Representative Democracy and Democratic Citizens: Philosophical and Empirical Understandings

Representative Democracy and Democratic
Citizens: Philosophical and Empirical

Understandings
SIDNEY VERBA
The Tanner Lectures on Human Values
Delivered at
Brasenose College, Oxford

May 10 and 11, 1999
Sidney Verba is Carl H. Pforzheimer University Pro-
fessor and professor of government at Harvard University,
as well as director of the Harvard University Library. He
was educated at Harvard College and at Princeton Univer-
sity, where he received his Ph.D. In 1993 he was awarded
the James Madison Award by the American Political Sci-
ence Association (APSA) for a career contribution to polit-
ical science. He is a member of the National Academy of
Sciences, a fellow of the American Academy of Arts and
Sciences, and has been both a fellow of the Center for Ad-
vanced Studies in the Behavioral Sciences and a Guggen-
heim Fellow. He is the author or co-author of many
published works, including Voice and Equality: Civic Vol-
untarism in American Democracy (1995, with Kay L. Schloz-
man and Henry E. Brady), Elites and the Idea of Equality: A
Comparison of Japan, Sweden, and the United States (1987),
and The Civic Culture: Political Attitudes and Democracy in
Five Nations (1963, with Gabriel A. Almond). Participa-
tion in America: Political Democracy and Social Equality
(1972, with Norman Nie) won APSA’s Kammerer Award
for the best book on American politics, and The Changing
American Voter (1976, with Norman Nie and John Pet-
rocik) won the APSA award for the best book in political
science.
The subject of human values is broad and daunting. When I began
to consider what I might say on this topic for the Tanner Lectures,
I looked at earlier lectures in this series. The lecturers were usually
political philosophers or legal scholars. They analyze systems of
values and beliefs and argue for one system over another. They
draw on the works of other thinkers, on the writings of other phi-
losophers, and on the opinions of judges for justiŠcation of the sys-
tem they prefer.
Lurking in the background of many of these accounts, espe-
cially those that deal with my area of concern, democratic gover-
nance, is another set of values and beliefs: those held by the public.
Reference is made to a “public philosophy,” to “common under-
standings in a community,” to a “public culture,” to the common
will, or to—the term used most often though with very varied
meanings—public opinion. Michael Sandel, in his book Democ-
racy’s Discontent, writes of the “public philosophy,” by which he
means “the political theory implicit in our practice, the assump-
tions about citizenship and freedom that inform our public
life. . . .”1 Michael Walzer in Spheres of Justice writes of something
similar: “sensibilities and intuitions” shared by those in a political
system.2 For theorists of democracy, like Sandel and Walzer, who
ground their democratic philosophy in an interpretation of moral-
ity as practiced in particular places by particular people, the values
and beliefs of the public would seem to be especially relevant.
Walzer, for instance, rejects a morality based on “what rational in-
dividuals . . . under universalizing conditions” would choose, in
favor of that which would be chosen by “ordinary people, with a
1 Michael Sandel, Democracy’s Discontent: America in Search of a Public Philosophy
(Cambridge, Mass.: Harvard University Press, 1993), p. 4.
2 Michael Walzer, Spheres of Justice: A Defense of Pluralism and Equality (New York:
Basic Books, 1983), p. 28.
[231]
232 The Tanner Lectures on Human Values
Šrm sense of their own identity, with their own goods in their
hands, caught up in their everyday troubles. . . .”3 Walzer cer-
tainly seems to invoke a real public with a real set of opinions.
What about those philosophers who seek to develop general,
perhaps universal, principles of political morality on the basis of
logic and thought experiments, rather than information about the
real world of citizens? John Rawls, in The Theory of Justice,4 starts
in an imagined world of rational people in a contrived state of ig-
norance as to who they are. Even here, one has the sense that one
will have sometime to face the issue of what people will actually
do and actually think. Rawls assumes a citizenry capable of reason,
of cooperation, and of conceptions of the good; reasoning, cooper-
ative, and moral citizens. These attributes are discussed at a most
abstract level, but they are attributes that are (or at least could be)
real. Rawls moves closer to the public in Political Liberalism, where
he asks how one can have a stable society when there exist multiple
conceptions of morality that are not compatible with each other.
In this book, the public is a more palpable entity. He states in the
opening of the book, “We start then, by looking at the public cul-
ture itself as the shared fund of implicitly recognized basic ideas
and principles.”5
A public philosophy is, thus, not a philosopher’s philosophy. It
seems to be something out there; something that needs to be dis-
covered and taken into account. But it is hard to deŠne and hard
to Šnd. “A public philosophy is an elusive thing, for it is con-
stantly before our eyes. It forms the often unrešective background
to our political discourse and disputes.”6 In saying this, Sandel
echoes comments by James Bryce a century ago about public opin-
ion in America. Bryce noted that nowhere is public opinion more
3 Ibid., p. 5.
4 John Rawls, The Theory of Justice (Cambridge, Mass.: Harvard University Press,
1971).
5 John Rawls, Political Liberalism (New York: Cambridge University Press, 1993).
6 Sandel, Democracy’s Discontent, p. 4.
[Verba] Representative Democracy 233
important than in America. He also noted that it was “difŠcult to

ascertain.”7
This is one reason for the ambivalence about public opinion
one Šnds in scholars, political actors, and the public itself. Public
opinion is important; at the most fundamental level democracy
rests on it. If democracy means anything, it means that in some
way public values and preferences ought to have an impact on gov-
ernment policies; more strongly, perhaps, ought to determine
what governments do. But public values and preferences are, as
Bryce says, difŠcult to ascertain. How can public opinion guide
policy when it is not clear what it is? And, even if we could as-
certain it, should it be a guide? Is it not too incompetent, ill-
informed, changeable, and irresponsible? You cannot follow it
and, if you could, you shouldn’t. Yet it is sovereign.
Empirical social scientists also address the issue of public be-
liefs and values and their relation to political life. There are nu-
merous empirical studies of citizen values on social and political
matters, of the ways citizens express those values. Many focus on
the speciŠc policy preferences of citizens, but there are also studies
of the broader and more fundamental values they hold. These
studies do not propose or justify one value system over another,
but seek to describe and explain systems of values. Normative phi-
losophers and empirical social scientists are, in some sense, in dif-
ferent businesses. Normative philosophers are concerned with the
structure of complex normative systems, and with the justiŠcation
of such systems. They pitch their arguments against alternative
value systems, claiming that the alternatives are wrongly derived,
or internally inconsistent, or unfortunate in their consequences—
or all of these. Systematic empirical researchers usually have a dif-
ferent agenda. They are concerned with method and inference.
They care about such matters as how one measures values or
behavior in a population. They also are concerned with causal
7 Quoted in James A. Stimson, Public Opinion in America: Moods, Cycles, and Swings
(Boulder, Colo.: Westview Press, 1991), p. 8.
explanation: What causes the values that people hold? What are
the consequences of those values? They pitch their arguments
against alternative methods or alternative descriptions or alterna-
tive explanations.
I hasten to add that we ought not to take the distinction too far.
Few normative philosophers ignore facts, and some pay a lot of at-
tention. Few empirical researchers deny the normative founda-
tions or implications of their work. Facts and values are not as
sealed off from one another as some once tried to make them. But
normative philosophers and empirical social scientists seem ani-
mated by different concerns and largely go about their businesses
differently.
But for both, the issue of the public philosophy or public opin-
ion is a problem. How then does one ascertain what the public
thinks and how the public thinks? Does it even make sense to ask
about such matters? Normative philosophers Šnd public opinion
elusive. No systematic social scientist who has worked on the sub-
ject would disagree.
There are several reasons for this elusiveness. Citizen values are
not immediately apparent. Philosophers tell us about their nor-
mative systems. That is their business. Citizens have no such obli-
gation. They do not spend time ruminating about their normative
systems, nor do they write treatises explicating and justifying
them. If one is interested in such values, one has to go out and Šnd
them.
There is a further problem in the empirical study of citizen
values and preferences: the fact that they must be studied on two
levels, that of the individual and that of the aggregate. The conse-
quence of citizen values for political outcomes depends not on in-
dividual values and preferences but on the aggregate of values and
preferences manifested in various ways—as votes, as public opin-
ion, as communications to political elites, and so on. This second
level of values—the values of aggregates—adds additional difŠ-
culty to the observation of citizen attitudes. Measuring the values
of an individual is hard enough. Aggregating them, especially if

they vary across individuals in both content and intensity, is
harder still.
How then can we know about this elusive phenomenon? The
answer takes us into questions of methodology, not the best sub-
ject for a public lecture on a big and important substantive issue.
But at least something needs to be said about it, for public opinion
is not right in front of our eyes—or, to put it bluntly, if you think
it is, as many do, you are wrong. It needs to be found, and what is
found depends on the method used. I do not intend to discuss
methods in detail, but need to digress for some general comments.
It is, in fact, not that much of a digression from my substantive
theme of how ordinary citizens think about political matters since
what I say about method is relevant to how citizens learn about
and reason about such matters.
A Consideration of Method
Let us distinguish broadly between two kinds of methods: system-

atic social scientiŠc methods and everyday commonsense meth-
ods. Systematic social science is, or at least should be, interested in
general knowledge about social and political reality; that is, in de-
scriptive and causal inference about the real world. I say descrip-
tive inference, not description, to distinguish that which we
observe from the more general inferences we make on the basis of
those observations. We observe some people doing things, saying
things, reacting in certain ways. What we observe depends on how
we observe it (are they saying things spontaneously that we over-
hear, are they answering our questions?). And it depends on whom
we observe (how were the people we observe selected?). Thus, sys-
tematic social science depends on good measures that get at the re-
ality in which we are interested and on a selection procedure that
allows us to assume that the people we observe represent the larger
population to whom we wish to generalize. If the selection is bi-

ased—gets us unrepresentative people because of the way we se-
lect them—we cannot generalize. The model for this is the
ubiquitous social science tool used for studying the public, the
sample survey.
Beyond descriptive inference, systematic social science aims at
causal inference: how things go together, what causes what. We
want to know not only what people say and do, but why they say
and do it. Why they vote as they do; why they prefer one policy
over another. And we want to know the consequences of what they
say and do. Does it matter how they vote? The model for answer-
ing such questions is the experiment. A perfect experiment can
answer the why question, because it can take into account alterna-
tive explanatory factors and isolate the ones that provide us with
an explanation. The ubiquitous surveys are analyzed as if they are
experiments. They are not; they don’t even come close. But good
analysis can get us somewhat closer to causal explanation.
Commonsense understanding of the public is less formal and
self-conscious about measurement or sampling. It relies more on
the vivid anecdote than on the systematic sample; on information
at hand, not on information carefully selected; it often relies on in-
trospection (“I am a member of this community so I know how we
feel”); it may use indirect evidence about the public (what others
say about it without asking how they know); it is intuitive.
Many—by no means all—philosophical works on democracy
that deal with the public or a public philosophy or ethos take this
approach. They are not too systematic in their sources. In many
cases, a public ethos may be mentioned or invoked, the evidence
for it being the intuitive understanding of the author. Philoso-
phers who ground their work in actual practice—a Walzer or
Sandel—often provide a rich but casually selected array of evi-
dence about those practices. Writings about the public are in-
voked: the writings of literary observers, social critics, earlier
philosophers, journalists, and others who put their thoughts on

paper. Anecdotes and interesting examples as to beliefs and behav-
iors abound. Reference is often made to court cases in which
judges base decisions on general values or community standards.
But if one looks to where the judges get their information about
the public, it usually comes from their own understanding of what
those standards are. A writer like Rawls, who begins from a more
abstract logical analysis, also Šnds little in social science data. He
invokes a vague common understanding of “the most reasonable
conception of the person that the general facts about human na-
ture and society seem to allow.” As for a social science contribu-
tion: “ . . . beyond the lessons of historical experience and . . . bits
of wisdom,” he tells us, “there is not much to go on” (Political Lib-
eralism, p. 87).
The Problem with Unsystematic Observation
The main fault of such unsystematic observation is that it is sub-

