Buchanan 2004

ALLEN BUCHANAN Political Liberalism and
Social Epistemology
I. A Different Kind of Political Liberalism
The Moral and Prudential Risks of Social Epistemic Dependency

I grew up in the American South during the 1950s and 1960s in a racist
family culture embedded in a society of institutionalized racism. Blacks
were relegated to separate and inferior schools, were effectively excluded
from voting, and could not use the same restrooms, hotels, or restau-
rants as whites. I was taught, by explicit dogma and by example, to regard
blacks as subhuman. Unlike my mother, I never witnessed a lynching,
but I did once see a desiccated, severed black ear of unknown prove-
nance, proudly displayed by a white junior high school classmate. I also
recall joking with my friends about the “Tucker telephone,” a crank-
operated dynamo that was used to deliver electrical shocks to the
genitals of black inmates of a nearby penal farm.
Largely through luck, I left this toxic social environment at the age of
eighteen and came to understand that the racist world view that had
been inculcated in me was built on a web of false beliefs about natural
differences between blacks and whites. My first reaction was a bitter
sense of betrayal: Those I had trusted and looked up to—my parents,
aunts and uncles, pastor, teachers, and local government officials—had
been sources of dangerous error, not truth.
Anger at being betrayed soon gave way to gratefulness for my good
fortune. I realized that I was young enough to have a chance to lead a
decent life and that my racist beliefs had not (yet) issued in any serious
injuries to black people. But to this day I tremble at the thought of the
I am grateful for excellent comments on earlier versions of this article provided by Sahar
Akhtar, Andrew Altman, Kristen Hessler, Cindy Holder, Avery Kolers, Alex Rosenberg, Kit
(Christopher) Wellman, and the Editors of Philosophy & Public Affairs.
©  by Blackwell Publishing, Inc. Philosophy & Public Affairs , no. 
96 Philosophy & Public Affairs
moral risk my upbringing imposed on me. Given the right circum-

stances, I might have perpetrated a hate crime.
Socially inculcated false beliefs can not only put one at moral risk,
they can also endanger one’s well-being—they can put one at what I shall
call prudential risk. False beliefs about an international Jewish conspir-
acy, about the inherent superiority and imperial destiny of the German
nation and the infallibility of the Führer helped motivate Germans to
support policies that resulted in their own deaths by the millions, and in
the destruction and division of their country.
The moral and prudential risks of socially inculcated false beliefs are
exacerbated by the systematic nature of the cognitive distortion. A
person brought up in a racist society typically not only absorbs an inter-
woven set of false beliefs about the natural characteristics of blacks (or
Jews, and so on), but also learns epistemic vices that make it hard for
him to come to see the falsity of these beliefs. For example, when a child,
who has been taught that blacks are intellectually inferior, encounters
an obviously highly intelligent black person, he may be told that the
latter “must have some white blood.” Along with substantive false
beliefs, the racist (like the anti-Semite and the sexist) learns strategies
for overcoming cognitive dissonance and for retaining those false beliefs
in the face of disconfirming evidence.
By acting on the belief that certain groups are inferior, those in the
thrall of distorted belief systems create an environment in which their
experience seems to confirm those very beliefs. For example, in a society
in which whites relegate to blacks the menial jobs and justify this in part
on the grounds that blacks have no mechanical abilities, blacks will have
few opportunities to develop competence with machinery, or to exhibit
it if they should somehow manage to develop it. Similarly, where teach-
ers believe that girls have little aptitude for the subject, they tend to have
lower expectations for girls’ performance in mathematics, for example,
with the result that girls may show less mathematical competence.1 Thus
1. Christine Reyna, “Lazy, Dumb, or Industrious: When Stereotypes Convey

Attribution Information in the Classroom,” Educational Psychology Review 12 (2000):
85–110; MaryAnn Baenninger and Nora Newcombe, “Environmental Input to the Develop-
ment of Sex-Related Differences in Spatial and Mathematical Ability,” Learning and Indi-
vidual Differences 7 (1995): 363–79; S.J. Spencer, C.M. Steele, and D.M. Quinn, “Stereotype
Threat and Women’s Math Performance,” Journal of Experimental Social Psychology 35
(1999): 4–28.
97 Political Liberalism and
Social Epistemology
the risk is not simply that one will come to have false beliefs, but that the
same social processes that instill false beliefs will make it difficult to
correct them.
To say that social institutions can put us at prudential and moral risk
by inculcating false beliefs and the epistemic vices and distorted expe-
riences that sustain them, still underestimates the problem. A person
growing up in a racist or sexist or anti-Semitic or fascist society suffers
affective, as well as cognitive, disabilities. One learns distorted emotional
responses, in part by modeling them on the responses of parents, other
authority figures and peers, and in part because one’s affective disposi-
tions are (mis-)informed by false beliefs. If one believes blacks or Jews
are unclean, inferior beings, one will find physical contact with them
repugnant. Instead of being reliable moral guides, one’s moral emotions
become both symptoms and sustainers of false beliefs.
Those who are cognitively and emotionally disabled by a socially
inculcated system of false beliefs are typically not sociopaths. They have
the usual moral powers, but their operation is systematically disabled.
The problem is not that they lack the stable dispositions to feel, to judge
and to act that are called moral virtues, but rather that they possess them
in a truncated and distorted form.2 For example, a Nazi may believe that
every human being’s rights are to be respected and may be capable of
being moved by sympathy for the suffering of someone he identifies as
a human being. Yet if he believes that Jews are dangerous subhumans,
he may also believe that they ought to be exterminated; and when he
sees them killed he may feel no sympathy.
As my description of my own predicament indicates, the social
processes that produce and sustain the sorts of false beliefs that disable
the virtues prominently feature what I have elsewhere called epistemic
deference, the tendency to regard certain persons as reliable sources of
beliefs.3 For example, in the Third Reich, deference, not only to govern-
ment officials, but also especially to scientists and to medical doctors,
as well as to teachers, played an important role in inculcating, dissemi-
nating, and sustaining racist beliefs and the social practices that
expressed them and helped fashion the experiences that seemed to
2. Allen Buchanan, “Social Moral Epistemology,” Social Philosophy and Policy 19 (2002):
126–52.
3. Ibid.
confirm them.4 Later I examine in some detail how the way in which
authorities are socially identified, and hence how epistemic deference is
allocated, can either exacerbate or diminish the moral and prudential
risks of socially inculcated false beliefs.
Of course, it is not only those who grow up in grossly repressive soci-
eties who are vulnerable to socially inculcated false beliefs. All societies
inculcate false beliefs (as well as true ones). Who we are depends in large
part upon what we believe, and what we believe depends, more than we
like to admit, upon our social environment. Among the two best con-
firmed sociological generalizations are that in all societies individuals’
beliefs are strongly influenced by their social surroundings, and that
most of us seriously underestimate the extent to which this is so.5
Not just in childhood, but throughout our lives, believing, and hence
knowing, are largely a collective enterprise. This is the central message
of social epistemology, which has been defined as the comparative eval-
uation of how well social institutions facilitate the formation, preserva-
tion, and transmission of true beliefs.6 The formation (or production) of
true beliefs can take place in either of two ways: by the creation of new
beliefs or the correction of false beliefs. Even when it does not achieve
the maximization of true beliefs, expertise can be extremely valuable in
avoiding or correcting false beliefs.
The dark side of this profound epistemic dependence is our vulnera-
bility to socially inculcated false beliefs that put us at grave prudential
and moral risk. The proper response to this vulnerability, contrary to
Descartes, is not to attempt to free ourselves of the epistemic influence
of society. Even if we could do so, the price would be much too high,
because we can know so little on our own. Instead, we should attempt
4. Robert Jay Lifton, The Nazi Doctors: Medical Killing and the Psychology of Genocide
(New York: Basic Books, 1986). Of course, this is not to say that epistemic deference is the
only cause of dangerous socially inculcated false beliefs, nor that an adequate causal expla-
nation of the prevalence of such beliefs would start with epistemic deference as the
primary cause. For example, many people may gain comfort from conformity of their
beliefs with those of their countrymen, especially in times of radical uncertainty, and epis-
temic deference can serve to facilitate such conformity.
5. See, for example, Jacquie D. Vorauer and Dale T. Miller, “Failure to Recognize the
Effect of Implicit Social Influence on the Presentation of Self,” Journal of Personality and
Social Psychology 73 (1997): 281–95.
6. Alvin Goldman, Knowledge in a Social World (New York: Oxford University Press,
1999). Note that by epistemology I mean what Goldman calls ‘veritistic epistemology’. In
other words, I am assuming that there are truths and hence that there can be true beliefs.
Social Epistemology
to determine which social arrangements are least prone to inculcate,

