Ancient History & Civilisation





The third book contains much of Aristotle’s best work on politics and much that is of permanent interest. His material is very diverse, and it looks as if he may have assembled and utilized part of some earlier studies.

The major part of the book is devoted to constitutions – democratic, oligarchic, monarchic – a topic which had long been the kernel of Greek political thought. The first chapter opens with a difficult passage, some of which looks more like notes for a lecture than continuous discourse. He has hardly mentioned that he is about to discuss the constitution of a state, when he makes a series of observations not only about that but about the polis or state, and the politēs or citizen. The connection between these observations is not clear, but the argument may have been something like this: One ought, before discussing politeia, to define polis; there is no unanimity about this definition, but, since a state is made up of its citizens, we should rather begin by defining citizen; there is just as much dispute about this definition, but it must be attempted, as a preliminary to discussing ‘constitution’, which is an organized system of relationships governing the state and the citizen.

Aristotle accordingly discusses the definition of a citizen. It may seem odd to us thus to make the citizen prior to the state, but we should remember, first, that to say merely ‘a citizen is a member of a state’ involves asking ‘what constitutes membership of a state?’ – which is just the question now to be discussed; and second, that Aristotle’s outlook is here, as so often, coloured by his biological studies: he is inclined to think of citizen as a kind of species and to look for the marks by which it may be recognized. He acknowledges in this first chapter, however, that the description of a citizen will vary according to the constitution he lives under, so that to look for the notion of the ‘citizen’ pure and simple, in some basic sense applicable to all citizens, may be mistaken in principle.

Nevertheless Aristotle attempts a general definition, namely that a citizen is one who ‘participates in giving judgement and holding office’. This definition ought to jolt us into realizing that Aristotle would regard most of us – who rely on representatives to run our public affairs – as barely citizens at all; he would condemn modern democracy as insufficiently ‘participatory’. On the other hand, even in a Greek democracy, a very large proportion of the population was excluded from office, notably women and slaves. For a stimulating discussion of these and other important questions, see M. I. Finley, Democracy Ancient and Modern (London, 1973), esp. Ch. I.

1274b32 In considering now the varieties and characteristics of constitutions, we must begin by looking at the state and asking what it is. There is no unanimity about this; for example, some say that an action was taken by the state, others that it was taken not by the state, but by the oligarchy or by the dictator. Now obviously the activities of statesman and legislator are wholly concerned with the state, and the constitution is a kind of organization of the state’s inhabitants; but like any other whole that is made up of many parts, the state is to be classed as a composite thing; so clearly we must first try to isolate the citizen, for the state is an aggregate of citizens. So we must ask, Who is a citizen? and, Whom should we call one? 1275a2 Here too there is no unanimity, no agreement as to what constitutes a citizen; it often happens that one who is a citizen in a democracy is not a citizen in an oligarchy. (I think we may leave out of account those who merely acquire the title indirectly, e.g. the ‘made’ citizens.)1

Nor does mere residence in a place confer citizenship: resident foreigners and slaves are not citizens, but do share domicile in the country. Another definition is ‘those who have access to legal processes, who may prosecute or be prosecuted’. But this access is open to any person who is covered by a commercial treaty2 – at any rate partially open, for a resident foreigner is in many places obliged to appoint a patron, so that not even this degree of participation is open to him unqualifiedly. (Likewise boys not yet old enough to be enrolled,3 and old people who have retired from duty, must be termed citizens in a sense, but only with the addition of ‘not fully’ or ‘superannuated’ or some such term – not that it matters which word we use, since what we mean is clear enough). What we are looking for is the citizen proper, without any defect needing to be amended. Similar difficulties may be raised, and solved, about persons exiled or with civic disqualifications.

1275a22 What effectively distinguishes the citizen proper from all others is his participation in giving judgement and in holding office. Some offices are distinguished in respect of length of tenure, some not being tenable by the same person twice under any circumstances, or only after an interval of time. Others, such as membership of a jury or of an assembly,4 have no such limitation. It might be objected that such persons are not really officials, and that these functions do not amount to participation in office. But they have the fullest sovereign power, and it would be ridiculous to deny their participation in office. In any case nomenclature ought not to make any difference; it is just that there is no name covering that which is common to a juryman and to a member of an assembly,which ought to be used of both. For the sake of a definition I suggest that we say ‘unlimited5 office’. We therefore define citizens as those who participate in this. Such a definition seems to cover, as nearly as may be, those to whom the term citizen is in fact applied.

1275a34 On the other hand we must remember that in the case of things in which the substrata differ in kind, one being primary, another secondary, and so on, there is nothing, or scarcely anything, which is common to all those things, in so far as they are the kind of thing they are.6 Thus we see the various constitutions differing from each other in kind, some being prior to others – since those that have gone wrong or deviated must be posterior to those which are free from error. I will explain later what I mean by ‘deviated’.7 A citizen, therefore, will necessarily vary according to the constitution in each case.

1275b5 For this reason our definition of citizen is best applied in a democracy; in the other constitutions it may be applicable, but it need not necessarily be so. For in some constitutions there is no body comprising the people, nor a recognized assembly, but only an occasional rally; and justice may be administered piecemeal.8 For example, at Sparta contract cases are tried by the Ephors, one or other of them, cases of homicide by the Elders, and other cases doubtless by other officials. Similarly at Carthage all cases are tried by officials.

1275b13 But our own definition of a citizen can be amended so as to apply to the other constitutions also. We simply replace our ‘unlimited’ office of juror or member of assembly by ‘limited’.5 For it is to all or some of these that the task of judging or deliberating is assigned, either on all matters or on some. From these considerations it has become clear who a citizen is: as soon as a man becomes entitled to participate in office, deliberative or judicial,9 we deem him to be a citizen of that state; and a number of such persons large enough to secure a self-sufficient life we may, by and large, call a state.

III ii

In this brief chapter, which coheres closely with III i, Aristotle points out that various common practical definitions of a citizen (e.g. as someone both of whose parents were citizens, or as a person who has been ‘made’ into a citizen by some state official empowered to do so) encounter certain difficulties. His own preference is to abandon such formal and legalistic criteria of citizenship, in favour of a functional one: do candidates for the title ‘citizen’ share in deliberative and judicial office, as explained in III i? If so, they are citizens, and may be properly described thus irrespective of their parentage or the manner in which they came to exercise the required functions; and that a man may come to exercise them unjustly does not in itself disqualify him from actually being a citizen in practice. Aristotle’s approach is thus both neat and pragmatic.

1275b22 For practical purposes a citizen is defined as one of citizen birth on both his father’s and his mother’s side; some would go further and demand citizen descent for two, three, or even more generations. But since these are only crude definitions, employed by states for practical purposes, some people pose the puzzle of how a great or great-great-grandfather’s citizenship can itself be determined. Gorgias of Leontini, partly perhaps in puzzlement and partly in jest, said that, as mortars are what mortar-makers make, so Larissaeans are those made by the workmen, some of whom were Larissaean-makers.1 The answer to such objectors is simple: if they participated in the constitution in the manner prescribed in our definition,2 they were citizens. Of course, the criterion of having citizen-parents cannot be applied in the case of the original colonists or founders.

1275b34 I think however that there is perhaps a more important puzzle here, namely about those who got a share in the constitution because it had changed – as for example after the expulsion of the tyrants from Athens, when Cleisthenes enrolled many foreigners and slaves in the tribes.3 The question here is not ‘Are these persons citizens?’, but whether they are citizens justly or unjustly. Some would go further and question whether anyone can be a citizen unless he is justly so, on the ground that unjust and false mean the same thing. But when persons exercise their office unjustly, we continue to say that they rule, though unjustly; and as the citizen has been defined by some kind of office (i.e. if he shares in such and such an office, he is, as we said, a citizen), we cannot deny the propriety of using the term even in these cases.

III iii


The pragmatic note struck at the end of the preceding chapter prompts Aristotle to resume a crucial practical difficulty he had raised at the beginning of III i: when a state’s constitution changes (e.g. from oligarchy to democracy), certain persons are admitted to or deprived of citizenship. If the new government repudiates debts contracted by the old, are we entitled to protest, on the grounds that only the constitution has changed, whereas the state remains in some sense the same? Or does a change of constitution entail a ‘different’ state, characterized by a complete lack of legal and civic continuity with its predecessor? In brief, what are the criteria of continuity of a state, and should the state for this purpose be identified with its constitution at a given moment?

Aristotle first reviews and dismisses some relatively simplistic criteria based on considerations of territory and population, and then seeks guidance from the plausible analogies of ‘choruses’ (the groups of actors who danced and sang in Greek drama), and musical ‘modes’ or ‘styles’. The same notes, differently arranged, produce different styles; similarly, while the people that make up a state may not change during a revolution, they may nevertheless, as a result of that revolution, be ‘associated’ differently to make a different constitution; and different constitutions, because some or all of those who are citizens under one constitution will not be citizens under another, thus entail different states. This is not such a weird conclusion as it may seem. Aristotle would not wish to deny that a state continues to exist after a change of constitution: clearly it is not the case that one state literally ceases to exist and another comes into being. His point is that since a change of constitution entails a change in the citizen-body – i.e. in those who hold deliberative and judicial office (see III i) – the state has become different in kind: it is not the state it was before, in spite of obvious material continuities of site, buildings, population, etc.

But this conclusion hardly settles the question of the justice of a regime’s repudiation of the commitments of its predecessor in power. Aristotle’s decision to make the form of constitution the criterion of continuity lends colour to the claim that a new form of government is not bound by all the contractual obligations of an old. Yet he does not want us to jump to this conclusion, for he indicates there is more to be said on the question; but in fact he never takes it up again. The first paragraph, however, suggests in passing a possible criterion: did the previous constitution promote the general, as distinct from a sectional, interest?

I276a6 This question of justice or the lack of it cannot be separated from the dispute we have already mentioned,1 which arises from the difficulty some people raise as to whether it was or was not the state that acted – for example when a change takes place from oligarchy to tyranny or democracy. There are those2 who after such a change claim that they are no longer obliged to fulfil the terms of a contract; for it had been entered into, so they say, not by the state but by the tyrant. Similarly they would disown other obligations, if these have been incurred under one of those types of constitution which rest on force and disregard the common interest. It follows that if there is a democracy of this type, we must say that the acts of this constitution are acts of the state to the same degree3 as those flowing from the oligarchy or tyranny are.

1276a17 And this topic seems to be part of yet another question – how are we to tell whether a state is still the same state or a different one? We might try to investigate this question using territory and inhabitants as criteria; but this would not carry us very far, since it is quite possible to divide both territory and population into two, putting some people in one part and some into the other. That is not a very serious difficulty: it arises from our use of the word polis in more than one sense.4 Such a puzzle is therefore resolved easily enough.

1276a24 Another question is this: when a population lives in the same place, what is the criterion for regarding the state as a unity? It cannot be the walls, for it would be possible to put one wall round the whole Peloponnese. Babylon is perhaps a similar case, and any other state with a circumference that embraces a nation rather than a state. (It is said of Babylon that its capture was, two days later, still unknown to a part of the city.) These questions of the state’s size – both how big it should be and whether it helps to have the population drawn from one nation or more than one – are problems to which it will be useful to return later,5 since the statesman has to keep them in mind.

