Deliberation as Self-Discovery and Institutions for Political Speech
| DOI | 10.1177/0951629807077573 |
| Author | Catherine Hafer,Dimitri Landa |
| Published date | 01 July 2007 |
| Date | 01 July 2007 |
| Subject Matter | Articles |
DELIBERATION AS SELF-DISCOVERY
AND INSTITUTIONS FOR POLITICAL SPEECH
Catherine Hafer and Dimitri Landa
ABSTRACT
We present a game-theoretic model of the social dynamics of belief change in
which the (relevant) logically non-omniscient audience becomes convinced that
the speakers’ messages are ‘true’ because its own prior beliefs logically entail them,
rather than – as in cheap-talk models – because the speaker is (endogenously)
trustworthy. We characterize the equilibria of the game and consider how their
aggregate informational properties change with the variation in the institutions
determining the ability of the speakers to reach their audience. We find that for
plausible restrictions on the distribution of arguments and on the corresponding
policy preferences in society, the informationally optimal institutions are first-best
implementable, inegalitarian with respect to the resource allocation across speak-
ers, and assign priority to the (more) extreme argument- and policy-holders.
KEY WORDS .deliberation .non-Bayesian learning .institutions
Introduction
One of the central claims of deliberative democratic theory is that deliberation –
an unforced exchange of reasons aimed at the development of the most justified
policy position or judgment – is a critical element of democracy. Provided that
our judgments are responsive to the morally and politically relevant information
communicated in the course of providing reasons for interlocutors’ positions,
individual and collective decisions untested by such exchanges are likely to be
deficient in the eyes of some, if not all, citizens. If so, then these decisions –
and, by extension, the institutional mechanisms generating them – may, in the
absence of deliberation, fail to gain the requisite degree of political legitimacy
(Manin, 1987; Habermas, 1996; Cohen, 1997).
To the extent that the quality of political decisions and the legitimacy of
political practices are fundamental concerns of both political philosophy and
This article was initially circulated under the title ‘Deliberation as Self-Discovery’ (2003). We
received valuable comments from audiences at MEDS, Kellogg School of Business (Northwestern
University), Department of Social Sciences, Carnegie Mellon University, and at NYU Law School,
and benefited from comments and conversations with David Austen-Smith, Ethan Bueno de Mes-
quita, Steve Callander, John Ferejohn, Sandy Gordon, Lewis Kornhauser, and John Patty.
Journal of Theoretical Politics 19(3): 329–360 Copyright 2007 Sage Publications
DOI: 10.1177/0951629807077573 Los Angeles, London, New Delhi and Singapore
http://jtp.sagepub.com
political science, an account of deliberation and of its determinants deserves a
central place on the agenda of both. But while deliberation has, indeed, become
one of the most actively explored issues in philosophy (Gutmann and Thomp-
son, 1996; Bohman and Rehg, 1997; Elster, 1998; Macedo, 2000), it has
received relatively little attention in rational choice social science. Not surpris-
ingly, the extensive conceptual argumentation for deliberation has been matched
with little headway in analyzing the optimal implementable institutional
mechanisms that would substantiate in non-ideal political environments the
properties claimed on its behalf (Johnson, 1998; Hardin, 2000).
Our aim is to contribute to closing the gap between these analyses of delib-
eration in two ways: by proposing an analytical model of deliberative practice
that captures some of the key features of empirical deliberations falling outside
the existing rational choice models, and by analyzing the equilibrium properties
of a particular class of institutions in its context.
Arguments and Disagreement
The standard rational choice model of verbal communication – cheap-talk sig-
naling – examines the extent to which agents can successfully communicate pri-
vately known empirical facts that are relevant to some collective decision.
