine doing the same thing in the early days of
electricity.  Asking what effects electricity will have might have seemed
to make sense when the only thing electricity could do was light a light
bulb.  Today we know electricity can be used for many different things, so
to have predicted the effects of all devices that have used electricity
would have required some technological vision about what those devices
might be.  The same is true of the Internet.  To evaluate the Internet's
potential requires technological vision about what kinds of software could
be developed using the Internet.  It is not sufficient to analyze the
software that currently uses it, the ways that that software is currently
used, or the gratifications sought by its users.  To assess the potential
of the Internet to achieve specific goals, we must attempt to design
Internet software specifically for those goals, and then attempt to test
whether those goals are achieved in actual use.

I have tried to do exactly that for the goals of facilitating collective
reasoning and collective consciousness in large groups.  The resulting
software is currently in alpha testing, and will soon be available for
public use and experimentation at www.meaningmap.com.  The same website
will also soon have more detailed descriptions of the system.  I give only
a brief overview of the system and some of its limitations here, in order
to demonstrate the feasibility of non-conversational deliberation and to
inspire other social scientists to try their hands at software design.

Overview

The Meaning Map is an online open polling system, where users explore
opinion statements either via a visual map of the network of statistical
relationships between opinions, or by traversing individual
consideration-links in a collaboratively created semantic opinion
network.  Unlike an opinion poll with fixed questions, anyone can at any
time post a new opinion statement, which other users can then cast agree or
disagree votes on.  The patterns of these votes determine the statistical
opinion network.  Instead of being forced to agree or disagree with each
opinion, users can also link it to another opinion that they feel is a
better way of framing the issue.  These "better frame" links appear as
considerations in a list visible to all users viewing that opinion in the
future.  In this same list are reasons to agree and reasons to
disagree.  Users can vote on the quality of each consideration, determining
the order of display of the consideration list (but not affecting the
opinion's position in the map).

Users are free to navigate on their own, and to post new opinion statements
or considerations regardless of what any other user is doing.  When viewing
any opinion statement, they have the benefit of a shared and rank-ordered
list of the considerations added by other users who have viewed that
statement in the past.  Reasoning together in this way makes getting along
with others far less important and ensures that the results of the
collective reasoning effort are preserved in a meaningful structure,
instead of being scattered in a disorganized history of replies.

Since statistical relationships in the patterns of votes on opinion
statements are visually displayed in a map for all to see, users of this
system should have a more accurate and up-to-date form of collective
consciousness than one can glean from anecdotal evidence of the known
opinion-sets of discussion partners.

Visualizing the statistical network

The technical challenge here is how to communicate (through both display
and interaction) to ordinary people that opinion "space" is best thought of
as a network of pair-wise relationships between individual opinions.  It
should not be presupposed that it is based on a handful of latent factors
(dimensions) or latent classes (categories).  Such categories or factors
may be deduced from exploration of the network, but a network structure
cannot be deduced from exploration of categorized or factored data.

In network visualization, the goal is usually to create a still image where
the distances between pairs of points in the image correspond as closely as
possible to their network distances.  In our system, we do not need to
limit ourselves to a still image.  Through animation and user interaction,
the process by which points find their positions (repulsion from negatively
related opinions and attraction towards positively related opinions) is
revealed to the user.   Users can drag and drop individual points around
the screen, allowing them to see other places where an opinion might come
to rest in the display (and see indirect effects on the rest of the space
due to that change).  Also, users can temporarily show only a subset of
opinions, which may result in a dramatic re-ordering of their positions
when outside influences are removed.  This allows the user to see the
internal structure of some meaningful set without reference to other
opinions (for example, seeing that socialism is central to the left when
viewed on its own, but peripheral in a broader view due to being strongly
repelled by the right).

Past evidence of public opinion being dominated by a single factor may be
moot due to the aforementioned problems with measuring public opinion via
private and pre-fixed questions, without adequate collective reasoning or
collective consciousness.  Analysis of change over time as a group uses the
system may shed light on this question.

