The following text has been automatically reproduced by an Optical Character Recognition (OCR) algorithm. It may not have been checked over by human eyes. For matters of precision please consult the original pdf.

Liberty, Authority and the Negative Dialectics of J.S. Mill

Liberty, Authority and the Negative
Dialectics of J.S. Mill
Trevor Pateman
I Overview
Social systems do not automatically reproduce themselves, and it may require an immense effort to create
or recreate conditions of possibility for the continued reproduction of a particular social order, or
component part of a social system. For at least part
of the nineteenth century in Britain it was no easy
task to sustain the conditions of possibility of both
capitalist production and a particular kind of parliamentary government, and for much more of the century
the task was perceived as existing. Its magnitude is
visible not only in the historical record of disorder,
moral and political, but also in the volume of theorising devoted to consideration of how to reproduce
both capitalism and parliamentary government.

In this article, I examine the internal structure
and intellectual sources of some aspects of the
utilitarian contribution to that theorising, as
represented in the work, principally, of John Stuart
Mill, but also of James Mill, John Austin and Sir
George Cornewall Lewis. I shall be concerned mainly
with Mill’s theory of authority and his theory of
liberty, and, more specifically, their internal interrelation. I shall argue that the kind of liberty
Mill defends in On Liberty has the characteristics
which it does, firstly, because it is a component of
a theory of authority adhered to by Mill from about
1830, formulated by him in The Spirit of the Age
(1831) [2] in terms borrowed from Austin, and developed at much greater length than it ever was by
Austin or Mill in Lewis’s An Essay on the Inj1uence
of Authority in Matters of Opinion (1849) [3]. This
theory of authority is not only specific to Utilitarianism (as argued by Richard B. Friedman in a very
important study to which I am indebted [4]) but
historically specific, for it is a theory of the
conditions of the possibility of political order in
a society in which the Agricultural and Industrial
Revolutions have already occurred, the truths of
Political Economy are established, and the establishing of truths in other branches of moral and political science is in prospect. This theory of authority
is the subject of Section 11 of this article.

But, secondly, Mill’s theory of liberty is as it
is because of his adherence to a particular theory of
the nature of moral and political truth and the means
of arriving at such truths, a theory which is, essentially, Platonism without a theory of Forms. Mill
takes over from Plato a dialectical methodology for
the pursuit of moral and political truth, divorced
from Plato’s correspondence epistemology of the
nature of truth as consisting in the adequation of
our ideas to the Forms which pre-exist them. In
section III below, I consider both the nature of
Mill’s Platonism and its consequences for his theory

of liberty. The utilitarians’ understanding of
Greek history and philosophy also undoubtedly influenced their thinking on the kind of political authority relationship possible in nineteenth-century
Britain, a topic I shall only briefly touch on [5].

11 Models of Authority

The Art of printing exists. And the irresistible
progress of the information which it diffuses
necessitates, not a change merely, but a perfect
revolution in the art of governing mankind.

In the times that are gone, the art of government has consisted in a mixture of fraud and
force in which, commonly, the fraud predominated.

In the times that are to come, as fraud will
be impracticable, and as knowledge of what is
good and what is evil in the mode of managing
the national affairs cannot be witheld from the
nation, government will be left either to
rational conviction, or to naked force. This
is the grand revolution of modern times [6].

(James Mill)
This quotation shows the author of the geometrical
theory of government thinking historically about the
conditions of possibility of political order – though,
admittedly, the history is schematic and represents
commonplace Enlightenment thinking. For my present
purposes, the quotation is of interest in two respects. First, because Mill takes for granted the
notion that mankind still needs governing. Ten
years previously, in the essay on Government, James
Mill had been explicit in ruling out the possibility
of mankind governing themselves in what we should
now call a direct democracy. Second, it is of interest because it advances ‘rational conviction’ as a
condition of the possibility of political order. In
this, Mill apparently departs from advocacy of the
deference politics he described in the final section
(section X) of Government, and which served to answer
the objection to his proposals for franchise extension, ‘That the People are not capable of acting
agreeably to their Interest’. In that section, he
argues that the working class does and will defer
in forming and expressing political judgements to
the views of the middle rank:

The opinions of that class of the people, who
are below the middle rank, are formed, and
their minds directed by that intelligent and
virtuous rank, who come the most immediately
in contact with them, who are in the constant
habit of intimate communication with them, to
whom they fly for advice and assistance in all
their numerous difficulties, upon whom they
feel an immediate and daily dependence, in

health and in sickness, in infancy, and in
old age; to whom their children look up as
models for their imitation, whose opinions
they hear daily repeated, and account it their
honour to adopt.

The trouble with this picture was that it was
already becoming anachronistic when it was painted.

The Agricultural and Industrial Revolutions, as
every sociology student knows, destroyed gemeinschaft
(organic) relations and substituted gesellschaft
(contractual) ones; there was no question of the
capitalist class playing out, as individuals, in
large manufacturing towns, roles modelled on those
of the Anglican clergy and rural squirearchy, and
this impossibility was created more by the steam
engine than the art of printing. Though some
thinkers – romantic reactionaries and Utopian socialists – continued to think in organic terms, for
Utilitarians who accepted the Agricultural and
Industrial Revolutions a major problem existed of
determining the kinds of political relations possible
in nineteenth-century Britain. The problem has an
analogue in that which Socrates and Plato and those
they spoke for confronted in’post-Homeric Greece:

the demos (the people) could only be tamed from a
new position, not by a reversion to Homeric models
[8].” The analogy was not lost on nineteenth-century

But in what way could ‘rational conviction’

provide a solution to the problem of governing mankind? To answer this question, it will be useful to
sketch in abstract three possible conceptions of
rational conviction, or, rather, three constellations
of conceptions.

