Tuesday, December 28, 2010

Scepticism about our knowledge of causation

David Hume’s sceptical argument about our knowledge of causation is the most important piece of reasoning in modern philosophy. Modern philosophy may be said to have started with Descartes trying to find a certain basis for reasoning – his cogito ergo sum – but his method turned out to be fruitless. It is Hume’s sceptical empiricism from which modern philosophy descends, with differing lines of philosophy deriving from different responses to it.

Accepting Hume’s sceptical conclusion about inference from experience led Kant to create his philosophical system of phenomena (what we humans perceive and can reason about) and noumena (things-in-themselves, not accessible to human knowledge or reason). From Kant comes Hegel and (eventually) post-modernism. Cultural anthropologists talking of “other ways of knowing” and science as a “patriarchal Western discourse” are ultimately, whether they know it or not, disciples of a dead white Scottish male (via various dead white German, Austrian and French males).

Scepticism about our knowledge of causation is not good for the status of science. This was demonstrated centuries before Hume was born. Compare Hume’s claim, first published in 1737:
When we look about us towards external objects, and consider the operation of causes, we are never able, in a single instance, to discover any power or necessary connexion; any quality, which binds the effect to the cause, and renders the one an infallible consequence of the other. There is required a medium, which may enable the mind to draw such an inference, if indeed it be drawn by reasoning and argument. What that medium is, I must confess, passes my comprehension; and it is incumbent on those to produce it, who assert that it really exists, and is the origin of all our conclusions concerning matter of fact. This question I propose as much for the sake of information, as with an intention of raising difficulties. I cannot find, I cannot imagine any such reasoning. But I keep my mind still open to instruction, if any one will vouchsafe to bestow it upon me.
With this passage from Muslim thinker al-Ghazali, in his C11th The Incoherence of the Philosophers:
our opponent claims that the agent of the burning is the fire exclusively; this is a natural, not a voluntary agent, and cannot abstain from what is in its nature when it is brought into contact with a receptive substratum. This we deny, saying: The agent of the burning is God, through His creating the black in the cotton and the disconnexion of its parts, and it is God who made the cotton burn and made it ashes either through the intermediation of angels or without intermediation. For fire is a dead body which has no action, and what is the proof that it is the agent? Indeed, the philosophers have no other proof than the observation of the occurrence of the burning, when there is contact with fire, but observation proves only a simultaneity, not a causation, and, in reality, there is no other cause but God.
Both thinkers are deeply skeptical about the capacity of human cognition to perceive the inner workings of reality. Of the two, al-Ghazali at least has some way of explaining causation – it is the will of God. In Hume’s case, causation is just a mystery we cannot ultimately know anything of. Hence Kant’s phenomena and noumena, the former being the realm of human reasoning (and theory), the latter being the unknowable reality of things. But a reality that we cannot truly know is no restraint on human reasoning. Which leads us to human reasoning as a playground of human theories and thus to the triumph of Theory: to all the pathologies of post-modernism with science just one “discourse” since all human discourses are equally impotent to discern reality and so none is privileged over any other.

In al-Ghazali’s case, it leads the only path to truth about the universe being knowledge of the will of God, via the revelations of the Prophet. The triumph of his view in Islam also undermined any hope for science in that civilisation, as any structures or patterns we perceive are just manifestations of the will of God, while encouraging a deeply fatalistic approach to human agency (as outlined in the work of Danish psychologist Nicolai Sennels) since we are not Aristotelian agents acting in an ordered universe but creatures of God utterly in His power. The further consequence of this is being deeply inclined to the conspiratorial mindset, since reality is, in a sense, one Great Conspiracy operating according to the will of the ultimate Hidden Power. If that is your ultimate and pervasive model of how the universe works, and you do not believe in the authoritative efficacy of ordinary (and individual) human action, of course it is going to be natural to explain the world in terms of hidden conspiracies.

But belief in powerful hidden agencies, the impotence of individual action and a tendency to postulation of conspiracies we can equally see in all the lines of thought which descend from Kant and (especially) Hegel. Marx may have been influenced by British empiricism, but he is ultimately too much of a Hegelian, and his disciples even more so, even more in the thrall of theory (and, eventually, Theory). Moving from Marxism (or post-modernism) to Islamism is much less of a shift in one’s mode of thinking that one might expect.

Both al-Ghazali and Hume were arguing against Aristotelianism, though al-Ghazali was doing so more directly. (Even so, there are some medieval precursors to Hume’s thought.) The difference, apart from al-Ghazali having God as his ultimate backstop, is that al-Ghazali’s skepticism about human epistemic capacity was operating a civilisation which had not undergone the Scientific Revolution – indeed, his intellectual triumph within Islam probably did much to forestall any chance of Islam experiencing such a sustained breakthrough – while Hume, Kant, Hegel and co were operating in a civilisation which not only had made the various breakthroughs which we know call the Scientific Revolution but which was, in many ways, increasingly defined (or, at least, profoundly transformed) by those breakthroughs in theoretical and practical knowledge and capacity.
Still, even if Hume’s scepticism about our knowledge of causation did not the civilisation-defining effects that al-Ghazali’s did, it has still been profoundly influential. And not in a good way.

