1. Hypothetical Publicity

1.1 Kant’s Hypothetical Publicity Test

Kant’s hypothetical publicity test appears in the second
appendix to his Zum ewigen Frieden (Perpetual Peace)
(1795 [1923:
He writes:

All actions relating to the right of other human beings are wrong if
their maxim is incompatible with

There are at least three interesting issues here. First, what is the
justification for such a publicity test? Is there any reason to
suspect any connection between the demands of publicity and the
rightness of a position? Second, which maxims of action and, more
generally, which moral theories are likely not to pass such a
publicity test? Third, what does this test entail with regard to the
need for actual publicity?

Some interpretative issues can quickly be disposed of. First, the
Kantian test applies specifically to political maxims and is
the equivalent in that sphere of the more general categorical
This is why Kant refers to it as the “transcendental principle
of the publicity of public law” (1795 [1923: 382]). Second, this
Kantian test is a merely negative one: passing the test is
only a necessary condition for a maxim’s rightness. It is
therefore not sufficient to guarantee that all maxims that pass this
test are right. This is suggested both by the expression “wrong
if” and by Kant’s explicit statement: “this
principle is, furthermore, only negative, i.e., it only serves for the
recognition of what is not just to
Such a reading entails a difference with the categorical imperative
insofar as the latter is a necessary and sufficient condition
for permissibility. Third, it is also clear from the principle that
neither an action itself, nor the actual motives underlying such an
action (Davis 1991: 418) must be capable of being successfully
disclosed in order to guarantee that they are not wrong. The test
applies only to the maxim of such an action, which indicates
one respect in which such a hypothetical test might leave room for
actual secrecy. This claim is true irrespective of whether David Luban
is correct in suggesting that:

the best way to make sure that officials formulate policies that
could withstand publicity is by increasing the likelihood
that policies will withstand publicity (Luban 1996:

Now, what does Kant mean when he uses the word
“publicity”? Were we to face an actual public, we
could give it a more or less demanding interpretation, ranging from
merely general knowledge (each of us actually knows or is able to
know) or mutual knowledge (each of us knows, and knows that the others
know), to critical debate where what we know is being actively
discussed (Luban 1996: 169–172). The latter form of publicity,
while seeming much richer and appealing, probably does not fit the
It could itself be understood in two ways. Either, publicity would be
taken as a synonym of debateability, i.e., the fact that the
maxim be such as to be deemed appropriate for public discussion, in
which case the test would exclude very little. Alternatively, were it
taken as referring to the ability of a maxim to pass the filter of
actual public debate and be accepted by its participants, the
test would then have a very contingent outcome, depending on the real
public we are facing.

We can try to ascertain whether publicity should be understood as
general knowledge, mutual knowledge, or publicity as the outcome of
However, in the context of understanding the nature of the Kantian
test, what matters much more is that Kant is referring to an ideal
and rational
public (Davis 1992:
The test is thus hypothetical, not only because it does not
necessarily require actual publicity (as we shall see), but
also because any outcome of actual publicity must inevitably remain at
best a rough approximation of what the test would lead us to with an

1.2 Sidgwick on Covert Utilitarianism

As we shall see in
Section 2,
classical utilitarians, such as Jeremy Bentham and especially Mill,
were generally sympathetic to the idea of publicity. Still, it is also
among utilitarians that we find one of the strongest principled
defenses of secrecy. Sidgwick defends the idea that it may be
necessary for utilitarianism to go under cover, and that this need not
be unjust. In his Methods of Ethics (1874), he writes:

[…] on Utilitarian principles, it may be right to do and
privately recommend, under certain circumstances, what it would not be
right to advocate openly; it may be right to teach openly to one set
of persons what it would be wrong to teach to others; it may be
conceivably right to do, if it can be done with comparative secrecy,
what it would be wrong to do in the face of the world; and even, if
perfect secrecy can be reasonably expected, what it would be wrong to
recommend by private advice or example […]. Thus the
Utilitarian conclusion, carefully stated, would seem to be this; that
the opinion that secrecy may render an action right which would not
otherwise be so should itself be kept comparatively secret; and
similarly it seems expedient that the doctrine that esoteric morality
is expedient should itself be kept esoteric. Or if this concealment be
difficult to maintain, it may be desirable that Common Sense should
repudiate the doctrines which it is expedient to confine to an
enlightened few. And thus a Utilitarian may reasonably desire, on
Utilitarian principles, that some of his conclusions should be
rejected by mankind generally; or even that the vulgar should keep
aloof from his system as a whole, in so far as the inevitable
indefiniteness and complexity of its calculations render it likely to
lead to bad results in their hands. (Sidgwick 1874 [1893:

As in the Kantian case, Sidgwick not only deals with secret actions,
but with the secrecy of doctrines underlying such actions. Two of his
points are worth stressing. First, the relative confidentiality of a
given practice or—more importantly for us—of its maxim may
not only be acceptable, but may even be a necessary condition
for the moral acceptability of such a practice, which is then the
reverse of what Kant defends. Indeed, utilitarians may consider some
practices acceptable on the condition that they remain marginal. And
keeping such practices confidential may be required for them not to
spread. Secrecy would thus not only be allowed in this case.
It would also be required. Second, there is a doctrine of
meta-secrecy at play as well, as Sidgwick’s argument
applies on top of actions to both maxims and meta-maxims. For him, the
very idea that utilitarianism may allow and even require such covert
actions should itself be kept relatively secret. As Luban puts it,

not only is it best if utilitarianism is not widely believed, but it
is best if the very fact that the belief in utilitarianism has been
suppressed by the “enlightened few” is not widely
believed. This is no longer Government House utilitarianism. It is
conspiratorial utilitarianism (Luban 1996: 167).

1.3 Sidgwick’s and Kant’s Examples

1.3.1 Sidgwick on Celibacy and Lying

With the aim of identifying the exact implications of Kant’s and
Sidgwick’s views, let us look more closely at the examples they
provide in the course of their discussions. We shall begin with
Sidgwick, who contrasts two cases: lying and
The argument unfolds as follows. First, some types of behavior are
morally acceptable only if they remain (very) restricted in
If lying and celibacy were to become widely practiced, this would
become a problem for utilitarians, according to Sidgwick. However, as
long as they remain marginally practiced, to an extent that is
justified on utilitarian grounds, they should be accepted. Second, the
question arises as to whether we should recognize all this publicly.
Albeit not raising any difficulty in the case of
publicly abandoning an absolute prohibition on lying might not be
acceptable for utilitarians. Sidgwick writes:

… assuming that general unveracity and general celibacy would
both be evils of the worst kind, we may still all regard it as
legitimate for men in general to remain celibate if they like, on
account of the strength of the natural sentiments prompting to
marriage, because the existence of these sentiments in ordinary human
beings is not affected by the universal recognition of the legitimacy
of celibacy: but we cannot similarly all regard it as legitimate for
men to tell lies if they like, however strong the actually existing
sentiment against lying may be, because as soon as this legitimacy is
generally recognized the sentiment must be expected to decay and
vanish (Sidgwick 1874 [1893: 486]).

It is worth emphasizing that Sidgwick clearly states that the need for
secrecy in the case of lying arises only because we are dealing with a
general public that is not constituted of only enlightened
utilitarians. For Sidgwick considers that

if therefore we were all enlightened Utilitarians, it would be
impossible for any one to justify himself in making false statements
while admitting it to be inexpedient for persons similarly conditioned
to make them; as he would have no ground for believing that persons
similarly conditioned would act differently from himself (Sidgwick
1874 [1893: 486–487]; Piper 1978: 190f).

However, once we assume that we do not live in such an ideal
world, populated by only enlightened utilitarians, we have reason to
keep our complex moral rules secret in order to avoid
counterproductive consequences from an utilitarian perspective. Such
consequences may flow from various sources, such as a lack of time
(e.g., people spending too much time finding out what the fair action
would be), misunderstandings possibly due to inadequate cognitive
abilities (e.g., construing too broadly the class of cases in which
lying would be acceptable from a utilitarian point of view, leading to
a spread of lying) or from distrust towards utilitarianism (e.g.,
principles having a larger number of exceptions generating more doubts
about their cogency than elegant and exceptionless ones).

