Professional Documents
Culture Documents
1
University of Hamburg, Institute of Philosophy, Überseering 35, 22297
Hamburg, Germany Room 02025, julian.mueller[at]uni-hamburg.de
1
Epistemic Democracy: Making Pluralism Productive
Julian F. Müller
1. Introduction
2
Institutional epistemology is about inquiring which set of institutions best
aggregates and produces the information that is needed to solve the specific
problems of specific subsystems of society. We might, for instance, ask what kind
of institution is best suited to the goals of science or economics. According to
Anderson, Hayek’s insights are an important contribution only to the economic
subfield of institutional epistemology (Anderson 2006, 8–9). This is a view, I
submit, that is shared by most epistemic democrats.2
In this paper, I want to challenge this view. Epistemic democrats are right in
believing that Hayek’s price theory – developed in “The Use of Knowledge in
Society” – does not contribute much to democratic reason. However, it is a mistake
to believe that this paper exhausts his epistemic contribution. If we want to
understand his contribution to epistemic democracy, I suggest, we should rather
look at “Freedom, Reason and Tradition” published in Ethics in 1958. Here, as in
ensuing work, he develops his thoughts on the epistemic value of freedom of
(collective) action3; a contribution that so far has largely gone unnoticed by both
his critics and sympathizers.
I will develop my argument in five parts. In the second part, I will give a brief
overview of what I understand to be the core commitments of epistemic democracy.
Next, I will situate what I see as Hayek’s central contribution to the debate about
epistemic democracy by contrasting it with Cass Sunstein’s view. While Sunstein
argues that Hayek’s central contribution is to flag the psychological limits of
deliberation, I will argue that his central contribution is to point out the fundamental
epistemic limits of deliberation. This part will also allow me to distinguish a
Hayekian critique of epistemic democracy from the better known and more frequent
objections by classical liberals such as Jason Brennan (2016) and Ilya Somin
(2013). The fourth part of the essay develops the core theoretical pieces of this
essay, the notion of the fundamental limits of rational deliberation and the notion
of (persistent) hidden champions. Building on these concepts, the essay
reconstructs a distinctive Hayekian challenge to epistemic democracy. However,
Hayek does not only diagnose the epistemic shortcomings of democracy, but also
presents a modicum. In the fifth part of the essay, I will present Hayek’s solution to
the problem of persistent hidden policy champions. Hayek argues that the only way
2
In a short introductory piece to epistemic democracy, Estlund (2008) presents a list of eminent
scholars on whose accomplishments the project of epistemic democracy builds on. The list
features such names as Rousseau, Mill, Peirce, Dewey, Habermas, Rawls, and Rorty. Hayek
on the other hand is absent. In Estlund’s (2009) own major contribution to the field Democratic
Authority, Hayek is just mentioned for his contribution to the epistemology of markets.
3
He mainly develops his thoughts on the epistemic value of freedom in The Constitution of
Liberty, first published in 1960, as well as in a 1968 lecture held at the University of Kiel titled
“Der Wettbewerb als Entdeckungsverfahren” and published in No. 56 in the series Kieler
Vorträge. The article was translated into English in 2002 as “Competition as a Discovery
Process.”
3
to discover persistent hidden champions is by making full use of the (reasonable)
perspectival diversity inherent to democratic institutions. However, to make full use
of a plurality of perspectives, minority factions not only need to be able to voice
their ideas and proposals freely; they also need to have the opportunity, in some
form (e.g. by devolving levels of collective choice), to act on their proposals. The
final part of the paper reconstructs Hayek’s central argument as a cascade model
and offers some supporting empirical evidence.
Before going into medias res let me add a caveat. The purpose of this essay
decidedly is not to defend a certain philosophical tradition against its rivals; it is
also not a piece in the philosophy of history. The goal of this essay is to reconstruct
Hayekian insights in order to make a genuine contribution to epistemic democracy.
2. Epistemic Democracy
4
For this reason, I will confine myself here to pointing out some of the more
obvious reasons that explain why deliberation does indeed often lead to correct
solutions. The epistemic virtue of deliberation is predicated on the epistemic nature
of human beings. Human beings usually possess only bits and pieces of the
knowledge relevant to a particular question. Moreover, people tend to be biased in
all kinds of ways and are prone to make basic errors in thinking through issues. In
such a world, deliberation is an important tool to compensate for the limited
knowledge of individual agents as well as our biases, and to correct the errors that
so easily sneak into our arguments.
Since deliberation under real-world conditions often fails to aggregate and process
knowledge in a desirable fashion, Sunstein proposes that we should think about
alternative ways of aggregating and processing knowledge. In particular, Sunstein
believes that prediction markets might do a better job than deliberative fora.
