1. Strawson's Basic Argument

Strawson first puts forward his argument in Freedom and Belief, (1986, revised edition 2010)Footnote ⁴ but I will examine a more recent formulation of his argument given in his essay ‘The Bounds of Freedom’.Footnote ⁵ What he calls the Basic Argument is stated in several versions. Here is the first and simplest one:

1. 1.1 When you act, you do what you do – in the situation you find yourself – because of the way you are.

2. 1.2 If you do what you do because of the way you are, then in order to be URD [ultimately and fully responsible and so deserving of praise or blame] for what you do you must be URD for the way you are.

But

1. 1.3 You cannot be URD for the way you are.

So

1. 1.4 You cannot be URD for what you do.Footnote ⁶

Strawson says that the argument is a priori and so is independent of the truth or falsity of determinism.Footnote ⁷ There is, he claims, just something deeply incoherent with the notion of an ultimately responsible individual – one who is at least partly but ultimately responsible for the way he or she is.Footnote ⁸ Why can't agents be ultimately responsible for the way they are? Because this would require the possibility that agents be, in some respects, self-originators. They would have to make themselves what they are without any way that they are to help originate the way that they are. There would have to be an impossible act of creatio ex nihilo – an act of pure bootstrapping. It is, therefore, metaphysically impossible to be a self-originator. Strawson quotes Nietzsche approvingly,

In a third formulation of the Basic Argument, Strawson tries to clarify and elucidate his argument. He uses A to refer to a ‘particular action or piece of deliberation that you engage in.’ He uses N to refer to ‘the way that you are mentally when you engage in [A] that leads you to engage in it in the way that you do.’Footnote ¹⁰ In this version premise 1.1 now becomes:

From 3.1 it ‘appears to follow immediately that’

The rest of the argument tries to show that the notion of URD requires either an infinite regress or an impossible act of self-origination. Strawson explains:

As I have said, the argument has proved difficult to refute. After all, it seems to employ such uncontroversial premises. The first premise (3.1) seems especially immune from dispute. Who can deny the seemingly uncontroversial assertion that any action you do, at a certain time, is a result of your overall mental makeup at that time? Indeed this premise is not defended by Strawson. He says that he will take it as ‘obvious’Footnote ¹² or ‘incontrovertible’.Footnote ¹³ He will probably have most readers' sympathy. It just seems so obvious that any state of the universe at any particular time arises from or is a result of the way the universe was just before. As we shall we, however, it will be precisely this premise that the Bergsonian critique will call into question. To see why we will now turn to an examination of Bergson's contribution to the free will debate.

2. Bergson on Free Will

Many of Bergson's most important ideas about free will can be found (unsurprisingly) in his book Time and Free Will.Footnote ¹⁴ In this book he sets out to counter some pervasive presuppositions. One of these is what he sees as the widespread but fundamentally mistaken notion that mental desires and acts of effort have an ‘intensive magnitude’ – that one desire can be stronger than another, that one pain can be greater than another, that the inner feel of a muscular effort can be greater or smaller than another. Bergson's crucial argument is that the notion of magnitude can only make sense when something is spatial. So one box can be greater than another if one box can be contained by the other, but how can one pain – something mental and without extension – be thought about as greater than another? It seems to make no sense to say a less intense pain is contained in a greater pain.Footnote ¹⁵ Extensive magnitude certainly makes sense since we can think of one spatial thing containing or being contained by another, but intensive magnitude, he thinks, cannot be thought about like that. Although this idea is important and extremely interesting, it is not something that I will pursue in this essay. Others have done an excellent job of explicating Bergson on this point.Footnote ¹⁶

My main concern is with Bergson's arguments in chapter three of the book, especially 174–183. He begins by defining the difference between libertarianism and determinism: ‘The argument of [John Stuart Mill and other determinists] implies that there is only one possible act corresponding to given antecedents: the believers in free will assume, on the other hand, that the same series could issue in several different acts, equally possible.’Footnote ¹⁷

Given Bergson's opposition to determinism we would expect his argument to be a defence of the latter position – that given a series of mental oscillations between one course or another, at the point of decision, the self is free to go one way or the other. There is more than one possible outcome. After all, that the self can go more than one way is what libertarians are keen to defend and determinists deny. But for Bergson it is this whole picture – a picture that both determinists and libertarians accept – which must be jettisoned.