Notice moreover the difference between the Sidgwickian argument, as
illustrated by these two examples, and another utilitarian argument
that would justify “self-effacement”, i.e., the idea that
to remain able to reach its goals, utilitarianism should go totally
undercover. In other words, even enlightened utilitarians should not
be motivated by their doctrine. The argument is as follows: being
motivated by a utilitarian goal may impede our ability
actually to reach this very goal. For instance, truly
believing in the value of an artistic project independently of its
ability to bring about happiness is more likely to maximize welfare
than being motivated towards this project out of a commitment to
utilitarianism. The upshot, then, is that a society committed to
utilitarian principles may turn out to be a less happy one than a
society relying on a given set of commonsense principles. If this is
true, we would not only have an argument, for “Government House
utilitarianism” (i.e., for utilitarian principles being relied
on only by the enlightened few), or for its stronger
“meta-secrecy” version (of which a few enlightened would
still be aware). Rather, we would have a case for something even more
radical: the full self-effacement of utilitarianism for utilitarian
reasons, i.e., even enlightened utilitarians would have to forget
about it if they are really committed

In short, Sidgwick’s argument explicitly refers to a
non-ideal public (“society as actually
constituted”, “mankind in general” or “the
vulgar”, as opposed to “an ideal community of enlightened
Utilitarians”). And his view then has to do with a logic of
confinement, which is necessary to preserve the practice’s
acceptability. Because of this, several of the reasons for secrecy
that underlie Sidgwick’s discussion of the lying case would not
apply if we were all enlightened utilitarians. However, it is worth
noting that, besides the “self-effacement” argument,
Adrian Piper has argued that there may be utilitarian grounds for
secrecy in an ideal community of enlightened utilitarians as
well. For knowing that in such a society, e.g., promises
might not necessarily be kept—keeping them depending in
each case on the outcome of a utilitarian calculus—will cause
significant coordination problems due to lack of predictability of
other people’s action. In the absence of secrecy, such an ideal
society would then become less efficient (Piper 1978:

1.3.2 Kant on Rebellion and Secession

Interestingly, while Kant’s test refers to an ideal public, the
examples he provides us with have to do—as in Sidgwick’s
discussion—with a non-ideal public. As we shall see,
however, the mechanisms on which Kant’s argument rest differ
from those on which Sidgwick’s is based. Two cases are
envisaged: rebellion and secession. Let us begin with the former. Kant
asks: “Is rebellion a legitimate means for a people to employ in
throwing off the yoke of an alleged tyrant (non titulo, sed
exercitio talis
)?” (1795 [1923: 382]). The publicity test
is used here in two ways. First, Kant claims that the open
acknowledgment of the maxim of rebellion “would make its own
purpose impossible” (1795 [1923: 382]). The argument is the
following: a person cannot be said to be the chief if the people may
sometimes employ force against him (chief-with-absolute-power
assumption). The establishment of a state requires a chief in this
sense (no-state-without-chief assumption). Therefore, publicizing the
maxim of rebellion would make the establishment of a state impossible.
Second, the chief himself may publish the maxim according to which
rebellion will be sanctioned with death, “for when he knows he
possesses irresistible power […] he need not fear vitiating his
own purpose by publishing his maxims” (1795 [1923: 383]). And it
is because of such irresistibility that publishing this maxim will not
be self-frustrating. Again, Kant does not say whether in a non-ideal
situation, in which nobody is in a position to have absolute power,
the fact that publishing the same maxim would frustrate its very
purpose would also mean that this is an unjust maxim. And what is also
worth stressing is that such purpose frustration does not imply
approval by an ideal and rational public. It merely derives from power
relations. Hence, were we to adopt the same line of argument, we could
conclude that the stronger a despotic power, the larger the spectrum
of maxims that may pass the publicity test, a conclusion that Kant
implicitly endorses in the case of maxims of international

The rebellion example illustrates the difficulty involved in moving
from the hypothetical test to real life situations. What is at stake
is our understanding of what “compatibility” could
plausibly mean once we move from an ideal to a non-ideal public. The
tension in Kant’s text is the following. As we keep reading the
general part of his argument, it looks as if reference is made to the
universal acceptability of a maxim in the eyes of an ideal and
rational public. The rebellion as well as the other examples he gives
however (1795 [1923: 383–384]) all rely on a
“self-frustration” (or counter-productivity) mechanism. Of
course, the latter may have a direct connection with what it may be
rational for the beneficiaries and victims of a maxim to do. However,
Kant remains unclear about the connection between such preoccupations
and the just or unjust nature of a maxim.

Let us then take the secession example:

“If a smaller State is situated as to break up the territory of
a larger one, and continuous territory is necessary to the
preservation of the larger, is the latter not justified in subjugating
the smaller and incorporating it?”. We easily see that the
greater power cannot afford to let this maxim become known; otherwise
the smaller states would very early unite, or other powers would
dispute the prey, and thus publicity would render this maxim
impracticable. This is a sign that it is illegitimate. (1795 [1923:

Again, Kant refers here to the risk that defense strategies of the
potential losers of a maxim’s implementation may render the
latter less efficacious, if not totally inefficacious. There are at
least two difficulties with this. First, in a real life context (e.g.,
Cuba or Chechnya type) in which the large state knows that other small
or large states are unlikely to intervene in such situations,
publicity would not make the maxim impracticable. Hence, the
maxim could in principle still be fair if we were to follow
the Kantian logic. Second (and conversely), let us take a domestic
analogy involving a state fighting a local mafia. Were the state to
announce in advance that it will run a special anti-mafia operation on
a given day, any benefits from a surprise intervention would vanish.
That is, a maxim stated precisely enough to include the date of the
police intervention could not be publicized. All the local Mafiosi
would seek to hide themselves on that very day, hence making the
intervention useless. This should not suffice to convince us
however—even if we were ideal and rational observers—that
the planned police intervention (and the maxim on which it would be
grounded) is unjust.

Both Sidgwick’s and Kant’s examples rely on non-ideal
publics. Admittedly they do so in different manners. Whereas
Kant’s ideal public apparently sticks to the role of a judge,
Sidgwick’s ideal community of utilitarians is taken as a set of
actors whose understanding of the rules is likely to affect their
actual behavior. However, the reasons justifying secrecy in the
examples each of them gives are of a very different nature. For
Sidgwick’s illustrations do not necessarily involve power
relation problems. Instead, they highlight motivational and cognitive
dimensions. For our purposes, what matters is that, whereas the
examples given by Sidgwick properly illustrate the logic of his
argument, those provided by Kant do not enlighten us much further, for
the idea of purpose frustration (or “self-frustration”)
mobilized in the rebellion and secession cases is not applicable once
we shift to an ideal public. Therefore, we shall now rely on four
examples distinct from those mentioned by Kant in order to ascertain
the meaning and implications that could be attached to the Kantian

1.4 Does the Kantian Test Prohibit Actual Secrecy?

1.4.1 Sidgwick’s Meta-Secrecy Doctrine

Let us now examine whether or not Sidgwick’s covert
utilitarianism is able to pass Kant’s publicity test. As we have
seen, this doctrine states that we should keep secret the very idea
that some principles and actions should remain secret. This means
that, even at a very high level of generality, there is no way we
could submit any maxim for assessment to the ideal public, unless such
a maxim makes no direct reference to the problem of publicity. But
what about the following maxim: “any means should be deemed
appropriate if they serve the utilitarian ideal”? Utilitarians
claim that an ideal public should at least consider utilitarianism a
plausible doctrine. And it is therefore unlikely that the Kantian test
will lead to a “necessary and unanimous” rejection of this
proposed higher-order formulation of the maxim. Hence the test may be
passed successfully, without any discussion on the issue of publicity.
Of course, one may find such a maxim totally unjust for other reasons.
But Kant himself does not seem to claim that his test should be seen
as a sufficient test for justice anyway.

1.4.2 Glomar Responses

Our second example comes from actual Freedom of Information (FoI)
regimes. That a public authority may have to justify each of its
significant decisions will seem largely uncontroversial. In principle,
this should also hold each time a public authority denies access to a
document requested by a citizen on the basis of his FoI right. The
citizen should know the grounds that access is being denied to the
document he wants to double-check, e.g., whether this is because of
privacy, trade secrets, or state security. However, there are cases in
which disclosing the exact reason why access to a document is denied
unavoidably entails the disclosure of the document’s very
content. One could deny access with no overt justification, implying
that disclosing the precise ground for denial would be
equivalent to disclosing the document’s very content. Along the
same lines, an authority could even refuse to deny or confirm the
very existence
of the requested document, which is known in US
law as a Glomar

Again, one may well propose a higher-order principle accounting for a
response of this kind:

whenever access to a document may be denied on a legitimate basis and
disclosing either the basis for denial or even the very existence of
the document would jeopardize the pursuit of this very goal, a
justification for denial or even the confirmation of the
document’s (non)-existence may be withheld.