Prediction markets or information markets are essentially information aggregation
mechanisms. Although there is no commonly accepted definition for prediction
markets, such markets can best be understood as “markets for contracts that yield
4
The full title reads: “Deliberating Groups versus Prediction Markets (Or Hayek’s challenge to
Habermas).”
5
payments based on the outcome of an uncertain future event” (Arrow et al. 2007,
1). Prediction markets, he claims, will “often outperform deliberating groups,
simply because they are so effective at pooling dispersed information among
diverse people” (Sunstein 2006, 193). Consequently, he argues, that for epistemic
reasons, we should think about basing at least some of our (political) decisions on
the results of prediction markets, rather than deliberative fora.
It is not quite clear what exactly Sunstein deems Hayek’s challenge to be. One
interpretation is that Hayek’s challenge consists of two parts. The first part is an
empirical critique of deliberation and the second one is the remedy, i.e., prediction
markets.5 I have two issues with Hayek’s supposed challenge. The first one
concerns the question of how far Sunstein’s challenge is really in a Hayekian spirit.
It is true that Hayek was concerned with the epistemic limits of deliberation, but he
was—as I will show in the next few sections—concerned with the principled rather
than with the psychological, that is, the contingent limits of deliberation. Indeed,
there is a range of scholars that have been deeply concerned with voter ignorance.
Importantly though, Hayek was not one of them.6 Moreover, basing democratic
decisions on the results of prediction markets rather than on inclusive deliberation
might be—at least in some interpretations—in the Hayekian spirit, but is surely not
a proposal that Hayek has suggested or defended himself.
The second worry has to do with the content of the challenge. As far as I can
see, this supposedly Hayekian challenge is a weak one, since it is easily rebutted by
democratic theorists. The results, summarized by Sunstein, should not be too
concerning for advocates of deliberation. In an important sense, these results just
confirm the suspicion of democrats; many democratic theorists have confidence
that deliberation under the right conditions should converge to unanimity. To put it
differently, if deliberations do not lead to an amicable result, democratic theorists
by default suspect that at least some party to the deliberation has been either
irrational, lazy or simply driven by ulterior motives.
Moreover, it is common sense that deliberation is futile if some parties are either
impatient or plain irrational. Advocates of deliberation have thus always
emphasized that the quality of the deliberative decision-making processes depend
on the rules under which they take place. Democrats thus expect that deliberation
that does not adhere to proper rules will produce exactly the effects that
psychologists measure in their experiments (Estlund 2009, 176), (Landemore
2014). At the same time, advocates of deliberation would also expect that these
5
Another interpretation would be that Hayek’s challenge just consists in offering an alternative
to deliberative decisions. That is, decisions based on prediction markets. For the purposes of
this essay, nothing hinges on which interpretation one prefers.
6
Within classical liberal thought, the concern that voters have bad incentives to get informed
for the purposes of voting is associated with public choice economics and such thinkers as
James M. Buchanan and Gordon Tullock. More recent contributions to this field of inquiry are:
Brennan (2016), Somin (2013).
6
effects can be ameliorated by addressing and reflecting on the underlying biases
during the treatments. Indeed, it turns out that advocates of democracy are right.
Studies show that if these biases are addressed in psychological treatments, the
results of deliberative groups significantly improve (Lu, Yuan, and McLeod 2011).
If Sunstein is right and Hayek’s challenge to the epistemic approach of
democracy consists in emphasizing the contingent limits of deliberation and
advocating prediction markets as a partial substitute, then epistemic democrats
seem to be indeed justified in their dismissal of Hayekian insights to democratic
reason.
The remarkable upshot of the last section is that epistemic deliberation—that is,
deliberation that aims for truth—emerges like a phoenix. The results of countless
empirical studies just reaffirms the suspicion of epistemic democrats: that whenever
the better argument does not deliver, some party to the deliberation has been either
irrational, lazy or just ill-willed.
What should we think then of the idea that rational deliberation, under the right
conditions, will usually lead the deliberators to converge on the best, most
appropriate solution?
In this section, I want to develop what I take to be Hayek’s epistemic challenge
to epistemic democracy. Unfortunately, his insights in the epistemic limits of
deliberation have never gained much traction with epistemic democrats or even in
Hayekian scholarship. The goal of the next sections is thus to reconstruct and
translate his insights into an argumentative form that is more accessible for
philosophers.
