Essays on Freedom of Action (Routledge Revivals)

Ted Honderich

• Publisher: Routledge
• Year: 2015
• ISBN:
  9781317516125
• Edition: 1
• Topic: Filosofia
• Subtopic: Libero arbitrio e determinismo in filosofia

Essays on Freedom of Action, first published in 1973, brings together original papers by contemporary British and American philosophers on questions which have long concerned philosophers and others: the question of whether persons are wholly a part of the natural world and their actions the necessary effects of causal processes, and the question of whether our actions are free, and such that we can be held responsible for them, even if they are the necessary effects of casual processes. This volume will be of interest not only to those who are primarily concerned with philosophy but also to students in those many other disciplines in which freedom and determinism arise as problems.

Towards a reasonable libertarianism

David Wiggins

Towards a reasonable libertarianism¹

One of the many reasons, I believe, why philosophy falls short of a satisfying solution to the problem of freedom is that we still cannot refer to an unflawed statement of libertarianism. Perhaps libertarianism is in the last analysis untenable. But if we are to salvage its insights, we certainly need to know what is the least unreasonable statement the position could be given. Compatibilist resolutions to the problem of freedom² must wear an appearance of superficiality, however serious or deep the reflections from which they originate, until what they offer by way of freedom can be compared with something else, whether actual or possible or only seemingly imaginable, which is known to be the best that any indeterminist or libertarian could describe.

A sympathetic and serviceable statement of libertarianism cannot be contrived overnight, nor can it be put into two or three sentences, which is all that some utilitarian and compatibilist writers have been willing to spare for the position. If they were more anxious to destroy or supersede libertarianism than to understand and improve it, this was natural enough; but time or human obstinacy have shown that the issue is too complex for such summary treatment. What follows is offered as a small step in the direction of a more reasonable exposition. It concentrates on two or three points, where many need attention. If the treatment of these two or three points has the final effect of making the position even less credible, or of making me sacrificial scapegoat for oddities which persist, I still hope to have shown that the libertarian perceived something which was missed by all extant compatibilist resolutions of the problem of freedom; and that the point the libertarian was making must bear upon any future reconstruction of our notions and practices.

I What the libertarian means by ‘He could have done otherwise’

The libertarian insists that a man is only responsible or free if sometimes he could do otherwise than he does do. It must be genuinely up to him what he chooses or decides to do. But what does this mean? Let us begin with three clarifications.

(i) It is characteristic of the libertarian to insist that for at least some of the things which the man with freedom does, or plans, or decides to do, he must have a genuine alternative open to him. That is, for some action A and some action B, where A^B, he must be able to do A and he must be able to do B. But does the same apply to what the man with freedom thinks, what he believes, and what he infers?³ In another place,⁴ I have given an argument, whatever it may be worth, whose purpose was to show that the notions open choice, decision, alternative, up to me, freedom have a different point in the realm of belief, the state whose distinctive aspiration it is to match or represent the world as it is, from their point in the realm of action and volition. For of action and volition the proper province is not to match anything in the physical world but to affect or act upon the world. The world and its causal properties, whether or not these constitute it a deterministic world, are the unarguable framework within which action takes place; but for the libertarian it is typical and proper to insist that nothing in that world should completely determine the ends, objectives and ideals with which the free agent, if he is truly free, deliberates to change that world. There is no question of requiring of ends and ideals some correspondence with, some sentence-like satisfaction by, the things in the world. (If the onus were anywhere then, as Miss Anscombe has suggested, it would have to be the other way about.) The libertarian ought, on the other hand, to be content to allow the world, if it will only do so, to dictate to the free man how the world is.⁵ Freedom does not consist in the exercise of the (colourable but irrelevant) right to go mad without interference or distraction by fact.⁶ Alternatives of the kind which the libertarian defines and demands are alternatives in the realm not of theory but of practice.

(ii) To say that an agent is doing B or will do B and not A, and that there is something else, A, which he can do, is to say something ambiguous, even though (ignoring permissive and epistemic contexts) ‘can’ itself is most likely univocal (see (iii) below). A may be something the agent can generally do, for instance, or something he can for such and such a stretch of time do, given the opportunity. It is true and important that the latter claim is confirmed if the agent’s wanting or trying to do A at an appropriate moment during that period is a sufficient condition of his producing a non-fluke performance of A. But read in this way he can do otherwise is irrelevant to what concerns the libertarian. What organises the whole dispute, and what holds the libertarian’s position apart from his present day opponents’ position, is rather his treatment of another question: if physical determinism is true, is there ever something different from what the agent will in fact do at some time ti such that the agent can at ti do that other thing at ti instead? If physical determinism is true, then the libertarian maintains that such an alternative is never really or truly available to the agent (see Section III). Sometimes earlier actions do completely determine successions of later events and actions. According to the libertarian, however, there can only be true alternatives if there are at least some movements or actions or mental events which, whether or not they completely determine their immediate successors, are not themselves entirely determined by some predecessor. (Of course, this is only a necessary condition of alternatives or freedom of action.) He readily allows that even if there were not such successions, we could, if we wished to ignore all sorts of relevant facts, mechanically continue to draw our conventional distinctions between different kinds of situations between acting voluntarily and acting reluctantly, between control and non-control, between freedom and constraint. But determinism undermines their whole point, he says. It whittles away too much that is important from the notion of responsibility. It transforms it out of recognition. True freedom cannot be maintained by holding onto distinctions for which there is no factual backing or consistent rationale.