ject to the main impediment to valid descriptive inference about
the public: selection bias. In the absence of systematic observa-
tion, it is easy for philosophers or politicians or other observers of
society to Šnd that which is most congenial to their position. This
is easily the case if one looks inside oneself for these values. It is
also easily the case if one extrapolates from those one knows best—
either personally or from their writings. And it is also easy if one
interprets the values of a culture; since cultures may have many
values, one can always Šnd those that suit one’s preconceptions. As
Brian Barry puts it: “ . . . claims to derive conclusions from the al-
legedly shared values of one’s society are always tendentious. If
they were not, it would have to be regarded as a remarkable coin-
cidence that the shared values a political philosopher says he has
detected always happen to lead to conclusions that he already
supports.”8 Barry says the task of understanding the public is a fu-

tile one; that there is nothing there to understand. I think this
goes too far. There is something there, but it is not easy to get at.
Survey Research and the Public
The main technique for getting at public opinion is the sample

survey. Public opinion polls are ubiquitous in politics and in the
press, as well as in the social sciences. Why are they used so much
by social scientists? One answer might be Abraham Kaplan’s law
of the hammer: give a little boy a hammer, and suddenly every-
thing around needs pounding. Give a social scientist the social
survey, and suddenly every topic needs a survey. But there are bet-
ter reasons. The main value of a good survey is that it is represen-
tative of an entire population. A good random sample eliminates
selection bias and allows us to generalize from the units observed
to the larger universe from which the units were chosen.
But about what does it let us generalize? Usually about rather
superŠcial measures. In a systematic random sample, one cannot
probe the complexity of individual value patterns. An individual’s
beliefs and values relevant to democracy are hard to elicit. They
would be revealed only across a range of choice situations, observed
over time, in varying contexts. Surveys, with their simple ques-
tions asked uniformly of a sample, run the danger of oversimplify-
ing what is there, missing the main points, and perhaps creating
rather than recording the reality we are studying. Furthermore, we
care about the aggregate, not only the individual; about public
philosophy or public opinion, not the values or thought processes
of one person or another. The complex patterns of beliefs and val-
ues held by an individual do not add up easily to form an aggre-
gate. You might understand individuals by writing biographies
8 Brian M. Barry, Justice as Impartiality (Oxford: Oxford University Press, 1995),
p. 5.
describing what each believes and how they got there. But it is not
easy, probably impossible, to sum up a multitude of such biogra-
phies to understand a large population. To add up values across in-
dividuals requires that we reduce those values to comparable and
countable items and, in so doing, we lose some of the substance of
that which we count.
The problem is a general one—not only for the social scientist
but for actual political choice and action situations across a large
population. James C. Scott has written most interestingly about
the attempt by governments to control societies by making them
“legible”—that is, by standardizing things (people, towns, trans-
actions) so that they can be counted or regulated. People and ob-
jects need to be put in categories (laws of course depend on
categories), observational techniques need to be standardized, ac-
counting schemes need to be devised. All of this smooths out the
edges of a reality that is much more complex and contextually var-
ied. Scott focuses on the pathologies this can create but realizes
that some such systematization is needed in any complex society.
The problem Scott identiŠes in relation to the action of states is
a well-known problem in the social sciences. He calls his book See-
ing Like a State.9 If it were a book on social science methodology, it
might be called Seeing Like a Social Scientist. Social scientists (or at
least some of us) consider the task of social science to be one of
making society legible, to Šnd order out of the chaotic business of
everyday life. We, like bureaucrats, tax collectors, and planners—
among other unpleasant people—like to have people neatly cate-
gorized. But, just as Scott worries about the pathologies associated
with this as states use such information to impose a rigid order
to society that is destructive of ordinary practices, so one might
worry that systematic social science bleaches out reality in its at-
tempt to systematize.
In democratic politics, the best example of the simpliŠcation
9 James C. Scott, Seeing Like a State: How Certain Schemes to Improve the Human Condi-
tion Have Failed (New Haven: Yale University Press, 1998).
of values and preferences is the vote. No one has found a substitute

as yet. Millions of individuals—each unique, with a complex set
of values, beliefs, and interests—can express a standardized prefer-
ence simultaneously. No form of citizen control would be possible
if it were not for that standardization. If all wrote long letters
about what the government ought to do, they could not be trans-
lated into action. Votes capture the entire voting body and reduce
it to a simple and coherent aggregate. You can vote or not vote
and, if you vote, you can choose from a limited set of options. The
multitudinous and varied reasons why citizens decide to vote
rather than stay home and then decide for whom to vote are re-
duced to a couple of stylized and blunt choices.
Surveys are better than votes in terms of the information they
give us about the individual. Indeed, they are used to interpret the
vote—to Šnd out why people voted as they did. But they still pro-
vide only limited insight into the public’s values and beliefs.
Nevertheless, surveys are as good as it gets when one wants to
typify a larger population in a valid manner. It is the technique
from which I will draw my evidence about the public, šawed
though the technique may be. Furthermore, good academic sam-
ple survey research—research that is not aimed at providing quick
descriptions of public opinion on Kosovo or some other current is-
sue—can go well beyond the simple report of what percent of the
public favors what. Survey research can deal with variation in
opinion, which is often more interesting than where the majority
opinion lies. Furthermore, the answer to a single question may tell
little. Such answers are sensitive to question wording. But pat-
terns across sets of questions are more reliable and can reveal a lot
about how people think. The fact that opinions change over time
and vary with question wording makes many doubt the solidity
and usefulness of survey results. But, properly analyzed, such
change over time and variation with the stimulus presented to the
respondent can give a good deal of insight into the nature of pub-
lic opinion.
Furthermore, ordinary surveys can be supplemented with

longer in-depth interviews or group discussions to get us closer to
an understanding of citizen reasoning processes. In addition, sur-
vey researchers have, in recent years, expanded their repertoire of
methods. Computer-assisted telephone interviewing allows one to
explore the reasoning behind a particular answer by varying the
source presented to the respondent, by varying the nature of the is-
sue, by asking and allowing citizens to rešect on their views, by
challenging views that they have previously expressed, and the
like. In this manner, something approaching a dialogue with an
individual is possible; more important, dialogues with samples of
individuals are possible.
The studies I report will come almost entirely from the United
States. The United States is a curious place with strange and some-
times frightening customs, some of which have been recently too
publicly displayed. I focus on it because that is where most of the
data I wish to refer to have been gathered. I am not sure of the gen-
eralizability of these data to other democracies—each of which is
curious in its own way. But my reading of such works as the Wil-
liam Miller, Annis May Timpson, and Michael Lessnoff work on
British political beliefs Šnds much that is consistent with what I
will discuss.10 I think there are general lessons here, but we must
be cautious in generalizing.
Democratic Citizens
Democratic Dilemmas
Having said something about the methods for ascertaining the

public’s values and preferences, let me turn to what can be learned
10 William L. Miller, Annis May Timpson, and Michael Lessnoff, Political Culture
in Contemporary Britain: People and Politicians, Principles and Practice (Oxford: Clarendon
Press, 1996).
about the public from surveys. The substantive issue I shall con-
sider relates to a major democratic dilemma associated with citi-
zen participation in a democracy. It is interesting how commonly
the words “democracy” and “dilemma” seem to go together.
When not acting as a social scientist, I am the director of the Har-
vard University Library, and I sometimes look at our electronic
catalogue just to see what one can learn from the catalogue re-
cords. I found over sixty records that contain the words “democ-
racy” and “dilemma.” Apparently authors like to put the terms
together. Democracy seems to pose a dilemma in relation to,
among other things, religion, bureaucracy, secrecy, science, social-
ism, nationalism, funding of the arts, school desegregation, au-
thority, national defense, in a number of titles Germany, and—for
whatever it means—the future. This is not surprising. Democracy
as a system spawns dilemmas and contradictions. The very term
“democratic government” is contradictory. “Democracy” refers to
the people, to control from below, to procedures for giving citizens
a voice over decisions. “Government” refers to policies, to author-
itative decisions, to control from above, to the means for making
effective policies. This is the dilemma I wish to explore from the
perspective of the role of the public in democratic governance: can
one have a democracy that gives the citizenry substantial and
equal voice over public affairs and at the same time produces effec-
tive and just policies?
Let me spell this out. Citizen voice and participation are at the
center of the democratic half of democratic government. Through
participation citizens express their preferences to governing ofŠ-
cials and induce them to respond to those preferences. They send
information about themselves (who they are, what they want,
what they need) and apply pressure on ofŠcials for some response.
Citizens do this in many ways: by voting, working in political
campaigns, writing letters, taking part in community actions,
protests, and on and on. The sum of these activities represents the
way in which democratic voice is expressed.
Democracy depends on such voice. It also depends upon equal

voice. One of the bedrock principles of democracy is the equal con-
sideration of the needs and preferences of all citizens, a principle
embedded in the notion of one person, one vote. Note that I do not
speak of equal response or equal beneŠt from the government.
That may be much too much to ask. Even if citizens spoke with
equal voice and governing ofŠcials gave each an equal hearing,
equal response would not be possible. Government resources are
limited; citizen preferences conšict with each other; everyone can-
not get an equal response. If they cannot get an equal response, can
they get a fair one? It is hard to know exactly what a fair response
would be, but let us assume it to mean, among other things, that
those who get somewhat less will not consider the result unjust (at
least, not wildly unjust). Without an equal voice and an equal
hearing, there is no chance of getting close to an equal response
and very little chance of an outcome that will be considered fair
and just. From this perspective equal political voice fosters beneŠ-
cial outcomes. Put another way, the outcomes will be better—
fairer—if everyone gets a chance to be heard. For one thing,
insofar as democratic outcomes are supposed to rešect the prefer-
ences and needs of the citizenry and insofar as equal political voice
fosters the communication of those preferences and needs, proce-
dures maximizing equal voice lead to beneŠcial outcomes. In ad-
dition, equal political voice should foster more supportive
political views among citizens. People who have a voice are more
likely to accept policies of which they disapprove than are those
who do not. And political voice—the chance to participate—also
fosters a more enlightened citizenry. All these things, and more,
suggest that a citizenry that has a real and equal voice in policy-
making is a better citizenry, and where that happens there will be
a better polity.
But there is another side to the story. The desirability of citizen
voice and, especially, equal voice depends on some qualities of the
citizenry. Why do we want to give voice to citizen preferences and
needs? And why an equal voice? For some the answer lies in the
belief that all people are equal before God. A more secular set of
reasons expressed by such democratic thinkers as Robert Dahl and
John Rawls is that people are reasoning beings, capable of under-
standing the world; and that they are moral beings, capable of
conceptions of the good. What does this mean for the ordinary cit-
izen, and why is it the principle on which equal citizen voice is
grounded? For one thing, it means that citizens should be good (or
at least adequate) moral reasoners. Their preferences and values,
whatever their substance, should have some coherence and be
somewhat stable. If preferences and values change from moment
to moment, or if they are incoherent and self-contradictory, it is
hard to know how they can be given meaningful expression. Also,
citizens should be good (or at least adequate) social scientists.
They should have enough information to know how to pursue
their goals through politics. They should be capable of making
causal inferences about how their activity might lead to desired
outcomes (for instance, to be able to Šgure out, in the light of their
preferred policy outcomes, whom to vote for or how one policy
rather than another will affect their welfare). If they cannot reason
about the political world, what do their expressed preferences
mean?
Thus, democratic voice would seem to rest on some capacity of
citizens to be moral reasoners and social scientiŠc reasoners. It may
rest on another quality of ordinary citizens: their willingness to
transcend their own values and preferences. There are two compo-
nents of this willingness. One is a willingness to look beyond their
own narrow self-interest to consider the common good and the
other their willingness to commit to (or at least accept) principles
of fair democratic procedure that allow a multiplicity of view-
points or doctrines to be expressed. Some would argue that the
Šrst is not necessary: let self-interest rule and an invisible hand
will provide good social outcomes. This may work in the market-
place of commodities; I believe (though I cannot argue the case
fully here) that it does not apply in the making of public policy
through the political process. The second is more unambiguously
tied to democracy. Rawls refers to this as a free-standing overlap-
ping consensus on a democratic process that involves tolerating al-
ternative doctrines. This, as he notes and as many democratic
theorists have noted before, is needed to maintain a stable democ-
racy given the inevitable plurality of competing doctrines sub-
scribed to by citizens in a democracy.
But what if citizens are not up to the task: if they are inade-
quate moral reasoners or social scientists—if they don’t know
what they want or how to get it? What if they can’t be relied upon
to go beyond their own preferred value system to allow others to
exist? Democratic voice becomes meaningless or undesirable. Or
what if citizens are unequally equipped for moral reasoning or so-
cial science analysis? Some are competent, others not. And what if
they are unequally committed to democratic procedures? Equal
democratic voice might be meaningless or undesirable. These con-
siderations have led many philosophers of democratic government
to express skepticism about the public, about too much democ-
racy, about populist democracy. The voice of the people, un-
checked, coming equally from those who might be expected to be
more democratically competent and supportive of democratic
principles and from others who are less so, might be dangerous for
democracy. And systematic social science studies have supported
that position. Yet without voice and without equal voice, what
happens to the equal consideration of the needs and preferences of
all?
This is the democratic dilemma of my lectures. Equal citizen
voice is needed for equal consideration of needs and preferences.
Equal citizen voice may be disruptive of effective government and
dangerous to democracy. I want to look more closely at what we
know about citizen voice and public opinion to see if this is the
case.
In this look at citizens, I am interested less in the substance of
what citizens think—the content of their values or their policy

