Summary of "The Problems of Philosophy"

Core Idea

Russell treats philosophy as the critical study of what can be known with certainty, especially where common sense and science leave room for doubt.
The book is mainly about knowledge: how we know, what we know directly, what we know only by description, and why much of our “knowledge” is really probable opinion.
Its deepest theme is that philosophy should not promise final metaphysical systems; its value lies in clarifying problems, resisting dogmatism, and enlarging the mind.

What We Directly Know

Russell begins with the ordinary table to show that sensory appearances vary with perspective, lighting, instruments, and the sense involved.
From this he distinguishes sense-data: the immediate items of awareness, from the act of sensation and from the inferred physical object.
What we perceive is not the table “in itself,” but changing appearances that function as signs of a table-like physical object.
This supports his anti-idealist point against Berkeley and others: confusing the act of apprehending with the thing apprehended does not show that the thing is mental.
Russell distinguishes knowledge by acquaintance from knowledge by description.
Acquaintance is direct awareness of things such as sense-data, memory contents, introspection, and probably some universals; description lets us think about what we are not directly acquainted with.
Many proper names and common terms work as disguised descriptions, so we can refer to Julius Caesar, Bismarck, or the Emperor of China without direct acquaintance.
The basic constraint is that any proposition we understand must be built from constituents with which we are acquainted.

Knowledge, Induction, and A Priori Truth

Russell says knowledge of truths falls into intuitive and derivative knowledge, but derivative knowledge always depends on already-known premises, so it cannot be the whole story.
Even when we know something by reading or inference, the psychological process of arriving at the belief is not enough unless there is also a valid logical connection.
Human knowledge is therefore mixed with uncertainty; there is no sharp line that cleanly separates knowledge from probable opinion.
The biggest problem is induction: no amount of past regularity logically proves that the future will resemble the past.
The principle of induction says that repeated association of A with B raises the probability of their recurring together, but this principle itself cannot be proven by induction without circularity.
Russell argues that some truths are known a priori, especially logical principles, arithmetic, geometry, and some ethical truths.
These are not truths about how the mind happens to work; they concern universals and relations that are not located in time and space.
Universals such as whiteness, resemblance, north of, before and after, and 2+2=4 are not particular things but what many particulars can share.
Russell stresses that relations are real and not merely mental; Edinburgh is north of London whether or not anyone thinks so.
In his view, all a priori knowledge is about the relations of universals, while empirical claims such as “all men are mortal” remain inductive and probable rather than necessary.
He rejects Kant’s attempt to make logic and arithmetic depend on the fixed structure of our minds, since that would make no stable guarantee that they remain valid.
Modern logic and mathematics, especially Cantor’s work and non-Euclidean geometry, weaken older claims that space, time, and infinity are impossible or purely subjective.
The result is that logic opens possibilities, while experience decides among them.

Truth, Self-Evidence, and Philosophy’s Value

Russell defines truth and falsehood as properties of beliefs, not of bare sense-data.
He rejects coherence as the definition of truth, because more than one coherent system may be possible and coherence already presupposes logic.
His preferred account is correspondence: a belief is true when the objects it concerns are actually related in the order the belief asserts.
His Othello example shows that belief is a complex relation to several objects, with direction and order mattering for truth.
He also distinguishes two senses of self-evidence: one that guarantees truth when we are acquainted with the relevant fact, and another that merely gives a stronger or weaker presumption.
Absolute self-evidence is private for particular sense-data, but relations among universals can be self-evident to more than one mind.
Much of what people call knowledge is really probable opinion, supported by degrees of self-evidence and by coherence with other probable beliefs.
Coherence is useful as a criterion of probability, not as a proof of truth; if dreams were as coherent as waking life, they would be harder to distinguish.
Russell rejects grand metaphysical systems, especially Hegel’s claim that contradictions in partial things prove that only the Whole is fully real and rational.
He argues that knowing some relations of a thing does not mean knowing all its relations, so fragmentariness does not imply unreality.
This undermines claims that space, time, matter, and evil are mere illusions of partial viewpoint.
Philosophy’s proper method is critical and piecemeal: to examine assumptions, test what is self-evident, and avoid overclaiming where experience cannot reach.
Its highest value is not practical utility but intellectual enlargement: it loosens prejudice, expands what seems possible, and restores wonder.
Philosophy makes us less centered on the self and more oriented toward the impersonal universe, which Russell treats as the mind’s highest good.

What To Take Away

Sense-data are immediate; physical objects are inferred, and much of ordinary realism is really knowledge by description.
A priori knowledge, for Russell, concerns relations among universals, not facts about particular things.
Induction is indispensable in practice but cannot be logically proved without circularity, so empirical knowledge always retains some probability.
Philosophy’s job is not to manufacture certainty where none is available, but to clarify distinctions, expose bad metaphysics, and widen the mind’s range of thought.