Social Power and Negative Freedom

Economists have recently begun to show interest in the concept of freedom—especially in the area of social choice theory—but have yet to turn their attention systematically to its relation to power. An exception is Braham (2006), but this is not concerned with the different forms of power, in the sense of ‘form’ I shall assume in this article.

For the sake of simplicity, I here assume that freedom is the absence of unfreedom, so that ‘not unfree’ entails ‘free’ (and ‘not free’ entails ‘unfree’). This bivalence assumption is not unproblematic, but I shall not discuss the issue here. For a critique, see Kramer 2003, pp. 41–60.

In Carter 2008, I apply this observation to an analysis of Philip Pettit’s notion of freedom as “discursive control” (Pettit 2001), arguing that freedom as discursive control is limited by offers, no less than by threats.

It should be added, however, that one may also go on to predicate freedom of agents (in the exercise sense) on the basis of the fact that they perform their actions freely.

The English terms ‘will’ and ‘voluntariness’ have different etymological roots. The connection between them is much clearer in Latin languages (their respective equivalents in Stoppino’s native tongue are volontà and volontarietà).

What is the exact meaning of ‘theoretical compossibility’ in this context? This issue is problematic and has given rise to some debate in the literature. For Steiner (1994, Chap. 2), it means ‘logically compossible’. In A Measure of Freedom I tentatively suggest that it might mean either ‘logically compossible’ or ‘technologically compossible’ or ‘possible according to laws of nature’ (Carter 1999, p. 173). For discussion, see van Hees (2000, pp. 131–133). Kramer (2003, Chap. 2) defines theoretical possibility, in this context, in terms of the agent’s abilities, identifying freedom with ability and unfreedom with the prevention of that which the agent would otherwise be able to do. On this view, those actions the agent would be unable to perform even in the absence of prevention on the part of others, are classified as actions the agent is neither free nor unfree to perform: if I am unprevented from doing x but am nevertheless unable to do x, then I am neither free nor unfree to do x.

If it were not a fairly reliable indicator, then threats would fail as instruments of generalized and stabilized power. I return to this point at the end of the present subsection.

For a more direct attempt to rebut this second counterexample, by showing that it fails to identify a threat that has no effect on B’s set of sets of available options, see Kramer (2003, pp. 195–204).

A useful account of this test is given in Braham and Holler (2009).

I present a critique of Skinner and Pettit along these lines, in part applying the analysis of the freedom-power relation contained in the present article, in Carter 2008. An earlier version of this critique can be found in Chap. 8 of Carter 1999. See also the writings of Matthew Kramer on the concept of freedom, in particular Chap. 1 of Kramer 2003, 2008.