spread, and above all sustain the most damaging false beliefs.
Although all carry the risk of false belief, some social institutions and
practices have greater resources than others for reducing this risk, in part
through mechanisms for the correction of false belief. My aim in this
article is to argue for certain key liberal institutions on the grounds that
they contribute significantly to the reduction of the moral and pruden-
tial risks to which we are all vulnerable by virtue of our ineliminable
social epistemic dependency, while at the same time allowing us to reap
the benefits of a sophisticated and flexible social division of epistemic
labor. I will articulate a social epistemological argument for some of the
most basic liberal institutions.
In the first instance I mean by ‘key liberal institutions’ (1) effective
institutional arrangements for freedom of thought, conscience, expres-
sion, and association, and democracy understood as the institutional-
ized opportunity for all to participate as equals, either directly or through
processes of representation, in the creation of the more important rules
of public order. But I also wish to emphasize the epistemic virtues of
liberal institutions more broadly construed so as to include another
feature: (2) a comparatively large role for merit in the social identifica-
tion of reliable sources of belief (epistemic authorities), where ‘merit’
means the possession of objective qualifications rationally related to the
functions of particular social roles and positions. Finally, I also show how
the epistemic virtues of these two types of liberal institutions depend in
part on the presence of (3) a broad culture of basic moral egalitarianism
that encourages a measure of what might be called epistemic egalitari-
anism, the tendency of ordinary people to think well enough of them-
selves to be willing to challenge socially identified authorities on
occasion, and to think well enough of their fellow citizens to be disposed
to listen to them when they criticize socially recognized authorities and
accepted practices.7
7. Meritocratic criteria are often qualified by giving weight to other considerations. For
example, seniority often counts in one’s favor in matters of promotion. My aim is not to
argue for strictly meritocratic criteria in all or any particular contexts. Instead, the social
epistemological argument I develop serves to make it clearer what the full costs—in par-
ticular the epistemic costs—of departures from meritocratic criteria are. More specifically,
departures from meritocratic criteria should be understood not only as possibly depriving
the best qualified candidate for a position of something that her qualifications entitle her
to, but as an instance of a practice that may be costly for all of us, from the standpoint of
reducing epistemic risk.
I will also suggest—although it is not part of my central argument—

that the same features of liberal societies that are appealing from the
standpoint of reducing the moral and prudential risks of socially incul-
cated false beliefs also make more likely the development of a discipline
of social epistemology. If this is the case, then the key liberal institutions
are doubly attractive: They include valuable mechanisms for controlling
the moral and prudential risks to which our social epistemic depen-
dence makes us vulnerable, while at the same time helping to generate
intellectual resources for identifying, refining, and improving those
mechanisms.
Social Epistemic Dependency, Liberal Institutions, and

Political Liberalism
I will argue that those who take seriously the moral and prudential risks
of social epistemic dependence ought to support liberal institutions, if
they already exist, or welcome the proximity of liberal institutions to
their own society, if they cannot create them there or do not wish to. A
chief virtue of this argument is that it achieves the goal of what Rawls
calls political liberalism: to provide a sound argument for key liberal
institutions without relying on any comprehensive moral conception,
secular or religious. The social epistemology argument for key liberal
institutions meets this criterion. It relies on the assumption that one
cares about reducing prudential and moral risk, not upon any particu-
lar specification of what welfare (or happiness) is or any particular con-
ception of morality. Moreover, the social epistemology argument makes
a case for the key liberal institutions without relying, as Rawls’s political
liberalism does, on a fundamental commitment to tolerance that some
have argued is as controversial as the comprehensive moral or religious
views the political liberal seeks to avoid.8
8. See, for example, Thomas Christiano, “Is There Any Basis for Rawls’ Duty of Civility?”
Modern Schoolman 78 (2001): 151–61. My reasoning in the current article, however, does not
assume that to be a sound justification, a defense of liberal institutions must meet con-
straints of political liberalism. My goal is only to show that the social epistemology argu-
ment provides a strong case for key liberal institutions and that this justification does
qualify as a political liberal justification. I happen to believe that sound justifications for
liberal institutions can be given that do not meet the constraints of political liberalism, but
I need not defend that position here.
Social Epistemology
Another feature of my argument is worth remarking. It does not base

the case for liberal institutions on the role of competitive markets in
producing true beliefs or correcting false ones. The social epistemology
argument I offer is compatible with, and might even be seen as strength-
ening that sort of argument for liberal institutions. However, it is non-
committal on the question of whether there could be a society in which
competitive markets play no significant role and yet in which the key
liberal institutions of freedom of expression, thought, conscience, and
association exist. Whether or not these liberal institutions could func-
tion effectively in the absence of an extensive role for competitive
markets, or at least a substantial system of private property rights that
serves as a limitation on potentially oppressive political power, is an
interesting question, but one on which I need not take a position here.
My concern is with the epistemic virtues of liberal institutions, not
with their preconditions. Nevertheless, in the penultimate section of the
article I consider the very real possibility that market institutions, unless
suitably constrained, can reduce the effectiveness of the key liberal insti-
tutions in mitigating epistemic risks. Under such conditions, I argue, the
social epistemological argument I provide in this article can serve as a
premise in arguments for intervening in markets.
Justifying Liberalism Without Over-Selling Autonomy

I have noted that the social epistemology argument for liberalism I
advance in this article does not rely upon any comprehensive moral con-
ception. In particular, it provides an argument for key liberal institutions
that does not assume what might be called autonomist perfectionism,
the position that the life of autonomy is best for all human beings. The
social epistemology argument for liberal institutions does not imply that
nonautonomous lives are inferior, from the standpoint of well-being or
of morality. This argument, unlike those based on autonomist perfec-
tionism, does not denigrate “traditional societies,” in which the ideal of
the autonomous individual is not prevalent. However, the social episte-
mology argument for liberalism does imply that living a nonautonomous
life is very risky, both morally and prudentially, unless, perhaps, that life
is embedded in, or at least appropriately connected to, a society that
includes key liberal institutions, institutions that facilitate the formation
and dissemination of true beliefs, and above all the correction of false
beliefs, through a comparatively reliable allocation of epistemic

deference.
My main concern is to argue that societies that include the key liberal
institutions are better, other things being equal, at reducing the moral
and prudential risks of socially inculcated false beliefs than societies that
do not include these institutions. I will use the term nonliberal societies
to cover any society that lacks the key liberal institutions.9 However, it
should be noted that a simple contrast between liberal and nonliberal
societies is somewhat misleading. The key liberal institutions can be
more or less developed. To that extent, being a liberal society is not an
all or nothing affair; it is a matter of degree. In addition, as I will elabo-
rate later, the societies that include the key liberal institutions may
sometimes include other features that reduce or counteract their epis-
temic benefits. My point is that societies that have more fully developed
key liberal institutions are, other things being equal, better at reducing
the moral and prudential risks of socially inculcated false beliefs.
My chief concern is with the comparison between liberal and nonlib-
eral modern societies. By ‘modern societies’ I mean those that exhibit the
robust division of labor characteristic of industrial and post-industrial
societies. Some of the arguments I will advance imply that liberal soci-
eties are superior, from the standpoint of reducing epistemic risks, to
nonliberal societies generally, whether they are modern or pre-modern
nonliberal societies. Others—those that assume a fairly extensive divi-
sion of labor—apply only to the comparative evaluation of liberal versus
nonliberal modern societies. This is not much of a limitation, however,
because it is fairly uncontroversial to assume that the chief task for the
justification of liberal institutions at this time is to make a case for them
in modern societies.
Reducing, Not Eliminating, the Moral and Prudential Risks of Social

Epistemic Dependency
Assuming that we are profoundly and unavoidably dependent for true
beliefs upon social institutions broadly defined, and assuming that no
human institutions could be constructed that would create, preserve,
9. I choose the term ‘nonliberal’ rather than ‘illiberal’ because the latter carries a neg-
ative connotation and to that extent is prejudicial in an article that makes a case in favor
of liberal institutions.
Social Epistemology
and transmit only true beliefs, it follows that the proper goal is not to
eliminate the moral and prudential risks of social epistemic dependency
but to achieve an optimal reduction of them. An optimal reduction of
the moral and prudential risks of social epistemic dependency, other
things being equal, is one that strikes a proper balance between reduc-
ing the risks of the production, dissemination and preservation of false
beliefs, on the one hand, and sustaining the features of social institu-
tions that facilitate the maximal production, dissemination and preser-
vation of true beliefs, on the other. Preventing the inculcation of false
beliefs by measures that would undercut the enormously valuable func-
tion of social institutions in producing true beliefs would not be an
acceptable way to reduce the risk of error.
Among the most important social institutions for the production of
true beliefs are (1) the social division of labor and (2) the social identifi-
cation of experts, that is, epistemic authorities, individuals or groups to
whom others defer as reliable sources of true beliefs. Although (1) and
(2) usually go together, they need not. A temporally limited social divi-
sion of labor would not produce the role of expert. For example, you and
I might divide between us a simple research task, making each of us
responsible for gaining certain information. This differs from a stable
social division of labor, in which persons occupy distinct roles for a suf-
ficient period of time to develop special expertise.
Even when the domain of expertise is a practical activity (such as auto
mechanics), rather than a theoretical pursuit (such as quantum mechan-
ics), experts are, inter alia, epistemic authorities, insofar as others defer
to their judgment and regard them as reliable sources of true beliefs.
They are objects of epistemic deference. In that sense, the division of
labor typically includes a division of epistemic labor.
The division of epistemic labor includes of a set of social processes
and norms that encourage individuals and groups to specialize in the
production of true beliefs about certain subjects. The social identifica-
tion of epistemic authorities picks out some individuals as being experts,
that is, relatively reliable sources of true belief, in those domains.
Social processes and norms that divide labor and identify experts can
achieve great epistemic gains, through the greater productivity that the
division of labor allows, by enabling individuals to develop special skills
for acquiring and transmitting truths, and by reducing the costs to indi-
viduals of gaining access to truths. But as I and all too many others have
learned to our sorrow, epistemic deference can be risky. Sometimes the

