POLITICAL THEORY and POLITICAL PRACTICE

Gordon Graham

In 1962 Isaiah Berlin published an essay entitled 'Does Political Theory Still Exist?'. This
appeared in the second volume of the series , Philosophy, Politics and Society and the question
which forms its title was prompted in part by a remark in the introduction to the first volume in
the same series that "for the time being political philosophy is dead". Berlin considers one
version of this claim; that political theory might finally have given way to political science in
something like the way astrology, say, came to be replaced by astronomy, Could empirical
inquiry into the actual workings of politics render redundant the general interest in political
ideas and conceptions that we find in a long list of thinkers from Plato to Mill? I am not here
directly concerned with the issue of political theory versus political science, but with a closely
related question: Is political theory normative? Since what Berlin has to say on the first issue
throws some light on the second, we will have occasion to return to his essay. Berlin is clear as
to why the future of political theory was at that time thought to be uncertain.

The principal symptom which seems to support this belief is that no

commanding work of political philosophy has appeared in the twentieth century.
By a commanding work in the field of general ideas I mean at least one that has
in a large area converted paradoxes into platitudes or vice versa. (The Proper
Study of Mankind p.59)

But, arguably, he wrote this too soon, because less than ten years later, in 1971, there
appeared what in retrospect does seem to be the commanding work of this century -- John
Rawls's A Theory of Justice. Few books in any period, in fact, have dominated the political
philosophy of their day quite so powerfully Rawls's Theory of Justice. Moreover, there is reason
to think that the enthusiastic reception it met with is to be explained partly by the fact that it
was perceived to be a work whose very existence refuted the doubts that Berlin addresses. A
Theory of Justice seemed to overturn the contention that political philosophy was dead. We

1
could say, in fact, that it converted the paradoxically normative character of traditional political
theory into a platitude. If someone were to suggest, post 1971, that political philosophy was
dead, the response would be 'Of course it isn't! Look at Rawls.' Yet the existence or non-
existence of normative political philosophy is not quite the same issue as the existence or non-
existence of political theory because the possibility or otherwise of normative political
philosophy has been a topic within political theory, and a subject upon which many of the best
known political theorists divide. So, even if we allow that political theory exists, we are left with
the question "Does it have any normative import?"
This is a question that may be divided into two parts. The first concerns the relation
between a philosophical education and political effectiveness, and the second to the way, if
any, in which political theorising feeds into political practice. To some minds, there is at least no
doubt on the first point. Here is Bishop Berkeley on the subject.

Those great men, Pythagoras, Plato and Aristotle, the most consummate in
politics, who founded states, or instructed princes, or wrote most accurately on
public government, were at the same time most acute at all abstract and sublime
speculations, the clearest light being ever necessary to guide the most important
action. And, whatever the world think, he who hath not much mediated upon
God, the human mind and the summum bonum may possibly make a thriving
earthworm, but will most indubitably make a sorry patriot and a sorry
statesman. (Siris 350)

In answer to the second point, political theorists have differed widely. Some are as clear
about it as Berkeley is on the earlier one. Coleridge, for example, says:

I have known men, who, with significant nods and the pitying contempt of
smiles, have denied all influence to the corruptions of moral and political
philosophy, and with much solemnity have proceeded to solve the riddle of the
French revolution by anecdotes! Yet it would not be difficult, by an unbroken

2
 chain of historic facts, to demonstrate that the most important changes in the
commercial relations of the world had their origins in the closets or lonely walks
of uninterested theorists. (The Statesman's Manual London, 1816 p.15)

I shall call the Berkeley/Coleridge position, political intellectualism. Its high priest is
probably Plato, who first argued that a training in mathematics and then in philosophy was an
essential route to political wisdom for the statesman. This is a position that may be sharply
contrasted with one I shall label 'political pragmatism', a conception whose high priest is usually
thought to be Machiavelli. Political pragmatism has a no less venerable history than political
intellectualism. It, however, attributes no importance whatever to visionary ideas, but sees the
foundation of political action as lying in the pursuit and exercise of power in accordance with
fairly low level motivation. Though a ruthless political pragmatism is indeed regularly associated
Machiavelli, his writings have in fact some decidedly intellectualist leanings and a more
unmistakable version of pure pragmatism is to be found in Henry St John, Viscount Bolingbroke,
author of The Idea of Patriot King. Bolingbroke's description of the Tory Ministry appointed by
Queen Anne in 1710, of which he was a part, runs as follows.

