Philosophy seems to be on a hiding to nothing. It has a 2,500-year history in the West and an extensive back-catalogue – of problems. There are questions about what exists, and what we know about it, such as: Do we have free will? Is there an external world? Does God exist? and so on. There are also questions of analysis and definition such as: What makes a sentence true? What makes an act just? What is causation? What is a person? This is a tiny sample. For almost any abstract notion, some philosopher has wondered what it really is.
Yet, despite this wealth of questions and the centuries spent tackling them, philosophers haven’t successfully provided any answers. They’ve tried long and hard but nothing they’ve said towards answering those questions has quite made the grade. Other philosophers haven’t been slow to pick holes in their attempted answers and expose flaws or dubious assumptions in them. The punctures in the attempted answers then get patched up and put up for discussion again. But what happens is that new punctures appear, or the patches fail and the old punctures are revealed again. Philosophy emerges as a series of arguments without end, and its questions settle into seemingly intractable problems.
Here is a little gem from the 18th century. It’s known as Molyneux’s problem in honour of the Irish scientist and politician William Molyneux (1656-98) who posed a question that has stumped philosophers ever since. Imagine someone completely blind from birth who’s been able to explore both a cube and a globe by touch. This person learns to identify and name these shapes. Now, suppose that this person is later able to see. Would they then be able to identify which is the cube and which is the globe, just by sight? Imagine them standing at a distance from the shapes. Would they be able to tell simply from looking which was the globe and which was the cube?
Here is a companion thought experiment, now called the knowledge argument. By reading the appropriate books, you could learn all about the chemistry of ammonia. By reading more books, you could learn all about how the human olfactory system works and, in particular, how it reacts in response to ammonia molecules – what distinctive changes occur in the mucous membrane and in the olfactory nerves. Given all this textbook information, though, could you then know all there is to know about the smell of ammonia? Or is there something about the smell of ammonia, the qualitative experience of that sharp, pungent aroma, that you won’t understand from this learning, independent of experience?
These thought experiments and others like them generate debates that run and run. It’s not simply that there are different sides to take on any one of these puzzles. It’s that a strong opening case can be made and sustained for each of these viewpoints, despite the fact that these viewpoints conflict. Take the second thought experiment. It seems that knowledge of the aroma of ammonia – what it actually smells like – is not the kind of information you can get from reading books. But then are there facts about human experience that can’t be captured by science and what it can report in its textbooks? Is there more to us than is scientifically describable? If so, it implies that humans aren’t purely physical systems – a remarkable exception to what the natural sciences otherwise tell us about the world.
Is the thought experiment illuminating about human nature or has it gone badly wrong? The jury remains out on this, and indeed on every other, problem in philosophy. Philosophy displays increasing ingenuity without an emerging consensus. Progress doesn’t require consensus, of course: some philosophers might have solved a given problem without this being acknowledged across the board. But the degree to which there is, or isn’t, consensus in a given field can be one indicator of how much progress has been achieved.
The contrast between science, which had a much later start date, and philosophy is striking. Philosophers can’t even agree about what they’ve achieved, other than remorseless argument and debate. Within the natural sciences, however, there’s widespread consensus and significant progress. Many scientific problems succumb to experimentation and hypothesis testing, whereas philosophy appears to be constantly faltering.
This contrast with science might prompt two rejoinders that each query this reassuring picture of science building up knowledge brick by brick. In The Logic of Scientific Discovery (1959), Karl Popper maintained that science is open to experimental disproof, falsification by experiment and observation. But, Popper continued, this consists in showing that proposed scientific theories are false, falsified by experience, and it never shows that any scientific theory is true or even probably true. We might extend Popper’s account of science to philosophy. Perhaps philosophy can be seen as following the same method of bold conjectures that, although they’re never confirmed, can be refuted by evidence.
Popper’s view that observation has only a negative, falsifying role in science has the starkly sceptical consequence that there’s no observational evidence for any scientific theory. But let’s set aside whether he has correctly described scientific method, and consider this extension of his account to philosophy. For the most part, philosophical theories don’t make predictions about what we observe. So those theories can’t be refuted by the discovery that they make false predictions about what we observe. When George Berkeley in 1710 set out his idealism, according to which physical objects are collections of ideas either in our minds or in God’s mind, he wasn’t making predictions about any particular observations we might have – predictions contrary to ones that rival philosophical theories make. According to Berkeley, it’s simply that whatever we observe are ideas in the mind. Observe all you like, and you won’t refute Berkeley. So we need to consider what the counterpart to observation would be if we were to extend Popper’s account of scientific method to philosophy. And that’s just where the problems start.
Any interesting philosophical view makes claims that aren’t obvious – otherwise there’d be little point in making them. The claims then need to be argued for, and that’s why the philosopher’s stock-in-trade is argument. Now, an argument has to have premises; that’s what the conclusion of an argument supposedly follows from. The premises provide reason to believe the conclusion. This raises two questions: what provides the premises of a philosophical argument? And why accept those premises?