First, there is a family of conceptions I shall
call monologic-[9] in which rational conviction is an
achievement of individuals not essentially dependent
on any community of inquiry or evaluation. Such
individuals engage in the collection and consideration
of evidence, which in the dominant empiricist tradition of monologism consists in facts (the products
o~ ohservation and introspection) expressible in an
observation language which allows theory-independent
access to both inner and outer Nature (human nature
and the external world), to which observation statements, if true, correspond. This activity generates
beliefs and is the source of rational conviction
(alternatively expresses as ‘moral certainty’ or
‘assurance’ in Enlightenment writings) provided that
all the evidence is gathered and subjected to unbiased
unprejudiced scrutiny (the requirement of ‘indifferency’): In other words, the characterisation of the
nature of evidence and truth in monological epistemologies is interdependent with an account of how the
business of enquiry ought to be conducted if it is to
yield rational conviction. In Locke, for example,
the theories advanced in the Essay Concerning Human
Understanding [10] are interdependent with the prescriptions to avoid bias and selectivity found in
Locke’s The Conduct of the Understanding, a work
little read nowadays, but held in high regard by
nineteenth century utilitarians [11].

In contrast, a second family of conceptions of
rational conviction, which I shall call dialogic,
departs from the monologic picture at either or both
of two levels. First, it is disputed that enquiry
could be in principle a private enterprise, either
because of the enquirer’s dependence on a public
language or on producing intersubjectively verifiable
observations (which banishes any use of introspection)
or both. These claims usually go along with a denial
of the possibility of .a neutral observation language,
and an insistence on the theory-dependence of all
enquiry, but I do not think this is a necessary
connexion. Wittgenstein’s work typifies this first
kind of dialogism. Second, the idea that individuals

could in isolation correct for their own biases and
arrive single-handedly at a rational conviction is
resisted by all those theories inspired by Plato’s
dialogues, notably the GOrgias. Plato insists that
rational conviction is something of which only the
person whose opinions have withstood the critical
scrutiny of others is capable, internal dialectic
being at best a poor substitute for living, external
dialectical argument. Such an approach is relevant
not only to scientific enquiry and the pursuit of
self-knowledge, but to the organisation of political
debate in a public sphere which is where citizens
strive to achieve a rational political will [12].

Hannah Arendt’s ideas are very much a development of
this second aspect of dialogism [13]. Both aspects
are, I think, welded together in the work of philOsophers like C.S. Peirce and Josiah Royce [14), and
Royce’s claim that ‘interpretation is a conversation
and not a lonely enterprise’ might serve as a mnemoni(
both for what dialogic conceptions are about, and for

what they contrast with. In section III below, I will
quote J.S. Mill expressing similar sentiments.

However, from the standpoint of nineteenth-century
utilitarians, both the monologic and dialogic account
of rational conviction could be accepted as coherent
and admirable – yet totally inapplicable, because not
generalisable as models of political or moral conviction in a society based on capitalist production and
division of labour. The reasons are quite straightforward. Both Austin and J.S. Mill argue that
realisation of the monologic conception in the political (and ethical) domains would demand a commitment
of time much in excess of the free time available to
wage labourers. Thus, Austin writes that, ‘ethical,
like other wisdom, “cometh by opportunity of leisure”.

And since they are busied with earning the means of
living, the many are unable to explore the field of
ethics, and to learn their numerous duties by learning the tendencies of actions’ [15]. And J.S. Mill
reiterates this in the second article in the series
Spirit of the Age arguing that the necessity of
labour for the great mass of mankind means that while
‘it is desirable that all should be firmly persuaded
[of the ground of opinion] … [the minority] alone
can entirely and philosophically know’ [16]. Lest
this should seem inconsistent with the doctrines of
Locke, to which utilitarians were committed, Mill say!

that Locke’s doctrines were intended for students ‘fo)
whom alone the great man wrote’ [17] – a view which
could be supported from Locke’s own equivocating
statement, ‘What colours may be given to this
[principling of pupils and students] or ·of what use
it may be when practised upon the vulgar, destined to
labour and given up to the service of their bellies,
I shall not here enquire’ [18], which at least

suggests that for all the polemical demagogy of the
Essay Concerning Human Understanding, Locke’s model
of the pursuit of knowledge is not intended to be
prescriptive for everyone.

As for the dialogic conception, that too is nongeneralisable because of the demands on time it would
make; as I noted above, James Mill in GOvernment rules
out direct democracy for this reason. Even if institutionalised through the written word (which Enlightenment thinkers like Condorcet tended to prefer to
the spoken as a medium of political discussion), the
dialogic conception can only lead to rational conviction among people who, apart from every other require. ment, know how to gather, evaluate and compare evidence. In other words, dialogic practice depends on
an intellectual culture which the mass of humankind
lack, and which, in virtue of a division of labour
sanctioned by the truths established by Political
Economy, they will continue to lack in the foreseeable future, even with the introduction (favoured by
utilitarians) of universal elementary education.

These considerations not only rule out the monologic and dialogic models as candidates for the kind
of rational conviction which could sustain a political
order, but point towards a third, generalisable, conception of rational conviction founded upon authority.

Austin puts the case for this as follows:

I.firmly believe (for example) that the earth
moves round the sun; though I know not a tittle
of evidence from which the conclusion is inferred.

And my belief is perfectly rational, though it
rests upon mere authority. For there is nothing
in the alleged fact, contrary to my experience
of nature; whilst all who have scrutinized the
evidence concur in confirming the fact; and
have no conceivable motive to assert and diffuse
the conclusiort, but the liberal and beneficent
desire of maintaining and propagating truth.