. . . as all parties have done . . . the principal spring of our actions was to have the
government of the State in our hands; . . . our principal views were the
preservation of this power, great employments to ourselves, and great
opportunities of rewarding those who had helped to raise us. It is true, that with
these considerations of private and party interest there were others
intermingled, which had for their object, the public good of the nation - at least
what we took to be such. (Works of Bolingbroke 'Letter to Sir William Windham',
Philadelphia 1841 I pp. 114-115)

From such a point of view, what goes on 'in the closets or lonely walks of uninterested
theorists' is of no political consequence, and to the pragmatists, this fact is no less obvious than

3
the power of ideas is evident to Coleridge. Thus, Elie Kedourie, commenting on the theorists of
nationalism says:

It is absurd to think that professors of linguistics and collectors of folklore can do

the work of statesmen and soldiers . . He who exercises power exercises it while
he can and as he can and if he ceases to exercise power he ceases to rule.
Academic research does not add a jot or tittle to the capacity for ruling, and to
pretend otherwise is to hide with equivocation what is a very clear matter.
(Nationalism 2nd edition London 1961 p.125)

Who is right about this? In this paper I want to attempt an adjudication between
intellectualists and pragmatists about politics, and this, as we shall see, will eventually bring us
back to Berlin on political theory.
I

One intuitively appealing picture of political action is this: the political theorist,
philosopher or visionary formulates the ends of politics and the skilled politician devises
practicable means of bringing these about. Whatever the merits of dividing political activity into
the formulating of ends and the devising of means, we can note straight away that, contrary to
a common assumption, this analysis actually assigns no special role to political theorising.
Suppose the task, and the skill, of the politician is indeed devising means to ends. There is no
obvious reason why those ends should be given by a political philosophy. As the quotation from
Bolingbroke illustrates, the ends of politics can be set by the private and party interests of those
who engage in it. Or by the demands of their supporters, and indeed this is the natural way to
employ the ends/means account in the modern world. Political ends, the democrat holds,
should be determined by what the people want. What the protagonist of normative political
philosophy needs to show, therefore, is that the ends set by political theory are in some sense
superior to those set by mere wants, whether these wants are the desires of politicians or the
people at large.

4
 How might this be shown? There seem to be two possible criteria of the superiority of
theoretically inspired ends -- either they are more likely to be successful, or they necessarily
have greater intrinsic merit. But consider each of these in turn. If we take success to be the
criterion, we have compounded the distinction between ends and means with which we began.
Success is a test of means. The successful politician is one who achieves what he or she sets out
to, and this test applies even if the aim is nothing more than remaining in power. Consequently,
successfulness cannot be a criterion by which we might distinguish some kinds of ends from
others. Of course, it is true that the chances of success in any given endeavour may be a
function of the clarity or coherence of the end in view, and that taking greater thought might
enable us to secure these virtues in the ends we pursue, but this applies directly to the aims of
the pure pragmatist. The pursuit or promotion of desire need not be blind. From this it follows
that success, even the success which results from greater clarity of thought, cannot be the mark
by which a preference for the ends political theories devise is to be justified.
The alternative criterion of superiority is intrinsic merit or worthiness. Bolingbroke
refers to the 'public good of the nation' as an end which transcends the preservation of power
and the acquisition of 'great employments' to those who exercise it. Suppose the public good of
the nation is indeed a worthier aim than mere party advantage. Is it in fact an alternative end,
properly so called, or is it not rather a normative standard by which specific political ends are to
be judged? There seems greater reason to interpret it in this second way because larger
conceptions of what is good in politics generally transcend the simple pursuit of political self-
interest not by the introduction of opposing 'higher' ends, but by the application of political
principles. For example, a self-serving course of action might be condemned as undemocratic or
a partisan policy criticised as illiberal. Both are clearly judgements of unworthiness, but the
unworthiness lies in the violation of principle, not the failure to pursue some alternative end. In
short, the ruthless pursuit of party interest is usually condemned, not because it has the wrong
goals, but because it pursues them in objectionable ways.
It seems that the simple means/end picture by which we might initially try to connect
the theoretical and the practical in politics is inadequate. Its inadequacy arises, at least in part,
from a general misconception of the relation between means and ends. Means and ends are

