Do animals have free will? Probably, the answer to that question would be agreed by most people to be a fairly obvious “no.” The concept of free will is traditionally bound up with such things as our capacity to choose our own values, the sorts of lives we want to lead, the sorts of people we want to be, etc. and it seems obvious that no non-human animal lives the kind of life which could make sense of the attribution to it of such powers as these. But in thinking about free will, it is essential, nevertheless, to consider the capacities of animals. Even if animals cannot be said to have full-blown free will, animal powers of various sorts provide a kind of essential underpinning for free will which philosophers who focus too exclusively on the human phenomenon are forever in danger of ignoring. And these simpler capacities are interesting enough to raise many philosophical issues all by themselves; indeed, I would argue that they raise the most discussed problem in this area of philosophy all by themselves. For they are, in my view, hard to accommodate within certain conceptions of the universe in which we live – what might be called mechanistic or deterministic conceptions of that universe. This makes it very useful and important to think about the simpler capacities from a philosophical perspective. Instead of asking, as philosophers constantly do, whether free will is compatible with determinism, we should first ask ourselves whether even the simpler powers which constitute what I call animal agency are consistent with it. And it might conceivably be that the answer to this question is “no,” which would shed a new and interesting kind of light on the free will debate.
What do I mean by “animal agency”? We can make a start on homing in on the concept by saying that an agent is something which can act. This, though, is not yet to say anything very clear, because it simply raises the question what it is to be able to act. We sometimes use the concept of action in a very general way – we can, for example, speak of the action of inanimate things such as the wind or the waves, or of the action of a chemical agent such as oxygen or water on a material such as iron when it rusts. So it might look as though almost everything should count as an agent, because almost everything can act – almost everything, that is, can impact causally in some way on something else, given the right conditions. But on the concept of action I am interested in, neither the wind, nor the waves, nor oxygen, nor water, would be allowed to count as having the power to act. Why not?
Well, if we think about what goes on when iron rusts, for example, we might say that oxygen and water can act on the iron. But we could surely equally well say that it’s the iron acting on the oxygen and the water. It’s not altogether easy to see why we should say either of these things in preference to the other when one reflects properly on what’s occurring. There is simply contact between different elements and compounds, and something happens as a result. But none of the reagents involved seems truly more active than any of the others. The distinction between activity and passivity is hard to maintain where we are considering truly inanimate agents – and my own view is that this is because the application of talk of agency in the inanimate realm is an extension of a concept which derives from, and belongs primarily to, our conceptualisation of the animate world. We have borrowed the basic idea of action which has its primary application to animals, for use in describing the more dynamic parts of inanimate nature. But the inevitability with which these interactions in inanimate nature occur robs the concept of action of strict and literal application in the inanimate world. It’s not as though, when oxygen and water cause iron to rust, we think that the oxygen or the water could do other than they do. You put iron and oxygen and water together and rusting is simply going to happen – it’s inevitable. Oxygen and water have the power to rust iron, but it’s not a power they can exercise exactly, precisely because it isn’t a power they can avoid exercising. We could say, if we like, that inanimate entities and stuffs can manifest their powers – but they cannot exert them. This means, in my view, that they cannot be true agents.
The medieval philosopher Duns Scotus distinguished between what he called one-way powers and two-way powers, and this distinction helps with explicating the conception of agency which gives rise to this verdict. Lots of things have one-way powers. Copper sulphate has the power to dissolve in water, my car has the power to move at 70 miles per hour under certain conditions, green plants have the power to photosynthesise in sunlight. But these are all just one-way powers, because given the right conditions, copper sulphate can’t do anything but dissolve in water, my car can’t do anything but accelerate to 70mph, green plants can’t do anything but photosynthesise.
Whereas it seems that in our ordinary thinking, we attribute to some sorts of things in the world a rather different sort of power – the power to do something or not, which is the type of power the medievals called a two-way power. For example, I now have the power to raise my arm. I could do it. But I also have the power not do it. It is up to me to settle whether my arm will go up within the next second or not. Either thing appears to be in my power at this very moment – to act or to refrain from acting. In this respect, then, I seem not to be like a bit of copper sulphate or a car or a green plant. What I do seems not to be entirely determined by the circumstances in which I find myself. What I do seems also to be dependent on me and on whether or not I exercise my power of action. And this gives the distinction between activity and passivity proper purchase. I am active, but the air which I move as I raise my arm is passive – because it is me and not the air that is the true source of the movement, its instigator. This isn’t just a matter of the source of the movement being internal to my body. It is a matter of what happens being up to me – a crucial idea in our conceptual scheme, the idea of an entity which has the power to make things go one way rather than another, even holding the conditions in which it finds itself fixed. And the concept of agency I want to utilise is closely associated with this idea of a two-way power – the power to do something or not. Only those things which possess two-way powers count as agents in my sense.