(iii) Though the sentence schemata he could have done otherwise and he could have done A instead of B may be used with varying truth conditions, one may hope to explain all these variations by differences of complementation with respect to (a) the time or period for which the ability subsists, (b) the particular fully specified value of the action variables, and (c) the time relevant for the acting itself. Can itself is, in my own provisional opinion, a unitary semantical element.⁷ But those who have distinguished, e.g. a ‘general’ can from ‘particular’ can have performed an important service in forcing us to be clear about what exactly it is that a man could or could not have done. The (b)-place must be carefully and fully specified. The provision of two slots (a) and (c), for the times of the ability and the performance respectively, may seem questionable. But consider the fact that I may now, in Baker Street at 9.55 a.m., be able to catch the train from Paddington to Oxford at 10.15 a.m. Eight minutes later, however, at 10.03 a.m., if I have not progressed from Baker Street, then, given the state of the Inner Circle line and Marylebone Road, I shall certainly be unable to catch the train. What we have in this example is not a special case but a specially clear case. Both slots are always there we cannot create them specially for the train case but when they both take the same temporal specification (as they must in ‘he could have done otherwise’ in at least some important occurrences) then the ellipse of one of them is surely natural and intelligible enough.

So much for the sentence he could have done otherwise as it figures in the dispute. The other urgent need is for a clarification of the determinism which the libertarian finds incompatible with his understanding of the sentence.

II What determinism signifies

J. L. Austin once maintained that determinism was ‘the name of nothing clear’.⁸ But as a second-level non-scientific theory that the world admits of explanation by a certain kind of ground level scientific theory, it seems to me that the thesis can be made as plain as ‘causality’ and ‘explanation’ can. Whatever his other difficulties, I think the libertarian must find it depressingly easy to indicate what it is that he is afraid of.

I propose to say that a scientific theory for a subject-matters is deterministic if and only if the theory possesses a store of predicates and relation-words for the characterisation of s-items (events, situations), and affirms lawlike general statements L_x, L₂, … L_n, such that for every s-item S_j it can find a description D_j, and s-item s_i with description D_i which occurred some t seconds earlier, and a law L such that L implies (if a Di event occurs, then a D_j event occurs t seconds later).⁹

A deterministic theory is adequate if its vocabulary of descriptions D₁ D₂, … D_n and its laws L₁ L2, … L_n together yield explanations which are of universal correct applicability and the statements L_x, L₂, … are true.

As a first attempt then let us say that determinism is the theory that for every event (.situation, state of the world or whatever) there is a true description D_j and an adequate and deterministic theory T which explains that event under Di. I suppose our reason for thinking that this might hold is science’s spectacular success in extending again and again the number and variety of events for which it can find theories with the title to be in my sense both adequate and deterministic.

Someone may comment that it is hardly surprising that we have discovered the regularities which were there to be discovered; that our success shows nothing about the residue; nor does the possibility of such success really guarantee the operational or empirical intelligibility of the thesis of determinism. Perhaps it is not intelligible, it may be said. The charge ignores falsification however; and those who persist in subscribing to determinism (in spite of, e.g. quantum phenomena) surely might reply to the whole objection with this question: ‘How big then is the residue? Can there really be, what the objection purports to achieve, an a priori estimation of it?’ At this point, however, we stumble upon the widespread idea, presumably shared by the objector, that every situation must be infinitely describable. If it were, then getting evidence for determinism would certainly be like filling a broken pitcher.

Those who claim to perceive this infinite describability in nature may try to support the objection in the following sort of terms. All we can do in a causal investigation is to pick out and test causal strands from a total physical background which is provisionally regarded as the ‘normal’ background;¹⁰ there is no logical question of this procedure, the only operational procedure they say, either terminating or issuing in finished law-like generalisations which are closed and not subject to a never ending process of qualification.

Let us first answer the difficulty about generalisations and then return to infinite describability. Adequacy in our first formulation required the strict and universal truth of the laws employed in deterministic explanation. If one says all F’s are G’s then one means all F’s; and if some restriction of the conditions under which all Dj events at t are followed by Dj events at ť is needed, then the restriction must, for purposes of this determinism, be made explicit. A body falling near the surface of the earth for t seconds will cover a distance of exactly 16t² feet, for example, provided it is in a vacuum and provided it is falling freely. The hypothesis of determinism which we are considering precisely entails that in due course such qualifications can be everywhere spelled out and completed.

It is true, of course, that the objection is making an important point about the discovery of physical laws, and about the way in which everyday conceptions of causality lead into scientific ones. In deference to it, and in deference to Hart and Honoré’s analysis of causality, we could meet the point in another way by saying that an explanation holds universally if either (1) there are no apparent exceptions to the predictions made by the use of the law or laws L which cover it; or(2) every such apparent exception can be explained in terms of an interference (a) describable by the vocabulary of the body of theory to which L belongs, and (b) for which there is in its own turn an explanation in terms of an adequate theory, this theory itself being compatible with L. Whatever one thinks of this strategy, it enables one to suppose...