preferences—than in how they think. I want to see how they oper-
ate as normative political philosophers; to consider, that is, the
structure of their systems of political preferences and values. And
I want to see how they operate as empirical social scientists; to
consider, that is, how they obtain information about the political
and social world, and how they deploy it to make causal inferences
about how the world works.
Both views of the citizen as participant—citizen voice as a
good thing and citizen voice as a bad thing—can be found in the
literature and in our common understanding of politics. In the ab-
sence of closer data on what citizens do and how they reason, one
might hope for the best but worry about the worst.
Democratic Citizens: What Are They Like?
So let us look more closely at the data—with all due caution about
the weakness of systematic data but, I hope, with respect for what
it can tell us. I will proceed as follows. I want to begin with work
I have been doing for a number of years on political equality, more
speciŠcally equality in political participation. This is not exactly
the same thing as political voice but close to it if we think of par-
ticipation as the means by which citizens exercise their political
voice—by which they communicate their interests and prefer-
ences to governing elites and induce them to respond. It is empir-
ical research based on large-scale surveys of what citizens do, work
done collaboratively with Kay L. Schlozman and Henry E. Brady.
I want to summarize this research briešy and then connect it,
more than it has been connected, to normative issues of political
equality.
Our work begins with the basic principle of equality men-
tioned earlier: that in a democracy citizens ought to receive equal
consideration of their needs and preferences. I’d like to summarize

what we found about the extent of political equality and then re-
turn to the normative principle from which we begin to look more
closely at the desirability of political equality—particularly from
the perspective of the capacity of the ordinary citizen to under-
stand and reason about political matters.
Are Citizens Equal in Politics?
To begin with, as is well known—and as our studies conŠrm with

a mind-numbing amount of data—citizen voices are not equal.
Americans are as active, or substantially more active, than citizens
elsewhere. But what is distinctive about political participation in
America is that participation is so unequally distributed, hewing
more closely to the fault lines of social class. In the United States
the skew introduced by the relationship between high levels of ed-
ucation or income and high levels of political activity—a bias
characteristic of political participation in democracies around the
world—is especially pronounced.
Why is this the case? In thinking about why some people are
active while others are not, my colleagues and I have found it help-
ful to invert the usual question and to ask instead why people do
not take part in politics. Three answers immediately suggest
themselves: because they can’t; because they don’t want to; or be-
cause nobody asked.
• “They can’t” suggests a paucity of necessary resources—time

to take part, money to contribute to campaigns and other po-
litical causes, and skills to use time and money effectively.
• “They don’t want to” focuses attention on the absence of po-
litical engagement—lack of interest in politics or little con-
cern with public issues, a belief that activity can make little
or no difference, little or no knowledge about the political

process, or other priorities.
• “Nobody asked” implies isolation from the networks of re-
cruitment through which citizens are mobilized to politics.
These three components—resources, engagement, and recruit-

ment—explain the stratiŠcation in political activity. In general,
the advantaged members of society are more politically moti-
vated. Educated, afšuent people are more likely to be politically
interested and informed; to care about political matters. They also
have more resources: not time (which turns out to be equally avail-
able—or perhaps equally unavailable—to all), but they obviously
have more money. And they are better endowed with the necessary
skills, which come from good education, a high-level job.
And a word on the last reason for inactivity, “Nobody asked.”
Some political activity is spontaneous, but a lot is recruited. We
become active because someone asked us to. One might imagine
that recruitment might overturn the stratiŠcation derived from
resources. After all, political movements often deliberately mobi-
lize the poor, the disadvantaged—people who would otherwise be
inactive. But our research shows that, overall across the popula-
tion, recruitment merely reinforces the stratiŠcation of participa-
tion. Recruiters are rational. They look for those who can partici-
pate effectively—they look where the money is—and they bring
into politics people who would have been active anyway.
In recent years, the stratiŠcation in political activity deriving
from resource disparities has been growing, the major reason be-
ing the relative shift in the place of time and money as the most
important resource for political activity. Money has become more
important, time less. The reason lies in supply and demand: time
has become in shorter supply as more and more families include
two earners. In addition, the professionalization of political cam-
paigns and the increased role of television have made money more
valuable to politics. Campaigns want money to buy a computer
and database or television time, more than they want people who
can walk the streets or stuff envelopes.
This increases the stratiŠcation of political activity for the sim-
ple and obvious reason that money is a stratiŠed resource, time is
not. The rich have money; the poor don’t. All of us, rich and poor,
seem to have as much—or as little—time.
From the point of view of our evaluation of political inequality,
it makes a difference whether the inequality in political voice is
based on a disparity in motivation or a disparity in resources. If the
former is the reason—people prefer to stay home and watch televi-
sion rather than turning out to vote or going to a community
meeting; or they prefer a few more hours in the ofŠce rather than
getting involved in a political campaign—we would be less con-
cerned with the fact that their voice is not heard. If it is a matter of
“they can’t”—because of resource inadequacy—then there is a
more serious problem. Our research on citizen participation shows
that, though both motivation and resources play a role, the latter
seem to be a much greater inhibitor of activity than the former.
Such resource disparities are, however, hard to remove. One
can imagine limiting the use of money in politics. That is what
campaign Šnance legislation in the United States attempts. It is
ineffective because of the inventiveness of campaign managers.
Furthermore, the U.S. Supreme Court has severely restricted the
extent to which one can limit the use of money in politics, arguing
that it inhibits free speech. There are many who disagree with that
interpretation and who might argue that speech is freer when the
ability to speak is more equal. But the Supreme Court ruling
stands. In American English—I don’t know about British Eng-
lish—there is the saying that “Money talks.” It is truly appropri-
ate for American politics, for money talks loudly and receives
constitutional protection as a form of speech.
In any case, even if there were limitations on money, it would
be difŠcult to go beyond limiting money to limit other resources.
Few of us would think it appropriate to limit the use of one’s
cognitive skills (the ability to understand issues and see connec-

tions among them) or rhetorical skills (the ability to make a com-
pelling case).
Michael Walzer has written approvingly of what he calls com-
plex equality, where there are different standards of fairness in dif-
ferent spheres of life, where the spheres are somewhat insulated
from each other, and where people may be at the top of the heap in
one sphere and lower down in another. Our data on citizen partic-
ipation do show some difference of standards in different spheres:
the norms in politics are more egalitarian than in the economy.
One person, one vote rules in politics; one person, one dollar or
pound does not rule in economic matters. But the boundaries be-
tween the spheres are porous, as inequalities in the economy—in
money and skills—spill over and create inequality in the political
sphere.
Thus, when it comes to political activity, the three factors of
motivation, resources, and recruitment work to reinforce each
other and give a cumulative political advantage to those better off
in other ways. The result of our analysis of political equality might
be summarized in the following manner: citizen voice ought to be
loud, clear, and equal: loud so ofŠcials pay attention, clear so they
know what people want, and equal so that the normative idea of
equal consideration of interests and preferences can be satisŠed.
Our data show that the public’s voice is often loud, sometimes
clear, but rarely equal.
Do We Want Political Equality?
That is our story in a nutshell. But let me come back to the open-
ing normative principle and make it a question: should all voices
be equal? Let’s look at that from two perspectives: the substance of
what is communicated to the government in terms of citizen pol-
icy preferences and citizen need (when citizens use their political
voice, what do they communicate?) and in terms of the quality of
decisions made by the government (what contribution does citizen

voice make to the justice and effectiveness of government policy?).
On the substance of what is communicated by active citizens,
the crucial point is that unequal voice makes a difference. Suppose
some people are active and others not, but in terms of the charac-
teristics that are relevant to the content and implementation of
public policy—the needs and the preferences people have—the
active and inactive citizens are not different. Inequality in activity
would not make much difference, because the subset of citizens
who are active would communicate the same messages about their
needs and preferences to the government as would the inactive
portion of the population. Some well-known research by Ray-
mond WolŠnger and Steven Rosenstone shows that, if you com-
pare voters and nonvoters in terms of their policy preferences on
standard U.S. National Election Study questions (along the lines
of should the government see that everyone has a good job and
standard of living or stay out of such things), there is little differ-
ence between the two groups—suggesting that it does not matter
much that some vote and others do not.11
But suppose we look at other political activities than the vote:
activities like working for a candidate, writing a letter to an ofŠ-
cial, or giving a large campaign contribution. There is a vast dif-
ference between those who are active and those who are not. Those
less active are groups with distinctive needs and preferences in re-
lation to government policy. They are the more disadvantaged
members of society: the poor, the less well educated, racial and
ethnic minorities. It is consequential that they are less active.
Thus, equal participation would enhance the likelihood of
equality in public policy. This is not to say that the expression of
political voice by a group necessarily results in government re-
sponse; or that government response to a group requires that the
group express its needs. There are many other parts to the political
11 Raymond E. WolŠnger and Steven J. Rosenstone, Who Votes? (New Haven: Yale
University Press, 1980).
process. People may speak up and get nowhere. Conversely, people

who do not speak up may have their views represented by proxy
participants who take up their cause, by nongovernmental organi-
zations, or by bureaucrats or legislators pursuing autonomous pol-
icy goals. Their needs and preferences can be sought out through
government or foundation research. All this is true. But, on bal-
ance, groups that put their own agenda on the table and press their
own case are likely to wind up better off in the policy-making and
policy-implementing arena than those who do not.
But that still leaves the democratic dilemma: the competence
of ordinary citizens. We want policies that are effective, that are
responsive to needs in a fair and reasonable manner, and that work
to preserve the democratic process. Suppose all speak up but few
are competent: will we get such policies? Suppose further that cit-
izens are self-centered, unwilling to pursue the common good,
and unwilling to tolerate those with whom they disagree. Can de-
mocracy survive? Yet, if we do not have something like equality in
political activity, how do we keep government policy from being
skewed to the beneŠt of a small, more advantaged portion of the
populace?
One obvious answer is representative government. It is a sys-
tem, as James Madison put it, that allows the views of the people
to be enlarged and reŠned by passing them through the consider-
ation of a few selected and wiser citizens. But if those selected and
wiser citizens—our representatives—hear only from some of the
people, they will be more likely to give consideration to those
from whom they hear. Views may be reŠned, but in the process the
views of signiŠcant groups may be reŠned away. Thus, citizen par-
ticipation remains important and citizen competence an issue.
A close look at the beliefs, values, and way of thinking of citi-
zens is in order to see if they can perform the tasks of citizenship.
How competent are they as reasoners about public matters and as
moral judges about such matters? How good are citizens as social
analysts? Do they have relevant, unbiased, and accurate informa-

tion or know how to get it? Can they use that information to make
valid causal inferences about the likely consequences of various ac-
tions? And do they have the moral commitments that would seem
to be relevant for an effectively functioning democracy in which
citizen voice is expressed? One would be a concern about public
matters and the public good rather than narrow self-interest. And
the other would be openness to the expression of views different
from their own.
Before systematic survey data were available, there were many
views as to the nature of the public. For some, the average citizen
was a wise and reasonable observer of politics, and public opinion
represented the sum of that wisdom. Among populist thinkers,
public opinion was far wiser and more moral than the opinion of
intellectual or economic or political elites. For others, the average
citizen was a pretty pathetic character when it came to reasoning
about public matters, and the aggregate of such opinions was even
worse. For the latter observers, political matters would be better
left to those more competent to deal with them—to the elites that
the more populistically inclined thinkers despised.12
The Incompetent Citizen
Perhaps the main indictment of the public—certainly the most

