Since the 1960s, questions about the definition and evaluation of freedom have been hotly debated, and in much more analytical detail than ever before. The 1970s and 1980s saw the publication of a large number of articles dealing exclusively with the definition of freedom, and in the late 80s and the 90s a number of important books have been published. This increased interest in the concept of freedom is often put down to the influence of Isaiah Berlin's seminal article "Two Concepts of Liberty", first published in 1958 (Oxford: Blackwell), and in a revised form in 1969, in his *Four Essays on Liberty* (Oxford: Oxford University Press). The article is reprinted in David Miller (ed.), *Liberty* (Oxford: Oxford University Press, 1991), which is a useful collection that contains a number of the pieces referred to below. It is true that Berlin's essay is still paradigmatic for current debates on freedom. But of course, the causes also go deeper. Perhaps the most important of these was the increased awareness in philosophical circles of the unnecessarily restrictive nature of the attitude of the logical positivists, who regarded the analysis of normative concepts as being of little worth. The work of Berlin and others in the 1960s gave political and moral philosophy a new kind of dignity, showing that it was indeed fruitful to try to analyze normative concepts and their relation to political ideologies. Thus, the increased interest in freedom is just one example - though an important one - of the rejuvenation of normative political theory over the last few decades.
To say, as many do, that political theory has been "rejuvenated" is of course to suppose that it had also been alive and well at some previous time. And indeed, there are some important historical landmarks for this subject that we need to bear in mind, rather than embarking on our analysis of the concept of freedom *ex novo*. Of particular interest to classical liberals is the history of the individualistic, "negative" view of freedom - of freedom as the absence of constraints on individual actions - which can be traced back at least to the levellers (see W. Haller ed., *Tracts on Liberty in the Puritan Revolution* [New York: Octagon books, 1965]), and to the work of John Locke (*Two Treatises of Government* [1690] [London: Everyman, 1924 / Cambridge: Cambridge University Press, 1970]) and Algernon Sidney (*Discourses Concerning Government* [1698] [Indianapolis: Liberty Classics, 1990]), all of whom connected freedom with the limitation of arbitrary power on the part of rulers, and with the idea of government based on consent. Benjamin Constant, in "The Liberty of the Ancients Compared With That of the Moderns" (*Political Writings* [1819] [Cambridge: Cambridge University Press, 1988]) contrasted this new, modern idea of freedom as the absence of constraints and interference in individuals' lives, with the ancient idea of freedom as consisting in one's status as a citizen, and as being fully exercised through political participation (see the reference to "republicanism" in part II of this essay). The individualist foundations of this modern concept of freedom were brought out especially clearly by Wilhelm von Humboldt who, in *The Limits of State Action* (1854) (Indianapolis: Liberty Classics, 1993), argued that it was only by keeping government intervention to a minimum that one would allow the space for the many-sided development of people's individual potentials. Humboldt's book also contains an important argument about the damaging effects on freedom of excessive state bureaucracy - advice that has been tragically ignored in the twentieth century. Humboldt's romantic individualism was an important influence on J. S. Mill's essay *On Liberty* (1859) (London: Everyman,1910 / New York: Prometheus,1986), which again argued for the minimization of coercion, and for an individual's right to do anything she pleased, provided she not do harm to others - what some contemporary libertarians have called "the right to do wrong". Another giant of the nineteenth century, sadly neglected today, is Herbert Spencer, who in *The Man Versus the State* (1884) (Indianapolis: Liberty Classics, 1982) presented some prophetic arguments about the dangers for individual freedom of socialist planning. In such writings we can see important influences on contemporary liberal theorists of freedom: we can see, for example, the influence of Locke on Nozick, and the influence of Mill and Spencer on Hayek. These and other contemporary liberals, along with their critics, will come up later, in part II of this essay.
The close study of historical texts such as these pays great
dividends. But despite the importance of these works, above all in
increasing our awareness of the *importance* of individual freedom,
there has never been so much precise, rigorous, analytical research
on the actual *meaning* of freedom as in the last three decades, and
it is therefore to this work that we should now turn for further
enlightenment. But first, let us take a brief look at the
methodological assumptions behind this recent work.
