![]()
1
Conceptualizing Freedom
We saw in the Introduction that the language of freedom applies on the side of agency – the principal focus in this book – and on the side of the environment where agency is exercised. And we distinguished three domains in which agency freedom is at issue: those of the action, the self and the person. The approach taken presupposes that there is a single, unequivocal concept of freedom at issue when we speak of a free action, a free self and a free person, and the question to which we now turn bears on what this is. What is the concept, assuming there is just one, that unifies talk of freedom in these three domains?
The connotations of freedom
To predicate freedom of an agent, in particular of something an agent did, is to suggest that at least three different sorts of thing hold (O’Leary-Hawthorne and Pettit 1996). The first is that the agent can be rightly held responsible for what he or she did; if the action was free then there can be nothing against thinking that the agent should have to answer for it. The second is that the action freely chosen is one that the agent can own, thinking: this bears my signature, this is me. And the third is that the agent’s choice was not fully determined by at least certain sorts of antecedents; it was not fully determined, for example, by a hypnotic suggestion or an unconscious complex or childhood conditioning.
Any account of the concept of freedom must look to these connotations of responsibility, ownership and underdetermination and ask after which of them, if any, is the most basic determinant of the way we apply the concept. The account offered must explain why each of the connotations holds, or appears to hold, in our ordinary usage of the term. It must explain how the concept of freedom in an agent can apply at the level of action, self and person. It must support the intuitions that we find broadly compelling in our view of human agency. And, ideally, it must fit with a plausible theory of what actually constitutes freedom in an agent. As I suggested in the Introduction, the account must look for a reflective equilibrium (Rawls 1971) – an equilibrium that may require reflectively developed adjustment on one or another side – between those different elements.
My plan in this chapter is to develop an approach that prioritizes the responsibility connotation. The approach goes back in recent thought to Peter Strawson’s essay, first published in the 1960s, on ‘Freedom and Resentment’ and it has been taken up in different ways by a variety of authors since then (Strawson 1982; Watson 1982; Wolf 1990; Fischer 1994; Pettit and Smith 1996; Wallace 1996; Scanlon 1998). There are a number of considerations that favour the approach but the best argument in its support is that it leads us towards a satisfying view of the area overall. Since that argument is available only in light of what comes later, the commitment to analysing freedom in terms of responsibility should be seen as an assumption that will be tested in the long haul of the book, not just the short haul of this chapter. The theme has already been sounded in the Introduction.
In this chapter we will be looking further at what is involved in thinking of freedom from the perspective of responsibility and we will be identifying some considerations that support the approach. One initial reason for favouring it is that it promises to do better than alternatives in saving all three connotations. There is little or no reason to think that the underdetermined action, or the action that is wholly owned by the agent, has to be a matter for which the agent can be held responsible. But there is reason to think that any action for which the agent can be held responsible is going to be under-determined in a significant way, and is going to be something that the agent can and must own.
Underdetermination will not connote responsibility, because a wholly chance event that occurs within me or by my hands will not necessarily be something for which I can be held responsible. Ownership will not connote responsibility, because I may identify as most distinctively and intimately mine a response or a state, a habit or a skill, that I cannot be held responsible for; it may be something that comes to me – happily, I think – by grace of genes or background. But the responsibility perspective saves those other connotations, because there is a certain underdetermination and a certain ownership implied in the very idea of being fit to be held responsible for something.
The connection with ownership appears in the fact that if I was fit to have been held responsible for doing A rather than B, and I did A, then that choice can be laid at my door – represented as mine – in a manner that connotes a degree of ownership. The self that is fit to be held responsible must be a self that is not alienated from the action of A-ing. The agent must be able to think of the thoughts that led to action, and of the action itself, in the first person: this is what I had in mind, and this is what I did.