True complexities
We can also see Hume’s epistemic scepticism as a manifestation of the retreat from Plato leading to the retreat of confidence in the notion of truth. To some extent at least, the notion of truth is inherent in language, since if words do not have specific meanings, we cannot have language. If we say ‘dog’ means dog but not cat, we are committed to some notion of truth. (I.e. that it is true that ‘dog’ means dog and false that ‘dog’ means cat.)

There is also little point to language unless it has some connection to reality. Why bother having language unless it is to communicate about the world around us? (Ask any parent: both babies and parents find the lack of language skills of small children frustrating.) There are a lot of situations where accuracy about the world is important: hence lying is broadly disapproved of and a reputation for dishonesty is a handicap. (Though whether it shows respect or disrespect for others makes a difference: lies to spare people’s feelings are often regarded as acceptable.) Perjury is a necessary crime for the functioning of law.

It is regarding general explanatory claims about the universe where we start to run into problems. With Plato, where ideas are more “real” than the reality that is a flawed reflection of them, truth is not a particularly problematic notion. The most “real” things are ideas and so the human grasp of such ideas is the deepest form of truth. Truth is almost equally unproblematic in Aristotle: even if Aristotelian Forms are immanent in reality, they are directly apprehensible by the human mind. Plenty of scope for argument about what is true, but the notion of truth is not problematic.

If, however, we can only get knowledge about the world from experience (empiricism), and we cannot make general inferences from experience (Humean scepticism), then the notion of truth – particularly general, explanatory truth about the world – becomes much more problematic.

That all observed swans are white makes it likely that the next observed swan will be white but, as the discovery of black swans in Australia demonstrated, certainly does not prove the next observed swan will be white. (In fact, of course, humans have been observing black swans for thousands of years. It is just that no European knew about this before the European discovery of Australia.) Hence confidence in the whiteness of swans suddenly turned out to be misplaced. Humean scepticism takes that lack of certainty and turns it into a disabling barrier.

One that looks particularly plausible when confronted with the collapse of confidence in previously taken-for-granted conceptions of the universe. Newtonian physics looked pretty good, until Relativity came along. We thought Newtonian physics was true: it turned out to be merely approximately true. (In effect, true except for very, very small and very, very large phenomena.) But the propositions of Newtonian physics do not much look like E=mc2. So how can they be (approximately) true and Relativity be true? Leaving aside this somewhat ambiguous example, the history of science is full of the debris of now rejected, but previously believed, theories.

It is hardly likely to be some coincidence that Hume’s scepticism, which arose in the wake of the collapse of Aristotelian physics, acquired a new wave of admirers (most famously Karl Popper) in the wake of the fall of the Newtonian empire in physics.

The concept of truth looks seriously shaky in such circumstances.

Not helped by the fact that Platonic and Aristotelian Forms both encourage a view of truth as a binary ‘yes/no’ matter. You either apprehend the Form or you don’t. The notion of partially true, or approximately true, does not recommend itself in such circumstances. If, however, one accepts the complex reality where nature abhors a category, approximately true, or partially true, looks just fine. But that also requires a more complex understanding of how propositions connect to reality, so what one means by ‘truth’.

If medieval philosophy was the period of metaphysics par excellence, modern philosophy is the period where epistemology is a central concern. But just as medieval metaphysics was abandoned when its claims about what we know turned out to be way overblown, so modern philosophy has failed to solve the problems of epistemology it attempted to wrestle after the failure of medieval metaphysics.

The detritus of post-modernism is, as Stephen Hicks demonstrated, a massive indictment of the failure of epistemology. If we can see that Hume’s (and al-Ghazali’s) scepticism about our knowledge of causation does not lead to good intellectual places – and does not remotely accord with our every day experience and confidence about our causal efficacy – then it is reasonable to ask what is wrong with their arguments. Something I will discuss in a later post.


  1. The exact reason that Hume couldn't understand causality is that he was trying to derive it from sensualism. Specifically, he was trying to derive it *directly* from perceptions apart from abstract universals, which did not fit into his epistemology.
    He was right that there is no way to discover a single instance in which causality can be directly perceived. If you throw a rock at a window, there is no little flag that comes up before the rock strikes which says, "The window is about to break," you merely perceive the rock next to the window and then the window breaking. In order to draw a necessary connection between the events, one must advance to the conceptual level, which is something that Hume opposed on principle.

    1. But is the connection usefully regarded as "necessary", which has a logical implication one probably does not want to embrace.