The intuitions of philosophers seem a poor counterpart to the observations of scientists
One answer to the first question that appeals to many philosophers is to say that the premises of their arguments are supplied by their ‘intuitions’ – by what they were inclined to think after being acquainted with a philosophical problem. After you’ve heard about Molyneux’s problem and (we’ll suppose) it strikes you that the person couldn’t identify each shape, that’s an intuition of yours. If it just strikes you that there has to be something irrational about people’s committing wrongdoing, that’s another intuition. As opinionated people, philosophers have lots of intuitions of their own. One difficulty, though, is that different philosophers have mutually contradictory intuitions, so not all their intuitions can be correct. Another difficulty is that, even where a majority of philosophers find that they share the same intuitions, the intuitions of non-philosophers from non-Western cultures apparently often diverge from these. And, lastly, even if we all shared our intuitions, so what? Unless we know what the sources of our intuitions are – and we don’t – we wouldn’t know what we would be relying on by appealing to our intuitions. We could all be wrong. The intuitions of philosophers seem a poor counterpart to the observations of scientists.
I said that there are two rejoinders that could be made to the way in which philosophy contrasts unfavourably with science. The second rejoinder draws on the very different work of another philosopher of science, Thomas Kuhn. In The Structure of Scientific Revolutions (1962), Kuhn rejected the popular picture of science as having a steadily progressive history in which the successive contributions of generations of scientists smoothly improve on one another, building up scientific knowledge in steady increments. Kuhn thought that such a view naively accepted a self-serving history written by the victors, where the victors are the scientists of whatever the reigning research programme happens to be. In place of this view, Kuhn defended a historical account whereby there’s no continuity in ideas between the research programmes on opposing sides of a scientific revolution. There’s no common currency of ideas, he claimed, between Aristotle and Galileo, or between Isaac Newton and Albert Einstein. The different parties talk past one another. Accordingly, where scientific progress occurs, it’s localised to the span of a given research programme. The programme is born, becomes dominant in its field for a time, then subsequently passes away.
Kuhn’s views are disconcerting and controversial. For the purposes of contrasting science with philosophy, it’s enough to make a restricted response. There is much that has changed over the decades in the scientific understanding of things. At the more speculative end of science, the cutting edge of its research, no doubt more will change in the future. But much in the scientific understanding of things hasn’t changed. Many scientific problems have been solved, the scientific community remains confident in the solutions, and there’s little prospect that the solutions will need to be revised or abandoned. Witness the raft of such reliable empirical laws as the inverse square laws governing light and sound, the Coulomb laws of magnetic and electric interaction, and Ohm, Ampère and Faraday’s laws of electricity. These physical principles are well confirmed and stable; can the same be said about anything in philosophy? Not really. It’s even up for debate what the laws of logic are.
Having got some sense of the state of play in philosophy, we can turn to the task of diagnosis. What’s gone wrong? Why do philosophical problems resist solution? I will consider five answers, the last being my own.
The first answer challenges the pessimism. The good news, it says, is that some philosophical problems have been solved. For example, Noam Chomsky claims that the mind-body problem was solved centuries ago. When René Descartes posed the problem, he took ‘body’ to be a substance that’s extended in space. Moreover, bodies can affect other things, or be affected by them, only by contact. ‘Mind’, by contrast, is a substance that’s conscious but lacks extension. Since minds can’t literally come into contact with bodies, they can’t interact. The problem then arises of how minds and bodies can interact. But, by positing a force of gravitational attraction, Newton allowed that things can affect one another without contact. The mind-body problem dissolves because there’s nothing answering to one side of the distinction: there’s no such thing as body.
If successful, this would be not only an example of a solved philosophical problem, but the solution would have been provided by science. Still, I’m unconvinced. As I see it, Newton exposed a deficiency in Descartes’s thinking about what body is. That’s to say, Descartes had a false theory of body. So there’s no such thing as body as Descartes construed it. But that’s not to say that there’s no such thing as body. There have been false theories of stars and of human beings, but that’s not to say that there are no such things as stars or humans. All that follows is that there are no stars or human beings as those false theories construed them. And there is such a thing as body, the physical, as typified by such things as planets and our heads. There remains Descartes’s problem about how minds with their remarkable properties are related to bodies and their apparently quite different properties. The persistence of this problem illustrates the more general fact that it isn’t easy to find clear examples where a philosophical problem has been solved.
The second diagnosis is dismissive: philosophical problems aren’t genuine problems. Philosophy’s failure to answer its own questions exposes it as vapid, a sham. Philosophers invent a bunch of artificial problems, parlour games, and then just kick them around among themselves.
To my mind, however, it’s exactly this diagnosis that’s vapid and glib. One point is that it does nothing to explain why philosophical problems are resistant to solution. If they were simple word games – games trading on puns or other forms of wordplay – it shouldn’t require much time or effort to solve and dispose of them. They should be polished off as easily as the morning crossword puzzle. The reality is that problems in philosophy are nothing like that. They resist cheap, easy answers. A second point is that this dismissive diagnosis, this incipient anti-intellectual response, seems especially wrongheaded when we think about philosophical problems that concern some of the things that most matter to us. These are issues about how we live our lives and how we are to live with others – issues about morality and politics.