The connection with underdetermination is a little less direct but still fairly compelling. If I was fit to have been held responsible for A-ing or not A-ing, then the awareness that I ought to A must have been capable of getting me to A, the awareness that I ought not to A must have been capable of getting me not to A. And that is to say that causal factors other than my awareness of what I ought to do cannot have fully determined what I did; they cannot have pre-empted my response. Those antecedents, however they are to be characterized, must have underdetermined my choice.
As the account given here of how we conceive of freedom starts from the responsibility connotation, so other accounts of freedom will start from the other connotations. We shall not be explicitly examining these rival approaches in this book, nor even looking at the objections that such approaches will sponsor to the conception of freedom as fitness to be held responsible. I put observations in place that pre-empt many obvious objections but I do not give much time to objections as such. The discussion of a number of themes will connect us with the literature developed in pursuit of those rival conceptions of freedom, however. In the discussion in chapter 3, for example, we shall be drawing extensively on Frankfurt’s (1988) work on the analysis of freedom as ownership. But for good or ill the book develops the responsibility conception of freedom, and tries to make its appeal and merit clear, without entering into a sustained dialectic with the alternatives.
The conundrums of freedom
Depending on which of the connotations is given priority in the conceptualization of freedom – if indeed any is prioritized – the problematic of free agency will be given a different cast; a different conundrum, as we might put it, will take centre stage. Under the responsibility connotation, freedom is problematic so far as it is recursive in character; under the ownership, it is problematic so far as it has a first-personal aspect; and under the connotation of underdetermined choice it is problematic so far as it involves a certain modal possibility. I will look at these problems in reverse order.
The modal problem that underdetermination raises is to explain how it can be that for anything that is freely done, the agent must have been capable of doing something else instead. Under the most radical interpretation this means that at the moment of choice it must have been possible for the agent, regardless of the causal regime and causal history of the world up to that point, to have done otherwise (Sartre 1958; Van Inwagen 1983). Unless it is to be taken in a downbeat way that will scarcely attract defenders (Lewis 1986, Essay 25), this condition is in tension with a naturalistic picture of the universe: that is, with a picture of the universe under which every aspect of the world, including the freedom of agents, is fixed in place by the way the world is constituted, and the way it is organized by law, in the microphysical realm postulated in physics (Pettit 1993b; cf. Jackson 1998). Let the world be governed by deterministic law, and the condition is straightforwardly ruled out. Let it be governed by indeter-ministic law, and it is still unclear how agents are supposed to fill the gaps that such law leaves open; and if they do fill them, it is unclear why this does not compromise indeterminism.
There have been other interpretations of the modal condition, designed to make the underdetermination involved in freedom seem less problematic. These construe it as underdetermination by a select set of antecedents, not by everything that happens up to the moment of choice. They represent the capacity of agents to have done otherwise as their capacity to have acted, if not counter to all the causal influences on the action, at least counter to those particular antecedents. But these interpretations run into a variety of problems, as is now well known (cf. Berofsky 1987).
One line takes the condition to be that had the agent chosen, then he or she would have done otherwise (Moore 1911, chapter 6; Ayer 1982). But this is no good. Choosing or willing or any such cognate is an action, so that there will always be a question as to whether it itself satisfies that condition; this question will open up an indefinite regress. Another, more popular line takes the condition to be that had the agent not desired to act in that way, then they would not have done so (Davidson 1980, Essay 4). This proposal is not subject to the same difficulty as the last, since desiring is not an action and the question does not arise as to whether it itself is a free action. But it is problematic in another respect. For all the condition now stipulates, it may be that in order for the agent’s desires to have gone the other way, the conditioning and drilling to which the agent was subject as a child would have had to be different from what it actually was (Chisholm 1982). It may be that the possible world where the agent’s desires are different from how they actually are is a very remote world indeed; in order to get to it from the actual world we would first have to make the actual past different from how it is.
So much for the modal conundrum that the conceptualization of freedom in terms of underdetermination puts in centre place. The conundrum that the conceptualization of freedom in terms of ownership puts in that position bears on the first-personal aspect of freedom. This consists in the fact that with anything freely done it must be the case that the agent is able, indeed compelled, to see the action as his or her own. The agent cannot be detached from the action, or from the process leading to the action, in the way they may be detached from a reflex or a pathology or even an obsession or compulsion. The agent must not be a mere bystander or onlooker of what happens; they must identify with what is done by their hands.
