Download Harvard Legal Theory Forum Working Paper Series

Document related concepts

Communication in small groups wikipedia , lookup

Democratic education wikipedia , lookup

Citizens' assembly wikipedia , lookup

Transcript
Harvard Legal Theory Forum Working
Paper Series
Paper No. 1
Qui ckTime™ and a
TIFF (U ncompr essed) decompressor
are needed to see thi s pi cture.
DELIBERATIVE DEMOCRACY, SOCIAL PSYCHOLOGY, AND
THE AUTHORITY OF LAWS
Jonathan Gingerich
March 3, 2009
This paper can be downloaded at the Harvard Legal Theory Forum
Working Paper Series Index:
http://blogs.law.harvard.edu/hltf/files/2009/03/gingerich0303200
9.doc
A Student Organization at Harvard Law School 
[email protected]
The Harvard Law School name and/or shield are trademarks of the President and Fellows of Harvard
College and are used by permission of Harvard University.
Gingerich
1
DELIBERATIVE DEMOCRACY, SOCIAL PSYCHOLOGY, AND THE AUTHORITY OF LAWS
DRAFT—NOT FOR CITATION
Jonathan Gingerich
March 3, 2009
N.B. This paper is rather long—if you don’t wish to read all of it but wish to read part, I would
recommend reading the Introduction (pp. 1-8), Part III-C (pp. 46-58), Part IV (pp. 58-71), and
the Conclusion (pp. 71-73). -JG
I. INTRODUCTION
A. The Problem of Deliberation and Authority
Many authors believe that democratic deliberation makes laws legitimate and
authoritative, even if they do not advance this argument explicitly.1 Some authors make this
argument explicitly. Particularly, many sophisticated political theorists argue that if a law is the
result of democratic deliberation, then there is a prima facie reason to treat it as having legitimate
authority. This is because, deliberative democrats argue, democratic procedures accurately
reflect the will of the people.
Perhaps this makes sense on a homo economicus model of rationality that assumes that
people reveal their preferences by their actions, that the motivational structures in which
preferences are revealed do not distort them, and that better access to information always
produces better results for a given individual. Modern economics has been dominated by “a

© 2008-2009, Jonathan Gingerich. I am grateful for the comments and questions of William
W. Fisher III, Yochai Benkler, Frank Michelman, Patrick Connolly, Stacylyn Dewey, and the
students in the Spring 2008 Motivation Seminar at Harvard Law School. I am particularly
grateful to Emily Wack for suggesting this topic. All errors, of course, are mine.
1
See, e.g., BRUCE ACKERMAN & JAMES S. FISHKIN, DELIBERATION DAY # (2004) (describing a
scheme for state-sponsored small group deliberations before elections).
Gingerich
2
conception of welfare based on the satisfaction of existing preferences, as measured by
willingness to pay; in politics and law, something called ‘paternalism’ is disfavored in both the
public and private realms.”2 However, social scientific literature that shows the diversity of noneconomically-rational motivational forces that structure human action casts this account into
severe doubt. If “design levers” like communication, empathy, solidarity, and motivation
crowding affect people’s preferences, and if different procedural frameworks for democratic
deliberation differentially affect these design levers, it is not apparent that democratic
deliberation can neutrally (that is, without ideological bias) reflect the beliefs, arguments, or
preferences of members of a democratic society. This means that laws cannot be authoritative
simply because they are the outcome of democratic deliberation; instead, any justification of
laws must stem from reasons other than their approval by democratic institutions (though
democratic institutions may still, for instrumental reasons, be the most desirable means of
producing and enforcing laws).
If different deliberative procedures affect people’s motivations in different ways and
generate different outcomes just because of their structures’, then choices about how to
deliberate are also choices about the outcomes of deliberation. If this is true, it discredits
arguments that substantive laws or procedural frameworks are legitimate simply because people
have agreed on them.3 Existing procedural frameworks reflect particular political commitments,
and must be justified with reference to these commitments, not with reference to the claim that
they are the outcomes of fair deliberation. More particularly, claims that certain structures are
2
Cass Sunstein, Democracy and Shifting Preferences, in THE IDEA OF DEMOCRACY 196, 198
(David Copp, Jean Hampton, & John E. Roemer, eds., 1993).
3
I am not sure that no deliberative framework could be truly procedurally neutral. There might
be some way of framing deliberation that I have not thought of that avoids the problems of
ideological infection that I describe, but I am very skeptical of the possibility.
Gingerich
3
legitimate because they simply “neutrally” aggregate citizens’ preferences should be disregarded,
as no aggregation of preferences is neutral. Whether it involves voting, discussion, bargaining,
or a mix of these, any democratic structure will shape as well as aggregate preferences or beliefs.
There are no systems of deliberation in which the aggregation of data will not feed back into the
system.4
The story I am telling here is rather procrustean in some ways. Of course, many
deliberative democracy theorists, and particularly Jürgen Habermas with his account of
“communicative rationality,” recognize that power, resources, and structures of communication
affect the outcomes and beliefs of people within communicative systems and treat these factors
as the proper focus of the analysis of the legitimacy of communication.5 Habermas’s “ideal
speech situation” is an attempt to ensure equal access to participatory democracy without
distortion by wealth or power.6 The ideal speech situation attempts to do this by defining the
characteristics of a deliberative practice that would make its outcomes legitimate.7 Thus,
Habermas recognizes that deliberation is not just about revealed preferences—people work
things out through communication and change their minds after deliberating with other people.
However, the goal of communicative rationality is still undistorted communication: “Whoever
enters into discussion with the serious intention of becoming convinced of something through
dialogue with others has to presume performatively that the participants allow their ‘yes’ or ‘no’
4
See MANUEL DELANDA, A NEW PHILOSOPHY OF SOCIETY: ASSEMBLAGE THEORY AND SOCIAL
COMPLEXITY # (2006) [PARENTHETICAL].
5
See JÜRGEN HABERMAS, BETWEEN FACTS AND NORMS # (William Rehg trans., 1996); Jürgen
Habermas, Constitutional Democracy: A Paradoxical Union of Contradictory Principles?
(William Rehg trans.), 29 POL. THEORY 766, # (2001); JÜRGEN HABERMAS, THEORY OF
COMMUNICATIVE ACTION # (Thomas McCarthy trans., 1984).
6
See JÜRGEN HABERMAS, ON THE PRAGMATICS OF COMMUNICATION 367-68 (Maeve Cook ed.,
1998).
7
See id. at 365-68; HABERMAS, BETWEEN FACTS AND NORMS, supra note 5, at 28.
Gingerich
4
to be determined solely by the force of the better argument.”8 Unlike psychoanalytic theories of
rationality the Habermasian theory is (largely) unsuspicious of the sources of the beliefs and
preferences that individuals articulate as the baseline of democratic deliberation. This means
that, on the Habermasian account, communicative rationality must have tremendous corrective
power—it must be able to reveal and overcome beliefs that are distorted before deliberation
begins.9 If, as I hypothesize, different deliberative structures lead to different political outputs
solely because of their procedural structure, the Habermasian theory must rely on some
substantive, pre-communicative political principles to resolve questions of legitimacy.10 It
would be one thing if real world deliberation were muddled by the pathologies of deliberation
only because of individual persons’ weakness of the will or failure to state criticizable public
political claims, but another thing altogether if deliberation were muddled because of
fundamental psychological features of deliberation. If the later is true, it seems unlikely that
participatory democracy could ever come even reasonably close to an ideal speech situation for
the purpose of providing and making public information necessary to meet the requirements of
justice.11 Furthermore, in spite of a critical awareness of the historical forces that shape practices
of legitimation, Habermas’s theory still gives legitimating force to deliberative practices that
8
HABERMAS, supra note 6, at 367.
See 1 HABERMAS, THEORY OF COMMUNICATIVE ACTION, supra note 5, at 398 (“The utopian
perspective of reconciliation and freedom is ingrained in the conditions for the communicative
sociation of individuals; it is built into the linguistic mechanism of the reproduction of the
species.”).
10
See JEAN-FRANÇOIS LYOTARD, POSTMODERN CONDITION 65 (Geoff Bennington & Brian
Massumi trans., 1984) (“There is no reason to think that it would be possible to determine
metaprescriptives common to all of these language games or that a revisable consensus like the
one in force at a given moment in the scientific community could embrace the totality of
metaprescriptions regulating the totality of statements circulating in the social collectivity. As a
matter of fact, the contemporary decline of narratives of legitimation . . . is tied to the
abandonment of this belief.”).
11
Cf. HABERMAS, BETWEEN FACTS AND NORMS, supra note 5, at #. I am grateful to Frank
Michelman for suggesting this argument.
9
Gingerich
5
reflect particular ideological commitments. Habermas would argue that certain deliberative
practices better approximate the ideal speech situation than do others because they decrease the
influence of wealth and power on participatory democracy, but I will argue that, ultimately, the
role of ideology cannot be eliminated from deliberation, only rearranged. Habermas could
overcome this by specifying substantive political principles, but any commitment to such
principles must, I will argue, come prior to deliberation.
B. Toward Recognizing the Ideological Structure of Deliberative Democracy
Political theorists and philosophers have undertaken significant studies of deliberative
democracy and its implications for democratic legitimacy in the past decade.12 Social
psychologists have also produced substantial literature on how group processes affect decisionmaking.13 Some authors have examined the impact that the psychological structures of group
decision-making can have on deliberation about politics.14 In this essay, I will use the
12
See ACKERMAN & FISHKIN, supra note 1 [PARENTHETICAL]; JAMES BOHMAN, PUBLIC
DELIBERATION (1996) [PARENTHETICAL]; ROBERT E. GOODIN, REFLECTIVE DEMOCRACY
(2003) [PARENTHETICAL]; DEBATING DELIBERATIVE DEMOCRACY (James S. Fishkin & Peter
Laslett eds., 2003) [PARENTHETICAL]; DELIBERATIVE DEMOCRACY (Jon Elster, ed., 1998)
[PARENTHETICAL]; DELIBERATIVE DEMOCRACY: ESSAYS ON REASON AND POLITICS (James
Bohman & William Rehg, eds., 1997) [PARENTHETICAL]; DELIBERATIVE POLITICS: ESSAYS
ON DEMOCRACY AND DISAGREEMENT (Stephen Macedo, ed., 1999) [PARENTHETICAL]; AMY
GUTMANN & DENNIS THOMPSON, DEMOCRACY AND DISAGREEMENT (1996) [hereinafter
GUTMANN & THOMPSON, DEMOCRACY AND DISAGREEMENT] [PARENTHETICAL]; AMY
GUTMANN & DENNIS THOMPSON, WHY DELIBERATIVE DEMOCRACY? (2004) [hereinafter
GUTMANN & THOMPSON, WHY DELIBERATIVE DEMOCRACY?] [PARENTHETICAL];
HABERMAS, BETWEEN FACTS AND NORMS, supra note 5 [PARENTHETICAL]; HENRY S.
RICHARDSON, DEMOCRATIC AUTONOMY: PUBLIC REASONING ABOUT THE ENDS OF POLICY (2002)
[PARENTHETICAL]; CASS SUNSTEIN, DESIGNING DEMOCRACY (2001) [PARENTHETICAL];
see also COLIN CROUCH, POST-DEMOCRACY (2004) [PARENTHETICAL]; MARK TUSHNET,
TAKING THE CONSTITUTION AWAY FROM THE COURTS (1999) [PARENTHETICAL].
13
See, e.g., BLACKWELL HANDBOOK OF SOCIAL PSYCHOLOGY: GROUP PROCESSES (Michael A.
Hogg & R. Scott Tindale, eds., 2001) [PARENTHETICAL].
14
See David Schkade, Cass R. Sunstein, & Reid Hastie, What Happened on Deliberation Day?
95 CAL. L. REV. 915, # (2007); SUNSTEIN, supra note 12, at #; CASS SUNSTEIN, INFOTOPIA #
(2006).
Gingerich
6
psychological literature to address whether and how democratic deliberation can provide reasons
for legal legitimacy, and I will explore the implications of empirical findings about group
decision making for the theoretical account of democratic legitimacy.15
In Part II, I will describe theories of democratic deliberation and legal authority. I will
set out a typology of theories of deliberative democracy, limit my scope to theories that claim
that democratic deliberation can provide constitutive reasons for legal authority, and explore the
requirements of democratic deliberation from the standpoint of political theory.
In Part III, I will summarize the social psychological literature, and will explore what
empirical evidence about group processes, and particularly deliberation about politics, has for
democratic deliberation in practice. I will describe different ways in which deliberation can
rearrange individual preferences and political commitments and explore how framing of
democratic deliberation inevitably shapes outcomes.
My principle argument, which I will develop in Part IV, has two prongs. The first,
weaker prong is that in any society even roughly resembling the United States in 2008,
democratic deliberation cannot provide a reason that any law is legitimate, because it must rest
on a fully adequate constitutional framework in order to reflect the values necessary for the
creation of political legitimacy. According to this argument, the procedural structure of
democratic deliberation will have predictable ideological impacts on the outcome of deliberation,
so to be truly democratic, any deliberative structure must be justified with reference to some
15
There is a fundamental tension in my argument: I advance a deep skepticism about the
possibility of deliberation and take a highly critical stance, but in doing so, I rely heavily on
“objective” social sciences, which constitute a much more liberal methodological approach. My
argument should, therefore, be conceptualized as an internal criticism of the coherence of the
system of thought that liberal deliberative democracy theory represents rather than a
methodologically foreign and independent assault on such theories. I am grateful to Yochai
Benkler for raising this point.
Gingerich
7
independent procedural rule for structuring the deliberation. In order for any foundational rules
of deliberation to be justified, they must ensure that every political agent is able to equally
participate in and influence the decision making process. Because of the pervasive effects of
geographical grouping of social groups, social hierarchy, and unequal resources, these conditions
necessary to justify any foundational rules for deliberation are not likely to obtain absent radical
social change.
My second, stronger claim is that democratic deliberation can never provide a reason that
a law is legitimate. Readers can accept my first claim without accepting my second claim, which
argues that any attempt to develop a framework for democratic deliberation that could generate
reasons for legitimacy will inevitably by infected by ideology. I will argue that any attempt to
establish neutral background rules will fail, because different structures of deliberation will
always have differential impacts on the motivational structures of participants, which will have
predictable political outcomes. Because of this ideological infection, any structure of
deliberation will be based on certain ideological structures, and cannot generate neutral reasons
for laws’ legitimacy. These reasons must, then, come from some source other than democratic
deliberation.
This is an attack on conseqentialist theories of democratic deliberation, not deontological
theories. Following Henry Richardson, I am concerned “with the will of the people as developed
and expressed in actual democratic procedures, and whether it is plausible to think of it as being
given a constitutive voice affecting what it is that we ought, democratically, to do.” 16 For
Henry S. Richardson, Our Call: The Constitutive Importance of the People’s Judgment, J.
MORAL PHIL. (forthcoming), manuscript at 5, available at
http://www1.georgetown.edu/departments/government/programs/speakerseries/27718.html (all
page references are to the version available at the Georgetown website); cf. Jon Elster,
Introduction to DELIBERATIVE DEMOCRACY 1 (Jon Elster, ed., 1998) [PARENTHETICAL].
16
Gingerich
8
instance, my argument does not touch on Rawls’s justice as fairness, nor does it touch on Jeremy
Waldron’s deontological account of deliberation.17 Likewise, my argument is not an attack on
purely normative accounts of deliberative democracy. It does not undermine theories that aim
only to set out what democratic society should look like on the basis of purely normative
reasons. It does, however, refute theories of democratic legitimacy that rely on any claims about
what democratic deliberation will accomplish in society as it is presently constituted and theories
that justify certain structures of democratic deliberation with reference to their actual political
outcomes. My argument is also not that deliberation is less legitimate than other modes of
decision-making; I argue that the laws it generates ought not to be considered authoritative
simply because they are outputs of deliberative democratic systems.
II. DELIBERATIVE DEMOCRACY AND AUTHORITY
In this section, I will first summarize contemporary theories of deliberative democracy
and explore how these theories relate to legal authority. I will then describe the institutions that
deliberative democracy requires.
A. What is Deliberative Democracy?
Deliberative democracy is (1) democratic and (2) deliberative. (1) means that a
deliberative democracy decides what laws to enact and enforce collectively, with every member
of the association who will be affected by the decision participating, either directly or through a
representative.18 (2) means that a deliberative democracy goes about making these decisions by
17
See JOHN RAWLS, POLITICAL LIBERALISM # (1996) [PARENTHETICAL]; JEREMY WALDRON,
LAW AND DISAGREEMENT # (1998) [PARENTHETICAL].