In this passage, Austin shows the utilitarians [20]
how belief in authority can be consistent with rationality, and indeed required by it – something unappreciated in what one might call popular Enlightenment
thought, and denied in some radical Enlightenment
thinking. Nor is the rationality of belief on authority a second-best rationality, which could be improved
upon if we were prepared to dispense with the division
of intellectual labour. For, first, it is a general
fact about human existence that people are located at
different points in space and time, so that the testimony of some people will always be authoritative
about matters to which they alone were and could have
been witnesses. And, second, the nature of the
scientific enterprise had by the nineteenth century
clearly turned out to be such that the discovery of
new truths was only possible on the basis of an
extensive intellectual division of labour. Thus, in
principle, the laity stood to be better off as a
result of believing on authority in the results
achieved by specialists in a particular branch of
science than they would have been by any pursuit of
intellectual self-sufficiency. This is only the
case, however, because it is less time-consuming to
believe on authority than otherwise and because
authorities are recognisable. I shall consider these
two aspects in turn.

That it is less time-consuming makes the model of
belief on authority generalisable, in a way that the
monologic and dialogic models were not. It is possible for wage-labourers to have rational convictions
founded upon authority, because it is only required
that they have time to hear or read the conclusions
arrived at by the authorities in the various branches
of knowledge in order forthwith to adopt them.

But are authorities always recognisable and how
indeed are they to be recognised? Do they always
exist? It is these latter questions which come to

preoccupy Lewis and J.S. Mill, though they are not the
first to be concerned with them: Condorcet, for
instance, whom Mill admired, had been led to think
about the marks of authority in leading the fight
against Mesmerism in Paris during the l780s, and
raised all the questions which reappear in Lewis and
Mill [21].

Austin writes of the astronomers as people who
(1) have scrutinized the evidence; (2) unanimously
agree that the earth moves round the sun); and (3)
have no conceivable motive to lie or express a
biased conclusion. In his Essay on the Influence of
Authority in Matters of Opinion Lewis develops at
much greater length than does Austin an account of
the marks of authority. He distinguishes four marks
of trustworthy authority, that is, criteria for distinguishing those whose opinions can rationally be
accepted on authority with respect to those subjects
for which they satisfy the criteria. First, says
Lewis, long study or practice of a subject or art is
such a mark and distinguishes ‘all the great luminaries of science, whether mathematical, physical,
metaphysical, ethical or political’ (p.28). Second,
it must be clear that the mental power of the person
whose authoritativeness is being assessed is equal
to the subject in question. (In passing, Lewis
mentions certification as a public mark that these
two criteria are satisfied.)
Third, the putative
authority, ‘ought to be exempt, as far as possible,
from personal interest in the matter’, or if not,
‘his honesty and integrity ought to be such as to
afford a reasonable security against the perversion
of his opinions by views of individual advantage’

(p.28). This view has an historical analogue in
Ancient Greek arguments over whether teachers should
accept payment for their teaching, and historical
precedents in the ideas of Locke, to whom Lewis refers
Fourthly, Lewis proposes a mark of authority which
does not refer to properties of individual subjects,
arguing that ‘with respect to subjects ~f¬∑speculation
and science, the existence of an agreement of the
persons having the above qualifications (the first
three marks above) is the most important matter. If
all the able and honest men who have diligently
studied the subject, or most of them, concur, and if
this consent extends over several successive generations, at an enlightened period, and in all or most
civilised countries, then the authority is at its
greatest’ (p.42), as it is with astronomy. In such
cases, we can feel confident of the rationality of
our believing on the simple authority of what others
say. The justification of the authoritativeness of
this last mark raises extremely interesting epistemological and political issues, discussion of which
occurs in both French and English political theory
[22] .

It is because the utilitarians believed in the
possibility of knowledge of ends, of a prescriptive
moral and political art, that they could adopt an
account of the authority relation between the natural
scientists and the layperson and propose it as a model
for the future organization of the relationship
between humankind and its governors – specifically,
between the working class and political leaders,
whose leadership would be justified in terms of
superior knowledge, and whose authority it would be
literally irrational not to accept. Such a proposal
is at the heart of J.S. Mill’s writings [23].

Of course, a case has to be made out that
people – and in particular the urban working class are not, or need not be, irrational, despite their
apparent predilection for all kinds of quacks and
charlatans, whether religious, medical or political.

Human rationality is seen by Lewis and J.S. Mill as
something which social organisation, and particularly
schooling, can increase: Mill, for example, in the
System o.f Logic defines a science of ethology, corres-

ponding to the art of education, which would study
the formation of character as a branch of general
psychology [24] and in On Liberty he makes numerous
remarks of an ethological character on the conditions
favourable and unfavourable to the production, among
much else, of ‘logical, consistent intellects’ [25].

There are many other loose ends to the theory
which utilitarians were generally aware of and sought
to tidy up. Let me simply note here that the intellectualist account of how political authority relations was essentially based on rational deference to
superior wisdom can clearly be applied in a gesellschaft society, since it does not suppose the existence of intimate, personal ties between those who
know and those who do not. It does not even require
personal contact, since the conclusions of those who
know can be diffused in printed form: free-floating
Mannheimian intellectuals can happily and effectively
operate in a society characterised by personal mobility and anonymity. Here there is again a parallel in
Plato’s and Socrates’ recognition of the changed
social conditions of post-Homeric Greece, and their
response to that. After all, The Republic is the
very first intellectualist theory of government, in
which those who know rule those who do not, though
rulers and ruled remain essentially strangers to
each other [26].

III Liberty, Negative Dialectics, and Rational Authority



The utilitarians firmly believed in the possibility
of knowledge of ends, and specifically in happiness
as the single end against which all actions, policies
and politics should be judged. J.S. Mill’s
Utilitarianism opens with a sentence in which it
is presupposed that the ‘controversy respecting the
criterion of right and wrong’ is decidable, and goes
on to stress the common ground between intuitionists
and utilitarians, for ‘the intuitive school affirm
as strongly as the inductive, that there is a science
of morals’ [27]. But, of course, the utilitarians
were perfectly well aware of the backward state of
both the prescriptive arts and the moral sciences
(excepting political economy) in contrast to that
achieved in the natural sciences (or their image of
them) and in particular aware of the extent of disagreement in the moral and political field. They had
an explanation for this, which was not a new one.