noticed in the social sciences—came in the important work of
Philip Converse on the nature of mass belief systems.13 He looked
for a structure to opinions and concluded that there was not much
there. A well-structured belief system would be one in which peo-
ple had a clear position on an abstract set of principles, and their
12 Of the many statements of this principle, Walter Lippmann’s is one of the clear-
est: Walter Lippman,, Public Opinion (New York: Macmillan, 1922).
13 Philip Converse, “The Nature of Belief Systems in Mass Publics,” in David
Apter, ed., Ideology and Discontent (New York: Free Press, 1964). pp. 206–61.
particular policy positions would derive from the more general

principles. The result would be a coherent and consistent set of
political positions—perhaps conservative, perhaps liberal.
Converse found that few people referred to broad ideological
positions when they discussed politics. Furthermore, there was
not much “constraint” in people’s attitudes on issues. By con-
straint he meant whether the attitudes of individuals were con-
nected to one another and whether speciŠc policy positions could
be deduced from the person’s more general principles. People did
not have that structure. One position did not predict another. Nor
did the answer to a question asked at one time predict the answer
to a question asked later; change the wording a bit and you change
the response. The general view was that Americans were nonideo-
logical. Their views were neither rich, nor Šrm, nor coherent.
In addition, there was evidence at that time that the public was
not morally competent; that they held views incompatible with a
cooperative, ongoing democratic system. Citizens were narrowly
focused on their own private lives and self-interest. And, as the
studies of Samuel Stouffer showed, they could not be relied upon
to support the basic freedoms of expression that underlie demo-
cratic rule in a pluralist society.
The result was disturbing to students of democracy. It sug-
gested that the average citizen was not up to the job. He or she was
neither cognitively competent to make the calculations needed for
a reasoned voice in the policy process nor morally competent to
seek goals consistent with a functioning democratic community.
If this is the case, it would imply that if we seek political equality
in order to get equality in the expression of citizen preferences and
needs, we may lose on the quality of citizenship.
Let’s see what more recent research tells us about both kinds of
competence—cognitive and moral. And in doing so we can see
how the reasoning of citizens resembles or differs from that of so-
cial scientists as well as normative philosophers.
The Somewhat Competent Citizen
Are citizens really as instrumentally incompetent as these early

data suggested? Though much of what earlier studies said about
the incompetence of citizens and the incoherence of their political
positions was and remains true, there are ways in which the judg-
ment may have been too harsh. For one thing, our research meth-
ods may overstate the absence of information among citizens or
the weakness of citizen reasoning. Here is a case where survey
methods and the simpliŠcation they create may do citizens a dis-
service by making them look more simple than they are. Surveys
ask citizens about the issues that the survey researcher cares
about—or thinks the public (or the people sponsoring the survey)
cares about. The agenda is set by the questioner who suddenly ap-
pears at the respondent’s doorstep or at the other end of the tele-
phone line. The citizen, with no prior warning or chance to study
is given a test on topics the poll-taker has chosen.
Furthermore, how competent a citizen appears in a survey may
depend not only on the questions the interviewer asks, but also on
the questions that the political process asks of citizens. When in-
terviewers began to ask citizens about their voting choice and how
they came to it, they found that citizens evaluated candidates on
the basis of personality or party, but rarely mentioned policy posi-
tions as a reason for preferring one over the other. One reason
might have been that citizens cared more about personality than
policy. Another might be that there was little difference in policy
between the candidates in the relatively unpolarized 1950s. When
my colleagues and I followed up these studies after the more polar-
ized election campaigns of Lyndon Johnson versus Barry Goldwa-
ter in 1964 and Richard Nixon versus George McGovern in 1972,
we found that respondents spoke more about issues when they an-
swered questions about these candidates, and that issues—the po-
sitions of the candidates and the positions of the voters—played a
bigger role in their voting decisions. In other words, by offering

the voters, as Goldwater put it, a choice not an echo—that is, a
candidate with a clear and well-articulated set of policy positions
away from the muddled median—he was posing a question to the
public that allowed for issue responses.14 (Incidentally, the elec-
tions of 1964 and 1972 show that if you offer the public a choice
rather than an echo, the public will vote for the echo. But that is
another story.)
Thus, it may be that these data overstate the limitations of or-
dinary citizens. They are not philosopher kings or even princes,
but they are not as benighted as the portrait might suggest. A
newer literature on citizen competence and citizen reasoning has
emerged, much of it based in cognitive social psychology. It is a
rich literature with many different schools that often seem, to me
at least, to be saying the same thing in different words while ar-
guing over who is right. This literature, while not rejecting the
substance of the Šndings about the limitation of ordinary citizens,
argues that they do better than one might expect. They have efŠ-
cient ways of using limited information, their views are not so-
phisticated but they do have a structure, and, in the aggregate,
they do better than their individual skills might suggest. The
thrust of this literature on citizen competence is that citizens and
the citizenry (individuals and the aggregate of individuals) do
pretty well with what limited resources they have.
Let’s begin with information. This newer literature focuses on
the issue of the way in which citizens can cope in situations of lim-
ited information. Though much classic analysis in economics as-
sumes perfect information for rational actors, much recent
theorizing and much empirical work recognizes that people oper-
ate with, to use Herbert Simon’s term, bounded rationality. We
cannot know it all—neither the facts nor the causal relationship
14 Norman H. Nie, Sidney Verba, and John R. Petrocik, The Changing American
Voter (Cambridge, Mass.: Harvard University Press, 1979).
among facts. As Anthony Downs famously pointed out, informa-

tion is costly. The ordinary citizen, faced with the complexity of
politics, would be irrational to gather too much—in part because
of the complexity of it all, more so because the individual citizen
can have only a tiny voice in what the government does.
How then do citizens with little information, limited time,
and limited inšuence on outcomes—but ultimate legitimacy—
navigate the complexities of politics? Let me outline some of
them. Much of the literature on citizen information has moved be-
yond asking what citizens know—which is painfully little—to a
focus on the issue of how citizens can make meaningful, rational
decisions under conditions of little information. They do so by us-
ing heuristics. These are shortcuts that allow someone to come to
a reasonable conclusion with limited information. There are many
types of heuristics; different authors use different categories. An-
thony Downs considered ideology and party to be the heuristics
most used in elections. You did not need to know more about an
issue than the position on it taken by your party or by those who
shared your general ideology in order to come to a decision “right”
for you.
There are other similar heuristics. If there are people you trust
who share your social position and your values, and if they are
more attentive to some issue than you choose to be, it is a pretty
reasonable approach to follow them. (We all do this in our day-to-
day lives; each of us has a friend who knows about restaurants and
another who can tell us what car to buy.) Or we follow some
trusted government ofŠcial. Members of the U.S. Congress follow
the lead of specialists, if they generally share the values of the spe-
cialist. Citizens can do that as well.
Heuristics have changed the ordinary citizen from the be-
nighted incompetent of a few decades ago to the rational user of
limited information of today. However, though a shortcut may be
an efŠcient way of getting somewhere, it is useful only if it takes
you where you want to go. The information acquired through the
use of a heuristic may not be valid. It may be irrelevant to the issue
at hand, wrong, or biased. Citizens who take information from
others often select those near at hand or those who share their
views. This is the road to biased information. There is, however,
evidence from an interesting study by Robert Huckfeldt that,
though individuals do use the criterion of shared views in choos-
ing information sources, they can tell who is informed and who is
not and they choose information also on that basis. They are fairly
good at choosing those who are in the know. They get by with the
help of their friends—and can choose friends who are informed.
This is not to convert ordinary citizens into social scientists.
They do place some quality control on information sources, but
they also look for that which is comfortable with what they al-
ready believe. (Social scientists sometimes do also.) Perhaps the
ordinary citizen differs from the systematic social scientist most in
the area where I suggest there might be some weakness in the in-
formation seeking by normative philosophers—in the area of sam-
pling and representativeness. Systematic social science depends on
representativeness (that’s why we like random samples and sur-
veys). The ordinary citizen samples information, not randomly
and perhaps not in terms of typicality, but in terms of the congeni-
ality of the information to what is already there, or the vividness of
an example, or the recency of acquisition of the information—
none of which criteria would satisfy the canons of proper research
design. The ordinary citizen always has an Aunt Edna—a person
well known to students of survey research. When someone is con-
fronted with statistics about health habits and longevity, espe-
cially when the statistics make the person uncomfortable, we hear
about Aunt Edna, who drank a quart of scotch and smoked two
packs every day but lived to be ninety-Šve. Ordinary citizens use
concrete examples—a particular news story, a particular personal
event—and analogize from that. Studies suggest that sample bias
in information acquisition represents the biggest challenge to ac-

curate inference on the part of the citizen.15 In general, ordinary
citizens misunderstand social reality because of statistical errors:
they do not know how to make valid inferences from evidence. For
instance, the notion of independent probabilities is unclear. One
of my favorite examples—indeed, a statistical heroine of mine—is
the woman who won the lottery in the state of New Jersey twice
(each time for over a million dollars). When interviewed by the
press she said she was not going to play the lottery anymore.
When asked why, she said that it would not be fair—she wanted
to give someone else a chance. Citizens also have logical problems
making causal connections. Their reasoning often involves such
statistical problems as omitted variable bias (thinking that A
causes B when something else, a variable not taken into account,
may cause both of them) or getting the direction of causality
wrong.
In sum, citizens are not statisticians or systematic social ana-
lysts. But they do have ways of getting by.
The Reasoning Citizen
What about citizen reasoning about political choices—whom to