So long as the set of such legitimate bases for denial is publicly
discussed and procedural checks are in place, there is no reason to
believe that publicizing such a maxim will be self-frustrating in the
presence of a real public, nor that such a maxim should be rejected by
an ideal public. Admittedly, the Glomar doctrine embodies some extent
of “meta-secrecy”. However, it does not go as far as a
meta-secrecy of the kind that Sidgwick favors. For the very
possibility that a state may sometimes act on secret grounds is not
publicly denied, nor are the grounds for justifiable Glomar responses
necessarily left outside the sphere of public

1.4.3 The Judiciary’s True Goal

For our third example, let us suppose that the true goal of the
judiciary were to promote peace and stability rather than justice.
However, let us assume as well that letting people believe that the
genuine goal of such institutions is justice would best guarantee
peace, for decision-acceptance would be enhanced if such decisions
were perceived as just (Luban 1996: 160). In such a world, there is no
question that peace would better be pursued by keeping it secret as
the true goal of judiciary. The maxim may be:

act in a way such that if peace is best pursued by letting people
think (wrongly) that it is in fact justice that is being pursued, you
should let people think so.

Publicizing such a maxim may be “self-frustrating” and
would therefore not pass the Kantian test.

However, besides asking ourselves whether such self-frustratingness is
what accounts for the possibly unjust nature of the maxim, we may
provide a higher order formulation of the maxim:

Act in a way such that if one legitimate goal is best pursued by
letting people think (wrongly) that it is in fact another legitimate
goal that is being pursued, you should let people think so.

This higher-order maxim’s publication is not self-frustrating,
even in a case in which we assume that the legitimacy of each goal has
been examined in detail, along with the legitimacy of other goals. Of
course, this does not end the discussion since, in case we were to
grant priority to justice over peace, the maxim’s implementation
could still lead to injustice, which the Kantian test does not

1.4.4 The Optimal Confidentiality of FoI Laws

Let us finally move to an example that deals again with real FoI
regimes. One of the surprises of such regimes is that, as a matter of
fact, their very existence remains rather confidential in many
countries. Citizens do not exercise their right of access to
state-detained documents as much as one might expect. This could be
regarded as unfortunate. One may as well consider it necessary,
however, at least if we find meaningful the idea of an optimal
level of confidentiality
of FoI regimes. This optimal level would
consist in some active citizens being maximally aware of its
existence, and all civil servants being as little aware of it as
possible. For if (non-ideal) civil servants know that they are
potentially being watched, they may admittedly try to change their
behavior for the better. However, they may simultaneously want to hide
even further the more problematic aspects of their behavior (e.g.,
through shifting from written to oral discussions). In other words, in
order for the fundamental FoI right to be maximally effective, its
existence and exercise may have to remain relatively covert, perhaps
unequally known.

What should an ideal public think about the following maxim then:
“In order for a legitimate goal to be pursued, while the goal
itself would be fully public (here: the accountability of public
authorities), while the social expectations of citizens towards civil
servants would be fully transparent, some of the means enabling such
citizens to ensure the accountability of public authorities may remain
covert”? Not each and every covert means is acceptable. However,
the one consisting in a right to ask for a copy of administrative
documents may very well be so. Of course, publicizing a lower order
version of this maxim may be self-frustrating, hence incapable of
passing the Kantian test. But again, this is not sufficient to discard
this lower order maxim as unjust, nor does it mean that a higher-order
translation of it (as suggested above) must be unable to pass the
Kantian test.

These four examples leave us with two important open questions
regarding the Kantian test. First, how should we assess the
test’s importance if the level of generality at which the maxim
of an action should be phrased cannot be normatively fixed? As we have
illustrated above, the lower the level of generality at which a maxim
is phrased, the higher the risk that its publicity would be
self-frustrating, and vice versa (Luban 1996: 189–191). No
doubt, the test would be over-exclusive if we were forced to
phrase our maxims at a very low level of generality. This is clear
when we think about keeping secret the date of a police intervention
to trap criminals or the location of a stock of weapons in times of
peace. Kant himself never suggests that all secret actions should be
ruled out by his test. The problem is that, as long as we do not have
any indication as to the appropriate level of generality at which an
action’s maxim should be tested, the opposite risk of
under-exclusiveness becomes hard to avoid as our four
examples above illustrate.

Second, it is not entirely clear why self-frustratingness (or
purpose-frustratingness) is an indicator of injustice. In other words,
whenever the Kantian test bites, why would this tell us anything about
the just or unjust nature of a maxim and the actions that fall under
Answering this question is even more crucial for those
who—contrary to us—view the publicity test as a necessary
and sufficient condition for permissibility. What we need
then is an account of the “intrinsic connection” between
self-frustratingness and injustice—in the same way as one may
ask why, for contractualists, there should be a significant connection
between justice and the hypothetical agreements of fairly-situated
individuals. Such an account is not provided by Kant, who claims about
the principle that “like an axiom, it is indemonstrably
certain” (1795 [1923: 382]). One sound suggestion is
Luban’s (1996: 191). Imagine that making the maxim of my action
public would frustrate my ability to reach this maxim’s very
goal. This can be taken as a sign of the injustice of this maxim
only if the frustration-generating mechanism is at least in
part based itself on morally justifiable grounds. But if publicizing a
doctrine generates distrust from a real audience merely because of
people’s attraction for simplicity or non-demandingness, the
frustration flowing from such a distrust might hardly be considered
relevant to the assessment of the maxim’s (in)justice.

2. Actual Publicity

Actual publicity can be looked at through the prism of hypothetical
publicity. However, philosophers approach the issue from other angles
as well. Their arguments have mobilized both empirical assumptions and
philosophical claims regarding the purpose of voting and political
representation, the importance of participation and deliberation, etc.
Here, we examine more closely arguments in favor and against actual
publicity in the political arena. In order to do so, we will restrict
ourselves to two modes of political action (voting and deliberation)
and two types of actors (electors and representatives). In so doing,
we consider two types of relationships: horizontal ones
(among voters or among representatives) and vertical ones
(between voters and representatives). Notice already that the view one
favors regarding the function attached to one’s vote as a member
of the electorate and to one’s activity as a representative will
have an impact on the reasons why publicity may matter and the extent
to which it does.

Before engaging in this philosophical discussion, let us clarify what
the arguments developed below may show for one significant practical
evolution, namely the development of FoI regimes in democratic
countries during the second half of the 20th
Such regimes have been set up with the view of allowing citizens to
participate more actively in public affairs, enabling them to have
access to the details of most administrative documents in due
Hence, not only MPs but civil servants become more effectively
accountable. This means that citizens are not forced simply to accept
package deals from their representatives at the time of elections.
They can question the practicalities of implementation at any

Not only does the enactment of FoI regimes provide opportunities for
multiple interactions between active citizens and state authorities,
they also force us to weigh the relative importance of open government
vis-a-vis other fundamental rights. Specific grounds for denying
access to administrative documents include the protection of privacy,
property rights (e.g., trade secrets), easing the gathering of
information from third parties (e.g., executive
or medical secrecy), paternalistic justifications (e.g., in the case
of technical documents that could be misinterpreted), state interests
(e.g., in defense and terrorism matters or international
negotiations), and economic efficiency (e.g., in public procurement,
the content of the offer made by competitors being disclosed only
after the final decision). Sometimes, such grounds for secrecy will
justify strict denial of access. On other occasions, the disclosure of
information will be deferred (e.g., through declassification
procedures) or be selective (e.g., to some types of actors only). This
variety of potential justifications of secrecy and the diverse ways of
accommodating the conflicting values constitute a largely unexplored
terrain for applied political philosophy. Here, we will focus only on
general reasons to promote or restrict actual publicity.

2.1 Voting

2.1.1 Mill’s (1861) Case Against Secret Ballot

Let us look at voting first. We may begin with the view according to
which the electoral process aims only at aggregating people’s
individual preferences regarding the type of decisions they consider
in their own interest (rather than in the public interest). Under such
a view, there is no strong case for making sure that my fellow citizen
be able to know what my actual preferences are. This is especially
significant when it comes to direct democracy (e.g., a referendum).
However, in democratic systems involving representation, there may be
weighty reasons for representatives to know which type of electors
they are representing, and hence to favor open ballots. This implies
that, even under an “aggregation of private interests”
view of democracy, there may be a case for open ballots. This case has
nothing to do with the need for fellow voters to know about their
neighbor’s actual preferences. Only their representatives need
to know about them.