This conundrum has been given particular attention since the work of Harry Frankfurt (1988) in conceptualizing freedom – strictly, as he would put it, freedom of the will – in terms of ownership. We naturally associate free agency with the operation of a certain process leading to action, and we naturally think that any objective process can be characterized in a third-person vocabulary. So how then can the agent be present in the first person; how can the agent see the action as distinctively his or hers? The question raises a serious challenge. Any objective process will lend itself to a third-person characterization – say in terms of a cause–effect sequence – that does not include mention of the self, only of events that occur within the self (pace Chisholm 1982). And so there will always be a question as to how the agent is able to see himself or herself in the process associated with free action, owning or claiming the action as something that bears their signature.
The conceptualization of freedom in terms of responsibility puts a different conundrum in centre position. This is associated with what I describe as the recursive character of responsibility. Suppose that I am responsible for an action. Presumably that will be so because that action is under the control of some other factor in me: say, my particular beliefs and desires. The recursive character of responsibility appears in the fact that, by ordinary intuitions, this means that I must be responsible in turn for those beliefs and desires. If I were not responsible for them, then presumably that would let me off the hook of responsibility with the action too: I could dismiss it as the product of something not within my sphere of responsibility. But if I am responsible for the beliefs and desires that are in control of the action, then presumably I am responsible for them in virtue of their being under the control of some further factor still in my make-up: say, my habits of forming and revising beliefs and desires. The recursive character of responsibility means that I must equally be responsible for those habits, for otherwise, once again, I would be off the hook. And so on, it appears, indefinitely. Let my responsibility be mediated by a certain sequence of controlling factors and it seems that I must be responsible for every link in that chain. I cannot be responsible for something in virtue of the operation of a controlling factor for which I am not also responsible. Responsibility is inherently recursive in nature.
The recursive nature of responsibility appears to entail an indefinite regress back along the lines of controlling influences in virtue of which an action is put down to an agent in the first place (cf. Klein 1990; Strawson 1994; Kane 1996; Hurley 1999). The problem is of a kind with the difficulties to which the other conceptualizations direct our attention. In each case free agency is associated with a condition that requires more than nature is apparently capable of providing; in each case free agency is depicted in a manner that makes it apparently inconsistent with a naturalistic picture of the universe.
Since we conceptualize freedom in terms of responsibility, we will usually focus in the discussion of free agency – particularly, in the discussion of free action – on the recursive conundrum. But the other conundrums will make an appearance too. The first-personal problem associated with the conceptualization of freedom in terms of ownership will come up here in the discussion of freedom in the self. And the modal difficulty that is associated with the conceptualization of freedom in terms of underdetermination will be explicitly addressed in chapter 4, when we defend the theory of freedom – particularly, the theory of free action – as discursive control. I will try to show at that point that the theory of freedom as discursive control provides resources sufficient to resolve the modal difficulty as well as the recursive problem.
Building on the responsibility connotation
The unity to talk of freedom, under the approach to be taken here, comes of the linkage between being free and being held responsible. ‘Ought’ implies ‘can’, so it has long been said, and whatever qualifications we may wish to impose on that principle, it gives nice testimony to the connection I have in mind. Wherever an ‘ought’ is addressed to a subject, it is assumed that there is a suitable ‘can’ present. Wherever there is an obligation imposed on a subject, it is assumed that the person has the capacity to discharge that obligation. You are a free agent and your action is a free action just to the extent that you are capable of being held responsible in the relevant choice. More specifically, you are free just to the extent that you are capable of being held rightly responsible, by the criteria implicit in the practice. You are free, so I shall say, just so far as you are fit to be held responsible.