Austin expresses the general view when he says, ‘It
was the opinion of Mr. Locke, and I fully concur in
the opinion, that there is no peculiar uncertainty in
the subject or matter of these [moral] sciences: that
the great and extraordinary difficulties, by which
their advancement is impeded, are extrinsick; are
opposed by sinister interests, or by prejudices which
are the offspring of such interests’ [28]. Mill
shared this view, and even if in On Liberty the day
of eventual consensus about morals and politics tends
to be indefinitely put off, Mill’s work shows a lifelong concern with the pursuit of truth about questions
most philosophers now regard as, ultimately, matters
for decision; and, in addition, his critical reviews
of rival moralists (for instance, Sedgwick and Paley
[29]) include demonstrations of the interests and
prejudices which motivate fallacious arguments.

What is, in my view, not sufficiently appreciated
is the position Mill takes with regard to the methodology he believes to be appropriate for the pursuit
of truth about questions of value.

Though it is never stated in one place [30], Mill
does have a consistent, developed doctrine about the
Method o~ Art, Mill’s name for the study of ends. In
his view, the appropriate method for Art is dialectics, as pioneered in both its positive and negative

parts by Socrates and Plato, and added to by

In the Inaugural Address to the University of St.

Andrew’s there is a clear statement of the crucial
methodological role of dialectics; Mill says:

Human invention has never produced anything so
valuable, in the way both of stimulation and
of discipline to the inquiring intellect as
the dialectics of the ancients …. No modern
writings come near to these, in teaching, both
by precept and example, the way to investigate
the truth, on those subjects, so vastly important to us, which remain matters of controversy,
from the difficulty or impossibility of bringing
them to a directly experimental test … with’

all this vigorous management of the negative
element, they inspire no scepticism about the
reality of truth, or indifference to its
pursuit [31].

Mill said this in 1867; in 1866 he had published a
lengthy essay on Plato in the form of a review of
George r,rote’s Plato and Other Companions of
Sokrates [32], an essay for the writing of which he
prepared himself by re-reading in Greek all of Plato
[33]. In this essay he says of negative dialectics,
the testing of every opinion by negative scrutiny,
that it ‘could only be done effectively by means of
oral discussion …. Its pressure was certain, in an
honest mind, to dissipate the false opinion of knowledge [Plato’S doxa] and make the confuted respondent
sensible of his own ignorance’, adding that negative
dialectics is ‘one branch of an art which is a main
portion of the Art of Living’ [34].

As for positive dialectics this involves precise
definition and description as well as argument for a
doctrine. In Utilitarianism Mill engages in positive
argument in support of the Principle of Utility,
advancing considerations ‘capable of determining the
intellect either to give or withhold its assent to
the doctrine; and this is equivalent to. proof’ [35].

Hany other remarks by Mill on dialectics could be
quoted from the Autobiography; from the late essay on
Aristotle; and, of course, from On Liberty [36]. Mill
rejects, however, the Platonic doctrines of Forms [37]
and this confronts him with the problem of finding an
alternative epistemology to complement his dialectical
(dialogic) methodology, a problem I am not sure that
he resolves. This will become clear in the course of
the remainder of this section.

Dialectics, as the method of pursuing truth about
questions of Art, has distinct social conditions of
possibility; it cannot be practised under all conceivable social circumstances. In particular, its negative element involves one person, A, criticising
(testing by negative scrutiny) the opinions of another,
B, as a result of which these opinions will be shown
either to be mere opinions (doxa) , though B falsely
believed them to be justified true beliefs (that is,
knowledge), or else the opinions will withstand the
dialogic test, and can henceforth be held by B in the
full, rational conviction that they constitute knowledge. But this process of testing is only possible
if A is at liberty to criticise B’s opinions. Dialectics is only possible in conditions of social freedom.

And in Chapter 2 of On Liberty, Hill defends liberty
of thought and expression as a condition of possibility
of the dialectical pursuit of knowledge. At the risk
of repeating the familiar, I want to say a little about

Consider, first of all, how Mill deals with the
objection to liberty of thought and expression that,
‘We may, and must, assume our opinion to be true for
the “guidance of our own conduct: and it is assuming
no more when we forbid bad men to pervert society by
the propagation of opinions which we regard as false
and pernicious’ (p.23l). To this Mill replies:

I answer, that it is assuming very much more.


There is the greatest difference between presuming an opInIon to be true, because with
every opportunity for contesting it, it has
not been refuted, and assuming its truth for
the purposes of not permitting its refutation.

Complete liberty of contradicting and disproving our opinion is the very condition
which Justifies us in assuming its truth for
purposes of action; and on no other terms
can a being with human faculties have any
rational assurance of being right.

Here ‘rational assurance’ has the sense which
‘rational conviction’ had in the second dialogic
model outlined in section II above. In a moment I
shall show how what Mill is saying in On Liberty
relates to the third model of rational conviction
on authority; but first of all let me note that Mill
recognises that consensus or unanimity, which is
a~ter all the goal of the practice of dialectics,
must have disadvantages if what Mill says in the
passage just quoted is so. For once a rational
consensus has been achieved, there will be no one
minded to contradict our opinions, though they are
free to do so. Mill takes the view that this consequence is, indeed, disadvantageous, and looks for
‘some contrivance’ to weaken its impact – a contrivance he finds in the ‘Devil’s Advocate’ use of negative dlalectics [38].