Let us now move to a different conception of democracy, advocated by

In any political election, even by universal suffrage (and still more
obviously in the case of a restricted suffrage), the voter is under an
absolute moral obligation to consider the interest of the public, not
his private advantage, and give his vote, to the best of his judgment,
exactly as he would be bound to do if he were the sole voter, and the
election depended upon him alone. This being admitted, it is at least
a prima facie consequence that the duty of voting, like any
other public duty, should be performed under the eye and criticism of
the public; every one of whom has not only an interest in its
performance, but a good title to consider himself wronged if it is
performed otherwise than honestly and carefully. Undoubtedly neither
this nor any other maxim of political morality is absolutely
inviolable; it may be overruled by still more cogent considerations.
But its weight is such that the cases which admit of a departure from
it must be of a strikingly exceptional character (1861 [1991:

Mill later discusses whether such reasons to depart from publicity are
forceful, which is an issue to which we shall return. However, what
matters here is that, once we consider that voters are supposed to
vote according to what they believe would promote the public interest,
rather than their own private
the case for publicity is reversed. On the one hand, if we can
associate to such a conception of voting one of representation
according to which representatives are supposed to defend the public
interest, rather than the class interests of their identifiable
the case for vertical publicity weakens. For knowing who their
electors actually are may induce MPs to stick to a defense of their
actual electors rather than to promote the public interest at large.
On the other hand, the very act of taking part in an election acquires
a non-private dimension, such that we should consider ourselves
accountable towards our fellow citizens (horizontal accountability).
And the claim that we should regard ourselves as horizontally
accountable towards those who may be affected by the consequences of
our choices clearly supports the case against secret ballots. For
Mill, this is a very significant reason for open ballots. In fact, he
even stresses that, as electors, we should see ourselves accountable
not only towards fellow electors, but also to non-electors
whenever the franchise is not as universal as it could be (which
generates the risk of electors voting only to defend their class
interests as
The example from Mill’s times is the one of women, who were
excluded from “manhood suffrage”. Present day
illustrations could refer to non-national residents excluded from the
right to vote or to those below the age of

2.1.2 Why Worry About Open Vote?

Mill was not the first to write about vote secrecy. Authors such as
Cicero (III.15),
and Alexis de
also expressed their views on the matter. And Mill’s
intervention is in fact part of a lively debate that had been going on
in England for several
He writes:

It may, unquestionably, be the fact that if we attempt, by publicity,
to make the voter responsible to the public for his vote, he will
practically be made responsible for it to some powerful individual,
whose interest is more opposed to the general interest of the
community than that of the voter himself would be if, by the shield of
secrecy, he were released from responsibility altogether. When this is
the condition, in a high degree, of a large proportion of the voters,
the ballot may be the smaller evil. When the voters are slaves,
anything may be tolerated which enables them to throw off the yoke.
The strongest case for the ballot is when the mischievous power of the
Few over the Many is increasing […].

But in the more advanced states of modern Europe, and especially in
this country, the power of coercing voters has declined and is
declining; and bad voting is now less to be apprehended from the
influences to which the voter is subject at the hands of others than
from the sinister interests and discreditable feelings which belong to
himself, either individually or as a member of a class. To secure him
against the first, at the cost of removing all restraint from the
last, would be to exchange a smaller and a diminishing evil for a
greater and increasing one […] (Mill, 1861 [1991:

Geoffrey Brennan and Philip Pettit (1990) revived Mill’s
argument, contending that, while we should not assume that people will
generally tend to vote according to their private interest whenever
voting is secret, unveiling the vote would certainly encourage further
the practice of voting in accordance with what one considers to be the
common good (Brennan and Pettit 1990: 325–329). They argue as
well that the case for open votes is especially strong in pluralistic
societies involving large

Still, Mill’s lesser-evil type of justification for open votes
has been rejected in the past on at least three grounds. First, one
could question the very purpose that Mill ascribes to voting. Hence,
some of the protagonists of the 19th century British debate criticized
the view according to which voting should be regarded as a trust
rather than a
In contrast with Mill’s view, this entails that ballot secrecy
is not even a lesser evil. It is no evil at all. Second, one may
disagree with Mill at the factual level rather than at the normative
one. James Mill, his father, clearly agreed with his son on the
purpose of voting, referring explicitly on various occasions to the
idea of a trust. In his long essay “On the Ballot”, he
states that “The voter for a member of Parliament has a trust
placed in his hands, on the discharge of which the highest interests
of his country
However, James Mill considered—in opposition to his son’s
view—that the negative consequences attached to the secret
ballot were not only less weighty than its advantages, but devoid of
any weight at
This entails that for James Mill as well, secret voting was not even
a lesser evil reform. It did not involve any evil at all. Third, let
us consider Bentham’s position. Although he does not refer
explicitly to the trust-based view of voting, he considers that voters
should be influenced by tutelary motives. For him,

in judging whether a motive ought to be referred to the class of
seductive or tutelary motives, it is necessary to
examine whether, in the case in question, it tends […] to
favour the greatest or the smallest number (Bentham 1843 [1999:

After balancing the pros and cons of publicity, Bentham concludes:

The system of secrecy has therefore a useful tendency in those
circumstances in which publicity exposes the voter to the influence of
a particular interest opposed to the public interest. Secrecy is
therefore in general suitable in elections.

In contrast with James Mill, Bentham’s argument is of the
lesser-evil type. But contrary to J.S. Mill, it is a lesser-evil
argument in favor of secrecy rather than

Such developments may be dismissed as being of merely historical
interest. And indeed, except for Brennan and Pettit’s (1990)
work, there is little contemporary debate about the justifiability of
the secret ballot. Rather, the discussion has shifted its focus to
parliamentary debate and to administrative activity, such as FoIs.
Still, there are at least three lessons to be learnt from older
debates about the ballot, over and above the fact that vote secrecy is
not unambiguously costless. First, already at that time, there were
attempts at envisaging simultaneously the questions of vote secrecy in
the case of electors and in the case of
The interesting fact is that we do not today have a clear theory
telling us why we should have secret ballots for electors and open
votes in plenary parliamentary sessions. Simply referring to the fact
that representatives are accountable to their electors (vertical
accountability) in a sense in which electors are not to their fellow
citizens (horizontal accountability) is insufficient in itself.
Ceteris paribus, the importance of horizontal accountability,
insofar as it emphasizes the need to justify our decisions to those
who may endure their consequences, could suffice to justify openness.
Moreover, it is probably also insufficient to claim that the fact that
representatives may lose their job as a result of electoral discontent
is not as problematic as the fact that electors could lose their job
if their employer were to know for whom they voted. This does not mean
that differences are absent. Ceteris paribus, a
representative has more power than a citizen, for
What we need to ask ourselves is whether this is sufficient to
justify treating the two categories so differently with regard to vote

Second, open voting in parliament may not only give room to pressure
from your actual electors, it may also lead to pressure from those who
did not actually vote for you or from well-organized lobbies defending
some specific interests. This is probably one of the concerns today in
many countries. Publicity allows the electors to use the threat of
non-reelection in full knowledge of what a representative actually
did. Still, it allows as well other actors to do so, who may neither
be the representative’s actual electors, nor lobbies acting in
the name of the general interest. Though employers and landowners may
not pressure employees and landless people anymore, lobbyists
representing the former do pressure the MPs elected by the latter,
with comparable results. This indicates how pressure made possible
through publicity may effectively force representatives to take
directions that do not fit with the view of the public interest
defended by those who elect them. Pressure is a double-edged sword.
This may call for at least mixed systems, a proposal to which we shall
return shortly. It may even justify taking more seriously the
possibility of vote anonymity in parliament, provided that the number
of votes for and against each proposal be known anyway.

Finally, one may consider that the only way of dealing with such
issues consists in obtaining empirical evidence regarding the
respective effects of secrecy and publicity and weighing them
More principled arguments could be developed however. Such principled
arguments could take at least two (non-exclusive) forms. According to
one strategy, even in a society where the level of coercion on voters
is very weak on average, it is reasonable to assume that those who
will suffer most from such risk of coercion (given their higher
dependence on others) are the least advantaged. There may thus be a
maximin case for secret ballots on such grounds, putting
priority on the protection of the most vulnerable. The other strategy
would consist in stressing the fact that the extent to which the
coercion of voters obtains in a given society may fluctuate, and hence
evolve at some point for the worse, in line with political or economic
degradations. A preventive approach would thus consist in
claiming that, once coercion increases to an extent such that it
becomes worrying (provided that there are levels of coercion that
would not be worrying in the first place), it would then be too late
to switch back to secret voting. This could justify preserving vote
secrecy from the start, even when the risk of coercion is currently

2.2 Deliberation

2.2.1 Can Publicity Civilize Representatives?

Let us now turn to the issue of public deliberation. Notice first that
the arguments for open ballots invoked above in the context of an
“aggregation of private interests” view of voting, equally
justify the publicity of debates taking place between representatives.
This is because these individuals have interests in learning about
each other. An implication of this fact is that, in a system in which
vertical accountability matters, it is essential both for citizens to
know what their representatives think and do, and for representatives
to know more than once every four years how the electorate reacts to
their arguments and decisions.

