The Proper Study of Mankind

by Isaiah Berlin

  DOES POLITICAL THEORY STILL EXIST?

  I

  IS THERE STILL such a subject as political theory? This query, put with suspicious frequency in English-speaking countries, questions the very credentials of the subject: it suggests that political philosophy, whatever it may have been in the past, is today dead or dying. 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 the very least one that has in a large area converted paradoxes into platitudes or vice versa. This seems to me no more (but also no less) than an adequate criterion of the characteristic in question.

  But this is scarcely conclusive evidence. There exist only two good reasons for certifying the demise of a discipline: one is that its central presuppositions, empirical, or metaphysical, or logical, are no longer accepted because they have (with the world of which they were a part) withered away, or because they have been discredited or refuted. The other is that new disciplines have come to perform the work originally undertaken by the older study. These disciplines may have their own limitations, but they exist, they function, and have either inherited or usurped the functions of their predecessors: there is no room left for the ancestor from whom they spring. This is the fate that overtook astrology, alchemy, phrenology (positivists, both old and new, would include theology and metaphysics). The postulates on which these disciplines were based either were destroyed by argument or collapsed for other reasons; consequently they are today regarded merely as instances of systematic delusion.

  This type of systematic parricide is, in effect, the history of the natural sciences in their relation to philosophy, and so has a direct bearing upon the question before us. The relevant consideration is this: There exist at least two classes of problems to which men have succeeded in obtaining clear answers. The first have been so formulated that they can (at least in principle, if not always in practice) be answered by observation and by inference from observed data. These determine the domains of natural science and of everyday common sense. Whether I ask simple questions about whether there is any food in the cupboard, or what kind of birds are to be found in Patagonia, or the intentions of an individual, or more complicated ones about the structure of matter, or the behaviour of social classes or international markets, I know that the answer, to have any genuine claim to truth, must rest on someone’s observation of what exists or happens in the spatio-temporal world. Some would say ‘organised observation’. I should be inclined to agree. But differences on this issue, while they are crucial for the philosophy of science and the theory of knowledge, do not affect my argument. All the generalisations and hypotheses and models with which the most sophisticated sciences work can be established and discredited ultimately only by the data of inspection or introspection.

  The second type of question to which we can hope to obtain clear answers is formal. Given certain propositions called axioms, together with rules for deducing other propositions from them, I can proceed by mere calculation. The answers to my questions will be valid or invalid according to whether the rules that I accept without question as part of a given discipline have been correctly used. Such disciplines contain no statements based on observation of fact, and therefore are not nowadays expected to provide information about the universe, whether or not they are used in providing it. Mathematics and formal logic are, of course, the best-known examples of formal sciences of this type, but heraldry, chess, and theories of games in general, are similar applications of the formal methods which govern such disciplines.

  The two methods of answering questions may be, very generally, denominated empirical and formal. Among the characteristics of both are at least these:

  1 Even if we do not know the answer to a given question, we know what kinds of methods are appropriate in looking for the answer; we know what kinds of answers are relevant to these questions, even if they are not true. If I am asked about the workings of the Soviet system of criminal law, or why Kennedy was elected President of the United States, I may not be able to answer the question, but I know within what region the relevant evidence must lie, and how an expert would use such evidence to obtain the answer; I must be able to state this in very general terms, if only to show that I have understood the question. Similarly, if I am asked for the proof of Fermat’s theorem, I may not be able to give it, indeed I may know that no one has yet been able to provide it, but I also know what kinds of demonstration would count as answers to this problem, even though they may be incorrect or inconclusive, and can discriminate these from assertions which are irrelevant to the topic. In other words, in all these cases, even if I do not know the answer, I know where to look for it, or how to identify an authority or expert who knows how to set about looking for it.

  2 This means, in effect, that where the concepts are firm, clear and generally accepted, and the methods of reasoning and arriving at conclusions are agreed between men (at least the majority of those who have anything to do with these matters), there and only there is it possible to construct a science, formal or empirical. Wherever this is not the case – where the concepts are vague or too much in dispute, and methods of argument and the minimum qualifications that constitute an expert are not generally agreed, where we find frequent recriminations about what can or what cannot claim to be a law, an established hypothesis, an undisputed truth, and so on – we are at best in the realm of quasi-science. The principal candidates for inclusion in the charmed circle who have not succeeded in passing the required tests are the occupants of the large, rich and central, but unstable, volcanic and misty, region of ‘ideologies’. One of the rough and ready tests for finding out which region we are in is whether a set of rules, accepted by the great majority of experts in the subject, and capable of being incorporated in a textbook, can be applied in the field in question. To the degree to which such rules are applicable, a discipline approaches the coveted condition of an accepted science. Psychology, sociology, semantics, logic, perhaps certain branches of economics, are in a no man’s land, some nearer to, some further from, the frontier which demarcates, less or more clearly, the territory of the established sciences.

