A Third Concept of Liberty

My starting point is one of the claims most widely accepted in current discussions about the theory of liberty. There is one overarching formula, we are told, under which all intelligible locutions about freedom can be subsumed. The prevalence of this belief appears to be due in large part to the influence of a single classic article, Gerald MacCallum’s ‘Negative and Positive Freedom’ (1967). Whenever the freedom of an agent is in question, MacCallum maintains, it will always be freedom from some element of constraint on doing or becoming (or not doing or becoming) something. Consequently, to speak of the presence of freedom is always to speak of an absence: absence of constraint on an agent from realising some goal or end. There is, in other words, only one concept of liberty.

These observations bring me to Isaiah Berlin, a thinker who devoted himself to many disparate themes, literary and historical as well as philosophical, but whose most important and influential work was on the theory of freedom. It is on that topic that I propose to concentrate. I shall focus in particular on Berlin’s most celebrated contribution to the debate, his essay entitled ‘Two Concepts of Liberty’, recently republished in a handsome new collection.*

The abiding merit of Berlin’s text is that, by contrast with the conventional wisdom I began by citing, Berlin succeeds in showing that a strong distinction needs to be marked between two rival and incommensurable concepts of liberty, which he, too, labels ‘positive’ and ‘negative’. When he discusses negative liberty, he gives an account closely resembling the analysis that, according to MacCallum and his numerous followers, must be given of any claim about freedom if it is to be intelligible. To see, therefore, where Berlin has something challenging to add to the argument, we need to turn to his account of what he calls positive liberty.

Berlin’s attempt to mark off this separate concept is admittedly dogged by several false starts. He begins by suggesting that, whereas negative liberty is freedom from constraint, positive liberty is freedom to follow a certain form of life. But this distinction cannot be used to disclose two different concepts of liberty, if only because all cases of negative liberty are at once cases in which I am free from constraint and in consequence free to act should I choose. Berlin next suggests that the positive sense of the word refers to the idea of being one’s own master as opposed to being acted on by external forces. But this, too, fails to isolate a separate concept of positive liberty. For the situation in which I am free to act in virtue of not being hindered by external forces is, according to Berlin’s own analysis, the situation of someone in possession of their liberty in the ordinary negative sense.

It soon emerges, however, that Berlin’s concern is not with the idea of being your own master. Rather, he is interested in the very different notion (although he sometimes runs them together) of mastering your self. When he first employs this formula, he uses it to refer to the thought – equally familiar to students of Plato and of Freud – that the obstacles to your capacity to act freely may be internal rather than external, and that you will need to free yourself from these psychological constraints if you are to behave autonomously. But this, too, fails to capture a separate concept of positive liberty. For while the notion of an internal obstacle extends the range of things that can count as constraints, we are still speaking about the need to get rid of an element of constraint if we are to act freely, and are still speaking in consequence about the idea of negative liberty.

The principal claim, however, that Berlin wishes to make about self-mastery proves to be a different and more convincing one. According to those who have wished to give a positive content to the idea of liberty, he suggests, the freedom of human agents consists in their having managed most fully to become themselves. Freedom is thus equated not with self-mastery but rather with self-realisation, and above all with self-perfection, with the idea (as Berlin expresses it) of my self at its best. The positive concept is thus that, as Berlin finally summarises, ‘whatever is the true goal of man . . . must be identical with his freedom.’

If there is any one philosopher whom Berlin had in mind in formulating this definition, I think it must have been Bernard Bosanquet. In The Philosophical Theory of the State, first published in 1899, Bosanquet speaks in so many words about the ‘negative idea’ of being ‘free from constraint’ and contrasts this juristic concept, as he calls it, with what he describes as the ‘fuller’ or ‘positive’ understanding of the term. Furthermore, when Bosanquet characterises the negative ideal as that of being preserved against trespass, and contrasts it with the positive view of the ‘real’ or ‘ideal’ self whose activity is identical with freedom, Berlin echoes his phraseology almost word for word.

Behind Bosanquet’s analysis, however, lies the overwhelming influence of T.H. Green. As Bosanquet acknowledges in the chapter I have been quoting, he makes ‘great use’ of the analysis of freedom offered by Green in his Principles of Political Obligation, originally published in 1886. Green does not explicitly speak in that work (although he does elsewhere) of ‘positive’ liberty, but he provides a subtler and more careful analysis than Bosanquet does of what might be meant by giving a positive content to the ideal. ‘Real freedom,’ according to Green, ‘consists in the whole man having found his object.’ To speak of the freedom of a man is thus to speak of ‘the state in which he shall have realised his ideal of himself’. Freedom is, in short, the name of an end-state; as Green concludes, it is ‘in some sense the goal of moral endeavour’.