18
See Elster, supra note 16, at 8; Richardson, supra note 16, manuscript at #.
Gingerich
9
means of the affected members or their representatives publicly reasoning with one another.19
This typically means that the members or their representatives must be committed to impartiality
and rationality.20
These are only the baseline requirements for deliberative democracy. Even within a
democratic ideological frame, it is far from clear what “democratic structures” are, and different
theorists flesh out these baseline requirements in different ways. Typically, they include some
constraint on the substantial fairness of the democratic deliberation, so that deliberation is not
unfairly skewed by unjustified distributions of resources or power. Theorists of deliberative
democracy typically require that the reasons given in public deliberation be “mutually
acceptable.”21 This means that the reasons given for adopting a particular law must be reasons
other than mere self-interest; they must somehow persuade other people that it is in the public
good to adopt a certain law.22
As Jon Elster notes, there are three aspects of democratic deliberation: arguing, voting,
and bargaining.23 Arguing is an attempt to appeal to reason to transform other people’s
preferences.24 It requires an appeal to impartial values, since simply asserting one’s own interest
19
See Richardson, supra note 16, manuscript at 5. It would be a mistake to think of deliberation
as an on/off switch; deliberation can take many different forms, and it is difficult even to image a
political situation in which there would be no deliberation.
20
See Elster, supra note 16, at 8.
21
GUTMANN AND THOMPSON, WHY DELIBERATIVE DEMOCRACY?, supra note 12, at 7.
22
Id. at 4.; see BRUCE A. ACKERMAN, SOCIAL JUSTICE IN THE LIBERAL STATE, 10 (1980) (“No
reason is a good reason if it requires the power holder to assert: (a) that his conception of the
good is better than that asserted by any of his fellow citizens, or (b) that, regardless of his
conception of the good, he is intrinsically superior to one or more of his fellow citizens.”);
Joshua Cohen, Deliberation and Democratic Legitimacy, in DELIBERATIVE DEMOCRACY: ESSAYS
ON REASON AND POLITICS, supra note 12, at 69, #.
23
Elster, supra note 16, at 5.
24
Id. (noting that both Habermas and Rawls believe that “political choice, to be legitimate, must
be the outcome of deliberation about ends among free, equal, and rational agents.”). Such
arguing might include appeals to emotion (and, in practice, it almost always does), but such
Gingerich
10
will probably not change other people’s minds. Arguing is also often about factual matters, since
people have the best chance of persuading each other by changing others’ opinions on causal
questions than on deeply held values.25 Bargaining occurs when participants trade off their
preferences against each other; outcomes depend on the resources that parties have to make
credible threats and promises.26 Voting, unlike bargaining and arguing, is not a form of
communication but is simply the aggregation of preferences.27 The preferences that are
aggregated through voting need not be narrowly self-interested preferences; they can also be
political preferences that refer to the public good.28
In this essay, I will concern myself primarily with arguing, rather than with voting and
bargaining, though it may not be possible to separate these from each other entirely, and in a
deliberative democratic structure, bargaining and voting are always present in the background.29
I will examine what role deliberation in particular has for legal authority, and how procedures for
arguing impact the ability of individual citizens to participate in deliberative democracy.
B. Deliberative Democracy and Authority
appeals must be criticizable, and they must be the sorts of claims that are likely to get other
people to change their minds. 1 HABERMAS, THEORY OF COMMUNICATIVE ACTION, supra note 5,
at 398 (“A process of self-preservation that has to satisfy the rationality conditions of
communicative action becomes dependent on the integrative accomplishments of subjects who
coordinate their action via criticizable validity claims.”).
25
Elster, supra note 16, at 6.
26
Id. at 6-7.
27
Id.
28
See, e.g., JEAN-JACQUES ROUSSEAU, THE SOCIAL CONTRACT # (Maurice Cranston, trans.,
1968) [PARENTHETICAL].
29
An exception to this would be a consensus-based system that does not rely on voting. See
HABERMAS, supra note 12, at #. However, since I am concerned with actual applications of
deliberative democracy, I will, for the most part, not concern myself with these theories.
Gingerich
11
Some theories of democratic deliberation think that the fact that deliberative democracy
generates a law is a reason that the law is authoritative or legitimate.30 In other words, they think
that deliberative democratic ratification provides a normative reason to follow a law.31 A fullblown coherentist, pragmatic theory of democracy might treat democratic ratification as a
necessary and sufficient condition for legal authority.32 Standing opposite such a theory are
objective theories of authority that hold that the simple fact that a law is the outcome of
democratic procedures can never be even a prima facie reason that the law has legal authority. 33
Between these two theories are mixed theories that view democratic ratification as a certain sort
of reason to treat a given law as having legal authority but maintain that there is also a balance of
reasons as to whether the law is authoritative independent of this reason.34
Additionally, on my reading there is a spectrum of approaches to deliberative democracy
that range from more theoretical accounts, concerned to a greater extent with ideal conditions of
deliberation, to more practical accounts, concerned to a greater extent with how deliberation
might take place in the real world. Jürgen Habermas lies toward the theoretical end of this
spectrum, while Amy Gutmann and Dennis Thompson lie near the center, and James S. Fishkin,
30
See, e.g., Joshua Cohen, Procedure and Substance in Deliberative Democracy, in
DELIBERATIVE DEMOCRACY: ESSAYS ON REASON AND POLITICS, supra note 12, at 407;
Richardson, supra note 7, at 18.
31
My argument is only about whether democratic deliberation can provide normative authority
to laws, not whether it can provide political authority. That is, I am not concerned with the
question of whether deliberative democratic institutions make people more likely in practice to
obey laws, only with whether such institutions themselves provide reasons that their laws ought
to be obeyed.
32
See, e.g., Tom Christiano, Authority, in STANFORD ENCYCLOPEDIA OF PHILOSOPHY (Edward N.
Zalta, ed., Fall 2004), http://plato.stanford.edu/archives/fall2004/entries/authority
[PARENTHETICAL].
33
See, e.g., Scott J. Shapiro, Authority, in THE OXFORD HANDBOOK OF JURISPRUDENCE AND
PHILOSOPHY OF LAW 432, # (Jules Coleman and Scott Shapiro eds., 2002) [PARENTHETICAL].
34
See, e.g., David Estlund, Beyond Fairness and Deliberation: The Epistemic Dimension of
Democratic Authority, in DELIBERATIVE DEMOCRACY: ESSAYS ON REASON AND POLITICS, supra
note 12, at 173, 182; Richardson, supra note 16, at 1.
Gingerich
12
Bruce Ackerman, and Cass Sunstein lie toward the practical end. The weak argument developed
in Part IV-A of this paper cuts against the theorists at the practical end of the spectrum. The
strong argument developed in Part IV-B cuts against all of these mixed theories, to the extent
that they claim that democratic deliberation can, by itself, provide normative reasons to treat a
law as legitimate or authoritative.
1. Coherentism
This position argues that each person ought to submit to the outcome of democratic
deliberation.
So if they advocate some policy on the grounds that it conforms to what they take to be
the correct principle of justice, J, and the majority chooses a different policy on the
grounds of an incompatible principle, L, the [coherentist] democratic theory says that
they ought to accept the policy that is grounded in L because only in this way do they
accord the proper equal respect to their fellow citizens.35
On this account, all that matters for legitimacy is that a law be ratified in accordance with an
acceptably fair procedure. According to this view, there are political reasons for and against any
policy that are independent of what the polity agrees on.36 However, no matter how strong one
believes the political reasons for a given policy to be, once a different policy based on conflicting
principles has been democratically ratified, the ratified policy is fully legitimate. If a law is
passed according to the correct procedures that says “don’t spit on the sidewalk” it not only gives
every member of the society a reason not to spit on the sidewalk but also excludes every other
reason that there could be for spitting on the sidewalk, at least to the extent that the topic of the
legislation is an appropriate issue on which to legislate.37 This is an extreme position, and it is
35
Christiano, supra note 32.
For any theory of deliberative democracy that requires public reasoning that amounts to
anything more than an aggregation of private preferences, there must be such reasons.
37
See infra note 38 and accompanying text. I follow Raz’s account of exclusionary reasons in
my description of the cohrentist account. While most theories of democratic deliberation do not
require the reasons for action provided by legitimate laws to be exclusionary, following
36
Gingerich
13
not clear that any advocates of deliberative democracy actually go this far.38 However, it does
represent the end of the spectrum of theories that grants the most constitutive authority to
democratic decision-making.
2. Objectivism
In contrast to this coherentist approach are theories that argue that the balance of reasons
for acting or not acting in a particular way is never affected by whether a deliberative democracy
has ratified a law.39 This approach says that any policy must be justified by reference to
objective political principles. “If there is a fact of the matter as to what we ought to do in a given
situation then, given the multiplicity of considerations that impinge on what we ought to do,
there must be objective reasons bearing on each alternative course of action.”40 Human
judgment is always fallible, and people often make mistakes about where the balance of reasons
lie, so a person or a group’s judgment about where the balance of reasons lies cannot itself affect
the objective balance of reasons. Theorists who take this stance tend to think that theories that
say that democratic ratification is itself a reason a law has legitimacy illegitimately bootstrap
reasons into existence. If there is no reason for me to take some action (say, to ), presumably
simply deciding to  should not itself provide a reason to . Likewise, it is unclear why having a
deliberative assembly decide I should  in the absence of any other reasons for me to do so gives
me a reason to . 41
This theory can still say that deliberative democracy is a good idea—however, the reason
that it is good is not that it provides legitimacy to laws, but that it is good for some other reason.
Christiano’s characterization of the coherentist position, I treat any such theories that do not give
exclusionary force to a legitimate law as “mixed theories” rather than coherentist theories.
38
See Estlund, supra note 34, at 176-77.
39
See Richardson, supra note 16, at #.
40
Id. at 4.
41
See id.
Gingerich
14
(For instance, it might be that democratic procedures are fair,42 develop desirable character traits
in citizens,43 or are simply more likely than any other form of government to generate outcomes
that are relatively close to where the objective balance of reasons lies.44) However, the reasons
that deliberative democracy might be good, on this account, do not arise from democratic
deliberation but rest on independent, objective political principles.
3. Mixed Theories
Mixed theories stand between the coherentist and objectivist theories, and hold that the
fact that democratic deliberation generates a law is a factor to consider in determining whether
the law is legitimate. The rules of democracy, then, give some constitutive authority to what the
people decide.45 This position accepts that democratic deliberation must be about something
other than the mere aggregation of preferences, so there must be some independent standards for
42
See Shapiro, supra note 33, at # [PARENTHETICAL].
See David Copp, in The IDEA OF DEMOCRACY, supra note 2, at 113 [PARENTHETICAL]; see
also David Dyzenhaus, The Legitimacy of Legality, 46 U. TORONTO L. REV. 129, 180 (1996)
[PARENTHETICAL].
44
See Estlund, supra note 34, at 183. This differs slightly from the position Estlund ends up
advancing, which he calls “epistemic proceduralism” and which is an instance of a mixed theory
because it provides that it is possible to think that a judgment is morally mistaken but still
morally binding on citizens as citizens for procedural reasons. Id. at 185. Like the theories of
deliberative democracy described below, the non-epistemic aspect of its legitimacy hinges on its
fairness: the reason to obey laws in non-epistemic cases (i.e. where a citizen has good reasons to
think that the majority was wrong) is the procedural fairness of deliberative democracy. Id. at
195.
45
Richardson, supra note 16, at 18; see also Habermas, Constitutional Democracy, supra note 5,
at 779.
[I]n the role of persons who act morally, legal persons must also be able to follow
the law out of respect for the law. For this reason, valid (in the sense of existing)
law must also be legitimate. And the law can satisfy this condition only if it has
come about in a legitimate way, namely, according to the procedures of
democratic opinion- and will-formation that justify the presumption that outcomes
are rationally acceptable. The entitlement to political participation is bound up
with the expectation of a public use of reason: as democratic colegislators,
citizens may not ignore the informal demand [Ansinnen] to orient themselves
toward the common good.
Id.
43
Gingerich
15
what constitutes the public good.46 This aligns with the restriction on deliberation that what
citizens’ want must be supported by reasons; deliberation entails making objective claims in
order to persuade other people.47
Henry Richardson explains how it is possible for democratic ratification to provide a
reason to regard a law as authoritative without providing a necessary and sufficient reason with
reference to the effect of wide scope normative operators: the “ought” of the proposition that “we
ought to treat a law that is the outcome of democratic deliberation as legitimate” comes outside
of a conditional.48 For instance, if we adopted the normative requirement that “[w]e ought, if the
duly-elected legislature has enacted a law on the basis of properly majoritarian procedures, to
accept that law as legitimate,” then acceptance of the wide-scope normative requirement does not
commit us to the conclusion that the law is legitimate.49 (That is to say, we could deny the
premise of the conditional statement.) On this account, individuals must make their own
decisions about the political justification of laws, but it is possible that the decision by the
democratic society to follow a particular course of action will be a reason for them to think that
that decision is legitimate.
4. What is Legitimate Authority?
While many theories of deliberative democracy treat democratic ratification as giving
some normative force to laws, it is often not fully clear what exactly this normative force is, and
theorists that focus on democratic deliberation often leave questions of legitimate authority
46
Henry S. Richardson, Democratic Intentions, in DELIBERATIVE DEMOCRACY: ESSAYS ON
REASON AND POLITICS, supra note 12, at 349, 349; see supra note 22.
47
See Sunstein, supra note 2, at 206.
48
Richardson, supra note 16, at 11.
49
Id. at 16. Richardson does not advance a comprehensive case for accepting this principle, and
does not necessarily defend it. It is simply an example of what a political principle regarding
political legitimacy of democratically made laws might look like.
Gingerich
16
under-theorized. One possible account of what sort of normative force democratic ratification
generates is that it makes laws authoritative in the sense that Joseph Raz uses the term
“authority.” The following illustration of Razian authority is simply one way that democratic
deliberation might generate normative force. This account is largely cabined in contemporary
legal positivism and particularly in the thesis that “what is law and what is not is largely a matter
of social fact.”50 Additionally, it is important that “it is an essential feature of law that it claims
legitimate authority.”51 Part of what makes law, law is its claim to have legitimate authority.
On this account, law is authoritative “if the existence of a law requiring a certain action is
a protected reason for performing that action; i.e. a law is authoritative if its existence is a reason
for conforming action and for excluding conflicting considerations.”52 To unpackage this a bit, a
law is authoritative if (1) it is, by its very existence, a reason for a person (say, A) to do
something (say, to ) and (2) it is, by its very existence, a reason to disregard reasons against
taking that action. To be authoritative, a law that says, “let A ” must be a reason for A to . It
cannot only be that A has a reason to  because of penalties for not ing, or because of social
approbation for not following the law, or for any other reason other than the law’s existence
(thought these might also be reasons for A to ).
Additionally, however, an authoritative law must be a reason to disregard reasons that A
might have for not-ing. As Raz says, “[t]he law’s claim to legitimate authority is not merely a
claim that legal rules are reasons. It includes the claim that they are exclusionary reasons for
disregarding reasons for non-conformity.”53 This means that the law has to be a reason for A not
even to consider some reason (say, P) for not-ing. This does not mean that A cannot not-, but
50
JOSEPH RAZ, THE AUTHORITY OF LAW 37 (1979).
Id. at 30.
52
Id. at 29.
53
Id. at 31.
51
Gingerich
17
that A cannot not- on account of P. This is compatible with not-ing for some other reason. It
follows from this that law must claim to be legitimate, because it claims to be a real reason to tip
the balance of reasons towards acting in conformity with it. If the law lacks legitimate moral
authority, then it cannot be a protected reason that tips the balance of reasons.
In order for a law to be legitimate, it must have some justification. This could be
provided in a number of ways. One might formulate political principles that justify laws (e.g.
political equality, Rawls’s two principles of justice, etc.). One might also formulate procedural
principles that say that a law generated by following a particular procedure is legitimate. Either
way, there must be a reason of some sort for a law to be legitimate, and it must not be overridden
by other reasons that the law should not be regarded as legitimate. This Razian story illustrates
how normative requirements generated by democratic deliberation might function, but they
might also function in other ways. (For instance, they might provide a reason for A to  without
providing an exclusionary reason.) I will use this story as an example of how normative force
generated by democratic deliberation might work in instances when theories of democratic
deliberation do not provide a theoretical account of this process.