experts are sources of error, even in their domains of expertise, and
sometimes we err by relying on them for truths about matters that are
not in fact within their domains of expertise.10
At least two factors encourage experts to make unwarranted claims to
epistemic authority. First, to the extent that their being identified as
experts enables them to garner control and reap social rewards, they
have an interest in convincing us that the domain of their expertise is
broader rather than narrower and that the reliability of their judgments
is greater than it is. Second, especially when the training or education
needed to gain expertise is extensive, individuals who have become
experts tend to be committed to and identify with their expertise, to
value it so highly as to be prone to exaggerate its usefulness. Hence the
old saw (so to speak) that for one who possesses a hammer, everything
is a nail. Similarly, for surgeons a remarkable number of medical prob-
lems turn out to be treatable by surgery, and for teachers a surprisingly
large proportion of social problems require an educational solution.11
The very specialization that grounds the claim to expertise makes it
difficult for nonexperts to ascertain the limits of another’s expertise. Yet
clearly—although we often only discover it after the damage is done—
we can be mistaken in deferring to others in the process of forming our
beliefs. I will use ‘unwarranted epistemic deference’ to cover both (a)
excessive trust in the reliability of judgments of an expert that actually
are within the expert’s domain of expertise and (b) misplaced trust, as
when one mistakenly believes an expert to have special knowledge about
an area that in fact is not within her domain of expertise.
Clearly, what is desirable is an optimum, not a maximum, of epistemic
deference. Too little epistemic deference is inefficient because it means
individuals and groups will have to bear the costs of rediscovering truths
that are already possessed by experts and of having false beliefs that they
would not have had if they had deferred to experts. Excessive epistemic
deference is also inefficient, and can often be prudentially or morally
dangerous as well.
The goal of social epistemology is usually said to be the identification
of a set of institutions or, more likely, a range of sets of institutions, that
10. Buchanan, “Social Moral Epistemology.”
11. See Allen Buchanan, “Toward a Theory of the Ethics of Bureaucratic Organizations,”
Business Ethics Quarterly 6 (1996): 419–40.
Social Epistemology
are both feasible and more reliable than any feasible alternatives for
creating, preserving, and transmitting true beliefs. This definition is
close to the mark, if it is understood to encompass the striking of an
optimal balance between maximizing the creation, preservation, and
transmission of true beliefs and reducing the risks of false beliefs. The
reason for this qualification should be obvious: It might be the case that
a set of institutions scores very high on the production, preservation, and
dissemination of true beliefs overall, but nonetheless tends to inculcate
an inordinately high number of false beliefs. Similarly, an institution
might score high on avoiding false beliefs, but only because it produces
few beliefs, and hence few true beliefs. To allow for these possibilities,
the goal of social epistemology should be defined so as to acknowledge
that the strategies for maximizing true beliefs and for maximally reduc-
ing the risks of false beliefs may conflict. In such cases, what is to be
sought is an optimum, that is, the best trade-off between these two
values.12
To achieve a fully adequate conception of social epistemology, one
further consideration must be taken into account: Some beliefs are more
important than others, theoretically or practically. Thus an institution
might produce a greater quantity of true beliefs, or even a higher ratio
of true beliefs to false beliefs, and still be inferior, if a very large propor-
tion of the true beliefs it produced were of little consequence, or if it
failed to produce some especially important true beliefs, or if it produced
and sustained enough theoretically or practically damaging false
beliefs.13 From the standpoint of practically relevant true beliefs, institu-
tions are preferable, other things being equal, if they do a comparatively
better job of producing, preserving, and transmitting those true beliefs
that are most important for acting prudently and morally.
Reflection on my earlier examples of racism, anti-Semitism, and
sexism suggest that among the more important practical beliefs are
those that concern fundamental moral status. These are of two types:
those that ascribe certain natural differences to different individuals or
12. My use of the economic language of trade-offs and optimization is, of course, an
idealization. Although these terms may suggest more precision than is ever likely to be
obtained, they are valuable to the extent that they make clear the importance of balanc-
ing a number of different considerations in overall assessments of the epistemic virtues of
institutional arrangements.
13. This point is due to Avery Kolers.
groups, and those that ascribe different moral statuses to individuals or

groups depending upon which natural characteristics they are thought
to have. Together these two types of beliefs determine the most funda-
mental moral status we accord to individuals. As I have already noted,
beliefs about fundamental moral status are of great practical importance
because they shape the way we apply moral principles and they influ-
ence the operation of our moral powers.
Without attempting anything so ambitious as a full-blown social epis-
temology, I will advance the following hypothesis: From the standpoint
of social epistemology, institutions are preferable to the extent that they:
(1) allow and encourage the free exchange of information and ideas; (2)
create and sustain an efficient epistemic division of labor, thus reaping
the benefits of reliance on experts; but at the same time (3) constrain
epistemic reliance on experts both by (i) merit-based competition for
expert status and by (ii) appropriate limits on epistemic deference
imposed by a broad-based critical attitude of epistemic egalitarianism.14
In addition, other things being equal, (4) institutions are preferable to
the extent that they do a better job of producing, preserving, and trans-
mitting the most practically important true beliefs. In the next section I
clarify these desiderata and argue that social arrangements that include
key liberal institutions and certain psychological dispositions—or in
more traditional terms, virtues—that are prevalent in liberal societies
are, generally speaking, more likely to realize them. But before laying out
that argument, I want to make more explicit what I have assumed thus
far about the connection between false belief, on the one hand, and
prudential and moral risk, on the other.
Truth, Well-Being, and Morality

The examples of racist, anti-Semitic, and fascist belief systems sketched
above vividly illustrate the moral and prudential risks of false belief. The
enterprise of social epistemology tacitly assumes that the production,
preservation, and transmission of true beliefs are desirable. It takes this
goal as a given and then compares institutions as to their effectiveness
in achieving it. In its more sophisticated forms, social epistemology
seeks to determine not only the comparative epistemic effectiveness of
14. Buchanan, “Social Moral Epistemology.”

Social Epistemology
alternative institutional arrangements, but their epistemic efficiency as

well. In other words, what matters is not just the ability to produce, pre-
serve, and transmit truths, but also to do so at least cost, where cost is
broadly construed. In any case, the assumption that drives the enterprise
is that having true beliefs is a good thing.
I shall make no attempt here to provide a general defense of this latter
assumption. Instead, I will only indicate, very briefly, why, generally
speaking, false beliefs—when adequate mechanisms for their correction
are absent—tend to put our welfare and our ability to act morally in
jeopardy.
First, and most obviously, both prudential action and attempts to
comply with moral principles require reliance on beliefs about how the
world is, at least to the extent that they both typically include some
means–ends reasoning as well as the ability to identify situations to
which principles of action are applicable. If prudential and moral action
requires true factual beliefs, and if the reliability of our access to true
factual beliefs depends upon the epistemic reliability of the institutions
within which we operate, then we should be concerned about whether
our institutions are epistemically reliable.
If some beliefs, including those that determine our judgments about
moral status, are especially important, then we should be particularly
concerned about how well alternative institutional arrangements
function so far as the production, preservation, and transmission of
those beliefs is concerned. Beliefs that influence our judgments con-
cerning fundamental moral status are particularly crucial for the proper
application of moral principles. To the extent that one cares about
acting morally, one ought to be concerned about the truth of the
beliefs that shape one’s judgments about fundamental moral status and
hence upon the reliability of the social processes that influence those
beliefs.
Second, if it is also the case that having access to true beliefs facili-
tates the evaluation of putative principles of prudence and/or putative
principles of morality, not merely their proper application, we have
another reason to try to ensure that our institutions are epistemically
reliable. In other words, if we strive to act prudently or morally, we
should be concerned not only about the truth of the factual beliefs on
the basis of which we apply principles of prudence or morality or select
appropriate means to achieve what those principles tell us to strive for,
but also about the validity of the principles themselves. Having true
beliefs can be relevant to this assessment.
Even if it turns out to be the case that the principles of morality, or of
morality and prudence, are not themselves capable of being true (or
false), it may still make sense to say that such principles can nonethe-
less be rationally evaluated, and that having true beliefs is necessary for
that evaluation. For example, if principles of prudence are those princi-
ples such that acting on them tends to maximize one’s own well-being,
then determining which principles really are principles of prudence, or
which are the best principles of prudence, will require reliable access to
true beliefs about what is in fact conducive to our well-being. If, among
whatever other characteristics they may have, moral principles facilitate
cooperation on mutually advantageous terms, or reduce mutually
destructive conflict, or are those principles that could not be reasonably
rejected by persons motivated to seek cooperation on terms that meet
that same condition, then true beliefs (about what arrangements are
mutually advantageous or reduce mutually destructive conflict, or about
what cannot be reasonably rejected) are necessary for knowing which
principles are moral principles and hence for knowing how to act
morally, so far as this requires acting on moral principles.
What sorts of true beliefs are needed to evaluate principles of pru-
dence or morality is an issue in metaethics, the depths of which I shall
not attempt to plumb on this occasion. For my defense of key liberal
institutions it is only necessary to note that: (1) true beliefs about the
world are needed to help ensure that we act appropriately on prudential
or moral principles, quite apart from whether having true beliefs and
avoiding false ones bears on the evaluation of the principles themselves;
and (2) true beliefs about moral status, or at least the avoidance or cor-
rection of false beliefs about moral status, are especially valuable. I will
proceed, then, on the assumption that those who desire to act prudently
or morally ought to be concerned about the epistemic reliability of the
social context in which their beliefs are formed, and then argue that
liberal institutions score better on epistemic reliability than nonliberal
ones.
The comparative nature of my argument is worth underscoring. I do
not argue that the key liberal institutions reduce the moral and pruden-
tial risks of socially inculcated false beliefs to zero, only that they are
likely to do a better job of risk reduction than nonliberal ones.
Social Epistemology
II. Liberal Institutions and Epistemic Reliability
Freedom of Thought, Conscience, Expression, and Association