But is everyone to submit to the negative scrutiny
of their opinions, as subjects in a dialogue, as Mill
seems to suggest? In my reading, Mill does not
regard this as necessary [39]. It is quite possible
for some people to take their opinions on the authority of others provided those others themselves practice dialectics (as Mill answers objections all the
way through On Liberty; criticises intuitionists in
numerous places; defends the principle of Utility in
Utilitarianism, and so on). Dialectics is compatible
with an intellectual division of labour, but the
rationality of belief on authority can only be sustained in matters of Art [40] if putative authorities
conduct their activities in conditions of liberty that is, in a communication situation free of domination, to use a Habermasian expression. Much of
Chapter 2 of On Liberty is devoted to the description
of the conditions which must be satisfied for the constitution of such a communication situation. In
addition to the liberty to practise negative dialectics, Mill says that it is also necessary that participants in argument have honest minds; must not be
too timid to state what they really think; and must
adhere to a ‘morality of public discussion’ (p.259).

In all of this, Mill is describing characteristics of
what Habermas would call an ideal speech situation
[41] .

But Habermas is right to argue [42] that Mill
describes a communication situation free of domination not only ‘from above’ (from the State) but also
‘from below’ (from public opinion). His attitude to
liberty is thus ambivalent [43]. For public opinion,
before it has submitted to authoritative guidance, is
not only the domain of unexamined opinions, but also
hostile to the unhindered pursuit of truth, and in
particular to having negative dialectics practised on
its opiniqns [44]. That Mill’s On Liberty is directed as much against public opinion as the State has
always been recognised; to cite a relatively unfamiliar source, Grote in Plato and the other Companions
of Sokrates says of On Liberty that i t ‘ stands almost
alone as an unreserved vindication of the rights of
the searching individual intelligence against the
compression and repression of King Nomos’ [45]. In
this connection, let me also point out that Mill’s
concern with individuality in On Liberty not only
parallels what Pericles singled out as an Athenian
virtue [46] but has a cognitive justification, as

comes out in the following passage:

Originality is not always genius, but genius
is always originality; and a society which
looks jealously and distastefully on original
people – which imposes its common level of
opinion, feeling and conduct, on all its
individual members – may have the satisfaction
of thinking itself very moral and respectable,
but it must do without genius.

In other words, those who can lead in intellectual as
well as political matters will often pursue unorthodox careers and lead unconventional lives, and
society is the loser by not tolerating those aspects
of their individuality. His liaison with Aspasia
does not alter the fact that Pericles is a greater
figure than the mediocre, but respectable, Nicias
[48] .

In sum, the Few whose business it is to search for
knowledge of Ends, as part of the division of intellectual labour, must practise – and be free to
practise – negative dialectics both upon each other’s
opinions, if trustworthy authorities are to be
formed, and upon public opinion, in pursuit of the
goal of a rational society and political order.

Theirs is a difficult task, the nature of which is
jointly defined by a conception of a political order
sustained by voluntary deference to superior wisdom
– rational conviction founded upon authority – and
a dialectical theory of the pursuit of truth, and
specifically the truths of Art.

The consequences of Mill’s Greek-inspired dialectics are quite dif~erent from the consequences of the
positivism of a Comte. The following passage from
Auguste Comte and Positivism brings together many of
the topics and arguments I have considered so far in
this article and shows how Mill is able to distance
himself from Comte:

It is, without doubt, the necessary condition
of mankind to receive most of their opinions
on the authority of those who have specially
studied the matters to which they relate. The
wisest can have no other rule, on subjects with
which they are not themselves thoroughly conversant; and the mass of mankind have always
done the like on all the great subjects of
thought and conduct, acting with implicit
confidence on opinions of which they did not
know, and were often incapable of understanding,
the grounds, but on which as long as their
natural guides were unanimous they fully relied,
growing uncertain and sceptical only when these
became divided, and teachers who as far as they
could judge were equally competent, professed
contradictory opinions … in order that this
salutary ascendancy over opinion should be
exercised by the most eminent thinkers it is
not necessary that they should be associated
and organized. The ascendancy will come of
itself when the unanimity is attained, without
which it is neither desirable nor possible.

It is because astronomers agree in their
teaching that astronomy is trusted, and not
because there is an Academy of Sciences or a
Royal Society issuing decrees or passing

[ 49]
Now in isolation, the last sentence of this quotation is compatible with the idea that scientists
could ‘agree in their teaching’, though working monologically. Each would arrive independently at the
same results, and the epistemological worth of this
would be assessable in Condorcetian terms. However,
it is clear that it is not Hill’s intention to sustain that position, since he extends the dialogic
process to the sciences. Furthermore, there is evidence that Mill took the view that dialogue was

necessary not only to testing an opInIon, but even
for formulatinq it. If this is the case, Mill has
as m~ch in co~on with C.S. Peirce and Josiah Royce
as he has with positivism. Consider the following

Why is it, then, that there is on the whole a
preponderance among mankind of rational opinions
and rational conduct?

it is owing to a
quality of the human mind, the source of everything respectable in man either as an intellectual or as a moral being, namely, that his errors
are corrigible. He is capable of rectifying
his mistakes, by discussion and experience.

Not by experience alone. There must be discussion, to show how experience is to be

The stress on corrigibility would be agreeable to
Popper. The idea that experience requires interpretation through discussion, however, suggests the
necessity of a realist, rather than a phenomenalistempiricist ontology and epistemology, which distinguishes appearance and reality, and makes the gaining
of access to the latter a collective enterprise.

The standard of truth is a reality which is revealed
through a social practice.

But though true statements of science are those
which correspond to a reality dialogically apprehended, to what could true statements of ends correspond, if to anything at all? Mill rejects the view
that they correspond to Platonic forms, and is often
thought to have accepted the view that they correspond to (are validated by) the same reality which
validates scientific statements. He is then accused
of committing a naturalistic fallacy in attempting to
derive the truth of ‘Happiness is alone desirable’ as
an entailment of the truth of ‘Happiness is alone
desired’ .

