At first glance it might seem like a peculiarly philosophical vice to ask whether or not our scientific theories are true. After all, every time we step onto an airplane, turn on the television, or go to the doctor, we are implicitly endorsing the truth of a myriad of different scientific theories. If these theories were not largely correct, it would be simply miraculous that our airplanes remain in the air, our favourite shows appear on the screen, and that we haven’t all managed to poison ourselves by now. The very fact that you are reading article, produced as it was on a computer, is testimony to our understanding of electronics, and in turn, our general grasp of the behaviour of subatomic particles. And indeed, this central intuition can be developed to drive a very simple argument in favour of our scientific theories – science works, therefore it is (at least approximately) true.
Another way to put the point is to think about how we might try to explain the remarkable success of contemporary scientific practice. Suppose for the sake of argument that our best scientific theories were actually false, and that there were no such things as electrons or other subatomic particles. Suppose that everything we thought we knew about electrical engineering and how to build a computer was just completely and utterly wrong. Nevertheless, we can’t deny the fact that our laptops still manage to behave in precisely the way that we predict that they will. You click the link, and the article loads. You enter your credit card details on Amazon, and 3-5 business days later the complete Jean-Claude van Damme box set arrives at your house. A stopped clock might give us the correct time twice a day, but for a false scientific theory to keep on correctly predicting the behaviour of a complex piece of equipment like a computer or a DVD player is surely a step too far. It would be what the Australian philosopher J J C Smart called a “cosmic coincidence” – a possibility he dismissed with characteristic antipodean bluntness.
The view that our scientific theories are more or less true is called scientific realism, and the principal argument offered in its support is therefore simply that it provides us with the most plausible explanation for the success of science. Do our computers work because there really are electrons whizzing around our circuit boards in more or the less the way we expect them to? Or is it all just one gigantic fluke? When the choice is put that starkly, it seems only rational for us all to be scientific realists. In fact, we can even go one step further. The general principle that the best explanation for a phenomenon is usually the correct explanation for that phenomenon is not only sound rational behaviour – it is also one of the basic principles underlying good scientific practice. To take a simple example, part of the reason why we believe in germs and viruses, rather than evil spirits and unbalanced humours, is precisely because modern medicine provides a better explanation for our ill-health than medieval superstition. The principal argument in favour of scientific realism is therefore not only just sound common-sense; it is also a scientific argument in favour of the truth of our scientific theories. As Hilary Putnam put it in what has become the classic statement of the position:
“the positive argument for [scientific realism] is that it is the only philosophy that does not make the success of science a miracle … these statements are viewed not as necessary truths but as part of the only scientific explanation of the success of science, and hence as part of any adequate description of science and its relations to its objects.”
In the literature, this has become known as the No-Miracles Argument for scientific realism. We should believe that our scientific theories are true, since any other possibility is just too unlikely to take seriously.
Since this is philosophy however, there is no reason to expect common-sense to prevail. One persistent challenge facing any scientific realist is of course the threat of radical scepticism. Our scientific theories may well appear to give us reliable knowledge about the world, and allow us to design and build all sorts of sophisticated technological devices, but we cannot rule out the logical possibility that we are being deceived. Maybe this is all a dream – or to take a more scientific scenario, some elaborate computer program designed to fool us into thinking we are experiencing the real world. Maybe we are all just brains in a vat, or the victim of implanted memories like in the sadly underrated and surprisingly philosophical Universal Soldier: Day of Reckoning. For the radical sceptic, the possibilities are endless, and indeed there is little that the philosophy of science can say to convince him otherwise.
But even if we leave these concerns to one side, there still remains a significant challenge for us to address. Our scientific theories may well be highly impressive with respect to the predictions they make and the devices they allow us to build. But this success has come at a cost. For every successful scientific theory that we possess, many more have been rejected and abandoned. Indeed, the history of science is littered with scientific theories that we now consider to be false. Consider for example the history of astronomy. Aristotle believed that the outer planets of the solar system were affixed to great crystalline spheres that revolved slowly around the Earth. Roughly two thousand years later, Copernicus argued that everything revolved around the Sun, and Newton proposed that they were all held in place by a mysterious and unmediated gravitational force like some kind of cosmic centrifuge. Three hundred years after that, the Newtonian picture was itself consigned to the great dustbin of history. According to Einstein, the planets are constrained by nothing more than the deformation of spacetime, and would wander off into the inky blackness of space if it weren’t for the vast astronomical trough caused by the mass of the Sun.