Learning about each other is one thing. Learning
from each other (about facts, what to think, or how to act)
is another,
Publicity is not supposed to allow only for citizens and their
representatives to act upon fixed preferences regarding their private
or even the public interests. The interactions that publicity enables
are also potentially rich in transformative power. Such a dynamic
impact can take various forms. Publicity of deliberations between
representatives may play an exemplary role for citizens (both
procedurally and substantively), as suggested by Bentham, who writes

the order which reigns in the discussion of a political assembly, will
form by imitation the national spirit. This order will be reproduced
in clubs and inferior assemblies, in which the people will be pleased
to find the regularity of which they have formed the idea from the
greater model (Bentham 1843 [1999:

Similarly, we have seen that open ballots can incentivize electors to
restrict themselves in public to positions that they see themselves as
being able to justify to other fellow citizens. And this is true
even when their fellow citizens are not especially educated,
as Mill argued:

The notion is itself unfounded, that publicity, and the sense of being
answerable to the public, are of no use unless the public are
qualified to form a sound judgment. It is a very superficial view of
the utility of public opinion to suppose that it does good only when
it succeeds in enforcing a servile conformity to itself. To be under
the eyes of others—to have to defend oneself to others—is
never more important than to those who act in opposition to the
opinion of others, for it obliges them to have sure ground of their
own. Nothing has so steadying an influence as working against
pressure. Unless when under the temporary sway of passionate
excitement, no one will do that which he expects to be greatly blamed
for, unless from a preconceived and fixed purpose of his own; which is
always evidence of a thoughtful and deliberate character, and, except
in radically bad men, generally proceeds from sincere and strong
personal convictions. Even the bare fact of having to give an account
of their conduct is a powerful inducement to adhere to conduct of
which at least some decent account can be given. If any one thinks
that the mere obligation of preserving decency is not a very
considerable check on the abuse of power, he has never had his
attention called to the conduct of those who do not feel under the
necessity of observing that restraint. Publicity is inappreciable,
even when it does no more than prevent that which can by no
possibility be plausibly defended—than compel deliberation, and
force every one to determine, before he acts, what he shall say if
called to account for his actions (1861 [1991:

Now, let us suppose that Bentham is correct to suggest that
representatives can civilize their electorate through public
deliberation, and that Mill is correct to suggest that voters may be
civilized by their not-necessarily-civilized fellow voters (and
non-voters) in case of open ballots. Elster presents still another
view according to which the publicity of deliberation can
assist citizens in civilizing their representatives. Elster refers to
this as “the civilizing force of hypocrisy”, and the
central passage is the following:

Generally speaking, the effect of an audience is to replace the
language of interest by the language of reason and to replace
impartial motives by passionate ones. The presence of a public makes
it especially hard to appear motivated merely by self-interest. Even
if one’s fellow assembly members would not be shocked, the
audience would be. In general, this civilizing force of hypocrisy is a
desirable effect of publicity. […] Publicity does not eliminate
base motives, but forces and induces speakers to hide them (Elster
1998a: 111; Naurin 2003).

The Filter

Elster’s argument occurs in two steps. The first one is the
filtering step, and it starts with the idea that the public holds
normative expectations regarding what representatives are supposed to
One of these expectations is that any overt reference to mere
self-interest is deemed unacceptable in the course of justifying their
which leaves open the question of whether it is acceptable to
refer to the self-interest of their class of
Another expectation is that representatives must justify their
position by referring to minimally general principles.
Finally, there is also a social norm favoring argument over
such that, whenever the debate takes place in public, deliberation
oriented towards the force of the better argument should be promoted.
At the very least, the entails the need to disguise overt threats as
warnings (Elster 1998a: 103–104). Normative expectations such as
these that regard acceptable discourse thus act as a filter on what
representatives are permitted to say. Importantly, it is not hypocrisy
as such that filters out the set of unacceptable speech elements.
However, it allows the representative to engage in a discourse in
which he does not necessarily believe. In other words, this
account does not necessarily presuppose a normative expectation among
the public that politicians ought to be
Actually, hypocrisy is rather an enabling factor (allowing for an
initial gap between what is said and thought), and not a cleaning up

Does this argument presuppose that voters be on average more civilized
than their representatives? At least two types of answers can be
provided. First, there are reasons of principle to doubt that
representatives are on average wiser or more committed to the common
good than their electorate. As Luban writes,

… the empirical validity of the publicity principle turns not
on whether the Many are ignorant or wrong-headed, but on whether their
leaders are less ignorant or less wrong-headed. No doubt the Wise are
few; and the leaders are few; but it hardly follows that the leaders
are wise. Before we reject the publicity principle because the leaders
know best, we must have reason to believe that the leaders know
better. And to find that out, we must look carefully at the variety of
mechanisms by which decision-making elites are actually selected. If
actual selection mechanisms choose randomly between the Many and the
Wise, or affirmatively disfavor the Wise, then the foolishness of the
many is irrelevant: the Few in official positions have no reason to
suppose that their policy brainstorms are any less foolish (Luban
1996: 193).

Second, the average level of civilization may be less relevant than
the presence of at least a minority of civilized members. As Mill puts
it, “[…] cases exist […] in which almost the only
restraint upon a majority of knaves consists in their involuntary
respect for the opinion of an honest minority” (1861 [1991:
362]). In line with this Millian view, were the average level of
civilization among representatives higher or equivalent to the one
obtaining within the electorate, publicity could still increase in
absolute terms the amount of civilized persons attending the
representative’s discourse, adding civilized members of the
public to those who are civilized in parliament. Hence, through such a
sampling effect, if we agree with Mill that the mere presence
of a civilized minority can make some difference, a larger absolute
size of such a minority might be relevant. As for Elster, he
explicitly relies more directly on a multiplier effect,
insisting that “a small group of impartially minded individuals
might induce many others to mimic their impartiality out of
self-interest” (Elster 1995: 249). And as Elster points out,
this is even more true whenever there is uncertainty as to who these
impartially minded people are (Elster 1995: 248).

Reducing cognitive dissonance

Let us now turn to the second step in Elster’s argument (Elster
1986, §II). There is an additional normative expectation that,
once representatives commit themselves in public to a given view, they
ought not to switch to another view unless they can justify such a
departure (Elster 1998a: 104). As Elster puts it,

public speaking is subject to a consistency constraint. Once a speaker
has adopted an impartial argument because it corresponds to his
interest or prejudice, he will be seen as opportunistic if he deviates
from it when it ceases to serve his needs (Elster 1998a:

Citizens may be more or less vigilant regarding compliance with such a
requirement that demands consistency through time. Still, it is likely
to have some weight. Furthermore, as representatives cannot publicly
depart from the principled views they have expressed earlier, they may
begin believing in what they say, even though they may not have been
holding such views at the time they began expressing
Hence, being forced to restrict oneself to publicly acceptable views
may turn out to influence what one eventually believes. This
is because, in order to reduce the dissonance between what one says
and what one believes, a representative may come genuinely to believe
what one is expected to say. Civilizing people’s speech will
eventually civilize their mind (and hopefully in turn their
non-linguistic actions) and—so Elster claims—“on the
average, […] yield more equitable outcomes” (Elster 1995:
251). What begins as a strategic use of a non-self-interested
arguments ends up leading to preference change (Fearon 1998: 54).

A related line of argument is offered by Robert Goodin, who contends
that individuals have a latent moral sense, and that having to state
arguments in public acts as a reminder of what they already
agree with. Crucially, they might not otherwise have connected the
issue at stake with the principles they
One way of contrasting Goodin’s position with Elster’s
view is the following. What is at stake would be as much a matter of
reducing dissonance between what one says and what one genuinely
believes, as one of reducing dissonance between two separate beliefs
one holds.

Having explored Elster’s claims concerned the civilizing force
of hypocrisy, there are two empirical questions to be asked. First,
are Elster’s factual assumptions empirically plausible? Second,
are there no other negative side effects relating to publicity that
might outweigh the benefits identified by Elster? In the next section,
we provide elements that may be relevant to answering the latter
question. Let us briefly address the first challenge here. Daniel
Naurin (2004) provides some interesting evidence in this respect. On
the one hand, he refers to a study by Christian Joerges and Jurgen
Neyer (1997) that looks closely at European Union comitology
committees. Their study indicates that public-spirited deliberation
can be present in such committees whose existence is unknown to most
people. This may seem reassuring as it suggests that publicity is
not necessary to foster truly deliberative behavior. However,
on the other hand, Naurin refers to evidence from Nina Eliasoph (1998:
7) suggesting that, in the case of activists, while in backroom
conversations these individuals tend to express themselves with
reference to justice and common good concerns, shifting to a front
stage context also increases their use of self-regarding arguments.
Naurin’s (2004) own empirical data illustrate a similar trend in
the case of industry lobbyists in the European Union context. In other
words, a forum-like type of behavior is already present in the lobby
corridor. Surprisingly enough, bringing lobbyists from the backstage
to the front stage therefore risks making things worse from the point
of view of reducing reference to mere self-interest. One should
admittedly be cautious in extrapolating such
Still, both Eliasoph and Naurin identify a phenomenon that goes in
the opposite direction of what Elster’s view implies. This is
clearly a serious challenge to his account.

