Introduction
There is a concept in neuro-physiology called the Readiness Potential. It can best be described by reference to the following experiment: a subject is asked to raise her hand at a time of her own choosing. She is wired up to a device that measures EEG potential – that is electrical activity in the brain. The results show spikes in brainwave activity about three-quarters of a second to a second before she raises their hand, whereas her conscious choice to raise her hand occurs more or less simultaneously with her actually raising it. The existence of the Readiness Potential would seem to show that it is possible to predict certain human actions, prior to the conscious decision to take them (I will refer to this as ‘Experiment 1’).
For the purposes of the discussion in this essay, I am going to assume that these spikes have some physical cause, and that there is no in principle reason why we could not go back further in time and make a measurement of the earlier actions or states of affairs that would lead to the subject raising her hand. Depending on the state of our knowledge of the laws of physics and the accuracy of our measuring equipment, we may be able to go back arbitrarily far in time and still be able to predict when she will raise her hand. In practice, the number of different physical factors forming part of the causal web will increase the further back in time we go, making it in all likelihood practical only to make accurate predictions immediately prior to the decision being made.
This account as given would appear to pose a serious threat our pre-philosophical notions of freedom and moral responsibility.[1] We would normally require that to be free, the actions we take are ours – that is, we own them in some way; and that we have control over them.[2] An essential element of control is that at the moment of choice we have the ability to do otherwise than that which we in fact do. Experiment 1 would appear to show that our willing of an action – which we might suppose grounds us with both ownership and control of an act – in fact comes after the act being bound to happen. Thus the subject’s willing of the act seems simply epiphenomenal and of no explanatory relevance to the occurrence of the act.
Furthermore, when we look at the subject from outside the whole causal system as a experimental scientist might, taking what one might call a God’s eye view of the system, we can see that the subject has no real choice at all. Her actions or choices seem to be completely determined by the rules under which the system operates. It is only when we are sitting inside the causal system that questions concerning freedom seem to arise.
However, as beings that live in the universe, placing ourselves outside the whole system – the system being the physical universe - is something that we can do only in our imaginations. The problem of being part of the system under consideration is a very real one: being part of the system we seem to be confronted with our own freedom and the freedom of other human beings – beings that freely choose their actions and are able to do otherwise.
Now consider the following putative experiment, which is a variation on Experiment 1. This experiment is the same, save that when the measuring device makes its prediction, the prediction is shown on a computer screen that the subject can see. In other words, the subject can see what it is that she is predicted to do (I will refer to this as ‘Experiment 2’).
As far as I am aware, this experiment has never been conducted but we can indulge our imaginations to suppose what might happen. For ease of exposition, suppose that the predictor predicts which hand - left or right - the subject will raise and the prediction is that the subject will raise her left hand.
Then there are a number of potential results: -
· the subject may feel a mysterious compulsion to raise her left hand in spite of her volition;
· believing the predictor to be infallible she may resign herself to her fate and raise her left hand;
· taking into account the prediction she may raise her left hand;
· taking into the prediction, she may raise her right hand, or, for that matter, do anything else you care to think of.[3]
Of these results, the first is a severe challenge to our notion of freedom. If we were to feel a mysterious compulsion such as this then we would not consider ourselves free. External compulsion is one of the factors that we uncontroversially use to excuse our responsibility for an action.
The second result appeals to the notion of fatalism. This is a confused notion, because, in resigning herself she does make a choice to do as the computer predicts - falsifying the contention that she in fact has no choice. In fact, she is simply confused and gives too much weight to the prediction when considering the factors to take into account when making her decision.
The third and fourth results are equivalent in that the subject makes a choice. However they can be distinguished in that in the fourth example the prediction is falsified. If this were to happen it would falsify the thesis that our behaviour is utterly predictable as the predictor would have been wrong.
This leaves the third result. In this case it appears that the subject can choose to raise her left hand, that she freely chooses to do so, and yet this choice is completely predictable under the rules of the system. This is, I think relevantly similar to what are called Frankfurt-style cases in which Frankfurt proposes that we can be in situations where are unable to do otherwise, and yet still be morally responsible (Frankfurt 1969).
However, this seems to me to be wrong. If this experiment were to be repeated a large number of times, and on each occasion the predictor correctly told us which hand we were going to raise, we would not only be severely unnerved, but our notion that we were in fact free would, I think, be seriously challenged. We might suppose that, after all, there was some mysterious compulsion acting on us. Rather than being directly aware of this compulsion, we might infer its existence and suppose it to be operating at a subconscious level.
So, in order to feel truly free, we seem to require the fourth result to be a possibility, and yet this result is incompatible with the success of the predictor. Attempting to reconcile these irreconcilables will be the aim of the remainder of this essay.
My purpose in writing about these experiments is that they highlight a number of the intuitive concerns we feel when we think about free will. More than that, they turn what have, until now, been merely theoretical threats to our freedom, into real, actual threats that are part of the world of scientific enquiry. However, taken together they can, perhaps, also give us an indication of how we might start looking for solutions. In the following few paragraphs I will sketch out how I propose to do this.
I intend to explore how freedom appears to be incompatible with determinism (and, for that matter, indeterminism). If we can find a way of way of reconciling freedom, free will and moral responsibility with the challenge presented to it by determinism, then we will have made an important contribution, as it is a practical question of some importance. Smart (1961:292) puts it as follows: ‘When in nineteenth-century England, the rich man brushed aside all consideration for his unsuccessful rivals in the battle for wealth and position and looking at them as they starved in the gutter said to himself, “Well they had the same opportunities as I had. If I took more advantage of them than they did, that is not my fault but theirs,” he was most probably not only callous but metaphysically confused.’ If we can dispel this metaphysical confusion, it may help us decide under what circumstances it is appropriate to assign blame and praise, when to punish behaviour and when to excuse it.