It is not difficult to see that the effective institutionalization of these key
liberal freedoms is generally a necessary condition for the reliable for-
mation and transmission of true beliefs and, at least as importantly, for
the detection and correction of false ones. John Stuart Mill famously
emphasized the conduciveness of these liberal freedoms to prudential
success, arguing that they increase the individual’s chances of deter-
mining what will in fact maximize his or her well-being.
However, Mill’s case for these key liberal freedoms is less than optimal
for three reasons. Pointing them out will make clearer what is distinctive
about the views advanced in this article. First, Mill sometimes writes as
if he were basing his case for freedom of thought, conscience, expression,
and association on an autonomist perfectionist view, as when he says
that traditional, nonautonomous lives are “cramped and dwarfed.”15 This
strongly suggests that the key liberal institutions are needed because it is
only through an individual’s active participation in them that he can
realize his human potential. In my judgment, such an autonomist per-
fectionist view is mistaken or, at the very least, has not been adequately
defended by Mill or anyone else. But more to the point for present pur-
poses, it is simply not necessary to assume an autonomist perfectionist
view in order to make the case that we all ought to support key liberal
institutions. Whether or not they are necessary if autonomy is to flour-
ish, they are valuable from the standpoint of epistemic reliability.
Second, Mill unnecessarily assumes a controversial radical pluralist
conception of the good for individuals, even going so far as to compare
the multiplicity of good lives to the variation in foot sizes.16 This assump-
tion is both problematic in its own right and unnecessary for the social
epistemological argument for liberalism. One need not assert that the
good is that various in order to appreciate the need for freedom of
thought, conscience, speech, and association for the production, trans-
mission, and preservation of true beliefs and the correction of false ones.
Even if there is considerably less diversity in the good for individuals than
15. John Stuart Mill, “On Liberty,” in On Liberty and Other Essays, ed. John Gray (New
York: Oxford University Press, 1991), Part III.
16. Ibid.
Mill thinks, the free exchange of ideas and opinions is extremely valuable
for the formation and transmission of true beliefs and the correction of
false ones. It is better, therefore, to remain more agnostic about the diver-
sity of the good than Mill is, while still emphasizing, as he did, the impor-
tance of access to a wide range of opinions and ideas for determining
what is best for us and what the most efficient means are for attaining it.
Third, and most important, Mill has too little to say about the specific
institutions and psychological dispositions that are needed for reliable
selection of true beliefs among the unruly mass of beliefs thrown up
by the exercise of freedom of thought, expression, and association. He
assumes, more than argues, that false beliefs will get winnowed out
through rational argument and through reflection on the effects of
acting on them, both in our own lives and through learning from the
experiences of others. But as social epistemology suggests, and as I shall
argue in the next two subsections, more specific institutional mecha-
nisms and psychological dispositions are needed for reliable selection of
true beliefs and correction of false ones.17 As I shall elaborate below,
among the most important of these are (1) effective processes for iden-
tifying epistemic authorities on grounds of merit, understood as the
possession of objective qualifications, and (2) a widespread, though not
necessarily universal, limited epistemic egalitarianism, an attitude that
fosters both a willingness to challenge putative authorities and a pre-
sumption that no one’s views are to be dismissed or discounted simply
because she is a woman, or belongs to a particular racial or ethnic group
or is a member of the working class, and so on. Both of these mecha-
nisms for the selection of true beliefs are distinctive of liberal societies,
or at least are likely to be much more robust in liberal societies.
Merit-Based Social Identification of Experts

I noted earlier that there are great advantages to a stable social division
of epistemic labor within a society, a set of processes for identifying
certain individuals or groups as experts, upon whom others can depend
as reliable sources of belief. I also noted, however, that reliance on puta-
tive epistemic authorities can impose grave prudential and moral risks,
if epistemic deference is excessive or misplaced. I now want to argue that
although liberal societies do not eliminate these risks, they are generally
17. Goldman, Knowledge in a Social World, pp. 123–59.

Social Epistemology
likely to do a better job of reducing them than nonliberal societies,

chiefly because liberal societies tend to rely more heavily on merit, that
is, on objective qualifications for the identification of epistemic author-
ities than nonliberal societies and also include more effective provisions
for ensuring that these qualifications are rationally related to the beliefs
for which putative epistemic authorities are identified as being reliable
sources. The crucial point is that where epistemic authority is based
chiefly on merit, epistemic efficiency is enhanced through a rational
division of labor, and the risks of surplus epistemic deference are
reduced. Obviously, this point applies only to the comparison of liberal
and nonliberal modern societies, those in which there is an extensive
division of labor.
Here the history of liberalism provides a valuable clue. Recall that a
major slogan the liberal revolutions of the late eighteenth to mid-
nineteenth centuries was carrieres ouverte aux talents (careers open to
talents): Access to desirable offices and positions should be based on
merit, that is, on objective qualifications rationally related to the role or
position in question, rather than hereditary status (religious or secular)
or kinship, race, or ethnicity, or loyalty to the rulers. Historically, the
liberal justification for determining access according to merit has been
that to do otherwise arbitrarily discriminates against individuals and to
that extent is a violation of the fundamental equality of persons.
However, from the standpoint of social epistemology, what makes the
appeal to objective qualifications valuable is that it facilitates a more reli-
able identification of experts and, with this, a more efficient allocation
of epistemic deference. To the extent that liberal societies tend to rely
more on the identification of epistemic authorities on the basis of objec-
tive qualifications than do nonliberal ones, they can be expected to do
a better job, not only of producing more beliefs through a freer exchange
of ideas and information, but also of effectively selecting true beliefs
from among them.
The same point can be made using a distinction between two types
of trust I have developed elsewhere.18 Status trust, which includes epis-
temic deference, is accorded to persons or groups simply on the basis
18. See Allen Buchanan, “Is There a Medical Profession in the House?” in Conflicts of
Interest in Clinical Practice and Research, ed. Roy G. Spece (New York: Oxford University
Press, 1996), pp. 105–36; “Toward a Theory of the Ethics of Bureaucratic Organizations”;
“Trust in Managed Care Organizations,” Kennedy Institute of Ethics Journal 10 (2000):
189–212.
of their being identified as having a certain status or of their being

a member of a certain group. Merit trust, in contrast, is individual-
performance based; it is conferred on an individual on the basis of an
appraisal of her own actions or attributes, so far as they are regarded as
exhibiting appropriate qualifications.
Yet merely believing that someone has expertise in a certain domain
is an insufficient basis for even minimally rational epistemic deference.
Also needed is trust in his or her veracity. Thus status trust, like other
forms of epistemic deference, includes not only confidence in its object’s
ability to make sound judgments but also the belief that the expert can
be relied upon to convey those judgments truthfully to others and to act
appropriately on them in carrying out her special tasks. For example,
physicians are typically regarded not only as experts on human health,
but also as being committed to serving the health interests of their
patients. When they are so regarded, the mere fact that an individual
is believed to be a physician will lead one to tend to defer to his or her
judgments about what one should do for the sake of one’s health more
than to the judgments of others who are not thought to be physicians.
Similarly for teachers: They are regarded as reliable sources of belief,
not simply because they are thought to possess truths, but also because
they are trusted to convey truths accurately to those in their charge, and
are assumed to be dedicated to conveying them in a way that can be
understood.
Where the medical profession is held in high regard, individual
physicians are accorded considerable status trust, and the epistemic
deference regarding matters of health care that this includes, by patients
who have no basis for conferring trust on them on grounds of their actual
competence as individuals. Where teachers are objects of status trust,
individual teachers are accorded epistemic deference, not just about the
information they convey to their students, but also about proper educa-
tional practices, independently of whether they actually merit such epis-
temic deference. Likewise, where priests or ministers or other occupants
of religious roles enjoy status trust, they are accorded epistemic defer-
ence that may not be warranted by their actual performance in supply-
ing reliable moral or spiritual guidance, sound interpretation of religious
texts, and so forth.
Status trust is valuable. It alleviates the anxiety of having to make
one’s own judgments concerning matters about which one has too little
Social Epistemology
information, and enables one to rely on the judgment of another without