There is another possibility. There is in Mill,
and specifically in Utilitarianism, a partially acknowledged consensus theory of truth in which consensus (about such propositions as ‘Happiness alone is
desirable’) in a communication situation free of
domination would itself constitute a standard of
truth. In Habermas’ language, there is a suggestion
that Mill believes normative validity claims to be
discursively redeemable and that their successful
discursive redemption is the standard or criterion
of truth.

There is one way, and a plausible way, of reading
Mill’s statement that though the principle of utility
is not capable of ‘proof in the ordinary and popular
meaning of the term’ (p.207), nonetheless ‘the subject is within the cognisance of the rational faculty;
and neither does that faculty deal with it solely in
the way of intuition. Considerations may be presented
capable of determining the intellect either to give or
withold assent to the doctrine; and this is equivalent
to proof’ (p.208). ‘The intellect’ here is the
intellect of the ‘thoughtful reader’ (-.239) to whom
Utilitarianism is addressed. The giving of assent
by readers to the author is ‘equivalent to proof’

provided that the author does not mislead or seduce
by rhetorical devices; and the reader is not blinded
by prejudices; and that the readers agree among themselves.

If such a rational consensus could be got, on what
grounds could it be said that for the principle of
utility ‘the equivalent of proof’ had not been
offered? If it was said that there can be consensus
in error as well as truth, this does not distinguish
the case~rom that of natural science, where the consensus of astronomers, even in a communication situation free of domination, does not prevent them from
being wrong. Nor does the general possibility of
error, without which claims to knowledge would not be
claims to knowledge, justify a general or particular

scepticism. Coming into a community of persons among
whom a rational consensus on some principle prevailed,
it would be irrational for the new individual to
reject the authoritativeness of that consensus unless
he or she could advance some evidence or argument
against the principle or could show that the consensus
was a constrained, not a free one. Otherwise, the
consensus is in itself authoritative, questions of
correspondence are irrelevant, and it would be irrational not to accept the consensus as authoritative.

Where in the case of science, it is reality which is
the source of a continuity which permits later scientific theories to be seen as revisions and corrections
of earlier ones, in the case of knowledge of ends it
is the continuity of the community itself which allows
the pursuit of such knowledge to be a rational and
potentially cumulative enterprise. A community is
linked to its past judgments by memory, and to the
future consequences of its present judgments by hope,
a hope which can be dashed if the consensual judgment
was a false one [51].

However, though in Utilitarianism Mill reaches out
to a wider community than that of the Few [52], belie~
ing (like Austin before him) that the very first
principle of morals and legislation is one which
citizens can come to know philosophically, and not
just accept on authority, nonetheless the distinction
in Utilitarianism between higher and lower pleasures
seals the division between the sphere of equality and
the sphere of inequality within which citizens of a
unitarian polity should consent to operate.

Utilitarianism underpins the claim that the Many
should allow themselves to be guided by the ‘counsels
and influence of a more highly gifted and instructed
One or Few’ [53].

IV The Passing of Utilitarianism
In conclusion, I should like very brie~ly -to consider
the fate of the Utilitarian theories of liberty and
authority which I have sketched out in this article.

They are theories likely to have some attraction for
contemporary’ advocates of Political Education
programmes, despite the general prejudice that
‘Anyone who still discusses the admissibility of
truth in practical [value] questions is, at best,
old-fashione.d’ [54]. But in terms of the historical
development of British political organisation, the
Utilitarians failed. Alternatives to rational conviction on authority were found in the party organizations, the popular press, and the continuation of a
deference politics based on a projected image of a
distant monarch, aristocracy and plutocracy, rather
than on any personal contact with them.

Among nineteenth century political writers it was
Bagehot in The English Constitution (written in response to Mill’s Representative Government) and
especially in the pre~ace to the second edition of
1872 [55], who emphasised the affective rather than
cognitive, irrational rather than rational, elements
of political life and their compatibility with
political order.

. The success in practice of an alternative to the
utilitarian vision of a rational society is marked by
the abandonment of the epistemological study of the
marks of science and charlatanism in favour of a
psychology of crowds. The Enlightenment notion of
educating the lower orders out of their credulousness
as an essential moment of social progress gives way
to the use of their credulity for reasons of state.

Equally, as liberty of thought and expression loses
its structural, political functions, tolerance becomes
no more than everyone’s entitlement to their own
opinion [56], an entitlement consecrated by twentiethcentury positivism, a way of thinking with which John
Stuart Mill has very little in common.












This article is loosely based on Chapter 3 of my unpublished M.Phil. thesis,
How is Political KnalJledge Possible? (University of Sussex, 1978). Roy
Edgley has been a constant source of critical encouragement; I have al so
made use of written comments on the draft of this article furnished by
Stefan Collini, Ben Gibbs, William Outhwaite, Hellmut Pappt’! and Carole

[J.S. Mill], ‘The Spirit of the Age’, The Examiner>, number 2,23 January

G.C. Lewis, An Essay on the Influence of Author>ity in Matter>s of Opinion,
London, J. Parker, 1849.

R.B. Friedman, ‘An Introduction to Mill’s Theory of Authority’, in J.

Schneewind, ed., Mill, London, Macmillan, 1969, pp. 379-425.

For an overview, see H.O. Pappt’!, ‘The English Utilitarians and Athenian
Democracy’, in R. Bolgar (ed.), Classical Influences on Wester>n Thought
A.D. 1650-1870 (Cambridge University Press, 1978), pp.295-307. Also
valuable is F. Sparshott, ‘Introduction’, in J.S. Hill, Essays on Philosophy
and the Classics, Ed. J.M. Robson, Toronto, University of Toronto Press,
1978), pp.vii-lxxv.

[James Mill], ‘The Ballot’, Westminster> Review, vol.XIll [1830], pp.1-39.