Certain philosophers - incompatibilists such as Peter van Inwagen – argue that freedom and determinism are incompatible. They then argue that since we are free, determinism must be false. Personally, I think that this is begging the question, but I concede that the direction of argumentation is not so obvious that it can simply be assumed, and I will discuss this further below when I discuss basic beliefs. If it is not obvious that we are free, then if an argument can be advanced that can reconcile freedom with determinism, this would be desirable: the question whether the belief that we are free is a basic belief loses some of its importance. Introducing indeterminism as a way out of the apparent incompatibility of freedom and determinism also raises problems as it undermines another element we consider essential to freedom – the element of control. The solution, I believe, does not lie upon that road.
We also cannot take the obvious escape route – asserting that we are in fact, not free. No philosopher, as far as I am aware, seriously argues that a fully rounded, sane individual, under normal circumstances - that is, in the absence of coercion, hypnosis or the like, lacks moral responsibility and/or free will. Even if they do, their behaviour would in all likelihood betray them: they would display Strawson’s ‘reactive attitudes’, and hoist by their own petard would have to concede that their moral reactions were either irrational or insane - “What would become of your laws, your morality, your religion, your gallows, your Paradise, your Gods, your Hell, if it were shown that such and such fluids, such fibres, or a certain acridity in the blood or in the animal spirits alone suffice to make a man the object of your punishments or your rewards?” (quoted by Smart 1961:292). As Smart says, ‘We do not accept the complete suspension of moral judgement implicit in de Sade’s view’. Similarly, Strawson (1963: 16) says, ‘it is useless to ask whether it would not be rational for us to do what it is not in our nature to (be able) to do.’ It would be brave indeed then to argue that we as humans are creatures that are neither free nor morally responsible.
Thus it seems that we are at an impasse, unable to reconcile our deep feeling – and indeed psychological need for freedom – with the challenge presented to it by the concepts of determinism and indeterminism. I will suggest that there is a way around this impasse and that it involves both a deeper understanding of the important distinction between observer and observed and of the importance of information feedback. This is hinted at in Experiment 2 but I will flesh it out further below. When I have done so, one question that will then arise is whether the freedom that then emerges is real, metaphysical freedom or simply a constraint on what we can know. I will suggest that, in fact, this is a false dichotomy and adduce in evidence certain aspects of Gödel’s Incompleteness Theorem.
Finally, I will suggest that this approach to the concept of freedom might also be applicable to other concepts in which we are presented with a stark difference in our intuitions depending on whether we take a subjective or an objective view: these problems include, but may not be limited to, the subjectivity of consciousness and the now of time.
Determinism and indeterminism as a threat to freedom
The account of Experiment 1 highlights the notion that determinism may be a threat to our freedom.[4] To clarify this, I will adopt what is essentially van Inwagen’s argument (1975: 186-187). He provides an sketch of determinism which is as follows: there is a state of the universe that can be described at time t1 and a different state of the universe at a later time t2; further, the laws of nature combined with the state of the universe at t1 necessitate the state of the universe at t2. The concept ‘laws of nature’ is necessary because otherwise one could posit any subsequent state of the universe and say that it was necessitated by earlier ones. In order for the notion of determinism to do some work, there needs to be some sense in which subsequent states of the universe are contingent on the laws of nature or some similar concept. Otherwise we simply have a universe where what happens happens.
I think that this van Inwagen’s need for a concept such as laws of nature is allied to our intuitive notion of causality – that earlier events or states of affairs cause later ones. One of the implications of determinism is this: if we had a complete description of the state of the universe at a particular time and a full knowledge of the laws of nature we should know the future. Furthermore, if we were to roll back time to time t1, and then to start it running again, the universe would always reach the same state at time t2.
This is how I understand the concept of determinism and I will adopt it as a provisional starting point. As my argument throughout this essay progresses, this way of looking at determinism will be undermined and through that, I think, a way of understanding freedom will emerge.
For the time being though we will adopt van Inwagen’s approach. Having arrived at a definition of determinism we can now assess its implications for freedom. These are simply that given the present state S1 of the universe at time t1, the laws of nature necessitate that at time t2 it will be in state S2 rather than, say, S2*. There is no way in which the universe can be in state S2* at time t2.
This is where the threat to our freedom arises. If I am to be held responsible for a certain state of affairs S2 then I must be able to control that event. But on our definition of determinism the universe must arrive at state S2 because previously it was in state S1. If we are to be held responsible, we must have control over state S1, yet this is a state over which we have no control. S1 is dependent on yet earlier states of the universe: states which existed prior to our birth and which we can trace back either to the Big Bang (if we are adherents of that theory) or ad infinitum if we believe that the universe had no beginning. In either case, S2 is not down to us.
It may be objected that Experiment 1 is not conclusive concerning the truth of determinism as defined. I happily concede this. It may be that the most we can conclude is that our actions are merely predictable. There is a another school of thought – labelled predestinarianism[5] which contends that if our actions are perfectly predictable then we are not free – and there is some strength in this view, as I think is revealed by our intuitive response to the thought experiment outlined above (specifically our response to the predictor telling us what we will do - see page 4). If you disagree with this imagine someone accompanying you around telling you precisely what you were going to do immediately before you did it. I contend that this would severely threaten your concept of yourself as a free being. [6]
Either way, whether we believe that there is a causal mechanism of some sort which necessitates our actions, or if we deny this and merely say that our actions are predictable, we are presented with a challenge to our concept of freedom.
However, the notion that determinism limits our freedom has been subjected to various attacks. Of these, one of the principal ones was by Harry Frankfurt, who in his famous article Alternate Possibilities and Moral Responsibility (1969) proposed a thought experiment in which a human subject appeared to have moral responsibility even though he did not have any ability to do otherwise (one of our supposed prerequisites for freedom). I originally considered Frankfurt’s argument in a previous MA essay and rejected it on the grounds that, despite initial appearances, the subject was not responsible for his actions. Having reconsidered the problem, I now think that the reason that his arguments appear to show how determinism and freedom may be compatible is that they implicitly assume that we must stand inside the system we are considering when we are thinking about moral responsibility, and not place ourselves outside it. I discuss this systemic kind of thinking further below.