support in an election; what policy to support; when and on what
to become politically active? Recent work on the way in which cit-
izens reason, closely related to the work on heuristics, takes what
seems to me to be a more realistic view than did earlier literature.
It turns the view of the public upside down. The earlier view was
that citizens had little in their heads; that was why they would
easily change from one position to another. The newer account of
why ordinary citizens’ positions on issues change over time or
change with changes in question wording is not that people have
15 Susan T. Fiske and Shelley E. Taylor, Social Cognition (New York: Random House,
1984), p. 253.
little in their minds, but that they have too much. John Zaller has
given the clearest expression of this position in relation to survey
responses. The reason responses seem to have a weak, almost ran-
dom character is that people have many values or principles in
their heads. Zaller calls them considerations—a broad and useful
term because it encompasses what we would call values or princi-
ples but also other considerations such as narrow self-interest, so-
cial pressures, and the like. When faced with a question about
public policy in a survey, people do not answer at random in order
to show that they have an answer (as many theories of the survey
response believe). They do answer “off the top of their heads”—
that is, without much rešection—but do so by sampling from the
set of considerations the question evokes. They see what kind of an
issue it is, they look at the considerations it relates to, they often
sample from the considerations they hold in their heads, and on
that basis they come to an answer.
This description of the process accords well with several facts
about the choice situation in relation to some issue of public pol-
icy. One is that almost all social issues involve a tradeoff among
values. Thus, an issue position can be affected by multiple consid-
erations. Most people have many such considerations in their
minds. Some of these are inconsistent with each other, but that in-
consistency is not apparent since citizens do not routinely critique
their values for logical coherence. Furthermore, when faced with
an issue where multiple and conšicting values may apply, individ-
uals can choose to frame the issue as involving one set of consider-
ations rather than another. In this way, they handle the inconsis-
tency without reconciling the various sides.
The work of Amos Tversky and Daniel Kahneman is most
prominent in this area. In a long series of experimental studies,
they and others have shown that the same choice situation leads to
a different decision depending on the way the choice is framed—
that is, on the context in which the choice is placed. Take the same
situation, change the symbolic description, and you change the re-
sponse.16 Often this is done by analogy: is intervention in Kosovo
like intervening to save the Jews under Hitler or intervening in
Vietnam? One of my favorite examples is an experimental prison-
ers’ dilemma study in which the likelihood of cooperative behav-
ior in exactly the same game was affected by whether the game was
named Community or Wall Street. Call it Community and players
are more likely to cooperate; call it Wall Street and they are more
selŠsh.17
The point here is that citizens have many standards in their
heads rather than one. Consider the self-regarding economic-type
calculations assumed in most economic modeling of human be-
havior. Such a standard may be applied in many situations, but
perhaps not in all. Experimental literature shows that the more
the experimental situation “approximates a competitive . . . mar-
ket” with anonymous buyers and sellers, the less other-regarding
behavior will be observed. As Samuel Bowles puts it, “ . . . market
like situations induce self-regarding behavior, not by making peo-
ple intrinsically selŠsh, but by invoking the self-regarding behav-
iors in their preference repertories.”18 What behavior individuals
choose depends on the context in which they are choosing.
The research on framing sheds light on what it means when
citizens appear to have changeable and uncertain views. Changes
in position from one question to another may be the result of sub-
tle changes in the frame that the question provides or change in
the context in which it has been evoked. Or it may be due to a
16 Amos Tversky and Daniel Kahneman, “The Framing of Decisions and the Psy-
chology of Choice,” in Roben Hogarth, ed., Question Framing and Response Consistency (San
Francisco: Jossey-Bass, 1982).
17 Lee Ross and Andrew Ward, “The Power of Situational Effects in the Prisoners’
Dilemma Game” (unpublished manuscript, 1993), cited in Cass Sunstein, “Social
Norms and Social Roles,” Columbia Law Review 96 (May 1996): 913n33.
18 Samuel Bowles, “Endogenous Preferences: The Cultural Consequences of Mar-
kets and Other Economic Institutions,” Journal of Economic Literature 36 (March, 1998):
75–111; quotation on p. 89.
change in the media that have primed people for one frame rather
than another. This changeability is not necessarily a manifestation
of the weak views of people but may be due to a change of the con-
text in which they see the issue.
This fact does not, however, necessarily create conŠdence in
the thoughtfulness and reasonableness of citizen views. Tversky
and Kahneman Šnd evidence for the existence of an availability
heuristic—that is, issues are framed by the set of considerations
that are nearest to the top of the individual’s mind.19 This may
be the most recent consideration to which they have been exposed.
(As an aside, one ought to note that this is not a monopoly of rank
and Šle citizens; each of us can certainly think of high political
leaders whose views depend on whom they spoke to most
recently.)
This is a major way in which media coverage affects how citi-
zens think about issues. Newspapers and TV have their major im-
pact by “priming” people to consider one consideration rather
than another more salient in relation to an issue—by giving it a
particular frame. They cue citizens to place an issue in one or an-
other category. If change of the symbol changes the position, or if
the media or elites can frame issues in one way or another and elicit
response from citizens, this suggests perhaps that citizens do not
act autonomously; that they are manipulated by symbol spin doc-
tors in the media and in public ofŠce.
Certainly there is evidence for such manipulation—and a large
number of people spend a good deal of time and money trying to
do just that. But one can put a somewhat more optimistic inter-
pretation on citizen sensitivity to the nature and origins of the
message and to context. Individuals may change their views de-
pending on the context within which an issue or decision arises.
But this is not mere reaction to irrelevant symbols. Context mat-
19 Amos Tversky and Daniel Kaneman, “Availability: A Heuristic for Judging Fre-
quency and Probability,” Cognitive Psychology 5 (1973): 207–32.
ters in terms of what values are applied. The games Wall Street
and Community differ. One ought to trust and cooperate in a com-
munity; on Wall Street less trust and more defection seems appro-
priate. We would not have much respect for the person who acted
like a Wall Street investor in all personal relationships. And the
investor who behaved like a communitarian on Wall Street would
likely soon go broke.
Furthermore, research shows that framing is not a manipula-
tion of passive individuals by sly framers. What frame an indi-
vidual accepts depends, at least in good part, on the values or
considerations the individual already has stored. There is a run-
ning interaction between that which is stored from earlier expe-
riences and the new input. There is evidence that individuals do
not accept just any frame. Frames are accepted if they are com-
patible with previously available considerations. And individuals
make judgments as to the credibility of a framer. Thus, the pro-
cess of framing is not merely a matter of passive citizens and
active elite manipulators. It may go too far to consider it a dem-
ocratic dialogue in which citizens interact with the media and
other elite communications, as well as with friends and others, to
consider and reconsider their positions on public matters in the
light of new arguments and changing contexts. But it may rep-
resent a quite reasonable decision process carried out by some-
what autonomous individuals. Certainly, there is evidence for
autonomous position-taking on the part of the mass public. Con-
sider the Bill Clinton/Monica Lewinsky affair. The public clearly
responded autonomously—that is, without cues from political or
media elites. When the story Šrst broke, all the pundits signaled
that Clinton was through since the public would not approve.
No one primed the public to make distinctions between private
and public life. Indeed, the only fully reasonable participant in
that madness seems to have been the American public.
This is consistent with several approaches to individual deci-

sion-making that stress on-line processing of political informa-
tion.20 Individuals develop overall evaluations of people or issues
or groups. New information modiŠes that stored impression—
creating an updated, running tally of favorable and unfavorable
judgments on President Clinton, on Kosovo, on the Republican
Party, or what have you. The point is that individuals will then
know where they are—that they favor or oppose some person or
policy—but they may not be sure how they got to that position. If
they are asked for reasons, they are likely to create them on the
spot. This presents a challenge to the observer—the journalist, or
politician, or social scientist—to try to Šgure out what was the
reasoning that went into the conclusion. But there is a structure of
reasoning behind it.
This analysis also sheds light on the difference between sets of
values held by philosophers and those held by ordinary citizens.
The value systems of philosophers—those people whose business
it is to propound such systems—are explicit, usually formal, logi-
cally consistent, and accompanied by explicit justiŠcations. Ordi-
nary citizens are under no obligation to explicate the logic of their
positions or to justify them. Popular moral positions are not ex-
plicit and not necessarily consistent. But inconsistency can remain
because there is no need to face it. The values come into play in re-
lation to particular moments—whether this be when the individ-
ual is a respondent answering a question, a voter in a voting booth,
or a citizen deciding whether to write a letter to a representative or
attend a protest meeting. In that coming together, there is no need
for the development of coherent justiŠcations or a reconciliation of
values. The only thing needed is the outcome—the vote, the pol-
icy position. If there is a reasoning pattern connected with it, it is
often left unarticulated.
The analysis illustrates another general difference between the
20 John M. Basili, ed., On-Line Cognition in Person Perception (Hillsdale, N.J.:
Lawrence Erlbaum, 1989).
reasoning process of philosophers and that of citizens. The com-

monsense reasoning of the latter appears to be more concrete,
more tied to speciŠc exemplars. Citizens will use as examples a
particular news story, a particular personal event, and analogize
from that. They recognize similarity of patterns. Philosophers also
use examples, but these tend to be abstractly constructed situa-
tions that illustrate a general point—how the willingness of
sports fans to pay extra to see Wilt Chamberlain play demon-
strates that free markets lead to just distributions or how the di-
lemma faced by a couple of criminals deciding whether to confess
or not demonstrates the difŠculty of collective action. Reasoning
by analogy is not as logical as reasoning by deduction—but it is
reasoning and can be reasonable reasoning at that.
In sum, citizens may appear inconsistent—the worst logical
sin—but it is not an unreasonable inconsistency. They are not phi-
losophers, but they get by. Citizens are also not social scientists;
they use shortcuts and research techniques that would not pass
muster in social science journals. And as social analysts, they also
manage to get by.
If ordinary citizens are not moral philosophers or social scien-
tists, what then are they? Perhaps the nearest analogy is that they
are legislators, legislating for themselves. They are policymakers
and must make decisions. The decisions are narrow; they can de-
cide policy only for themselves. They decide how to vote, they de-
cide whether and on what issues to become politically active, they
decide what policies to prefer, they decide how to answer a survey
question on some public issue.
Legislation involves both normative and empirical analysis.
One needs to know what it is that one wants, which involves hav-
ing values and goals and, one would hope, the ability to reason
about such values and goals. And one needs to be able to analyze
the means of achieving one’s goals, which involves knowledge of
relevant information and the ability to use that information to an-
alyze the consequences of various acts. Thus, the ideal citizen
should be both philosopher and social scientist. But given that no

one is paying for their time, given that they have other things to
do, citizens may take many shortcuts in their normative rešec-
tions and in their social science analyses. The result is not the
highest level of reasoning either about values or about facts (or
their combination). But it is not unreasonable.
Let’s turn to some examples of the public philosophy to see
how this works out.
The multiplicity of values and the absence of a need to recon-
cile them would lead to citizen support for fairly balanced views
on issues when there is a conšict among generally accepted values.
They may wind up in the middle, a position more consistent with
those normative theories that seek to balance or reconcile various
views than with those theories that take a clearer position on one
side or the other. Consider the following democratic dilemmas.
I began by saying that the terms “democratic” and “govern-
ment” contain a built-in tension—between democracy, which im-
plies power from the bottom up, and government, which implies
power from the top down. It parallels the tension between the leg-
islator as delegate and the legislator as trustee and many other
democratic dilemmas involving the need for a citizen voice and
the need for effective government that cannot respond to every cit-
izen demand. Citizens—if one reads the results of polls—agree
with both positions. They do not want leaders who do nothing but
follow the polls or leaders who ignore them. They want both lead-
ership and responsiveness.
Similarly, in relation to their own competence as citizens they
are relatively balanced. The average citizen is not a populist—
at least that is what survey studies show. Most members of the
American public seem to be quite modest about their own capa-
bilities. They believe that most citizens don’t know what is best
for them, that they don’t understand issues and arguments, and
that they are too uninformed to make sensible choices in public
matters. Nonetheless, more support the view that “every citizen
must have an equal right to decide what is best for the country”
than accept the proposition that those of character and intelli-
gence ought to have more voice.21 They are not sure they know
what’s best, but they do not think anyone else has the right to de-
cide that for them. I Šnd that a somewhat contradictory but actu-
ally quite reasonable position.
The Competent Citizen and the Competent Public
Individual citizens may be better citizens than one might expect.

However, democratic choices depend, not on the views of any in-
dividual citizen, but on the aggregate of citizens. As with the issue
of the competence of the individual citizen to function well within
a democratic system, there are conšicting views of the capacity of
the aggregate of citizens to do so. Various arguments can be and
have been educed as to the analytical and/or moral incompetence
of citizens in the aggregate:
1. The aggregate is immoral. The argument here is that peo-

ple in the aggregate may behave in ways that they would
Šnd unacceptable if they were acting as individuals—either
because their responsibility (to themselves or to others) for
acting can be submerged in a group or because group pres-
sures push them to act this way. Mobs, “mass publics,” can
be undemocratic or unjust in ways that individuals would
not be.
2. The aggregate can distort the positions of the individuals
who make it up. Pluralistic ignorance and spirals of si-
lence exist when individual members of a group mistake
the views of their fellow citizens but, because of weak
21 Herbert McClosky and John Zaller, The American Ethos: Public Attitudes toward
Capitalism and Democracy (Cambridge, Mass.: Harvard University Press, 1984), chap. 3;
quotation on p. 79.
communications or fear of rejection, act in ways that dis-

tort what would be the sum of their individual positions.
3. There may be no stable summary of the position of a group
due to Arrow problems of aggregation. It may be that a col-
lective view is an oxymoron.
There may be a good deal of validity to each of these reserva-

tions about the mass public. But let me focus on the opposite ar-
gument: that the public is superior to the individual. Benjamin
Page and Robert Shapiro in their book The Rational Public argue
that individual instability of opinion need not be incompatible
with a relatively stable set of reasonable preferences among the
public as a whole. “Even though individuals may hold only weak
and poorly informed opinions, subject to measurement error and
random change due to new information, there can still exist a sta-
ble, meaningful public opinion based on the underlying central
tendencies of individuals’ opinions. And sample surveys eliciting
the expressed opinions of many individuals at a given moment
may quite accurately reveal what collective public opinion looks
like.”22
How does this happen? As they put it, “Individuals exposed to
random bits of information may err or be misled about an issue
and may form policy preferences not well suited to their needs and
valuers; but the public as a whole, so long as the errors are ran-
domly distributed, will make use of all the available information
and choose the appropriate policies.”23 The well-known argument
comes from Condorcet’s jury theorem and was made 200 years
ago. The argument is that given modest amounts of information,
imperfect individual judgments are more likely to lead to correct
aggregate judgments the larger the group of judges. Thus, the
jury does better at coming to the truth than would any individual
22 Benjamin I. Page and Robert Y. Shapiro, The Rational Public: Fifty Years of Trends
in Americans’ Policy Preferences (Chicago: University of Chicago Press, 1992), pp. 25–26.
23 Ibid., p. 26.
member, and public opinion may be sounder than the opinion of