  3 But besides these two major categories, there arise questions which fall outside either group. It is not only that we may not know the answers to certain questions, but that we are not clear how to set about trying to answer them – where to look – what would constitute evidence for an answer and what would not. When I am asked ‘Where is the image in the mirror?’ or ‘Can time stand still?’ I am not sure what kind of question it is that is being asked, or whether indeed it makes any sense at all. I am in not much better plight with some traditional questions which have probably been asked since the dawn of thought, such as ‘How did the world begin?’ and, following that, ‘What happened before the beginning?’ Some say that these are not legitimate questions; but then what makes them illegitimate? There is something that I am trying to ask; for I am certainly puzzled by something. When I ask ‘Why can I not be in two places at once?’, ‘Why can I not get back into the past?’ or, to move to another region, ‘What is justice?’ or ‘Is justice objective and absolute?’ or again ‘How can we ever be sure that an action is just?’, no obvious method of settling these questions lies to hand. One of the surest hallmarks of a philosophical question – for this is what all these questions are – is that we are puzzled from the very outset, that there is no automatic technique, no universally recognised expertise, for dealing with such questions. We discover that we do not feel sure how to set about clearing our minds, finding out the truth, accepting or rejecting earlier answers to these questions. Neither induction (in its widest sense of scientific reasoning), nor direct observation (appropriate to empirical enquiries), nor deduction (demanded by formal problems) seem to be of help. Once we do feel quite clear about how we should proceed, the questions no longer seem philosophical.

  The history – and indeed
the advance – of human thought (this is perhaps a truism) have, in fact, largely consisted in the gradual shuffling of all the basic questions that men ask into one or the other of two well-organised compartments – the empirical and the formal. Wherever concepts grow firm and clear and acquire universal acceptance, a new science, natural or formal, comes into being. To use a simile that I cannot claim to have invented, philosophy is like a radiant sun that, from time to time, throws off portions of itself; these masses, when they cool down, acquire a firm and recognisable structure of their own and acquire independent careers as tidy and regular planets; but the central sun continues on its path, and does not seem to diminish in mass or radiance. The ‘status’ and vitality of philosophy is another matter, and seems to be directly connected with the extent to which it deals with issues that are of concern to the common man. The relation of philosophy to opinion and conduct is a central question of both history and sociology, too large to be considered here. What concerns us is that philosophy in one state of development may turn into a science in the next.

  It is no confusion of thought that caused astronomy, for example, to be regarded as a philosophical discipline in, say, the rime of Scotus Erigena, when its concepts and methods were not what we should today regard as firm or clear, and the part played by observation in relation to a priori teleological notions (such as the yearning of each body to realise the full perfection of its nature) made it impossible to determine whether the amalgam that went under the name of the knowledge of celestial bodies was empirical or formal. As soon as clear concepts and specific techniques developed, the science of astronomy emerged. In other words, astronomy in its beginning could not be relegated to either compartment, even if such compartments as the empirical and the formal had been clearly distinguished; and it was, of course, part of the ‘philosophical’ status of early medieval astronomy that the civilisation of that time (Marxists would say ‘the superstructure’) did not permit the distinction between the two compartments to be clearly demarcated.

  What, therefore, is characteristic of specifically philosophical questions is that they do not (and some of them perhaps never will) satisfy conditions required by an independent science, the principal among which is that the path to their solution must be implicit in their very formulation. Nevertheless, there are some subjects which clearly are near the point of taking flight and divorcing themselves from the main body in which they were born, much as physics and mathematics and chemistry and biology have done in their day. One of these is semantics; another is psychology; with one foot, however reluctantly, they are still sunk in philosophical soil; but they show signs of a tendency to tear themselves loose and emancipate themselves, with only historical memories to tell them of their earlier, more confused, if in some respects richer, years.

  II

  Among the 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. Ethics, aesthetics, criticism explicitly concerned with general ideas, all but the most technical types of history and scholarship, still live at various points of this limbo, unable or unwilling to emerge by either the empirical or the formal door. The mere fact that value judgements are relevant to an intellectual pursuit is clearly not sufficient to disqualify it from being a recognised science. The concept of normal health certainly embodies a valuation, and although there is sufficient universal consensus about what constitutes good health, a normal state, disease and so on, this concept, nevertheless, does not enter as an intrinsic element into sciences such as anatomy, physiology and pathology. Pursuit of health may be the strongest sociological and psychological (and moral) factor in creating and promoting these sciences; it may determine which problems and aspects of the subject have been most ardently attended to; but it is not referred to in the science itself, any more than the uses of history or logic need be mentioned in historical or logical works. If so clear, universally accepted, ‘objective’ a value as that of a desirable state of health is extruded from the structure of the natural sciences, this fact is even more conspicuous in more controversial fields. The attempts, from Plato to our own day (particularly persistent and numerous in the eighteenth century), to found objective sciences of ethics and aesthetics on the basis of universally accepted values, or of methods of discovering them, have met with little success; relativism, subjectivism, romanticism, scepticism with regard to values keep breaking in.