incurring the costs of investigating her qualifications. In simplest terms,
status operates in much the way a brand name does in the market.
But status trust can be misplaced or excessive, and this can be danger-
ous. If one overestimates either one’s physician’s competence or her
commitment to one’s health interests, one may accept, as sound, judg-
ments that are not only false, but dangerous to one’s health (and one’s
pocketbook). If one overestimates the reliability of religious or govern-
ment authorities or teachers, the result may be moral or prudential
disaster, or both.
Because of their commitment to careers open to talents, liberal soci-
eties, more so than nonliberal ones, tend to reduce the role of status trust
in the epistemic division of labor, or at least to tie status trust more
closely to merit. To return to the previous example, in a liberal society,
where there is an emphasis on objective qualifications for positions of
authority generally, including professional roles, status trust in physi-
cians will largely be a function of the belief that the status of physician
is conferred only on individuals who merit trust so far as health care is
concerned, because they have undergone a rigorous education and
training that confers objective qualifications regarding the provision of
health care and that inculcates a sincere commitment to the patient’s
well-being. Status trust, under these conditions, is a kind of indirect
merit trust, trust that is derived from beliefs about characteristics of the
institutions that confer status.
Furthermore, the extensive exchange of ideas and information made
possible by the key liberal institutions of freedom of thought, con-
science, expression, and association when combined with the promi-
nence of merit as the criterion for being identified as an expert also
contributes to an efficient allocation of epistemic deference. Access to
information empowers individuals and groups to combat excessive or
misplaced status trust and thereby to avoid the moral and prudential
risks they entail. For example, suppose that it becomes common knowl-
edge that physicians who own CAT scan machines order CAT scans two
to three times more frequently than those who do not, even when
dealing with patients that are medically indistinguishable. If many other
particular instances of physician conflicts of interest become generally
known, and if the media make clear that the structure of the private
health care delivery system creates powerful financial incentives for
physicians to serve corporate interests, then patients may come to doubt

their assumption that physicians put their patients’ interests first, and
this may lead them to be less likely to rely on a physician’s judgment
simply because she is a physician. Patients may come to base their epis-
temic deference more on their beliefs about the actual skills and dedi-
cation of a particular physician than upon status.
Similarly, if it becomes general knowledge that the processes by which
priests are educated, trained, and supervised do a poor job of weeding
out pedophiles, then status trust in priests is likely to decrease. Erosion
of status trust in the case of physicians may be beneficial to patients,
resulting in the avoidance of unnecessary and expensive procedures.
Erosion of status trust in priests may also be beneficial, to the extent that
it results in parishioners no longer omitting to take reasonable pre-
cautions to protect their children from sexual abuse. The freedom to
attain and exchange information provides opportunities for determin-
ing whether individuals traditionally regarded as reliable sources of
belief and moral or practical guidance are in fact reliable, and to that
extent reduces reliance on status trust.
Here it might be objected that, although pre-modern nonliberal soci-
eties tend not to rely on access to positions according merit, modern
nonliberal societies, such as those of Nazi Germany or the Soviet Union
or present-day Saudi Arabia or China do, and that because this is so, the
epistemic benefits that accrue to “careers open to talents” are not unique
to liberal societies. It is of course true that modern nonliberal societies
do not typically allocate positions of authority, and hence epistemic def-
erence, according to caste or hereditary or religious status, in the way
that traditional nonliberal societies do. Instead, such modern nonliberal
societies typically purport to allocate positions of authority, and hence
epistemic deference, on grounds of merit.
However, there are two features of modern nonliberal societies that
tend to make them less successful than liberal ones in reaping the epis-
temic benefits of “careers open to talents.” First, they tend to obliterate
or blur the distinction between the public and private spheres, or at least
to diminish the private sphere, as compared with liberal societies. The
larger the domain of the political, the more opportunity there is for party
loyalty or ideological purity to usurp the role of objective qualifications.
Second, although both liberal and nonliberal societies are subject to
departures from “careers open to talents,” it is more difficult to constrain
Social Epistemology
the influence of nonmerit considerations in nonliberal societies. In

nonliberal societies people have less access to information relevant for
determining whether positions are being allocated according to merit.
In addition, the lack of political liberty and freedom of expression and
association makes it difficult if not impossible for people in nonliberal
societies to hold the government accountable for departures from the
meritocratic ideal.
In a liberal society, freedom of information and the freedom to criti-
cize authorities, under conditions of freedom of association, enable indi-
viduals to challenge the assumption that the process of credentialing
public officials, physicians, priests, or teachers does in fact allocate posi-
tions of authority according to objective qualifications. Such challenges
may result in an erosion of status trust, or the beginning of a process
through which status trust is established on a new basis, if there are
reforms in the credentialing process. In societies that lack the free
exchange of ideas and opinions among the public at large it will be more
difficult to avoid the unwarranted epistemic deference that excessive or
misplaced status trust tend to promote, because it will be more difficult
to get information relevant to determining whether or to what extent
status trust is in fact warranted or to gain the information needed to
replace status trust with merit trust.
In a liberal society, the general public will have access to information
on the basis of which to determine whether the processes for creating or
identifying objects of epistemic deference are in fact performing as those
who control those processes say they are. But members of liberal soci-
eties are also better able to raise more basic questions, from time to time,
about the scope and limits of the roles of those who are designated as
experts. For example, where there is a free exchange of opinions and
information and a willingness to criticize accepted values that distin-
guishes liberal from traditional societies, parents may question not only
whether the process by which teachers are trained is equipping them to
discharge the role of teacher as it has been traditionally conceived, but
also whether the role has been properly conceived. Parents may come
to reject the assumption that teachers are to inculcate Christian moral-
ity or patriotism in students or to teach girls to cultivate the “feminine
virtues,” and this change in how the role is conceived may be achieved
from the outside, as it were, by the initiative of persons who are not
teachers, or trainers of teachers, or the religious or governmental
authorities who previously have been thought to be the proper arbiters

of what the teacher’s role is.
Where the lack of liberal institutions makes it difficult to counteract
the influence of religious affiliation, hereditary status, race, ethnicity,
clan or family connections, party loyalty or ideological purity on the allo-
cation of authoritative roles, the risk of unwarranted epistemic defer-
ence will be greater, other things being equal. This risk will be especially
high if the same processes that allow nonmerit considerations to influ-
ence the allocation of epistemic deference not only restrict the public’s
access to information needed to assess the reliability of those who are
identified as authorities, but also silence or penalize those who question
either particular occupants of authoritative roles or the accepted con-
ception of the nature of the roles. The free exchange of information and
the freedom to criticize authority that liberal institutions facilitate do not
eliminate the risk of the surplus epistemic deference that the social
identification of experts inevitably entails, but they do ameliorate it
significantly.
In liberal societies, many of the more important government officials
are not selected according to merit criteria; they are elected. Although
they typically present themselves as having expertise in appealing to cit-
izens for their votes, voters are free to apply other criteria, such as how
well the individual is likely to further their interests or values if elected.
The literature on the epistemic benefits of democratic selection of gov-
ernment officials is extensive and it is not my aim to canvass or add to
it here.19 My focus, instead, is on the role of those who are identified as
experts, where this includes not only nonelected government officials,
but also individuals who are not part of government but nonetheless can
play an important role in the social formation of beliefs. In part because
elected officials are not identified primarily as experts, but depend in
various ways upon those who are thought to be experts, the role of
19. Among the most important contributions to this literature is Amartya Sen’s finding
that democratic societies tend not to have famines because democratic accountability, at
least under conditions of freedom of information about the behavior of government, tends
to prevent government from persisting in policies that lead to famines. See Amartya Sen,
Poverty and Famines: An Essay on Entitlement and Deprivation (Oxford: Clarendon Press,
1981). The more general point is that the availability of information about what government
is actually doing, along with accountability through electoral processes, tends to produce
better policy, or at least to reduce the risks of persistency in bad policies.
Social Epistemology
experts in liberal societies can be especially significant in the social for-

mation and correction of beliefs.
Thus far I have emphasized the importance of the freedom of
expression and information for curbing unwarranted epistemic defer-
ence. It is worth emphasizing that the mere availability of information
and the absence of penalties for challenging authority, although neces-
sary, are not sufficient to curb the tendency toward unwarranted epis-
temic deference. Whether enough people will be sufficiently motivated
to seek out the relevant information and to use it to challenge and
thereby constrain status trust will depend in part upon how pervasive
an attitude of epistemic egalitarianism is. Liberal institutions, as I shall
now suggest, both require this crucial attitude in order to function, and
nurture it.
Egalitarianism as a Constraint on Epistemic Deference