Versions of the cheap-talk model fix primitive preferences over outcomes and
address the impact of communication on induced preferences over instrumental
choices. Individuals may exchange messages about the content of their private
information and, because that information is private, they can, within limits,
engage in misrepresentation of their information. The determination of the extent
or quality of deliberation in such models, then, turns largely on the credibility of
the sent messages, which is induced by the proximity of the sender’s primitive
preferences to those of the individual receiver (Crawford and Sobel, 1982; Austen-
Smith, 1990) or, in the case of a group of receivers, to those of the expected
majorities within the group (Meirowitz, 2007), and by the sender’s pivotalness both
as a source of information and as a voter (Austen-Smith and Feddersen, 2002).
When the information transmitted in deliberation consists of privately known
empirical facts relevant to the application of particular moral principles (self-
interest, fairness, etc.), cheap-talk models capture important features of the delib-
erative interaction.However, as the following examples drawn from recentpublic
discourse illustrate, deliberation commonly takes a form that cannot plausibly be
construed as an exchange of previouslyunknown empirical evidence:
(1) Although statistical justifications of racial profiling are not racist, they are wrong
because they ignore theindirect costs of a practice that relies on racial selectivity:
Statistics abundantly confirm that African Americans – and particularly young
black men – commit a dramatically disproportionate share of street crime in
the United States. This is a sociological fact, not a figment of the media’s
330 JOURNAL OF THEORETICAL POLITICS 19(3)
(or the police’s) racist imagination ... [Nonetheless] politicians ... must be
willing to demand equal treatment before the law even under circumstances in
which unequal treatment is plausibly defensible in the name of nonracist goals
...[The reason is] the recognition ...that the presence of a racial factor in gov-
ernmental decisionmaking gives rise to the presumption that officials may be
acting in violation of someone’s civil rights. (Kennedy, 1999)
(2) The consenting adults argument for the legal recognition of gay marriage is con-
sistent with the legal recognition of incest, and so cannot be a sufficient policy
principle:
If it’s just about two consenting adults who love each other, why would we
then deny a father and a daughter who’s of age to get married? ...Why do we
ban such things? It’s because it’s not good for society. (Wood, 2003)
(3) Although the critics of the legal recognition of gay marriage are correct about the
status of the consenting adults argument with respect to incest, it need not follow
that the consequence is, therefore, necessarily socially deficient:
Santorum and Scalia and the rest are absolutely correct that there is no princi-
ple which would deny two brothers from marrying that can allow two unre-
lated gay men from marrying. There is only taboo and prejudice .... All of the
arguments about deformed children and abusive parents are red-herrings since
the law could certainly protect against those problems while still allowing for
gay same-sex siblings to do whatever they want. (Goldberg, 2003)1
Although these arguments include references to empirical facts, states of the
world, etc., the deliberation they connote is, in effect, an exercise in ‘experiential
self-discovery’. When these arguments convince, they do so because they invoke
as antecedents knowledge that hearers already have or that can be made self-
evident to them as its deductive or inductive extensions.
2
In such deliberation,
1. Other, somewhat more esoteric, examples include most moral philosophical arguments, includ-
ing John Rawls’ (1978) for the uniquely fair character of his principles of justice. The example of
Rawls’ argument is particularly interesting in the context of our article if one interprets the device of
the original position as de-biasing the perspectives of individual citizens, whose social and economic
engagements make it difficult for them to see their way to the correct principles of justice (Hurley,
2000). On this account, as in our model, the normative argument is addressed to those whose ability
to draw implications from their commitments (in Rawls’ case, to fairness) is suitably impaired.
2. Calvert and Johnson (1998: 6) emphasize the prominence of non-cheap-talk arguments in moti-
vating their view of deliberation as a mechanism of coordinating expectations in the context of multi-
ple distributionally non-equivalent equilibria. We believe that deliberation quite likely serves that
function but, in our view, it does so through preference transformation – effectively re-defining the
initial game (though not the game-form). Significantly, our model of this process meets Calvert and
Johnson’s criticisms of preference-transformation views of deliberation: we assume neither the uni-
versal post-deliberative consensus (as in most primitive preference transformation theories) nor that
cheap-talk communication of private information exhausts the deliberative practice.
HAFER & LANDA: DELIBERATION AS SELF-DISCOVERY 331
To continue reading
Request your trial