In order to show the illegitimacy of this shared understanding of the processes of thought, Bergson imagines himself hesitating between ‘two possible actions X and Y, and [going] in turn from one to the other’.Footnote ¹⁸ He represents this by a diagram symbolizing deliberation, point of decision and possible (forking) paths. It is the kind of diagram familiar to any follower of the modern free will debate. For many it seems an entirely natural way of representing the history of a decision:

Here, then, is a classic picture of the ‘self which after a series M O of conscious states…reaches the point O [and] finds before it two directions O X and O Y equally open’.Footnote ¹⁹ Bergson explains that the libertarian will say that if the self chooses X, then, since it was hesitant and deliberated beforehand, it could have chosen Y instead. The other path was equally open at the time of decision. The determinist will ‘answer: you chose X, therefore you had some reason for doing so, and those that declare that Y was equally possible forget this reason: they leave aside the conditions of the problem.’Footnote ²⁰ Bergson comments that both participants in the debate share a common assumption: ‘both take up their position after the action X has been performed, and represent the process of my voluntary activity by the path M O which branches off at the point O, the lines O X and O Y symbolizing the two directions which abstraction distinguishes within the continuous activity of which X is the goal.’ The determinist notes that the line M O X ‘has been traversed’ and has to take this into account. In contrast, the libertarian concentrates her energy on the point O noting that at that point no decision had been taken. The libertarian is likely to concentrate on that point, and take the self back ‘to the point O to oscillate until further orders’.Footnote ²¹

Bergson thinks that the diagram is seriously misleading. He argues that it misrepresents the reality of what happened. Crucially, it tries to represent as determinate what was, at the time of its occurring, not determinate. It looks at the train of thought and the eventual decision in retrospect after a certain determinate outcome has occurred. Bergson says the diagram ‘does not show me the deed in the doing but the deed already done. Do not ask me if the self, having traversed the path M O and decided in favour of X, could or could not choose Y: I should answer that the question was meaningless, because there is no line M O, no point O, no path O X, no direction O Y. To ask such a question is to admit the possibility of adequately representing time by space and a succession by a simultaneity.’ He continues by saying that the diagram offers us a ‘stereo-typed memory of the whole process of deliberation and the final decision arrived at…’ He asks the question, ‘…how could it give us the least idea of the concrete movement, the dynamic progress by which the deliberation issued in the act?’Footnote ²²

In order to understand Bergson's idea we have to appreciate that one of Bergson's main philosophical contentions is that time is utterly different from space. We have already seen this in his rejection of intensive magnitudes – this is to spatialize something that is inherently non-spatial. Roughly speaking, for Bergson, space is the realm of the quantitative, whereas time is the realm of the qualitative. But one of the main tendencies in science and in the pragmatic demands of everyday life is that we confuse the two and spatialize time. We see it as laid out, for example, on a time-line with events happening in succession as if an earlier event is somehow quasi-spatially adjacent to one immediately afterwards. Bergson certainly recognises that, for pragmatic purposes, we cannot help but represent temporal events and processes by laying them out in time lines and other essentially spatial representations. This is, acknowledges Bergson, acceptable for everyday purposes, but he thinks that it is poor tool for philosophical study since it misrepresents reality.Footnote ²³ It is a pragmatically necessary, but potentially misleading symbolic notation. It is misleading in that we lose much of what a process was actually like. It is to impose determinate spatial adjacency on something which was a (non-spatial) flux and flow of intermingled thoughts and feelings each one, at any moment, evolving and indeterminate. Indeed, even to pick out a thought and treat it as an individual when we try to retrace the movements of our decisions is to impose a certain artificiality since at the time of the thought ‘it’ was surrounded by a complex multitude of other thoughts the boundaries of which were unclear and indistinct. Additionally, the next thought did not yet exist, and it is only when it comes to exist that we can see, in retrospect, something determinate.