2.2.2 The Quality of Deliberation

The variety of effects of publicity on the nature of the deliberation
process is an understudied matter. We have already looked at some
considerations, asking ourselves whether publicity substitutes
public-interest-oriented to self-regarding arguments, and whether it
leads to a shift from bargaining to deliberation. Here, leaving
Elster’s account aside for a moment, we go on with examining
other dimensions of what can be referred to—in an admittedly
vague manner—as the quality of the debate. We start with the
possibility that publicity risks modifying the nature of a discussion
in ways that are not desirable. One such negative side-effect, as
identified by James Madison in 1787 in the case of the American
Constitutional Convention, is the following:

Had the members committed themselves publicly at first, they would
have afterwards supposed consistency required them to maintain their
ground, whereas by secret discussion no man felt himself obliged to
retain his opinions any longer than he was satisfied of their
propriety and truth, and was open to the force of

The idea of sticking to a position and being unwilling to depart from
it, despite the existence of good arguments, can be a problem for at
least two
Consensus will be harder to reach because of such stubbornness, which
matters whenever agreement can be reached only by consensus (be it in
a deliberation or in a bargaining type of context). More importantly,
we may be worried not (only) by actors’ lack of sincerity, but
by the loss in debate
Unless the deliberating parties are able to try out ideas, to show
and to re-consider the issues again and again with a fresh eye,
actual deliberations may be no more than the juxtaposition of
pre-prepared statements with no genuine interaction taking
Part of the problem here is that the public may view hesitation and
trial and error as signs of a lack of commitment, which may
disadvantage those representatives truly taking part in the
discussion, hence inhibiting them from adopting a genuinely
deliberative posture. And as Naurin puts it,

[…] if deliberation is about transforming preferences, and
publicity forces you to know what you want and stand by your position,
then “public deliberation” is something of a contradiction
in terms (Naurin 2003: 32).

Because of this, if we value deliberation as a living interaction with
real transformative power, there may be a strong case for at least
some proportion of political deliberation to take place behind
closed-doors, despite the latter’s negative

Ellen Meade and David Stasavage (2008) document another way in which
publicity might negatively affect the quality of deliberation. They
deal with the specific case of central banks and whether minutes of
board meetings should be released. Based on empirical evidence, they
show how, under conditions of publicity, the fact that expert advisors
may have career concerns reduces the chances of dissent being
expressed during deliberation. Here, what is at stake is not whether
one sticks to one’s own initial position. Rather, it
has to do with whether one is ready to show disagreement with the
position taken by another expert who is highly regarded and
has spoken
Hence, publicity might generate disincentives both to openly changing
one’s mind and, in some contexts, to dissenting with other
(authoritative) people’s views.

Simone Chambers (2004) points at still another negative impact of
publicity on the quality of deliberation. Discussion behind closed
doors will admittedly tend to allow for the expression of
particularistic reasons that would not pass the test of actual
publicity. However, while making deliberation public might force the
deliberating parties both to provide the best possible reasons in
support of their claims (Socratic effect) and to provide public
reasons (democratic effect) in the sense of reasons that should refer
to the public interest and be acceptable to a wide audience in a
pluralistic society, one problem
It is that publicity also tends to favor reasons that are
“plebiscitary” in the sense of shallow or
manipulative (Chambers 2004: 393). Moving from private,
particularistic reasons to truly public reasons may then often have an
associated price, i.e., shifting from profound (private) reasons to
shallow (public) reasons. Admittedly, the reason why such shallow
reasons prosper might have to do with the need for a common
denominator, which could be needed in a deliberation behind closed
doors as well. However, it is plausible that it may also have to do
with an attempt at pleasing the public and thus giving way to

These are just a few illustrations of possible negative side effects
of bringing deliberation under the public eye. It may become less
lively, leave less room for dissent in the presence of an
authoritative member, and favor shallower types of arguments, no
matter how rational and public they are. It is important to identify
exactly which benefits and which costs are attached to actual
publicity and secrecy, and what their respective weight might be. In
this respect, it is true that the argument may be of a different
nature or at least carry a different weight depending on whether one
deals with voting or deliberation. For instance, the importance of
learning from each other (rather than simply about each other) may
provide a stronger argument for open public deliberation than for open

Once all normative and empirical dimensions have been carefully
considered, we may want to adopt a firm standpoint regarding the need
for publicity or secrecy in deliberation and voting in general.
However, we may also want to try to mix secrecy and publicity
in ways that allow for more optimal schemes, providing us with the
best of both worlds. Let us illustrate the availability of this option
by pointing at two of such schemes. First, Bentham refers to a system
obtaining in the Polish Permanent Council in the late 18th century by
which an open vote was followed by a secret vote on the very same
issue (Bentham 1843 [1999: 147–148]). Second, Elster argues that
a constitution-making process

ought to contain elements of both secrecy (committee discussion) and
publicity (plenary assembly discussions). With total secrecy, partisan
interests and logrolling come to the forefront, whereas full publicity
encourages grandstanding and rhetorical overbidding. Conversely,
secrecy allows for serious discussion, whereas publicity ensures that
any deals struck are capable of withstanding the light of day (Elster
1998a: 117).

One can thus mix both publicity and secrecy in the case of voting
(Bentham) or deliberation (Elster), through a succession of openness
and secrecy
or secrecy and openness
One could also advocate for mixing public deliberation with a secret
vote (which is probably a correct description of what happens during
electoral periods in many democratic countries), or, conversely,
secret deliberation and a public vote (which is what actually takes
place in parliamentary systems with secret commissions in which
plenary debates take a purely formal dimension). Each of these regimes
has specific properties that may warrant separate study.

3. Rawls on Public Reason and Public Rules

Having looked at both hypothetical publicity and actual publicity, let
us now turn to examine the place of publicity within the political
philosophy of John Rawls. We shall analyze two distinct conceptions of
publicity that arise in his writings. The first relates to the
doctrine of public reason. The second relates to the doctrine
of public rules. Though we begin with some remarks about
public reason, we shall focus predominantly on public rules, as this
idea is less familiar.

3.1 The Doctrine of Public Reason

Stated in a quite general way, the doctrine of public reason holds
that, when justifying the exercise of political power, we ought to
appeal exclusively to reasons that all “citizens as free and
equal may reasonably be expected to endorse in the light of principles
and ideals acceptable to their common human reason” (Rawls 1996,
137). So as to highlight their shared nature, we can call these
public reasons.

In contrast with public reasons, there are non-public
. These reasons are controversial, in the sense that they
are not reasons that we can expect all reasonable citizens to accept.
Their possibility emerges from the fact of reasonable pluralism (Rawls
1996: 36–39), which refers to the idea that, in any society in
which basic rights are protected, citizens will inevitably disagree
about the worth of competing conceptions of the good life. The
doctrine of public reason states that, because certain reasons are
controversial, we ought not to appeal to them when making the case for
the political legislation we support. In this respect, the doctrine
calls for a kind of justificatory restraint (Rawls 1996:

To illustrate, let us consider the case of a Christian, who is opposed
to abortion for religious reasons. The doctrine of public reason holds
that, even though this citizen may act on these reasons in her own
personal life, she ought not to appeal to her religious beliefs in
order to oppose the legal right to abortion. This is because her
Christian reasons are not ones that we can reasonably expect others to
endorse. Instead, if she wishes to oppose the legislation, she must
offer reasons that are shared by all reasonable citizens (Rawls 1996:
243 n. 32).

One concern with the doctrine of public reason is that, by asking
citizens to set aside certain controversial reasons in political
decision-making, it invites them to misrepresent their beliefs in
various ways. This is the Insincerity Objection. For example, the
Christian who is prohibited from appealing to her religious beliefs in
order to oppose the legal right to abortion may decide to offer an
argument that she does not find persuasive, but that she hopes others
will, in order to advance her political agenda. This may be
objectionable for a number of reasons, including because it may reduce
the quality of political deliberation (Quong 2011: ch. 9; Schwartzman

A further concern is that the doctrine of public reason proves too
much. This is because we disagree about almost all values, including
political values. Even when we use political power in seemingly just
ways, such as in the name of equality of opportunity, say, we must
appeal to controversial reasons that are not widely accepted. Does it
follow that we cannot reasonably expect citizens to endorse those
reasons, and, if so, isn’t the doctrine of public reason too
restrictive? Or, despite such disagreement, can we still reasonably
expect their endorsement? If so, why can’t we also reasonably
expect citizens to endorse other controversial reasons, such as those
that derive from religious beliefs? Proponents of the doctrine must
answer these questions by explaining why we may treat disagreements
about political values differently from disagreements about
conceptions of the good life. For this reason, we can call it the
Asymmetry Objection (Caney 1995; Fowler and Stemplowska 2015; Sandel
1994; Waldron 1994).