There are other arguments in favour of determinism and freedom being compatible: one view is that it is in realising that we have no choice over what we do that we become free. I reject this argument. There is a suppressed premise in this argument that we are in fact able to reach that conclusion that we are not free. Of course, in a deterministic universe this is not necessarily the case. The laws of nature may be such that we never reach such a conclusion. Furthermore, it requires a radical revision in what we consider freedom to be: it supposes that we become free by realising that we are not free. This to me is simply incoherent.
Where does this leave us then? Bok (1998) sums it up thus: ‘on Van Ingwagen’s view determinists who recognize these facts must therefore either give up their belief in determinism, learn to live with the fact that they have beliefs that they know to be inconsistent[7], or give up on deliberation altogether - in which case … they will either “move about in random jerks and scuttles, or …withdraw into catatonia”’.
What of indeterminism then? There are philosophers who argue that since we are free, and since determinism threatens freedom, there must be an element of indeterminism in the universe.
An indeterministic universe is one in which, given state S1 at time t1, it is not necessarily the case that the universe will be in state S2 at time t2. Since the universe will always be in some state at time t2, I find that the most appropriate way of thinking about the notion is to imagine that if we were to turn back time to time t1, and start the clock running again, then, when we reach time t2 again, we should find that the universe is not necessarily in state S2, but might be in some other state, perhaps S2*. Some philosophers talk of probabilities of reaching either state S2 or S2*. The crucial element is that it is not necessarily the case that state S2 will be arrived at. It might be, but it might not be. I have avoided using the concept of laws of physics in the definition of indeterminism here because, depending on how one understands them, particularly if one views them as Humean regularities or some variation thereon, then in an indeterministic universe, it may not make sense to say that there are such things as laws of physics.[8] In order to avoid this problem I choose to think of indeterminism in a way which avoids using the laws of physics.
I believe that indeterminism does not bring us any closer to a useful account of freedom. If certain acts are truly undetermined then some of our actions are not completely under our control. I don’t wish to give a full account of control here because there are clearly cases where our notions of control are undermined but we would still say we have free will. For instance, if we were tied up, our bodily movements would be restrained and our control of them undermined, but we would, I think, still wish to say that we had free will (see Albritton 1985). More generally, if our actions are undetermined, then there is, or may be, a possibility that whatever we decide to do, we will be unable to guarantee that it will come off.
You may still say that, whether or not we have this kind of control, it doesn’t threaten our notion of freedom of will. So far I concede this. However, the crucial point and the element that does, I believe, undermine it, is that in an indeterministic universe as defined, our mental states (or our ability to make decisions, or what we will) are not completely under our control. There is the possibility that if we were to run events over again, then, despite all the knowledge of the world that we possess, and despite our character and intellectual capacities leading us to make Decision A, we would in fact make decision B, due to factors beyond our control. While this seems to be in accordance with one necessary element of freedom - the ability to do otherwise – it certainly does not accord with the requirement that our actions are owned by us.
Normally, if someone does an act that is out of character, we would see this as an indication that there may be some element present which undermines moral responsibility – for instance, an insane person is one who acts severely out of character, if indeed they display any consistent character at all. In an indeterministic universe, some of our choices are inevitably out of character; at the moment of decision there is an element of randomness about what we choose to do. Thus it seems that we do not own our actions in the way that we require, or at the very least that our notion of ownership is undermined.
Freedom as a basic belief
‘Sir, we know our will is free, and there’s an end on’t’ Dr Johnson.
Is there another approach that we can take? It appears that by considering the dichotomy between deterministic and indeterministic universes, we have exhausted all the possibilities as to how the universe might be and found that they are both unsatisfactory for believers in freedom. However, perhaps we can come at the problem from the other end.
Some philosophers - Sartre, for example - consider that we are clearly free but subject to constraints throughout our lives. I would suggest that we all have a very strong intuition that we are truly free in a very important sense, even if we struggle to clearly define that freedom. As I have said earlier, as far as I am aware, no philosopher seriously argues that we are not free in a moral-responsibility-undermining sense. Adopting Bok’s (1998) account, an incompatibalist/determinist in conversation with Sartre’s young man who is deliberating whether to join the Free French or help his ageing mother, can tell him that the universe is subject to inexorable rules and that in fact the choice was made long ago. No amount of explanation though will prevent the young man responding, “the truth of determinism gives me no reason to abandon either my mother or my country; and as for my ‘alternatives’, it implies only that whichever choice I make will have been the only choice that I could possibly have made…I am left with exactly the same decision, to be made among exactly the same alternatives, whether or not determinism is true. And now, if you will excuse me, I’ll try to make it.”
On a foundationalist account of knowledge, there are a number of things that we believe, which we believe not on the basis of evidence but simply because we are presented with the world in a certain way. Plantinga gives examples of such beliefs – which he calls basic beliefs – such as ‘I see a tree’, ‘I remember eating breakfast this morning’ and ‘that person is angry’ (1981). A very short logical jump leads to abstractions based on these beliefs; these might be ‘there is a world out there’ (abstracted from, ‘I see a tree’), the world is more than five minutes old (abstracted from ‘I remember eating breakfast’), and ‘there are other minds’ (abstracted from ‘that person is angry’).[9] Our other beliefs are built on the foundations of these basic beliefs.
Secondary beliefs, such as the belief that the world is round, or that Durham is north of London, are not basic but based on evidence. These secondary beliefs should be traceable back to our basic beliefs if we are to give them any credibility. Eventually, we find that we have a whole network of beliefs, some of which are more fundamental than others, that go to make up our knowledge of the world. We might also concede that we can be mistaken about some of our beliefs - such as the belief, held for much of human history, that the sun revolved around the earth.