1276a34 But when the same population continues to dwell in the same territory, must we say that the state remains the same so long as there is continuity of race among that population, even though one generation of people dies and another is born – just as a river or spring is commonly said to be the same, although different water passes into and out of it all the time?6 Alternatively, ought we to speak of the population as being the same for the reasons stated, but say that the state is different?

For the state is a kind of association – an association of citizens in a constitution; so when the constitution changes and becomes different in kind, the state also would seem necessarily not to be the same. We may use the analogy of a chorus, which may at one time perform in a tragedy and at another in a comedy, so that we say it is different – yet often enough it is composed of the same persons. And the same principle is applicable to other associations and combinations, which are different if the combination in question differs in kind. For example, we say the same musical notes are fitted together differently, to produce either the Dorian or the Phrygian mode. If this is right, it is clear that the main criterion of the continued identity of a state ought to be its constitution. This leaves it quite open either to change or not to change the name of a state, both when the population is the same and when it is different.

But whether, when a state’s constitution is changed, it is just to disown obligations or to discharge them – that is another question.

III iv

Aristotle has discussed, but has not really finished discussing, citizen and state in relation to constitution. He has had difficulty in defining citizen in such a way as to be applicable to all forms of constitution. This same difficulty now reappears in another form: in attempting to find out what agood citizen is, he discovers that here too the answer depends largely on the politeia, constitution, in which the citizenship is held. Perhaps this is our cue to remind ourselves of the inadequacy of ‘constitution’ as a translation ofpoliteia, which embraces the whole social, political and economic organization of the state; and also that ‘virtue’ (aretē, ‘excellence’, ‘efficiency’) is often conceived in terms of civic function rather than of character or mental or spiritual state: ‘what I can do’ as well as ‘what I inwardly am’. What follows now, about the good man, ‘sound’ citizen, and good ruler or official, will be clearer if these points are borne in mind. (In this chapter, spoudaios, ‘sound’, usually of the citizen, is distinguished in translation from agathos, ‘good’, usually of the individual man – though what if anything hinges on this point of vocabulary is by no means clear; indeed, in this chapter the two terms are commonly interpreted as virtual synonyms – but see R. Develin, ‘The good man and the good citizen in Aristotle’s Politics’, Phronesis, 18 (1973), pp. 71–9.)

Aristotle argues that the good citizen can be a good man only in the ‘best’ state, and only then when he is performing the function of ruling, not when he is being ruled. Such a doctrine sounds to us arid and unhelpful. Why does Aristotle hold it?

The first four paragraphs argue that what makes a good citizen is the contribution his ‘virtue’ makes to the stability and well-being of the constitution; but of constitutions there are many different kinds, so the ‘virtue’ of a good citizen will vary according to the constitution under which he lives. It cannot, therefore, be the same as the invariable ‘perfect’ virtue of the good man. Even in the ‘best’ state, the inevitable differentiation of function among the citizens prevents the virtue of the good citizen from being identical with that of the good man, which is single, and the same in all good men.

The next two paragraphs concede that the two virtues may coincide in particular cases, i.e. when the good man is, in his capacity as good citizen, ruling well – for that entails the exercise of wisdom (phronesis) which is an essential part of the perfect’ virtue of a good man. But part of the virtue of a good citizen is to be ruled well also, which does not require wisdom, and is less praiseworthy. Only when the good citizen is ruling is his virtue identical with that of the good man. Two more paragraphs then distinguish the citizen’s alternation of ruling and being ruled from the permanent relationship of ruler and ruled seen in slavery. A final paragraph points out that the particular virtues, e.g. wisdom, differ according as to whether they are exercised by ruler or ruled. This consideration too militates against the view that the virtue of the good citizen and the good man are identical, for the virtue of the latter is exercised only when ruling.

The last paragraph of the chapter perhaps suggests that Aristotle found these conclusions somewhat unpalatable. His dilemma arises, as so often, from his belief in natural teleology. Man has certain faculties to use, including the noblest, the intellectual. A man cannot be ‘good’ if he is not using his faculties: there is a kind of moral imperative to use them (cf. introduction to III vi). Now a man who is ruled has to hold his reasoning faculty in abeyance, to the extent that he must do as he is told and not exercise judgement or take a decision of his own or direct other people. He cannot, therefore, be a ‘good’ man. He may be a good citizen, of a particular constitution; but unless he is ruling well over fellow-citizens, having first learned to do so by being ruled, he is not exercising his human faculties to the full, and is not ‘good’. The ‘good man’ looks as though he will be a rather rare phenomenon (cf. the exacting criteria and apparent rarity of phronēsis, practical wisdom, in Nicomachean Ethics, VI xiii); and at the end of the present chapter we learn that men who are being ruled will need mere ‘right opinion’, not phronēsis (cf. Plato, Republic IV). Aristotle is thinking in terms of an intellectual and political élite.

Aristotle could have escaped his dilemma simply by challenging his own assumption that ‘goodness’ or ‘virtue’ (aretē) cannot be complete without the exercise of certain functions. In so far as ‘good’ is a term of moral approbation, why should it depend on the exercise of the ‘noble’ faculty of reason for the purpose of ruling? Could not an unintelligent person, who never rules, be ‘good’? Back would come Aristotle’s answer: ‘As a citizen, yes; as a man, no’. We may think this queer; but at least his very acute discussion does have the considerable merit of focusing our attention on the question, ‘What do we mean when we say a man is “good”?’.

These seem to be the central issues of the chapter, but there is much else of interest: for a full discussion, see the commentaries, especially Newman I 234 ff. Some sentences are very difficult, partly because of textual cruces.

1276b16 Connected with the matters just discussed is the question whether we ought to regard the virtue of a good man and that of a sound citizen as the same virtue, or not. If this is a point to be investigated, we really must try to form some rough conception of the virtue of a citizen.

1276b20 So then: we say a citizen is a member of an association, just as a sailor is; and each member of the crew has his different function and a name to fit it – rower, helmsman, look-out, and the rest. Clearly the most exact description of each individual will be a special description of his virtue; but equally there will also be a general description that will fit them all, because there is a task in which they all play a part – the safe conduct1 of the voyage; for each member of the crew aims at securing that. Similarly the task of all the citizens, however different they may be, is the stability1 of the association, that is, the constitution. Therefore the virtue of the citizen must be in relation to the constitution; and as there are more kinds of constitution than one, there cannot be just one single and perfect virtue of the sound citizen. On the other hand we do say that the goodman is good because of one single virtue which is perfect virtue. Clearly then it is possible to be a sound citizen without having that virtue which makes a sound man.

1276b35 Look now at the problem from another angle and consider the same point in relation to the best constitution. That is to say, if it is impossible for a state to consist entirely of sound men, still each of them must do, and do well, his proper work; and doing it well depends on his virtue. But since it is impossible for all the citizens to be alike, there cannot be one virtue of citizen and good man alike. For the virtue of the sound citizen must be possessed by all (and if it is, then that state is necessarily best). But if it is inevitable that not all the citizens in a sound state are good,2 it is impossible for all to have the virtue of the good man.

1277a5 Again, a state is made up of unlike parts. As an animate creature consists of body and soul, and soul consists of reasoning and desiring, and a household consists of husband and wife, and property consists of master and slave, so also a state is made up of these and many other sorts of people besides, all different. The virtue of all the citizens cannot, therefore, be one, any more than in a troupe of dancers the goodness of the leader and that of the followers are one.

1277a12 Now while all this shows clearly that they are not the same in general, the question may be asked whether it is not possible in a particular case for the same virtue to belong both to the sound citizen and the sound man. We would answer that there is such a case, since we maintain that a sound ruler is both good and wise,3 where as wisdom is not essential for a citizen. Some say that from the very start there is a different kind of education for rulers. They instance (a) the obvious training of the sons of royalty in horsemanship and war, and (b) a saying of Euripides, which is supposed to refer to the education of a ruler: ‘No frills in education, please… only what the state doth need.’4 But though we may say that the virtue of good ruler and good man is the same, yet, since he too that is ruled is a citizen, we cannot say in general that the virtue of citizen and man are one, but only that they may be in the case of a particular citizen.5 For certainly the virtue of ruler and citizen are not the same. And that doubtless is the reason why Jason of Pherae said that he went hungry whenever he ceased to be tyrant, not knowing how to live as a private person.

1277a25 But surely men praise the ability to rule and to be ruled, and the virtue of a citizen of repute seems to be just this – to be able to rule and be ruled well. If then we say that the virtue of the good man is to do with ruling, and that of the citizen to do with both ruling and being ruled, the two things cannot be praiseworthy to the same degree.

1277a29 So since on occasions they seem different, and ruler and ruled ought not to learn the same things, where as the citizen ought to know both and share in both, one could see from the following.6

1277a33 For there is such a thing as rule by a master, which we say is concerned with necessary tasks; but the master has no necessity to know more than how to use such labour. Anything else, I mean to be able actually to be a servant and do the chores, is simply slave-like. (We speak of several kinds of slave, corresponding to the several varieties of operation. One variety is performed by manual workers, who, as the term itself indicates, live by their hands; among these are the skilled mechanics.) Hence, in some places, only with the arrival of extreme democracies have workmen attained to participation in office. The work then of those who are subject to rule is not work which either the good statesman or the good citizen ought to learn, except occasionally for the personal use he may require to make of it. For then the distinction between master and slave just ceases to apply.

1277b7 But there is another kind of rule – that exercised over men who are free, and similar in birth. This we call rule by a statesman.7 It is this that a ruler must first learn through being ruled, just as one learns to command cavalry by serving under a cavalry-commander and to be a general by serving under a general, and by commanding a battalion and a company. This too is a healthy saying, namely that it is not possible to be a good ruler without first having been ruled. Not that good ruling and good obedience are the same virtue – only that the good citizen must have the knowledge and ability both to rule and be ruled. That is what we mean by the virtue of a citizen – understanding the governing of free men from both points of view.

1277b16 Returning now to the good man, we find the same two qualities. And this is true even though the self-control and justice exercised in ruling are not the same in kind.8 For clearly the virtue of the good man, who is free but governed, for example his justice, will not be always one and the same: it will take different forms according to whether he is to rule or be ruled, just as self-control and courage vary as between men and women. A man would seem a coward if he had only the courage of a woman, a woman a chatterbox if she were only as discreet as a good man. Men and women have different parts to play in managing the household: his to win, hers to preserve. But the only virtue special to a ruler is practical wisdom; all the others must be possessed, so it seems, both by rulers and by ruled. The virtue of a person being ruled is not practical wisdom but correct opinion; he is rather like a person who makes the pipes, while the ruler is the one who can play them.

These considerations have made clear whether the virtue of the good man and that of the sound citizen are the same or different, and the sense in which they are the same and the sense in which they are different.


Aristotle now returns to citizenship. He is aware that in some states the banausoi, mechanics, have full citizenship; but in his view they cannot, by the nature of their occupation, possess the qualities and abilities necessary for a citizen. Does this call for a new definition? Aristotle does not think so. He holds that citizens are a particular class of men, to which no one who is constantly engaged in commercial or manual labour can belong, at any rate in the ‘best’ state. Such people simply do not have the time and opportunity to fulfil the essential function of a citizen, to rule (while holding office) and to be ruled by turn. Mechanics, labourers, etc. are no doubt essential to a state; but that is not in itself a qualification for citizenship, which is a matter not merely of life but of the ‘good’ life, in which a man exercises those high faculties possessed by him as a human being, notably that of reason.