5
not two distinct types of entity, and hence cannot be distinct spheres of concern. Any action or
policy can be both means and end, though not at the same time. What is an end at one point in
a sequence may be a means at the next. For instance, my taking a job might be the means to
the end of raising a sum of money, and possessing the sum of money a means to the end of
buying a bicycle, which in turn is a means to entering the Tour de France; and so on. If this is
correct, means/end cannot characterise the relation between the theoretical and the practical.
If one and the same thing can be both a means and an end, it cannot be the exclusive concern
of theory or practice, and the distinctiveness of theory and practice cannot, accordingly, we
characterised in terms of their objects. We need, therefore, to start thinking about the the
relation between political theory and political practice afresh.

II

Not everything about the simple means/end model is mistaken however, which is why it
has an initial plausibility. Some normative political theories can indeed be construed as
attempts to formulate generally validated aims in politics, and when this is so the practical
activities -- leading popular campaigns, seizing power by force of arms or passing legislation, for
example -- can be understood as means to pursuing these ends. But even where the
ends/means model does apply the important point to notice is the reference to general
validation of the ends in view. What distinguishes ends that are validated in this way, we may
suppose, is that they have a justification that self-interested or partisan ends do not have. On
what does this justification rest?
One possibility is rationality. The point of engaging in political theory, one might
imagine, is to arrive at truth and coherence. By this reckoning, a course of action is justified if it
rests upon an explanation which is validly derived from a combination of indubitable first
principles and true propositions. This, I take it, is what Hobbes was aiming at. There is of course
a quite general doubt about whether perfect rationality in this sense is possible, and whether,
even if it is, Hobbes (or anyone else) has ever reached it. Such sceptical doubts are familiar and
important, but even if we discount them, two further problems arise. First, there is a question

6
about the power of reason to govern action. Prescriptions refer to a future yet to be made, and
cannot in themselves, it seems, determine how it shall be. Hobbes thinks he has formulated a
theory which shows the rationality of absolute sovereignty. Suppose he has. This will not by
itself bring about the existence of any absolute sovereign.
This gap between present prescription and future events is the one which Hume exploits
to support his doctrine that reason is the slave of the passions. Even if a political theory can be
devised which is rational in every particular, its power to govern political activity depends upon
actual agents being motivated exclusively by pure practical reason, and this is what Hume
questions. Secondly, and more importantly for present purposes, even if we concede the
possibility of a perfectly rational political theory and complete rational motivation on the part
of those who subscribe to it, there seems to be a further gap between rationality and efficacy.
Why should perfect rationality guarantee success? Coleridge believes, somewhat implausibly,
that "it would not be difficult, by an unbroken chain of historic facts, to demonstrate that the
most important changes in the commercial relations of the world had their origins in the closets
or lonely walks of uninterested theorists". Even if, despite its implausibility, this is true it does
not follow that the impact of these ideas was a result of their greater rationality. For all
Coleridge says to the contrary, the maddest ideas of isolated theorists could have caused the
most important changes.
It is hard to see how, even in the abstract, this second problem could be solved. Given
that the world of politics is peopled by forces of darkness as well as forces of light, the
possibility of wilfulness, wickedness and error renders it impossible to guarantee that perfectly
rational schemes of government, if such there can be, will also be effective enough to prevail.
An alternative basis for the preference of theory over pure pragmatism is public
consensus. It is here that the Rawlsian revival of political philosphy has been most influential.
Rawls attempts to uncover an 'overlapping consensus' composed of principles which actually
underlie the basis of common political activity. Such an overlapping consensus is historically
contingent rather than abstractly rational, but this is precisely its strength, because a political
philosophy based upon it can command assent among otherwise competing political agents
without appeal to contentious conceptions of rationality. A political theory of this kind