We are able to question most of our beliefs and this, of course, is what Descartes did, when he attempted to put knowledge on a firm foundation. So we can question whether there is an external world, or whether the world is more than five minutes old. It is more open to debate whether we can question the specific belief – such as ‘I see a tree’ – on which the abstraction is based. However, we might concede that we could be suffering an hallucination. What we could not question, I think, is that we had had a certain sensory experience: in the case of ‘I see a tree’, a visual sensory experience of brown and green patches. These beliefs are beyond challenge. An important point to note here is that the truth of these basic beliefs is difficult if not impossible to establish. For instance, I don’t think that it would be possible to prove the truth of ‘I see brown and green patches,’ and similarly, attempts to prove that there are other minds have failed; further I cannot see how one could convincingly establish that the world could not have been manufactured five minutes ago with all of us implanted with fake memories – although, no one is likely to convince me that it is true. One simply holds these beliefs in order to function in the world. The objective truth of the belief is of minor importance.
In fact, there are a number of false beliefs that we do adopt because they seem to benefit us despite the fact that they are not true. One famous example of this is the belief adopted by soldiers in battle situations of their own invulnerability. Clearly they are not invulnerable, but, in the heat of battle, this belief would appear to help them to act in situations where they might otherwise freeze. Another example is the belief most people have that they are of above average intelligence. Only fifty percent of the population can be above average intelligence, but perhaps the belief helps people to lead happier and more successful lives irrespective of their actual intelligence level, so this belief might be a useful one to hold even if it is false. Belief in a benevolent God might be yet another example: believing in God may give us purpose and a sense of having a place in the universe which is beneficial, whether or not He actually exists.
Leaving aside what might constitute a basic belief, the question that I wish to raise is whether the belief, ‘I am free,’ is, or might be, a basic belief. If so, the objective truth of that belief is not important. It simply is the case that we cannot but help think of ourselves as free in order to function in the world. The question of whether we are in fact free then becomes a quasi-problem in much the same way as the question of whether the world was created five minutes ago is.
But I believe that we cannot simply assert the basicality of our belief in freedom. We can and do draw distinctions between greater and lesser degrees of freedom; we still find ourselves presented with cases where we would want to absolve a party of responsibility for their actions because their freedom was constrained, whether or not they believed that they were free. For instance, we would generally want to absolve the insane from responsibility for their actions. But an account of freedom based on the notion of basic beliefs, does not help us here. A man could believe that he was Hitler on a mission to invade Western Europe, and still believe himself to be acting freely, yet we would not want to hold him morally responsible for his actions.[10] We need an account of freedom that provides us with excusing conditions.
How are we to draw this line? This problem presents itself to us in any one of the accounts outlined above – determinist, indeterminist or foundationalist. We still need to feel that our actions are ours for them to be free; we still need to feel that we have some control over our actions; we still need to feel that we could do otherwise; we need to feel that our actions are unpredictable to the extent that no one could say precisely what we are going to do next, and yet they need to remain in character.
The fundamental problem of the will
The fundamental problem, as far as I am concerned, is this: ‘(1) Nothing can be causa sui – nothing can be the cause of itself. (2) In order to be truly morally responsible for one’s actions one would have to be causa sui, at least in certain crucial mental respects. (3) Therefore nothing can be truly morally responsible.’[11] In order for us to be free, our wills need to be a kind of prime mover unmoved or uncaused cause. But in a causally closed deterministic universe, it seems that there is no room for this; no room, in fact, for the human will to have any say at all. If everything that occurs is causally necessitated by previous events, then there is need for or room for the human will. Conceding indeterminism does not help for the reasons I have already given. According to Campbell (1951), freedom appears to be ‘contra-causal’; there is no ‘unbroken causal continuity’.
The kind of will that remains then, if it exists at all, appears to be something non-material: a kind of Kantean noumenal-self outside time and space or Buddhist-like connection the universal whole. A will of this kind may be free, but it has little if any connection to the world. No account is given as to how the will affects what we do, if indeed it does at all. An account of the will such as this does not help us because it seems to have no causal interaction with the universe, which is required if we wish to hold people responsible for their actions. Parenthetically, of course, Kant also says that we can behave perfectly by sheer force of will irrespective of our personalities, but this kind of will is clearly one that does interact with the universe.
A non-material will, then, is not the sort of will that we need if we wish to find a compelling account of freedom. It merely moves the mystery of the will to a point beyond exploration. Instead, I wish to attempt to find an account of freedom that is materialist. In such an account, whatever happens in the universe is as a result of the forces of nature acting on matter – nothing more nothing less. Such an account of course is difficult easily to reconcile with the complex reasons and motives - such as our desire to avoid pain or gain pleasure - that we humans have for our actions. However, I am going to assume that given a full understanding of the universe, these reasons and motives could be traced back to these laws of nature acting on matter within the universe. What we think of as our reasons for doing something are either epiphenomena or are agglomerations of these forces manifesting themselves as what we think of as reasons, character, motivation and so on. I believe that this is not dissimilar to a view such as Nagel’s (2003: 233) when he says that we can give an account of our actions that is ‘motivationally explicable’, and that ‘my reason for doing something is the whole reason that it happens’ at least from my perspective (my italics). But it also allows that why we do something is subject to influences in the world.
This is preferable to assuming a non-material notion of will. I suspect that even if we were to allow such a non-material notion, an exploration of the precise nature of how it were to work freely would reveal exactly the same kinds of problems as we are beset with anyway. I am unwilling to adopt such a solution and will instead assume that if there is a solution it a materialist solution.