[James Mill], ‘Government’ [1820], in Essays on Gove1’nlllent, JUl’ispr>udence,
Liber>ty of the FTess, and Law of Nations, London, J. Innes, 1825; New York,
A.H. Kelley [reprint edition], 1967, pp.3l-32.

For a recent discussion, see E. Wood and N. ‘Vood, Class Ideology and
Ancient Political Theor>y (Oxford, Basil Blackwell, 1978). J.S. Mill
contrasts the two situations in his 1846 review of Grote’ s Histor>y of
Gr>eece.Thus, ‘the sense of obligation in the Homeric period is exclusively of a personal kind’, whereas ‘in the conceptions of the citizens of
historical Athens [quoting Grote now] “the great impersonal authority
called the Laws stood out separately both as guide and sanction, distinct
from religious duty or private sympathies”.’ Ouoted from J.S. Mill,
‘Grote’s History of Greece [Il’, in Mill, Essays on Philosophy and the
Classics, p. 297.

This term (and its twin, dialogic) is suggested for the work of J. Habermas,
but I do not aim at fidel i ty to his use of it, for which see J. Habermas,
Knowledge and Human Inter>ests (London, Heinemann, 1972).

J. Locke, An Essay Concer>ning Human Under>standing [1960], ed. J.W. Yolton,
two vols.,(London, Dent, 1961);. J. Locke, Of the Conduct of the Under>standing [1706], ed. F.W. Garforth (New York, Teachers College Press, 1966).

See the encomium in [James Mill], review of ‘Essays on the Formation and
Pub 1 ication of Opinions’ [by S . Bailey], Westminster> Review, vol. VI [1826],

Compare J. Habermas, Str>uktUl’lJandel der> tJffentZichkeit (Neuwied: Luchterhand,
1962; French translation, L ‘Espace public, Paris, Payot, 1978); J. Habermas,
Legiti=tion Cr>isis (London, Heinemann, 1976)
For relevant discussion, see R.J. Bernstein, ‘Hannah Arendt: Opinion and
Judgement’, unpublished paper for the American Political Science Association
Annual Meeting, 1976; J. Habermas, ‘Hannah Arendt’s Communications Concept
of Power’, Social Resear>ch, vol.44 (1977), pp.3-24.

On Peirce, see, for instance, J .E. Smith, ‘Community and Reality’, in R.J.

Bernstein (ed.), Per>spectives on Peir>ce (New Haven, Yale University Press,
1965), pp.92-119. Very relevant to the present article is J. Royce, The
FToblem of Chr>istianity, vol.2 (New York: Archon Books reprint edition,
1967) .

J. Austin, The FTovince of JUl’ispr>udence Determined [1832], ed. H.L.A. Hart
(London, Weidenfeld and Nicolson, 1954), p.60.

Mill, Spir>it of the Age, p.5l.

ibid., p.52.

Locke, Conduct of the Under>standing, p.1l6.

Austin, FTovince of JUr>ispr>udence, p.63.


Cornewall, Lewis and J.S. ~lill attended the lectures on which the book is
based. See R. Friedman, op.cit., p.4l9 and pp.422-23.

See R. Darnton, Mesmer>ism and the End of the Enlightenment in Fr>ance
(Cambridge, Hass., Harvard University Press, 1968), esp. pp.189-92.

The co-text for Lewis’ s discussion of this question is Condorcet, Essai sur>
l ‘application de l ‘analyse Cl la pr>obabiliU des decisions r>endues Cl la
plUl’aliU des voix (Paris, L’Imprimerie Royale, 1785), mediated to Lewis
through the popular work of Laplace, Essai philosophique sur> les probabilites (many editions from 1814). See Lewis, Influence of Author>ity in
Matter>s of Opinion, p.2l6. I discuss Condorcet’s argument in Chapter 2 of
Pateman, How is Political Knowledge Possible? See also K. Baker, ConMr>cet:

from Natu:rol Philosophy to Social Mathe=tics (Chicago, Uni versi ty of
Chicago Press, 1975); G. Granger, La Mathematique sociale du Mar>quis de
Condor>cet (PariS, Presses Universitaires de France, 1956); 1. White,
‘Condorcet: Politics and Reason’, in S.C. Brown (ed.), Philosophy of the
Enlightenment (Brighton, Harvester Press, 1979); and Section I Il of the
present article.

J.S. Hill, A System of Logic [1843]’ two volumes, edited by J.M. Robson
(Toronto, University of Toronto Press, 1974), p.869.

See, ‘most obviously, J.S. ~lill, Consider>ations on Repr>esentative Gove1’nlllent,
[1861]’ in J.S. Mill, Essays on Politics and Society, edited by J.M. Robson
(Toronto, University of Toronto Press, 1977), pp.371-577. There is an
excellent discussion of this work in D. Thompson, John Stuar>t Mill and
Repr>esentative Gove1’nlllent (Princeton, Princeton University Press, 1976).

J .S. Mill, On Liber>ty [1859], in Hill, Essays on Politics and Society, p.242.

Plato, The Republic, in Collected Dialogues of Plato, pp.575-844. Benjamin
Gibbs says that my remark is ‘fair comment’, but draws attention to Plato,
The Republic, pp.462b-464a.

J .S. Mill, Utilital’ianism [1861]’ in J .S. Mill, Essays on Ethics, Religion
and Society, edited by J .M. Robson (Toronto, University of Toronto Press,
1969), pp.205 and 206.

Austin, FTovince of Jur>ispr>udence, p.74.









J.S. Mill, ‘Sedgwick’s Discourse’ [1835], in J.S. Mill, Essays on Ethics,
Religion and Society, pp. 31-74.