7
continues to allow for a good deal of rivalry between specific policies, certainly, but according
to Rawls it provides a philosophical foundation for the design of basic institutions.
The consensual, as opposed to rational, model of political philosophy, has this to be said
for it: it provides some sort of link between theoretical adequacy and political efficacy. In a
modern populist democracy, political programmes derived from a theory which actually
commands underlying agreement are more likely to meet with success. The connection is not
certain, of course, because rival political groups who do as a matter of fact share a good deal of
common ground may fail to see this, or perversely they may prefer party interest. But even if
we concede this strength to the consensual account, there is a further major problem, which
afflicts both it and the rational model, and in a way that is fatal. Or so I shall argue.

III

Political action, and hence political planning, is highly subject to contingency. This in
itself is not a peculiarity of politics, because there is a plain sense in which all action and
planning is subject to contingency. How could it be otherwise, since nothing that happens in
time is necessary? This might be disputed by an absolute determinist who would contend, on
the contrary, that if we knew the full causal story, we would see that everything is determined
and nothing is an accident. However, as Kant argued persuasively, from the point of view of
deliberative action, metaphysical determinism, even if true, is irrelevant. Suppose I accept that
all my desires and choices are determined. Faced with the dinner menu, I still have to make a
choice, and nothing in my deterministic beliefs will make it for me.
To say that action is subject to contingency can be to say no more than that the best
laid plans can go astray. I propose to travel to France, for instance, and I make my plans
accordingly. On the day I intend to leave, there is a rail strike, or a violent storm, or I break my
leg, and so my trip never comes off. Whether absolute determinism is true or not, such
disruptive events are not usually predictable. Even if they were, this would do nothing to alter
their fortuitous nature, because their predictability does not put them any more under our
control. This is a point of some importance to the present topic. It shows that, though well-

8
formed intention may be a necessary condition for effective action, it is not sufficient, and in
politics this problem of contingency is compounded by scale. Political action takes place in the
public realm, and its scope is for the most part medium to long term. In other words, a political
policy or programme is intended to govern the actions of large numbers of individuals over
some considerable period of time. Take, for example, a policy intended to realise universal
primary school education. The immediate action may be the securing of a vote in some
assembly, but the efficacy of the policy requires co-ordinated and sustained action by very
many individuals over a considerable period of time. We might think of the implementation of
such a policy, in fact, as consisting in two very large chains of actions and decisions, one
spreading outwards through the agency of individuals, the other spreading over time through
consequence and response. Every connection in each chain is subject to contingency such that
we can only expect any given action to be followed by its intended outcome with a certain
probability.
This is, of course, stating the obvious, but it has an interesting consequence. Such chains
of actions and events are called Markov chains (after the 20th century Russian mathematician
Andrei Markov). That is to say, to discover the likelihood that the last event in the intended
sequence will actually occur, given the first, we multiply the probabilities. Now, given that the
probability of one event's being followed by its intended outcome is less than 1, and that this is
true of every link in the chain, it does not take many steps before we reach a cumulative
probability of less than 0.5. What this means is that, the longer or more extensive the chain of
actions and events, the less likely it is that the intention with which it began will be realised,
and that for many such chains we can say in advance of embarking on them that it is more likely
they will not produce the intended result than that they will. And to embark on a course of
action of which we can say in advance, not merely that it is unlikely to succeed, but that it is
likely not to succeed, is the mark of irrational action.
If this is correct, we should conclude that given the public and long range nature of
political action on the one hand, and a general subjection to contingency on the other, there is
good reason to eschew what Samuel Johnson called 'schemes of political improvement'. But
what are the theories of the political philosophers, whether rational or consensual, but just

9
such large scale schemes? In this way, pace Coleridge, the ruminations of isolated theorists,
however rational or likely to receive general consent, become as nothing worth in the world of
political practice.