C. Fairness and Deliberative Democracy
Most theories of deliberative democracy claim in some fashion that democratic
deliberation justifies treating laws that it generates as legitimate because it is procedurally fair.
Though these arguments could, in principle, be coherentist accounts, they tend in actuality to be
part of mixed accounts, and since I take that to be more persuasive than the coherentist account, I
will address them in their mixed theory formulation.
1. Deliberative Democracy and the Value of Fairness
Gingerich
18
Many of the basic formulations that theorists offer of deliberative democracy in some
way incorporate considerations of the equality of participants in democratic deliberation and the
fairness of the process.54 These basic commitments are specified very differently in different
theories. Some theories impose comprehensive requirements of civic education and welfare55;
others merely require that everyone be allowed one vote and that no restrictions be placed on
individuals’ ability to use their own money to pay for political advertisements.56
Regardless of how, precisely, the fairness requirement is formulated, it is central to the
claim that outcomes of democratic deliberation have normative authority. Joshua Cohen’s
account of deliberative democracy is one example of the importance of fairness in the
deliberative process to its legitimating function. For Cohen, in a just society, political
opportunities and powers must be available equally to everyone without reference to economic
or social position—they must be “manifestly fair.”57 This manifest fairness is explicitly tied to
the legitimacy of laws: “[t]he fundamental idea of democratic, political legitimacy is that the
authorization to exercise state power must arise from the collective decisions of the equal
members of a society who are governed by that power.”58 If democratic deliberation could not
ensure that its outcomes were the collective decisions that equally represented all the affected
members of a society, it would not only fail to legitimate laws or state action, but it would be a
less legitimate system than any other system that better represented all members of society.
See, e.g., Cohen, supra note 30, at 412 (“The deliberative conception of democracy is
organized around an ideal of political justification. According to this ideal, to justify the exercise
of collective political power is to proceed on the basis of a free public reasoning among
equals.”).
55
See, e.g., GUTMANN & THOMPSON, DEMOCRACY AND DISAGREEMENT, supra note 12.
56
See, e.g., Buckley v. Valeo, 424 U.S. 1, # (1976) (holding that spending money on a political
campaign is a form of free speech protected by the First Amendment).
57
Cohen, supra note 22, at 69.
58
Joshua Cohen, Democracy and Liberty, in DELIBERATIVE DEMOCRACY 185, # (Jon Elster, ed.,
1998).
54
Gingerich
19
Cohen goes further to explain how it can be that this requirement of rough equality in
deliberative democracy can justify legitimacy: deliberative democracy requires that every person
governed by a collective decision must find that the political bases of the decision acceptable,
even if they disagree with the details of the decision.59 This is an outgrowth of the deliberation
requirement of deliberative democracy, which requires that citizens of a democracy engage in
public reasoning designed to advance reasons that could be accepted by anyone.
Similarly, in Amy Gutmann and Dennis Thompson’s view, democratic deliberation is a
reason-giving activity aimed at “free and equal persons seeking fair terms of cooperation.”60
This conception of fairness is connected with strong claims that “[d]eliberation is the most
appropriate way for citizens collectively to resolve their moral disagreements not only about
policies but also about processes by which policies should be adopted.”61 The foundational
importance of fairness and equality to Gutmann and Thompson’s account makes this claim about
the appropriateness of deliberation hinge on the ability of deliberative democracy really to
deliver on its promise of political equality.62
James Fishkin emphasizes a similar importance for fairness in deliberative democracy.
For Fishkin, the four principle values of democracy are deliberation, non-tyranny, political
equality, and participation.63 Fishkin argues that the reason that his “Deliberative Polls” are
“representative of the public the people would become if everyone had a comparable opportunity
to become more like ideal citizens and discuss the issues face to face with other voters and
59
Id. at 222.
GUTMANN & THOMPSON, WHY DELIBERATIVE DEMOCRACY?, supra note 12, at 3.
61
GUTMANN & THOMPSON, DEMOCRACY AND DISAGREEMENT, supra note 12, at 4.
62
Additionally, Gutmann and Thompson are committed to a practical account of decisions about
moral issues. See id. at 5 (“In our use of this method, the principles operate in the middle range
of abstraction, between foundational principles and institutional rules; and the judgments apply
as much to particular decisions and policies as to the basic structures of society.”).
63
JAMES S. FISHKIN, THE VOICE OF THE PEOPLE 173 (1994).
60
Gingerich
20
political leaders”64 is that it offers a representation of a democracy that embodies the four values
he identifies.65 For Fishkin, as for Cohen and Gutmann and Thompson, the deliberation’s
capacity to ensure a fair and equitable distribution of political power is central to what makes
deliberative democracy legitimate; it is also a reason to support his and Ackerman’s concept of
“Deliberation Day.”66
Cass Sunstein’s description of a democratic system also requires that a “democratic
system must be built on various safeguards to ensure that its decisions are in fact a reflection of
deliberative processes” that can be “favorabl[y] characteriz[ed].”67 The preconditions of political
sovereignty are violated if the government power is not justified by reasons given in a
discussion.68 This means that the problems that “such processes are distorted by the fact that
some groups are more organized than others, by disparities in wealth and influence, and by
public and private coercion of various kinds” must be overcome.69 In his account of the
American Constitution, then, such problems must be overcome so that the constitutional system
can “ensure discussion and debate among people who are differently situated, in a process
64
Id. at 163.
Id. at 173 (“With a deliberative atmosphere of mutual respect, tyranny of the majority is
unlikely. When all the citizens are effectively motivated to think through the issues, when each
citizen’s voices count equally, and when every member of the microcosm participates, the other
three values are realized as well.”). But see Henry S. Richardson, Public Opinion and Popular
Will, in DELIBERATIVE DEMOCRACY: THEORY AND PRACTICE (David Kahane, Dominique Leydet,
Daniel Weinstock & Melissa Williams, eds., forthcoming) # (arguing that that deliberative polls
are not good proxies for actual democratic deliberation because authority not simply to discuss
but actually to make decisions is central to the idea of democratic deliberation).
66
See FISHKIN & ACKERMAN, DELIBERATION DAY, supra note 1, at #.
67
Sunstein, supra note 2, at 209.
68
CASS R. SUNSTEIN, REPUBLIC.COM 2.0, at 40 (2007).
69
Sunstein, supra note 2, at 209.
65
Gingerich
21
through which reflection will encourage the emergence of general truths.”70 For Sunstein, it
seems, democratic deliberation can generate “general truths,”71 but it can do so only if the
requisite requirements of deliberative fairness and impartiality are satisfied. Sunstein also argues
that citizens of democracies must make certain that “deliberative forces should prevail over the
arbitrary,” implying that the outcomes of democratic deliberation must, in some important way,
be non-arbitrary.72
Roberto Gargarella, who thinks that democratic deliberation can help political systems to
make impartial decisions and believes that the “Founding Fathers” of the United States thought
the same thing, further elucidates the mechanisms through which deliberation might work to
increase impartiality (and fairness) in democratic decision-making.73 First, it can decrease
ignorance about the actual preferences or interests of other individuals. Second, it can force
people to filter out arguments that are merely self-interested as they engage in the process of
public reasoning. Third, it can educate people to act impartially by training people to exchange
opinions and better listen to other people’s arguments. Fourth, it can help each participant in
deliberation better formulate her own position (e.g., resolving transitivitiy problems in one’s
preference ordering, correcting erroneous factual beliefs, etc.).74 These seem to be the
mechanisms at play in most accounts of deliberative democracy that rely on its ability to
generate equality or fairness to justify its outputs as legitimate and authoritative. 75
70
CASS R. SUNSTEIN, THE PARTIAL CONSTITUTION, 253 (1993); see also SUNSTEIN, supra note
68, at xii (“Democracy . . . is at risk whenever people sort themselves into enclaves in which
their own views and commitments are constantly reaffirmed.”).
71
Presumably, these truths are context-independent truths.
72
SUNSTEIN, supra note 68, at 43.
73
See Roberto Gargarella, Full Representation, Deliberation, and Impartiality, in DELIBERATIVE
DEMOCRACY 260, # (Jon Elster, ed., 1998).
74
See Gargarella, supra note 73, at 261.
75
See supra note 4.
Gingerich
22
The proposition that proper democratic deliberation is fair and impartial is, therefore,
central to the belief that ratification by democratic deliberation can provide a reason that a law is
legitimate. In order for the foregoing theories of deliberative democracy to succeed in their
justificatory projects, the deliberative processes must at least be capable of producing the sort of
impartiality and equality that they envision.76 I will examine this proposition in some detail in
Parts III and IV of this essay.
2. Pragmatic Theories
The foregoing accounts of deliberative democracy require significant commitment to
political principles of equality, and tend to be committed to deliberation as a justificatory activity
only in a society that looks substantially different than contemporary North American society.
There are, however, more pragmatic theories of deliberative democracy that are willing to grant
that democratic deliberation can generate legal authority without such radical moves toward
greater political fairness and impartiality. For instance, on Mark Warren’s account, democratic
authority is “generated neither by the authoritative status of expertise or beliefs as such nor by
the authoritative status of rules and procedures; instead it comes from a set of institutionalized
protections and securities within which the generative force of discursive challenge is
possible.”77 On this account, the discursive challenge need not be exercised in order for a law to
be authorized; the possibility of discursive challenge is sufficient. Theories such as this may
provide more accurate descriptions of how activists or revolutionaries might think of a society in
76
I say that they only need to be capable of doing so, not that they must do so in actuality,
because many of them envision a world substantially different from the status quo in which
deliberation should take place. They must, however, be capable of producing the sort of
impartiality and fairness that they envision with procedural changes to the status quo (and
perhaps limited substantive changes), since all of the theories that concern me purport to be
practical, rather than merely ideal, theories of democratic deliberation.
77
Mark E. Warren, Deliberative Democracy and Authority, 90 AM. POL. SCI. REV. 46, 57 (1996).
Gingerich
23
which they are able to disrupt decision-making practices by issuing discursive challenges to
majority proscriptions. They might see a deliberative democracy as a society that gives room to
critical consciousness, rather than one that has deliberative procedures that are per se fair and
impartial. I will not focus on these theories in my press on deliberative democracy theories, and
I think that such accounts of deliberation might offer a fruitful alternative to accounts of
deliberative democracy that make hopelessly idealistic assumptions about the possibility of
political impartiality.
3. Alternative Justifications of Democratic Deliberation
There are justifications of deliberative democratic authority within a mixed theory
framework other than that democratic deliberation generates political legitimacy. For instance,
James Fearon identifies six reasons that deliberation is desirable:
1. Reveal private information
2. Lessen or overcome the impact of bounded rationality
3. Force or encourage a particular mode of justifying demands or claims
4. Help render the ultimate choice legitimate in the eyes of the group, so as to contribute
to group solidarity or to improve the likely implementation of the decision
5. Improve the moral or intellectual qualities of the participants
6. Do the “right thing,” independent of the consequences of discussion78
Diego Gambetta offers four reasons:
7. It render the outcomes of the decisions Pareto-superior by fostering better solutions
[since it will introduce more imagination into discussions];
8. It can make the outcome fairer in terms of distributive justice by providing better
protection for weaker parties;
9. It can lead to a larger consensus on any one decision
78
James D. Fearon, Deliberation as Discussion, in DELIBERATIVE DEMOCRACY 44, 44-45 (Jon
Elster ed., 1998).
Gingerich
24
10. It can generate decisions that are more legitimate (including for the minority). 79
Most of the reasons identified by Fearon and Gambetta are independent of equality, fairness and
legitimacy; they have to do with the decisions made by deliberative institutions being simply
better by objective criteria. But both Fearon and Gambetta link these prudential considerations
to legitimacy: when a political system typically generates good outcomes, the fact that it
generates an outcome might be a reason to suppose that the outcome should be regarded as
legitimate.80
The existence of these reasons indicates that deliberative democracy could give authority
to its ratified decisions by virtue of its ability to usually generate good outcomes. Setting aside
for the moment whether it actually tends to produce good outcomes, these theories do not
actually describe a situation where a law can have authority simply by virtue of being the
outcome of democratic deliberation. A law might have a higher probability of being legitimate,
but that is all. On this account of legitimacy, except to the extent that fairness is involved,
recourse to external principles of political justice is always necessary to explain why deliberation
is a reason for a law to have authority, so I take this to simply be a variant of the objectivist
approach to democratic deliberation, and I will bracket it for the remainder of this essay.
III. THE MOTIVATIONAL STRUCTURE OF DELIBERATION81
Diego Gambetta, “Claro!”: An Essay on Discursive Machismo, in DELIBERATIVE
DEMOCRACY 19, 24 (Jon Elster, ed., 1998).
80
Of course, these theorists also recognize that there are times when deliberation does not lead to
better decisions. For instance, under certain circumstances, deliberation can waste precious time,
lead people to be duped by sophists, allow self-interested lobbies to manipulate information in a
debate, or make choice indeterminate. Gambetta, supra note 79, at 21. So these theorists do not
attempt to claim that deliberation is always good. Nor, for that matter, do any other deliberative
democracy theorists of whom I know.
81
In this section, I examine very concrete empirical evidence, which I then apply to fairly
abstract political theory. Following Kwame Anthony Appiah, I do not claim that the empirical
data can resolve normative questions, but that it can usefully inform abstract theory. See
generally KWAME ANTHONY APPIAH, EXPERIMENTS IN ETHICS (2008) (exploring how
79
Gingerich
25
David Schkade, Cass Sunstein, and Reid Hastie conducted a study in which they created
an experimental Deliberation Day modeled on James Fishkin’s Deliberative Polls.82 They
assembled five person groups in Boulder, Colorado, a predominately liberal city, and Colorado
Springs, a predominantly conservative city. They asked the deliberators their individual views on
global warming, affirmative action, and civil unions for same-sex couples. Then they had them
deliberate in five person groups about those issues and reach a decision as a group.83
Subsequently they were again asked their individual opinion on the issues. Schkade et al. found
that the groups from Boulder became more liberal after deliberation and those from Colorado
Springs became more conservative, so deliberation increased extremism.84 Second, every group
moved towards a consensus of individual opinions and diversity of individual opinions among
group members decreased, even in the anonymous post-deliberation expression of individual
views.85 Third, deliberation increased the political gap between the predominantly liberal
residents of Boulder and the predominantly conservative residents of Colorado Springs.86 What
happened, Schkade et al. concluded, was that “deliberation among like-minded people produced
ideological amplification—an amplification of preexisting tendencies, in which group discussion
leads to greater extremism.”87
These outcomes contrast with the results that James Fishkin obtained from deliberative
polls conducted in England, Australia, and the United States. His data showed that individual
opinions changed significantly, indicating that deliberation has a large effect, but he did not find
experimental psychology might influence how moral philosophers construct normative
arguments).
82
Schkade, Sunstein, & Hastie, supra note 14.
83
Id. at 918-20.
84
Id. at 918-19.
85
Id. at 921.
86
Id. at 922-23.
87
Id. at 917.
Gingerich
26
any significant evidence of systematic ideological amplification.88 Schkade et al. noted several
important factors that they think accounted for the difference between their data and Fishkin’s
deliberative polling data. First, Fishkin’s groups had moderators, who were charged with
ensuring openness of the deliberation and altering some of the dynamics of deliberation, while
Schkade et al.’s groups did not have moderators.89 Second, Fishkin’s polls presented
deliberators with a set of written materials that “attempted to be balanced and that contained
detailed arguments supporting sides.”90 Third, Fishkin instructed participants in the deliberative
polls not to reach a group decision, which may have limited the effects observed in Schkade et
al.’s study.91
If Schkade et al.’s explanation for the difference between their data and Fishkin’s data is
right, which seems to at least be plausible, then if an account of deliberation entails citizen
discussions in a forum that looks anything like Fishkin’s Deliberation Day, or even if it just
envisions occasional discussions about policies among citizens in somewhat formal institutional
settings, deciding whether to provide written materials and whether to have a moderator will
have important political consequences. Although it is hard to know in advance how any group of
deliberators will react to a set of purportedly balanced materials, one possibility is that
deliberation will reduce the tendency toward ideological amplification. There is nothing
inherently good or bad about ideological amplification, so if you want more ideological
amplification (say, if your views are toward one end of the political spectrum) you would want to
avoid moderators and written materials in your ideal Deliberation Day, but if you want less
88
Id. at 934.
Id.
90
Id. at 934-35.
91
Id. at 935.
89
Gingerich
27
ideological amplification (say, your political opinion is right at the center of the political
spectrum) then you would want impartial moderators and “unbiased” written materials.