1277b33 There remains still a question about the citizen. Is a citizen really ‘one who has the chance to participate in offices’, or are we to count mechanics too as citizens? If we do the latter, i.e. give them the title citizen though they do not share in government, then the virtue of the citizen ceases to be that of every citizen, since the mechanic too is a citizen. On the other hand, if he is not a citizen, where does he belong, since he is not a foreign resident or a visitor either? But perhaps this kind of reasoning does not really result in any absurdity. After all, slaves do not belong to any of the above-mentioned categories, nor do freed slaves: true it is that we must not give the name citizen to all persons whose presence is necessary for the existence of the state. (Nor yet are children citizens in an unqualified sense, like grown men; children can be called citizens only in a hypothetical sense:1 they are citizens, but incomplete ones.) Indeed, in ancient times in certain countries the mechanics were slaves or foreigners, and therefore mostly still are. But2 the best state will not make the mechanic a citizen. But if even he is to be a citizen, then at any rate what we have called the virtue of a citizen cannot be ascribed to everyone, nor yet to free men alone, but simply to those who are in fact relieved of necessary tasks. Some tasks of this kind are discharged by services to an individual, by slaves, others by mechanics and hired labourers, who serve the public at large.

1278a13 A little further examination will show how it stands with these people, and our earlier statement of the position will itself suffice to make matters clear:3 as there are several constitutions, so there must be several kinds of citizen, particularly of citizen under a ruler. Thus in one constitution it will be necessary, in another impossible, for the mechanic and the hired labourer to be a citizen. It would, for example, be impossible in any constitution called aristocratic or any other in which honours4 depend on merit and virtue; for it is quite impossible, while living the life of mechanic or hireling, to occupy oneself as virtue demands. In oligarchies it is not possible for a hireling to be a citizen, because of the high property-qualifications required for participating in office; but it may be possible for a mechanic, since in fact most skilled workers become rich. In Thebes, however, there was a law requiring an interval of ten years to elapse between giving up trade and participating in office.

1278a26 In many constitutions the law admits to citizenship a certain number even of foreigners; in some democracies the son of a citizen mother is a citizen, and in many places the same applies to illegitimate children. Lack of population is the usual reason for resorting to laws such as these. But when, after making such persons citizens because of a dearth of legitimate citizens, the state has filled up its numbers, it gradually reduces them, dropping first the sons of slave father or slave mother, then sons of citizen mother but not father, and finally they confine citizenship to those of citizen birth on both sides.

1278a34 From all this two points emerge clearly: first, that there are several kinds of citizen, but second, that a citizen in the fullest sense is one who has a share in honours.4 We are reminded of Homer’s ‘Like some immigrant settler, without honour’.5 For he who has no share in honours4 is no better than a resident alien. (Sometimes such efforts are concealed, so that the fellow-inhabitants6 may be deceived.)

1278a40 We have now answered the question whether it is the same or a different virtue that makes a good man and a sound citizen, and have shown that in one state they will be the same person, and in another different; and that where they are the same, not every sound citizen will be a good man, but only the statesman, that is one who is in sovereign control, or capable of being in control, either alone or in conjunction with others, of the administration of public affairs.

III vi

In this chapter Aristotle first briefly classifies constitutions according to the ‘sovereign’ or ruling element in each; then, in the light of his teleological view of the state as existing to serve the common good, pronounces judgement on them according to whether the ‘sovereign’ element serves its own interests only or those of all the citizens.

His contrast between political rule and the rule of master over slave is revealing: whereas political authority should be exercised in the interests of the ruled, a master’s authority is primarily for his own benefit, and only incidentally for that of the slave. The crucial difference, evidently, is that citizens are equal, and those in power are not essentially superior to the ruled; but a slave is inferior to his master (in point of rationality and hence capacity for natural ‘political’ life as a citizen), so that his master’s interests are paramount. Not that Aristotle means to license ill-treatment or unscrupulous exploitation of slaves: as he remarks, authority over slaves in bad condition is a poor thing, and slaves in fact benefit, in virtue of their natural capacities and functions, from their relationship with their masters; in this sense master and slave have the ‘same’ interest in maintaining the relationship. Obviously Aristotle’s argument from inferiority of interests is bound to be repugnant to anyone with egalitarian views about the rights of man. It does however make sense within the framework of a belief that the ‘political’ life lived by a citizen of a polis is supremely ‘natural’ and ‘best’, for ‘natural’ interests are the highest interests of all. ‘Nature’ seems to carry strong social and political imperatives.

1278b6 Having settled these questions, we must proceed to our next and ask whether we are to posit only one constitution or more than one; and if more than one, what they are and how many, and what the differences are between them. The ‘constitution’ of a state is the organization of the offices,1 and in particular of the one that is sovereign over all the others. Now in every case the citizen-body of a state is sovereign; the citizen-body is the constitution. Thus in democracies the people are sovereign, in oligarchies the few. That, we say, is what makes the one constitution differ from the other; and the same criterion can be applied to the others also.

1278b15 We ought at the outset to state the purpose for which the state has come to be, as well as the number of kinds of authority controlling men and their life as members of an association. At the beginning of this work, when we drew a distinction between household-management and mastership, we also stated that by nature man is a political animal.2 Hence men have a desire for life together, even when they have no need to seek each other’s help. Nevertheless, common interest too is a factor in bringing them together, in so far as it contributes to the good life of each. Thegood life is indeed their chief end, both communally and individually; but they form and continue to maintain a political association3 for the sake of life itself. Perhaps we may say that there is an element of good even in mere living, provided that life is not excessively beset with troubles. Certainly most men, in their desire to keep alive, are prepared to face a great deal of suffering, as if finding in life itself a certain well-being and a natural sweetness.

1278b30 But to return to authority: it is not difficult to distinguish its recognized styles (I often speak about their definition in my public lectures).4 First, although the natural slave and the natural master really have the same interest, rule of master over slave is exercised primarily for the benefit of the master and only incidentally for the benefit of the slave, because if the slave deteriorates the master’s rule over him is inevitably impaired.

1278b37 Then there is the authority of a man over his wife, his children, and his whole household, to which we give the name ‘household-management’. This is exercised either for the benefit of those subject to the authority, or for some benefit common to both parties. In itself it is for the benefit of the subjects, as we see by the analogy of the other skills, such as that of a doctor or of an athlete’s trainer, who would only incidentally be concerned with their own interests. (For of course there is nothing to prevent a trainer on occasion being himself a member of the team in training, as the man who steers the ship is always one of the members of the ship’s company. The trainer or pilot looks to the good of those under his authority, but when he himself is one of them he gets the same benefit out of it incidentally as they do, in that the pilot is a member of the ship’s company, and the trainer becomes one of those in training, while yet remaining their trainer.)

1279a8 That is why,5 whenever authority in the state is constituted on a basis of equality and similarly between citizens, they expect to take turns in exercising it. This principle is very old but in earlier times it was applied in a natural and proper manner: men expected each to take a turn at public service,6 and during tenure of office to look after the interests of someone else, who then did the same for him. But nowadays there is more to be gained out of public affairs and offices, so men want to be in office continuously. They could hardly be more zealous in their place-hunting if they were ill and their recovery depended on securing office.

1279a16 It is clear then that those constitutions which aim at the common good are right, as being in accord with absolute justice; while those which aim only at the good of the rulers are wrong. They are all deviations from the right constitutions. They are like the rule of master over slave, whereas the state is an association of free men.

III vii

This chapter carries forward the theme of its predecessor, by distinguishing three ‘correct’ constitutions and three ‘deviations’. The three correct kinds aim at the common interest: kingship, aristocracy, ‘polity’; the three others aim at the sectional interest of the rulers: tyranny, oligarchy, democracy. In each set the first type is the rule of one, the second the rule of a few, and the third the rule of the many. These distinctions are somewhat rough-and-ready, as Aristotle himself recognizes in the next chapter.

1279a22 Having drawn these distinctions we must next consider what constitutions there are and how many. We begin with those that are correct, since when these have been defined it will be easy to see the deviations. As we have seen, ‘constitution’ and ‘citizen-body’ mean the same thing, and the citizen-body is the sovereign power in states. Sovereignty necessarily resides either in one man, or in a few, or in the many. Whenever the one, the few, or the many rule with a view to the common good, these constitutions must be correct; but if they look to the private advantage, be it of the one or the few or the mass, they are deviations. For either we must say that those who do not participate are not citizens, or they must share in the benefit.1

1279a32 The usual names for right constitutions are as follows: (a) Monarchy2 aiming at the common interest: kingship, (b) Rule of more than one man but only a few: aristocracy (so called either because the best men rule or because it aims at what is best for the state and all its members).3 (c) Political control exercised by the mass of the populace in the common interest: polity. This is the name common to all constitutions.4 It is reasonable to use this term, because, while it is possible for one man or a few to be outstanding in point of virtue, it is difficult for a larger number to reach a high standard in all forms of virtue – with the conspicuous exception of military virtue, which is found in a great many people. And that is why in this constitution the defensive element is the most sovereign body, and those who share in the constitution are those who bear arms.

1279b4 The corresponding deviations are: from kingship, tyranny; from aristocracy, oligarchy; from polity, democracy. For tyranny is monarchy2 for the benefit of the monarch, oligarchy for the benefit of the men of means, democracy for the benefit of the men without means. None of the three aims to be of profit to the common interest.

III viii

The classification of constitutions in the preceding chapter was on the basis of the number and aims of the rulers.

Now, in a somewhat Marxist vein, Aristotle argues that the really crucial criterion of classification is an economic one (cf. IV iv init.): an oligarchy is the rule of the wealthy, whether they are few or many (though in practice they are invariably few), and democracy the rule of the poor, whether they are many or few (though in practice they are invariably many). ‘Democracy’ is in fact to Aristotle and the Greeks in general a brutally realistic word: its literal meaning is ‘power of the people’ (dēmos), and it means what it says – rule by a particular class, the numerous poor, and in their own interests. Today we think of ‘democracy’ as being, ideally at any rate, ‘rule by and for everybody at large’, whether rich or poor (cf. p. 362).

1279b11 We must however go into a little more detail about what each of these constitutions is. Certain difficulties are involved, which one whose aim is strictly practical might be allowed to pass over; but a man who examines each subject from a philosophical standpoint cannot neglect them: he has to omit nothing, and state the truth about each topic.