An attempt to propose a solution
‘When we view action from an objective or external standpoint, some of its most important features seem to vanish under the objective gaze. Actions seem no longer assignable to individual agents as sources, but become instead components of the flux of events in the world of which the agent is a part’ (Nagel 2003: 229)
We are happy to absolve people from moral responsibility under certain circumstances. We do so in situations in which a person cannot do otherwise in a local situation. For example, if the driver of a car is attacked by a swarm of bees and loses control of his car, we absolve him of responsibility for his car swerving across the road and crashing head on into another vehicle because he is under a form of constraint; similarly if a gun were held to the head of the driver or if he were hypnotised and ordered to swerve, we would say that he was coerced and absolve him from moral responsibility. Dennett (Elbow Room 1984: 104) calls such situations, ‘local fatalism…consider the man who has thrown himself off the Golden Gate Bridge and who thinks to himself, as he plummets, “I wonder if this is really such a good idea”’. No amount of deliberation will help him at this point.
This is uncontroversial. What is, however, is the notion that in a deterministic universe we are under a kind of global constraint and that it is this global constraint that seems to absolve us of moral responsibility. The problem is that there doesn’t seem to be any logical cut off point at which it is fair to say that the car driver suddenly acquires moral responsibility. For instance, if we ditch the bees and suppose instead that the driver had some kind of brain seizure, we would still, I presume, absolve him. What if, instead, he had had an argument with his wife and was so preoccupied with that, that he failed to concentrate on his driving? Now I suspect our intuitions would be that we would not absolve him – because here he does not appear to be acting under a constraint. But on the model of the universe that I have presented, what is the distinction? He either lacks control over what he thinks about to exactly the same extent that he lacks control over his actions if he were subject to a seizure or he does not. In favour of the proposition that he lacks control: there is just as much a causal explanation, which is equally not under the agent’s control. On the model of determinism that I have presented, I find it difficult to see how he does have the kind of control necessary for moral responsibility – how does the distinction arise without assuming a will of the kind that has already been ruled out? In favour of the proposition that he does have control: under those specific circumstances he is free to do otherwise – but this is simply begging the question.
Nagel puts it as follows: ‘there is no way of preventing the slide from internal to external criticism once we are capable of an external view. It needs nothing more than the ordinary idea of responsibility. The problem of free will, like the problem of scepticism, does not arise because of a philosophically imposed demand for external justification of the entire system of ordinary judgments and attitudes. It arises because there is a continuity between familiar “internal” criticism of the reactive attitudes on the basis of specific facts, and the philosophical criticisms on the basis of supposed general facts. When we consider the possibility that all human actions may be determined by heredity and environment, it threatens to defuse our reactive attitudes as effectively as does the information that a particular action was caused by the effects of a drug … the same is true when we expand the point to cover every way in which our lives can be seen as part of the course of nature, whether determined or not. No new standards come into it; in fact no demand of justification comes into it, since the challenge depends only on generalizing familiar standards of criticism. We cease to resent what someone has done if we cease to see the alternatives as alternatives for him (Nagel 2003: 244).
Thus, according to Nagel, it seems that unless we can find a relevant distinction between local and global determinism, we have to rule out moral responsibility and freedom. Fortunately, I think we can, and a consideration of Experiment 2 will aid us in this.
When we discuss freedom, it is noticeable that there is one element that is essential to the discussion – all freedom relevant discussions involve human beings (or other intelligent beings, possibly including the higher animals). We do not think of a system devoid of humans as free, or potentially free, in any philosophically interesting sense. We do not raise questions about the freedom of the planets to move in orbits other than those that they in fact do. We do not discuss their moral responsibility for so doing. All of our freedom relevant problems relate to humans.
Now let us return to the experiments referred to at the start of this essay. If we were to be told after we had performed a series of actions that every one of them had been predicted and were shown what those predictions were, we would no doubt reflect that our freedom was challenged; we may not be willing to concede that our actions were unfree, but I suspect we would feel a little uneasy. I have already referred to predestinarianism, which contends that we are unfree precisely in circumstances in which our actions are completely predictable.
However, if the predictions are shown to us before we make our decision, this very act of showing us the prediction gives us another factor to take into account as we make our decision. I have already suggested that when we make a decision the factors that determine what we decide are ultimately dictated by the laws of physics and previous states of the universe. At a more human level, we would say that what we do is determined by our character, our physical and mental abilities, our state of mind, our state of knowledge and so on. At another level of explanation we might say that the decision is determined by the firing of certain neurons within the brain, which are determined by the movement of atoms and so on. In this case, crucially, one of the factors that will affect our decision at any one of these levels will be the prediction itself. If we look at the point of decision as a confluence of various mental factors such as state of mind, knowledge and so on it is easy to see that a prediction about what we will do is one of the items of knowledge that we might take into account. At the atomic level, adopting a materialist account, we can suppose that new information will affect the firing of particular neurons within the brain and the movement of various atoms. One could also attempt to cache out the explanation in terms of states of the universe and the laws of physics. I will, however, refrain from doing so and I hope that the reason for this will become clear below. Before that I will take a short detour examining other proposed accounts of freedom, and suggest that they can be subsumed into the account I propose.
An account proposed by Harry Frankfurt (1971) is that freedom is related to second-order desires: ‘first order desires are desires to do or to have various things; second order desires are desires about what desires to have or what desires to make effective in action. In order for an agent to have both freedom of action and freedom of will, that agent must be capable of governing his or her actions by first-order desires and capable of governing his or her first-order desires by second-order desires. Charles Taylor (1976) says ‘our characters and desires are not so brutely implanted – or, at any rate, if they are, they are subject to revision by our own reflecting, valuing, or second-order desiring selves. We human beings - and as far as we know, only we human beings - have the ability to step back from ourselves and decide whether we are the selves we want to be.’ Acting on first-order desires does not make us free – if we did we would be like Pavlov’s dogs. This position ‘must be distinguished from the position of a creature that doesn’t suffer from the sense of helplessness because it can’t take up the external view toward itself. When a cat stalks a bird, no element of the cat’s self can remain outside as a detached observer of the scene, so there is no sense in which the cat can feel that he is not doing it. But because there is more to me than there is to a cat, I am threatened by the feeling that I do not really act when I act only on the basis of that internal view which suffices for a cat’. Simply reacting to various sensory stimuli then does not make us free - ‘the cat is more trapped than we are by instinct, fear and so on’. On Frankfurt’s account, then, our ability to reflect on our desires and either to act or refrain from acting on the basis of them, taking into account other conflicting desires and other more distant aims and goals is what gives us freedom.