It is hard nowadays to recapture how disquieting this analysis seemed to many Anglophone philosophers writing in the aftermath of the First World War. L.T. Hobhouse, for example, whose critique of Hegel, Green and Bosanquet appeared in 1918, went so far as to assert that in the bombing of London he had witnessed ‘the visible and tangible outcome’ of this ‘false and wicked doctrine’. To anyone of Berlin’s generation, however, these anxieties about Hegelian philosophy remained remarkably acute, and these are the feelings that Berlin is registering, I think, in his account of positive liberty and the dangers to which it gives rise.

I do not wish, however, to press the historical point. My reason for quoting Green and Bosanquet is to lend further support to what seems to me Berlin’s most important contention, and my reason for wishing to add this support is that Berlin seems to me to miss the force of his own argument. This becomes clear when he responds to MacCallum’s insistence that all claims about freedom conform to the same triadic structure, since they are all claims about the need to be free from constraint to do or become something. Berlin merely returns the suggestion – which I have already shown to be confused – that some pleas for liberty reflect a simpler dyadic structure, since they are nothing more than pleas to be liberated. What Berlin should have retorted, it seems to me, is that the positive conception of liberty he rightly isolates cannot be made to conform to the triadic structure on which MacCallum and his followers insist. The crux of Green’s and Bosanquet’s argument is that the freedom of human agents consists in their having succeeded in realising an ideal of themselves. But this is not to speak of a condition in which someone is free to do or become something, as required by MacCallum’s analysis. It is to speak of a condition in which someone has succeeded in becoming something. Freedom is not being viewed as absence of constraint on action; it is being viewed as a pattern of action of a certain kind.

Berlin’s argument can be carried yet another step if we recognise that what underlies these theories of positive liberty is the belief that human nature has an essence, and that we are free if and only if we succeed in realising that essence in our lives. This enables us to see that there will be as many different interpretations of positive liberty as there are different views about the moral character of humankind. Suppose you accept the Christian view that the essence of our nature is religious, and thus that we attain our highest ends if and only if we consecrate our lives to God. Then you will believe that, in the words of Thomas Cranmer, the service of God ‘is perfect freedom’. Or suppose you accept the Aristotelian argument that man is a political animal, the argument restated as a theory of freedom by Hannah Arendt in Between Past and Future (1961). Then you will believe that, as Arendt maintains, ‘freedom . . . and politics coincide’ and that ‘this freedom is primarily experienced in action.’

Faced with these equations between freedom and certain forms of life, how can MacCallum and his followers hope to rescue their contention that all intelligible claims about liberty must be claims about absence of constraint? As far as I can see, their only recourse will be to suggest that the arguments I have cited from Green and Bosanquet are not intelligible as claims about liberty; that they must either be muddled or be talking about something else. As Berlin excellently points out, however, there is no difficulty in seeing how the neo-Hegelians took their thesis, without any incoherence, to be one about human freedom. The claim they are advancing is that if, and only if, we actually follow the most fulfilling way of life shall we overcome the constraints and obstacles to our realisation of our full potential, and thereby realise our ideal of ourselves. The living of such a life alone frees us from such constraints and, by making us fully ourselves, makes us fully free.

As the title of his essay indicates, Berlin’s main concern is to contrast this positive ideal of freedom with what he describes, following much precedent, as negative liberty. As we have seen, by negative liberty Berlin means absence of constraint, and the specific interpretation he believes must be given to the concept of constraint is that it must consist in some act of interference, by some external agency, with the capacity of another agent to pursue ‘possible choices and activities’. These obstacles or hindrances need not be intentional, for Berlin allows that they may be the result of – as opposed to being deliberately caused by – the actions of others. But his fundamental contention is that the absence marking the presence of liberty must always be absence of interference.

While this is a familiar vision of human liberty, it is a matter of no small difficulty to state it with precision, and it is worth noting that Berlin’s statement embodies a valuable qualification often omitted in more recent accounts. Berlin adds that I am unfree ‘if I am prevented by others from doing what I could otherwise do’. I may be physically obstructed in such a way that an action within my powers becomes impossible to perform. Or I may be subjected to such a degree of coercion that the action is rendered, in Jeremy Bentham’s phrase, ineligible. But in either case my loss of freedom stems from ‘the deliberate interference of other human beings within the area in which I could otherwise act’.