The precise formulation of the doctrine of public reason is subject to
further variation. Among other factors, it will depend upon (1) how we
discriminate between the reasons that we can and cannot reasonably
expect others to endorse; (2) whether its demands apply
broadly to any matter of public concern or narrowly
only to the essentials of a political constitution (Kelly 2013; Quong
2011: ch. 9); and (3) whether we acknowledge that there may be
conditions under which the doctrine is defeated by other
considerations, such concerns for religious salvation, for example
(Rawls 1996: 152; Clayton and Stevens 2014).

Though the arguments we discuss below provide partial answers to these
questions, we shall not explicitly discuss these issues here. Instead,
our goal is to outline the various arguments in defence of the idea of
public reason. These arguments clarify the demands of public reason,
as well as inform our later analysis of public rules.

We can distinguish three distinct justifications for the doctrine,
each of which finds support within Rawls’s texts:

  1. The Political Autonomy Argument
  2. The Civic Friendship Argument
  3. The Stability Argument

The Political Autonomy Argument begins with the claim that each
citizen has a weighty interest in being politically
—that is, in freely identifying with the
constraints that she faces, such that she fully comprehends and
endorses the justification of those constraints (Rawls 1996:
Only by justifying the exercise of political power by appeal
exclusively to public reasons can we protect citizens’ interest
in political
After all, what distinguishes public reasons from non-public reasons
is that the former are ones that we can reasonably expect everyone to
endorse and to act upon.

One objection to this argument comes from those who deny that citizens
have a weighty interest in being politically autonomous. These critics
target the fact that it is possible to justify the kind of
justificatory restraint imposed by the doctrine of public reason only
if this interest is sufficiently weighty. If the interest were not
very weighty, or if there were no interest at all, then it is
difficult to see why we ought to exercise the kind of justificatory
restraint that the doctrine of public reason calls for, especially
since it prevents us from appealing to a certain class of sound
reasons (Raz 1986; Wall 1998).

The second justification for the doctrine of public reason is the
Civic Friendship Argument. On this view, we can justify the doctrine
by appeal to its role in enabling citizens jointly to maintain just
institutions, and thereby realise a form of civic friendship.
Public reason is central to this task since the joint pursuit
of justice is possible only if citizens “affirm the same
political conception of justice” and so “share one very
basic political end” (Rawls 1996: 202).

Political autonomy and civic friendship are distinct ideals. Unlike
political autonomy, civic friendship is an essentially common
good—its benefits cannot be enjoyed by an individual in
isolation but can be enjoyed only with others. Only citizens who
relate to each other in a particular way—that is, as joint
pursuers of a shared conception of justice—can enjoy this kind
of civic friendship (Rawls 1996: 204; Leland and Wietmarschen 2017;
Lister 2013; Ebels-Duggan 2010; Schwarzenbach 1996, 2005; Watson and
Hartley 2018). Elaborating upon the idea of essentially common goods,
Rawls writes,

That there should be such a political and social good is no more
mysterious than that members of an orchestra, or players on a team, or
even both teams in a game, should take pleasure and a certain (proper)
pride in a good performance, or in a play of the game, one that they
will want to remember (Rawls 1996: 204).

The success of the Civic Friendship Argument depends upon two claims.
The first is that it is not possible, or at least it is much more
difficult, for citizens to realise civic friendship in ways other than
by upholding the doctrine of public reason. Some deny this claim,
instead maintaining that forms of democracy that do not uphold the
doctrine of public reason can fare equally as well in this regard
(Billingham 2016; Caney 1995; Kogelmann 2017). The second is that
citizens’ interest in realising civic friendship is sufficiently
weighty so as to justify the kind of restraint imposed by the doctrine
of public reason (Caney 1995).

The third justification of the doctrine of public reason is the
Stability Argument. This argument emphasizes the alleged psychological
fact that citizens are more likely to support principles and
institutions that they themselves see as just (Rawls 1996:
140–4; Weithman 2011). This implies that, if we want to extend
the life-expectancy of our institutions, then we should favour
principles and institutions that will predictably generate their own
support (Freeman 2007b: ch. 3). This counts in favour of the doctrine
of public reason, whose defining claim is that we may appeal only to
those reasons that we can reasonably expect others to endorse.

Unlike the previous two arguments, the Stability Argument is purely
instrumental. It sees citizens’ support for the principles and
institutions to which they are subject as valuable only insofar as it
enhances the stability of those principles and institutions. Of
course, this is not to say that stability is always valuable: instead,
the stability of a set of institutions is valuable only if those
institutions meet other moral standards, such as if they treat
citizens justly.

3.2 The Doctrine of Public Rules

Whereas the doctrine of public reason restricts the reasons to which
we may appeal, the doctrine of public rules regulates the character of
the requirements that principles of justice may
More specifically, it states that we have weighty reasons to prefer
principles of justice that can serve as a public standard, in the
senses that they are (i) not self-effacing; and yield demands (ii)
that are sufficiently determinate and accessible, and (iii) compliance
with which is sufficiently verifiable to others (Rawls 1996:
66–67; Rawls 1999a: 48–49; see also Williams 1998: 233).
We explain each of these properties in turn.

First, principles of justice provide a public standard only if they
are not self-effacing—that is, only if the success of
the principles from which the rules derive does not depend upon some
citizens being ignorant of these principles, or upon some citizens
being ignorant of others’ knowledge of these principles (see
In other words, the idea of publicity requires mutual knowledge, and
not mere general knowledge (see
In clarifying what he means by a public system of rules, Rawls

I mean then that everyone engaged in it knows what he would know if
these rules and his participation in the activity they define were the
result of an agreement… He also knows that the others know this
and that they know that he knows this, and so on (1999a:

Second, principles of justice provide a public standard only if they
provide sufficiently determinate and accessible standards of moral
permissibility, such that a “person taking part in an
institution knows what the rules demand of him” (Rawls 1999a:
48). This has two components. First, principles pass the test of
only if they are not so vague that they are incapable
of producing clear
Second, principles pass the test of accessibility not merely
when citizens can attain knowledge of how they would be required to
act if they were to have the relevant empirical information, but
instead only if citizens are in fact able to attain knowledge of that
empirical information (Rawls 1996: 182; Lippert-Rasmussen 2008:

Third, principles of justice provide a public standard only if
compliance with their demands is sufficiently verifiable to
. For Rawls, what matters is not only that “the
institutions of the basic structure are just” but also that
“everyone recognizes this” (1996: 66). This test of
verifiable compliance
therefore requires the possibility of
widespread knowledge of the extent to which citizens discharge their
duties. This test is most plausibly understood as an aggregative one.
For example, the duty not to litter is verifiable to others in virtue
of the fact that we can attain knowledge of the extent to which others
on aggregate refrain from littering, even though we may not
be able to attain knowledge of the extent to which any given citizen
refrains from

Rawls appeals to the doctrine of public rules in order to justify
several features of his view, including the idea that the principles
of justice should govern the distribution of social primary goods
rather than the distribution of welfare. Part of what justifies this
verdict is that, because it is not feasible accurately to measure
welfare, principles that refer to welfare fail the test of
accessibility. Accordingly, these principles “run afoul of the
[publicity] constraint”, such that “In political life
citizens cannot sensibly apply or follow it” (Rawls 1996: 182 n.
11; Rawls 1999b: 13 n. 3). Because of this, principles of justice that
govern the distribution of welfare abrogate their chief purpose, which
is “to provide a publicly recognized point of view from which
all citizens can examine before one another whether their political
and social institutions are just” (Rawls 1996: 9).

In the same vein, Andrew Williams writes,

…we have reason to reject conceptions of justice which, given
the fact of limited information, are too epistemically demanding to be
public and stable. We should, like Rawls, favour conceptions whose
scope is restricted to publicly accessible phenomena (1998: 245).

An implication of the doctrine of public rules, therefore, is that we
have reasons to object to highly epistemically demanding principles of
justice that fail the tests of determinacy, accessibility, and/or
verifiable compliance. As an example of this, let us consider G. A.
Cohen’s claim that justice requires citizens to internalize an
egalitarian ethos that condemns self-serving behaviour and,
subject to a personal prerogative and special labour burdens, requires
citizens to be guided by a concern for the least advantaged when
making occupational choices (2008: part 1). Rejecting this possibility
on the grounds that it violates the doctrine of public rules, Williams
writes that

it is quite possible that, in typical cases, we will lack a
sufficiently precise public standard by which to justify, or
criticize, each other’s self-serving behavior (1998: 240).

This is because, in order to calculate whether a citizen is acting in
accordance with the demands with the egalitarian ethos, we would need
to know which occupation available to her best serves the interests of
the least advantaged, as well as whether her choice to pursue an
alternative occupation would be protected by either her personal
prerogative or a concern for special labour burdens.

The success of Williams’s reply to Cohen depends partly upon the
existence of weighty reasons to endorse the doctrine of public rules.
For this reason, we shall now turn our attention to the various
arguments in defence of this doctrine. In particular, we highlight
three distinct justifications, each of which corresponds to an
argument in defence of the doctrine of public reason.