Let us start by taking a look at one of the central passages in which Hayek talks
about the limits of deliberation in general:
It is worth our while to consider for a moment what would happen if only
what was agreed to be the best available knowledge were to be used in
all action. If all attempts that seemed wasteful in the light of generally
accepted knowledge were prohibited and only such questions asked, or
such experiments tried, as seemed significant in the light of ruling
opinion […]. We might conceive of a civilization coming to a standstill,
not because the possibilities of further growth had been exhausted, but
because man had succeeded in so completely subjecting all his actions
and his immediate surroundings to his existing state of knowledge that
there would be no occasion for new knowledge to appear (Hayek 2010,
34).
7
There is a lot going on in this quote. The main point that Hayek tries to drive
home here, I submit, is that if society limited itself to undertake projects that are
chosen in accordance with deliberative standards of rationality, the progress of
society would come to a halt or be at least severely hampered. Hayek fears then, to
put it simply, that society will miss out on something important if it limits itself to
collective rational deliberation.
8
evidence. To put it differently, we can state that rational deliberation is particularly
well suited to pick out salient champions.
Definition: Salient Champion
• A proposal is a salient champion if and only if it is the best proposal within
a choice set and it is the proposal that is best corroborated by the available
evidence.
The first crucial point I take Hayek to make is that choosing alternatives in
accordance with the demands of rational deliberation cannot ensure that we pick
out the best solution in a given choice set. What he alerts us to is that sometimes
the most justified and the best choice do not coincide but diverge. Sometimes it is
the case that even if we flawlessly aggregate, critically examine and systematize the
evidence, the processed evidence will fail to point us to the best solution. In such
cases, a choice set features, what I call a hidden champion.7
Definition: Hidden Champion
• A proposal is a hidden champion if and only if it is the best
proposal within a choice set, but there is at least one proposal in
the choice set that is better corroborated by the available
evidence.
We can state then that a choice set features a hidden champion if and only if the
evidence is tilted against the best solution. Now, the notion of rational deliberation
as well as the notion of hidden champions presupposes something like an evidential
standard. An evidential standard among other things determines what counts as
evidence and how to weigh different kinds of evidence. For this essay, I will assume
that there are, with regards to moral but also social scientific questions, always a
number of reasonable evidential standards.8 This means that two agents judging on
the basis of the same evidence but on the basis of distinct evidential standards might
come to different conclusions.
To illustrate the notion of hidden vis-à-vis salient champions, we might think
of a moral public problem. A group of deliberators, sharing one epistemic standard,
needs to make up its mind about which of the three policies A, B or C best solves
7
I want to thank Matthew Braham for pointing out that the term “hidden champion” is used in
the business literature to denote small, but highly successful companies especially in the
German context. These companies are hidden in the sense that despite their success the general
public is not aware of them. In contrast, the term has a much more limited meaning in this
paper.
8
For reasons of space, I cannot defend this assumption here in any detail. For a defence of the
view that rational agents are justified in judging evidence according to different evidential
norms, compare Peter (2013), Goldman (2010); for the more general claim that the social
sciences play different explanatory games and thus rely on different standards of what counts
as evidence, compare Mantzavinos (2013), (2016).
9
the problem at hand. The task of the deliberators is then simply to create a ranking
of the policies on the basis of the available evidence. Such a ranking I will also call
‘a perspective,’ because the ranking reflects the underlying epistemic standard of
an individual or group of individuals.
Now, consider the case in which A enjoys high evidential support, B enjoys
medium evidential support and C enjoys low evidential support.
Evidential
High Medium Low
Support
Applying the concept of salient and hidden champions, it follows that if Policy A
is objectively the best solution, then Policy A is a salient champion. On the other
hand, if Policy C is objectively the best solution, even though it is not corroborated
by the evidence, then Policy C is a hidden champion.
In general, policies can be understood as a means to achieve a certain state of
affairs. A policy proposal might then be a hidden champion in two distinct ways.
First a Policy Z might be a hidden champion because a certain deliberative body
might be, even after working diligently through the evidence, unable to recognize
that the Policy Z is the best proposal to achieve a certain desired state of affairs A.
Thus a deliberative body might fail to recognize the instrumental value of Z.
However, a Policy Z might also be a hidden champion because a certain deliberative
body might be, even after working diligently through the evidence, unable to
recognize that the state of affairs A that the Policy Z is able to realize is morally
desirable.9
Building on both the definition of rational deliberation and the concept of
hidden champions, we can now delineate the conditions under which rational
deliberation will fail to converge to the best solution in a given choice set.
Definition: The Fundamental Epistemic Limits of Deliberation
• A group of deliberators, sharing one epistemic standard, that deliberates in
accordance with the standards of rational deliberation, will always fail to
pick out the best solution in a given choice set, if the best solution is a
hidden champion.