IV

The argument which invokes the idea of a Markov chain is a powerful one, but there is a
danger that it is too powerful. Since any action which comprises a number of stages could be
construed in this way, the argument seems to imply a wholly global scepticism about sustained
and collective action which simply must be unwarranted. One can imagine, for instance, the
conduct of a war being analysed in this way, with the result that we can say, in advance, of any
large scale military action that it is likely not to succeed. But we know in fact that war can be
chosen as an instrument of policy and fought to good effect, and from this it surely follows that
there is something wrong with the Markov chain argument.
Now it is true that there is a danger of excessive scepticism here, but such scepticism is
not borne out by the Markov chain argument. Rather, the argument merely provides an
analytical expression of the scepticism. What the argument itself tells us is that long enough
chains of action intended to produce large enough schemes of change cannot be undertaken
rationally, if we hold that anticipated marginal probability of success is a necessary condition of
rational action. What the argument does not tell us is what the standard of 'large enough' and
'long enough' should be.
It would probably not be profitable, however, to be more precise in this matter. It is
sufficient to note that the political theories and ideological aspirations that Coleridge has in
mind, that is to say, theories drawn from sources other than the detail of practical politics and
conceived at high levels of abstraction, almost certainly fall into this category. To seek to bring
about the communist revolution, the anarchist ideal or even a Rawlsian system of political
justice is to aspire to a goal of such a long term and large scale nature that it will inevitably fall
foul of the argument.

10
 If this is true, does it show that political intellectualism is false, and correspondingly,
that pragmatism is true? To answer this question something more needs to be said about just
how we should understand political pragmatism. In so far as etymology is any guide, the
distinction between pragmatism and dogmatism (in its original sense) may be said to be a
preference for action over opinion. It is better or more important from the pragmatic point of
view to act effectively than to have sound opinions about principles upon which to act. Thus
stated, there is certainly something to the contrast, but it would be a mistake to conclude from
this that pragmatists can be indifferent to political principles. It depends upon the world in
which they aim to act effectively, and in amplification of this point there is reason to return to
Berlin.

According to Berlin the reason political theory has not and is unlikely to cede to political science
is because:

Among the many topics that remain obstinately philosophical, and have, despite
repeated efforts, failed to transform themselves into sciences are some that in their
very essence involve value judgements . . . Among the problems which form the core of
traditional political theory are those, for instance, of the nature of equality, of rights,
law, authority, rules. . . When we ask what is perhaps the most fundamental of all
political questions 'Why should anyone obey anyone else?' . . . we are asking for the
explanation of what is normative in such notions as authority, sovereignty, liberty . . .
These are words in the name of which orders are issued, men are coerced, wars are
fought, new societies are created and old ones destroyed (The Proper Study of Mankind,
pp.63-5)

Now, if, as a matter of fact, there were no disagreements about these normative notions and
the values they imply, then political practice could be left to the experts.

