An Enquiry Concerning Human Understanding

Overview

A hundred and twenty pages in twelve sections, written by a man who had published a much longer and more difficult book at twenty-six and watched it, in his own phrase, fall dead from the press. The Treatise of Human Nature of 1739 to 1740 is the great work; the Enquiry of 1748 is the same philosophy recast for people who read essays. Hume later asked that the Treatise be disregarded and that this book be taken as his considered statement. Scholars do not obey him, but for a first reading he was right: the arguments are here, the prose is better, and two sections on religion are here that the Treatise does not contain.

The style is the opposite of German philosophy. Hume writes short paragraphs, uses homely examples, billiard balls, bread, the sun rising, a man who has seen water only in its liquid state, and never invents a term when an ordinary word will do. This is disarming and slightly dangerous, because the arguments are far more radical than the manner suggests and it is easy to nod along through a demolition.

He is also a comic writer. Section 11 puts the attack on natural religion into the mouth of an unnamed friend whom Hume then pretends to argue with, and Section 10 contains a long straight-faced account of the miracles claimed at the tomb of a Jansenist in Paris.

The argument

Section 1 defends doing philosophy at all, and distinguishes the easy kind that paints virtue attractively from the difficult kind that maps the mind. Hume wants the second done with the care Newton brought to the heavens.

Sections 2 and 3 give the materials. Every idea derives from a prior impression, either of sense or of internal feeling, and this becomes a test: when a philosophical term seems to mean nothing, ask which impression it came from. Ideas combine by resemblance, contiguity and cause and effect.

Section 4 is the centre of the book. All enquiry divides into relations of ideas, the province of mathematics, where denial produces contradiction, and matters of fact, where the contrary of anything is always conceivable. Our knowledge of matters of fact beyond present experience comes entirely through cause and effect. But how do we know the causes? Not by reasoning from the nature of the objects, since no inspection of bread reveals that it nourishes. Only by experience. And experience tells us what has happened, so to get from that to what will happen we need the premise that the future resembles the past. That premise cannot be demonstrated, because its falsehood is conceivable, and it cannot be established by experience without assuming itself. The circle does not close.

Section 5 offers what Hume calls a sceptical solution. The inference is made not by reason but by custom. Repetition produces a habit of transition, and belief is the extra force and liveliness that habit lends to the resulting idea. Section 6 adds probability, and Section 7 asks where the idea of necessary connection comes from, since no single observation ever shows a power. The answer is that it comes from the felt determination of our own minds after repeated cases, and Hume gives two definitions of cause, one in terms of constant conjunction in the objects and one in terms of the transition in the observer.

Section 8 applies this to human action. If causation is just regular succession, then the dispute between liberty and necessity is verbal: actions are as regular as anything else, and freedom means acting from one’s own will rather than under compulsion, not acting from no cause.

Section 9 notes that animals learn by experience exactly as we do, which supports the account, since nobody credits a dog with a hidden argument.

Sections 10 and 11 are the applications to religion. Testimony is itself something we assess from experience, and since a miracle is by definition a violation of the most uniform experience we have, the testimony supporting it must be weighed against the whole weight of that experience. Hume then argues that the actual testimony is always weak in specific ways. Section 11 argues that from a world of mixed good and evil one cannot infer a designer with attributes beyond what the world displays, and so can draw no consequences for conduct.

Section 12 sorts scepticism into a useless total variety and a useful mitigated one: modest claims, close attention to evidence, and abandonment of questions the mind is not built to answer.

How to read it

Read Sections 2, 4, 5, 7 and 8 first, in that order, and read Section 4 twice. Everything Hume is famous for is in Section 4 part 2, and it takes about fifteen minutes.

Then read Section 10 on its own terms, as a piece of reasoning about evidence. It has been attacked for two and a half centuries and the attacks are worth more once you have the argument straight.

Do not translate Hume’s conclusion as the claim that induction is irrational or that science is unjustified. He says the inference has no foundation in reasoning, and that we make it anyway, must make it, and are right to. His scepticism is about the credentials of a practice he takes to be unavoidable.

Watch how much the argument depends on the claim that every idea traces back to an impression. That principle is doing more work than its one paragraph of defence supports, and it is where most later philosophers push back.

Questions it raises

• If reason cannot justify inductive inference, is calling it custom an explanation or a change of subject?
• Hume gives two definitions of cause, one about the world and one about the mind. Are they definitions of the same thing?
• Does the argument against miracles rule them out in principle, or only make them unprovable by testimony?
• Is Hume’s compatibilism about freedom a solution, or does it concede the thing people wanted when they asked the question?
• The copy principle holds that every idea comes from an impression. Which of our concepts survive that test?
• If we should proportion belief to evidence, what proportion does the copy principle itself deserve?