The argument here is straightforward. The rationality requirement requires
deliberators always to pick the solution that is best recommended by the evidence.
9
Hayek, interestingly, did not subscribe to the credo of economists that preferences are simply
given. Sometimes, Hayek (2010, 27) argues, we only know what we “want when we see it”.
10
Sometimes the best evidence, however, points in the wrong direction. Whenever
this is the case, rational deliberation will fail to pick out the best proposal.10
10
Some might wonder whether we need Hayek to explain that sometimes justified and true belief
can come apart. The short reply to this worry is this: part of Hayek’s contribution consists in
translating this insight to the realm of social problem-solving especially in the context of
discovering socially useful innovations. Hayek’s full contribution, however, should be
understood as a bundle consisting of both a diagnosis and a treatment of the epistemic
shortcoming of deliberation.
11
at a certain moment in time. This view, I submit, is common to a whole range of
important democratic theorists. The underlying idea here seems to be something
like this: It must be conceded that factions, that hold hidden champions, might not
be able to convince the majority at a certain point of time. Deliberation itself,
however, should be viewed as an open process that extends over time. Within the
process, it is the duty of deliberators that are unable to convince the majority to go
back to the workshop and work on their arguments until the next debate.11
Habermas thus implicitly recognize that there are—at least sometimes—a range of
reasonable views and that the (reasonable) minority is justified in holding on to their
opinion (cf. Courtois (2004), McCarthy (1998, 123–124)) in order to come up with
better arguments.
To get a better sense of the issues at stake, I want to propose a simple thought
experiment. Say there are three factions deliberating about how to achieve a certain
state of affairs A. Let us assume that the factions are of different size: the Majority
Faction has 65% of the votes, the Big Minority Faction has 25% and the Small
Minority has 10%. As, we have argued, Habermas acknowledges—as it is
commonly assumed—that at least for some issues there are many reasonable
perspectives. Accordingly, we want to assign each faction a different evidential
standard. In the next step, every faction ranks the policies according to their
evidential standard. Assume this leads to the following result:
To complete the picture, assume further that Policy C is the objectively best policy
proposal in the option space. Putting the thought experiment in a matrix form gives
prominence to a further important detail of the hidden champion concept. A hidden
champion, as I mentioned, is always defined against a specific evidential standard.
In the thought experiment at hand, Policy C is thus a hidden champion given the
11
This view is also explicitly defended by Elizabeth Anderson. In her 2006 piece, she defends
her approach against the one of Hélène Landemore.
12
specific perspective of the Majority. However, in the perspective of the Small
Minority, the Policy C is a salient champion.12
Earlier, we talked about making pluralism productive. I also mentioned making
use of a diversity of reasonable perspectives. The idea of making use of a diversity
of perspectives in this paper has a rather specific sense that can be appreciated now.
If various individuals or groups have different epistemic standards, this means that
they will have different likelihoods to pick out the best alternative in a given choice
set. The problem, however, as Hayek pointed out, is that we usually do not know,
ex-ante, which epistemic standard best fits the task. Because of that, he argues, we
need a reliable process by which we can find out which epistemic standard will
allow us to find the best solution Hayek.
Now, to continue the thought experiment, at a certain time, t1, the debate about
the policy comes to a conclusion and ballots are cast. The majority, as expected,
wins the vote and Policy A is introduced. Imagine further that Policy A, as the
majority expected, ameliorates the problem at hand somewhat but doesn’t solve it
as Policy C would.
According to Habermas, the small minority which failed to convince the
majority of the objectively best Policy C, after the ballots are cast and counted, now
has the task to go back to the drawing board and enhance their arguments in such a
fashion that eventually the arguments for Policy C will convince the majority in
rational collective deliberation.
This brings an important question to the forefront that so far has received little
attention. To rationally convince the majority, the minority needs to acquire a
certain new bundle of evidence E, a bundle that it originally, that is, in t1, did not
possess. The question then becomes: what mechanism within democratic society
ensures that the minority—at least in principle—will be able to produce the
evidence it needs to convince the majority of a hidden champion at a later point?
Current discussions in epistemic democracy have focused mainly on the issue
of knowledge aggregation and have so far neglected the issue of knowledge
production almost entirely (Mueller 2017). That the issue of knowledge production
is neglected in current debates is reflected already in the basic notion of democracy
employed by epistemic democrats. Democracy is mainly understood as a certain
choice mechanism. Hélène Landemore (2013, 10) for instance emphasizes that she
defines democracy as “a procedure for collective choice decisions” rather than “a
set of political institutions.” David Estlund (2006, 65) similarly employs a rather
thin notion and understands democracy primarily as the “collective authorization of
the laws by voting.”