In this part of this essay, I will attempt to show that this is just one of many political
decisions to be made about the structure of deliberation. Different deliberative structures result in
different deliberative outcomes. In III-A, I will set out the empirical data that show that different
deliberative structures produce different outcomes, and in III-B, I will argue that an important
(though not the only reason) that different deliberative structures generate different outcomes is
that they pull different motivational levers and thereby shape and transform people’s preferences
and political opinions. In III-C, I will argue that because of the effects described in III-A and IIIB, decisions about how to structure deliberation always have political stakes, and will always
have an impact (often a predictable one) on the outcome of deliberation. In order to win my
weak argument, I only need to show that in practice, deliberative institutions always distort
preferences (which is my argument in III-A). 92 In order to win my strong argument, I also need
to show that deliberation is not only distorting preferences and beliefs, in which case (at least in
theory if not in practice) a deliberative structure might avoid distorting those preferences and
beliefs, but that deliberation determines what people’s preferences are by motivating people to
change them, in which case, there is no neutral baseline of preferences and beliefs for
deliberation to avoid distorting (which is my argument in III-B and III-C). 93
A. System Design Distorts the Neutrality of Deliberation
92
My weak argument is that deliberative democratic ratification does not give a law normative
force in the status quo. See Part IV-A, infra.
93
My strong argument is that deliberative democratic ratification cannot give a law normative
force. See Part IV-B, infra.
Gingerich
28
Generally, the impact of deliberation on group decision-making about politics is highly
context dependent.94 How it affects individual opinion will depend on the purpose of
deliberation, the subject of deliberation, the participants, how deliberation is connected to
authoritative decision makers or how much authority it has, what govern how deliberators
interact with each other, what information is provided to the deliberators, the prior beliefs of the
deliberators, and the material conditions of the society in which the deliberation takes place.95
This extreme context dependence indicates that choices about what forms democratic
deliberation should take will have a big impact on the outcomes of deliberation.
Disregarding context, when it comes to judgments for which there is a right answer,
group judgments end up about as accurate as the mean judgment of their members, more
accurate than the judgment of typical individuals, and less accurate than the judgments of the
most accurate individuals.96 Of course, deliberating about normative political questions is not
the same as making judgments about questions that have right/wrong answers, but if there is an
objective balance of reasons about what, politically, should be done (as any non-coherentist
model of democratic deliberation must suppose), then it would make sense that to the extent that
political deliberation hinges on causal questions like “does pornography cause rape?” or “which
law best reduces transaction costs?,” we can expect the same distortions that affect non-
94
See Michael X. Delli Carpini, Fay Lomax Cook, and Lawrence R. Jacobs, Public Deliberation,
Discursive Participation, and Citizen Engagement: A Review of the Empirical Literature, 7 ANN.
REV. POL. SCI. 315, 336 (2004) [PARENTHETICAL].
95
See id.
96
Daniel Gigone & Reid Hastie, Proper Analysis of the Accuracy of Group Judgments, 121
PSYCHOL. BULL. 149, 153 (1997) (surveying empirical studies on group judgments across
contexts).
Gingerich
29
normative reasoning to influence political deliberation.97 This means that political judgment
might be made more or less accurate by a variety of mechanisms at play in group judgment.
Generally it is also possible to say that when individuals are biased in a particular way
because of a widely held belief system or heuristic, “group interaction will enhance the bias,”
while if the underlying beliefs are not widely shared and deliberative groups contain at least one
member who does not share the basis, group interaction lets “more accurate members . . .
persuade (or correct) less accurate members.”98 Depending on the background belief systems of
group members, group decision making is likely to either increase or decrease individually held
biases. The impact of group deliberation on individual bias also follows certain patterns.99
Particularly: (1) as group size increases, the direction of bias is unchanged but the magnitude of
the bias increases, (2) the direction and magnitude of group bias can change the magnitude of the
bias of individuals in the group, but not the direction of individual bias, and (3) different group
processes can produce radically different group biases.100
1. Homogenization
Internal group diversity is, at least in some instances, decreased by deliberation. Solomon
Asch conduced an experiment in which he showed two cards to subjects, the first card showing a
standard line and the second showing thee lines, one of which was obviously the same length as
the standard and asked the subjects which lines were the same length. The subjects were placed
97
Thanks to Terry Fisher for suggesting this point.
Garold Stasser & Beth Dietz-Uhler, Collective Choice, Judgment and Problem Solving, in
BLACKWELL HANDBOOK OF SOCIAL PSYCHOLOGY: GROUP PROCESSES, supra note 13, at 31, 49.
99
Norbert L. Kerr, Geoffrey P. Kramer, & Robert J. MacCoun, Bias in Judgment: Comparing
Individuals and Groups, 103 PSYCHOL. REV. 687, # (1996).
100
Id. at 713.
98
Gingerich
30
in rooms with a set of confederates instructed to give the wrong answer at certain times. 101 Asch
found that when a subject was placed with a single confederate who gave the wrong answer, the
subject continued to answer correctly in almost all trials, but when two confederates gave the
wrong answer, the subject went along with them 13.6 percent of the time, and when three
confederates all gave the wrong answer, the subject also gave the wrong answer in 31.8 percent
of trails.102
Muzafer Sherif conducted a similar experiment in which he placed subjects in a darkened
room with a pinpoint of light and instructed them to gaze at a pinpoint of light. After a few
minutes seemed to “move” because of a perceptual illusion called the “autokinetic effect.” He
then asked the subjects to estimate how far the light had moved. Sherif found that when the trial
was conducted with individuals, the answers that they gave were highly variable from person to
person and trial to trail. 103 But when he conduced the experiment with groups of two or three
people “the subjects’ estimates invariably began to converge and a group norm quickly
developed.104 Sherif then introduced a confederate with one naïve subject and had the
confederate give estimates of the light’s movement that were consistently much higher or lower
than the subject’s, and found that the subject quickly adopted the standard set by the
confederate.105
101
Solomon E. Asch, Opinions and Social Pressure, in READINGS ABOUT THE SOCIAL ANIMAL
13, # (Eliot Aronson, ed., 7th ed., 1995).
102
Id. at 17. Asch also found that the presence of a single confederate who gave the correct
answer greatly decreased the pressure of the incorrect majority (it made the subjects four times
less likely to give the wrong answer) and increasing the incorrect majority beyond three did not
increase the subjects’ error rate.
103
LEE ROSS & RICHARD E. NISBETT, THE PERSON AND THE SITUATION: PERSPECTIVES OF SOCIAL
PSYCHOLOGY 28-29 (1991).
104
Id. at 29.
105
Id.
Gingerich
31
These studies are not about deliberation over norms and policies, but they might still have
important implications for political deliberation. Many political decisions rely on causal claims
(Will a certain policy decrease or increase the GDP? Will a certain criminal law actually
decrease the prohibited behavior’s frequency?) These claims often, perhaps usually, do not have
clear answers, so individuals who do not care about a policy except in light of, say, its impact on
the GDP may not feel very strongly about the policy itself and may be open to other group
members’ causal arguments. The Asch study indicates that the size of deliberating groups might
impact their outcomes, since individuals are more likely to go along with other people say if they
are in a big group and everyone disagrees with them. The Sherif study indicates that when
people are highly uncertain about a claim, they tend to be quite willing to go along with what
other people tell them, so if people deliberate about a complicated policy, the causal
consequences of which they cannot accurately evaluate, they may be likely to simply agree with
what other people say, even if other people do not have any special expertise or information
about the policy that they lack.
A study by Sunstein, Schkade, and Ellman bears out that homogeneity effects can crop up
in policy deliberations: they found a tendency toward “collegial concurrences” on panels of
American federal circuit court judges. When judges are ideologically predisposed toward
reaching a particular opinion but two other judges on their panel reach the opposite opinion, they
are less likely to write a dissent than ideological data alone would predict.106 The study also
found that on highly ideologically charged issues, like abortion and capital punishment, there
was no such tendency toward collegial concurrences.107 This substantiates the findings of the
106
See Cass R. Sunstein, David Schkade, and Lisa Michelle Ellman, Ideological Voting on
Federal Courts of Appeals: A Preliminary Analysis, 90 VA. L. REV. 301, 337 (2004).
107
See id. at 339.
Gingerich
32
Sherif study in a political context—less firmly held opinions are more likely to be swayed by the
opinions of other group members than are more firmly held ones.
Similarly, a study of the German Mediation Committee (Vermittlunsausschuss) found
that a high level of discourse on a given issue correlated with a high probability of the Mediation
Committee reaching a unanimous decision about the issue.108 This study also found that appeals
to the common good framed in egalitarian terms enhanced the likelihood of egalitarian outcomes
(although the effect was only marginally significant).109 This study showed that better and more
extensive deliberation on a given political issue led a political body (composed of representative
of the parties represented in the German Bundestag and Bundesrat) to issue more unanimous
rulings. So designs of deliberative institutions that allow for more discussion might be more
likely to produce unanimous results (and hence greater group homogeneity).
2. Polarization
Empirical research also supports the point made by Schkade et al. that deliberation about
politics can increase group polarization.110 A study of French students in 1969 by Serge
Moscovici and Marisa Zavalloni found that discussion to unanimity in a small group setting led
to more polarized attitudes toward De Gaulle and toward Americans.111 The study also found
that individuals tended to adopt the opinions endorsed by the group consensus as their own
108
JÜRG STEINER, ANDRÉ BÄCHTIGER, MARKUS SPÖRNDLI & MARCO R. STEENBERGEN,
DELIBERATIVE POLITICS IN ACTION: ANALYZING PARLIAMENTARY DISCOURSE 162-63 (2004).
The Mediation Committee is a body that mediates conflicts between the Federal Diet
(Bundestag) and the Federal Council (Bundesrat). Its objective is to bridge the differences
between the Diet and the Council and to reconcile conflicting interests. Id. at 140-41. This study
also found no link between the quality of political discourse and the substantive outcomes of
political deliberation. Id. at 157.
109
Id. at 163.
110
Schkade, Sunstein, & Hastie, supra note 14, at 922-23.
111
Serge Moscovici & Marisa Zavalloni, The Group as a Polarizer of Attitudes, 12 J.
PERSONALITY & SOC. PSYCHOL. 125, 130-32 (1969).
Gingerich
33
opinion.112 Deliberating about political issues in small groups can increase extremism and group
polarization.
Research has shown that ideological dampening and amplification also happens on panels
of circuit court judges.113 The Sunstein, Schkade, and Ellman study found that circuit court
judges appointed by a Democrat, when sitting with one Republican appointee and one
Democratic appointee, voted in a stereotypically liberal fashion 51 percent of the time. When
sitting with two Republican appointees, however, a Democratic appointee cast liberal votes in
only 45 percent of cases, and when sitting with two Democratic appointees, cast liberal votes 63
percent of the time. A Republican appointee, when sitting with one Republican and one
Democratic appointee, cast liberal votes in 35 percent of cases. A Republican appointees cast
liberal votes only 30 percent of the time when sitting with two other Republican appointees and
45 percent of the time when sitting with two Democratic appointees.114 Creation of groups of
liberals and conservatives, either intentionally or inadvertently, may lead the opinions of such
groups to polarize away from each other toward political extremes.115 Deliberative structures
112
Id. at 134.
Sunstein, Schkade, & Ellman, supra note 106, at 305.
114
Id. at 316-17.
115
These accounts of group polarization do not seem to account for cases of what Duncan
Kennedy describes as “loopification.” See Duncan Kennedy, The Stages of the Decline of the
Public/Private Distinction, 130 U. PA. L. REV. 1349, 1354 (1982).
One's consciousness is loopified when the ends of the continuum seem closer to one
another, in some moods (for some purposes, in some cases), than either end seems to the
middle. Otherwise stated, one's consciousness is loopified when one seems to be able to
move by a steady series of steps around the whole distinction, ending up where one
started without ever reversing direction. Like wow, man.
Id. Certain ideological issues, particularly extreme political positions might loop from one end
of the spectrum to the other, in which cases, it would not make much sense to describe the
political views generated as polarized. In some cases, this might mean that “group polarization”
can generate political consensus from two sides of the spectrum. An example might be
protectionist policies favored by both cultural conservatives and pro-union liberals. It is
113
Gingerich
34
that favor or permit such ideological groupings will, then, have a predictable impact on the
political outcomes of deliberation.
Another study shows that less institutionally bounded and clearly delimited can make
political attitudes more extreme. Theodore Newcomb studied the political preferences and
voting habits of students at Bennington College in the late 1930s.116 Newcomb found that when
they entered college, the students had predominantly conservative Republican political views and
voting preferences, and reflected the views and preferences of their parents. After two years at
Bennington, the student’s views were far to the left of those of their parents and other Americans
of their social background.117 Newcomb also found that there was a high degree of social
solidarity among the Bennington students, a high degree of exposure to political discourse, that
Bennington was largely isolated from the surrounding community.118 Group membership, group
identity, and structures of informal deliberation might be expected to play a substantial role in
shaping individual beliefs and political preferences.
Sunstein makes the important point that while group deliberation leads to group
polarization, it is not at all apparent that this is bad or good. Enclave deliberation, or deliberation
within small groups of like-minded citizens, may produce more polarized political opinion.119 It
may undermine social stability, promote social fragmentation, and threaten security.120 At the
same time, it may be the only way to ensure that the views of low-status groups and individuals
are heard and developed, because large heterogeneous groups often give less weight to the
instructive to keep in mind that political extremes are “extreme” only in comparison to the
background political culture.
116
Ross & Nisbett, supra note 103, at 35-6.
117
See id. at 35.
118
See id. at 36-7.
119
SUNSTEIN, supra note 12, at 15-16.
120
See id. at 15.
Gingerich
35
opinions of low-status members, including (in some contexts) women, racial minorities, and
poorly educated people.121 So whether group polarization is a good thing depends on the context
and it depends on what groups you want to have political power.
3. Extremism
Some effects of group deliberation are better described as generating more extreme
positions rather than simply generating group polarization. Particularly, the ideological
amplification identified by Schkade et al. as a possible outcome of deliberation may operate even
on groups that are not conceived primarily as in opposition to other ideologically aligned
groups.122
For example, Robert MacCoun examined the impact on mock criminal juries of extraevidentiary biases, like defendant unattractiveness or pretrial publicity biased against the
defendant. MacCoun found that when the bias was weak and the prosecution’s case was also
weak, the jury was more likely to return a fair outcome than an individual juror, “[b]ut when
antidefendant bias is strong, juries are considerably more biased than [individual] jurors.” 123 So
decisions about how to structure deliberation will have predictable impacts on whether and what
biases individuals hold.
Another instance of group extremism is jury deliberation about damages. A study by
Schkade, Sunstein, and Kahneman found that once juries have decided to award punitive
damages, “deliberation causes awards to increase, and it causes high awards to increase a great
121
See id.
See Schkade, Sunstein, & Hastie, supra note 14, at #.
123
Robert J. MacCoun, Comparing Micro and Macro Rationality, in JUDGMENTS, DECISIONS,
AND PUBLIC POLICY 116, 128 (Rajeev Gowda & Jeffrey C. Fox eds., 2002).
122
Gingerich
36
deal.”124 In some cases, post deliberation jury verdicts were even higher than the highest
individual pre-deliberation verdict.125 This study found that small awards increased as a result of
deliberation as well as large awards.126 The authors explained this “severity shift” toward larger
verdicts as the result of a “rhetorical asymmetry”: after the jury agreed on some award larger
than zero, “the arguments for the larger award have a rhetorical advantage and are more
persuasive.”127 Deliberation in certain contexts, might systematically privilege certain political
positions for which particularly rhetorically compelling arguments exist and are easy to deploy.
Group deliberation also predictably impacts judgments about risk. Studies of risk
assessment have shown that individual and group risk tolerance can be either heightened or
decreased by group discussion.128 A study by James Stoner found that when people were asked
to deliberate about risk in groups, items for which widely held values favored the risky
alternative, “unanimous group decisions were more risky than the average of individual
decisions.”129 On the other hand, for items for which widely held values favored caution, group
decisions were more cautious than the average of individual decisions.130 Group deliberation and
decision making can, therefore, generate risk tolerances that differ from the mean of individual
risk tolerances. Many political decisions (particularly environmental and public health
124
David Schkade, Cass R. Sunstein & Daniel Kahneman, Debating About Dollars: The Severity
Shift, 100 COLUM. L. REV. 1139, 1155 (2000).
125
Id.
126
Id.
127
Id. at 1161.