1279b16 Tyranny, as has been said, is a monarchy which is exercised like a mastership1 over the association which is the state; oligarchy occurs when the sovereign power of the constitution is in the hands of those with possessions, democracy when it is in the hands of those who have no stock of possessions and are without means. The first difficulty concerns definitions. Suppose the majority to be well-off, and to be sovereign in the state; then we have a democracy, since the mass of the people is sovereign. So too, if it is somewhere the case that those who do not own property, while fewer in number than those who do, are more powerful and in sovereign control of the constitution, then that is called an oligarchy, since the few are sovereign. It looks therefore as if there were something wrong with our way of defining constitutions.2

1279b26 Even if we try to include both criteria of nomenclature, combining wealth with fewness of numbers in the one case (calling it oligarchy when those who are both wealthy and few hold office), lack of wealth with large numbers in the other (calling it democracy when those who are both poor and numerous hold office) – even then we are only raising a fresh difficulty. For if there is not in fact any other constitution than those with which we have been dealing,3 what names can we give to the two just mentioned, one in which the wealthy are more numerous, and one in which the poor are less numerous, each category being in its own case in sovereign control of the constitution? The argument seems to show that it is a matter of accident whether those who are sovereign be few or many (few in oligarchies, many in democracies): it just happens that way because everywhere the rich are few and the poor are many. So in fact the grounds of difference have been given wrongly: what really differentiates oligarchy and democracy is wealth or the lack of it. It inevitably follows that where men rule because of the possession of wealth, whether their number be large or small, that is oligarchy, and when the poor rule, that is democracy. But, as we have said, in actual fact the former are few, the latter many. Few are wealthy, but all share freedom alike: and these are the grounds of their respective claims to the constitution.4

III ix (1280a7–1281a10)

Another possible criterion for the distribution of political power is justice. Now differences in ethical standards – between what counts as ‘just’ in one state and what counts as such in another – could indeed provide an excellent way of classifying and of comparing the various forms of society. But Aristotle’s use of the new criterion, which he discusses in Chapters ix–xiii, is hardly what we expect. By ‘justice’ in this chapter he means distributive justice, in a political context: what share of political power is it ‘just’ that each man should have? An oligarch will answer, ‘the rich should have more, the poor less’; the democrat will say, ‘all men who are equal, i.e. in being of free birth, should have share in political power on an equal basis’. Aristotle objects to both these answers as resting on at best a partial notion of justice. He accepts the notion that the state should confer political power, privilege and status in proportion to ‘value received’, i.e. in proportion to the contribution men make to the total purpose for which the state exists, the good life, which entails the exercise of all the distinctly human virtues. Such a distribution would be ‘just’ in a ‘complete’ sense, because it depends on a man’s contribution not to partial ends such as wealth, but to the overall end of the association we call a state.

So the question that faces Aristotle is, What kind of superiority is to be regarded as constituting a claim to office and privilege? As we would expect, he puts good birth and ownership of property on the list, and the moral qualities of justice and courage; a high level of culture and education too will be a token of merit in one who is to take part in the working of a state which aims at securing the good life. Men are not equal in these respects, and any state which ignores this fact and thinks in terms of absolute equality must be one of the wrong types, a ‘deviation’ (cf. III xiii). The upper groups will always be superior in education and ability; but the numerical superiority of other sections of the population must also be taken into account.

The description of ‘deviation’ given in ix–xiii differs from that given in vi–vii, but does not contradict it: Aristotle is drawing out the implications of his earlier account. ‘Deviated’ constitutions, according to vi–vii, are those in which the interests of one section of the community take precedence over those of the others. Consequently, absolute equality has to be condemned, since it fails to satisfy the criterion of a ‘correct’ constitution, that it should seek the good of all.

These considerations prompt Aristotle to discuss why, on this view, mere geographical proximity of residence, or trading agreements or a military alliance between various groups of people, do not make a state in the true sense of the word. Such arrangements may indeed serve the material interests of all, but they suffer from the crucial limitation that they do not, at least directly, help to make the citizens more ‘politically’ virtuous, i.e. more able to play their part in a ‘good’ state (polis). A state will no doubt make trading and defensive treaties etc.; but they are not its raison d’être. Or are they? An objector may well retort that historically at any rate these were precisely the purposes for which states were founded – for mutual protection and to ensure a food supply. Why should we attach to the state any other purpose than these limited utilitarian ones? In any case, he might complain, the notions of ‘political’ virtue and the ‘good life’ are very hard to describe exactly. It must be admitted that on this point Aristotle is in this chapter both brief and vague; however, he does give fuller accounts elsewhere, notably in the first three chapters of VII. But just how far does he mean to go when he insists (in the third paragraph) that a ‘true’ state must concern itself with the moral virtue of its citizens? For the controversy on this point, see e.g. l–lii in Barker’s translation of the Politics, and D. J. Allan, ‘Individual and state in the Ethics and Politics’, in La Politique d’Aristote, Fondation Hardt, Entretiens XI (Geneva, 1965) (see Select Bibliographies).

1280a7 First we must grasp what definitions of oligarchy and democracy men put forward, and in particular what is the oligarchic and what is the democratic view of justice. For all adhere to a justice of some kind, but they do not proceed beyond a certain point, and are not referring to the whole of justice in the sovereign sense when they speak of it. Thus it is thought that justice is equality; and so it is, but not for all persons, only for those that are equal. Inequality also is thought to be just; and so it is, but not for all, only for the unequal. We make bad mistakes if we neglect this ‘for whom’ when we are deciding what is just. The reason is that we are making judgements about ourselves, and people are generally bad judges where their own interests are involved. So, as justice is relative to people, and applies in the same ratio to the things and to the persons (as pointed out in my Ethics),1 these disputants, while agreeing as to equality of the thing, disagree about the persons for whom,2 and this chiefly for the reason already stated, that they are judging their own case, and therefore badly.

1280a21 There is also this further reason, namely that both parties are talking about justice in a limited sense, and so imagine themselves to be talking about justice unqualifiedly. Thus it is an error when men unequal in one respect, e.g. money, suppose themselves unequal in all, just as it is an error when men equal in one respect, e.g. in being free, suppose themselves equal in every respect. To argue thus is to neglect the decisive point. If persons originally come together and form an association for the sake of property, then they share in the state3 in proportion to their ownership of property. This is the apparent strength of the oligarchs’ view that it is not just that out of a sum of a hundred minae he that contributed only one should receive equal shares with him who found the remaining ninety-nine; and that this applies equally to the original sum and to any profits subsequently made. But a state’s purpose is not merely to provide a living but to make a life that is good. Otherwise it might be made up of slaves or animals other than man, and that is impossible, because slaves and animals do not participate in happiness,4 nor in a life that involves choice.

1280a34 A state’s purpose is also to provide something more than a military pact of protection against injustice, or to facilitate mutual acquaintance and the exchange of goods, for in that case Tyrrhenians and Carthaginians, and all others with commercial treaties with each other, would be taken as citizens of a single state. Certainly they have import agreements, treaties to prevent injustice, and written documents governing their military alliance. But in the first place each has its separate officials: there are none in common to which they are both equally subject for these purposes. Secondly, neither side is concerned with the quality of the other, or with preventing the behaviour of any person covered by the agreements from being unjust or wicked, but only with the prevention of injustice as between each other. But all who are anxious to ensure government under good laws make it their business to have an eye to the virtue and vice of the citizens. It thus becomes evident that that which is genuinely and not just nominally called a state must concern itself with virtue. Otherwise the association is a mere military alliance, differing only in location and restricted territorial extent from an alliance whose parties are at a distance from each other; and under such conditions law becomes a mere agreement, or, as Lycophron5 the sophist put it, ‘a mutual guarantor of justice’,6 but quite unable to make citizens good and just.

1280b12 That this is true will be clear from some further illustrations. Suppose you merge the territories into one, making the walls of Corinth and Megara contiguous: that still does not make a single state of them, nor would it even if they established rights of marriage between the two, though this is one of the ties peculiarly characteristic of states. Or again, suppose you had 10,000 people living apart from each other, but near enough not to become dissociated: carpenter, farmer, shoemakers and suchlike are there, and furthermore they have laws prohibiting injustice in their transactions with each other; yet, so long as their association does not go beyond such things as commercial exchange and military alliance, that is still not a state. And why not? you may ask. The reason is certainly not that the association is loosely knit. For even if they actually moved close together, and maintained an association such as I have described, with each man still treating his own household like, a state, and if they mutually supported each other, as in a defensive alliance, against injustice only, even then that would not be considered a state, not at any rate in the strict sense, since the nature of their intercourse is the same whether they move close together or stay apart.

1280b29 It is clear therefore that the state is not an association of people dwelling in the same place, established to prevent its members from committing injustice against each other, and to promote transactions. Certainly all these features must be present if there is to be a state; but even the presence of every one of them does not make a state ipso facto. The state is an association intended to enable its members, in their households and the kinships,7 to live well; its purpose is a perfect and self-sufficient life. However, this will not be attained unless they occupy one and the same territory and intermarry. It is indeed on that account that we find in states connections between relatives by marriage, brotherhoods, sacrifices to the gods, and the various civilized pursuits of a life lived together. All these activities are the product of affection, for it is our affection for others that causes us to choose to live together; thus they all contribute towards that good life which is the purpose of the state; and a state is an association of kinships7 and villages which aims at a perfect and self-sufficient life – and that, we hold, means living happily and nobly.8

1281a2 So we must lay it down that the association which is a state exists not for the purpose of living together but for the sake of noble8 actions. Those who contribute most to this kind of association are for that very reason entitled to a larger share in the state than those who, though they may be equal or even superior in free birth and in family, are inferior in the virtue that belongs to a citizen. Similarly they are entitled to a larger share than those who are superior in riches but inferior in virtue.

All this makes it clear that all those who dispute about constitutions are using the term ‘justice’ in a limited sense.


This short chapter is a good example of Aristotle’s ‘aporetic’ style, in which he makes a number of suggestions in swift succession, only to dismiss them. The opening question, ‘What should be the sovereign authority of the state?’ links the chapter to the beginning of III vi; and the question of the justice of the actions of various possible ‘sovereign authorities’ links the discussion as a whole to the preceding chapter, on the just distribution of power. The candidates for sovereignty Aristotle lists for consideration in this chapter are (i) the mass of the people, who are poor; (ii) a tyrant; (iii) the few, who are rich; (iv) ‘respectable’ people (epieikeis); (v) the one most worthy man (spoudaiotatos); (vi) law.

In the case of (i)–(iii), Aristotle is at pains to point out that an authority may commit unjust acts; he instances the taking of the property of other sections of the population, and makes the useful point that sovereignty does not actually confer justice on anything the sovereign power may do, even if it is ‘just’ that the rulers in question should hold the sovereignty in the first place. (He perhaps implies, but does not make, a reverse point: that few and slight unjust acts of rulers holding sovereign authority do not in themselves make it unjust that they should hold it.) The justice of holding authority and the justice displayed in exercising it in action are indeed different, and the distinction is important; but the chapter is too staccato to reveal Aristotle’s views on their exact relationship. The chapter closes with the sober reflection that making law sovereign instead of men solves no difficulties: law can and does reflect the bias of its framers.

1281a11 Another question is ‘Where ought the sovereign power of the state to reside?’ With the mass of the people? With the rich? With the respectable? With one man, the best of all? Or a tyrant? There are objections to all these. Thus suppose the poor use their numerical superiority to make a distribution of the property of the rich: is not that unjust? ‘No, by Zeus,’ it may be said, ‘it has been done justly, by a decision of the sovereign power.’ But what else can we call the very height of injustice? And if the majority, having laid their hands on everything, again distribute the possessions of the few, they are obviously destroying the state. But virtue does not destroy its possessor, nor is justice destructive of the state. So it is clear that this law too cannot be just. Or, secondly, if it is just, any actions taken by a tyrant also must be just: his superior strength enables him to use force, just as the mass of the people use force on the rich. Thirdly, is it just for the few and the wealthy to rule? If so, and they too do this and plunder and help themselves to the goods of the mass, then that is just. And if that is so, then it is just in the former case also. The answer clearly is that all these three states of affairs are bad and not just.