Berlin’s way of articulating this distinction is strongly reminiscent of Thomas Hobbes’s analysis of free action in his Leviathan of 1651. Hobbes compares the predicament of two men who are unable to leave a room. One possesses the power to leave, but has been ‘restrained with walls, or chains’ and thereby disempowered; the other straightforwardly lacks the ability, because he is ‘fastened to his bed by sickness’. According to Hobbes’s analysis, the first man is unfree to leave, but the second is neither free nor unfree; he is simply unable. The reason, Hobbes explains, is that the idea of free action presupposes the idea of deliberating between alternatives. But it makes no sense to deliberate as to whether to perform an action we already know to be beyond our powers.

Contrast this understanding with the view, currently prevalent, that we need to distinguish between the formal and the effective possession of negative liberty. One of the examples Berlin gives in distinguishing lack of freedom from inability is the case of a man who cannot read because he is blind. If we apply the distinction between formal and effective freedom, we arrive at the view that the blind man is formally free to read, because no one is interfering with him in this pursuit. But he is not effectively free, since he is not in a position to make use of his formal liberty.

Berlin’s Hobbesian approach enables us to see that this kind of analysis results at best in confusion and at worst in a kind of mockery of freedom. There are two contrasting points to be brought out here. One is that, on Berlin’s account, the blind man is neither formally nor effectively free to read. As Berlin insists, I am free only if I am capable of exercising an ability, should I choose, without interference. But the predicament of the blind man is that he is incapable of exercising the ability to read under any circumstances. The contrasting point is that, on Berlin’s account, the blind man is neither formally nor effectively unfree to read. To be unfree is to have been rendered incapable of exercising an ability I possess. But the blind man has not in this way been disempowered; he is simply not in possession of the relevant ability.

Although Berlin’s analysis of negative liberty is exceptionally acute and valuable, it nevertheless seems to me to suffer from a serious limitation of coverage. This weakness, moreover, is one that it shares with almost every other recent statement of the theory of negative liberty I have come across. This being so, the nature of the weakness seems to me well worth trying to identify and remedy, and this is the task to which I shall devote the rest of these remarks.

When Berlin first introduces his view of negative liberty, he rightly observes that ‘this is what the classical English political philosophers meant’ by freedom, and he specifically refers us to Hobbes’s definition in Leviathan. What Berlin misses, however, is the fiercely polemical character of Hobbes’s analysis. When Hobbes announces, in words that Berlin echoes closely, that our liberty consists of nothing more than absence of external impediments, he is attempting at the same time to discredit and supersede a rival and strongly contrasting understanding of negative liberty. This rival theory had risen to prominence in English public debate in the early decades of the 17th century, and it appeared to Hobbes to be extremely dangerous as well as hopelessly confused.

I can best bring out the significance of Hobbes’s critique if I try to answer a question raised by Berlin in the introduction to his revised edition of ‘Two Concepts of Liberty’. He asks when the idea of freedom as nothing other than non-interference was first explicitly formulated, and what prompted its rise to its present hegemonal prominence. I would answer that it is very hard to find an explicit statement of such a theory any earlier than Hobbes’s in Leviathan, and that what prompted him to articulate it was his sense of the need to respond to the ‘democratical gentlemen’, as he called them, who had deployed their very different theory to promote the cause of Parliament against the Crown and to legitimise the execution of King Charles I in 1649.

Hobbes’s counter-revolutionary challenge eventually won the day. To cite Berlin’s own litany, we find his basic line of argument taken up by David Hume, Jeremy Bentham, to some degree by John Stuart Mill and even more closely (Berlin might have added) by Henry Sidgwick. This great tradition of classical utilitarianism proved impressively successful at occupying the entire conceptual space, thereby managing to dismiss any rival interpretations as either pernicious or confused. As a result of this profound and enduring ideological success, the alternative vision of negative liberty that Hobbes originally set out to discredit has virtually sunk from sight. What I now want to do is to try to lift it back to the surface.

As I have indicated, the theory in question rose to prominence in Anglophone political theory in the course of the disputes between Crown and Parliament in 17th-century Britain. Critics of the royal prerogative began to argue that, to the extent that they were obliged to live in dependence on the power of the king, and obliged in consequence to rely on his goodwill for the continuation of their rights and liberties, they were living in a state of servitude. They insisted, in other words, that freedom is restricted not only by actual interference or the threat of it, but also by the mere knowledge that we are living in dependence on the goodwill of others. These writers are not making the obvious point that the possibility of arbitrary interference renders our liberty less robust or secure. They are arguing that our mere awareness of living under an arbitrary power – a power capable of interfering with our activities without having to consider our interests – serves in itself to limit our liberty. Knowing that we are free to do or forbear only because someone else has chosen not to stop us is what reduces us to servitude.