Filtering and statistical zoom
"Any form of filter imposes its own biases.  But the absence of any filter
also has its own bias.  It causes public opinion expression to break down
into a babel of voices, with only the loudest achieving some level of
recognition" (Davis 1999:166).

Not all opinions are displayed in the opinion space at once, unless the
topic is very new.  This is not so much due to computational constraints as
to the fact that unless the network structure is "simple," beyond a certain
number of opinions, their positions become so stressed (pulled and pushed
in so many directions) as to be meaningless.  By "simple" I mean driven
almost entirely by a handful of underlying categories or dimensions
(instead of the complex webs of semantic relationships implicit in the idea
of considerations).  If the opinion space actually is dimensionally simple
(for example a left-right continuum or the left-right /
authoritarian-libertarian "Nolan chart"), network analysis methods can
uncover these dimensions more accurately than factor analysis (Brazill and
Grofman 2002).

In case the space is not simple, we need tools to explore it as a
network.  First, this means displaying a reasonable number of points at
once.  The Meaning Map displays the most prominent opinions as the "top
level" view.  The system allows the user to specify by what combination of
criteria each opinion is rated as prominent.

For the same reasons, the system also indicates node-level stress.  The
color of opinions indicates the extent to which they "fit" in their current
position.  Opinions that fit well are green, and as stress increases the
color fades to yellow and then red.  This highlights opinions that serve as
bridges between clusters or as "wormholes" in the dominant dimensions.

Finally, the system allows for exploration of the network by providing a
"statistical zoom."  Since underlying dimensions cannot be assumed from the
outset, the idea of "zooming in" to view less prominent opinions "near" a
certain opinion can only be based on their statistical nearness.  When a
user selects an opinion to view in detail (in order to read considerations
and/or give an informed vote), they see a map of a different set of
opinions, selected not just for prominence but also for their statistical
relatedness to the current opinion.

Statistical zoom can result in a dramatic re-ordering of the space – for
example, "god exists" may be a highly stressed bridging opinion between
left and right, both negatively and positively related to many opinions in
those two dominant clusters.  However, when this opinion is made
artificially dominant by statistically zooming to it, one might see a
completely different ordering, where religious and non-religious (or
moralistic and non-moralistic) statements are two dominantly opposed
clusters.  This can be a powerful (if indirect) way of seeing different
worldviews in the same data.


Limitations

The Meaning Map is a relatively simple first step.  The set of problems in
deliberative democracy it attempts to address do not include those of
constructing a full alternative political system.  Apart from issues of
translating the results of national deliberation into national policy, many
technical issues would need to be addressed before this system could even
be capable of creating one integrated national deliberation.  And then
there are security, identity and privacy issues.  This system merely
elevates the possible deliberative group size from on the order of tens, to
perhaps tens of thousands, and makes deliberative outcomes more likely
among diverse groups of strangers.  Even this simple system would be very
useful to citizens trying to pool their abilities to evaluate existing or
proposed legislation.  It would also be useful to visually see candidates'
positions (and their change over time) on the most prominent issues, where
that agenda is determined by a deliberation among citizens instead of being
driven by candidate strategy or by media agenda setting.

The Meaning Map does not attempt to directly facilitate empirical
verification or the evaluation of risks (automating the combination of
probabilities and severities), although it clearly can help large groups of
people do both of these things more efficiently than discussion.  Some
potentially important work on these possibilities is taking place on the
topic of deliberation for environmental risk regulation.  Payne (1996)
points out the suitability of deliberation for global environmental issues,
and McBurney and Parsons (2001) formulate an environmental-risk-specific
ideal deliberative system, which, although conversational in structure,
represents a compelling argument for customized public sphere facilitation
systems for specific domains.

At the time of this writing, the opinion mapping methods have been tested
with real data, but the system has not yet been tested with a large group
of users.  Unforeseen problems of all varieties could arise through real
use.  It is conceivable that exposure to the opinion map may lead to an
increase instead of a decrease in polarization.  Experiments should be
conducted to assess the impact of four different experimental conditions:
the complete system, the system without the statistical map, the system
without considerations, and the system without either (just a simple open
polling system).  Collaboration on these or other experiments with this
software is very welcome.