Of course, no-one is going to deny that science makes mistakes. It is a process of trial and error. Progress is often slow and gradual, but nevertheless we presumably know now far more about the fine details of the universe than we did before, and consequently our contemporary scientific theories must therefore be more accurate and more successful than those of our scientific predecessors. The problem however is that both Aristotelian and Newtonian accounts of the universe were also considered to be highly successful at the time, and offered a considerable improvement over anything that had come before. For a scientific realist writing in the early modern period, Newtonian Mechanics offered without doubt the best explanation for the observed order of the heavens. It would have been inconceivable for him to suppose that things could behave exactly as Newton predicted they would, and yet for the theory to be false – it would have been a “cosmic coincidence” of such proportion that it did not bear serious consideration. In short, any such individual would have been just as convinced of the truth of Newtonian mechanics as we are today of the theory of relativity. They were wrong. But how can we be so sure that own scientific theories will not suffer a similar fate in the future?
In fact, once we delve deeper into the history of science, the problem only seems to get worse. The problem of course is not simply limited to astronomy. Rather, it would seem that every scientific discipline is full of theories that were once considered to be true, but which have subsequently been abandoned as false. Consider for example the phlogiston theory of chemistry, the caloric theory of heat, or the nineteenth century view that light propagated through an all pervasive luminiferous ether. Henri Poincaré summed up the situation when he wrote:
“The man of the world is struck to see how ephemeral scientific theories are. After some years of prosperity, he sees them successively abandoned; he sees ruins accumulated on ruins; he predicts that the theories in vogue today will in a short time succumb in their turn, and he concludes that they are absolutely in vain. This is what he calls the bankruptcy of science.”
In contrast to what the scientific realist would maintain, it would seem that sound common-sense should actually tell against the truth of our scientific theories. The history of science cautions us against an unduly optimistic assessment of our scientific practice. It is a case of once-bitten, twice-shy.
Moreover, such a line of thought can also be shown to be an instance of good scientific practice. The argument from history notes that many if not all of our most successful scientific theories have eventually turned out to be false. It then proceeds to offer a simple extrapolation from this fact to the conclusion that our contemporary scientific theories will also probably turn out to be false. The argument is therefore a straightforward instance of enumerative induction,which is of course exactly the sort of argument that we rely upon in science. We observe how things have happened in the past, and infer they these patterns will continue in the future. Just like the No-Miracles Argument then, the argument from history is therefore also a scientific argument – although in this case it is a scientific argument against the truth of our contemporary scientific theories. For fairly obvious reasons, this argument has come to be known in the literature as the Pessimistic Induction.
Much of the contemporary debate over scientific realism is framed around these two competing considerations. The No-Miracles Argument appears to give us good reasons to believe that our contemporary scientific theories are true, whereas the Pessimistic Induction appears to give us good reasons for a more cautious assessment. There is naturally enough a lot that can be said about either argument. The approximate truth of our scientific theories may well offer a better explanation for their success than a mere cosmic coincidence, but we might still question whether it offers the best explanation for that success. Similarly, we might agree that the history of science is crammed full with failed scientific theories, but we might question whether or not the evidence is representative of our scientific practice as a whole. There might even be ways of reconciling the two arguments, and arguing that we have good reasons to believe that our scientific theories are true in some respects, but perhaps not others.
What is perhaps most interesting about this debate however is the way in which both arguments can be seen as providing instances of the very scientific practice that they seek to evaluate. On the one hand, this is good news. There is always a risk when doing philosophy of losing touch with one’s subject matter. It is often complained about radical scepticism for example that the arguments are too far-fetched, and that the whole issue of whether or not we are the victims of a persistent and utterly compelling illusion can only be of interest to the most disinterested of philosophers. As the readers of this magazine well know, it can often be difficult to convince friends and family why they should care about Descartes’ evil demon, or the fake barn that we might have passed by the side of the road. In the case of the No-Miracles Argument and the Pessimistic Induction however, these are arguments that are directly related to our everyday scientific practice. The hope then is that these are arguments with a far more direct relevance to our daily situation, and may even be of interest to the practicing scientist.