individuals.
But the condition for a Condorcet outcome is that the errors of
the several individuals are random. If the individual voters are ex-
posed to a systematic bias in what they hear in one direction or an-
other, the aggregate may not come to the right decision more
frequently. It is one of those areas where there is much reasoning,
but little empirical work. Larry Bartels has attempted to study
this empirically using election studies. His results are not conclu-
sive. “How likely is it that the effects of voter ignorance would
persist even in the aggregated choice of a mass electorate? The
simple answer is that no one knows.”24
One of the more interesting approaches to the relationship
between the attitudes of individuals and the aggregate of those
attitudes is found in the work of James Stimson. Using some inno-
vative ways of summing up the positions of individuals across a
range of issues, he traces the movement across time of the average
position of the American public in broad policy orientation—
essentially its average position on a left to right dimension of gov-
ernment policy that might be thought of as running from the
unfettered free market to the full-šedged welfare state. Though
many individuals have unclear and inconsistent positions on these
issues, the movement of the position of the citizenry has an inter-
esting regularity. In general the public moves away from ex-
tremes. When government policy is off to the left, the public
begins to move right; as policy moves right consistent with public
opinion, the public moves back to the left. Thus, the public fa-
vored a diminution of social services before the Ronald Reagan ad-
ministration, but when Reagan started to implement those
policies it moved back. And this pattern has been repeated—more
recently in relation to the Republican Party’s movement to the
right and its alliance with the religious right in the 1994 election.
24 Larry M. Bartels, “Uninformed Voted: Information Effects in Presidential Elec-
tions,” American Journal of Political Science 40 (1996): 200.
The public wants change, but when it sees change going too
far, it backs off. Stimson likens the effect to that of a thermostat.
When things get too cool, it raises the temperature until it gets
too hot, and then it lowers it. Such moderation can look like in-
consistency, and to some extent it is. And it can be maddening, I
assume, to politicians who think they have won the support of the
public for a new direction to policy only to Šnd the public turning
against them as soon as they try to move in the new direction. But
the position seems not unreasonable. In this sense, the civic aggre-
gate may indeed be better than the civic individual.
The Private and the Public Citizen
Let us consider the other set of reasons why one might oppose
equal citizen voice: that citizens are morally incompetent and hold
values incompatible with an ongoing democratic polity. It is hard
to delineate what values are needed for democratic functioning.
Let me focus on two that seem basic and have the advantage of
having been studied empirically. These are having some consider-
ation for the public good and having tolerance for opposing and
unpopular views.
Let us begin with the issue of the consideration of the common
good. The ordinary citizen is commonly believed to be a narrow,
somewhat selŠsh, parochial person, concerned only with his or her
narrow world and seeing the broader society as a projection of his or
her inner needs. If that is the case, who will think of the common
good? I realize that this is an issue only for those who believe that
citizens can and ought to think beyond their narrow self-interest.
From a more economistic point of view, narrow self-interest is what
we would expect and, probably, what we would want. I cannot get
into that debate, so let me state that I think it good if people think
about the common good. In any case, let’s see if they do. And for
this I can return to my own research on citizen activism.
The interest in this subject for myself and my collaborators ac-

tually came from a different direction: from our concern with the
paradox of participation. This is a puzzle derived from the conclu-
sion of rational actor theory that it is irrational for the individual
citizen to participate in politics unless there is some selective ben-
eŠt. To participate for some collective outcome (something in the
public good) makes no sense from this perspective since the indi-
vidual, through, say, a vote, cannot have much effect on a policy
outcome and can beneŠt from it even if inactive. Thus, the indi-
vidual ought to take a free ride. The argument is well known, and
I won’t elaborate it here. And the paradox is that, despite the
seemingly impeccable logic, millions in fact participate.
Our attempt to understand the paradox led us to a better un-
derstanding of the ordinary citizen as private and as public person.
We did something rarely done: we asked individuals why they
were active. More speciŠcally, we asked them what gratiŠcations
they obtained from their activity.25 It is, of course, tricky to ask
the kind of question I am referring to: to ask people to tell you
why they did something or what they got out of some action. Cit-
izens may give what they assume to be the socially correct answer.
One way we obtained what I think are compelling answers to the
question of why people participate is by asking them to name the
issue that animated their activity. Why did they take the time and
effort to do what they did? For some activities such as voting the
reasons for acting may be many and somewhat uncertain. But we
were asking about speciŠc acts: thus the motivation may be easier
to recognize and the purpose ought to be clearer to the actor. Most
of us know what it is we want when we write a letter to a represen-
tative or when we take part in a protest or go to a community
meeting.
We found their answers compelling. Sometimes, they describe
25 The data from these questions and the full analysis of the material discussed here
can be found in Sidney Verba, Kay L. Schlozman, and Henry E. Brady, Voice and Equality
(Cambridge, Mass.: Harvard University Press, 1995), chap. 4.
the issues on which they are active as involving narrow selective

beneŠts, which would seem to Št the expectations of rational actor
theory. But narrow issues were relatively rare. The issues or prob-
lems they say animated their activity are ones they describe as af-
fecting many people, the community, or the nation, not just
themselves. Indeed, they are usually issues easily labeled collec-
tive. And when we ask for descriptions of these issues, they Št that
categorization. We believe that these are sincere descriptions of
their motivation.
In general, we found that some of the reasons why people are ac-
tive Št the narrow, material, self-interested, selective beneŠt cate-
gories that social choice theory might lead us to expect: to get a job,
a contract for one’s Šrm, or a particular government beneŠt. Some
activity is animated by such concerns, but not much. More is ani-
mated by other concerns. Some are active for what can be called so-
cial reasons: they enjoy working with others, or they want to please
a friend or someone else with whom they want to be in good repute.
More still get gratiŠcation from doing what they consider to be
right—they are active because it is their civic duty, or because, as
they say, they think of themselves as the kind of person who cares
about social matters. From a rational actor perspective, such moti-
vations can be—and have been—considered self-interested gratiŠ-
cations. But that makes self-interest a rather broad and not very
useful rubric. And, even if we subsume social and psychic gratiŠca-
tions under the rubric of self-interested beneŠts, it does seem to
make a difference whether someone who is active is so for a material
beneŠt or to satisfy a feeling of good citizenship. The latter feels a
lot more like civic virtue. Lastly, many of the answers as to why
people participate explicitly invoke a public purpose: to make the
community or nation better, to inšuence public policy. And such
public purposes clearly do not Št rational actor calculations.
In sum, the people we studied act more often as civically
minded citizens than as narrowly selŠsh individuals. Aristotle
may not be right that humans are by nature political animals, but
the economists are not right either that they are rational calcula-
tors of narrow, individual self-interest. In sum, if by civic virtue
we mean “the disposition to further public over private good in
deliberation,”26 citizens appear to manifest it—indeed, they pro-
claim their commitment to it.
This is all consistent with another body of literature that shows
that citizens evaluate policies and candidates as public issues, not
in terms of the speciŠc effects on them. In research on attitudes to-
ward unemployment policy, Kay L. Schlozman and I found that
individuals (in deciding what policies to support or what candi-
dates to choose) were guided more by the national situation in
terms of unemployment than by their own experience.27 Similar
Šndings exist about attitudes toward racial matters, schooling,
American involvement in war, medical care, and other issues.
General commitments (self-identiŠcation as liberal or conserva-
tive, party afŠliation, general attitudes on matters of race, etc.) are
more likely than the impact of the issues on a person’s narrower
self-interest to predict the speciŠc policy preferences of an indi-
vidual or a voting decision.28
A comment on rational actor theory and the understanding of
ordinary citizens: Surely, if there is a candidate for a theory of citi-
zen involvement in public life that would universally explain citi-
zen behavior, it is rational actor theory. It is, in one form or another,
appearing as an explanation of behavior in many Šelds outside of
economics, from whence it came, and certainly in political science.
26 Quoted in Richard Dagger, Civic Virtues: Rights, Citizenship, and Republican Liber-
alism (New York: Oxford University Press, 1997), p. 14
27 Kay L. Schlozman and Sidney Verba, Injury to Insult (Cambridge, Mass.: Harvard
University Press, 1979).
28 See, for example, Richard R. Lau, Thad A. Brown, and David O. Sears, “Self-In-
terests and Civilians’ Attitudes towards the War in Vietnam,” Public Opinion Quarterly 42
(1978): 464–83; David O. Sears, C. P. Hensles, and L. K. Speer, “Whites’ Opposition to
Busing: Self-Interest or Symbolic Racism,” American Political Science Review 73 (1979):
369–84; and David O. Sears, Richard R. Lau, Tom R. Tyler, and Harris M. Allen, Jr.,
“Self-Interest vs. Symbolic Politics in Policy Attitudes and Presidential Voting,” Ameri-
can Political Science Review 74 (1980): 670–84.
As I indicated, the concern that my colleagues and I had for the

motivations behind citizen activity derives from the difŠculty that
rational actor theory has with citizen participation: the fact that
many people who ought to be free riders are in fact active. Our
analysis of the motives that individuals have for such activity sug-
gests that, if we take their statements to be true—which I argue we
can more or less—they show that citizens are active for many rea-
sons. Some but not many of the reasons Št a rational actor, narrow,
self-interested cost-beneŠt calculation model. Some might Št if
the scope of beneŠts is extended to include the psychic beneŠts of
being a good citizen, though if that is done the theory loses dis-
criminatory power. And some motivations—such as the desire to
inšuence public policy, which is quite common—do not Št the
theory at all.
But earlier in my lectures, I mentioned our analysis of the
strategies used by those people who act as political recruiters—the
people who make telephone calls or in other ways contact citizens
to get them to give money or time to a campaign or a political
cause. I said that they act as rational recruiters. We Šnd we can
model their strategies quite well using cost-beneŠt calculations.
We show, for instance, that they deploy information in a rational
manner to Šnd those people who are most likely to say yes to a re-
quest and, having said yes, have the capacity to make contribu-
tions to the campaign or the cause. They quite rationally follow
the advice of Willie Sutton, the famous American bank robber,
who, when asked why he robbed banks, replied: “That’s where the
money is.” In addition, they use personal connections in a ration-
ally calculating manner.
How do these parts go together—our inability to model the
activity of citizens in general in rational actor terms and our abil-
ity to model the activity of political recruiters when they seek ac-
tivists? The answer sheds light on the circumstances under which
rational actor theory works and is useful for understanding citizen
political behavior. We could not model the reasons why individu-

als become recruiters—why they decide to spend their time on the
telephone calling others to support some political cause. They
have many and varied motives. Some do so out of narrow self-in-
terest—they are hired to do that work. Others do it for social rea-
sons—perhaps their friends are working in the same campaign.
Others do it out of commitment to the cause. As with political ac-
tivity in general, there are many reasons for becoming a recruiter.
But once people have decided to be active as recruiters, they
pursue their recruitment activities in a carefully calculated and ef-
Šcient manner. They look for people who have the money to con-
tribute, have past records of contribution, are known to be
supporters of the cause—all characteristics that make them useful
targets of requests.
The reason, we believe, that cost-beneŠt calculations work in
the recruitment case is that there is a well-deŠned task (to mobi-
lize support) and such calculations can be made. It is similar to a
market decision where one wants to get the most for the least cost.
In general, we found that rational actor approaches to citizen
activity Št some contexts and not others. They are sometimes
quite useful, but not always.29 We have made that point at several
conferences in the United States, and the reaction has been inter-
esting. The conclusion has been attacked. The grenades have been
thrown from both directions—and, in a way, for the same reason.
Rational actor fundamentalists have objected to the idea that the
theory is sometimes useful; they would prefer to think it is always
useful. The rational actor rejectionists—of whom there are many,
though they are not as well organized—have also objected to the
notion that the theory is sometimes useful. They would prefer to
think it is rarely or never useful. So be it.
29 For a fuller discussion of the uses of rational actor theory for the study of politi-
cal participation, see Sidney Verba, Kay L. Schlozman, and Henry E. Brady, “Rational
Action and Political Activity” (unpublished manuscript).
Free Speech and Tolerance
There is another basic value that underlies democracy and that we