128
SUNSTEIN, supra note 12, at 25.
129
James A. F. Stoner, Risky and Cautious Shifts in Group Decisions: The Influence of Widely
Held Values, 4 J. EXPERIMENTAL SOC. PSYCHOL. 442, # (1968). For instance, “A football captain
must choose between gaining a tie on the last play of the game or attempting a play which will
bring either victory or defeat.” Id. at 447.
130
Id. For instance, “A couple must choose between allowing a complicated pregnancy to
continue, with danger to the mother’s life, or having the pregnancy terminated.” Id. at 447.
Gingerich
37
decisions) require deciding how much risk to tolerate, so this effect might impact the policy
choices that democratic deliberation makes.
4. Error Rates
It cannot be shown that deliberating groups generally arrive at the truth; groups often
aggregate information poorly. Sometimes groups do better than their individual members, but
sometimes they do worse.131 If groups make errors about facts, they might also make errors
about norms. For instance, certain types of decision errors limit the efficacy of majority rule as
an aggregator of individual information and amplify individual error rates.132 Groups are also
more likely than individuals to escalate commitment to a cause of action that is failing.133
Glen Whyte conducted an experiment that substantiates this theoretical point in which he
asked subjects to make a risky investment (one with an expected return near zero) in a project. 134
When subjects were told that there were no “sunk costs” in the project, only 29 percent chose to
make the risky investment. When the subjects were told that there were significant “sunk costs”
in the project, 66 percent made the risky investment, and when they were informed that they
were personally responsible for the sunk costs, 72 percent made the risky investment.135 Whyte
then assembled subjects into groups, and had the groups deliberate about whether to invest.
Groups made the risky investment in the no sunk costs condition 26 percent of the time, the sunk
costs condition 77 percent of the time, and the sunk costs plus personal responsibility condition
131
CASS R. SUNSTEIN, INFOTOPIA 57-8 (2006).
See, e.g., William P. Bottom, Krishna Ladha & Gary J. Miller, Propagation of Individual Bais
Through Group Judgment: Error in the Treatment of Asymmetrically Informative Signals, 25 J.
RISK & UNCERTAINTY 147, 154-60 (showing how individual errors about Bayesian reasoning are
not corrected by majority rule).
133
Stasser & Dietz-Uhler, supra note 98, at 48.
134
Glen Whyte, Escalating Commitment in Individual and Group Decision Making: A Prospect
Theory Approach, 54 ORGANIZATIONAL BEHAV. & HUM. DECISION PROCESSES 430, # (1993).
135
Id. at 442.
132
Gingerich
38
94 percent of the time.136 Whyte also checked private individual preferences after the group
deliberation, and found that after group deliberation, individuals were more likely to make the
unwise investment: 30 percent in the no sunk costs position, 74 percent in the sunk costs
condition, and 87 percent in the sunk costs plus personal responsibility condition.137 Thus, group
decision-making significantly magnified the individual preferences for escalation in a failing
project.138 Moreover, even when a majority of a group initially favored not escalating but where
a minority favored escalation, 50 percent of the groups ultimately decided to escalate.139 (In
groups where an initial minority opposed escalation, the initial majority won out 98 percent of
the time.140) These results led Whyte to conclude that while “[g]roups may perform better than
individuals on some tasks . . . decision making in escalation situations is apparently not one of
them.”141 If group decision-making tends to amplify commitment failing projects in which one
has invested, this suggests that group decision-making might systematically increase
conservatism (in the form of commitments to settled political courses of action) and entrench
failed policies.
B. Non-Economic Levers of Motivation Explain these Distortions
One story about these effects of group deliberation says that the impact of deliberation on
outcomes just has to do with cognitive distortions. If the distortions could be eliminated,
deliberative structures might accurately reflect what people prefer or believe, and a neutral
deliberative framework could accomplish this. A different story, which I will advance here, says
136
Id.
Id. at 446.
138
Id. The study also found that when the dominant tendency among individuals was to forgo
the risky investment (the no sunk costs condition), “group decisions increased this conservative
tendency.” Id. at 447.
139
Id. at 446.
140
Id.
141
Id.
137
Gingerich
39
that deliberation changes and transforms people’s opinions rather than simply distorting them, so
there is no neutral baseline for deliberation to preserve. Specifically, a variety of motivational
processes are at play in these instances of deliberative distortion. Different designs of
deliberative frameworks will make different uses of these motivational structures and will impact
the outcomes of political deliberation in different ways. Four mechanisms that might function in
group deliberation are: communication and information availability, empathy/humanization,
solidarity with group members and desire for acceptance, and motivation crowding.142 Acting in
a manner influence by these mechanisms might enhance one’s self-interest, as measured by a
complete, transitive preference ordering. But it might not. So, since it affects these design
levers, “deliberation may lead people to hold beliefs that are not in their best interest.”143
1. Communication and Information Availability
The data described in the foregoing section does not, by itself, explain what impact
communication has on how people shape their beliefs. One possibility is that views become
homogenous because people have imperfect information about their decisions and they rely on
the decisions of other people who face similar decisions to guess what the right decision is, and
this “pervasive but fragile herd behavior” then generates informational cascades.144 When such
cascades occur, as the number of people expressing a belief or preferences increases, the number
of uncommitted decision makers who treat this information as a good basis to make a decision
142
See Yochai Benkler, Law Policy and Cooperation (forthcoming) (manuscript at 10). These
mechanisms are drawn from Benkler’s table of “Design Levers for Cooperation.”
143
Adam Przeworski, Deliberation and Ideological Domination, in DELIBERATIVE DEMOCRACY
140, 140-41 (Jon Elster ed., 1998). From the standpoint of system design, it is not at all clear in
the abstract whether it is good for people to act contrary to their self interest—this is context
dependent. Sometimes, public welfare (or some other measure of what is desirable) might be
increased when certain individuals act in a fashion contrary to their preferences.
144
Sushil Bikhchandani, David Hirschleifer & Ivo Welch, Learning from the Behavior of
Others: Conformity, Fads and Informational Cascades, 12 J. ECON. PERSP. 151, 168 (1998).
Gingerich
40
increases, creating a ongoing positive feedback effect. 145 The shift toward more extreme views
might reflect the informational bias that people become more confident of their own beliefs when
they know that other people share them. 146 These informational externalities can then distort
people’s beliefs.147
This is not, however, the whole story. It is also true, as David Salley notes, that “the act
of communicating itself influences people’s preferences.”148 Salley found that in experimental
prisoner’s dilemma settings where communication between decision-makers should not
influence the decisions that they reach on a self-interested rational actor model (where “talk is
cheap”), the opportunity to communicate was the most important factor that led participants to
act in fashion not guided by self-interest.149 Salley suggests that this may be because “meaning
arises from the mutual consideration of the utterer and the addressee, and this common
identification may spill over into actions other than speaking and listening.”150 Therefore,
communication may induce people to change their preferences so that they are better aligned
with the preferences of other people in decision-making groups. Communication does not
simply provide people with information about which decisions are most likely to lead to
outcomes they prefer, it also constitutes preferences. Deliberative institutions that promote
communication among a particular group of people may lead those people to change their
preferences so that they are more homogenous, even if they have perfect information. Whether
145
See SUNSTEIN, supra note 12, at 20.
See Schkade, Sunstein & Hastie, supra note 14.
147
See SUNSTEIN, supra note 12, at 17.
148
David Sally, Conversation and Cooperation in Social Dilemmas: A Meta-Analysis of
Experiments from 1958 to 1992, 7 RATIONALITY & SOC’Y 58, 69 (1995).
149
Id. at 78.
150
Id. at 87.
146
Gingerich
41
such an institution generates correct outputs depends on whether the median or mean predeliberation beliefs of the group on the issues about which they deliberate are correct. 151
2. Solidarity
Michael Bacharach noted “a cardinal feature of group identification is to take the goals of
the group to be your goals.”152 When people are part of groups that make decisions, they might
change their beliefs because of an intrinsic motivation to conform to the beliefs of the group if
they define themselves by their group membership. Being part of the group may become more
important to individuals than are their opinions about particular decisions that the group needs to
make. When assigned to a group, people almost reflexively think of that group as better for them
than an out-group, and do this because they are motivated to create a positive self-image.153 This
socially constituted self-identity can then affect a person’s individual identity as she tries to
ensure that who she is matches who she is supposed to be as a group member.154 Individuals
may therefore attempt to conform to group identity when a group is deliberating about how to
decide and when they have information about what deliberative position best represents the
group identity.155 If individuals have a view that they think many other people in their group
strongly dislike, they may abandon rather than advance it in deliberative processes both to
preserve their identities to the group (not wanting to lose the approbation of the group) and
preserve their self-identities (not wanting to risk their internal beliefs that their individual
Knowing whether this is true requires independently evaluating the validity of the group’s
opinions prior to deliberation.
152
MICHAEL BACHARACH, BEYOND INDIVIDUAL CHOICE 75 (2006).
153
See ROGER BROWN, SOCIAL PSYCHOLOGY: THE SECOND EDITION 551 (1986)
[PARENTHETICAL].
154
See id.; Miller McPherson, Lynn Smith-Lovin & James M. Cook, Birds of a Feather:
Homophily in Social Networks, 27 ANNU. REV. SOCIOL. 415, # (2001) (arguing that people are
more likely to cooperate and agree with people who are like themselves in some salient way).
155
See Schkade, Sunstein & Hastie, supra note 14, at #.
151
Gingerich
42
identities are congruent with their social identities).156 The flip side of this pursuit of conformity
to group identity is the exclusion or shunning of non-group members.157 This may, then,
contribute to a disregard in group decision-making of alternative possibilities that might be seen
as contrary to the identity of the group.158 An (at least partial) explanation of many of the
outcomes of the experiments described above is that people change their preferences to conform
to what they think fellow group members want.
This solidarity effect is likely to affect individuals’ preferences and beliefs differently
depending on how deliberative democratic institutions are designed. It is possible for
governments, political parties, or other interest groups to manufacture identities that support their
ideological positions.159 If such identities simply produced cognitive distortions (that is, if they
made people think that they wanted something different from what they really wanted) the threat
of political infiltration might not be too great for deliberative systems—at least in theory, a
system could be designed that would not allow for such cognitive distortions. But such identities
do not appear to simply distort preferences. Rather, they appear to create or change them.
People are always part of groups, and the groups of which they are members will have a big role
in determining what they like and value and believe.160 Because people invariably form groups
and connections with others who share salient features with themselves, no deliberative system
could ever hope to preserve the pre-group preferences of individuals (if it is even possible for
156
See SUNSTEIN, supra note 12, at 26.
See Samuel Bowles & Herbert Gintis, Persistent Parochialism: Trust and Exclusion in Ethnic
Networks, 55 J. ECON. BEHAV. & ORG. 1, 2 (2004) [PARENTHETICAL].
158
This seems particularly likely in deliberative democratic settings, since the organizations most
frequently discussed as the locus of deliberative democracy (nation-states) tend to inculcate
particularly strong (nationalistic) group identifications.
159
Susan C. Stokes, Pathologies of Deliberation, DELIBERATIVE DEMOCRACY 123, 134 (Jon
Elster ed., 1998).
160
See McPherson, Smith-Lovin & Cook, supra note 154, at #.
157
Gingerich
43
people to really have preferences outside of groups). As such, deliberative institutions must cope
with the inevitability that depending on the sorts of groups that they create, they will support the
development of certain ideological values and opinions rather than others.
3. Empathy
Even in the absence of a solidarity effect arising from a strong group identity,
deliberation may encourage participants to alter their beliefs through the motivational
mechanism of empathy.161 Iris Bohnet and Bruno Frey found that in a dictator game allowing
for one-way identification in which a potential benefactor received some information about the
potential grantee significantly increased gifts by the benefactor.162 This one-way identification
occurred in one-shot games, so cannot be explained as a reciprocity effect. Instead, it is the result
of an increased concern with the well-being of the possible recipients of a gift arising from
increased social closeness.163 When people recognize the humanity of a counterpart, they are
more likely to act in ways that increase the well-being of their counterpart.164
This empathy/humanization effect is further evidence by neurological evidence. James
K. Rilling et al. conducted fMRI scans on subjects playing prisoner’s dilemma games. This
experiment found that when subjects were introduced to a human partner before playing the
game and were told that they were playing the game with a human partner, mutually cooperative
social interactions were associated with activations of the anteroventral striatum, which
161
See Benkler, supra note 142 (manuscript at 11); Sally, supra note 148, at 69 (conversation
may create empathy among participants, including participants who are not members of the same
in-groups).
162
Iris Bohnet & Bruno S. Frey, The Sound of Silence in Prisoner’s Dilemma and Dictator
Games, 38 J. ECON. BEHAV. & ORG. 43, 53 (1999).
163
Id.
164
See Benkler, supra note 142 (manuscript at 11-12).
Gingerich
44
neuroimaging and electrophysiological evidence has linked to reward processing. 165 In contrast,
when subjects were told that they were playing the game with a computer, activations of striatal
mechanisms related to reward were not observed.166 Empathetic motivations appear to be salient
in normative decision-making, too. Norman Frohlich and Joe A. Oppenheimer conducted a set
of experiments asking subjects to endorse or reject particular principles of justice.167 They found
that when even high performing subjects were able to access information about the real effort
and income of low performers, they were happy to endorse a floor-constraint principle of justice,
requiring that the income of low performers not fall below a certain level, even when that
principle would sometimes result in a significant redistribution of income.168
These data indicate that democratic deliberators will be differentially motivated in
different institutional settings, and are more likely to adopt political and normative principles that
benefit other people with whom they can most easily empathize. Because spatial proximity
plays an important role in encouraging empathetic responses, the geographic organization of
deliberative structures may play an important role in determining which empathetic motivations
communicative processes of deliberation trigger.169 Again, this effect seems not simply to distort
preferences but rather to change them: people express strong preferences for acting
empathetically toward humanized others. So deciding how to design deliberative procedures
165
James K. Rilling, David A. Gutman, Thorsten R. Zeh, Giuseppe Pagnoni, Gregory S. Berns &
Clinton D. Kilts, A Neural Basis for Social Cooperation, 35 NEURON 395, 403 (2002)
[hereinafter Rilling et al.]; see also Alan G. Sanfey, James K. Rilling, Jessica A. Aronson, Leigh
E. Nystrom & Jonathan D. Cohen, The Neural Basis of Economic Decision-Making in the
Ultimatum Game, 300 SCIENCE 1755, 1755 (2003) [PARENTHETICAL].
166
Rilling et al., supra note 165, at 403.
167
NORMAN FROHLICH & JOE A. OPPENHEIMER, CHOOSING JUSTICE: AN EXPERIMENTAL
APPROACH TO ETHICAL THEORY (1992).
168
Id. at 181.
169
See McPherson, Smith-Lovin & Cook, supra note 154, at 429-30.
Gingerich
45
requires making a (political) decision about what sorts of empathy these institutions should
encourage.
4. Motivation Crowding
Motivation crowding out occurs when the presence of one motivational lever diminishes
or eliminates the importance of another motivational lever.170 In certain situations, encouraging
people to act in a certain way by using a particular motivational tool (say, economic punishment
or reward) may have a perverse effect of diminishing other motivations for people to act in the
desired manner.171 This suggests that the design of deliberative democratic institutions may
shape deliberators preferences by causing certain motivations to trade-off with other motivations.
For instance, Bruno S. Frey and Reto Jegen found that laws that suggest that citizens should be
trusted enhance civic virtue while laws that imply a distrust of citizens decrease civic virtue and
make citizens less likely to support the law.172 Specifically, they found that in Swiss cantons that
are pure direct democracies, tax morale (an instantiation of civil virtue) is substantially higher
than it is in Swiss cantons that are pure representative democracies.173 In the context of Swiss
tax morale, laws that seek to discipline citizens or that reduce direct political control of a
canton’s laws by its citizens crowd out intrinsic motivations to pay taxes.174 This is another
170
See Benkler, supra note 142 (manuscript at 15-16).
See, e.g., Ernst Fehr & Bettina Rockenbach, Detrimental Effects of Sanctions on Human
Altruism, 422 NATURE 137, # (2003) (finding that in investment games played in a laboratory
setting, altruistically motivated sanctions to benefit the group increased cooperative behavior but
sanctions imposed to enforce an unfair distribution of resources decreased subjects’ motivation
to cooperate).
172
Bruno S. Frey & Reto Jegen, Motivation Crowding Theory, 15 J. ECON. SURVS. 589, 604
(2001).