1281a28 The fourth alternative, that the respectable should rule and have sovereign power over everything, means that all the rest must be without esteem, being debarred from the honour of holding office under the constitution. For offices, we say, are honours; and if the same persons hold office all the time, the rest must be without honour. Is then the fifth alternative better, that one man, the most worthy, should rule? But this is yet more oligarchical, because it leaves still larger numbers without honour. It might be objected that it is a bad thing for any human being, subject to the affections that enter the soul, to have sovereign power, which ought to be reserved for the law. But that will not make any difference to the problem-cases we have been discussing: there may be a law, but it may have a bias towards oligarchy or democracy, so that exactly the same results will ensue.

III xi

Aristotle now takes one of the candidates for sovereignty which he had listed and discussed summarily in the last chapter, the mass of the people, and seems, at first blush, to be espousing a major democratic principle: ‘Many heads are better than one.’ There is, he argues, some advantage in giving the people political power, at least to the extent of allowing them to elect officials and scrutinize their conduct: for the fact is that collective decisions, even by persons not individually distinguished for practical wisdom (phronēsis),are commonly at least as good as those reached by a select group of ‘sound’ men. It is interesting to note that Aristotle here concedes, in the first paragraph, a degree of phronēsis and aretē (moral virtue) to the ordinary man; contrast Nicomachean Ethics, VI xiii, where both seem to require rare intellectual qualities (and cf. introduction to III iv).

Socrates and Plato had been inclined to argue that politics is or should be a skill, with recognized modes of procedure, a precise set of ends, and agreed criteria for assessing whether these ends have been achieved. Such a skill they believed to be a rare accomplishment, not to be found among the populace at large. In general terms, Aristotle shares these beliefs; hence his judicious recognition of the wisdom of collective decisions by the people may well seem more surprising than it really is. His point is not that collective decisions are or must be good because they are made by the people, who, being a majority, have some sort of entitlement to have their own way; it is that decisions about the ‘good’ life in a polis are analogous to those made by a skilled practitioner in an art or profession, but that to judge from experience they are as likely to be made correctly by the many as by the few. In this chapter, then, Aristotle merely recognizes a limited efficiency in popular judgement. For a discussion, see the reference to Braun’s first 1959 article in the second collection of essays in the Select Bibliographies.

Why is popular judgement efficient? Aristotle’s answer is brief and somewhat mechanical: one man judges one ‘part’ (of a work of art or political problem) better than others, and another man another part, so that the many collectively may judge better than an individual expert. Is it not also a matter of uncanny group-intuition? (Compare Plato’s recognition of the curious moral intuitions of even rogues, at Laws 950bc.) It is hard to resist quoting one of the stately nineteenth-century quotations Newman prints in his commentary (III 215): ‘Canning used to say that the House of Commons as a body had better taste than the man of best taste in it, and I am very much inclined to think that Canning was right’ (Lord Macaulay, Letter of February 1831).

Like its predecessor, this chapter concludes with a ringing declaration of faith in the rule of law; but the discussion of it is again brief and inconclusive.

1281a39 The other possibilities are to be discussed elsewhere;1 at the moment it would seem that one view put forward – namely that the mass of the people ought to be sovereign, rather than the best but few – is not without difficulty, but has perhaps some truth in it. For it is possible that the many, no one of whom taken singly is a sound man, may yet, taken all together, be better than the few, not individually but collectively, in the same way that a feast to which all contribute is better than one supplied at one man’s expense. For even where there are many people, each has some share of virtue and practical wisdom; and when they are brought together, just as in the mass they become as it were one man with many pairs of feet and hands and many senses, so also do they become one in regard to character and intelligence. That is why the many are better judges of works of music and poetry: some judge some parts, some others, but their collective pronouncement is a verdict upon all the parts. And it is this that gives sound men their superiority over any individual man from the masses. Handsome men differ from ugly, it is said, and paintings from actual objects, just because they draw together into one what was previously scattered here and there, though any one of the features taken separately, the eye of one man, some other part of another, might well be more handsome than in the picture.

1281b15 But it is not at all certain that this superiority of the many over the sound few is possible in the case of every people and every large number. There are some, by heaven, among whom it would be impossible: otherwise the theory would apply to wild animals – and yet some men are hardly any better than wild animals. But there is no reason why in a given case of a large number we should not accept the truth of the point we have made.

1281b21 These considerations enable us to solve the former problem,2 and also another and related question – in what spheres is this sovereignty of the free men, the mass of the citizens, to be exercised? We must remember that they are not men of wealth, and have no claim to virtue in anything. To let them share in the highest offices is to take a risk: inevitably, their unjust standards will cause them to commit injustice, and their lack of judgement will lead them into error. On the other hand there is a risk in not giving them a share, and in their non-participation, for when there are many who have no property and no honours they inevitably constitute a huge hostile element in the state. But it can still remain open to them to participate in deliberating and judging.

1281b32 It was for this reason that both Solon3 and some of the other lawgivers give to the people power to elect officials and to demand an account from them at the end of their tenure, but no right individually to hold such offices. This was on the principle that the whole body acting together has the necessary perception,4 even though each is individually only partly qualified to judge. By thus mixing with the better sort, they render good service in their states, in something like the way that a combination of coarse foods with refined renders the whole diet more nutritious than a small amount of the latter.

1281b38 But such an arrangement of the constitution raises a number of difficulties. First, the proper person to judge whether a piece of medical work has been properly done is the same sort of person as is actually engaged on such work, on curing the patient of his present sickness – in other words the medical practitioner himself. And this is equally true of the other skills and empirical crafts. As then it is among doctors that a doctor should give an account of himself, so also should other professional men among their peers. By doctor I mean not only the ordinary practitioner and the master-craftsman, but also those who have been trained in the art, such as are to be found in pretty well all skills. And we let these trained persons judge no less than we let the experts.

1282a7 Second, in the matter of elections the same would seem to apply. Choosing aright is a task for the experts – those in surveying choose surveyors, those in navigation choose navigators, and so on. Admittedly in some jobs and some skills you will find laymen who share in the choice, but not more than the experts. So it would seem that on this argument the mass of the people should not be given the sovereign power either of choosing officials or of calling them to account.

1282a14 Perhaps, however, not all these arguments are right. First, there is the argument which we used a while back5 – that provided the mass of the people is not too slave-like, each individual will indeed be a worse judge than the experts, but collectively they will be better, or at any rate no worse. Secondly, there are tasks of which the actual doer will be neither the best nor the only judge, cases in which even those who do not possess the skill form an opinion on the finished product. An obvious example is house-building: the builder can certainly form an opinion on a house, but the user, the household-manager, will be an even better judge. So too the user of a rudder, the helmsman, is a better judge of it than the carpenters who made it; and it is the diner not the cook that pronounces upon the merits of a dinner. I think perhaps these two arguments are sufficient to resolve the question.

1282a24 But there is another and connected difficulty. It seems odd that inferior persons should have sovereign control over more important matters than the respectable sort; yet the choice of officials and the scrutiny of their work are a very important matter indeed. But as we have said, in some constitutions they are assigned to the people, the assembly having sovereign control of all such matters; yet although members of the assembly, and also of the council, and of the panels of jurymen, are recruited on low property-qualifications and are of any age, nevertheless for treasury officials, generals, and the supreme offices a high property-qualification is demanded.

1282a32 This difficulty too can be met, and this practice6 too perhaps justified, by arguments similar to those which we have just used: it is not the individual juryman, councillor, or member of assembly who rules, but the court, the council and the people; and of these each individual mentioned – I mean the councillor, assembly-member and juryman – is a part. So it is quite just that the mass of the people should be in sovereign control of more important things, since people, council, and law-court all comprise many persons. And as for property-qualifications, the sum total of property owned by them is larger than that of those who, singly or with a few colleagues, hold the highest offices.

1282a41 These matters may be regarded as settled in that way; but we must look back at our original problem,7 from which nothing emerges so clearly as the fact that the laws, if rightly established, ought to be sovereign, and also that officials, whether individually or as a body, ought to have sovereign power to act in all those various matters about which the laws cannot possibly give detailed guidance; for it is never easy to frame general regulations covering every particular. We said ‘laws rightly established’, but we have not yet discovered what sort of laws these ought to be, so the old problem8 remains. For as constitutions vary, simultaneously and in like manner the laws too inevitably vary, and are sound or bad, just or unjust; but this much is clear, that the constitution must set the pattern for the laws. If however that is so, laws framed in accordance with one of the right types of constitution will inevitably be just, but if according to one of the deviations, unjust.9

III xii

This chapter, being reminiscent of I i ad init. and of the opening chapters of the Nicomachean Ethics, looks like something of a fresh start; but it turns out to raise much the same problem as was discussed in III ix, and to treat it in much the same way (see introduction to that chapter, and cf. V i). On the notion of ‘proportionate’ equality (i.e. greater shares for greater merit, as distinct from identical shares irrespective of merit), useful background reading would be Plato, Laws 756e–758a.

1282b14 In every kind of knowledge and skill the end which is aimed at is a good. 1 his good is gieatesl, and is a ‘good’ in the highest sense, when that knowledge or skill is the most sovereign one, i.e. the faculty of statecraft.1 In the state, the good aimed at is justice; and that means what is for the benefit of the whole community. Now all men believe that justice means equality in some sense, and they are in limited agreement with the philosophy of justice which I explained in my Ethics:2 they hold that justice is some entity which is relative to persons, and that equality must be equal for equals. The question we must keep in mind is, equality or inequality in what sort of thing? For this is a problem, and one for which we need political philosophy.3

1282b23 It is possible to argue that superiority in any good whatever justifies unequal distribution of offices, given that in all other respects than this the persons are not different but similar – for differences in them would mean different justice and different deserts. But surely, if that be granted, we shall have to allow that superiority in height or complexion or any other good thing will confer an advantage in political rights.4 Is not the fallacy here pretty obvious? A comparison with other kinds of knowledge and faculties shows that it is. For if, say, pipe-players are equal in skill, we must not give an advantage in instruments to those of better birth, for that would not enable them to play any better. The use of the better instrument ought to belong to the better performer.

1282b34 If this is not sufficiently clear, it will become so if we sharpen the example. If one man is outstandingly superior in pipe-playing. but far inferior in birth or good looks (even supposing that birth and good looks are each a greater good than the skill of pipe-playing, and its superiority to them is greater in proportion than the superiority of this player’s ability to that of the rest), even then, I say, he should still get the best pipes. For superiority both in wealth and in birth ought to contribute to the quality of the performance – to which these qualities in fact contribute nothing at all.

1283a3 Moreover, according to that way of reasoning, every good thing would be commensurable with every other good thing. For if marks are given for a particular degree of tallness, then tallness in general would be in competition with both wealth and freedom. So if we say that X has greater superiority in height than Y has in virtue, then even if in general virtue is of greater importance than height, we are making everything commensurable with everything; since if one amount is greater than some other, clearly there is another which is equal.5 But such mensuration is quite impossible here, so it is clear that in matters relating to the state men are quite right not to take any and every kind of inequality into account in competing for offices, and only those differences which contribute to making up the state as a whole should be urged in the competition. Such qualities as superior swiftness of foot, however important that may be in winning honour in athletic contests, should not entitle one to take more than the next man.