What this account and Experiment 2 have in common is that there is an element of feedback in both cases. That is, in Experiment 2 the knowledge of what one is predicted to do is one of the factors that is taken into account by the agent when he decides what he will do. In Frankfurt’s account, the subject’s wants and desires are essential factors taken into account when he makes his deliberations. In other words, they are fed back into his considerations.
Now let us consider an observer O who is sitting outside the system, watching a subject S, and who as observer has a complete knowledge of the state of the universe and the laws of physics. It would seem that O would have complete knowledge of what S is going to do. But what if O were to then provide S with a prediction of his actions? We have already seen that S may then do something different because this prediction is one of the factors that S will take into consideration when making his decision. If so, then it seems in fact that O’s knowledge is not actually complete after all. Alternatively, O may say to himself, ‘if I provide him with the information then he will change his mind, if I don’t then he will not’. Again, his knowledge of what will happen within the system is limited (unless he knows what he as observer will do, but I am assuming that since he outside the system he does not have full knowledge of himself in the same way as he does what goes on within the system – you may object to this, but I consider it further below). It seems to me that there is no escape from this limit on the knowledge of the system on O’s part. He might try to avoid the limitation by telling S not only of the prediction, but also of his predicted reaction to it, but this leads to an infinite regress. For instance, O might realise that, ‘as a result of my telling him prediction x he will do y’. But what happens if O then thinks, ‘I will tell S the following: “as a result of my telling you prediction x you will do y”’. In this instance, O may very well do z! And so on…
This raises the question: can we know what we will choose to do while we deliberate? Nagel (2003) talks of a blind spot that prevents us from knowing what we will do while we deliberate. I suggest that it is this feedback element that constitutes the blind spot – ‘there is a sense in which even within classical physics some events must be unpredictable. If there are two predictors P and Q, they cannot predict one another’s behaviour. For by the definition of a predictor, small changes in P must lead to large changes in Q and vice versa. So in order for P to predict Q it must predict itself, but it cannot do this for reasons similar to those in Ryle’s Concept of Mind’ (Popper: 1951).
Unpredictability arises in the system by the very act of feeding back relevant information. The only way an observer can avoid this is to locate himself outside the whole system and refrain from interacting with it. In our universe that would mean an observer that was not spatio-temporally located – ‘if the Laplacian demon is to predict the universe it cannot itself be part of the universe, nor can it interact with the universe. The notion of a Laplacian demon is thus a physically unrealisable one’ (Smart: 294).
But for us humans, finding ourselves inside the universe, we are able to make predictions about what we will do and act on those predictions; however, no matter how accurately we know the rules by which the universe works, we find that unpredictability will arise from this concept of feedback. Of course, the examples of feedback I have used a stark and extreme, and in more general cases they are more likely to be like the second-order desire types of feedback I have previously discussed. It is in this ability to feedback information into the system that I believe we can say that freedom of the sort that we are looking for arises. The upshot of this is that other people can predict my actions, but I cannot, and they cannot tell me for certain what I will do.
Metaphysical or epistemic?
One might be tempted to object that this account of freedom is merely an epistemic constraint on what we can know. It says nothing about fundamental metaphysical freedom. After all, on this model of explanation, it seems that God as observer can sit outside the system and know exactly what I will do, as long as he does not interfere with it. He can do what I discussed earlier: he can turn back time, re-run it and know that the outcome will be the same. He can observe an experiment, such as Experiment 2, and as long as he doesn’t interfere with it, he will know exactly what will happen.
However, his freedom is constrained: the moment that he becomes an actor within the system, then he loses his ability to know the system completely.
Bok’s approach to this is to suggest that there are two types of possibility: the sort that compatibilists allow, which is the ability to choose between relevant possibilities, and ‘possibility tout court’, or determinist possibility. ‘If a free agent must have alternatives which are possible only in the compatibalists’ general sense, then determinism is compatible with her freedom. But if she must have alternatives which are possible tout court, then determinism implies that she is not free.’ and, ‘this argument shows only that no narrower conception of our alternatives is available to us while we deliberate, it might seem to show only that we must use this conception faute de mieux: that while ideally we might hope to narrow our conception of our alternatives to our possibilities tout court, our epistemic limitations prevent us from doing so, and thereby force us to settle for the broader compatibilist conception’(Bok 2003: 157 )
So perhaps God can see reality as it truly is and we are not fundamentally free. As long as it is possible to step outside the system and see it for what it truly is, then those of us inside it are constrained. However, I suggest that there is a move that provides an escape from this. We can suppose that there are two areas: the first is the universe that we inhabit; the second is that I will call God-space – that is, the realm of being that God inhabits. If we then consider this system as a whole, the same question arises for God. Is he free? Yes, if he considers himself simply as an actor within the system; but perhaps not if he were to imagine a yet greater being sitting outside the ‘universe plus God-space’ system as a whole, observing it, but not interfering with it. However, since God is the ultimate being, there is nowhere else outside this system to go. Even if there were, the same argument could be applied to that system. As Nagel says, ‘all accounts fail to allay the feeling that, looked at from far enough outside, agents are helpless and not responsible’ (2003: 231). I submit that looking at the system from that far outside is ultimately a nonsensical move.
Another argument in support of this view comes from logic, specifically Gödel’s Incompleteness Theorem (“the Theorem”).[12] I am not sure whether this argument is merely an argument by analogy - in which case I would be loathe to rely on it - or something more fundamental. I suspect it may be the latter.