173
Id. at 605.
174
At first glance, it may seem that it would be desirable for legal structures to always promote
civic virtue, and therefore that it would be simple to stipulate that deliberative democratic
institutions should always be designed to do so, but this is far from clear. For one thing, it may
not be the case that people should always be encouraged to contribute to the overall welfare of a
171
Gingerich
46
motivational lever that can change individuals’ political preferences or beliefs that will have
different effects in different systems of democratic deliberation.
C. Framing Shapes the Outcomes of Deliberation
The motivational levers that help to explain how preferences and beliefs are changed by
participation in group decision-making processes suggest that how democratic deliberation is
institutionalized and framed will determine which motivational levers deliberation pulls, which
in turn will play a significant role in determining the outcomes of democratic deliberation. As
Tony Honoré recognized, “[a]ll law is the law of a group of individuals or of groups made up of
individuals.”175 What law is depends on the groups by and for which law is made. Many
preferences that play a role in democratic deliberation are “not fixed and stable but are instead
adaptive to a wide range of factors—including the context in which the preference is expressed,
the existing legal rules, past consumption choices, and culture in general.” 176 Many different
factors will be impacted by how deliberative democracies are structured. Do people have
political debates with other citizens in their neighborhoods? Do people discuss politics on the
Internet? Does citizen deliberation take place in a formalized setting, as in Fishkin and
Ackerman’s Deliberation Day? Is deliberation closely connected to authoritative decisionmaking? Are there spending limits on campaign advertisements in elections? Is political
representation based on geography or on something else? Are communities structured so that
state rather than of some other entity (for instance, the world). Additionally, to the extent that
civic virtue is associated with purely democratic political structures, encouraging civic virtue
may require abandoning institutional structures designed to preserve political equality (like
judicial review of government actions), which could undermine democratic decision making and
therefore civic virtue in the long run.
175
TONY HONORÉ, MAKING LAW BIND: ESSAYS LEGAL AND PHILOSOPHICAL 33 (1987).
176
Sunstein, supra note 2, at 197.
Gingerich
47
like-minded individuals can form politically influential enclaves? Are individuals’ self-identities
strongly connected to the identity of the deliberating group?
1. Geography and Scope
The outcomes of democratic deliberation are, to some extent, determined by decisions
about the geographic arrangements of deliberative institutions. Where national or state borders
are drawn plays an important role, to the extent that democratic deliberation is an enterprise
carried out by geographically bounded nation-states. Deciding how to count votes in elections is
a related question. Questions of scope also matter a great deal, and are closely connected to the
problems of geography: how many people get to deliberate? One of the basic issues of scope is
that where the boundaries of citizenship lie will determine who participates in democratic
deliberation, which will, in turn, determine what interests and opinions are advanced in
deliberation. Another issue is that size of deliberative groups will itself determine, in part, what
interests are represented in deliberation, which is particularly important in the context of
deliberation by the people’s representatives or proxies. (A small jury will likely leave some
important opinions and interests unrepresented177; at the other extreme, the Israeli Knesset may
provide for representation of too many preferences and opinions to effectively govern.178)
Fishkin and Ackerman’s Deliberation Day is an example of how geographic constraints
can impact deliberation. Fishkin and Ackerman suggest that Deliberation Day participants at
each school or community center where deliberations are conducted be drawn from a three- to
177
See BROWN, supra note 153, at 285-86.
See, e.g., Israel Ministry of Foreign Affairs, FOCUS on Israel: The Knesset (Mar. 1, 1999),
http://www.mfa.gov.il/MFA/MFAArchive/1990_1999/1999/3/FOCUS+on+Israel+The+Knesset.htm [PARENTHETICAL].
178
Gingerich
48
five-mile area.179 This geographic structure may determine the views that are presented during
and adopted as a result of deliberation. Schkade et al.’s experiment suggests that these are
background system design factors that might play important roles in determining what outcomes
deliberation generates. They suggest that if they had mixed people from Colorado Springs with
people from Boulder before having them deliberate, the predeliberation median would probably
have been predictive of the outcome—conservatives would have become more liberal and
liberals more conservative.180 They also note, however, that in such a scenario the positions of
the conservative and liberal groups could become entrenched, and that the more extreme and
divergent the pre-deliberation positions are, the more likely such entrenchment is.181 Thus, it is
possible that the order of deliberation could affect the likelihood of entrenched extremism. Since
the sorted deliberation among liberals in Boulder and conservatives in Colorado Springs led
participants to adopt more extreme views, they might be less willing to surrender such views if
inter-group deliberation were to be conducted after intra-group deliberation. Furthermore, if
group identity is particularly salient (if, for instance, it is announced or somehow made explicit)
group polarization effects are likely to be more extreme.182 Sorted political deliberation might
also encourage entrenchment by facilitating the development of a political group identity,
making compromise with political out-groups less likely. Enclave deliberation might therefore
encourage the development of extreme, uncompromising political views. Practical applications
of deliberative democracy always require eventually voting on outcomes, since consensus cannot
179
ACKERMAN & FISHKIN, supra note 1, at 70 (2004) (suggesting that this arrangement might
help ensure that conversations on Deliberation Day will not be dominated by highly educated,
relatively wealthy white males).
180
Schkade, Sunstein & Hastie, supra note 14, at 927 (finding that the predeliberation median is
the best predictor of the postdeliberation shift).
181
See id.
182
See id. at 932.
Gingerich
49
always be reached.183 Because of this, such entrenchment might support the adoption of political
principles supported by a majority of citizens but significantly removed from the median
political principles rather than the adoption of near-median political principles.
This possibility suggests that the political stakes of designing deliberative institutions are
quite high. Deliberation is always embedded in local power structures rather than autonomous
from them. 184 As long as mechanisms of deliberation are geographically structured, and as long
as deliberation involves fewer than all of the members of a decision-making organization, the
geographic effects of deliberation are likely to be unavoidable, since migration patterns tend to
produce more homogenous subcultures, decreasing people’s exposure to diverse opinions.185
2. How Deliberation Proceeds
Deliberative decisions will also, in part, be determined by the methods of deliberation
that an organization uses to deliberate. Different methods will likely encourage participants to
develop different political preferences or beliefs by appealing to different motivations. Because
different methods of debating will allow certain arguments to be developed more fully and gain
183
See Elster, supra note 16, at 7.
Günter Schönleiter, Can Public Deliberation Democratise State Action?: Municipal Health
Councils and Local Democracy in Brazil, in POLITICISING DEMOCRACY: THE NEW LOCAL
POLITICS OF DEMOCRATISATION, 75, 105 (eds. John Harriss, Kristian Stokke, and Olle Tönquist,
2004).
185
See Schkade, Sunstein & Hastie, supra note 14, at 938; see also Alan Abramowitz, Brad
Alexander & Matthew Gunning, Incumbency, Redistricting, and the Decline of Competition in
U.S. House Races, 68 J. POL. 75, 86-87 (2006) (finding that in the past thirty years there has been
a substantial increase in the partisan polarization of U.S. House districts and that there has been a
substantial increase in partisan voting in House elections); McPherson, Smith-Lovin & Cook,
supra note 154, at 429-30 (noting that even seemingly trivial geographic factors can influence
the formation of homophilic ties).
184
Gingerich
50
wider audiences, and because different methods of aggregating or synthesizing political opinions
will yield different results, democratic procedures have various distributional properties. 186
The existence of mass media, and the forms that it takes can have an important impact on
a democratic society’s deliberative processes. Different structures of mass media will result in
different structures of political communication and deliberation, which in turn help decide what
laws or political actions people agree on. Thus, “[t]he BBC or the state-owned televisions
throughout postwar Western European democracies . . . constituted the public spheres in
different ways than did the commercial mass media that dominated the American public
sphere.”187 The political implications of different designs of media and communications systems
arise, in part, from the necessity that in any system with a very large number of participants (and
hence any system that has any potential to mitigate the ideological effects of geographic
arrangements), arguments must in some way be filtered for “political relevance” and
“accreditation.”188 A deliberative system must determine what concerns are “plausibly within
the domain of political action and those that are not” and exclude non-political concerns from
democratic deliberation.189 Such systems must also, in some way, exclude subjects that lack
credibility (like implausible conspiracy theories). A variety of institutional actors help to provide
accreditation, including professional norms for journalists, political parties, academia, the civil
service, large corporations, and nongovernment organizations.190 These filtering functions are
necessary for any large organization that conducts democratic deliberation because participants
have limited resources of time and attention to devote to deliberation, and it is impossible for
186
See Thomas Christiano, Social Choice and Democracy, in THE IDEA OF DEMOCRACY, supra
note 2, at 173, 182-83.
187
YOCHAI BENKLER, THE WEALTH OF NETWORKS 181-82 (2006).
188
See id. at 183.
189
Id.
190
See id. at 183-84.
Gingerich
51
every individual to evaluate every single argument or claim made by every other individual.191
System design decisions about who performs the filtering and how it is performed will play a
role in shaping the political debates that a society has and determining what voices will be heard
in the debate. The BBC is not inherently more democratic than American commercial television
broadcasting, nor vice-versa, but the two systems privilege different voices, and shape
discussions of political questions differently.
It is possible that systems could be designed to allow for filtering institutions that are
more participatory than mass media. Benkler has argued that the Internet provides such a
system, in which individual groups focus on ideas that interest them in specialized online
communities and then pass the best and most relevant ideas on to larger audiences through nodes
that connect to larger, less specialized audiences.192 However, while the Internet may restructure
political deliberation to give a larger number of individuals a larger role in shaping the public
agenda, as Benkler argues, such a restructuring could take on a variety of forms, each with
distinct political implications. The following examples are purely hypothetical, but show what
the political stakes might be. First, image a network system in which electronic communication
is, because of regulations or because of the physical infrastructure of the network, very cheap
when done locally (i.e. within a particular geographic neighborhood) and very expensive or
191
The absence of such filtering would be just as ideologically colored as its presence, since the
absence of filtering mechanisms would necessitate a random selection of the arguments that an
individual with limited time and attention is able to listen to, which would benefit political
interests that make claims that are regarded as irrelevant or non-credible that would normally be
removed from large scale deliberation by filtering mechanisms. This might also simply benefit
speakers who happen to be charismatic. See ARTHUR LUPIA AND MATHEW D. MCCUBBINS, THE
DEMOCRATIC DILEMMA: CAN CITIZENS LEARN WHAT THEY NEED TO KNOW? 62 (1998) (“A
person’s ability to persuade depends on how he or she is perceived by others.”).
192
See BENKLER, supra note 187, at 232, 242, 260. This argument seems consistent with
Sunstein’s argument that enclave deliberation can improve the overall argument pool, if enclave
deliberation is connected to broader, more heterogeneous discussions. See CASS R. SUNSTEIN,
REPUBLIC.COM 2.0, at 77-80 (2007).
Gingerich
52
impossible when done nationally or globally. Compared to a system in which communication
with people all over the world is just as easy as communication within a neighborhood, different
interest groups would emerge to filter for relevance and accreditation. It might be easier for
groups concerned with physical infrastructure to grab media attention, since it might be easier to
pull together a group of people who share the same physical environment even if it is not
possible (say, because of their work schedules) or organize them on the streets or in parks. And
urban interests might better be able to develop and vet ideas than would be rural interests. Local
solidarity effects on motivation would play a larger role in shaping preferences in the localized
network, while particular topical interests would be more likely to generate solidarity effects in
the global network. Second, imagine a network system in which communication was conducted
aurally. Compared with a system in which most communication was visual, the aural network
system might generate stronger empathy effects among participants in deliberation (by
humanizing the other participants) and it might require stronger filtering efforts, since people
typically read more quickly than they listen. Third, imagine a system in which accreditation is
performed primarily by a single node (a search engine). Imagine that, on the model of the BBC,
this single node is state owned or state sponsored (say, the Federal Government nationalized
Google), and that therefore democratic controls exist on how the search engine functions and its
operations are not driven by profit motives. Compared to the extant system of the Internet as
described by Benkler, such a system might give fewer individuals a direct role in filtering
arguments for inclusion in a public debate, but it might also be less market driven (maybe the
single node would adopt a system making it easier for web pages produced by educational or
non-profit organizations to pass through the filters of relevance and accreditation).
Gingerich
53
In any case, a variety of network structures could increase participation in filtering, and
could do so in a number of ways, each of which would have political implications. A network
can never decide solely by itself how it should be structured, because networks always exist
against a backdrop of external conditions including markets, technologies, and laws and
regulations. There is not a single democratic method of structuring deliberation, since different
institutional structures will motivate people to develop and subscribe to different political
arguments and ideas.
3. Deliberation Daze
Two hypothetical constructions of a Deliberation Day demonstrate how different
deliberative systems might, by way of their institutional design, generate different political
outcomes.193 Both of the following hypotheticals are modeled on Ackerman and Fishkin’s
Deliberation Day. Imagine that these hypothetical deliberation days are held in conjunction with
elections to an American state legislature. On Deliberation Monday, shortly before the election,
citizens of the American state of Ames gather in small groups at neighborhood schools and
community centers. Participation is mandatory, and there is a substantial tax penalty imposed on
no-shows. No deliberation center has more than a few hundred participants, and after
participants arrive at the schools, they are divided into smaller, ten to twelve person, geographic
groups, so that citizens deliberate with their (geographically) closest neighbors. There are no
moderators, and no set limits on the topics for deliberation. After deliberating for four hours,
each small group issues a consensus report on the issue that they think is most important for the
193
This is not to say that Deliberation Day is a necessary element of any system that claims to
effectuate democratic deliberation, but simply an example of how structuring a particular
institution in the deliberative system could have ideological repercussions. However, the
decision whether or not to have a Deliberation Day is itself an element of deciding how to design
a deliberative system, so to the extent that decisions about this particular institution are
ideologically colored, the broader system may also be so colored.
Gingerich
54
upcoming election. After the reports are issued, there is another four hours of deliberation as a
large group (in a school cafeteria or gymnasium), after which the large group chooses the three
political issues that it thinks are most important and decides, roughly, what its positions are on
these issues. The group selects a reporter who writes a brief, formal statement describing the
issues that the group thinks are most important, and what the group thinks should be done about
the issues (e.g., “the gas tax should be raised to 10%”). Deliberation completed, the formal
statement is published online, transmitted to the local paper, and sent to the candidates for the
Ames Legislature, who can then take positions on the issues and recommendations of the
deliberative groups before the election.
In contrast, on Deliberation Sunday citizens of Ames who wish to deliberate go to a
website set up for their state house district, where a number of chat rooms exist dedicated to a set
of issues drawn from lists prepared by all of the candidates on the ballot for the Ames House of
Representatives and the Ames Senate as well as any topic that at least thirty citizens of a given
house district request to be included. There is no requirement that citizens participate, and no
financial incentive or disincentive for them to do so. In advance of Deliberation Sunday, a
bureaucratic agency conducts research on every topic included on each website and prepares
short briefings on each issue, which are distributed to participants to read before Deliberation
Sunday. Each participant can choose to participate in one discussion, and after choosing a topic
about which she wishes to deliberate, is randomly assigned to a chat room with no more than 30
participants. Each group elects a foreperson, who is able to moderate comments in the chat
room, and a reporter who, after four hours of deliberation, prepares a brief statement on the issue
the group discussed and posts it on a deliberation blog for every deliberator from the Ames
Gingerich
55
House district to view. All of the deliberators for the district can then post comments on any
issue they please, and the reporter who prepared each entry on the blog moderates the comments.
It seems likely that even if exactly the same people participated in Deliberation Monday
and Deliberation Sunday, the results would likely differ, perhaps particularly on questions of
how to use scarce resources and other tradeoff questions. I suspect that Sunstein and Fishkin and
Ackerman would agree that these two designs would yield substantially different outcomes. It is
by appealing to differences like those between Sunday and Monday that Sunstein and Fishkin
explain how the deliberative polls that Fishkin has conduced avoid the polarization effect
detected by Sunstein.194 That Fishkin and Sunstein would agree that these two procedures would
generate very different outcomes might alone be sufficient to show that there is no neutral point
from which to deliberate without reference to the motivational levers that I have discussed.
However, I suspect that Fishkin might argue that one design is more neutral than the other and
that they differ because one achieves the goal of deliberative neutrality less well than the other.
It may be that the best response to such an argument, albeit a rather impressionistic response, is
that Sunday and Monday look quite similar, and both designs appear to embody many of the
same values. It would be difficult at best to formulate neutral standards that would enable
Fishkin to determine that one is actually more neutral than the other.195 Moreover, the fact that
each design benefits some groups and hurts others makes it very difficult to think that it is
possible for one design to be more neutral than the other. This impressionistic response to
194
Schkade, Sunstein & Hastie, supra note 14, at 934-35.