12
For purposes of expedience, I will speak of a proponent of a hidden champion in order to talk
about a proponent of a policy proposal that is objectively the best, but given the evidential
standards of the majority a hidden champion.
13
The current discussion in epistemic democracy thus does not present any answer
to the question of what mechanism ensures that hidden champions can—at least
over time—be discovered. Historically, as Karl Popper noted, philosophers like
John Stuart Mill and August Comte subscribed to—what Popper calls—the
unconditional theory of epistemic progress. However, the idea that epistemic
progress is unconditional is rather implausible. As Popper (1957) convincingly
argues, epistemic progress in any area is for the most part a function of the rules of
the game. To see this, we only need to consider what levers we would need to pull
to slow down or arrest scientific progress; e.g. decreasing public funding, reducing
academic freedom.
In order to guard against the problem of hidden champions, epistemic democrats
need to defend something like the evidence creation claim.
Evidence Creation Claim:
• The basic epistemic structure of democracy is such that necessary evidence
to uncover hidden champions can and over time will emerge.
Epistemic democrats have yet to give an account of the mechanisms in
democracy that ensure that the evidence needed to unlock hidden champions can
and with some regularity will be produced. However, it seems plausible that
epistemic democrats would endorse the concept of the “basic epistemic structure”
as developed by Kurtulmus and Irzik (2017). The concept of the basic epistemic
structure of a well-ordered democracy denotes the institutions of democracy that
are tasked with creating and disseminating knowledge. The authors develop the
concept within a Rawlsian framework, but it fits the current issue well enough. The
branches of the basic epistemic structure that are tasked with knowledge creation,
according to Kurtulmus and Irzik (2017, 129), are the “institutions of science and
… those government agencies and offices that carry out research or publish basic
statistics.” However, the creation of new knowledge is at least to some extent
dependent on the creation of new data. One of the main sources of data creation is
the democratic political process itself. By introducing new laws or discarding old
ones, the government changes the social ontology of society. This, in turn, produces
new data on the effects and desirability of policies.13
In my reading, Hayek takes issue with the Evidence Creation Claim. He seems
to claim that, given majorities are stable, the proponent of a hidden champion will
often be unable to create the evidence to convince his peers, simply because he is
prohibited, in a sense soon to be explained, from creating the evidence in the first
place. However, if proponents of hidden champions are prohibited (rather than just
contingently unable) from generating the required evidence, this might lead to an
issue of persistent hidden champions. Hayek’s argument for the existence of
13
Another major source for the creation of new data are of course the sciences themselves.
14
persistent hidden champions and thus the epistemic impairment of democracy
revolves around a specific epistemic claim:
14
Compare for instance Zollman (2010). Moreover, in the innovation diffusion literature, that
studies how innovations across domains (innovative social rules, policies, products, etc.)
diffuse through society, it is well-established that the trialability and demonstrability of an
innovation of an ex-post successful innovation is positively correlated with its rate of adoption
(Rogers 2003).
15
option simply because the advocate of the best alternative lacks the evidence that is
needed to convince the majority in rational discourse. Moreover, in cases of a
persistent hidden champion, the advocates will be permanently unable to convince
the majority because they are precluded from producing the evidence that is needed
to convince the majority.
The Hayekian analysis of the epistemic limits of democracy then highlights three
important issues: First of all, the argument highlights that epistemic democracy as
a concept builds on a mostly implicit but problematic premise, namely the premise
that over time, the knowledge that we need to unlock hidden champions will just
emerge. Secondly, it highlights that “knowledge creation” is not an unconditional
process but rather a function of the rules of the game. Moreover, Hayek makes the
point that protecting free speech and academic freedom might be necessary but not
sufficient conditions for unlocking persistent hidden policy champions. Indeed, as
we will see in the next session, to make full use of diverse perspectives, reasonable
minorities must be able—in some form or another—to act upon their epistemic
standards in order to produce the evidence that is necessary to unlock the hidden
champions in the majority perspective.
15
There are various institutional ways to create these ‘spaces.’ I will come back to this point later.
16
Though we must always strive for the achievement of our present aims, we
must also leave room for new experiences and future events to decide which
of these aims will be achieved (Hayek 2010, 22–23).
[N]obody can know who knows best and [...] the only way by which we can
find out is through a social process in which everybody is allowed to try and
see what he can do (Hayek 1980, 15).
It is because every individual knows so little and […] because we rarely know
which of us knows best that we trust the independent and competitive efforts
of many to induce the emergence of what we shall want when we see it (Hayek
2012, 27).
Democracy is not just one form of social life among other workable forms of
social life; it is the precondition for the full application of intelligence to the
solution of social problems (Putnam 1992, 180).