The Theorem states that for logical systems of a certain level of complexity there are certain statements the truth of which cannot be established. An example is the Turing halting problem. There are certain cases in which it is impossible for a universal Turing machine (a theoretical machine which is analogous to today’s computers) to know whether a particular sequence of tasks that it has been programmed to perform will come to an end. This lack of knowledge is not merely a fault in the program or in the complexity of the machine. It cannot ever know, and the reason for this is fundamentally connected with the logic of the system under which it operates. This lack of knowledge applies not solely to computer programs but to certain other logical problems and a number of hypotheses in mathematics, such as, perhaps, Goldbach’s conjecture (the conjecture that every even number is the sum of two primes). Not only might there not be a solution, but it might be that we cannot know whether there is a solution or not. It may be the case that it is simply impossible, within the rules of the mathematics, to find a solution or know whether there is a solution. According to the Theorem, there are some propositions that we cannot know, and more worryingly, we cannot know that we cannot know them.
This relates to my thesis, because this unknowability arises only with logical languages of a certain complexity. At levels below this complexity, this problem does not arise, but here the language is constrained and the number of logical statements that can be made is limited – thus a complete description of the system cannot be given. Thus with logical languages there are two possible outcomes: either the language is simple and so the number of statements that can be made is limited (and hence we can only give a partial description of the logical universe) or alternatively the language is complex and then there are certain propositions the truth of which cannot be established.
The tipping point between the two is when the language becomes sufficiently complex that it can make self-referential statements, that is, statements such as ‘this statement is true’ or, alternatively ‘this statement is false’.[13] In other words, the point at which the system becomes complex but unknowable is when it becomes able to talk about itself, or adopting the language we have used previously, when it becomes capable of feeding knowledge of the system of which it forms part back into the system itself.
Is the Theorem simply limited to logic and mathematics? To some extent the answer to this depends on whether you are a scientific realist or not. If you believe that these laws, which are fundamental to logic and mathematics, do map on to the physical world then it may be that it applies to the universe as a whole. If so, then the universe is a Gödelian system and the extent to which it can be known is fundamentally limited. Insofar as the universe does appear to be a rule following system, then there does appear to be evidence in support of this proposition. However, a fuller examination is outside the scope of this essay.
My answer then to the metaphysical / epistemology dichotomy then is this: in metaphysics we are concerned with how reality is; in epistemology we are concerned with what we can know. But in the present discussion, what we know will to some extent determine how reality is, and reality determines what and how much we can know, so this may be a case where the distinction between the two disciplines actually breaks down. What we know affects and determines what is.
It should be noted that this account may require a revision of the definition of determinism that I outlined previously. Any system that includes all of its own elements within it - that is, any system that has no outside observer, will be subject to uncertainty should feedback conditions arise. If so, then determinism as a fixing of the all states of the universe for all future times is incoherent as it supposes that we can look at the whole system objectively from the outside. The system can still be rule following, and prior states can still lead to subsequent states following the laws of nature, but there is a limit to how much we can know about these states.
Final Thoughts
On this thesis then, freedom arises because of our ability to feed back knowledge into the system of which it forms part. One objection I can see to this is that one might argue that there are systems which are capable of feedback and yet which are not free - for instance, computers of a certain level of complexity. I acknowledge this, but point out that there are other relevant factors: it may be that our wish to deny computers freedom of the relevant sort is merely human parochialism; or it may be that such a computer is ‘free’, but not yet a moral being. Perhaps the level of complexity needed for moral responsibility is greater than for freedom. I do not imagine that computers can yet be supposed to be morally responsible, but would not wish to rule it out as their level of complexity increases.
The account of freedom that I have advanced also allows us to utilise all of the normal excusing conditions that we normally use to exempt people from responsibility. One failing thus far is that it does leave open the question of where responsibility begins and ends, but an investigation into different types and levels of feedback may provide us with a starting point. I will not address that here however.
I will finish this essay of by raising some questions. Why is it that feedback of the sort that I have discussed is so important? We have seen that it helps the subject of an experiment break out of his ‘programming’ and also that it leads to a kind of freedom in logical systems. Why should this be so? Is there an underlying unity to ideas such as freedom, feedback and, leading on from that, self-reference? If so, might this unity apply to other areas of philosophical enquiry where there appears to be a stark contrast between subjective and objective accounts. For instance, objective ‘b’ series accounts of time give a full description of events in time, yet they fail to capture what to a human is an essential element of time – the ‘now’. When we consider freedom, the importance of ‘the ability to do otherwise’ is a crucial factor, but this point of choice is something that occurs in time. If we express regret it is because we wished we had done otherwise at a crucial moment of decision. The concept of a particular special moment in time – the ‘now’ seems of particular importance in relation to freedom: it would be absurd to say that Napoleon would be free if he could now [in 2004] choose to invade or refrain from invading Russia. It is only at the moment of decision that the ability to do otherwise is relevant. This seems to require the notion of a ‘now’, and seems not dissimilar to the requirement for subjectivity - that is, being a part of the system - that I have suggested is essential for freedom. In the alternative, in the objective case – either sitting outside the system (in relation to our discussion on freedom) or adopting a ‘b’ series view of time (in relation to the latter), we are able to produce a system that is predictable and scientific, yet which fails to capture an essential element – the feeling of freedom or the concept of the now. It is this similarity that leads me to suppose that the search for any unifying factor may be found by clarifying our distinction between objectivity and subjectivity. Similarly, there are few accounts of consciousness that provide adequate explanation of the subjective element. Could an explanation of this also be found in our ability to feed back ideas into each other? These are questions for another time.
Gary Lorrison
Department of Philosophy
University of Durham
garylorrison@aol.com
References
Albritton, R. 1985. Freedom of Will and Freedom of Action, from The Proceedings and Addresses of the American Philosophical Association 59/2: 239-251.