Fishkin might reply that both are ideologically tainted, so a different design altogether is
needed, but this strikes me as an unsatisfactory reply because non-ideal theory is already in the
business of picking between imperfect alternatives. If such a theory cannot provide practical
answers that let deliberators know which system is better but instead only describes an ideal
system without telling them how close they are to that ideal, the theory has effectively ceased to
be a non-ideal theory.
195
Gingerich
56
Fishkin is not a knockdown, killer argument. It is certainly possible that someone could
formulate neutral principles that show Sunday to be more neutral than Monday or vice-versa or
could design some other Pareto-optimal Deliberation Day that benefited every group and did not
hurt any group, but these possibilities strike me as highly unlikely.
An additional response to Fishkin is that the presence of motivational levers that the
design of deliberative systems affects undermines the possibility of developing principles that
allow theorists to determine that one deliberative design is more neutral than another. On
Monday, we might expect that geographical solidarity is likely to play a big role and that
individuals’ political decision are likely to be motivated in part by empathy for their
neighbors.196 The tax penalty for non-participation might crowd out civic virtue motivations to
see the deliberative process as legitimate, which might make participants less committed to the
deliberative endeavor but might also increase critical self-awareness of how the deliberative
process might replicate the bias of participants or of the dominant political system. On Sunday,
deliberators might be more strongly motivated by feelings of solidarity with people who share
their interests and, because the means of deliberation do not provide for seeing other people in
person, might be less motivated by empathy.197 Substantial communication among a small group
on single issues might help build agreement about what the solutions to particular problems
should be, but less agreement than on Monday about the relative priority of different political
problems.198 Another important difference between Monday and Sunday, perhaps even more
196
See McPherson, Smith-Lovin, & Cook, supra note 154, at #.
It seems likely that these two designs will impact solidarity and empathy in very different
ways and to very different degrees. See infra note 201.
198
Another possibility is that certain religious communities would be less likely to participate on
Sunday than on Monday, which would have ideological effects on the outcomes of deliberation
provided that the group of religious non-participants had political beliefs that differ significantly
from the political views of the deliberating community as a whole.
197
Gingerich
57
important, is the presence of a moderator on Sunday and the limitation of the discussion on
Sunday to a single, pre-determined issue.199 On Sunday, deliberators might become committed
to a particular issue about which they have spent time deliberating, and may be less willing to
trade-off this issue with other issues.200
Further suppose that in Ames, there are two main ideological factions; call them the
Evens and the Odds. The Evens think that the basis of a strong community is having clean and
orderly streets and ensuring that there are lots of neighborhoods with tight knit communities.
The Odds, on the other hand, think that the basis of a strong community is home ownership, and
they are very worried about deceptive mortgage lending. If you are an Even ideologue who
cares a great deal about neighborhoods having clean streets, you should prefer the structure of
Deliberation Monday, while if you are an Odd ideologue who cares about the technical
intricacies of a plan to regulate mortgage lenders, you should prefer Deliberation Sunday.
Suppose there is also a third, very small ideological faction; call them the Primes. The Primes
consist of thirty scientists in Ames City who have been studying a near extinct Paraguayan insect
and have come to believe that there is some chance that this insect’s genes hold the key to curing
malaria. The Primes think that preserving the habitat of the Paraguayan insect is the most
important global political issue, and think that all local considerations pale in comparison, and
199
As I have set up the scenario, it is tricky to parse out what precise aspects of the design cause
the differences between Monday and Sunday. I have set up the scenario in this way in part to
demonstrate how many design features confound the possibility of neutrality and in part because
the non-motivational differences I have incorporated seem, to me, to magnify the salience of the
motivational differences. (The salience of solidarity can be magnified when a group is able to
coalesce around a single issue and develop more homogenous views because, in part, of the
presence of a moderator.) The designs could be modified to isolate differences between Sunday
and Monday, which could allow a more precise determination of what causes the different
designs to generate different outcomes. (For instance, Sunday could be given the same design as
Monday, except for the geographical approach to grouping.)
200
See Whyte, supra note 134, at #.
Gingerich
58
they want the State of Ames to do everything possible to pressure the Paraguayan government to
save the insect’s habitat. The Primes will certainly prefer Deliberation Sunday to Deliberation
Monday, since it will give their esoteric but possibly quite important cause a better change of
garnering public attention. In any case, the issues that are emphasized in the media and in
popular discussion after Deliberation Monday are likely to differ from those emphasized after
Deliberation Sunday, no matter how much work is done to ensure that the institutions set up for
deliberation are politically neutral. 201 This presence of politics in institutional design questions
about democratic deliberation has important repercussions for democratic theory.202
IV. IMPLICATIONS FOR DEMOCRATIC THEORY
Consequentialist deliberative democracy theorists tend to assume that the conditions of
fairness are possible in modern society.203 If democratic deliberation works correctly, its
outcomes should reflect a synthesis of the beliefs of the deliberators rather than the ideological
201
It is entirely possible that these two designs would have radically different effects on levels of
empathy and solidarity. However, the effect might be much more pronounced when the group
identity is generated through face-to-face contact with geographic neighbors than through online
meetings with strangers. See supra note 185. It is entirely possible that further empirical studies
would be needed to gather data that would show how solidarity would play out differently on
Sunday and Monday, but I think that this does not necessarily weaken my point. Interest groups,
like the Evens and the Odds, could theoretically gather this data so that they could determine
which design would best promote their ideological objectives. At the very least, it seems likely
that the solidarity effect would be at least marginally better for the Evens on Monday than on
Sunday and for the Odds on Sunday than on Monday. Furthermore, whenever a salient group
identity exists, some solidarity effect can be expected. See McPherson, Smith-Lovin, & Cook,
supra note 154, at #.
202
It is possible that, at some point, deliberative systems interactions with ideological systems
become too complex to make accurate predictions about systems’ ideological effects. If it is too
difficult to predict whom a particular design feature will benefit or burden, ideological interest
groups may not care about the structure of deliberation. However, this possibility of excessive
complexity (which I cannot discount) would cut against the goals of deliberative democracy
theory, since it would suggest that the decisions generated by democratic deliberation are, at least
in some fashion, no better than random political decisions.
203
See, e.g., Shapiro, supra note 42, at 437 (“I will assume, however, that such conditions are
attainable in modern society, although I admit that such a proposition is far from self-evident.”).
Gingerich
59
predispositions, whether about certain issues or about how political institutions should work, of
the system designers. Thus, deliberative democracy theorists identify principles to structure
deliberative systems to ensure that they remain ideologically neutral. For example, Sunstein
identifies four ways in which neutrality can be not objectionable: when it is a call for internal
consistency, when it is a requirement that legal outcomes and the distribution of social costs and
burdens be justified with public regarding arguments, when it imposes a requirement of
impersonality or abstraction on decision-makers in certain contexts, and when it refers to a
suitable baseline for determining the existence of selectivity and bias.204
However, democratic deliberation is always “in danger” of distortion by political power,
because it trucks in political power.205 This means that certain deliberative set ups will likely
have particular political consequences. I will argue in this section that deciding how the
deliberation will proceed has important ideological and political ramifications and that neutrality
is impossible to achieve in deliberative systems, both as they might be implemented in the real
world and as they might be theoretically conceived. I will argue that Sunstein’s principles, for
example, would not succeed in creating a neutral deliberative system, because even within the
constraints for deliberation that he specifies, ideology can still predetermine deliberative
outcomes. I will first argue in IV-A that democratic deliberation cannot be neutral in a world
that roughly resembles the status quo, and I will then extend the argument in IV-B to more
theoretical accounts of democratic deliberation. 206
A. Real World Democratic Deliberation Does Not Generate Ideologically Neutral Outputs
204
SUNSTEIN, supra note 70, at 351-53.
Richardson, supra note 65 (manuscript at 7, on file with author).
206
My argument is not, of course, that enabling citizens to deliberate is a bad idea, only to
suggest that simply because a law is the result of democratic deliberation does not, by itself,
provide any normative reason to treat that law as a law that should be followed.
205
Gingerich
60
My argument that democratic deliberation cannot avoid ideological infiltration as it might
be practiced in the real world only requires showing that as it is practiced, deliberation does
distort preferences, not that this distortion is inevitable or could not be corrected in an abstract
theory.207 It is only necessary to show for this argument that given the substantive background
conditions of society, what deliberation does and how it works have important political
implications. Inequalities of wealth, power, and education are so vast in contemporary American
and global society that it would be ludicrous to suggest that democratic deliberation could
proceed in a manner that provided fair access to all participants without vast, systemic
changes.208 In order to generate normative reasons that a law has authority, democratic
deliberation must be supposed to give the right answer to political questions. But the impact of
education and socio-economic status on the ability to deliberate is immense, and political
participation in the modern world invariably depends upon the possession of resources.209
Because of the important role that these factors play in deliberating about policy and principles,
the outcome of any democratic deliberation given the conditions of society today would likely be
far different than the outcome of any deliberation in which all of the participants had equal
207
The best arguments about deliberation accept that the political status quo is far from what is
needed to generate legitimate political outcomes. My argument in this section should, therefore,
be read primarily as laying the ground for my argument in Section IV-B. I think that my
argument in IV-A does, however, cut against certain naïve constitutional theories that maintain
that a law is legitimate if it is passed in accordance with the procedures specified in the
American constitution, a position which, while uncommon in academic writing on deliberative
democracy, is not entirely absent from academic discourse on other topics and is not at all absent
from the popular media.
208
See, e.g., THE ECONOMIST, POCKET WORLD IN FIGURES passim (2009) [PARENTHETICAL].
209
See Jack Knight and James Johnson, What Sort of Political Equality Does Deliberative
Democracy Require?, in DELIBERATIVE DEMOCRACY: ESSAYS ON REASON AND POLITICS, supra
note 12, at 279, 306 [PARENTHETICAL]; Aileen Kavanagh, Participation and Judicial
Review: A Reply to Jeremy Waldron, 22 L. & PHIL. 451, 480 (2003) [PARENTHETICAL].
Gingerich
61
access to education, to wealth or free time, to charisma, and to other resources to enable them to
persuade other deliberators.210
Perhaps democratic deliberation could provide normative force to laws by simply
avoiding extremism. Gutmann and Thompson suggest that democrats should seek to avoid
processes that promote extremism. 211 Similarly Christopher McMahon argues that democratic
values should foster debates that reduce differences of opinion.212 However, deliberative
systems that simply spit out the median political view of the status quo do not seem any more
likely to generate reasons that a law is authoritative. It would be a mistake to assume that the
status quo is in any way neutral simply because it is the status quo.213 The status quo is a
baseline that is historically contingent, and just like any other baseline it must be justified.214 A
deliberative system that simply shifts all political views toward the center does not provide any
reason to treat its outcomes as correct unless there is some independent reason to believe that the
present center has at least roughly the best answers to political questions. Ironically, Gutmann
and Thompson’s view would require such a dramatic reform of the existing procedural system in
order for its preconditions of fairness to hold that it would itself be “extreme” compared to most
210
An additional problem is ensuring not only that everyone has equal access to the resources
necessary to enable successful participation in political deliberation, but to ensure that everyone
actually participates. If many people are not participating, it might be because the ground rules
for deliberation intentionally or inadvertently exclude them, or because they disagree with the
ground rules. So, as Roberto Gargarella suggests, the problem of deliberative institutions
securing “full representation” affects the impartiality of political decisions. See Gargarella,
supra note 73, at 274 (“[W]e need to know who deliberates, and we should be worried if most
people are kept at the margin of political deliberation.”).
211
GUTMANN & THOMPSON, WHY DELIBERATIVE DEMOCRACY?, supra note 12, at 54.
212
Christopher McMahon, Autonomy and Authority, 16 PHIL. & PUB. AFF. 303, 328 (1987).
213
Social psychologists have, however, identified a general tendency to rationalize the status quo
and treat it as good, fair, and rational, even in the absence of good reasons to do so. See, e.g.,
Aaron C. Kay et al., Panglossian Ideology in the Service of System Justification: How
Complementary Stereotypes Help Us to Rationalize Inequality, 39 ADVANCES EXPERIMENTAL
SOC. PSYCHOL. 305, # (2007) [PARENTHETICAL].
214
SUNSTEIN, supra note 70, at 353.
Gingerich
62
extant beliefs about deliberation in the North American status quo.215 Any attempt to find
normative force by moderating political beliefs therefore seems suspect.
Furthermore, the empirical realities recited above in Section III appear to necessarily be
at work in any system of deliberation that does not assume that deliberative institutions would
transform the preferences, beliefs and attitudes of deliberators. It seems reasonable that theories
of democratic deliberation should accept extant knowledge about practical reasoning and
political motivation rather than simply assuming that participants in deliberation have
normatively desirable characteristics and beliefs.216 To the extent that this background
knowledge about political motivations is accepted, the effects of communication, solidarity,
empathy, motivation crowding, and the informational effects of deliberative processes will result
in deliberative outcomes with an ideological tilt. That said, it is possible that democratic
deliberation could be at least theoretically (if not practically) structured to be ideologically
neutral with respect to its outcomes, and I will now turn to the question of whether such a
theoretical account of deliberative neutrality is plausible.
B. Deliberative Democracy Is Always Distorted by Ideology
Ideological considerations play a significant role in determining how societies design
deliberative institutions. Different institutional structures encourage participants in deliberation
to develop different preferences and beliefs than do other structures. In Ames, the Odds, the
Evens, and the Primes would design different institutions for deliberation because of different
215
Frederick Schauer points out that for what Gutmann and Thompson see as legitimate political
deliberation to occur, we would probably need a much better substantive political environment
that we have now, and to the extent that we had a substantive better political environment, the
need for legitimate public deliberation would, presumably, be greatly reduced. Frederick
Schauer, Talking as a Decision Procedure, in DELIBERATIVE POLITICS: ESSAYS ON DEMOCRACY
AND DISAGREEMENT, supra note 12, at 17, 24-26.
216
James Johnson, Arguing for Deliberation: Some Skeptical Considerations, in DELIBERATIVE
DEMOCRACY 161, 174 (Jon Elster ed., 1998).
Gingerich
63
ideological commitments. The ideological commitments that determine how to shape
deliberative institutions would not, however, need to concern discrete political questions—they
could also be ideological commitments to principles regarding how government institutions
should function, including deontological commitments to certain deliberative structures, or they
could be principles about efficacy and the cost of governing.217 Choosing one institutional
design rather than another will reflect certain ideological considerations rather than others. If
people had permanently fixed preferences and political beliefs, this would not matter much to
theoretical accounts of democratic deliberation, because a system could be designed, at least in
theory, in which the deliberative process would not distort any preferences. But this is
impossible. As I argue in Section III-B of this Paper, preferences and political opinions are
endogenous, and are constituted in part by political institutions. Different systems will motivate
people in different ways, and therefore will encourage people to develop different political
commitments. Such constitutive effects are inevitable, and there is no undistorted baseline of
preferences or beliefs that democratic deliberation could preserve. Of course, institutional
designs could be developed to reduce or eliminate certain motivational forces. For instance,
dispersing power among deliberative bodies could reduce group polarization, decreasing the
salience of solidarity and empathy mechanisms.218 However, whether this would be good or bad
depends on one’s political goals. Reducing the salience of some motivational mechanisms will
simply increase the salience of other motivational mechanisms. So a fundamental dilemma for
217
I am not, however, attacking theories that justify democratic deliberation on deontological
grounds, like Jeremy Waldron’s theory. I am only attacking consequentialist accounts of
deliberative democracy that argue that democratic deliberation gives moral force to laws because
it generates good outcomes.
218
See COLIN FARRELLY, JUSTICE, DEMOCRACY AND REASONABLE AGREEMENT 213 (2007).
Gingerich
64
deliberative democracy is that it must presuppose certain substantive political principles in order
to resolve basic institutional design questions adequately.
This severely threatens democratic deliberation’s ability to generate normative reasons to
treat a law as legitimate, because if someone disagrees with the political commitments that
underlie the institutional design of deliberation, then she does not have a reason to treat the
outcomes of deliberation as legitimate, and this suggests that reasons to treat the outcome of
deliberation as authoritative must precede deliberation rather than arise out of it. This is a
notable problem because any theory of deliberative democracy that adequately copes with the
problems of inequality present in real world deliberation must develop a rather thorough baseline
account of the conditions necessary for democratic deliberation to proceed. Determining which
politically relevant capacities are beyond individual control will depend on society’s
understanding of personal relationships, which are controversial subjects themselves.219 This
means that there will be much room for disagreement with the fundamental principles structuring
deliberation, as these principles will have very wide ranging effects.