On the other hand though, there is a danger in relying too closely upon our scientific methodology. Think about this in the case of the Pessimistic Induction. If this really is to be understood as a scientific argument for the claim that our contemporary scientific theories are probably false, then it runs the risk of undermining itself. To argue that our contemporary scientific theories are probably false is to argue that our scientific methodology is not in fact very reliable – after all, it has only managed to deliver us scientific theories that we expect to fail. But if our scientific methodology is not in fact very reliable, then we can hardly appeal to those very same scientific methods when offering a philosophical evaluation of our scientific theories. Yet that is precisely what the Pessimistic Induction invites us to do. It is the intellectual equivalent of trying to prove mathematically that mathematics doesn’t work, or me writing here that you can’t believe anything that you read. Even better – it would be like the genetically enhanced clone of a serial killer engaging in a series of bone-crunchingly choreographed martial arts contests in order to prove that violence is never the answer.
And similarly in the case of the No-Miracles Argument. There is of course nothing incoherent in offering a scientific argument for the truth of our scientific theories. Such a line of reasoning does not undermine itself in the same way as the Pessimistic Induction. But it does beg the question. Think about it this way. If we were genuinely unsure as to whether or not we should believe that our scientific methodology was reliable, it would not help us very much to be given an argument that explicitly relies upon precisely the sort of reasoning that is in doubt. If you’re not convinced of the validity of a logical inference, I can hardly give you an argument that is itself an instance of that logical inference. Or to take a less abstract example, imagine an astrologer forecasting that astrology will be a successful method of predicting the future, or trusting a politician just because he told us that he was honest. Thus as many philosophers have complained, the No-Miracles Argument seems to presuppose what it is in fact trying to show.
All of this points to an interesting difficulty lying at the heart of the scientific realism debate. Our scientific methods are not just one of the ways in which we try to understand the world around us. They are our most successful way of understanding the world around us. One way or another, our scientific theories have helped us to survive and prosper in a hostile environment, and therefore enjoy an intellectual prestige that far overshadows anything of which professional philosophy can boast. To offer a purely philosophical evaluation of our scientific methods can therefore seem somewhat presumptuous. Science does not require philosophical legitimacy. In fact, it seems as if the relationship should be running in the opposite direction. Philosophers should look to our best scientific theories as a guide for their own investigations. They should take seriously what modern science has to tell us about cognitive bias, the evolution of the optic nerve or the behaviour of light when formulating their own theories about the nature of knowledge and the justification of belief. As W V O Quine famously argued, our philosophical investigations should be continuous with our scientific investigations – we should practice what he called “naturalized epistemology”.
The contemporary scientific realism debate has evolved into a perfect example of what Quine had in mind. After all, both the No-Miracles Argument and the Pessimistic Induction purport to offer scientific arguments for their respective conclusions. But while a suitably naturalized epistemology may make sense when dealing with the philosophy of perception or analysing Gettier cases, it seems to hit a bit of snag when dealing with the philosophy of science. The problem is that when we ask about the truth of our scientific theories, we are asking about the relationship between our theories and the world they seek to describe – and this relationship is itself one of the many aspects of the empirical world which it is the business of science to investigate. Human beings are after all just another animal running around in the great wide world, and our scientific theories just another kind of tool that we use to help us reproduce and survive. If our scientific theories do accurately represent the world therefore, this will have something to do with human psychology, our evolutionary background, and the causal properties of whatever it is out there in the world that we talk about when constructing our theories. It follows then that any attempt to explain how our scientific theories manage to provide reliable knowledge about the world will just end up appealing to those very same scientific theories in framing its response.
So maybe philosophy cannot tell us anything about science after all. If our philosophical investigations are to be guided and constrained by our best scientific theories, then we cannot hope to show anything of philosophical substance about those scientific theories. We would just be arguing in a circle, as the case of the No-Miracles Argument and the Pessimistic Induction have shown. The scientific realism debate thus illustrates the awkward middle-ground that can exist between our more esoteric philosophical speculations, and our common-sense epistemology. On the one hand, we can entertain ourselves with evil demons and computer simulations and highly convoluted thought-experiments about what we might be said to know in such-and-such a situation. These are paradigmatic philosophical problems that we can really get our teeth into without fear of being undermined by empirical science – but at the risk of alienating the man on the street who fails to see the point. On the other hand, we can attempt to construct ever more concrete philosophical considerations that explicitly appeal to the well-respected methods of the natural sciences – in which case we risk having already presupposed what we were attempting to show.
Maybe what the scientific realism debate really shows us then is not whether or not our contemporary scientific theories are true, but rather something about the limits of our philosophical investigations. And I think that might be even more interesting.