The Political Autonomy Argument justifies the doctrine of public rules
by appeal to its role in protecting political autonomy (Williams 1998:
244). If citizens have no clear sense of what the principles of
justice demand of them, then they lack awareness of the moral
constraints to which they are subject, and so are prevented from
freely identifying with those constraints. A similar threat to
political autonomy arises when principles of justice fail the test of
verifiable compliance. If citizens are freely to identify with the
constraints to which they are subject, it is important that they can
recognise these constraints as a product of others’ compliance
with principles of justice.

The second justification is the Civic Friendship Argument, according
to which we should uphold the doctrine of public rules because of the
role it plays in realising civic friendship (Williams 1998: 244).
According to this view, civic friendship can exist within a society
only when citizens are able justifiably to regard themselves and
others (and justifiably be regarded by themselves and others)
as doing what is necessary to sustain reasonably fair terms of social
cooperation. Publicity facilitates each of these ends. Just as members
of an orchestra are denied a common good when they cannot hear, or
otherwise appreciate, their own or others’ contributions to
their joint performance, so too citizens are denied the good of civic
friendship when they are denied awareness of their own or
others’ compliance with the principles of justice.

Finally, the Stability Argument justifies the doctrine of public rules
on the grounds that its acceptance enhances “the long-term
probability of a society conforming with its conceptions of
justice” (Williams 1998: 244). Again, this instrumental argument
relies upon the alleged psychological fact that citizens are more
likely to support principles and institutions that they perceive as
just (Rawls 1996: 140–4). This supports the doctrine of public
rules insofar as citizens are more likely to perceive principles and
institutions as just when the demands of justice are sufficiently
determinate and accessible, and compliance with them is sufficiently
verifiable to others (for objections: Caney 1995; Lippert-Rasmussen
2008: 45–46).

3.3 The Good Faith Efforts Objection

The most sustained critique of the doctrine of public rules comes from
Cohen (2008: ch. 8). Though he advances a number of distinct
objections to the doctrine, for reasons of space we shall focus on
only one element of his analysis, in which he disputes the claim that
principles of justice must pass the tests of determinacy,
accessibility, and of verifiable compliance. Cohen denies that, in
order to be valid, principles of justice must be capable of serving as
a public standard. Instead, he maintains that we can rely solely upon
citizens’ good faith efforts. This is the Good Faith
Efforts Objection (Casal 2015; Lippert-Rasmussen 2015: ch. 8).

In support of this objection, let us consider a gender egalitarian
ethos that condemns women having to do more housework than men, for no
good reason. As with Cohen’s egalitarian ethos, this ethos may
fail to provide a public standard by which to justify and to criticize
our own and others’ sexist behaviour. This is because, in order
to determine whether a husband who refuses to do the washing up
because of backache is acting in accordance with the demands of the
gender egalitarian ethos, we would need to know how severe his
backache is, as well as to compare the other domestic burdens that he
has borne with those that his wife bears. This makes the gender
egalitarian ethos highly epistemically demanding, such that it may
fail the tests of determinacy, accessibility, and/or verifiable
compliance (Cohen 2008: 359–60). However, it is intuitively
implausible to reject the gender egalitarian ethos on this basis.
Rather, in this case, justice requires “good-faith efforts, not
omniscience” (Casal 2015:

Cohen offers a second case in support of the Good Faith Efforts

During World War II in Britain, a social ethos induced people to
sacrifice personal interests for the sake of the war effort, and
everyone was expected, as a matter of justice, to “do his
bit,” to shoulder his just share. It is absurd to suppose that
someone could have stated precisely what amount of sacrifice that
injunction required, and it is true, therefore, that, with respect to
many people, one couldn’t tell, and with respect to some, they
couldn’t even themselves tell, whether they were sacrificing on
the required scale (2008: 353).

Cohen maintains that, rather than reject such an ethos on the grounds
that it is inconsistent with the demands of the doctrine of public
rules, we should instead abandon the latter. What matters in these
cases, Cohen thinks, is that citizens are able to make “a
reasonably efficacious societywide good faith effort
(2008: 352 [emphasis in original]).

In reply to this objection, proponents of the doctrine of public rules
might circumscribe their claims, holding that we should prefer
principles of justice that provide a public standard only under
certain conditions
. One possibility is to relinquish the doctrine
of public rules in those cases in which citizens are required to bear
costs that are much smaller than the benefits generated by good faith
efforts to comply with those principles. This would explain why the
indeterminacy, inaccessibility, and lack of verifiable compliance is
not troubling in the two cases we have just considered. In the first,
the gender egalitarian ethos requires citizens to bear minimal costs
in the fight against sexism; and in the second, Cohen’s
sacrifice ethos requires citizens to bear potentially larger costs,
but the reward is enormous.

This would not amount to an abandonment of the doctrine of public
rules. Rather, according to this reply, we should prefer principles of
justice that provide a public standard, except when there is not this
discrepancy in costs. As the costs that citizens are required to bear
increase in comparison with the benefits they are required to
generate, it becomes increasingly important that citizens know both
what their duties demand of them and the extent to which others
discharge their duties.

Two reasons support this conviction. First, without determinate and
accessible standards of moral permissibility, there is a risk that
some citizens will make larger sacrifices than what is morally
required of them. When the costs involved are comparatively high, this
is likely to be unfair. To illustrate this unfairness, we can consider
the case of an individual who internalises the egalitarian ethos, and
who takes on some grueling occupation only because she fails to
recognise that, because of her personal prerogative, she would be
morally justified in pursuing a more enjoyable occupation.

Second, it is sometimes reasonable for a citizen who makes
considerable sacrifices to have assurance that other citizens
are discharging their duties. This is because she knows that she is
not being exploited by non-compliers, and because she knows that, if
she were not to make such sacrifices, she would be free-riding on
benefits that others’ compliance brings (Lister 2020; Mason
2012: 541). Again, these concerns are likely to be more powerful when
the sacrifices required of citizens are comparatively high.

This reply to the Good Faith Efforts Objection is incomplete in
multiple ways. Proponents of this reply must specify the implications
of their views in a larger range of cases, including those in which
the costs and benefits involved are of comparatively equal values
(both high and low). Moreover, it also remains important to explain
more precisely how this reply relates to the three justifications for
the doctrine of public rules that we considered above. Meeting these
challenges is central to establishing the plausibility of the doctrine
of public rules.

4. Conclusion

We began our exploration by looking at Kant’s hypothetical
publicity test. We asked ourselves whether there were good reasons to
believe that hypothetical publicity should be considered a sound
requirement to test maxims of action. We also wondered whether
hypothetical publicity entailed significant restrictions from the
point of view of actual publicity. The answer appears to be rather
negative in both cases. This is because, first, there does not seem to
be any intrinsic relationship between self-frustratingness and
justice. To put things in another way, the less the effect of
publicity is contingent, the more we can doubt about its validity as a
test of justice. And second, if no criterion is provided as to the
level of generality at which the Kantian test should operate, we run
both risks of under- and over-exclusiveness. Which maxims of action
should be deemed unjust is therefore left rather undefined by the test
in most cases.

While it is unclear whether the Kantian hypothetical publicity test
could (and should) be rescued, there is no doubt that actual publicity
issues raise interesting philosophical challenges. One of them has to
do with the articulation between, on the one hand, the function that
political philosophy ascribes to voting, representation, and
deliberation; and, on the other hand, the types of effects that could
be expected from going public. In the course of our discussion, we
also tried to draw parallels and to identify differences between
voters and their representatives (e.g., through the idea of horizontal
accountability) and between the voting and the deliberation debate.
The meaning of the publicity requirement may differ significantly as
well when we compare standard representative bodies (e.g.,
parliaments) with the civil service or with “independent”
bodies acting within a democratic context (e.g., central banks or
courts). Publicity in the latter case is likely to have both a meaning
and consequences quite different from the ones attached to public
deliberation, e.g., in a parliament’s plenary session.

Moreover, while we did not question the possibility of developing
principled views in the field of actual publicity, factual assumptions
present in most of the key arguments in this respect (e.g., in those
of Mill or Elster) require more extensive empirical research. The
results that we reported from some of these empirical investigations
are potentially rich in consequences for our normative theories. The
data gathered by Naurin (2004) and Meade and Stasavage (2008) are
especially important in this respect. As is the case with other issues
in political philosophy, there is certainly a need here for
philosophers to engage in a dialogue with social scientists.

Finally, we briefly presented Rawls’s uses of the idea of
publicity. The variety of ways in which he refers to it certainly
shows the need for a more substantial mapping enterprise, at a
strictly conceptual level. This could in turn indicate reasons to
develop philosophical arguments in new directions.

Read More

Axel Gosseries and Tom Parr