might hope would be shared by ordinary citizens: tolerance of the
views of others. Here is an issue that has been studied a good deal
by social scientists and that is closely related to normative issues.
The issue is that of stability in a society with multiple, incompat-
ible doctrines. John Rawls assumes that in any pluralistic society,
there will inevitably be a number of comprehensive doctrines that
will conšict with each other. How then can they live side by side?
His assumption as to the nature of the doctrines held by a demo-
cratic population and his description of the solution Št the clear
and comprehensive principles often found in political philosophy:
differing comprehensive and internally consistent conceptions of
the good exist side by side. This would cause instability in a soci-
ety committed to peaceful and noncoercive means of political
decision-making—that is, in a democratic society. Stability is
maintained and the strain of multiple incompatible doctrines is
moderated by a free-standing overlapping consensus on a demo-
cratic process that involves tolerating alternative doctrines. The
conception of a fair democratic procedure is free standing in that it
is not derived from any of the competing conceptions of the good.
This is an area in which there is a fairly long tradition of empir-
ical research on what citizens actually believe about the variety of
doctrines to be found in a democratic society (not cited by Rawls),
and about whether they are committed to some overarching con-
ception of a democratic process that would lead them to tolerate
doctrines incompatible with the ones in which they believe.30
What does the empirical literature tell us about the issue? The
30 See, among others, Samuel A. Stouffer, Communism, Conformity, and Civil Liber-
ties: A Cross-section of the Nation Speaks Its Mind (Garden City, N.Y.: Doubleday, 1955);
John L. Sullivan, James Piereson, and George E. Marcus, Political Tolerance and American
Democracy (Chicago: University of Chicago Press, 1982); Paul M. Sniderman et al.,
“Principled Tolerance and the American Mass Public,” British Journal of Political Science
19 (1989): 19–44. For a discussion of the relation of Rawls’s work to empirical material
on free speech, see George Klosko, “Rawls’s Political Philosophy and American Democ-
racy,” American Political Science Review 87 (1993): 348–60.
matter is a subject of debate; there are various studies done with

somewhat differing assumptions and different methods, so I can-
not give a deŠnitive reply. But the studies are illuminating about
the issue.
To begin with, it appears that political philosophers such as
Rawls—as one might expect—take doctrines of the good more se-
riously than do ordinary citizens. It may be that comprehensive
competing doctrines of the good exist, but most citizens do not
hold such a doctrine. Competing doctrines may not be that hard
edged, so there can be overlapping conceptions of the good. Such
conceptions need not be mutually exclusive among individuals or
groups of individuals. If people have different conceptions of what
is right in relation to the family and the economy and religion and
the polity, they may disagree with their fellows on some things
but agree on others. This overlapping applies to social groups as
well. African Americans in the United States are quite at odds
with many of the religious right when it comes to matters of eco-
nomic policy—but not that far away when it comes to such issues
as abortion and homosexuality. Such overlapping consensus—per-
haps better called by an older name in political science, cross-cut-
ting cleavages—can hold a system together.
What about evidence of a free-standing commitment to toler-
ance of competing conceptions of the good? Much of the earlier lit-
erature suggested that this was not something on which to count
as a support for democracy. The studies of Samuel Stouffer and oth-
ers found that most citizens did subscribe to what seemed to be a
free-standing commitment to the general principles expressed in
the Bill of Rights. The hitch was that these citizens did not seem
to deduce from those principles a more speciŠc commitment to
tolerate unpopular groups. Almost all supported a general right to
free speech, but many opposed allowing someone to speak in their
community if that person espoused communism or atheism.31
There are several possible interpretations of this phenomenon:
31 Stouffer, Communism, Conformity, and Civil Liberties.

citizens may merely be giving lip service to support for the Bill of
Rights (a positive symbol), or may be inconsistent in their posi-
tions (an illustration of inadequate logic), or may be placing one
value (free speech) against another (the desire to protect democ-
racy or religion from attack). But whatever the interpretation,
there did not seem to be much comfort for those who would base
democratic stability on a free-standing commitment to a more
open society.
The work of John Sullivan and his associates modiŠed Stouf-
fer’s design and carried the research forward. He and his associates
argued that Stouffer’s examples of groups to which people might
deny free speech rights were mostly groups that would be anath-
ema to conservatives—socialists, Communists, atheists. If one ex-
pands, and lets the individual choose his or her enemy (that is,
respondents are asked to name groups they really dislike and then
asked whether they would accord them free speech opportunities),
then intolerance is found to be even more endemic. Similar re-
search in Britain comes to a similar conclusion. What Sullivan’s
research seems to show is that people report commitment to some
free-standing principles of free speech and tolerance and may in-
deed have such commitments—but when it comes right down to
it, their opposition to people espousing doctrines they dislike
swamps their commitment. Again this can be seen not as intoler-
ance, but rather as a greater concern about social stability. The
least-liked groups in the United States include the Ku Klux Klan
and the American Nazi Party; in Britain, the National Front and
Sinn Fein.
The studies did suggest several mechanisms by which this gen-
eral expression of principle coupled with speciŠc rejection of free
speech did not create instability or a loss of such freedoms. The
mechanisms include the fact that those citizens least committed to
freedom of speech for speciŠc groups were also the least educated
and the least active citizens. Their views may be more threatening
to democratic freedoms, but they are not likely to act on them. The
argument was that intolerance among the mass public did not
harm democracy—except perhaps in times of mass political
arousal—because it was irrelevant. The apathy and ignorance of
the intolerant mass neutralized its potentially deleterious effects.
Elites—more active and more tolerant than the mass—saved de-
mocracy. Stouffer found community leaders to be much more sup-
portive of democratic free speech than was the average citizen. And
the impressive massive survey work of Miller, Timpson, and Less-
noff in Britain conŠrms this.32 Similarly, John L. Sullivan et al.
found legislators in the U.S.A., the U.K., and several other coun-
tries to be much more supportive of civil liberties for unpopular
groups.33
One more body of research is relevant to this topic. Paul Sni-
derman and his associates, using some sophisticated analytical
techniques, found what strikes me as good evidence for a free-
standing commitment to tolerance. People who are tolerant of one
group are tolerant of others regardless of the closeness of the group
to the individual’s own preferred position. Thus, tolerance is free
standing and does not depend (certainly not fully) on one’s views
of the substance of the doctrine one is tolerating or not. Put an-
other way, among citizens who are generally tolerant, they are
more supportive of the free speech of groups they oppose than in-
tolerant citizens are tolerant of the free speech of groups of which
they approve. A generally tolerant racial liberal is more tolerant of
free speech for a racial bigot than an intolerant racial bigot is for
other racial bigots.
Sniderman and his associates call such general tolerance across
groups and doctrines principled tolerance. But it is interesting
32 Miller, Timpson, and Lessnoff, Political Culture in Contemporary Britain.
33 John L. Sullivan et al., “Why Politicians Are More Tolerant: Selective Recruit-
ment and Socialization among Elites in Britain, Israel, New Zealand, and the United
States,” British Journal of Political Science 55 (November 1993): 51.
what they mean by this. It is not that these principled tolerators

have explained that they hold to such a position because it derives
from some general view of society or a general theory of democ-
racy. They have given no such justiŠcation, and Sniderman et al.
suggest they would not be able to do so if asked. Rather their
analysis suggests that principled tolerance means general toler-
ance; tolerance that transcends one’s support for or opposition to
the views tolerated. It is, thus, a good example of a Rawlsian free-
standing commitment to a political principle of justice.
What still seems to be missing is a principled commitment ar-
rived at by reasoning from the nature of a pluralistic democratic
system. Rather than coming to their general tolerant position by
reasoning from a more abstract set of principles, Sniderman and
associates present evidence (based on how one instance of support
for free speech relates to others) that these tolerant citizens reason
by analogy. They focus on some vivid example of tolerance—on an
example of free speech for some group—and then support toler-
ance for other groups because they see the situation as similar. It is
an example of framing. A particular situation—tolerance for some
group—is categorized as an example of free speech. If tolerance of
another group is seen to Št the category, the individual favors free
speech again. Principled consistency is not a matter of a logically
connected structure of argument, but of a connected set of recog-
nizable objects.
But this still leaves out reasoning. Is there really none? The
new standard view of citizens derived from the cognitive social
psychological literature is that citizens reason their way to conclu-
sions through a commonsense, results-oriented process. They rec-
ognize speciŠc patterns in speciŠc cases; they put together a
variety of considerations and come to a conclusion. It is a form of
logic and reasoning. It differs, however, from the reasoning of phi-
losophers in that citizens do not tell you how they got there. They
may reason, but do not give you their reasons.
The difference is not trivial. Reasons and their public state-

ment are crucial if democratic decision-making is to involve de-
liberation and is to be more than a clash of preferences or wills or
interests. We are supposed to be able to reason together—to argue
and persuade. It is too bad if citizens come to decisions that are
reasonable, that are consistent with their set of stored considera-
tions, but they cannot tell you why.
But again, maybe we do not hear citizens reason because they
have not had the chance. We ask speciŠc questions and get their
responses, but do not ask them why they said what they said.
However, experimental studies where one modiŠes the wording of
questions and the presentation of issues do seem to show us a rea-
soning process. We change the frame and the respondent changes
the response. This, as I have suggested, is not necessarily weakness
of belief, but what we expect from democratic discourse—desic-
cated though the examples may be. In a sense, changing the frame
of the questions is saying to someone, as we say in discussion: you
are saying yes, but have you thought about the following consid-
eration? Oh yes, you respond, now that you mention that, I think
the opposite.
But in these experiments, the individuals do not explicate.
Maybe they can’t, but maybe they can if we ask them, and maybe
they can if we ask them and they have some time to think about it.
They are faced with important issues that they have not con-
fronted, and need some time to rešect.
What if we give people time to think and ask them to explain
themselves? There is a tradition of this kind of research, the tradi-
tion of depth interviews. They make it difŠcult for us to generalize
because the numbers are too small. But if they are selected more or
less at random—not interviews of our professional colleagues or
our students—they give us a good hint at the reasoning processes
that would be found in a larger sample of the population. This is
the tradition of the work of Robert Lane and Jennifer Hochschild,
of the discussion groups of William Gamson and Roberta Segal.34