17
Hayek affirms the epistemic value of free speech and concomitantly the value
of epistemic deliberation. His point, however, is that this is not the whole
(epistemic) story. To unlock the gains of diversity, he claims, society not only needs
to secure a space in which people can freely talk about their ideas, but also space in
which they can act upon their ideas. What Hayek seems to have in mind here is that
for epistemic reasons, democracy not only needs to secure a robust right to freedom
of speech but also a robust right to freedom of pursuit.16 According to Hayek, the
epistemic value of the latter freedom, roughly, then consists in enabling society (in
whichever domain it is granted) to exploit cognitively peripheral ideas. In general,
we can protect various values through different institutional arrangements. This is
also true for the freedom of pursuit. In this essay, I want to stay agnostic about how
to implement such a freedom in the political realm and how extensive it should be.
I want to focus instead here on the mechanism that explains how freedom to act on
one’s views can unlock (persistent) hidden champions.
16
Both rights have epistemic and moral aspects, to be sure. For reasons of brevity, however, I
will only concern myself with the epistemic properties of these rights.
18
majority of its favourite candidate by presenting all the available evidence for
Policy C, they do not succeed. Moreover, whenever there is an election, the small
minority collects all the newly available evidence for Policy C but always gets
turned down, since Policy C is a persistent hidden champion. This leads again to
the figure showcased earlier.
Now, what is the Hayekian solution to this problem? The Hayekian solution to
the problem of persistent hidden champions consists in granting the reasonable
factions a right to act on what their evidential standards suggest.
In the present case, this means that all three factions are permitted to implement
their policy for their constituency.17 Assume then that every faction implements its
favourite policy and by the time of t2, Policy C has already produced certain
desirable effects, (which is to be expected since it is a hidden champion). Imagine
that Policy C proves to be effective with regard to increasing the supply of kidneys
under the constraint of not exploiting the weak. Let us further imagine that Policy
B is not making much of a dint. Having arrived in t2 and evaluating the policy
effects in t1, the Big Minority is confronted with a new set of evidence in favour of
Policy C and is therefore forced to re-evaluate its policy ranking. Given the new
evidence, the Big Minority judges that the total evidence (given their evidential
standard) now speaks in favour of Policy C. The Big Minority, therefore, adopts
Policy C in t2.
Assume further that the Majority Faction is still not convinced. The reason
being that Policy A is doing reasonably well and, in addition to that, there are
reasonable worries that Policy C’s positive effects might be not sustainable. One of
the prominent worries by representatives of the Majority Faction, we might
imagine, is that Policy C might decrease kidney donations over the long-term
because of the specific incentive scheme it employs. Thus the reasonable worry is
that Policy C, in the long run, might be unable to keep up its efficacy in providing
a certain high level of kidneys without violating the third criterion.
However, a couple of years later, well into t3, there are still no signs that the
worries by the Majority might come to fruition. Thus, re-evaluating its policy,
adding the new evidence to the existing stock, the Majority finally decides to adopt
Policy C as well, since the total body of evidence (applying the majority’s evidential
standard) now speaks in favour of Policy C. Figure three depicts the policy diffusion
just described.
17
For the purposes of the model, it is unimportant how this right is implemented.
19
T1 T2 T3
What is the upshot of the Hayekian cascade model? Earlier, we demonstrated that
a population of agents defined by a certain distribution of epistemic standards—
under certain conditions—is persistently unable to identify its hidden champions.
What the Hayekian Cascade Model demonstrates is that the same population of
agents under the same conditions will in principle be able to find its hidden
champion, if each epistemic faction is permitted to act on their evaluative standards
and learn from the experience of others.18
To put it differently, the Hayekian cascade model suggests that if we want to
make full use of the reasonable and diverse perspectives inherent to democracy, we
need not only provide citizens with a robust right to free speech, but also enough
space to act on their epistemic standards.19 Against this background, I think, we
18
Hayek, I want to suggest, thought that this argument is applicable across various domains. As
such the argument might be used to argue on epistemic grounds for a robust freedom to act on
one’s reasonable epistemic standards across various domains in which we hope to achieve
epistemic progress. That Hayek was aware of the general applicability of these results comes
out when he approvingly cites W. A. Lewis, who writes: “These innovators are always in a
minority. New ideas are first put into practice by one or two or very few persons, whether they
be new ideas in technology, or new forms of organization, new commodities, or other novelties.
These ideas may be accepted rapidly by the rest of the population. More probably they are
received with skepticism and unbelief, and make their way only very slowly at first if at all.