Finally, at present the system does not provide opportunities for
brainstorming.  All user interaction is relatively composed; people are
likely to extensively edit a contribution before releasing it to the
group.  Brainstorming benefits from composition-times as short as possible,
but also from the expectation that nobody is going to hold you to an
opinion you blurt out.  Therefore, perhaps the system should have anonymous
"co-edited chat" rooms (meaning every character a user types is immediately
visible to others).  These could be used to collaboratively write an
opinion statement or consideration, or to collaboratively interpret an
existing statement or consideration.  In the former case, it might be
useful to select "the right kind of discussion partner" by choosing a set
of opinions with which the partner must agree and/or a set of opinions with
which they must disagree.  In the latter case, it might be useful to pair
people who fall on opposite sides of the issue.


Conclusion

Richard Davis argues that "the notion that the public will take control of
agenda setting is absurd.  … Someone must organize the discussion and frame
the alternatives.  Then, and only then, can the public respond
intelligently" (1999:170, emphasis added).  If we can complete the work
this project begins, making mass deliberation technically feasible, the
situation will be reversed.  It will instead be absurd to claim that the
public should not participate in setting its own agenda, that aggregates of
privately-measured agreements with fixed opinion statements framed by
elites can be called public opinion, or that elections conducted under such
conditions can be called legitimate democracy.

Which sphere is virtual?

Recent work on the deliberative potential of the Internet has used the term
"virtual sphere" to refer to online discussion (Papacharissi, 2002).  Zizi
Papacharissi's overview of the subject highlights the differences between
the Habermasian ideal public sphere and the realities of discussion, both
online and offline.  She summarizes Michael Schudson's 1997 critique of
deliberation-as-discussion: "there is little evidence that a true ideal
public ever existed, and … public discourse is not the soul of democracy,
for it is seldom egalitarian, may be too large and amorphous, is rarely
civil, and ultimately offers no magical solution to problems of democracy"
(2002:11).  It's high time for us optimists to admit that discussion itself
may not be an adequate tool for creating the ideal public sphere.

However, our goal can no longer be seen as a "tragic and stoic pursuit of
an almost impossible rationality, recognizing the impossibility of an ideal
public sphere and the limits of human civilization, but still striving
toward it" (11).  More important than whether communication occurs online
or offline are the questions of whether communication is guided by
arbitrary, oversimplified, and self-reproducing mental maps of opinion
space, whether people must reason in lock-step together in order to benefit
from (and come to understand) each other's reasoning, and whether everyone
can participate equally in the framing and agenda setting processes.  The
possibility of removing these limits creates a whole new set of
opportunities and problems, both theoretical and technical, which may lead
to eventually creating a more real public sphere than has ever existed
before.