A liberal society is not distinguished exclusively by its institutions, but
also by a widespread, although not necessarily universal, attitude of
basic moral egalitarianism.20 Members of a liberal society tend to believe
themselves to be the equal of anyone, when it comes to fundamental
moral status. This belief in equal fundamental moral status has profound
implications for how they see themselves as citizens. To the extent that
liberal individuals believe that a proper recognition of their fundamen-
tal moral equality requires a commitment to attempting to base social
cooperation on principles for which reasons can be given, and reasons
that are accessible to others who are similarly committed, they also view
themselves as competent to form and revise their beliefs and to give
reasons capable of prompting others to form and revise their beliefs.
Although this fundamental epistemic egalitarianism is compatible
with the recognition that one should defer to others’ judgments in any
number of different domains of expertise—and for that reason I have
emphasized that it is limited—it nonetheless provides a significant
20. In his insightful book Liberal Virtues: Citizenship, Virtue, and Community in Liberal
Constitutionalism (New York: Oxford University Press, 1991), Stephen Macedo argues that
liberal society is properly characterized not only by certain institutions but by distinctive
virtues as well, including the commitment to defend individual liberty. My suggestion is
that the list of liberal virtues be expanded to include certain epistemic virtues, including
what I have called epistemic egalitarianism.
constraint on epistemic deference and hence on the moral and pruden-

tial risk that results from excessive or misplaced epistemic deference.
Epistemic egalitarianism also includes a tendency to resist discount-
ing or dismissing other persons’ views simply because they happen to
be women, or members of another racial or ethnic group or social class.
Where the commitment to the fundamental moral equality of all persons
is absent, it is easier, other things being equal, to dismiss or discount the
views of certain categories of persons due to assumptions about their
natural inferiority. The moral egalitarianism of liberal societies therefore
helps to mitigate the risk of unwarranted moral status beliefs and the
moral risks they carry.
It is perhaps in its commitment to democratic politics, understood as
including the entitlement of all to participate as equals in the creation
of the most important rules of public order, that a liberal society most
concretely and convincingly expresses this fundamental moral egalitar-
ianism and its implications for the limits of epistemic deference. Demo-
cratic institutions, at least in their deliberative processes, can be viewed
in part as attempts to proceduralize the commitment to intersubjective
justification that lies at the heart of liberalism’s fundamental moral
egalitarianism. In other words, institutionalized support for democratic
political participation both expresses and nurtures the conviction that
everyone’s opinion counts because we are all accountable to each other
in the justification of rules of public order, and this in turn supports the
sense of self-worth that is required for ordinary people to be able to
challenge putative epistemic authorities.
My point is not that epistemic egalitarianism is something to be max-
imized, i.e., that the more ordinary people challenge putative epistemic
authorities the better. Unrestrained epistemic egalitarianism would
deprive us of the benefits of the division of epistemic labor. Instead, the
point is that the right combination of (limited) epistemic egalitarianism
and publicly accountable, merit-based processes for identifying experts
and ascertaining the proper limits of their expertise is extremely valu-
able from the standpoint of epistemic reliability. Liberal societies do not
necessarily achieve this optimum, but to the extent that they feature a
measure of widespread epistemic egalitarianism empowered by the free
exchange of ideas and a highly developed division of epistemic labor,
they at least supply the ingredients for approximating it.
Social Epistemology
Moral Status Judgments

I observed earlier that the comparative evaluation of institutions should
take into account not only the quantity of true beliefs produced or the
ratio of true to false beliefs, but also how well a set of institutions does
in producing, preserving, and transmitting the most practically impor-
tant true beliefs. I also noted that among the latter are beliefs that shape
our judgments concerning fundamental moral status. I now want to
elaborate what I have already suggested: that liberal institutions, when
combined with the attitude of moral egalitarianism characteristic of
liberal societies, provide powerful protections against erroneous judg-
ments concerning fundamental moral status.
Where all persons are accorded the basic civil and political rights that
constitute a liberal social order, and where the principle of careers open
to talents is realized, it is much more difficult to sustain social practices
that systematically distort our experience of what blacks are like. To the
extent that the inculcation and preservation of beliefs about the nature
of blacks that consign them to an inferior moral status depend upon dis-
torted experiences of what blacks are like, liberal institutions, by reduc-
ing the inequalities that make such distorted experiences possible,
provide a check on false beliefs about the moral status of blacks. Fur-
thermore, so far as the prevalence of false moral status judgments
depends upon their being inculcated by persons who are objects of
status trust—such as religious and political leaders—those features of
liberal society that provide checks on unwarranted status trust can also
serve to reduce the risk of false moral status judgments. Where there is
freedom of information and the free exchange of ideas, and where many
individuals feel competent to question authorities, the dangers of
unwarranted status trust will be reduced. Where status trust is con-
strained by freedom of information and an attitude of epistemic egali-
tarianism, it will be more difficult for religious and political leaders to
convince people that certain groups are of a lower moral order, espe-
cially if liberal institutions have undercut the distorted experiences that
seem to confirm what these objects of status trust say is true.
Here it might be objected that the attitude of moral egalitarianism by
itself eliminates erroneous moral status judgments and that therefore
there is no need to appeal to the role of the key liberal institutions. This
is not the case, however. First, the attitude of moral egalitarianism in a

society may not be sufficiently widespread to rule out prejudice toward
certain groups and the distorted experience of what their members are
like that prejudice helps to produce. Second, it is a mistake to assume
that once an attitude of moral egalitarianism is widespread in a society,
it cannot diminish. Especially in times of economic depressions and
other situations in which insecurity is widespread and the inclination to
single out some group for blame is strong, there can be a tendency to
regard some of our fellow citizens as less than our moral equals. Institu-
tionally supported freedom of expression and the free exchange of infor-
mation make it harder to create and sustain the false beliefs about
natural differences among groups that can erode the moral egalitarian-
ism of a liberal society. Similarly, where merit plays a fundamental role
in access to a full range of desirable positions, there will be fewer oppor-
tunities for sustaining erroneous beliefs about the moral status of
minorities, because they will not be systematically barred from positions
in which their true abilities can be exercised and observed. In that sense,
meritocratic institutions and freedom of expression and access to infor-
mation can help ensure that people regard all their fellow citizens, not
just some, as their equals.
III. The Scope and Limits of the Social Epistemology Argument for
Liberal Institutions
Not Truth or Knowledge for Its Own Sake

Thus far I have argued that liberal societies are preferable to the extent
that they are more likely to achieve an optimal reduction of moral and
prudential risk from socially inculcated false beliefs. But my arguments
also appear to support a more general conclusion: that liberal societies
are more epistemically efficient simpliciter. In other words, liberal insti-
tutions not only are better at reducing the risk of damaging false beliefs,
but also are superior from the standpoint of producing true beliefs, and,
assuming that knowledge includes true belief, are therefore better gen-
erators of knowledge.
However, I have deliberately not based the case for liberal institutions
on the claim that they are more likely to generate true belief, and to that
extent, knowledge, because my goal is to provide a compelling argument
Social Epistemology
for liberal institutions within the constraints of political liberalism, that

is, without relying on any comprehensive conception of the good. Basing
the case for liberal institutions on their superior capacity to generate
knowledge would violate this constraint by assuming that knowledge is
the highest good or at least is an important component of the highest
good. Instead, I have only argued that one will support liberal institu-
tions if one duly appreciates both the value of social institutions in the
production of true beliefs and the moral and prudential risks of social
epistemic dependency. My argument for this conclusion does not
assume any particular conception of the good, whether religious or
philosophical, including those that view truth or knowledge as intrinsi-
cally valuable.
Not Autonomy for Its Own Sake

My argument suggests that what many proponents—and critics—of lib-
eralism have said about it is true: Liberal institutions both require and
nurture individual autonomy. Some degree of individual autonomy, in
some significant portion of the population, appears to be necessary if
the key liberal institutions are to function well enough to reduce the risks
of our unavoidable social epistemic dependency. At least some people
must be capable of thinking for themselves and of critically evaluating
what they and others have taken to be sufficient reasons for believing
and acting.
But from this it does not follow that one must or should base argu-
ments for liberal institutions on a conception of human good that places
autonomy at the summit. I have argued that anyone who takes seriously
her own well-being or the commitment to acting morally, and who
understands the implications of our ineliminable social epistemic
dependency, has a stake in the existence of liberal institutions. If there
must be some individuals who live an autonomous life for those insti-
tutions to function effectively, then anyone who takes her own well-
being or morality seriously has an interest in there being some
autonomous lives and in there being the sorts of social arrangements
that make autonomous lives possible. However, this is not to assume that
autonomy is valuable for its own sake, much less that it is the highest
good. The social epistemology argument, then, does not rest the case for
liberal institutions on autonomy-perfectionism. For political liberals,
and for those who reject the constraints of a political liberal justification
but are not convinced that autonomy is the highest good, this is an
attractive feature.
Room for Nonliberal Associations and Nonautonomous Lives?