In a moment, I shall turn to consider an important challenge that might be mounted to the idea that the concept of agency I have tried to describe is a coherent one. But supposing for the moment we accepted that the distinction between one-way and two-way powers was a legitimate and interesting one, an obvious question rears its head. So far, I have put plants and inanimate things on the non-agent side of the divide. I have put human beings on the side of the agents. But what about non-human animals? Where do they go? Do they have two-way powers?
A very familiar way of thinking about animal nature returns the answer “no” to this question. There is a long tradition in the history of philosophy of separating ourselves off from the other animals – of trying to argue that we are nobler or better or more important than they are in some very fundamental respect. It was always part of Christian doctrine that human beings were uniquely close to God, being made, indeed, in his image, and in the seventeenth century, Descartes gave this Christian view a metaphysical underpinning by insisting that human beings were to be distinguished completely from the other animals because only humans were possessed of souls. The movements of mere animals, Descartes insisted in his Passions of the Soul, could all be explained by means of purely mechanistic principles:
“It seems reasonable, since art copies nature, and men can make various automata which move without thought, that nature should produce its own automata much more splendid than the artificial ones. These natural automata are the animals.”
Like others in the seventeenth century, Descartes was very impressed by the technological innovations of his age – particularly clockwork. He seems to have been especially struck by a statue of Neptune which stepped forward and threatened with a waving trident those who approached it as they stepped unwittingly on a button, which he mentions several times in different writings. The thought naturally occurred to him, as it did to many others in the seventeenth century, that animals might just be actuated by a biological sort of clockwork – and hence that they did not require truly psychic powers to explain their functioning. But what is interesting, I think, is that this seventeenth century philosophical view of non-human animals as basically a biological sort of clockwork has lasted in some quarters with surprisingly little alteration well into the twenty-first. The only really major change which has taken place in philosophical orthodoxy since then is that we human beings tend not to be any longer accorded the privileged status that Descartes insisted upon, because very few Western philosophers now accept the existence of an independent immaterial entity called the soul. And as a result, it has been argued by many that Descartes’ basically mechanical conception of animality must be extended also to human beings. We, as well as the other animals, must be conceived of as complex mechanisms, the interactions of whose parts is sufficient to explain the functioning of the whole.
Some of those who embrace the mechanistic picture would therefore argue that the distinction I have been trying to draw between true agents, which have two-way powers, on the one hand, and non-agents, possessors merely of one-way powers, has simply been rendered redundant by the progress of modern science. We now know, it might be said, that our own actions are produced by neural firings which in their turn are produced by further neural firings, or by changes in the concentrations of various brain chemicals, or by similar physiological events. And these changes which go on inside us and which determine what actions we will perform are just the realisations of the one-way powers of things like chemicals and cells.
And so it might be inferred that it must just be an illusion to suppose that humans have any genuine two-way powers – that I possess, for example, as I earlier asserted I did, the power right now to raise my arm and also not to raise it. If I don’t actually raise it, it might be said, that just means the causal conditions necessary for making my arm go up weren’t there in my brain immediately beforehand. Holding prior conditions fixed, these people would argue, it simply isn’t true that either thing could happen. On this view, then, there really is, at bottom,no such distinction as the agent/non-agent distinction I have tried to characterise. Rather, the whole of the universe, including humanity itself, is basically an enormous mechanism, and the agent/non-agent distinction is either unfounded, or must be characterised in some different way (for example, by differentiating the deterministic causes of actions from the deterministic causes of other sorts of events, perhaps by saying that they are “mental” causes, and hoping that this will do the trick. For various reasons I can’t go into here, it won’t).