  What, we may ask at this point, is the position of political theory? What are its most typical problems? Are they empirical, or formal, or neither? Do they necessarily entail questions of value? Are they on the way to independent status, or are they by their very nature compelled to remain only an element in some wider body of thought?

  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. We demand the analysis of these concepts, or ask how these expressions function in our language, or what forms of behaviour they prescribe or forbid and why, or into what system of value or outlook they fit, and in what way. When we ask, what is perhaps the most fundamental of all political questions, ‘Why should anyone obey anyone else?’, we ask not ‘Why do men obey?’ – something that empirical psychology, anthropology and sociology might be able to answer – nor yet ‘Who obeys whom, when and where, and determined by what causes?’, which could perhaps be answered on the basis of evidence drawn from these and similar fields. When we ask why a man should obey, we are asking for the explanation of what is normative in such notions as authority, sovereignty, liberty, and for the justification of their validity in political arguments. 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 – expressions which play as great a part as any in our lives today. What makes such questions prima facie philosophical is the fact that no wide agreement exists on the meaning of some of the concepts involved. There are sharp differences on what constitute valid reasons for actions in these fields; on how the relevant propositions are to be established or even rendered plausible; on who or what constitutes recognised authority for deciding these questions; and there is consequently no consensus on the frontier between valid public criticism and subversion, or between freedom and oppression, and the like. So long as conflicting replies to such questions continue to be given by different schools and thinkers, the prospects of establishing a science in this field, whether empirical or formal, seem remote. Indeed, it seems clear that disagreements about the analysis of value concepts, as often as not, spring from profounder differences, since the notions of, say, rights or justice or liberty will be radically dissimilar for theists and atheists, mechanistic determinists and Christians, Hegelians and empiricists, romantic irrationalists and Marxists, and so forth. It seems no less clear that these differences are not, at least prima facie, either logical or empirical, and have usually and rightly been classified as irreducibly philosophical.

  This carries at least one important implication. If we ask the Kantian question ‘In what kind of world is political philosophy – the kind of discussion and argument in which it consists – in principle possible?’ the answer must be ‘Only in a world where ends collide.’ In a society dominated by a single goal there could in principle only be arguments about the best means to attain this end – and arguments about means are technical, that is, scientific and empirical in character: they can be settled by experience and observation or whatever other methods are used to discover causes and correlations; they can, at least in principle, be reduced to positive sciences. In such a society no serious questions about political ends or values could arise, only empirical ones about the most effective paths to the goal. And, indeed, something amounting to this was, in effect, asserted by Saint-Simon and Comte; and, on some interpretations of his thought, by Marx also, at any rate after ‘prehistory’, that is, the class war, is ov
er, and man’s true ‘history’ – the united attack on nature to obtain goods upon whose desirability the whole of society is agreed – has begun. It follows that the only society in which political philosophy in its traditional sense, that is, an enquiry concerned not solely with elucidation of concepts, but with the critical examination of presuppositions and assumptions, and the questioning of the order of priorities and ultimate ends, is possible, is a society in which there is no total acceptance of any single end. There may be a variety of reasons for this: because no single end has been accepted by a sufficient number of persons; because no one end can be regarded as ultimate, since there can, in principle, exist no guarantee that other values may not at some time engage men’s reason or their passions; because no unique, final end can be found – inasmuch as men can pursue many distinct ends, none of them means to, or parts of, one another; and so on. Some among these ends may be public or political; nor is there any reason to suppose that all of them must, even in principle, be compatible with one another. Unless political philosophy is confined to the analysis of concepts or expressions, it can be pursued consistently only in a pluralist, or potentially pluralist, society. But since all analysis, however abstract, itself involves a critical approach to the assumptions under analysis, this distinction remains purely academic. Rigid monism is compatible with philosophical analysis only in theory. The plight of philosophy under despotism in our own times provides conclusive concrete evidence for this thesis.

  III

  Let me try to make this clearer. If we could construct a society in which it was believed universally (or at least by as many people as believe that the purpose of medicine is to promote or maintain health and are agreed about what constitutes health) that there was only one overriding human purpose: for example, a technocratic society dedicated to the single end of the richest realisation of all human faculties; or a utilitarian society dedicated to the greatest happiness of men; or a Thomist or communist or Platonic or anarchist society, or any other society which is monistic in this sense – then plainly all that would matter would be to find the right roads to the attainment of the universally accepted end.