Perhaps, in theory, it would be possible to remedy these pathologies of deliberation by
justifying a certain baseline procedural rule for deliberation and then following that baseline
procedure to generate outcomes that have legitimacy because they are the product of democratic
deliberation. For instance, Sunstein argues that while in a constitutional system “there are
constitutional limits on what citizens may decide . . . [and] these limits are themselves set down
in advance by the citizenry.”220 However, my argument undermines this constitutionalizing
move as well. Suppose that system designers decide that in order for legitimating democratic
deliberation to take place, the government must provide free education from preschool through
219
220
See Knight & Johnson, supra note 209, at 305.
SUNSTEIN, supra note 70, at 373 n.21.
Gingerich
65
graduate school to anyone who wants it and must require that ever citizen complete at minimum
two years of college or equivalent post-high school vocational training. If I think that providing
so many resources to education fundamentally misallocates resources, I will not have a reason to
agree with the outcomes of deliberation by virtue of their being the outcomes of democratic
deliberation.221 On the other hand, if I think that fair democratic deliberation is impossible
unless the government offers free education through graduate school to any citizen who wants it,
I will not have a reason to treat as legitimate the outcomes of deliberation in any system that does
not provide such educational opportunities, since I will think that not everyone has fair and equal
opportunities to participate in the system. In fact, almost every political issue will have some
implication for the ability of different individuals to participate in democratic deliberation.222
Since baseline requirements for deliberation could be so substantial and could significantly affect
decisions about a wide range of political issues, it is not plausible that everyone in a society
could agree on basic ground rules for deliberation even if they could not agree on discrete
political issues. Arguing that deliberation must be among equal citizens is, by itself, meaningless
absent any specification of what it means for citizens to be equal. If anything, choosing ground
rules should be harder than deciding about discrete, substantive political issues rather than easier.
A further difficulty with picking a baseline for deliberation before designing deliberative
institutions with that baseline in mind is that “outside of institutional settings, the people cannot
221
Although I might think that such outcomes are legitimate for other reasons, unrelated to their
being the outcomes of democratic deliberation.
222
A very partial list of laws that could differentially affect the ability of certain individuals to
participate in democratic deliberation includes: regulation of communication networks, sexual
harassment laws, anti-discrimination laws, laws regulating and/or guaranteeing health insurance,
occupational safety laws, wage and hour laws, and landlord-tenant laws.
Gingerich
66
coherently be conceived of as thinking anything.”223 There is no natural, pre-institutional
method of determining what the popular will is, and thus no way to generate baseline principles
for deliberative democracy without referring to external, objective political values. There cannot
be democratic control of anything until there are democratic institutions.224 Because of this
inevitability, any claim that democratic deliberation provides normative force to a law by virtue
of democratic ratification necessarily rests on more fundamental political principles that justify
the mechanisms of democratic deliberation. Constitutional constraints cannot be the outcome of
democratic deliberation alone. Deliberative democracy is required to bring forth fundamentally
valid laws, but this can happen only when it is the outcome of a conceptually prior procedural
event, itself framed by valid laws.225 Any constitutional framework for deliberation cannot itself
possibly be the outcome of the popular will.226 As Knight and Johnson argue, “whatever form
the actual mechanism for assessing effective participation takes, the requirements of political
equality will themselves be fundamentally political questions.”227
The empirical evidence shows that deliberation is never neutral: deciding whether to have
a moderator, deciding how big the group should be, deciding how to select the group, deciding
how debate should proceed (will materials be provided by some outside actor?), deciding how to
group people for deliberation, deciding what background conditions are necessary for equality,
and deciding whether structures should be designed to increase or to decrease polarization are all
223
Richardson, supra note 65 (manuscript at 13, on file with author) (arguing that aggregations
of preferences, like those generated in Fishkin’s deliberative polls, cannot be taken as the popular
will because there are many different ways of aggregating a set of preferences, all yielding
different results).
224
Richardson, supra note 65 (manuscript at 12, on file with author).
225
See Frank I. Michelman, How Can the People Ever Make the Laws?: A Critique of
Deliberative Democracy, in DELIBERATIVE DEMOCRACY: ESSAYS ON REASON AND POLITICS,
supra note 12, at 145, 162-65.
226
See Richardson, supra note 46, at 353.
227
Knight & Johnson, supra note 209, at 309-10.
Gingerich
67
difficult questions that cannot be resolved by deliberation. Maybe they could be resolved by
deliberating within an acceptable procedural framework for deliberation (say, a constitutional
convention) but then that framework would also need to be justified without reference to its
democratic nature.
Perhaps Bruce Ackerman’s idea of constitutional moments offers a way out for
deliberative democracy to create a basic structure on the basis of the popular will and then to
allow for deliberation within that structure.228 On this account, the popular will might emerge
during moments of perturbation when the political beliefs of the people become clear, and
through this process, political institutions might earn the authority to speak for the people.229
There is a two-track lawmaking system, the lower track of which is designed for normal political
decisions that are made in the absence of highly mobilized majority sentiment and the higher
track of which imposes rigorous tests on political movements that can speak for the people and
allow for considered judgments to be made by the people, rather than simply by the
government.230
The idea of constitutional moments, however, fails to overcome the problems of
ideology. First, every political deliberation, even one that takes place with minimal formal
See BRUCE ACKERMAN, WE THE PEOPLE (1991). This appears to agree with Habermas’s idea
of a reflexive and recursive legitimating legal procedure that is itself legitimated by a
conceptually prior procedural event. See Habermas, Constitutional Democracy, supra note 5, at
# (“[T]he allegedly paradoxical relation between democracy and the rule of law resolves itself in
the dimension of historical time, provided one conceives the constitution as a project that makes
the founding act into an ongoing process of constitution-making that continues across
generations.”); Frank I. Michelman, Book Review, 93 J. PHIL. 307, 308 (1996)
[PARENTHETICAL].
229
See 1 ACKERMAN, supra note 228, at 44.
230
See 2 ACKERMAN, supra note 228, at 5. Habermas envisions a similar extra-institutional
circulation of public political discourse. See HABERMAS, BETWEEN FACTS AND NORMS, supra
note 5, at 408. The line between the two spheres is not, however, as sharp for Habermas as for
Ackerman. See Michelman, supra note 228, at 313.
228
Gingerich
68
structure, has certain baseline requirements. Background conditions of education, wealth, and
social network memberships (carrying with them ability to leverage solidarity) shape the ability
of different individuals to mobilize people (or resist mobilizations) and thus to participate in
decision making on the higher of the two tracks described by Ackerman. Second, group
polarization is likely even if equality and public reasoning requirements are somehow
satisfied.231 In fact, polarization seems particularly likely when ideological affinity forms the
basis for political organization, as typically occurs in large-scale movements. Because of the
likelihood of such polarization, even among citizens whose make conscientious efforts to
achieve truth and understanding, institutional design questions about the extent to which
deliberative structures should channel popular impulses into more heterogeneous deliberative
bodies will retain significant ideological importance.
Third, and finally, deciding who counts as a participant is also a political question that
precedes the possibility of moments of popular perturbation that provide ongoing legitimacy to
government institutions. The ability to participate in political life in a meaningful way and to
mobilize popular opinion may depend on acceding to the legitimacy of foundational laws before
deliberation even has a chance to begin. For instance, on one reading of the history of Federal
Government interactions with Native Americans in the late nineteenth century, tribes were
required to accede to the legitimacy of the American government and its institutions.232 If they
failed to accede to this authority, they were subjected to spatial isolation in the form of
confinement to reservations. While a mere denial of citizenship might still allow a group of
people to mobilize popular support to gain rights as citizens, spatial exclusion dramatically
231
SUNSTEIN, supra note 12, at 42.
This history is certainly contested, and I do not purport to offer a definitive historical account,
only a plausible one in order to illustrate this point. I would like to thank Kyle Doyle for
suggesting this example.
232
Gingerich
69
reduced the possibility of participating in broad, popular political deliberation for certain groups
of Native Americans. Decisions about who is allowed to live where and who counts as a citizen
can differentially influence the ability of interest groups to secure popular support for laws or
legal reforms that they desire. Such decisions loom in the background of any popular
mobilization that could spark constitutional moments, so ideological coloration cannot be
avoided simply by claiming that largely informal political movements that crop up from time to
time provide ongoing legitimacy to institutional mechanisms for day-to-day deliberation and
political decision-making.
Furthermore, even if the system of democratic deliberation is just, or at least as just as it
reasonably could be given the circumstances, this does not provide a reason to treat its results as
having legal authority. Because deciding what sort of procedural system to set up must make an
all things considered judgment about the outcomes it will generate (unless it is constrained by
some prior procedural framework that was itself established by asking such an unconstrained
question), it must account for every tendency that a deliberative system will have to produce one
outcome rather than another, and accord it some weight in the decision about what procedural
system to adopt. Because the procedural system adopted is based on political commitments, any
normative claim it generates should only be thought of as a wide-scope ought, and members who
adhere to the political values that created the system should be free, to use John Broome’s term,
to put their reasoning in reverse, and reject the political premises that they relied on rather than
accept an unpalatable outcome of their reasoning.233 (If, for instance, they though a system of
233
See John Broome, Reasons, in REASON AND VALUE: THEMES FROM THE MORAL PHILOSOPHY
28, # (R. Jay Wallace, Philip Pettit, Samuel Scheffler, & Michael Smith, eds.,
2004); see also JAMES BOHMAN, PUBLIC DELIBERATION: PLURALISM, COMPLEXITY, AND
DEMOCRACY 238 (1996) (arguing that public deliberation must take place in a revisable
institutional and interpretive framework in order to be legitimate).
OF JOSEPH RAZ
Gingerich
70
deliberative democracy would work out very well when they set up its institutions but the system
then decided that slavery is permissible.) As Raz points out, approving of a democratic
institution does not mean acknowledging that every decision it makes is just.234
This point goes further, too: if you think that a system of deliberation is neutral and really
allows everyone to participate equally in a democratic decision making process, then the fact that
democratic deliberation ratified a certain law might be a reason to treat that law as a wide scope
normative requirement. But the empirical evidence makes it difficult to believe even this. The
empirical evidence shows that there is no such thing as neutral deliberation: even if you could
agree on the baseline requirements of equality for participation in deliberation, it is still
necessary to decide procedural questions that cannot be decided simply by reference to equality.
For instance, are divided institutions desirable in order to discourage extremism? This depends
on deciding if it is desirable to have extreme opinions, which does not necessarily have to do
with equality.235 Should deliberation take place at a local level? A national level? A global
level? Again, in a world of perfect resource equality, the answer to this question would seem not
to be dictated by principles of equality; some other political commitments will be needed to
decide what scale deliberative democracy should take. As long as these decisions involve
political commitments other than agreement to abide by the decisions made by fellow-citizens
following fair procedure and playing equal roles, they cannot provide the basis for finding that a
law has authority, since democratic deliberation can only provide such a reason if the decision it
generates is uniquely fair. The empirical evidence demonstrates that it is possible because of
framing or polarization effects that two different deliberative democratic procedures could
234
RAZ, supra note 50, at 242.
It might, however, have to do with equality depending on the substance of the extreme
beliefs.
235
Gingerich
71
generate different outcomes even though both met all of the requirements of political equality
and fairness and both adhered to the requirements of public reason.236
The argument that choices about ideological structures are inevitably subject to
ideological infiltration does not only undermine arguments that ratification by democratic
deliberation generates normative reasons to treat law as authoritative, it also cuts against
epistemic arguments for democratic deliberation and as wisdom-pooling arguments that
democratic deliberation is desirable because it successfully synthesizes knowledge from a large
number of people. Such theories say that democratic deliberation is good because it produces the
right political outcomes. But the arguments that these theories rely on about the outcomes of
democratic deliberation are wrong if deciding what sort of deliberative structures to create is an
ideological decision. Choosing one framework for deliberation and democratic decision making
rather than another will produce different outcomes. It cannot be the case that if Odd ideologues
secure a deliberative framework like Deliberation Sunday which subsequently ratifies the Odd
political agenda then a reason to accept Odd supported laws as true is thereby generated. This
would not only bootstrap reasons into existence but bootstrap political truth into existence—
simply by virtue of believing in the Odd ideology and securing an institutional framework
favorable to that ideology, the Odds would be able to generate reasons that their political beliefs
are true. Since ideological choices about political institutions result in certain outcomes rather
than others, and possibly radically different outcomes, there is no single answer or even a single
set of true political answers that deliberation can provide.
V. CONCLUSION
236
Of course, anyone who thinks that reason is unitary might not think that this is possible. See,
e.g., G.W.F. HEGEL, PHENOMENOLOGY OF SPIRIT # (A.V. Miller, trans., 1977)
[PARENTHETICAL].
Gingerich
72
If people were really economically rational and their preference were not disturbed by
feelings of empathy and solidarity, then it might make perfect sense to say that democratic
deliberation can generate laws that are prima facie (or even all-things-considered) legitimate. 237
People would engage in deliberation which would solve any problems about limited access to
information they might have, allow them to see how strong their preferences are compared to the
preferences of other people, and agree to cut bargains with other people so as to produce Paretooptimal political outcomes. However, if people are influenced by various and idiosyncratic
motivational structures like solidarity, empathy, and communication, it is much harder to believe
that democratic deliberation necessarily produces legitimate laws.238 People do not merely have
their preferences distorted by deliberative structures; rather, while they remain themselves, they
are “changed to the very marrow of [their] bones” by the process of deliberation.239 As Johnson
notes, “reasonable pluralism” should then be an outcome of, rather than a precondition for,
democratic deliberation.240 Nevertheless, the theoretic refusal to relax rational expectations
assumptions has led to contorted theoretical models.241 Given this understanding, it is better to
see the justification for majority rule as resting on independent principles of justice.242
However, Philip Soper raises the point that if our theory does not accord legitimacy to
laws, this is a reason to try different theories because law so persistently claims to have authority
I say “might” both because the objectivist challenge would still obtain, and because it is not
clear to me that the requirement of public reason is compatible with a view of people as
economically rational actors.
238
That is, laws that we have a reason to believe to be legitimate by virtue of their democratic
ratification.
239
JOHANN WOLFGANG GOETHE, ITALIAN JOURNEY 147 (W.H. Auden & Elizabeth Mayer, trans.,
1970).
240
Johnson, supra note 216, at 177.
241
BRYAN CAPLAN, THE MYTH OF THE RATIONAL VOTER: WHY DEMOCRACIES CHOOSE BAD
POLICIES 207 (2007).
242
See, e.g., JOHN RAWLS, A THEORY OF JUSTICE, 313 (rev. ed. 1999) [PARENTHETICAL].
237
Gingerich
73
and the intuition that the law has authority is so strong.243 One possibility is a social contract
model of legitimacy rather than a democratic deliberation model. Paul J. Weithman notes that
while the democratic deliberation model has recently become quiet popular in the academy, the
social contract tradition is just as rooted in American history and just as viable.244 Another
alternative, perhaps more intellectually satisfying and certainly more responsive to inequities of
the North American and global status quo, is to conceive of democracy as less confined to state
institutions and claim that democracy consists in a constant critical questioning of empire and
elite power.245 On this account, democracy is critique, the democratic impulse calls into question
every assertion of legal authority or legitimacy, and deliberative democracy is the constant,
insistent instantiation of this critical consciousness.
243
Philip Soper, Legal Theory and the Claim of Authority, 18 PHIL. & PUB. AFF. 209, 237 (1989).
Paul J. Weithman, Contractualist Liberalism and Deliberative Democracy, 24 PHIL. & PUB.
AFF. 314, 343 (1995).
245
See CORNEL WEST, DEMOCRACY MATTERS: WINNING THE FIGHT AGAINST IMPERIALISM, 1622 (2004) [PARENTHETICAL]; see also LYOTARD, supra note 10, at 65-67 (suggesting that the
end of discussion is not consensus but “paralogy,” which is a sort of “inventor’s knowledge” that
destabilizes and reinvents systemic, institutionalized knowledge); Habermas, Constitutional
Democracy, supra note 5, at 774 (suggesting that a democratic constitution is a “traditionbuilding project with a clearly marked beginning in time” and that later generations have the task
of critically actualizing the norms laid out in the founding document).
244