Now, some other philosophers, of course, have resisted this mechanistic outlook. They are sure it must be wrong because it just seems so obvious to them that we have two-way powers – that I am able, for example, right now, in these precise present conditions, to raise my arm and also not to raise it. It just seems preposterous to them to deny that we are agents in this full and robust sense. They have stood up for what they call libertarian free will. But on the whole, they have been disinclined to suppose that any animals other than human ones might possess this special capacity. Only humans have generally been accorded the status of free willed entities, even by those philosophers who have been inclined to insist that the idea of free will is not just an enormous confusion.
These two camps – let us call them mechanists and free willists, just to have convenient names – now represent two of the main positions philosophers tend to take with respect to the metaphysics of agency. Either agency must be regarded as a special variety of deterministic causal mechanism, or humans are a special exception to the normal workings of the natural world. But surely neither of these positions is at all attractive. The mechanists, on the one hand, seem to present us with a deeply non-intuitive conception of ourselves as creatures completely in the grip of electrical and chemical reactions in our parts, which looks, on the face of it as though it might require extraordinarily radical revisions on our everyday world view. This vision of our place in reality seems to turn us into mere arenas for the playing out of chemical and electrical reactions, all of which are determined, in their turn, by preceding reactions of the same sort. The free willists, on the other hand, seem still to be insisting on the old Cartesian differences between ourselves and absolutely everything else in nature – something which, in a post-Darwinian age, it’s hard to feel comfortable with. What I want to suggest, therefore, is that we need an alternative to both views. And the key to finding an answer lies with the question I raised earlier – namely, what we are to say about where non-human animals fit into the picture.
It becomes much easier to accept the existence of the interesting phenomenon I call agency once we allow that it might be much more widespread than the free willists have tended to suppose, and that it characterises a large number of non-human animals, not just ourselves. It becomes easier because it no longer looks as though the view just amounts to a sort of special pleading for the human species. Rather, one might argue that the development of a robust conception of agency constitutes serious recognition of the surely indubitable specialness of certain modes of functioning that appear to be present throughout much of the animal kingdom.
Like us, I want to suggest, many animals have two-way powers. They are of course driven to do certain of the things they do by instinct (as are we) – but this is not in the least inconsistent with the admission that they have two-way powers. Instincts govern animals, as they govern us, in a general way – setting a specific range of goals, aims and behaviours, but often leaving it up to the animal itself to settle many of the precise details of where, when and how those goals will be pursued. More complex animals have greater freedoms in this respect than less complex ones – a dog, for instance, can decide whether to urinate now, or to run after the stick which has just been thrown for it instead; a spider seems likely to be far more narrowly constrained. But still, even a spider might be able to decide whether here or there is a better place to build its web, to take this or that route to a particular place, etc. There is evidence, indeed, that certain spiders possess quite remarkable abilities – for instance, there is a wonderful account of some observations made in the rainforests of northeastern Australia of the jumping spider, Portia, by Stim Wilcox and Robert Jackson. This spider preys on other spiders, and Wilcox and Jackson describe an occasion on which, having tried and failed to tempt a second spider out of the centre of its web by means of various strategies, Portia disappears from view for about an hour and finally reappears on a rock projection, high above the web of the target prey spider. She then lets herself down from the rock projection, and swings in on a thread to eat her prey. What seems remarkable about this account of Portia’s behaviour is the potential it suggests for forward planning, for spatiotemporal awareness, for the maintenance of intention over a significant amount of time, and in the absence of continuous perception of the desired object. Clearly, Portia’s general goal – to catch and eat spiders – is instinctively given; but equally clearly, I should argue, the choice of means to this end in a given case is settled by Portia, on the spot, in a manner sensitive to the affordances of the environment in which she finds herself.
Agency, then, should be regarded neither as an illusion, nor as the peculiar possession of a single species. Rather, it is a power which characterises biological entities of a certain degree and type of complexity, an evolutionary innovation of quite extraordinary importance, which develops as creatures emerge which have a range of complex and sometimes competing goals. Such creatures may need to utilise perceptually acquired information from a number of sources in order to solve problems, to which there is not necessarily any unique best answer, about how to distribute themselves and their efforts through space and time, in pursuit of their ends. Agential powers are nature’s way of solving these problems; and we need to understand them, and the metaphysics they require, before we have any hope of finally resolving the question whether or not we humans could possibly have free will.