The immediate inspiration for this way of thinking appears to have stemmed from a number of medieval common-law texts, above all those of Bracton and Littleton. But the most striking feature of these discussions (although later common lawyers did their best to ignore the fact) is that they in turn owe their phraseology entirely to the analysis of freedom and slavery to be found in the Digest of Roman law. There we are first informed that ‘the fundamental division within the law of persons is that all men and women are either free or are slaves’. Next we are given a formal definition of slavery. ‘Slavery is an institution of the law of nations by which someone is, contrary to nature, subjected to the dominion of someone else.’ This in turn is held to yield a definition of individual liberty. If everyone in a civil association is either free or a slave, then a free citizen must be someone who is not under the dominion of anyone else, but is capable of acting in their own right.

By the time these distinctions were definitively summarised in Justinian’s Codex, they had been the common coin of Roman political theory for generations. They had been popularised above all by the sequence of great historians – Sallust, Livy, Tacitus – who had traced the subversion of the republican constitution and its collapse into the servitude of the principate. If you turn to any of these authorities, you will find it argued once again that what it means to possess your liberty is, as Livy puts it, ‘to be in your own power’, not obliged to live at the mercy of anyone else.

It was this understanding of political liberty that a number of spokesmen in the English Parliament began to deploy in criticism of the Crown in the early decades of the 17th century. They were partly protesting against what they took to be straightforward violations of their fundamental rights. But they were also objecting to what they saw as a deeper affront to liberty. They were fearful of the underlying principle that, in times of necessity, the Crown possesses the discretionary power to override civil rights. The objection they developed was that, if the Crown is the bearer of any such prerogatives, this is as much as to say that our property and personal liberties are held not ‘of right’ but merely ‘of grace’, since the Crown can take them away without injustice at any time.

What troubled these critics was the view of rights implied by this understanding of the prerogative. To claim that our basic liberties are subject to being taken away with impunity is to declare that they do not have the status of rights; it is to say that they are mere licenses or privileges. This was the insight that prompted these critics to reach for their Bracton – and indeed their Livy and Tacitus. To accept that we hold our liberties at discretion, they retort, is to accept that we are living in dependence on the will of the king. But to admit that we are living in such a state of dependence is to admit that we are living not as free citizens but as slaves. The mere knowledge that the Crown possesses such prerogatives is what reduces us to servitude.

The moment at which these arguments provoked a fatal crisis came in 1642. When the House of Commons brought forward a proposal early in February to take control of the militia, Charles I made it clear that he would veto any such legislation by exercising his so-called prerogative of the Negative Voice. Parliament then took the revolutionary step of claiming that, at least in times of emergency, it must possess the right to legislate even in the absence of the royal assent. The reason why this must be so, a number of spokesmen now proclaimed, is that the alternative is national servitude. The most influential statement of the Parliamentary case was furnished by Henry Parker in his Observations of July 1642. If the Crown can block any legislation with the Negative Voice, this will reduce the Parliament to a state of dependence on the will of the king. But if we permit the king ‘to be the sole, supreme competent Judge in this case, we resign all into his hands, we give lives, liberties, Laws, Parliaments, all to be held at mere discretion’ and thereby consign ourselves to slavery.

Parker was not the first to put forward this argument, but his Observations offered the most confident statement of the case, and did much to make it central to the rhetoric of the ensuing Civil War. We encounter the same argument in Parliament’s call to arms of August 1642, and we encounter it yet again after the Parliamentary victory, when it was used to justify not merely the regicide but the abolition of the monarchy. The charge against Charles I at his trial was that he had ruled by his arbitrary will, and hence tyrannically. The Act of March 1649 abolishing the office of king confirmed that monarchy is ‘dangerous to the liberty, safety, and public interest of the people’, and added that in England the effect of the prerogative has been ‘to oppress and impoverish and enslave the subject’.

This neo-Roman analysis of what it means to possess our freedom carried with it a distinctive view of the relations between the liberty of citizens and the constitution of the state. The essence of the argument is that freedom is restricted by dependence. To be free as a citizen, therefore, requires that the actions of the state should reflect the will of all its citizens, for otherwise the excluded will remain dependent on those whose wills move the state to act. The outcome is the belief – crucial alike to the English Revolution of the 17th century and to the American and French Revolutions of a century later – that it is possible to enjoy our individual liberty if and only if we live as citizens of self-governing republics. To live as subjects of a monarch is to live as slaves.