After a while the new ideas are seen to be successful, and are then accepted by increasing
numbers” (Hayek 2010, In fn9 on p. 371). However, I cannot delve deeper into the question of
the applicability of the cascade model here.
19
Of course, with regard to public policy epistemic progress is only one value that needs to be
counterbalanced against a host of other considerations. For reasons of space, I cannot go into
the question of how to conceptualize this trade-off. Moreover, I also need to postpone any
discussion of how to best implement a Hayekian inspired version of epistemic democracy
(XXX 2019).
20
now come to understand much more clearly what Hayek meant in his concluding
remarks to “Freedom, Reason, and Tradition”:
“None of the conclusions are arguments against the use of [democratic, the
present author] reason but only arguments against such uses as require any
exclusive and coercive powers of government; not arguments against
experimentation as such, but arguments against all exclusive, monopolistic
power to experiment in a particular field—power which brooks no alternative
and is in its essence based on a claim to the possession of superior wisdom—
and against the consequent right to preclude the emergence of better solutions
than the ones to which those in power have committed themselves.” (Hayek
1958, 242).
Let me finish this section by situating Hayek’s argument both in a historical and
contemporary context. I take Hayek’s insight that to make full use of diversity
innovations need to have space to prove themselves, as a genuinely new
contribution to the field of political philosophy. John Stuart Mill and Wilhelm von
Humboldt, both famous for their epistemic approach to questions of political
philosophy—like their modern counterparts—have mainly focused their
intellectual energies on presenting an epistemic defence of freedom of speech. It is
true, however, that both thinkers at least en passant touched upon the epistemic
value of freedom of action. John Stuart Mill even explicitly discusses the value of
experiments in living. Nevertheless, the epistemic value of letting people act on
their epistemic standards, and the issue of evidence creation, remains understudied
in both Mill and Humboldt.
Another way to contextualize the Hayekian Cascade Model is to contrast it with
the Diversity Trumps Ability (DTA) theorem introduced to epistemic democracy
primarily by Hélène Landemore (2003, cf. Landemore and Page 2015, Anderson
2006.) The theorem states that under certain conditions, a group of randomly
selected diverse problem solvers beats a group of experts in problem-solving tasks.
From the Hayekian point laid out in this paper, the crucial question is what exactly
the DTA is supposed to model. Usually, the DTA is meant, at least in philosophy
circles, to model inclusive deliberation. When we apply the DTA to deliberative
problem-solving tasks, the problem is that the very distinction between knowledge
exploitation and knowledge generation, on which the present essay builds on, tends
to get leveled. Take for instance a case that Landemore discusses in a paper co-
authored with Page (2015, 235). In the case discussed, a deliberative body of locals
is attempting to solve “a recurrent safety issue on a dark bridge that separates it
from the city’s downtown.” Initially everybody in the community has some sub-
optimal proposal for solving the issue, but by building on each other’s proposals,
they eventually find the best one. The whole model, however, only functions on the
premise that there is no disagreement about the objective value of each of these
21
proposals. Landemore and Page (2015, 234) are cognizant of that fact and thus
introduce the premise of an oracle: “a machine, person, or internal intuition that can
reveal the correct ranking of any proposed solutions” to their thought experiment.
Now, by introducing the concept of an oracle, what they are essentially doing is to
assume that every participant by definition is already in possession of the best
evidence for her very proposal (even if she just came up with the proposal a second
ago). The model thus assumes away the very epistemic issues that this essay has
grappled with.20
20
In principle, however, we could also use the DTA to model epistemic progress of federalized,
democratic decision making over time. In such a model, each agent would represent a political
community that implements policies in accordance with their epistemic standard. After each
round of implementing policies, the agents would meet and discuss the newly gathered
evidence. If one community is successful in improving the local optimum of the status quo
ante, then everybody else implements the new local optimum and continues their search for the
global optimum from there. In principle thus, it is possible to capture the distinction of evidence
exploitation and generation within the model spearheaded by Scott Page (2006). The Hayekian
critique of epistemic democracy then should lead us to rethink how to best employ the Scott
Page model for the purposes of pushing forward the concept of epistemic democracy.
22
coercion by other governments, pure imitation of successful first movers and
competitive pressures (Shipan and Volden 2008).
That political units take into account the experience of similar or even quite
different political units to aid them in their decision process for or against a certain
policy should be quite uncontroversial to begin with, since learning from the
experience of others seems to be a simple demand of rationality.
6. Conclusion
In this essay, we set out with a simple question: what—if anything—is Hayek’s
contribution to epistemic democracy? Posing this question was in part motivated
by the insistence of the foremost scholars in the discipline that Hayek has little to
contribute to the debate.