1283a16 Hence those of noble birth or who are free or have wealth are quite right to lay claim to honours, since the members of the state must be free and must have taxable property (you could no more make a state out of paupers than out of slaves). But obviously something more is needed besides: I mean justice, and the virtue that is proper to citizens.6 For without these additions it is not possible for the state to be managed. More exactly, whereas without free population and wealth there cannot be a state at all, without justice and virtue it cannot be managed well.

III xiii

In the early paragraphs of this chapter Aristotle examines how the various claims to political power (notably wealth, birth and superior strength of numbers) may compete. He admits that each claim has a certain partial or relative ‘justice’, but he denies that any of them is ‘just’ in an absolute sense. Each demands exclusive consideration at the expense of all the others: whichever is adopted, any man, or set of men, who excelled in that respect, would have to be given power, even if he excelled in no other. For instance, on the criterion of wealth, one supremely rich man would have to be allowed to rule, even if he were also a supreme rogue. Aristotle discusses various cases where difficulties could occur: a single man could deploy greater strength than the masses, whose claim to rule normally depends on precisely this – superior strength; or a few persons claiming to rule on the grounds of overwhelming wealth may nevertheless be poorer than the masses collectively (though not individually). All such problems arise if we adopt one claim in preference to the rest: each taken by itself is, as we might say, ‘socially divisive’.

This latter point is the key to Aristotle’s solution. The only claim that is absolutely just is that of ‘virtue’ (aretē), that is, the moral and intellectual ability always needed in a ‘statesman’, a citizen who is ruled and rules by turn in the interests of the whole state, not in those of just one part which might advance one of the ‘partial’ claims (e.g. the masses, who might advance the claim of superior numbers). When a man excels to a superlative degree in ‘virtue’, he must be reckoned as a ‘god among men’ and given supreme power above the law. It is not clear, however, here at least, why superlative ability in ruling and being ruled ought to exempt one from the latter. But Aristotle follows whither he thinks the argument leads him, and comes right out and says that such a person ought to be a permanent king in the state, claiming the glad obedience of all others.

One could argue that Aristotle is here showing himself to be Platonist in spite of himself: Plato too had envisaged (notably in the Republic and Politicus) a supreme individual or individuals above the law and ruling in the interests of the state. But the positions of the two philosophers are instructively different. Plato’s supreme rulers would have owed their power to their metaphysical insight into the supreme moral values, Aristotle’s to (as it were) a claim for a quid pro quo from the state. If each person with a ‘partial’ claim to power (wealth, etc.) deserves some political power in return for the partial contribution he makes to the common good (see III ix), the person with the ‘absolute’ claim (virtue) apparently deserves absolute political power in return for his total contribution to that good. Aristotle’s approach is thus less metaphysical than ‘transactional’.

The chapter closes with the tantalizing statement that those who excel in ‘the virtue of a statesman’ (in the special sense of this latter word) should be permanent kings, but Aristotle does not enlarge on how they may be identified, appointed, or dismissed. He may perhaps even see outstanding individuals as more of a problem than an asset: certainly he seems to have a certain sympathy with the practice of ostracism, of which he has a brief but interesting historical discussion towards the end of the chapter.

1283a23 Some or all of these things would seem to have a proper claim to be contributions to a state’s existence; but I repeat that, in order to secure the good life, education and virtue would have the most just claim of all. But since those who are equal in one particular ought not to enjoy equality in all things, nor those who are unequal in one respect have inequality in all, it follows that all constitutions in which such a state of affairs prevails must be deviations.1

1283a29 It has already been stated that while all men have some kind of justice in their claims, not all of them have a claim that is just in an absolute sense.2 (a) The rich argue that they have a greater share in the land, and the land is of social3 interest; and further, that they are more to be relied upon to fulfil their contracts, (b) The claims of the free and well-born are closely related: the more nobly born are more fully citizens than the non-noble, good birth being held in esteem in every country; and the offspring of the better sort are likely to be better men, for good birth is excellence4of stock, (c) Next we shall mention the equally just claims of virtue, for we always speak of justice as a social3 virtue, and one which is sure to bring all the other virtues along with it. (d) And surely themajority have a better claim than the minority, as being stronger, richer and better, if we balance the larger numbers against the smaller.

1283a42 Now suppose all these to be present in a single city – that is to say, the good, the rich and the well-born, and beside them a mass of citizens – will there or will there not be dispute as to which should rule? Now in the three types of constitution of which we spoke earlier the decision provokes no dispute, because they differ from each other in just this respect, sovereignty being exercised in oligarchy by the rich, in aristocracy by the sound, and so on. But we have to ask ourselves how to reach a conclusion when these elements are present at one and the same time. Suppose for example that those who have virtue are exceedingly few in number – how is the matter to be settled? Are we to regard their fewness in the light of the work to be done, asking whether they are strong enough to run the state? Or are we to ask whether their numbers are sufficient to make a state?

1283b13 The problem arises in regard to all claimants to honours in the state. Those who base their claim to rule on wealth would seem to have no just claim at all, nor those who base themselves on birth; for if one man is very much richer than the rest, then clearly by the same principle of justice,5 he will have to be sole ruler over them all, and similarly one who is superior in good birth will have to rule over all whose claim is based on free status. This same thing could well happen where the constitution is an aristocracy, based on virtue; for if one man is better than all the sound men in the citizen-body, then on the same principle of justice5 he ought to be sovereign over them. Again, suppose that the multitude ought to be sovereign because they are stronger than the few, and suppose one man, or more than one but still fewer than the many, to be stronger than the rest – then these would have to be sovereign rather than the multitude.

1283b27 All these considerations seem to show that none of these criteria is right by which one set of men claim that they themselves should rule and all the rest be subject to them. For surely, whether their claim to sovereignty over the citizen-body rests on wealth or on virtue, it remains true that against their arguments the multitude will have some justice on their side; for it is quite possible on occasion for the multitude to be better than the few, and richer too, when considered not singly but together.

1283b35 So it is possible in this fashion to meet also a difficulty which some people pose and debate, namely whether a lawgiver, who seeks to lay down the laws that are most right, ought, given the circumstance mentioned,6 to legislate for the benefit of the majority or for the benefit of the better sort. By ‘right’ we ought to mean ‘equally’ right, i.e. right with respect to the benefit of the whole state and the common3 interest of the citizens. (A citizen is in general one who has a share in ruling and in being ruled; but he will not be identical in every kind of constitution. So far as the best constitution is concerned, he is a man who is able and who chooses to rule and to be ruled with a view to a life that is in accordance with virtue.)

1284a3 But if there is one man (or several, but not enough to make up the whole complement of a state) of such superlative virtue that the capacity for statecraft7 and the virtue of all the rest are simply not to be compared with his (or theirs), such men we must take not to be part of the state. To judge them worthy of mere equality with the rest would be to do them an injustice, so far unequal to them are they in virtue and in the capacity for statecraft.7 We may reasonably regard such a one as a god among men8 – which shows, clearly, that legislation too must apply only to equals in birth and capacity. But there is no law that embraces men of that calibre: they are themselves law, and anyone who tried to legislate for them would be snubbed for his pains. They might well say what the lions in Antisthenes’ fable said to the hares who asserted their claim to equality with them.9

1284a17 It is for this kind of reason that democratically organized states establish also the practice of ostracism. They appear to attach such immense importance to the principle of equality above all else that they ostracized and removed out of the state for fixed periods anyone whose power was deemed to be excessive, whether this power was due to wealth or popularity or any other influence in the state. (Here is an example from mythology: the Argonauts left Heracles behind for some such reason – because the Argo would not have on board one so vastly bigger than the rest of the crew.) This is why critics of tyranny who disapprove of the advice given by Periander are not to be considered fully justified. It is said that to Thrasybulus’ messenger, who had come for advice, Periander returned no answer; but while walking in a field, reduced all the ears of corn to one level by lopping off the tallest. The messenger did not understand the motive for this action, but reported the action to Thrasybulus, who perceived that he ought to remove the outstanding men.10 The method is useful not only to tyrants, and tyrants are not alone in practising it: oligarchies and democracies are in just the same position, for ostracism has very much the same effect as lopping off and exiling the leading men. And it is the regular practice of the holders of sovereign power, in their dealings with other states and with foreign nations. For example, the Athenians, as soon as their empire11 was strong enough, reduced Lesbos, Chios, and Samos to submission,12 contrary to the terms of the agreement; and the Persian king often trimmed the Medes, Babylonians, and others who prided themselves on their earlier supremacy.11 Indeed this whole question concerns all constitutions, not merely the divergent ones, which resort to such methods for their own advantage, but also right forms of constitution, which aim at the common3 good.

1284b7 This same point may be observed also in the other skills and fields of knowledge. A painter would not allow his representation to have one foot disproportionately large, however magnificent the foot might be. A shipbuilder would not let the stern, or any other part of the ship, be out of proportion. A chorus-master will not allow among the members a performer whose voice is finer and more powerful than all the others. On this showing there is no reason at all why monarchs should not remain on good terms with their states, provided that in taking this action13 their own rule is beneficial to those states. Therefore the theory behind ostracism has some measure of political justice,14 in cases of admitted disproportion. Of course if the lawgiver can so construct the constitution from the start that there will never be any need of this kind of medicine, so much the better. Otherwise the best we can do, if occasion arises, is to put matters right by some such method. (In fact this is not the way it worked out in the states, because instead of seeking the advantage of their own constitution, men used ostracism as a weapon in factional strife.)

1284b22 So in the deviation forms of constitution ostracism is obviously exercised for some personal benefit,15 and is just; though it is perhaps equally obvious that it is just in a limited sense only. But when it comes to the best type of constitution, there is a real difficulty, not in the cases of superiority in the other goods, such as strength16 or riches or popularity, but where a man is pre-eminent in virtue. What are we to do then? Men will not say that such a person ought to be banished or deported; nor yet that they ought to rule over him, for that would be like claiming to rule over Zeus, by dividing up the offices.17 It only remains therefore to let nature take its course; he will govern and we will all gladly obey him. Thus such men will be permanent kings in their states.

III xiv

In III vii six forms of constitution were set out, three ‘right’ (kingship, aristocracy, polity), and three ‘deviations’ (tyranny, oligarchy, democracy). Kingship and tyranny were both ‘rule by one man’, the literal meaning of the Greek word monarchia, monarchy. The intervening discussions about justice and sovereignty in the state have been carried on with the rule of the few or of the many as a constitutional background, and there has been no mention of monarchy – naturally enough, since it would be only slightly relevant to questions about sovereignty and the just distribution of political power, and not relevant at all if the monarchy’s powers are unlimited. But obviously not all monarchies have such powers, and most of this chapter is devoted to discussing four types of limitedmonarchy (the Spartan, the non-Greek, ‘aisumnēteia’, and the ‘heroic’), in the light of four main criteria: (a) whether they are subject to law, (b) whether their tenure is for life or for a set term, (c) whether they are elective, (d) whether they are over willing subjects. Confusingly, Aristotle calls all these four monarchies ‘kingships’, but admits that two are effectively tyrannies. Lastly he describes a fifth kind of monarchy, which presumably looks back to the unfettered kingship of the ‘god among men’ envisaged at the end of the last chapter, and forward to the ‘absolute’ kingship mentioned at the beginning of the next. The discussion is pleasingly enriched by Aristotle’s intimate knowledge of Greek history, and by his felicitous quotations from Homer and Alcaeus.