We can clarify these difficult thoughts by applying them to Strawson's argument. For Bergson, Strawson's first premise is anything but obvious. Strawson claims that at a time t₁ there is a state N – this is the ‘way you are’. This whole way you are is ‘the way that you are mentally when you engage in [A] that leads you to engage in it in the way that you do.’ ‘After all’, says Strawson, ‘it is your overall mental makeup that leads you to do what you do when you act or deliberate.’Footnote ²⁴ There is, then, a whole way that you are at any particular time t₁ which leads you, at that moment, to behave and think in the way that you do. This might seem obvious, but, according to Bergson, it is completely false. Bergson would claim that if we take the putative whole-way-that-you-are at a time t₁, it is not adequate for understanding. The way that you are at t₁ – whatever that isFootnote ²⁵ – is simply not understandable on its own. It is for him an entirely artificial representation – a snap shot of a moving life. As a still life (which is, of course, an oxymoron) it can be accepted that it captures something of who you are – an aspect or part, but it is an aspect or part of something larger. Considered on its own or as a ‘part’, it is incomplete and indeterminate.

Bergson's contention is firmly holistic – the so-called ‘whole way that you are’ is simply not comprehensible without seeking a wider, more inclusive view. For Bergson, your whole body and all its psychological attributesFootnote ²⁶ at a particular moment is an aspect of something more complete. F.C.T. Moore has an excellent example to illustrate what Bergson is saying. In Bach's Violin Double Concerto there is a sequence of three notes identical to those familiar notes that begin Three Blind Mice. But it is nonsense to take those three notes on their own and say that the opening of Three Blind Mice is a part of Bach's masterpiece. The mistake here is to understand music as merely a succession of notes rather than a melody. Music (and selves) are more holistic in nature than that.Footnote ²⁷ The mistake Strawson makes, as Bergson would see it, is twofold. First, there is the mistake of thinking that slices of a life can be studied on their own, and that they can give us a comprehension of the whole. Secondly, there is the related mistake of thinking that a part can easily be distinguished or picked out from its surroundings as if the phrase ‘at t₁’ really picks out something determinate.

To see all this more clearly, imagine that you entered the time slice of the world at t₁. Bergson would claim that it would be bereft of almost all that you needed in order to understand it. Let us keep the picture of the so-called whole-way-things-are at t₁ simple, and imagine, in this world, you see a football on the ground.Footnote ²⁸ What could be known about this football? You might ask some questions: Does this inanimate lump of matter represent a moving football, and if so where is it going? Is it going up or down or from side to side? Is it expanding or contracting, vibrating or even bursting? Is it spinning, and so at the beginning of a wonderful curve into the distant net? Where has it come from? What is the story or history of this frozen lump of matter in a particular position? But none of these questions seem answerable by concentrating our attention on the state of the world at t₁. It simply does not seem that an inspection of the world at t₁ can supply us with the missing information. Why not? Because, Bergson insists, the real world – which contains lives and feelings as well as footballs – is one that is moving, dynamic and holistic in character. The state of the world at a particular time, in contrast, is static and razor thin, and consequently too much is lost that we need in order to understand it. Again a musical analogy is apposite: listening over and over again to one note in Wagner's Ring Cycle will not enable you to understand the music. That note, at that particular snapshot of time, will not give you an insight into the whole way the music is. Similarly, to stare fixedly at one moment in time reveals very little about what is going on.

Bergson would be ready to acknowledge the practical usefulness of this artifice of freezing the moving into slices of time such as t₁. We might, for example, be investigating the spin of the ball in order to understand why that kind of object curves in flight. And so we might look at a photograph of the ball at one position and another photograph at another position. We will note the differences in location and the gap between the photographs. All the time we do this however, we know that these two discrete frozen moments are of the same spinning ball in flight. Without that prior knowledge, there would be little point in the photographs and the mathematics of our deductions. We would be like the visitor to the inanimate world of our example.

Is there, then, a whole way the world is at a particular time, t₁? For Bergson the answer is a firm No. The world is more melody than mechanical. At every moment in a melody there are notes (or if there is a brief silence we have the echo of the last sound and an anticipation of the next), but that does not license us to analyse the music as merely made up of notes, as if there is a way the music is at t₁. Melody simply does not allow this kind of temporal atomism. The ‘parts’ coalesce and affect each other in complex ways. Snapshot ways of thinking are useful (as in the photograph example) but they miss out so much that is real and relevant. If something as comparatively simple as a football (or music) cannot be properly comprehended by being broken down into atomic, temporal units, why then should we be tempted to think that the psychology of a person can be?