References

Bell, D. (1999).  Democratic Deliberation: The Problem of
Implementation.  In Macedo,
S. (ed.), Deliberative Politics: Essays on Democracy and Disagreement, 70-87.
Brazill, T. and Grofman, B. (2002). "Factor analysis versus
multi-dimensional scaling:
binary choice roll-call voting and the U.S. Supreme Court." Social Networks
24:201-229.
Cohen, J. (1993).  Moral Pluralism and Political Consensus.  In D. Copp, J.
Hampton,
and J. Roemer (eds.), The Idea of Democracy.  Cambridge University Press,
270-91.
(1997). Procedure and Substance in Deliberative Democracy.  In J. Bohman
and W. Rehg (eds.), Deliberative Democracy: Essays on Reason and
Politics.  Cambridge, MA: MIT Press
(1998). Democracy and Liberty.  In J. Elster (ed.), Deliberative
Democracy.  Cambridge University Press, 185-231.
Dahlberg, L. (2001).  Computer-Mediated Communication and The Public Sphere: A
Critical Analysis.  Journal of Computer-Mediated Communication, 7(1), 1-28.
Davis, R. (1999).  The Web of Politics.  New York: Oxford University Press.
Denver, D., Hands, G., & Jones, B. (1995). Fishkin and the deliberative
opinion poll:
Lessons from a study of the Granada 500 television program.  Political
Communication, 12, 147-156.
Eliasoph, N. (1996).  Making a Fragile Public: A Talk-Centered Study of
Citizenship and
Power.  Sociological Theory, 14, 262-289.
  (1998).  Avoiding Politics: How Americans Produce Apathy in Everyday
Life.  New York, NY: Cambridge University Press.
Elster, J. (1998).  Introduction.  In J. Elster (ed.), Deliberative
Democracy.  Cambridge
University Press, 1-18
Fearson, J. (1998).  Deliberation as Discussion.  In J. Elster (ed.),
Deliberative
Democracy.  Cambridge University Press, 44-68
Fishkin, J., & Luskin, R. (1999). Bringing deliberation to the democratic
dialogue. In M.
McCombs, A. Reynolds, & Eds. (Eds.), The poll with a human face: The
National Issues Convention experiment in political communication, Mahwah,
NJ: Lawrence Erlbaum, 3-38
Fishkin, J. (2000).  Virtual Democratic Possibilities: Prospects for
Internet Democracy.
Conference Papers, "Internet, Democracy, and Public Goods," Belo
Horizonte, Brazil, November 6-10, 2000.
Fruchterman, T. and Reingold, E (1991). "Graph Drawing by Force-Directed
replacement." Software--Practice and Experience 21:1129-1164.
Gastil, J. (2000).  Is Face-to-Face Citizen Deliberation a Luxury or a
Necessity?
Political Communication, 17, 357-361
Gutmann, A. and Thompson, D. (1996).  Democracy and Disagreement.  Cambridge,
MA: Harvard University Press.
Habermas, J. (1984).  The Theory of Communicative Action, Vol. 1. (T.
McCarthy,
Trans.). Boston, MA: Beacon.
  (1991) The Public Sphere.  In C. Mukerji & M. Schudson (Eds.),
Rethinking popular culture.  Berkeley: University of California Press.
Hardin, R. (1999).  Deliberation: Method, not Theory.  In Macedo, S. (ed.),
Deliberative
Politics: Essays on Democracy and Disagreement, 70-87.
Klein, H. (1999).  Tocqueville in Cyberspace: Using the Internet for
Citizen Associations.
The Information Society, 15, 213-220.
McBurney, P. and Parsons, S. (2001). Intelligent systems to support
deliberative
democracy in environmental regulation. Information and Communications
Technology Law, 10 (1): 33-43.
Noelle-Neumann, E. (1984).  The spiral of silence: Public opinion – our
social skin.
Chicago: University of Chicago Press.
Papacharissi, Z. (2002).  The virtual sphere: The Internet as a public
sphere.  New
Media and Society, 4, 9-27
Payne, R. (1996).  Deliberating Global Environmental Politics.  Journal of
Peace
Research, 33 (2), 129-136.
Price, V. and Capella, J. (2001).  Online Deliberation and its Influence:
The Electronic
Dialogue Project in Campaign 2000.  Conference Papers, annual meeting of the
American Association for Public Opinion Research, May 2001.
Resnick, P. (2000) Beyond Bowling Together:  SocioTechnical Capital. In J.
Carroll,
HCI in the New Millennium. New York: Addison Wesley, Chapter 29.
Sanders, L. (1997). Against Deliberation. Political Theory. 25(3), 347-364.
Shapiro, I. (1999).  Enough of Deliberation: Politics is about Interests
and Power.  In
Macedo, S. (ed.), Deliberative Politics: Essays on Democracy and
Disagreement, 28-38.
Warren, M. (2002).  Deliberative Democracy.  In Carter, A. and Stokes, G.
(eds.),
Democratic Theory Today.  Malden, MA: Polity Press.  173-202
Witschge, T. (2002).  Online Deliberation: Possibilities of the Internet
for Deliberative
Democracy.  Conference papers, "Prospects for Electronic Democracy," Carnegie
Mellon University.
Zaller, J. (1992).  The Nature and Origins of Mass Opinion. New York:
Cambridge
University Press.