1284b35 After what has just been said it will be a good thing to change the subject and consider kingship, since we hold that it is one of the right forms of constitution. We have to inquire whether or not a king’s rule, rather than some other constitution, is advantageous for the good management of any country or state, or if it is advantageous for some and not for others. But first we must decide whether there is only one kind of kingship or several different varieties of it. It is easy to see that there are several kinds, and that the mode of rule is not the same in every case.

1285a3 The clearest example of kingship according to law is the one to be found in the Spartan constitution. It is, however, not sovereign over everything; though when a king goes out of the country he is leader in all matters relating to the war, and to the kings1 is committed also the care of religious matters. So such a kingship is like a perpetual generalship held on terms of personal authority, in that the king has no sovereign power to put anyone to death except for cowardice, as on military expeditions of old, by law of force. There is a clear example in Homer: Agamemnon put up with being abused in meetings of assemblies, but once an expedition had begun, he had sovereign power of life and death. So much is implied in his words: ‘Anyone whom I shall catch absent from the fighting… his body shall become the prey of dogs and birds, and no escape, for the power of death is in my hand.’2 This then is one type of kingship – a generalship tenable for life, which may be acquired either by birth or by election.

1285a16 Alongside this there is another type of monarchy, such as kingships found among certain non-Greeks. All these have power approximating to that of tyrannies, but they are legally established and ancestral. For it is because non-Greeks are by natural character more slavish than Creeks (and the Asiatics than the Europeans) that they tolerate master-like rule without resentment. Therefore, while such kingships are for these reasons like tyrannies, their legality and ancestral status make them safe. And for the same reason the ruler has a royal, not a tyrant’s bodyguard; for a king’s bodyguard is composed of citizens carrying arms, a tyrant’s of foreigners. And the king rules over willing subjects according to law, the tyrant over unwilling subjects; so that whereas the one draws his bodyguard from among his citizens, the other uses it against them.

1285a29 These then are two types of monarchy; and there is a third, which used to exist among Greeks of old. This third type is called aisumnētēs, and was in rough terms an elective tyranny. It differs from the non-Greek monarchy only in not being ancestral; it is equally subject to law. The rulers held office sometimes for life, sometimes for a stated period or until certain things should be accomplished; for example the people of Mytilene elected Pittacus for the purpose of repelling the exiles who tried to come back led by Antimenides and the poet Alcaeus.3 That Pittacus was chosen is clear from one of Alcaeus’ banqueting songs in which he grumbles that ‘With mass-adulation they appointed low-born Pittacus to be tyrant of their easy-going and unlucky state’. Of these kinds of rule, we may say that by being like that of a master, they are and were like tyrannies; but royal, in being elective, and in being exercised over willing subjects.

1285b3 There is a fourth kind of royal monarchy, which existed in heroic times. It was both ancestral and subject to law, and willingly accepted by its subjects. The first of these kings had been benefactors of the mass of the people in the arts of peace or in warfare, or in welding the people together, or in providing them with land. So they became kings, willingly accepted by their subjects, and ancestrally established in the eyes of their successors in the next generation. They held sovereign control of leadership in war, and of sacrifices not reserved for priests. They also gave judgements at law; some did this on oath, some without oath, the oath being the raising aloft of the royal sceptre. In early times these kings ruled continuously over the affairs of the city, of the country, and across the borders; but later in some cases they themselves relinquished some of their duties, in others they were deprived of them by the populace. The duty of offering sacrifices was in some states the only one which was left in the hands of the kings; and even where one could justifiably say a kingship did still exist, they retained only the leadership of armies on expeditions beyond the borders.

1285b20 These then are the four forms of kingship: (a) that of heroic times, willingly accepted by the people and exercised on certain specified conditions, the king being general, judge, and religious head; (b) the non-Greek rule, acquired by birth, exercised as by a master, and subject to law; (c) what men call aisumnēteia, i.e. an elective dictatorship; and (d) the Lacedaemonian, in rough terms an hereditary permanent generalship. These then are the distinguishing marks of these four kinds. But there is also a fifth, in which one man single-handed is in sovereign control of everything, in the same way as each state or foreign nation controls its own public affairs. This fifth kind comes under the same heading as household-management, for just as household-management is, as it were, the kingship of the household, so this kingship is the household-management of a state, or of a foreign nation or nations.


In this chapter a number of related problems are raised, discussed and dropped in somewhat bewildering succession. The central themes are (a) the contrast, indeed the tension, between the personal rule of a king and the rule of law, and (b) the assessment of their relative merits.

The first four types of kingship described in the preceding chapter were all in some way restricted; the fifth was unrestricted. Hence there are really only two main classes. Of the four limited monarchies the Lacedaemonian is now taken for comparison because it is the most restricted of all. But for that very reason it is not typical of monarchy; and since limitation of royal authority involves some distribution of power, which takes place also under any other system of government, it seems that only absolute kingship needs to be discussed. Now this at once raises a problem. When Aristotle was discussing distribution of power under oligarchy or democracy, he came to the conclusion (at the end of III xi) that where possible the law should be sovereign. But at the end of the last chapter and the beginning of this, he postulates a king who makes his own laws and whose superlative excellence appears to entitle him to do so. The answer to the question what is to be done in such an event is not given till the last two chapters of the book (xvii–xviii), and the discussion in the present chapter and the next has relevance to any form of slate in which personal authority and legal authority could exist side by side. On the Platonic background, see e.g. Politicus 294a ff.

Towards the end of the chapter Aristotle digresses to speculate on the origin of kingship and on the course of early constitutional development. The reconstruction he gives in IV xiii is different; but in an age when adequate source-material for writing exact and detailed history of early times barely existed, it is hardly surprising that both Plato’s and Aristotle’s historical accounts are somewhat generalized and free, and commonly adjusted to the theme in hand.

The chapter closes with a discussion of some practical difficulties of kingship; the observation that even the rule of law needs to be backed by a degree of force is chilly and realistic.

1285b33 We may say then that there are really only two types of kingship to be considered – the foregoing and the Lacedaemonian. Most of the others fall between these two extremes, as having sovereign power over fewer matters than absolute kingship has, but over more than the Lacedaemonian. So there are two questions to be asked, first whether or not it is expedient for states to have a perpetual general, appointed either on grounds of birth or by turns, and second whether or not it is expedient that one man should have sovereign powers in all matters. The former question, relating to generalship of the kind mentioned, is of a type concerned with laws rather than constitutions, since it is possible for such a thing to exist under any constitution. I am therefore leaving it aside and concentrating on the other style of kingship, since it is a type of constitution. We must therefore examine it and run over the problems that we find there.

1286a7 We begin by asking whether it is more expedient to be ruled by the best man or by the best laws. Those who believe that to be ruled by a king is expedient think that the laws enunciate only general principles and do not give day-to-day instructions on matters as they arise; and so, they argue, in any skill it is foolish to be guided always by written rules. In Egypt a doctor is allowed, rightly, to depart from his treatment if the patient is not well after four days; if he does so earlier, he does it at his own risk. For the same reason, it is obvious that the constitution which goes by laws and written rules is not the best. On the other hand, rulers cannot do without that general principle in addition: it provides something which, being without personal feelings, is better than that which by its nature does feel. Every human soul must have feelings, whereas a law has none; but in compensation, one might say, a man will give sounder counsel than law in individual cases.

1286a21 It seems clear then that this man must be a lawgiver, and that laws must be laid down, which shall be sovereign in all cases, except those in which they go awry.1 But when the law either cannot decide at all or will only decide badly, ought the power to rule to rest with the one best man or with all? In our own day verdicts are given, deliberations are carried on, and decisions are arrived at by men acting together; and all these decisions refer to separate problems and individual cases. Now any one of these individuals, measured by his own contribution alone, may be inferior. But the state consists of many men, and this gives it the same kind of superiority as a communally provided banquet has over a single simple meal. So a crowd is, on numerous occasions, actually a better judge than one man, whoever he may be.2 Again, the many are less easily corrupted. As a larger amount of water is less easily polluted, so the multitude is less easily corrupted than the few. The judgement of one man is bound to be corrupted if he is in a bad temper or has very strong feelings about something. But in the other case it would take a lot of doing to arrange for all simultaneously to lose their temper and go wrong.

1286a36 But we must make sure, first, that the multitude are also the free-born, and second that they depart from the provisions of the law only in cases which the law itself inevitably fails to cover. This second proviso may not be at all easy where numbers are large; but if the good, that is good men and good citizens, are in the majority, then when we put the question, ‘which is the less liable to be corrupted, the sole ruler or the numerical majority who are all good?’, the answer is obviously, ‘the majority’. And if it is urged that the many will split into factions, which the one cannot do, perhaps the answer must be that they too are sound souls,3 no less than your ‘one man’. If then we are to describe this rule of the majority who are all good men as aristocracy,4 and the rule of one as a kingship, then aristocracy in a state will be preferable to kingship, whether or not rule is exercised by armed force, provided it is possible to get a homogeneous majority.

1286b8 Perhaps we have here a clue to the reason why royal rule used to exist formerly, namely the difficulty of finding enough men of outstanding virtue, all the greater since in those days the states they inhabited were small. An especial function of good men is to confer benefits, and it was in recognition of the benefits that they had conferred that men were appointed to be kings. Then, when a large number of men of similar virtue became available, people no longer tolerated one-man rule but looked for something communal, and set up a constitution. But the good men did not remain good: they began to make money out of that which was the common property of all. And to some such development we may plausibly ascribe the origin of oligarchies, since men made wealth a thing of honour. The next change was to tyrannies, and from tyrannies to democracy. For the struggle to get rich at all costs tended to reduce numbers,5 and so increased the power of the multitude, who rose up and formed democracies. And now that there has been a further increase in the size of states, one might say that it is hard to avoid having a democratic constitution.

1286b22 If anyone should hold that it is best for states to be ruled by kings, he will have to consider a question relating to the king’s children. Are his offspring also to be kings? Considering what kind of persons some of these have turned out to be, we would have to say that hereditary succession is harmful. You may say the king, having sovereign power, will not in that case6 hand over to his children. But it is hard to believe that: it is a difficult achievement, which expects too much virtue of human nature.

1286b27 Then there is this question of armed force: is the intending king to have about him a force with which he will be able to impose his will on those who seek to resist his rule? How else is he to exercise his authority? For even if his sovereignty is such that he can act only in accordance with law, and do nothing of his own volition that is illegal, it will still be necessary for him to have sufficient armed force to give the laws protection. This question, in so far as it relates to this kind of king,7 is perhaps not difficult to answer. He must indeed have a force, and it should be made strong enough to overpower one man or a band of men, but not the multitude. This is the principle which was followed in earlier times, when guards were assigned to a man who was being appointed as ‘tyrant’ or ‘aisumnētēs’, as they called him, of the state; so too, when Dionysius asked for these guards, somebody advised the Syracusans to limit them to just the number required for that purpose.