So far I have argued that anyone who takes acting morally or her own
well-being seriously has a stake in the existence of the key liberal insti-
tutions. From this alone it does not follow that everyone who takes acting
morally or her own well-being seriously should support only liberal insti-
tutions or should participate only in liberal social institutions or live
an autonomous life. The social epistemology argument for liberal in-
stitutions, strictly speaking, is not an argument for everyone being
autonomous or for prohibiting nonliberal associations, institutions, or
even societies.
It is worth considering whether people could live nonautonomous
lives, committing themselves to nonliberal communities, but nonethe-
less reduce the moral and prudential risks of doing so, if they and their
communities are connected to liberal institutions in such a way that they
benefit from the latter’s epistemic risk-reduction mechanisms. In other
words, just as there can be an epistemic division of labor within a society,
might there not be a division of epistemic risk reduction between soci-
eties or social groups? There are three main ways in which a nonliberal
society or social group might be able to enjoy, to some extent, the reduc-
tion of epistemic risk provided by liberal institutions.
(1) Ensure that information of many sorts and an awareness of the
possibility of challenging epistemic authorities penetrate into nonliberal
social environments. In some cases, merely becoming aware that others
challenge the accepted view can reduce unwarranted epistemic
deference to those who are regarded as authorities within one’s own
community.
Just as importantly, promoting freedom of movement between liberal
and nonliberal social environments, as well as access to personal elec-
tronic communications, may do much to chip away at the prevalence of
false beliefs, diminish the unwarranted epistemic deference that fosters
them, and even make it harder for those whose beliefs and attitudes are
systematically distorted to continue to succeed in sustaining their errors.
To take but one prominent example of this phenomenon, the marked
Social Epistemology
increase in freedom of movement and expression facilitated by the

Helsinki Accords of 1975, combined with the electronic communication
revolution, did much to erode the trust of citizens of nonliberal state
socialist regimes in their leaders and in the viability of their forms of
social and economic organization. Correction of false beliefs resulted
in loss of legitimacy, depriving these regimes of the popular support (or
at least acquiescence) necessary to persist in policies that were both
immoral and against the best interests of the general public.
(2) Provide viable exit options. Hirschman emphasizes that the exis-
tence of a viable exit option can facilitate improvement of a malfunc-
tioning organization, whether it is a corporation, a community, or a
state.21 If members of an nonliberal society or a nonliberal enclave within
a liberal society can leave it, then the desire to avoid defection can lead
those who control the nonliberal social environment to improve the
performance of their epistemic authorities.
How effective the exit option is in reducing the moral and prudential
risks of socially inculcated false beliefs will depend upon how success-
ful strategy (1) above is. Unless information about the relative attractions
of liberal arrangements penetrates the nonliberal society or enclave, its
members will not seek to leave it or be able to make a credible threat to
leave it.
(3) Reduce the moral and prudential risks of persisting in false,
socially inculcated beliefs by limiting the power of nonliberal groups to
implement their views, thus ensuring that whatever damage is done by
the nonliberal group’s socially inculcated false beliefs is limited in scope.
The most obvious way in which this form of damage control could be
achieved is if the nonliberal society or group were embedded in a larger
liberal society. The same goal could be achieved, however, if a powerful
liberal state, or a coalition of liberal states, were willing and able to exert
influence, and even in the last resort to threaten or use force, to prevent
nonliberal groups from exporting dangerous policies or implementing
them on a large scale. Other things being equal, the comparatively
limited ability of nonliberal societies to avoid or at least to correct false
beliefs argues for preventing them from becoming so powerful that they
can evade the constraints on epistemic risk that their being subject to
21. Albert O. Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Orga-
nizations, and States (Cambridge, Mass.: Harvard University Press, 1972).
the influence of liberal societies places on them, either in the form of

ensuring an exit option or ensuring that they are penetrable by infor-
mation and ideas from the outside.
Notice that my concern here is not to argue that members of liberal
societies ought to or may impose such connections on nonliberal soci-
eties. That is an interesting proposal, but I am not advancing it. My point,
rather, is that it is in the interest of members of nonliberal societies to
ensure that such connections exist. Thus, my argument in the present
article is for self-binding, not paternalism. In other words, the conclu-
sion I seek to establish is that those who attempt to sustain their engage-
ment in nonliberal associations ought to support, or where necessary
create, the sorts of connections to liberal institutions that are needed to
reduce the moral and prudential risks they bear, not that liberals ought
to impose such connections in order to protect nonliberals. However, to
the extent that the unchecked prudential and moral risks to which non-
liberal societies are prone do pose a serious risk to liberal societies, the
social epistemology argument also provides a premise in a possible
justification for liberal intervention in nonliberal societies, if such
intervention is needed to achieve the connections required for reducing
that risk.
Might it not turn out that connections to liberal institutions that are
strong enough to achieve a significant reduction of epistemic risks would
eventually destroy or seriously damage nonliberal forms of association?
Whether this is so is in the end an empirical question that we are prob-
ably not now in a position to answer. Notice, however, that if efforts to
forge or sustain such risk-reducing connections to liberal institutions do
result in damage to nonliberal societies or forms of association, those
who lament the loss cannot say that it occurred as a result of the effort
to eliminate nonliberal forms of association or to force all to live the life
of autonomy.
The social epistemology argument proceeds on the assumption that
epistemic risk reduction is valuable for all who take seriously their own
well-being and the commitment to acting morally. But the argument can
also acknowledge that the goods provided by nonliberal communities
can be valuable and perhaps even unique to them. Because it can rec-
ognize both values, the argument allows for trade-offs between epis-
temic risk reduction and protection of nonliberal societies, or, more
accurately, the distinctive goods that they provide. What the argument
Social Epistemology
shows is that even those who do not value autonomy or liberal institu-
tions for their own sake at the very least have an important interest in
ensuring that the forms of association and modes of living they seek to
sustain are connected to liberal institutions in such a way as to provide
some significant reduction of the moral and prudential risks of being
subject to deeply entrenched, socially inculcated false beliefs.
The social epistemology argument therefore implies that it is irra-
tional for anyone who cares about avoiding moral and prudential error
as a result of having socially inculcated false beliefs to support the com-
plete destruction of liberal institutions and their replacement by non-
liberal ones. If my argument is sound, then those religious or fascist
imperialists who seek to eradicate liberal social orders from the earth
cannot consistently and sincerely say that they are committed to acting
morally, even according to their own conception of morality, unless they
are willing to claim that their understanding of what morality requires is
totally immune to distortion by false beliefs of any kind, including false
factual beliefs that can lead to erroneous applications of moral princi-
ples. To put the point somewhat ironically, the social epistemological
argument shows that nonliberals ought to tolerate the existence of
liberal institutions—indeed ought to be willing to help ensure that they
exist if necessary—not because they value or ought to value tolerance for
its own sake or as an expression of respect for persons, but rather for the
sake of their own (nonliberal) commitment to morality.
The social epistemological argument makes a case for supporting
liberal institutions that relies only upon a commitment either to one’s
own well-being or to morality, not upon any particular conception of
individual well-being or any particular moral view. However, the argu-
ment does not purport to show that everyone ought to support liberal
institutions regardless of what views he or she holds about truth. If one
holds a conception of the good that includes the belief that one has
access to all the truths that are needed (and ever will be needed) for
acting prudently or morally, or believes that one can (and always will be
able), infallibly, to identify those who do have such access, then the
social epistemology argument will leave one cold.
It is worth emphasizing just how extreme, and extremely implausible,
such an assumption of immunity to false belief is. By way of illustration,
consider the case of a member of a religious community who believes
that there is no need for the existence of liberal institutions (or of
autonomous individuals) because he knows that the sacred text or the

deliverances of the priesthood are immune from error. To conclude that
there is no problem of prudential or moral risk due to socially inculcated
false beliefs, he must assume not only that the sacred text is infallible,
but also that his epistemic authorities are infallible not only in explicat-
ing the meaning of the text, but also in their factual beliefs about how
the principles in the text are to be applied.
The social epistemology argument does not show, then, that even the
most extremely “fundamentalist” of religious types ought to support the
existence of liberal institutions. It does show that most of those we now
call religious fundamentalists have good reason to do so, since it is pre-
sumably only a minority of those who are given this label that claim the
extreme epistemic infallibility I have just described.
The argument also remedies a defect of those justifications for liberal
institutions that appeal to an overriding commitment to preserving civil
peace under conditions of deep religious or ideological pluralism. Such
arguments have no force against those who are willing to put their com-
mitment to ensuring that all people live under the nonliberal social order
they endorse ahead of their commitment to peace, even if this means
killing or coercing a good many people. The social epistemology argu-
ment for liberal institutions, in contrast, does not assume that all of
those to whom it is addressed give priority to the avoidance of massive
bloodshed or coercion over the creation of an nonliberal social world. It
gives anyone who appreciates the risk of prudential or moral error—
even those who are willing to impose their conception of the good by
lethal force—a weighty reason to support the existence of liberal insti-
tutions. How weighty is this reason? It is as weighty as the commitment
to acting morally or to pursuing our own good effectively.
Similarly, the social epistemological argument for liberal institutions
should have force even for those who deny what Locke held, namely, that
forcible conversion is incompatible with true religion.22 Even a religious
imperialist who denied Locke’s thesis and would otherwise be willing to
exterminate all those who do not convert voluntarily in order to realize
his dream of a world populated exclusively by believers should not desire
the total destruction of liberal institutions.
22. John Locke, A Letter Concerning Toleration [1689] (New York: Prometheus Books,
1990).
Social Epistemology
The difference between Rawls’s political liberalism and the social

epistemology argument should now be clear. The social epistemology
argument is not based on any commitment to tolerance as such. It does
not assume that either morality or reasonableness requires us to refrain
from imposing rules of public order on those with whom we disagree,
out of respect for them as persons who hold different but not unrea-
sonable conceptions of justice or the good. In my judgment, this is an
advantage, because it avoids an obvious objection to which Rawls’s posi-
tion is vulnerable. Most simply put, the objection is that basing the case
for liberalism on this overriding commitment to tolerance of not unrea-
sonable views of others vitiates the political liberal project. The difficulty
is that this overriding commitment to tolerance of not unreasonable
views is as controversial as the comprehensive conceptions of the good
that the political liberal seeks to avoid, and indeed may make sense
only against the background of a Kantian comprehensive conception of
morality that equates respect for persons with a rather extreme inter-
pretation of respect for their reasons.23 For those who find this objection
telling, the social epistemology argument provides a way to preserve the
political liberal project.
Like other arguments for liberal institutions, the social epistemology
argument will not persuade those who hold their belief in the desirabil-
ity of a nonliberal social order to be literally infallible. But unlike those
other liberal arguments, it does not assume that everyone, or even every-
one who is reasonable, puts civil peace above the realization of their reli-
gious or political goals or believes, as Locke did, that genuine religious
belief cannot be coerced.
The Epistemic Effects of Economic Institutions