What would Strawson say to this critique? Indeed, what would anyone say who wants to say that there actually is a fully (determinate) whole way the world is at a time t₁? I think that Strawson would reply to this kind of critique by insisting that a proper description of the whole way that one is at a particular time would be richer than a fairly uninformative snapshot. Strawson could say that an adequately enriched description of the way that you are at t₁ would eliminate the mysteries, and provide us with any missing information.Footnote ²⁹ According to this reply, the football in that frozen moment would be described as moving at such and such a rate with a particular vector. This reply would have in mind physics text books where we are used to seeing pictures or diagrams depicting objects moving at certain rates in particular directions. We see arrows representing the direction of the motion of the ball, and written on the page we are informed that it is moving at, say, 30 miles per hour or spinning or heading up a slope. So when Strawson refers to the ‘way that you are – in the situation you find yourself’ he means a snapshot with information added that tells us of the trajectories and movements of an agent's thoughts as well as the physical movements of his body. We will have to say something like this: ‘the person at t₁ is anxious, sweating and twitching as he moves towards the examination hall’ – of course, there will be a whole host of other bits of information in order to give us a full description of the whole way one might be at t₁. These additional pieces of information will be added to the frozen picture to give us a fuller description of the person at t₁.

Is this an adequate reply to Bergson's critique? It seems to me that Bergson would counter by saying that the seemingly innocent addition of information is to go outside the strict atomistic confines of t₁. It is being implicitly acknowledged then that the so-called whole way you are at t₁ has to include much about what is happening either side of that thin time slice. The time slice is being surreptitiously thickened. We need, in other words, that moment's temporal context. But this addition of the temporal context is already to have admitted that the world cannot be broken up into temporal atoms, and that movement is not wholly understandable as a series or collection of still lives. We see the movements and processes first and only then can we break them down, so to speak, into more mathematically amenable moments of infinitesimal smallness.Footnote ³⁰

Bergson is saying then that the world and life is not made up of temporal atoms. Consequently there is no way you are at t₁. Let us now connect this idea to another part of Bergson's critique. Recall that Bergson uses the decision diagram to represent the history of a decision. He says that these kinds of diagram are always looked at after a decision has been taken. It shows the person's future and decision as if completed. We have seen that the determinist insists that, if M O was taken, then, there must have been a reason for this. The libertarian, in contrast, says that at point O the person could have still taken either path. Bergson, however, refuses to accept the presuppositions represented in the diagram – the presupposition that the future can be represented at t₁ as a determinate path or paths. There is, he insists, at point O, no paths. The paths are simply not there yet. Of course, once the path has been taken we have something determinate to contemplate, and think about and try to comprehend, but that is to apply our thoughts from the present into a past where, at the particular point O, there was no future.Footnote ³¹ Bergson comments upon our tendency to impose a determinate structure on the past, when at the time it was evolving Footnote ³² there was indeterminacy:

Bergson entirely rejects the notion that prior to X's realization, there was a possible X – that backstage to reality there is a world of determinate possibles waiting for their chance to stand upon the stage of full blown reality.Footnote ³⁴ For Bergson the future simply does not exist. The present moment, to use Peirce's words, is ‘that Nascent State between the Determinate and the Indeterminate’.Footnote ³⁵ Our mistake, as Bergson sees it, is to look upon a now determinate past and think of an already determinate future towards which it was heading. The libertarian wants there to be many determinate paths towards which the past indefinitely heads, while the determinist thinks that there is only one and that its heading towards it is definite. Both, however, are mistaken – there are no paths. The paths evolve and come to be, but until that coming to be, they do not exist. Every moment is, in Bergson's understanding, to walk on the edge of a kind of precipice and take a step that creates a path rather than taking a step which finds a path which is already there.Footnote ³⁶ The future is not a world of definite possibility. It only becomes definite once it evolves.

Let us now apply this part of Bergson's ideas to Strawson's approach. It is clear that Strawson thinks of the future as a world of definite possibility. Indeed, as Bergson points out, this picture of definite future paths is one that dominates the whole debate. In Bergson's terminology, Strawson sees the world as unfurling, rather than evolving.Footnote ³⁷ Strawson does not think of anything as radically novel, but as the result of a previously given determinate moment. There is a definite, determinate past which can be broken down into temporal atoms and these result in the next determinate temporal atom. Definite slice is added to a slice which is, before it is actual, already definite.