It’s hard to think of a problem with more consequence than how to live our lives
Our lives are regulated by, among other things, moral codes, codes prescribing what’s off-limits (what’s morally wrong) and what isn’t (what’s morally permissible). Just what is a moral code though? What is the source of morality? Is it our emotions or our reason or something else again? And there are further questions: why should anyone be moral? What’s in it for them? Plato gave these questions close attention. He took the view that a wrongdoer is someone who makes a cognitive mistake by not thinking things through clearly enough. Plato thought that, if only we had a clear idea of what moral goodness is, if only we could know it for what it is, we’d be bound to avoid wrongdoing. To know the good is to love it.
Other philosophers disagreed and found no route from reason to morality. David Hume thought that only emotion, not reason, could provide direction to our lives. There’s nothing contrary to reason, Hume provocatively said in his Treatise of Human Nature (1739), to care more about scratching your finger than the fate of humanity. Something we should take from this debate between Plato and Hume is that it’s not at all like a parlour game on which nothing of consequence hangs. In fact, it’s hard to think of a problem that could have more consequence than one about how we’re to live our lives. Dismissing this debate as empty wordplay would be a cop-out, an evasion of an especially difficult intellectual problem. It is, moreover, far from being an isolated example. Debates about the reality of moral responsibility, the rationale for punishment or the moral status of animals raise other intellectually and morally pressing issues.
A third diagnosis says that philosophical problems are just much harder than science problems – that’s why no one has solved them yet. But the claim that philosophical problems are hard would be a poor explanation of why none of these problems have been solved. The degree to which a problem is hard just means the degree to which it resists solution. I don’t see by what other measure every philosophical problem should be rated as harder than any scientific one.
The fourth diagnosis takes up where the third leaves off. According to it, philosophical problems are genuine, but it’s a serious understatement to say that they’re hard. The problem is that we’re cognitively not up to solving them. The hardwiring in our brains makes us good at some things – like learning a language or judging where a tennis serve will land – at the expense of others. Solving philosophical problems is among these other things. We’re just not cut out to be good philosophers – not any of us.
This is an interesting piece of speculation. Just what is cognitively closed to humans seems an open empirical issue. But the diagnosis is awkward in claiming that solving philosophical problems is cognitively closed to us while allowing that everything else we do in philosophy – understanding the problems, offering hypotheses about them, criticising or refining those hypotheses – is cognitively open to us. That seems a curiously uniform and neat split.
The fifth diagnosis, the one I think explains the most, doesn’t single out any one factor to explain philosophy’s lack of progress. Instead, it takes this to be the interaction effect of a cluster of things. As we saw in the case of intuitions, there’s controversy not only about the theories that philosophers devise but also about many of the methods or kinds of data that they appeal to in support of their theories. Also, philosophical problems have ‘entangled’ natures: proposed solutions to one problem require contentious assumptions about other live problems. For example, there’s a problem in saying what morality is about – what it is for actions or people to be morally good or bad. But this problem is not compartmentalised. Accompanying this problem about the nature of morality, there’s a problem about why we should accept some moral views rather than others. And, as we’ve seen, there’s also a problem about why anyone should care about morality. So, we have a nest of problems here: a definitional problem (what is morality?), an epistemological problem (how can we tell what’s moral?), and a motivational problem (why does morality matter?). Solutions to these problems will make assumptions about reality and our minds that raise fresh problems of their own, and so the issues ramify.
If that’s the diagnosis of what impedes philosophical progress, what’s the remedy? How might we do philosophy better? It goes without saying that we should try harder, but that doesn’t tell us which methods to rely on and which to renounce. Perhaps advances in artificial intelligence could help. As the saying goes, predictions are difficult, especially about the future. What would be wanted is software that executes patterns of reasoning. The difficulty of formalising some of these patterns raises difficulties in programming. Moreover, the reasons being assessed would need to be assigned weights in various ways, and that would shift epistemic responsibility back to human programmers. In a related point, a greater employment of formal methods has enhanced rigour and precision in philosophy. Decision and game theory, for instance, have sharpened up thinking in areas of moral philosophy concerned with rationality and the making of contracts.
Input from the sciences might not settle philosophical problems but it’s a valuable reservoir on which to draw. Empirical psychological research (involving, for example, cataract surgery) has supplemented armchair thinking about Molyneux’s problem. Adopting scientists’ practice of working collaboratively in research teams might also benefit philosophers. The individualistic and contrarian streak of many philosophers, however, might generate in-house disagreements of an all-too-familiar character.
We have, then, something to add to our stock of philosophical problems: why is philosophy so difficult and how can we get reliable results in it? Reflecting on this gives us all the more reason to form our philosophical views tentatively and provisionally. And, I might tentatively add, that’s not such a bad approach to forming views about anything.