I have tried to take care to qualify the main conclusion of my argument.
It is not that all societies that include what I have called the key liberal
institutions nor all societies that are called liberal will fare well in reduc-
ing the moral and prudential risks of our inevitable social epistemic def-
erence. My conclusion, rather, is that societies that contain the key
liberal institutions will fare better, other things being equal. For it seems
likely that there are other features of a society, in particular the nature
23. Christiano, “Is There Any Basis for Rawls’ Duty of Civility?”
and influence of its economic institutions, that can counteract the

beneficent operation of the key liberal institutions. Suppose, for
example, that the most effective media in a society systematically distort
the information available to the public and create serious biases in the
way public issues are identified and framed, because powerful economic
interests control them. Suppose that the media also more directly serve
the same economic interests by contributing to a culture of con-
sumerism that diverts the attention and energies of many citizens from
political issues. Under these conditions, some of the most important
benefits of the key liberal institutions may be significantly reduced and
many people will simply not be motivated to take advantage of what
benefits there are. Epistemic egalitarianism is not likely to be an effec-
tive check on unwarranted status trust if an insufficient number of
people care about whether the ‘information’ they are getting is correct
or desire to seek out and employ good information in order to evaluate
critically the claims of expertise.
Similarly, even if experts are selected largely on the basis of expertise,
they will not be reliable if powerful economic incentives serve to corrupt
their judgment. For example, at present there is the concern that
although reliance on the highly technical expertise of leading genomic
scientists is essential for informed public policy deliberations about the
proper application of genomic technologies, their judgment is likely to
be tainted by the fact that they are, increasingly, entrepreneurs and
shareholders in biotech businesses as well as scientists.
What such considerations show is that for the key liberal institutions
to work well in reducing epistemic risks, they must be properly sup-
ported within the larger framework of social institutions. This conclu-
sion strengthens, rather than weakens, the case for the importance of
social epistemological analysis in the defense of the key liberal institu-
tions. By focusing on the conditions necessary for effective epistemic
risk reduction, the social epistemology argument can help us avoid
the mistake of understanding the key liberal institutions too narrowly
or conflating them with only one form they may take. For example,
where the powerful economic interests produce distortions in the free
exchange of information and biases in the identification and framing of
public issues, it is inadequate to identify freedom of expression with a
legal-constitutional principle that focuses exclusively on preventing
Social Epistemology
interference with freedom of expression. Instead, social epistemological

analysis can help make it clear that without corrections for the influence
of powerful economic interests, freedom of expression thus narrowly
defined may not be effective in reducing epistemic risks. Understanding
the vital role of these institutions in reducing the risks of socially incul-
cated false beliefs provides strong arguments for attempting to under-
take the economic and political reforms needed to ensure that their
potential for risk reduction is realized. The greater our understanding of
how the key liberal institutions tend to reduce epistemic risk—and of the
conditions under which they can effectively do so—the better we will be
able to target the needed reforms. This article takes the first step, by iden-
tifying some of the ways in which the key liberal institutions can reduce
epistemic risks. Proceeding to a full analysis of the conditions that can
reduce their effectiveness and proposals for how to mitigate these coun-
tervailing influences is beyond its scope.
IV. Conclusion
I have argued that key liberal institutions, when combined with the
prevalence of access to positions of epistemic authority according to
objective qualifications and the broad-based fundamental egalitarian-
ism that is characteristic of liberal society, provide important safeguards
against the moral and prudential risks of socially inculcated false beliefs,
safeguards that are not present in nonliberal social orders. The inference
I have drawn from this is not that all associations must be liberal or that
all should live the life of autonomy often associated with liberalism.
Instead, the point is that even those who cherish nonliberal forms of
association and modes of living and who find no intrinsic value in
autonomy need to be connected in appropriate ways to liberal institu-
tions and to people who have the attitude I have characterized as moral
egalitarianism.
This social epistemology argument for liberal institutions satisfies the
requirement of political liberalism: It makes a strong case for liberal
institutions without relying upon any particular conception of the good
or any comprehensive moral conception, grounding it instead in the
commitment to ameliorating the moral and prudential risks to which we
are all liable by virtue of our social epistemic dependency. Furthermore,
as I noted earlier, the social epistemology argument has the virtue of

avoiding reliance upon what some take to be a problematic conception
of tolerance.
I now want to offer a conjecture about the relationship between social
epistemology and liberal institutions: An effective discipline of social
epistemology is only likely to exist in liberal society, because only in
liberal society is there likely to be the free access to information about
the social processes of belief formation and the willingness and ability
to investigate the conditions of rational epistemic deference that the
enterprise of social epistemology requires. Liberal societies, therefore,
are doubly superior from the standpoint of reducing the moral and pru-
dential risks of socially inculcated false beliefs: (1) they not only contain
institutional mechanisms and widespread attitudes that are needed to
detect and correct false beliefs and to achieve an efficient allocation of
epistemic deference; but (2) they also help foster the development of a
discipline that can reflect on these epistemic resources and provide
intellectual tools for improving them.
I introduced this rather abstract investigation of the comparative
epistemic virtues of liberal institutions with a concrete, personal experi-
ence—the bitter indignation and profound sense of vulnerability I felt
when I came to understand that my epistemic deference to authority
figures had placed me at great moral risk. This experience not only moti-
vated my investigation, but also confirms one of its central conclusions:
It was only because the nonliberal society in which I lived was connected
to and constrained by a larger, more liberal society and was to some
extent penetrated by its free exchange of ideas that it became possible
for me to come to understand the need to free myself from the web of
false beliefs that were endangering me. The nonliberal society in which
I grew up was not so thoroughly nonliberal as to make exit from it impos-
sible; and because of that, I was able to accelerate the process of divest-
ing myself of debilitating socially inculcated beliefs, by escaping the
influence of the social environment that produced and sustained them.

Buchanan 2004

Uploaded by

Document Information

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

Buchanan 2004

Uploaded by

Copyright:

Available Formats

ALLEN BUCHANAN Political Liberalism and

I. A Different Kind of Political Liberalism

The Moral and Prudential Risks of Social Epistemic Dependency

moral risk my upbringing imposed on me. Given the right circum-

1. Christine Reyna, “Lazy, Dumb, or Industrious: When Stereotypes Convey

to determine which social arrangements are least prone to inculcate,

I will also suggest—although it is not part of my central argument—

Social Epistemic Dependency, Liberal Institutions, and

Another feature of my argument is worth remarking. It does not base

Justifying Liberalism Without Over-Selling Autonomy

beliefs, through a comparatively reliable allocation of epistemic

Reducing, Not Eliminating, the Moral and Prudential Risks of Social

learned to our sorrow, epistemic deference can be risky. Sometimes the

groups, and those that ascribe different moral statuses to individuals or

Truth, Well-Being, and Morality

14. Buchanan, “Social Moral Epistemology.”

alternative institutional arrangements, but their epistemic efficiency as

II. Liberal Institutions and Epistemic Reliability

Freedom of Thought, Conscience, Expression, and Association

Merit-Based Social Identification of Experts

17. Goldman, Knowledge in a Social World, pp. 123–59.

likely to do a better job of reducing them than nonliberal societies,

of their being identified as having a certain status or of their being

information, and enables one to rely on the judgment of another without

physicians to serve corporate interests, then patients may come to doubt

the influence of nonmerit considerations in nonliberal societies. In

authorities who previously have been thought to be the proper arbiters

experts in liberal societies can be especially significant in the social for-

Egalitarianism as a Constraint on Epistemic Deference

constraint on epistemic deference and hence on the moral and pruden-

Moral Status Judgments

is not the case, however. First, the attitude of moral egalitarianism in a

Not Truth or Knowledge for Its Own Sake

for liberal institutions within the constraints of political liberalism, that

Not Autonomy for Its Own Sake

Room for Nonliberal Associations and Nonautonomous Lives?

increase in freedom of movement and expression facilitated by the

the influence of liberal societies places on them, either in the form of

autonomous individuals) because he knows that the sacred text or the

The difference between Rawls’s political liberalism and the social

The Epistemic Effects of Economic Institutions

and influence of its economic institutions, that can counteract the

interference with freedom of expression. Instead, social epistemological

as I noted earlier, the social epistemology argument has the virtue of

You might also like