It is not discussed in Book VI of The System of Logic, where you would
expect to find it. As Roy Edgley pointed out to me, it is surprising that
Mill does not discuss in Book VI, Chapter XII what he says he is going to
discuss. He says at the beginning of the chapter that: ‘The Method,
therefore, of Ethics can be no other than that of Art, or Practice, in
general: and the portion yet uncompleted, of the task which we proposed to
ourselves in the concluding Book is to characterise the general Method of
Art, as distinguished from Science’ (p.943). A few pages on, however, he
says that: ‘The theory of the foundations of morality is a subject which
it would be out of place, in a work 1 ike this, to discuss at large’ (p. 951) ,
and that he will not attempt to justify his utilitarian position ‘or even
to define the kind of justification which it admits of’ (p.951). Since the
Method of Art is inseparable from the justification of ends, Mill’s procedure
is, to say the least, surprising. However, a footnote to the above passage
added in the 1865 edition of the System of Log1:C says: ‘For an express
discussion and vindication of this principle [of utility], see the little
volume entitled Uti litar>iani sm, (p.95l).

()uoted from J. Schneewind, editor, Mill’s Essays on Liter>atUl’e and Society
(New York, Collier Books, 1965), pp.370-71.

J. S. Mi 11, ‘Grote’ s Plato’ [1866], in Mi 11, Essays on philosophy and the
Classics, pp.375-440. G. Grote, Plato and Other> Companions of sokl’ates,
three volumes (London, John Hurray, 1865).

F. Sparshott, ‘Introduction’, in Mill, Essays on Philosophy and the Classics,

Mill, ‘Grote’s Plato’, p.405. See also p.4l0.

Mill, Utilit=ianism, p.208.

J.S. Mill, Autobiogr>aphy, in Essential ~lor>ks of John Stuart Mill, edited by
~1. Lerner (New York, Bantam Books, 1965), especially p.2lf.; J .S. Mill,
‘Grote’s Aristotle’ [1873], in Mill, Essays on Philosophy and the Classics,
especially pp.505-5l0; Mill, On Liber>ty, especially pp.250-52.

As is clear in Mill, ‘Grote’s Plato’, pp.4l2-l3.

See Mill, On Liberty, p.232 and p.251. Also Mill, Autobiogr>aphy, p.2l;
Mill, Ar>istotZe, pp.509-l0.

See T. Pateman, Language, Truth and Politics, second edition (Lewes, Jean
Stroud, 1980), Chapter VI.

He says the same of science: ‘truth, in everything but mathematics, is not
a single but a double question; not what can be said for an opinion, but
whether more can be said for it than against it. There is no knowledge,
and no assurance of rip,ht belief, but with him who can both confute the
opposi te opinion, and successfullY defend his own against confutation’

(Mill, ‘Grote’ s Plato’, p.4ll).

For a fuller discussion, see Pateman, HOlJ is Political KnOlJledge possible?

Chapter 3.

Habermas, Struktu:rwandel deI’ tJffentZichkeit, Ch. IV, section 15.

As is his attitude to the Athenian demos and to the Sophists. See, for
example, J.S. Mill, ‘Grote’s History of Greece [Il]’ [1853], in Mill,
Essays on Philosophy and the Classics, pp.307-337, especially p.3l6 and
p.329. I exclude all discussion of and reference to Mill’s relation to
de Tocqueville here, both for reasons of space and because of my main aim
of indicating the ‘Greek connexion’ of Mill’s political and epistemological

Compare the following:

‘Socrates, in morals, is conceived by Mr Grote as the parallel of
Bacon in physics. He exposed the loose, vague and misleading character
of the common not ions of mankind on the most famil iar subjects ….

It is the fashion of the present day to decry negative dialectics; as
if making men conscious of their ignorance were not the first step and an absolutely necessary one – towards inducing them to acquire
knowledge…. The common notions of the present tillle on mental and
moral matters are as incapable of supporting the Socratic crossexamination as those of his own age: they are, just as much, the
wild fruits of undisciplined understanding’ (Mill, ‘Grote’ s History
of Greece [ll], pp.309-l0. This passage was added in the 1859 version.)
G. Grote, Plato and Other Companions of Sokl’ates, new edition (London, John
Murray, 1888), Volume 1, p.395. The quotation is from the chapter in which
Grote discusses dialectics
See Hill, ‘Grote’s History of Greece [Il], pp.3l8-20.

ibid., pp.320-2l.

ibid., pp.333-36.

J. S. Mill, Auguste Comte and Positivism, in Mill, E3says on Ethics,
Religion and Society, pp.3l3-l4.

Mill, On Liber>ty, p.23l. See also J.S. mll, FTinciples of Political
Econonry [1848], Books Ill-V, edited by J.M. Robson (Toronto, University
of Toronto Press, 1965), Book V, Chapter XI, especially section 8 (drawn
to my attention by Hellmut PappE!); and, of course, the argument about
higher and lower pleasures in Mill, Utilitarianism. Compare Royce, ‘I
cannot find or even define the truth in terms of my individual experience,
wi thout taking account of my relation to the community of those who know.’

(Royce, FTobZem of Christianity, p.3l2).

These ideas are based on FTobZem of Chr>istianity. In a talk to the
University of Sussex Philosophy Society, October 1980, Max Black illustrated
the claims of a wholehearted or critical rationalism by pointing out that
the real continuity of the Supreme Court makes it possible genuinely to
progress from the decision of 1950 to that of 1960, and so on. Compare
R. Dworkin, Taking Rights Ser>iousZy (London, Duckworth, 1977).

Utilit=ianism originally appeared as a series of articles in Fr>aser’s
Magazine for 1861.

Mill, On Liberty, p.269.

Habermas, Legitimation Crisis, p.123. The criticism has been aimed at
Habermas, notably by Niklas Luhmann.

W. Bagehot, The English Constitution [1867; 1872], introduced by R.H.S.

Crossman (London, Fontana/Collins, 1963).

T. Pateman, Language, Truth and Politics, Chapters I and VI.

Download the PDFBuy the latest issue