It would be wrong to imply that Isaiah Berlin failed to recognise the existence of this tradition of thought. It is true that he never discusses it with the same historical specificity as he brings to bear on the other two traditions he examines, and that he never singles out any particular theorist or movement capable of being associated with this alternative standpoint. Given, however, that he was writing at the height of the debate about decolonialisation, he could scarcely have been unaware that nations as well as individuals often claim to be unfree when they are condemned to social or political dependence. He devotes considerable attention at the end of his essay to what he describes as the resulting ‘search for status’, and he explicitly asks himself whether it might not ‘be natural or desirable to call the demand for recognition and status a demand for liberty in some third sense’.

Having raised the question, however, Berlin confidently answers that no such third concept of liberty can be coherently entertained. To speak of dependence as lack of liberty, he writes, would be to confound freedom with other concepts in a manner at once misleading and confused. Stating his grounds for this conclusion, Berlin goes on to enunciate his most general claim about the concept of liberty. He insists that it is true not merely of any coherent account of negative freedom, but of any concept of freedom whatever, that it must embody, at least as a minimum, the idea of absence of interference. If we are to speak of constraints on our liberty, we must be able to point to some visible act of hindrance, the aim or consequence of which was to impede us in the exercise of our powers.

It is precisely this assumption, however, that the writers I have been considering reject. The distinctive claim they defend is that a mere awareness of living in dependence on the goodwill of others serves in itself to restrict our options and thereby limits our liberty. The effect is to dispose us to make and avoid certain choices, and is thus to place clear constraints on our freedom of action, even though our rulers may never interfere with our activities or even show the least sign of threatening to interfere with them.

The exploration of this argument had been a leading preoccupation of the classical historians I have singled out. Tacitus in particular speaks with an unforgettable combination of agony and contempt about the psychological impact of living under tyranny. If you are subject to unaccountable power, you will find in the first place that there are many things you are not free to say or do. Above all, you will need to ensure that you avoid saying or doing anything that might be construed by your rulers as an act of challenge, emulation or reproach. You will likewise find that you lack the freedom to abstain from saying and doing certain things. When required to advise your rulers or to comment on their behaviour, you will find yourself constrained to endorse whatever policies they already wish to pursue. Yet more serious is the long-term psychological damage inflicted by such forms of self-censorship. As Tacitus bitterly emphasises, servitude inevitably breeds servility. When a whole nation is inhibited from exercising its highest talents and virtues, these qualities will begin to atrophy and the people will gradually sink into an abject condition of torpor and sluggishness.

It was this analysis that exercised perhaps the most formative influence on the democratical gentlemen who challenged the government of Charles I and instituted the first and only British republic. So far. They, too, were deeply preoccupied by the dangerous implications of the fact that unaccountable rulers are inevitably surrounded by servile flatterers, and have little hope of hearing frank advice. As in the case of the classical historians, however, their principal anxiety was that, under such rulers, no one will perform any deeds requiring public spirit or courageous and great-hearted qualities. At first they lacked the vocabulary in which to express this Tacitean insight, but they gradually popularised a series of neologisms that enabled them to refer directly to the loss of spirit, courage and great-heartedness that tyranny brings in its train. The final effect, as they put it, is that everyone becomes dispirited, discouraged, disheartened.

For all the power of this analysis, contemporary political theory has largely neglected it. Berlin’s view that negative liberty must be construed as absence of interference remains the orthodoxy, and nowhere more so than in Great Britain and the United States. But this is deeply ironic, especially in the American case, for the United States was born out of the rival theory that negative liberty consists of absence of dependence. When Congress adopted Thomas Jefferson’s Declaration in July 1776, what they decided to call it, no one needs reminding, was a Declaration of Independence. But do we ever pause long enough over that word? Independence from what? From living in dependence on the arbitrary power of the British Crown. And what made Congress believe that this justified revolution? Their acceptance of the classical contention that, if you depend on the goodwill of anyone else for the upholding of your rights, it follows that – even if your rights are in fact upheld – you will be living in servitude.

Given our current predicament, it is unfortunate that this way of thinking about freedom has become so widely discredited. We are again being urged to recognise that, in times of emergency, civil liberties must bow to national security. We are being urged, that is, to acknowledge that our liberties are held not as rights but by grace of our rulers, and that it is for them to tell us what counts as an emergency. These arguments are of course being put to us in the name of freedom and democracy. But it is worth recalling that, according to the American Founding Fathers, and to the democratical gentlemen by whom they were so greatly influenced, this is to speak the language of tyranny.