The two parts of Bergson's ideas can be connected. In the first part we saw that he insists that an evolving, dynamic world cannot be broken up into temporal atoms. If we do so, we reduce reality to a still life, and much of what is needed to understand it fully has been lost. As we have seen, we have to import information from the (real) world of dynamism and change if we want to understand the way that things are. The second part of the critique claims that the future is not definite until it occurs. But the import of information, in order to contextualize t_1, adds where it is going at t₁ as if where it is going is, at that moment, a definite fact. The still life or snapshot is certainly determinate if we only look at positions in space, but its movement, its direction or travel, its doings, are not – unless, that is, in retrospect, we add that information. The edge of reality – the present – is not yet determinate. Perhaps this picture will help. The present is always shimmering, waiting to take shape. The shape that it takes on is a definite shape since it stops shimmering, but the new present is a new present of shimmering indefiniteness.Footnote ³⁸ Each present is indeterminate, and the shapes it adopts are, until actual, not existent even as possibles.

This brings us to the third and most perhaps most important aspect of a Bergsonian critique of Strawson's reasoning. As we have seen, Strawson's first premise – one that seems ‘obvious’ or ‘incontrovertible’ to him – goes like this:

Our Bergsonian critique now hones in on the word ‘because’. Note, incidentally, that Strawson's use of the word has to waver indefinitely between two interpretations. He cannot mean the word in a deterministic sense since he is not claiming a kind of incompatibilism – that the deterministic facts oust moral responsibility. Neither is he saying that indeterminism ousts moral responsibility.Footnote ³⁹ Rather the claim is that ultimate moral responsibility is incoherent. Nevertheless, that the world is the result of its immediately previous state in some interpretation of the word ‘because’ seems to many to be obvious. The Strawsonian philosopher will point out that, after all, such thinking works. We know where a physical body will end up if we describe the state of that body at t₁ with sufficient richness. We draw the diagram, label it, do the calculations, and find we can predict the body's path. It is now contended that it is similar with persons. Of course, the analysis and consequent predictions are here much more difficult – they may even have to be probabilistic. We are dealing with systems that are much more complex than the simple physical bodies of physics text books, but the principle is the same. Why does such thinking work? Because the next state of the world, they will insist, is a result of its previous state. The inference follows that the world is not like music at all, and its temporal parts are not akin to musical notes. In music one note does not result in the next one – if it did, and we knew the laws governing music, then, considering one note's exact pitch and exact frequency in Wagner's Ring Cycle might well lead to a comprehension of the whole.

But for Bergson this is completely wrong. To say that the world at a particular time is a result of the way the world was at an immediately previous time is already to have accepted temporal atoms (and the spatialization of time) but also to have rejected the possibility that the next ‘moment’ in time can be a creation of the first ‘moment’. For him the musical analogy is apposite. No one will want to say that each note in a piece of music results in its successor or that because the music is in state N the next note is the consequent.Footnote ⁴⁰ A wholly different way of seeing the music is needed – first we have to understand the whole in order to understand the appropriateness (or inappropriateness) of that particular note at a particular juncture in the music. And we have to see the ‘parts’ not causing each other but as having relations of appropriateness or inappropriateness given the whole.Footnote ⁴¹

So let us apply Bergson's way of thinking to Strawson's apparently metaphysically innocent premise. I quote the ‘incontrovertible’ or ‘obvious’ premise again:

Let us try a Bergsonian inspired re-writing – a rewriting where we excise what Bergson would see as Strawson's metaphysically contentious and imprecise presuppositions:

I think that the first reaction of the Strawsonian would be to accuse this rewriting of an unneeded rather florid and undisciplined colourfulness, a kaleidoscopic confusion that stands in stark contrast to the austere black and white precision of language such as t₁, N, and A. But Bergson asks: is this language so very precise, or does it misrepresent? Is life a series of states, each of which is the consequent of its immediate predecessor as if N₁ results in N₂, or is it something different in kind altogether – indeed, for Bergson is the whole evolution of the world something entirely different from the kind of analysis we find on the pages of physics text books? Ponder for a moment the word ‘ponder’ – a word which carries with it a drawn out process of careful thinking – a thinking with perhaps an element of gentle whimsy as well as concentration. Bergson's contention would be that it is impossible to be in a state of pondering at a snapshot of time, and so the word ‘ponder’ is the more precise word precisely because it is so much more colourful and less austere.