Bok, H. 2004. Freedom and Practical Reason from in Gary Watson (Ed.) Free Will. Oxford University Press.
Campbell. 1951 Is Free Will a Pseudo-Problem? Mind, LX.
Dennett, D. C. 1984. Elbow Room. MIT Press. Cambridge, Mass.
Dennett, D. C. 1984. I Could not have Done Otherwise - - So What? The Journal of Philosophy, Volume 81, Issue 10: 553 – 565.
Frankfurt, Harry G. 1969. Alternate Possibilities and Moral Responsibility. The Journal of Philosophy, Volume 66, Issue 23: 829-839.
Frankfurt, Harry G. 1971. Freedom of the Will and the Concept of a Person. Journal of Philosophy, Volume 68: 5-20.
Fischer, John.1999. Recent Work on Moral Responsibility. Ethics, Vol. 110, Issue 1: 93 – 139.
Hofstadter, D.R. 1979. Gödel, Escher, Bach: an Eternal Golden Braid. Basic Books. New York.
Horwich, Paul. 1987. Asymmetries in Time: Problems in the Philsophy of Science. MIT Press, London.
Hume, D. 1936. Enquiries Concerning the Human Understanding. Oxford University Press. Oxford.
Kant, I. 1963. The Critique of Pure Reason, translated from the German by Norman Kemp Smith. Macmillan Press. London.
Kant, I. 1997. Critique of Practical Reason, translated and edited by Mary Gregor. Cambridge University Press.
Locke, John. 1975. An Essay Concerning Human Understanding (1690). Oxford Clarendon Press.
McTaggart, J.M.E. 1908. The Unreality of Time. Mind 17: 457-474.
Nagel, T. 2003. Two Problems, in Gary Watson (Ed.) Free Will. Oxford University Press.
Plantinga, Alvin. 1981. Is Belief in God Properly Basic? Nous. 1981: 41-51.
Popper, K. 1951. Indeterminism in Quantum Physics and Classical Physics. British Journal of the Philosophy of Science 1 179-188.
Ramachandran, Vilayanur. 2003. The Emerging Mind. Reith lectures. www.bbc.co.uk/radio4/reith2003.
Ryle, G. 1949. The Concept of Mind: 195-198. Barnes & Noble Books. New York.
Sarte, J.P. The Humanism of Existentialism, in Wade Baskin (ed.) in Sartre: Essay in Existentialsim: 42-43. Citadel Press, 1965. Sexaucus, NJ.
Taylor, C. 1976 Responsibility for Self: 281-299, in A. E. Rorty, ed. The Identities of Persons. University of California Press.
Smart, J. J. C. 1961. Free Will, Praise and Blame. Mind 70: 291-306.
Strawson, P.F. 1963. Freedom and Resentment. Proceedings of the British Academy, Volume 48: 1-25. Oxford University Press. London.
Van Ingwagen, Peter 1975. The Incompatibility of Free Will and Determinism. Philosophical Studies 27: 185-199.
Van Ingwagen, Peter 1978. Ability and Responsibility. Philosophical Review 87: 201-224.
Watson, G. 1996. Two Faces of Responsibility. Philosophical Topics 24: 227 – 248.
Widerker, David. 1995. Libertarianism and Frankfurt’s Attack on the Principle of Alternate Possibilities. Philosophical Review 104: 247-261.
Wyma, Keith. 1997. Moral Responsibility and Leeway for Action. American Philosophical Quarterly 34: 57-70.
[1] Although this essay is concerned primarily with freedom and free will, I will use both the phrases ‘freedom’, ‘free will’ and ‘moral responsibility’ where appropriate. Although I am loathe to define them, I take freedom to be a prerequisite for moral responsibility. It is possible to be free without being morally responsible, but not vice versa in the relevant sense.
[2] I have adopted these from Fischer (1999).
[3] There is another possibility: due to the fact that the subject is aware of the prediction, the predictor may fail to make any prediction at all. This is in fact what I suppose might happen. However, in order for this discussion to proceed, I will suppose that it does make a prediction and examine the implications
[4] One objection to this account may be the notion of quantum indeterminacy. I do not propose to deal with this here. Insofar as it may be a valid objection to the determinist thesis, I believe that it is adequately covered by the subsequent discussion on indeterminacy.
[5] Here I am adopting the label used by van Inwagen in ‘Incompatibility’ (1975: 196)
[6] Experiment 1 seems to show that the universe (or a limited part of it) is in fact predictable. This element of predictability in our actions poses problems for those who rely on indeterminism to locate our freedom. However, there may be cases where an action may be predictable and yet still be undetermined. As an example, as far as we understand the laws of physics at present there does not appear to be any causal mechanism determining the time at which a particular atom of a radioactive isotope decays – it is not a deterministic process. However, the proportion of a particular isotope that decays in a given time can be predicted to an extremely high degree of accuracy. Furthermore, if we concede the logical possibility of time travel (and it seems permissible under the laws of physics as we presently understand them, even though severely impractical), we can imagine a time traveller who observes the decay of a single atom at a particular time and who then goes back in time and is able to state precisely when that decay will occur even though it is undetermined – assuming it is (see Horwich).
[7] That is, their belief in determinism and their belief that they are morally responsible beings.
[8] Although I am sure that there are formulations of the laws of physics that allow an element of indeterminism.
[9] I leave open the question of whether these abstractions are basic beliefs too or are secondary beliefs.
[10] This is similar to Plantinga’s Great Pumpkin problem in which the proponent of basicality of belief in God is presented with the believer in the Great Pumpkin: what is it that makes belief in God appropriate but belief in Great Pumpkin not so?
[11] I adopt the formulation given by Strawson, G. (1994)
[12] I have adapted this argument from Hofstadter (1979)
[13] This is a simplification. The complete description of the process is much more complex and not necessary for the purposes of this essay.