One way of opposing the essential contestability view is by saying that it is at least possible to make progress towards agreement over the definition of concepts like freedom by appealing to our *linguistic intuitions*. A much discussed example of Berlin's is that of the so-called paradox of the *contented slave*. We start by noting that on the face of it a reasonable way of defining freedom is by saying that it means "being allowed to do what you want". We then notice, however, that this would imply that a slave who wants to be nothing other than a slave is totally free. Yet we have a strong linguistic intuition that slaves are very unfree. So the above definition must be wrong. Here we have a very simple illustration of the method used by most of the writers on freedom who disagree with the essential contestability thesis. Their idea is that by appeal to (a) the rules of logic and (b) our linguistic intuitions, progress can be made in arguments over the definition of freedom.
Methodological arguments of this sort against essential contestability are to be found in Felix Oppenheim's *Political Concepts: a reconstruction* (Oxford: Blackwell, 1981) and J.P. Day's *Liberty and Justice* (London: Croom Helm, 1987). More recently, see Oppenheim's "Social Freedom and its Parameters", *Journal of Theoretical Politics*, Vol. 7 (1995), which argues that agreement can be reached about an *empirical* definition of freedom without any implications for political ideologies. A detailed argument against different forms of the essential contestability thesis can be found in Christine Swanton's "On the 'Essential Contestedness' of Political Concepts", *Ethics*, Vol. 95 (1984-5), and in her *Freedom: a Coherence Theory* (Indianapolis: Hackett, 1992). In the latter, as an alternative to the essential contestability thesis, Swanton argues in favor of an application of John Rawls's method of "wide reflective equilibrium" to the task of defining freedom. Some interesting methodological reflections are also contained in Peter Morriss's book, *Power: A Philosophical Analysis* (Manchester: Manchester University Press, 1987), ch. 25, which argues against the essential contestability thesis, but also against Oppenheim's purely empirical approach. Morriss emphasizes the distinction between a "word" and a "concept", where a single word can be used to refer to a plurality of different concepts, depending on the purpose one has in mind in using the word. Further observations on method, including a discussion of some of the above works, can be found in Kristjan Kristjansson's recent book, *Social Freedom: the responsibility view* (Cambridge: Cambridge University Press, 1996), ch. 7.
As we shall see, both below and in part II of this essay, most
classical liberals disagree with the essential contestability thesis,
because they believe that engaging in arguments about the definition
of freedom will help to promote liberalism itself. They believe, in
other words, that the apparent espousal of freedom on the part of
non-liberals can, through conceptual argument, be shown to be no more
than linguistic sleight-of-hand.
Much subsequent work on positive freedom has concentrated on attempting to avoid this counterintuitive implication. Some have argued for weaker versions of positive freedom, identifying it with acting on one's desires, but not necessarily with acting on universally valid principles. Here, though, one still has to face up to the problem of the "contented slave": because an individual's positive freedom depends on her desires, there is the danger, as we saw, of concluding that people who learn to desire fewer things thereby make themselves freer. Richard Arneson, in "Freedom and Desire", *Canadian Journal of Philosophy*, vol. 3 (1985), and John Christman, in "Liberalism and Individual Positive Freedom", *Ethics*, vol. 101 (1991), try to avoid this problem by emphasizing that the desires one should be free to act on should be "home grown", which is to say, autonomously developed by the agent herself. If one accepts this and similar conditions, the actual remaining cases of free (because contented) slaves will not be so counterintuitive. John Christman's edited collection, *The Inner Citadel: Essays on Individual Autonomy* (Oxford: Oxford University Press, 1991), also contains a number of contributions which defend the connection between freedom and desire. In the negative camp, which is opposed to linking freedom with desire, the most important work is probably still the rigorous and much cited article by J. P. Day, "On Liberty and the Real Will", *Philosophy*, Vol. 45 (1987), reprinted in his *Liberty and Justice*, mentioned above. Other useful ideas on the link between freedom and desire can be found in work dealing with the concept of *autonomy*. For a general introduction and guide to the literature, see R. Lindley, *Autonomy* (London: Macmillan, 1986).