11
 If we could construct a society in which it was believed universally. . . that there
was only one overriding human purpose . . . a utilitarian; or a Thomist or
communist or Platonic or anarchist society . . . then plainly all that would matter
would be to find the right roads to the attainment of the universally accepted
end . . . To find roads is the business of experts. It is therefore reasonable for
such a society to put itself into the hands of specialists of tested experience,
knowledge, gifts and probity, whose business it is, to use Saint-Simon's simile, to
conduct the human caravan to the oasis the reality and desirability of which is
recognized by all. (The Proper Study of Mankind,66-68)

Equally plainly, however, we do not live in such a society, and the fact that we do not is
one which the political pragmatist, as a realist, must accommodate. It would be as fatal to
effective action to deny the competing variety of political values as to become immersed in
theoretical speculation alone. Political activity is communal just in the sense that it is the co-
ordinated activity of many agents. Robinson Crusoe on his desert island can do many things, but
he cannot engage in politics. It follows that effective political action requires deliberative
communication between political agents and, in the world we actually live in, this implies a
normative and not merely a technical language. There are three implications of this fact that
are of some importance here. First, such language cannot arise from nowhere, it cannot, that is
to say, be derived by pure reason from self-evident first principles. Second, it must provide and
sustain reasons for action. Third, it must be capable of generating a measure of consensus.
Agents can only act together if they agree at some level upon what to do.
It follows from these three features that political deliberation relies upon beliefs and
conceptions which, if they are not agreed upon exactly, are nonetheless recognized as
established parts of a shared political vocabulary, and which carry normative force in the
business of arriving at conclusions, taking decisions and commanding the compliance of others.
Such a vocabulary will in large part be made up of moral beliefs, idealized models of society,
and general political principles. But what beliefs, models and principles these are in any given

12
society at any given time will depend upon a wider cultural sphere than that of politics, a
cultural sphere that includes Coleridge's 'uninterested theorists'. It is of course possible to
exaggerate their role, as no doubt Coleridge does, and to make the related mistake of
supposing that the importance and interest of political ideas lies chiefly or even solely in their
impact upon political life. Both views are mistaken. Political philosophy is neither the master
nor the servant of political action and a political philosopher may raise serious questions of
great intrinsic interest. For instance, there are important philosophical doubts about the value
of democracy, the existence of rights and the concept of social justice. By uncovering such deep
conceptual difficulties amongst our moral ideas, political philosophy provides occasion for
powerful and perplexing intellectual engagement. But just because these are our moral ideas,
uncovering these difficulties will not of itself cut much ice in the world of political deliberation.
Anyone who approaches the question of electoral reform by doubting whether we ought to be
striving to be democratic at all, or who argues about the management of the welfare system by
doubting the very existence of welfare rights or social justice, simply will not get a hearing in
the real world of legislatures and government offices.
For all that, both the common formulations and the general acceptance of these same
ideas, which together allow them to act as normative reasons for action and deliberative full
stops in political decision making, unquestionably owe something to the writings of
'uninterested theorists'. Conversely, the fact that slavery and the divine right of kings are
political non-starters in our world is in part attributable to the reflections of philosophers and
theologians upon them. To quote Berlin again:

In as much as each rests on a vivid vision of permanent human attributes and is

capable of satisfying some enquiring minds in each generation, the models of
Plato, or of Aristotle, or of Judaism, Christianity, liberalism, romanticism
historicism all survive . . . the debate will continue in terms set by these visions
and others like them (The Proper Study of Mankind, p87)

13
 In summary: the intellectualist account of the relation between political theory and
political practice is certainly deeply mistaken. But though it is true that this means the
pragmatists in politics need not trouble to take courses in political philosophy, the pragmatic
requirement to use the political vocabulary of the day means that they cannot ignore the world
of ideas around them, and hence cannot altogether ignore the continuing contribution of
political theorists to the formation of that world.

BIBLIOGRAPHY
Isaiah Berlin, The Proper Study of Mankind (London, Pimlico, 1998)
George Berkeley, Siris (London, 1744)
S T Coleridge, The Statesman’s Manual (London, 1816)
H S J Bolingbroke, Works of Bolingbroke, (Philadelphia, 1841)
Elie Kedourie, Nationalism 2nd edition (London, Hutcheson, 1961)
D Freedman, Markov Chains (San Francisco, Holden-Day, 1971)

14