The Problems of Philosophy

Overview

A hundred and seventy small pages in fifteen short chapters, published in 1912 in the Home University Library, a shilling series aimed at the general reader. Russell was forty, had just finished the three volumes of Principia Mathematica with Whitehead, and knocked this out in a matter of weeks. He later called it his shilling shocker.

It has never been out of print and it is still the standard first book. The reason is that Russell does not explain philosophy; he does it, slowly, on a single example, and lets the reader watch the problems appear.

The philosophical position is of its moment. Russell had recently broken with British idealism and was arguing for a realism in which the world is made of many independent things, and the sense-datum theory that dominates the first half of the book was the going account among his circle. He abandoned parts of it within a decade. That does not damage the book, because what survives is the way the questions are set up.

The prose is famously good: short sentences, concrete examples, no jargon introduced without being earned. Russell won the Nobel Prize in Literature in 1950 and this is a fair sample of why.

The argument

Chapters 1 to 3: the table. Russell looks at a table. It appears brown, but the colour differs from different angles and under different lights, and a photograph would show a different distribution again. It feels smooth, but under a microscope it is not. It sounds solid when struck, but the sound depends on the striker.

So the colour, shape, hardness and sound we are directly aware of are not properties of the table itself; they are what Russell calls sense-data. And the table, if it exists, is something inferred behind them.

He then asks the question the whole book turns on: is there a real table at all? Berkeley said no, only ideas in minds. Leibniz said yes but it is a colony of souls. Russell’s answer is that there is no logical proof, but that the hypothesis of an independent physical object is by far the simplest explanation of why our sense-data behave as they do, and particularly of why other people report the same ones. He accepts it as an instinctive belief that survives scrutiny rather than as a demonstration, and he is honest that this is a weaker footing than certainty.

Chapter 3 works out what physics says the table is: something in a space of its own, related to the space of our senses but not identical with it, whose intrinsic nature is unknown to us. We know the structure and not the quality.

Chapter 4: idealism. A short treatment of Berkeley’s argument that to be is to be perceived, and of the idealist claim that whatever can be known must be mental. Russell locates the error in an equivocation on the word idea, between the act of thinking and the thing thought about, and points out that from the fact that we can only know what we can know, nothing follows about what there is.

Chapter 5: acquaintance and description. The most influential chapter, and the one that made the book matter to professionals.

We have acquaintance with a thing when we are directly aware of it without inference: sense-data, memories, our own inner states, and universals. We have knowledge by description when we know something only as the so-and-so, the thing that has such-and-such a property, without ever having met it. Bismarck’s acquaintances knew him one way; we know him as the first Chancellor of the German Empire.

The chapter’s principle is that every proposition we can understand must be composed entirely of constituents we are acquainted with. This is what allows us to make meaningful statements about things we have never encountered, and it connects to Russell’s theory of descriptions, published in 1905, which showed how to analyse a sentence like the present King of France is bald without committing to a bald king.

Chapter 6: induction. Short, clear, and the most quoted thing in the book after the last chapter.

We expect the sun to rise because it always has. But what licenses the move from all observed cases to the next case? Not logic, since it is perfectly consistent for the sun not to rise. Not experience, since experience only tells us about cases already observed, and to argue that induction has worked before and therefore will work again is to use induction to defend induction.

The chicken makes the point. It is fed every morning and forms a reasonable expectation, right up to the morning the farmer wrings its neck.

Russell concludes that the principle of induction cannot be proved from experience and must be accepted on the strength of its own self-evidence, or else all expectation about the future is groundless. Hume had shown this and been read as a sceptic. Russell states it as a standing problem, and that is how it has been treated ever since.

Chapters 7 to 10: a priori knowledge and universals. How can we know anything in advance of experience? Russell rejects Kant’s answer, that the mind imposes the forms of space, time and causality, on the ground that it makes the uniformity of nature depend on a fact about us that could change.

His own answer runs through universals. When we see two white patches we are acquainted with whiteness, which is not either patch. He then presses the case with relations, which he thinks nominalists cannot handle: Edinburgh is north of London, and the relation is not in Edinburgh, not in London, not in the mind of anyone thinking it, and does not exist in time or space at all. It subsists.

Knowledge of the relations between universals is a priori. This is why arithmetic and logic can be known without further experience: they are about what holds among universals, and no fact about tables enters into them.

Chapters 11 to 14: knowledge, error and truth. Intuitive knowledge, the self-evident truths on which everything else rests, and the degrees of self-evidence. Truth as correspondence, defined so that beliefs are true when they correspond to a fact and false when they do not, with the machinery worked out on Othello’s belief that Desdemona loves Cassio. Knowledge as true belief properly derived, and Russell’s frank admission that saying exactly what properly means is very difficult, a difficulty that has occupied the subject since.

Chapter 15: the value of philosophy. Four pages, and the most reprinted thing Russell wrote.

Philosophy cannot show a balance sheet of established results, because as soon as a question becomes answerable it stops being philosophy and becomes a science. Its value is not in its answers. It is in what the questions do to the person asking them: they free the mind from the tyranny of custom and of what he calls the prejudices derived from common sense; they show that familiar things can be made strange; and they enlarge the self by directing attention outward, at what is not us and not for us.

The uncertainty, Russell says, is the point. A person who has never asked these questions lives confined in instinctive interests, and philosophy keeps the sense of wonder alive by showing how little we know.

How to read it

Read it in a week, a chapter or two at a time, and do the thinking as you go. It is written to be argued with.

Chapters 1, 5, 6 and 15 are the ones that matter most. If you only read four, read those.

Do not accept the sense-data theory just because Russell states it confidently. It was the dominant view in 1912 and it came under sustained attack afterwards, notably from those who argued that treating the immediate objects of perception as a special kind of thing creates a veil between us and the world rather than explaining our access to it. Russell himself moved on. Reading the argument and then asking what is wrong with it is better training than accepting it.

Notice how much of the book is Hume rewritten. The problem of induction, the doubts about the self, the inference to external objects: Russell is working over Hume’s ground with better logical tools and a less sceptical temperament.

Read chapter 15 twice and then again when you have finished some harder philosophy. It is the answer to the question everyone asks about the subject, and it is more honest than most answers because it concedes the premise.

Questions it raises

• If we only ever encounter sense-data, what entitles us to believe in the table behind them?
• Can induction be defended without using induction? If not, what is its status?
• Is the relation between Edinburgh and London something real, and if so where is it?
• Knowledge by description lets us talk about what we have never met. How much of what you know is of that kind?
• Russell says philosophy’s questions lose the name as soon as they can be answered. Does that make it a residue, or a method?
• If the value of philosophy is in enlarging the self rather than in results, how would you tell good philosophy from bad?