III xvi

Aristotle has wandered away from the question posed at the beginning of III xv. All these discussions about the relations between monarchy and law seem to have been concerned not with absolute but with ‘legal’ monarchy – a subject which he professed to have put aside in order to consider ‘absolute’ monarchy. To this he now returns, repeats something of what he said at the beginning of Chapter xv, and then once again discusses the authority of the law and the authority of the individual. It looks as if there may have been at one time a double set of Aristotle’s notes; yet although xv and xvi overlap a good deal, xvi does contain a number of new points, notably: (a) That law is in some sense a ‘mean’. Aristotle was in general inclined to think of virtues as means between extremes (seeNicomachean Ethics, II vi–ix), and we have here evidently an attempt to extend the analysis to law. From the drift of his discussion, it looks as if he thought of law as a mean in the sense that it constitutes a compromise between unfettered personal discretion on the one hand and slavish adherence to written rules on the other (cf. Plato, Laws 875d ff, 925–926a). Or is his point simply that law is a mean in the sense that it is impartial? (b) That law is in some sense ‘natural’. Again, Aristotle does not discuss the point at length. The argument seems to rest on the natural equality of men, which entitles them all to a share of ruling according to some agreed and hence ‘legal’ system of alternation.

However, it should be noted that the entire chapter is written rather confusingly: ostensibly, at any rate, it is an account of the arguments of certain anti-monarchical polemicists, and it is not clear precisely where Aristotle’s own comments, if any, begin and end.

1287a1 We must now turn to consider our next subject, the king whose every act is in accordance with his own personal volition. For the king who is ‘subject to law’ does not, as has already been said,1 amount to a form of constitution, because perpetual generalships may exist in democracy or in aristocracy or in any other constitution, and it is not unusual to put one man in sovereign control of the whole administration – witness the government of Epidamnus and, to a lesser extent, of Opus. But we are now speaking of ‘absolute’ kingship, meaning by that one whereby the king rules over everything according to his own volition.

1287a10 There are some who hold that it is not even in accordance with nature2 that one man should be sovereign over all the citizens, when the state is made up of persons who are alike. For, they say, those who are by nature alike must get the same natural justice and deserts; and so, if it is bad for the health of unequal persons to have equal food or clothing, this is also applicable to honours; and the converse also is true. Justice therefore demands that no one should do more ruling than being ruled, but that all should have their turn. So we are back again with law, for organizationis law. It follows therefore that it is preferable that law should rule rather than any single one of the citizens. And following this same line of reasoning further, we must add that even if it is better that certain persons rule, these persons should be appointed as guardians of the laws and as their servants. Offices there must be, they say, but it is not just that there should be only one man in office, at any rate where all men are alike

1287a23 Again, though there are matters about which the law appears incapable of giving a decision, in such cases a human being too would be unable to find an answer. It is in order to meet such situations that the law expressly educates the officials, and empowers them to decide and to deal with these undetermined matters to the very best of their just judgement.3 Moreover, it allows for amendments to be made, wherever after experiment a new proposal is thought to be better than the established practice. Therefore he who asks law to rule is asking God and intelligence and no others to rule; while he who asks for the rule of a human being is importing a wild beast too; for desire is like a wild beast, and anger perverts rulers and the very best of men. Hence law is intelligence without appetition.

1287a32 Another argument is that which employs an analogy with medicine or other professional skills, and alleges that it is a bad thing to practise medicine according to written rules, it being preferable to call in those who possess the skills. But this analogy is false. The doctor does not do anything for friendship’s sake that is against his rational judgement: he cures his patient and takes his fee; but people in offices of state usually do all manner of things to show favour or disfavour. If you suspected that the doctors had been bribed by your enemies to make an end of you, then you would naturally prefer to get treatment according to written rules. Again, doctors when ill call in other doctors to treat them, and trainers other trainers when they themselves go into training – on the principle that it is impossible to give true judgement when their own interests and their own feelings are involved. So it is clear that the search for what is just is a search for the mean; for the law is the mean. Further, customary laws influence us more crucially,4 and in more crucial4 matters, than written laws; so that though a man, as ruler, is less fallible than written laws, he is no less fallible than customary laws.

1287b8 Of course, if a man is the only ruler, there will be much that he cannot easily supervise; he will therefore need to have several other officials appointed by him, so what difference does it make whether they were there at the beginning of his rule or were appointed subsequently by him in virtue of his position as sole ruler? Besides, if, as has been said before, a sound man has a right to rule because he is better, then two good men are better than one. Hence the expression ‘Let two go together’, and Agamemnon’s prayer, ‘Would that I had ten such counsellors.’5

1287b15 In our own day too the officials (for instance judges) have the sovereign power to give decisions about some matters on which the law is not able to decide. But it is universally agreed that where the law is capable, it is the law’s rule and decisions that will be best. But there are things which can and others which cannot be included in laws; and it is the latter that give rise to difficulties and raise the old question: Which is preferable, the rule of the best man or the rule of the best law? Among the matters which cannot be included in laws are those which are the subjects of deliberation. Yet they6 do not deny the inevitability of such decisions being taken by a man; they merely say that there should be not one man only but many.

1287b25 An individual ruler, if he has been well educated by law, gives good decisions; but he has only one pair of eyes and ears, one pair of feet and hands, and it would be a paradox if he had better vision in judgement and action than many men with many pairs. Monarchical rulers, as we see even in our own times, appoint large numbers of men to be their eyes and ears, hands and feet; for such people as are friendly to themselves and to their rule, they make sharers in it. If they are not friends, they will not act according to the monarch’s intentions. On the other hand if they arefriends, both of the monarch and his rule, then, since a friend is equal and similar, if the monarch thinks that these men ought to rule, he thinks thereby that equal and similar persons ought similarly to rule.

These, roughly stated, are the contentions of the opponents of kingship.

III xvii

In this chapter Aristotle completes his discussion of kingship and law, and decides, as he has already suggested at the end of III xiii, that when a single individual is utterly outstanding in virtue, he ought to be given a permanent all-powerful and ‘absolute’ kingship, above the law. Two related points (among many it would be possible to make about this fascinating thesis) may be indicated here: (i) Aristotle seems not to envisage such a ruler ceasing to be wholly virtuous: indeed he apparently thinks that such virtue may be transmitted genetically, or at least by education within the same family. ‘Once wholly virtuous, always so’ would seem to us political naïveté; or does Aristotle conceive such a king merely as a theoretical possibility? (2) In the very suggestive second paragraph, Aristotle describes three kinds of society, and the kind of ruler which each naturally produces and which is suitable for it; one of these three is the ‘royal’ society, which produces the wholly virtuous and absolute king. Here again, the possibility of change seems not to be allowed for: could not societies, by internal change, progress to or regress from this ideal? Aristotle was certainly aware of the fact that a state may change its constitution over a period, and his analysis here seems curiously static – though it is admittedly brief and exploratory.

1287b36 But these observations, though valid in some cases, are perhaps not valid in others. For while in the natural order of things ‘just’ and ‘expedient’ vary in sense according to whether the ruler is a master, king or statesman,1 there is no natural sense in which they can apply to the rule of a tyrant, nor to any of the other constitutions which are deviations, for they develop in a manner contrary to nature. But from what has already been said, it is clear that, among those who are equal and alike, it is neither just nor expedient that one single man should be sovereign over all the rest, whether he rule with laws or without (he himself then being law instead), or whether he is a good man ruling over good men, or not-good over not-good; nor even if he is superior in virtue – except in certain circumstances. What these circumstances are I must now state, though to some extent they have been stated already.2 But first I must define what is meant by the terms ‘royal’, ‘aristocratic’, and ‘of statesmen’.1

1288a8 A populace is ‘royal’ if it is of such a kind as naturally to produce a breed3 of outstanding virtue fitting it for leadership in the state. A populace is ‘aristocratic’ if it is such as naturally to produce a breed able to be governed as befits free persons by those whose virtue equips them to be leaders in the holding of office in the state. A populace is ‘of statesmen’1 if it is one in which there naturally grows up a breed of citizens able to rule and be ruled according to a law which distributes offices on a basis of merit to those who are well-off financially.

1288a15 When therefore either a whole family or a single individual among people at large can be found, whose virtue is so outstanding as to outstrip that of all the rest, then it becomes just that this family should be royal and sovereign over all things, and that this one man should be king. For, as has been said earlier, this meets the requirements of that justice which men are wont to demand in setting up constitutions, whether they are making them aristocratic or oligarchic, or, again, democratic. For they all base claims on superiority, though the superiority is not the same in each case. It also accords with the position adopted previously,2 for of course it would be improper that such a man4 should be put to death or exiled or ostracized or required to be ruled over in his turn.5 For while the part5 is not naturally superior to the whole, yet superiority is exactly the position of one so outstandingly excellent as I have described. There is therefore nothing for it but to obey such a man and accept him as sovereign, not in alternation5 but absolutely.

1288a30 On kingship then and on its various forms and whether or not it is advantageous for states, and for which states, and in what circumstances, to all those questions let these be our answers.

III xviii

In the preceding chapters Aristotle has given really very little detail about absolute kingship, either the kind of man he needs to be, or the circumstances in which absolutism is a good thing. This makes it impossible to say with certainty whether he had Alexander the Great in mind or not. On the whole it seems unlikely: see V. Ehrenberg, Alexander and the Greeks (Oxford, 1938), ch. 3, but cf. H. Kelsen, ‘Aristotle and Hellenic-Macedonian policy’, in Articles on Aristotle II (see Select Bibliographies). At least we can say that there is no mention of great conquests as a necessary qualification for monarchy.

The chapter is by way of a tail-piece to the book and looks forward to the discussions of Book VII. It makes a single and valuable point about the education which produces the ideal monarch: that it will be virtually identical to that of the ‘sound’ man who is a citizen of the best state. It is this person, somehow writ large, who seems to be the absolutely virtuous absolute king Aristotle envisages.

1288b32 But since we say that there are three right constitutions,1 and that of these the best must be of necessity that which is managed by the best men (and that may mean one man alone, or one entire family,2 or a populace outstanding in virtue, some able to rule, others to be ruled, with a view to the most desirable life), and since further it was shown in the initial discussions3 that in the best state it is bound to be the case that the virtue of a man and of a citizen are identical, then it follows clearly that the method and means one would employ to establish a state under an aristocracy or king, and those by which a sound man is produced, are the same, so that the education and morals4 that make a man sound, and those that will make him fit to play the part of a statesman or of a king, are also more or less identical.

1288b2 Now that these matters have been settled we must next endeavour to describe how the best constitution naturally comes about and how it is naturally established.

The promise to examine the best constitution is redeemed in Books VII and VIII (hence it was for a time customary for editors to renumber these two books and print them after Book III). In the Books IV–VI the emphasis is however not on the absolutely best constitution but on the most serviceable kind for actual use. To this the name ‘polity’ is given, as in III vii. There is a general cohesiveness about Books IV, V and VI (see, in the Reviser’s Introduction, ‘The Contents and Structure of the Politics’), and in the second chapter of Book IV Aristotle gives a partial table of contents. But when it comes down to detail there is much disarray.

If you find an error or have any questions, please email us at Thank you!