Let me illustrate what can be found by looking behind some of the
answers to questions about civil liberties.
Herbert McCloskey conducted pilot interviews in order to
design the forced choice questions for his large study of attitudes
toward free speech. Dennis Chong unearthed this trove of conver-
sational material on free speech, in which people were asked about
their views and were allowed to reason about the issues.35 Issues of
free speech almost always involve conšicting considerations. Free
speech is desirable for lots of reasons, but the problem arises, of
course, when it is speech for people who will say bad things:
things that offend others, that hurt others (sometimes children),
that lead to social unrest. Ambivalence about free speech ought to
be expected from most people who are rešective. And this is what
is found in the discussions with interviewers. People change their
minds. They say they favor free speech, but they are afraid that
people will be convinced by Nazis or by Communists. As one
woman says, she believes in free speech but opposes letting groups
of this sort open a bookstore. As she puts it, “. . . that’s trampling
on their rights a bit, but maybe to protect other people, I would
say no [to allowing them to open a bookstore].” Another moves
the other way, opposing free speech rights for various radical
groups. But when the interviewer asks in general about people
“who stand up and advocate the forcible overthrow of our govern-
ment,” she replies, “. . . that would be in a sense what the Nazis
and communists were trying to do. They do have the right, they
34 Robert E. Lane, Political Ideology: Why the American Common Man Believes What He
Does (New York: Free Press of Glencoe, 1962); Jennifer L. Hochschild, Facing Up to the
American Dream: Race, Class, and the Soul of the Nation, Princeton Studies in American
Politics (Princeton: Princeton University Press, 1995); William A. Gamson, Talking
Politics (New York: Cambridge University Press, 1992); and Roberta S. Sigel, Ambition
and Accommodation: How Women View Gender Relations (Chicago: University of Chicago
Press, 1996).
35 See Dennis Chong, “How People Think, Reason, and Feel about Rights and Lib-
erties,” American Journal of Political Science 37 (1993): 867–99.
do have the right. I know they do. . . . That’s what freedom is all
about. Freedom of speech. Yeah. I’m all just doing a topsy turvy
here. I know they have the right and our government isn’t perfect
and they might, you know, bring out a lot of things that would
change for the better, but I would not want to go to the extreme of
turning all of a sudden communist. . . . ”
One further example illustrates how people are likely to come
out if they rešect on such matters. James Kuklinski carried out an
interesting experiment.36 He asked about free speech rights for
members of the Ku Klux Klan and tried to get answers that were
more or less rešective. Some respondents were told to answer with
whatever they Šrst thought; others to rešect on the consequences
of the position they espoused. He found a not insigniŠcant differ-
ence. Interestingly, it was not that the more rešective respondents
were more favorable to free speech opportunities for the KKK;
they were less so. More rešection did not raise the importance of
the free speech perspective in the balance of values, but apparently
led them to think of the social harm coming from speech whose
content they found hateful. Rešection did not lead to more toler-
ance, but to less. There is a point to be made here. I am not sure
whether these results would always hold up. But they contradict
the assumption we sometimes make that people, if they only re-
šected, would become enlightened and more committed to free
speech. They may indeed come to a different position on rešec-
tion—but maybe it will not be the one we imagine.
In sum, the reasoning found in these interviews is not elegant;
the individual has not thought much about it and is deciding on
the spot. But one cannot expect most people to have spent much
time mulling over such issues, and the considerations are reason-
able.
36James Kuklinski et al., “The Cognitive and Affective Bases of Political Tolerance
Judgments,” American Journal of Political Science 35 (1991): 1–27.
The Educated Citizen and Participation
One last point about the relationship between equal participation

and quality participation: Though I have suggested that the ordi-
nary citizen is not as benighted as one might imagine, this does
not mean that all citizens are equally competent. Some are more so
than others, the biggest predictor of such competence being edu-
cation. It is no wonder that almost all students of citizenship have
hoped for a more educated citizenry, and some, of course, have ar-
gued for special participatory rights for the better educated—
whether that be achieved through extra votes for university
graduates or through literacy tests. In fact, one need not have pol-
icies that favor the educated as participants, because they partici-
pate more than others anyway.
This fact illustrates the dilemma of unequal participation.
Let’s explore this a bit further. The best predictor of political activ-
ity is education. In our research on participation, my colleagues
and I tried to answer the question: why is education so potent a
predictor of political activity? Was it that it made people more in-
formed? Made them feel more efŠcacious? Made it easier for them
to see connections between their values and preferences and gov-
ernmental action? Was it that it inculcated the values of citizen-
ship—that one ought to be a participant? Did it increase the store
of resources that people had—skills that made one an effective
participant or income useful in making political contributions?
Did it put the individuals into networks so that they were sur-
rounded by others who were active and by others who could help
them act effectively? The answer was—to use the forced choice
language of tests and surveys—all of the above. Education fosters
activity though its effect on information, skills, values, resources,
networks, and more. No wonder it is so potent. Furthermore, the
potency grows after education ends.
The educated citizen is not only more active; he or she is a bet-
ter citizen. The educated citizen is more informed, has a more co-
herent or consistent set of political values, and can make better

connections between means and ends. Educated citizens have
other virtues: they are in general more supportive of the rules of
democracy, more tolerant of unpopular voices, more committed to
cummunal rather than individualistic goals. Some simple data
from our studies will make this clear. Compare individuals with
no college education with college graduates. The latter are more
active in any of the activities we measure, and by a long shot. They
are, for instance, four times as likely to have contacted a govern-
ment ofŠcial (a good way to express political voice) than is some-
one with only a high school education. They get twice as many
information items right. When they contact, they are four times
as likely to deal with an issue that has broad relevance, rather than
some issue limited to themselves and their family; they have, thus,
more civic concern. And in other ways they are exemplary. They
are Šve times as likely to support the right of someone who op-
poses religion to speak in the community. And on and on. They
are better citizens.
I don’t want to overstate this case. There is the possibility that
we give educated citizens more beneŠt of the doubt when it comes
to civic virtue than they deserve. And this may be because they
speak our language. Some early research on the civic-mindedness
of citizens found that better-educated citizens were more likely to
have a long time horizon and to think of the public good when
contemplating social and economic issues. Less well educated citi-
zens and newer immigrants had a shorter time horizon and
thought of policies from the point of view of their immediate im-
pact on themselves.37 I think there is some truth to this, but many
years of looking at what respondents say about public issues have
also made it clear that those who are more articulate have little dif-
Šculty in expressing their own self-interest in broad social terms.
In our research on unemployment we found that less well educated
37 James Q. Wilson and Edward C. BanŠeld, “Public-Regardingness as a Value
Premise in Voting Behavior,” American Political Science Review 58 (1964): 876–87.
people need a job and ask for a job. More educated respondents
need a job and discuss the problem of unemployment—
while also asking for a job. Neither I nor my colleagues who have
sent in research proposals to the National Science Foundation ever
argued that the research funding would allow us to do what we like
better to do (research) than what we might otherwise have to do
(teach) or that the funding would increase the prestige of our re-
search institute, etc. But it may be (just may be) that some of us
have such things in mind—as well as having a sincere commit-
ment to the substance of the research and its value to scholarship
and the understanding of society.
It is to the good that the educated are more active. That educa-
tion fosters activity means it fosters better democratic participa-
tion. But the educated do not differ from the rest of the citizenry
only in their greater competence and commitment to democracy.
They are also wealthier, more likely to be male, more likely to
come from the dominant race and ethnic groups. They are less
likely to support spending on programs to aid the poor. More im-
portant, perhaps, they are less likely to face the deprivations faced
by those with less education. On a large number of measures of
need (the need to put off medical treatment, the need for better
housing, etc.) they differ substantially from those who did not
attend college. In sum, educated activists are more civically com-
petent, which makes for a more enlightened input into the demo-
cratic policy process; they are wealthier and more advantaged,
which means they have policy preferences and needs different
from those of the population as a whole, and that makes for a more
biased input into the policy process. Insofar as the participation of
the educated is driven by the democratic values of tolerance or by
the civic beliefs they acquire in school, this creates a better citi-
zenry. Insofar as their activity is driven by the components of their
social class position (their income, the networks of connections
that come with various jobs), this creates the more biased polity.
We may want an educated citizenry, but we wind up with a

wealthy one.38
Conclusion
Where does this all take us in relation to citizens as social scien-

tists and moral philosophers? Individual citizens fall far short of
the reasoning of either group. But they have ways of making not
unreasonable choices about what to support and what to do. In
sum, the individual citizen is not a philosopher; he or she could
not articulate a coherent social philosophy, nor give full justiŠca-
tions for the positions held or the political choices made. But the
individual citizen is—as the political scientist V. O. Key made
clear—not a fool.
Is the ordinary citizen a rational actor, calculating from a self-
interested perspective the costs and beneŠts of his or her own ac-
tions and the actions of others? The answer may depend on how
narrowly or broadly one deŠnes self-interest and what one consid-
ers to be costs and to be beneŠts. But using any deŠnition not so
broad as to make the question vacuous, the answer would seem to
me to be: sometimes. It depends on the venue. Citizens do think
about the collectivity when it comes to public issues. They think
about their narrow interests as well. And they balance the two. It
38 My colleagues and I, in our studies of political activity, tried to disentangle the

effects of education and income on civic activity and, in turn, on the consequences of that
activity for the kinds of messages sent to the government. We linked three components
of a model: what social characteristic is connected to the activity (allowing us to distin-
guish the effect of education from the effect of income); what was the nature of the polit-
ical activity (distinguishing activities that involve giving time from those that involve
giving money); and what were the preferences or values that were given greater voice in
the participatory process. We found that time-based activity derived directly from edu-
cation while money-based activity came from income and only indirectly from educa-
tion through its effect on income. In turn, the income-driven, money-based activity led
to a greater voice for the afšuent and policies that favored the wealthy; the education-
driven, time-based activity led to a greater voice for the informed and for those who were
more tolerant.
could not be otherwise. Rabbi Hillel wrote: “If I am not for my-
self, who will be for me? If I am only for myself, what am I?” The
average person agrees.
Is the ordinary citizen the free and equal person that Rawls as-
sumes, with a capacity for a conception of justice and of the good?
Again the answer may depend on exactly what such conceptions
entail. But if one looks at what we know about individual citi-
zens—and looks at both the representative but sketchy informa-
tion we get from surveys and the more rounded picture we get
from depth studies—the answer would seem to be pretty much
yes.
Does this mean that representative government with its ability
to reŠne the preferences of the people can or should be replaced by
more direct citizen control—through referendums, or polls, or
other direct democratic means? I hope no one takes that to be the
implication of my remarks. There is a long distance between the
subtle complexities of policy-making and the preferences of ordi-
nary citizens as revealed by surveys and others means. The public
itself, so it would seem, does not favor that much unŠltered citi-
zen voice. But I do suggest that the collectivity of citizens can add
useful input to the political process, input that will make the gov-
ernment more democratically responsive. I am not sure that the
input ought to be labeled wisdom, but it is not foolishness. Wis-
dom may, in any case, be in somewhat short supply throughout
the political process—among ordinary citizens, and among our
leaders as well. But that is another topic.

Representative Democracy and Democratic Citizens: Philosophical and Empirical Understandings

Uploaded by

Copyright:

Available Formats

Representative Democracy and Democratic Citizens: Philosophical and Empirical Understandings

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Representative Democracy and Democratic Citizens: Philosophical and Empirical Understandings

Uploaded by

Copyright:

Available Formats

Representative Democracy and Democratic

Citizens: Philosophical and Empirical

The Tanner Lectures on Human Values

Brasenose College, Oxford

important than in America. He also noted that it was “difŠcult to

of an individual is hard enough. Aggregating them, especially if

Let us distinguish broadly between two kinds of methods: system-

population to whom we wish to generalize. If the selection is bi-

philosophers, journalists, and others who put their thoughts on

The Problem with Unsystematic Observation

The main fault of such unsystematic observation is that it is sub-

supports.”8 Barry says the task of understanding the public is a fu-

Survey Research and the Public

The main technique for getting at public opinion is the sample

of values and preferences is the vote. No one has found a substitute

Furthermore, ordinary surveys can be supplemented with

Having said something about the methods for ascertaining the

Democracy depends on such voice. It also depends upon equal

what citizens think—the content of their values or their policy

Democratic Citizens: What Are They Like?

consideration of their needs and preferences. I’d like to summarize

Are Citizens Equal in Politics?

To begin with, as is well known—and as our studies conŠrm with

• “They can’t” suggests a paucity of necessary resources—time

or no difference, little or no knowledge about the political

These three components—resources, engagement, and recruit-

cognitive skills (the ability to understand issues and see connec-

Do We Want Political Equality?

decisions made by the government (what contribution does citizen

process. People may speak up and get nowhere. Conversely, people

analysts? Do they have relevant, unbiased, and accurate informa-

The Incompetent Citizen

Perhaps the main indictment of the public—certainly the most

particular policy positions would derive from the more general

The Somewhat Competent Citizen

Are citizens really as instrumentally incompetent as these early

bigger role in their voting decisions. In other words, by offering

among facts. As Anthony Downs famously pointed out, informa-

in information acquisition represents the biggest challenge to ac-

The Reasoning Citizen

What about citizen reasoning about political choices—whom to

This is consistent with several approaches to individual deci-

reasoning process of philosophers and that of citizens. The com-

should be both philosopher and social scientist. But given that no

The Competent Citizen and the Competent Public

Individual citizens may be better citizens than one might expect.

1. The aggregate is immoral. The argument here is that peo-

communications or fear of rejection, act in ways that dis-

There may be a good deal of validity to each of these reserva-

member, and public opinion may be sounder than the opinion of

The Private and the Public Citizen

The interest in this subject for myself and my collaborators ac-

the issues on which they are active as involving narrow selective

As I indicated, the concern that my colleagues and I had for the

political behavior. We could not model the reasons why individu-

Free Speech and Tolerance

There is another basic value that underlies democracy and that we

matter is a subject of debate; there are various studies done with