Epistemic democrats hold that part of the reason why we should value
democracy (and rational deliberation) is that it is an unparalleled mechanism for
making use of diverse perspectives. Against this view, Hayek mounted a challenge
that—to my knowledge—has so far gone unnoticed. The goal of this essay was to
reconstruct Hayek’s core argument.
Hayek, I argued, can be understood to make two arguments. His first argument
is challenging the claim by epistemic democrats that rational deliberation (i) at
every instance or at least (ii) over time will converge on the objectively correct
policy proposal. Hayek argues that whether democratic institutions can live up to
the expectation of epistemic democrats depends on the epistemic basic structure of
democracy. Moreover, he argues, that the way we usually conceive of the core
institutions of democracy (majoritarian rule and civil liberties, among them a right
to free speech, academic freedom and publically-funded research facilities) are a
necessary but not a sufficient condition for ensuring that democracies can over time
make true of the epistemic democrat’s promise.
His second argument ties in with the first one. Epistemic democrats claim that
the principal way to make use of a diversity of reasonable perspectives is inclusive
deliberation over time. Hayek agrees that inclusive deliberation plays an important
role for making use of a diversity of perspectives, but he contends that to make full
use of a diversity of perspectives for the purposes of political decision making,
society not only has to provide citizens with a robust right to free speech, but also
with a robust right to act on their epistemic standards. In the last section of this
essay, I offered a simple cascade model to explain Hayek’s contention.
23
It is my hope that this essay was able to demonstrate that Hayekian thought
presents not only a good challenge to epistemic democracy, but that epistemic
democracy as a paradigm can grow by facing up to Hayek’s challenge and by
incorporating his insights into the institutional preconditions for making pluralism
productive.
References
24
Lu, L., Y. C. Yuan, and P. L. McLeod. 2011. “Twenty-Five Years of Hidden Profiles in
Group Decision Making: A Meta-Analysis.” Personality and Social Psychology
Review 16 (1): 54–75.
Machiavelli, Niccolo. 2003. The Prince. Cambridge: Cambridge University Press.
Mantzavinos, Chrysostomos. 2013. “Explanatory Games.” The Journal of Philosophy, 110
(11): 606–632.
Mantzavinos, Chrysostomos. 2016. Explanatory Pluralism. Cambridge: Cambridge
University Press.
McCarthy, Thomas. 1998. “Legitimacy and Diversity: Dialectical Reflections on
Analytical Distinctions.” In: Habermas on law and democracy: Critical
exchanges. Michel Rosenfeld and Andrew Arato (eds). Berkeley: University of
California Press, 381–452.
Misak, Cheryl, and Robert B. Talisse. 2014. “Debate: Pragmatist Epistemology and
Democratic Theory: A Reply to Eric MacGilvray.” Journal of Political Philosophy
22 (3): 366–76.
Mitchell, J. L. and Elizabeth Petray. 2016. “The march toward marriage equality:
Reexamining the diffusion of same-sex marriage among states.” Public Policy and
Administration 31(4): 283-302.
Mueller, Julian F. 2018. “Epistemic democracy: Beyond knowledge exploitation.”
Philosophical Studies 175 (5): 1267–1288.
Page, Scott E. 2007. The difference: How the power of diversity creates better groups,
firms, schools, and societies. Princeton: Princeton University Press.
Peter, Fabienne. 2013. “Epistemic Foundations of Political Liberalism.” Journal of Moral
Philosophy 10 (5): 598-620.
Popper, Karl R. (2002): The Poverty of Historicism. Routledge, London.
Putnam, Hilary. 1992. “A Reconsideration of Deweyan Democracy.” In Renewing
philosophy, ed. Hilary Putnam. Cambridge, Mass., London: Harvard University
Press, 180–200.
Rawls, John. 2005. Political Liberalism. New York: Columbia University Press.
Rogers, E. M. 2003. Diffusion of Innovations, New York: Free Press.
Sager, Lawrence G. 2005. “Cool Federalism and the Life-Cycle of Moral Progress.”
William & Mary Law Review 46(4): 1385-1398.
Shipan, C. R. and C. Volden. 2008. “The Mechanisms of Policy Diffusion.” American
Journal of Political Science 52 (2): 840-857.
Somin, Ilya. 2013. Democracy and political ignorance: Why smaller government is
smarter. Stanford, California: Stanford Univ. Press.
Sunstein, Cass R. 2006. “Deliberating Groups versus Prediction Markets (or Hayek's
Challenge to Habermas).” Episteme 3 (03): 192–213.
Zollman, Kevin J. S. 2010. “The Epistemic Benefit of Transient Diversity.” Erkenntnis 72
(1): 17–35.
25