Russell's Paradox

Russell’s paradox is a contradiction arising from the assumption that every definable condition determines a set. Consider the purported set of all sets that are not members of themselves: its defining condition implies that it belongs to itself if and only if it does not. Discovered by Bertrand Russell in 1901, the paradox exposed a fundamental problem with unrestricted set formation and helped motivate axiomatic set theory and type theory. (plato.stanford.edu)

Formulation and proof

The problematic principle is unrestricted comprehension: for any formula $\varphi(x)$, there exists a set containing exactly those objects satisfying that formula. In the language of first-order logic, it has the form

where $y$ must not occur free in $\varphi$. This is an axiom schema, with an instance for each permitted formula. Taking $\varphi(x)$ to be $x\notin x$ produces the purported Russell set:

Its defining property is

Substituting $R$ for $x$ gives

If $R\in R$, the defining condition entails $R\notin R$. Conversely, if $R\notin R$, it satisfies the condition for membership and therefore belongs to $R$. Neither alternative is consistent. (plato.sydney.edu.au)

The argument concerns membership, not inclusion as a subset. It does not dispute the fact that every set is a subset of itself. Rather, it shows that no set can have exactly the proposed membership condition under the usual logical rules. (plato.stanford.edu)

Historical background

Russell discovered the contradiction while working on The Principles of Mathematics, published in 1903. On June 16, 1902, he communicated it to Gottlob Frege, whose second volume of Grundgesetze der Arithmetik was approaching publication. Frege added an appendix discussing the problem. (plato.stanford.edu)

Frege’s system was not simply ordinary set theory with unrestricted comprehension as an explicit axiom. It distinguished concepts from objects and associated concepts with object-like extensions. His Basic Law V required extensions of concepts to be identical exactly when the concepts applied to the same objects. Combined with his comprehension principles for concepts, this permitted a Russell-style contradiction. The difficulty therefore concerned the interaction of these principles, rather than the ordinary principle that sets with the same members are equal. (plato.sydney.edu.au)

Why the contradiction matters

In classical logic, contradictory premises entail any proposition whatsoever. This property is called the principle of explosion:

Consequently, a theory deriving the Russell contradiction cannot distinguish provable propositions from their negations: every sentence becomes derivable. The problem is therefore not merely an unusual set with puzzling properties, but the collapse of a proposed foundational theory as a discriminating system of mathematical proof. (plato.stanford.edu)

Russell’s paradox establishes an inconsistency in unrestricted comprehension, not an inconsistency in every possible theory of sets. Different foundational systems avoid it by changing which collections exist, which expressions are well formed, or which logical inferences are permitted. (plato.stanford.edu)

Restricted set formation

In Zermelo–Fraenkel set theory (ZF), unrestricted comprehension is replaced by more limited existence principles. Separation permits a definable subset of an already existing set:

Thus, for any set $A$, one may form

This produces no contradiction. Instead, it implies that $R_A\notin A$: if $R_A$ belonged to $A$, substituting it into its defining condition would reproduce the Russell contradiction. This is a direct deduction from separation. In particular, no set can contain every set, because such a set would have to contain $R_A$. (plato.stanford.edu)

ZF also includes the axiom of foundation, which rules out self-membership. However, foundation is not needed for the preceding argument: separation already excludes a universal set. Avoiding Russell’s paradox is therefore not merely a matter of prohibiting sets from belonging to themselves; unrestricted comprehension must also be rejected. (plato.stanford.edu)

Types and stratification

Type-theoretic approaches restrict the kinds of objects to which predicates or membership relations may apply. In a simple hierarchy, individuals occupy one level, collections of individuals another, and collections of those collections a further level. A collection cannot be treated as an object of the same type as its members. Expressions corresponding to $x\in x$, or to a predicate applying to itself, are consequently not well formed. (plato.stanford.edu)

This differs from the ZF response. ZF allows the formula $x\in x$ to be written but restricts set-existence principles; simple type theory blocks the relevant self-application through rules governing expressions. Type theory subsequently became a framework for formal logic and computational foundations, not merely a device for excluding this particular paradox. (plato.stanford.edu)

Another approach is stratified comprehension, used in New Foundations. A formula qualifies for comprehension only if its variables can be assigned levels so that, in every membership expression $x\in y$, the level of $y$ is one higher than that of $x$. The formula $x\notin x$ cannot satisfy this requirement. By contrast, $x=x$ can, allowing a universal set without licensing the Russell set. This illustrates that a universal set is not itself the decisive problem: the contradiction depends on which comprehension principles accompany it. (sites.math.rutgers.edu)

The barber analogy

The barber paradox illustrates the same logical pattern. Suppose a barber belongs to a group and shaves exactly those members who do not shave themselves. Does the barber shave himself? Either answer conflicts with the stipulated rule.

The conclusion is that no such barber exists. Russell’s paradox is foundationally more consequential because unrestricted comprehension supplies a general principle apparently guaranteeing the existence of the corresponding set. The contradiction therefore requires abandoning or restricting that principle, rather than merely rejecting a fictional person’s description. (plato.stanford.edu)

Alternative logical responses

Paraconsistent logic rejects explosion: a contradiction need not entail every proposition. Some approaches to set-theoretic paradoxes accordingly retain forms of broad comprehension while modifying the underlying logic. Such an approach differs from ZF or simple type theory: rather than necessarily preventing contradictory conclusions, it seeks to prevent them from trivializing the entire theory. Whether a particular system succeeds depends on its precise axioms and inference rules; paraconsistency alone does not settle which sets exist. (plato.stanford.edu)

References

1. Russell’s Paradox — Stanford Encyclopedia of Philosophyplato.stanford.edu
2. Frege’s Theorem and Foundations for Arithmetic — Stanford Encyclopedia of Philosophyplato.sydney.edu.au
3. Set Theory — Stanford Encyclopedia of Philosophyplato.stanford.edu
4. Type Theory — Stanford Encyclopedia of Philosophyplato.stanford.edu
5. Paradoxes in Set Theorysites.math.rutgers.edu
6. Paraconsistent Logic — Stanford Encyclopedia of Philosophy, Summer 2026 Editionplato.stanford.edu