Other defenders of positive freedom, such as Amartya Sen, have tended to identify an agent's freedom with her effective *ability*. See Sen's many writings on freedom, including "Well-being, Agency and Freedom", *The Journal of Philosophy*, vol. 82 (1985), "Justice: Means versus Freedom", *Philosophy and Public Affairs*, vol. 19 (1990), and *Inequality Reexamined* (Oxford: Oxford University Press, 1992), ch. 3. Defences of a similar conception of freedom-as-ability are to be found in Lawrence Crocker's *Positive Liberty* (London: Nijhoff, 1980), Horatio Spector's *Autonomy and Rights* (Oxford: Clarendon Press, 1992), and Philippe Van Parijs's *Real Freedom for All: What (if Anything) Can Justify Capitalism?* (Oxford: Oxford University Press, 1995). For Van Parijs, "real" freedom contrasts with the "merely formal" freedom generally defended by classical liberals. Van Parijs's "real freedom" is again a kind of *positive* freedom: according to authors like Sen and Van Parijs, a tramp is not "really" free to dine at the Ritz, because he lacks the necessary means to do so. This, despite the fact that no one is actively preventing him from doing so. It is a conception of freedom that appears to date back to the "New Liberalism" of the turn of the century, as exemplified in Hobhouse's *Liberalism* (London: Greenwood, 1911). Those who favor the negative conception of freedom point out that freedom and ability are different things. Freedom is a *social* concept - a relation between persons - and should be something which *political* action (rather than the action of medics or engineers) can do something to augment. The tramp is thus *free but unable* to dine at the Ritz, as long as no one stops him from doing so. This position is argued for by Berlin, Oppenheim and Day (in the works already cited), and by W.A. Parent in "Some Recent Work on the Concept of Liberty", *American Philosophical Quarterly*, vol. 11 (1974), which, incidentally, provides an excellent overview of the work prior to that date.
One author who continues to identify freedom with rational action is Charles Taylor (see his "What's Wrong with Negative Liberty", in A. Ryan (ed.), *The Idea of Freedom*, London: Oxford University Press, 1979, and reprinted in the Miller collection), who claims that once we abandon the "crude" Hobbesian psychology on which the negative view of freedom rests, we have no other alternative than to progress a fair way down Berlin's slippery slope - from an "opportunity concept" of freedom (negative - where freedom is about how many doors are open) to an "exercise concept" (positive - where freedom is about how one actually acts). Taylor's theory has been criticized by Hillel Steiner, who has argued with the aid of a simple mathematical formula that it leads straight to Berlin's paradox whereby people can be "forced to be free" (see his "How Free: Computing Personal Liberty", in A. Phillips Griffiths (ed.), *Of Liberty*, London: Cambridge University Press). A more detailed discussion of Taylor's theory, which links Taylor's work on freedom to his work on Hegel, and which discusses the notion of "internal" constraints, is to be found in Richard Flathman's *The Philosophy and Politics of Freedom* (Chicago: Chicago University Press, 1987), pp. 15-107. Flathman usefully sets out a negative-positive spectrum of conceptions of freedom, with Steiner's "pure negative" conception at one end, and Taylor's "fully virtuous" (and in that sense, fully positive) conception at the other. Taylor's notion of internal constraints is further discussed, and ultimately rejected, in Kristjansson's *Social Freedom* (see above), ch. 5. Another defense of a strongly positive conception of freedom is Benjamin Gibbs's *Freedom and Liberation* (London: Chatto and Windus, 1976), which is wittily attacked by the classical liberal Antony Flew, in "'Freedom is Slavery': a Slogan for Our New Philosopher Kings", in A. Phillips Griffiths, *Of Liberty*, cited above. See also Gibbs's reply in the same volume.
The distinction between negative and positive freedom was challenged
some time ago in Gerald MacCallum's seminal article, "Negative and
Positive Freedom", *Philosophical Review*, Vol. 76 (1967), reprinted
in Miller's *Liberty* (see above). According to MacCallum, it is
wrong to see the alternative definitions of freedom as divided into
two categories indicating the *absence* and *presence* of something.
Instead, we should say that *any* conception of freedom must define
three things: an agent, X, who is free from constraints, Y, to do an
action (or to be something), Z. All disagreements about the
definition of freedom therefore boil down to questions about the
extension of factors X, Y and Z. While MacCallum did not do away with
the positive-negative distinction, his formula has provided theorists
with a clear and useful framework. It is useful, for example, to see
debates such as those mentioned above as debates over whether the
extension of the Z factor should be "any actions whatsoever" or only
"actions one desires to perform", over whether the Y factor should
indicate merely "external" or also "internal" constraints, and over
whether Y includes only constraints imposed by other agents or also
those imposed by nature. In each of these debates, the supporters of
the negative conception of freedom tend to favor the first
alternative over the second.