The notion of fitness to be held responsible is grounded in the ordinary practice whereby we hold one another responsible for things we do, and proceed to impute blame for those actions we see as bad, praise for those we see as good. To be fully fit to be held responsible for a certain choice is to be such that no matter what you do, you will fully deserve blame should the action be bad and fully deserve praise should the action be good. How much blame or praise you actually get will vary, depending on orthogonal factors such as the expectations held of you and the performance of others; being fully fit to be held responsible means being fully qualified to receive whatever level of praise or blame is on offer. You will be a free agent so far as the way you are – in your self and in your person – allows you to make choices for which you are fully fit to be held responsible in this sense. And your action in a given case will be free so far as it materializes in a way that enables you to count as fully fit to be held responsible.
The practice of imputing praise and blame and holding people responsible is not an intellectual exercise in which we draw up an audit of one another’s behaviour and then ponder on appropriate responses. It is written into some of our most basic reactions to one another, as when we feel resentment for the harm that someone does to us, or gratitude for a good that they provide. It appears also in the indignation that we feel on behalf of others who are maltreated by a third party or the approval that we feel on behalf of others who are treated well. But it is in every instance a matter of sensibility and affection as much as it is a matter of cognition and judgement. The practice is rooted deeply in the architecture of our psychology, engaging with some of our most robust emotions (Strawson 1982; Wallace 1996).
The picture adopted here is that when we think of an agent as free, we do so just to the extent that we think of them as worthy of this sort of reaction, be it a negative or a positive response. Being reaction-worthy in that sense is what it means to be fit to be held responsible. Not every free action will engage our reactions, of course; some may be neutral enough not to elicit resentment or gratitude, indignation or approval. But even such a neutral action, if it is the free initiative of a free agent, will issue from a choice such that had the agent done something good, our reaction would have been positive, and had the agent done something bad, our reaction would have been negative.
We see only human beings as worthy of reactions like resentment and gratitude. To feel such reactions towards natural phenomena like the weather or the business cycle, or even towards non-human animals and their doings is, so most of us think, quite inappropriate. Yet human beings and the choices of human beings are not inevitably reaction-worthy. We spontaneously identify people who are out of their mind or not themselves, in vernacular phrases, as inappropriate objects of resentment or gratitude. And we discriminate without difficulty between cases where ordinary subjects should be held reaction-worthy and responsible – if they do something bad then, as we say, there is little or nothing to excuse them – and cases where they should not.
The general idea behind the responsibility approach, then, is this. We engage with other human beings in a distinctive manner that involves the spontaneous attribution of responsibility, and we conceive of freedom as that property of human beings, and of the actions performed by human beings, that makes such an attribution appropriate under the rules of the practice. Our reactive attitudes are the lens in which the image of the free agent and the free action first takes shape and when we endorse the deliverances of that lens – when we think that the agent and the action are fully reaction-worthy – then we think that it is proper to predicate freedom. Being free is being such that the reaction is appropriate; it is being fit to be held responsible.
When will freedom in this sense fail? Intuitively, you will not be fully free in respect of a choice between A and B, if you are not aware of the availability of those options in your environment of choice, do not have the conceptual resources to evaluate them, or are not functioning in a way that would allow that evaluation to affect what you do. You will not be fully free if, as a self, you are subject to problems that make it impossible for you – or just particularly difficult for you – to claim A or B as something you did. And you will not be fully free if, as a person, you are the victim of an unwelcome form of pressure or duress or coercion that makes it more difficult to do one or other of those things. Such conditions generally serve to exculpate or at least excuse an agent; they remove or reduce the responsibility. And so if freedom is just fitness to be held responsible, they will also count as conditions that destroy or diminish the agent’s freedom.
I speak for convenience, here and throughout the book, as if the practice of holding responsible always involves more than one person. But it should be clear that we also hold ourselves responsible for the things we do; we do not limit the practice to other people (Watson 1996; Oshana 1997). This shows up in the fact that we feel guilty about what we did in the past, and blame ourselves for our actions, as of course we may equally feel innocent and think that there...
