In set theory, the union of a collection of sets is the set of all elements that belong to at least one member of that collection. For two sets and , their union is written . An element belongs to if it belongs to , to , or to both. Union is a fundamental set operation, used throughout mathematics to combine collections without retaining repeated occurrences of an element.
Definition and notation
The union of two sets is defined by
The word “or” is inclusive: membership in both sets is allowed. In the language of propositional logic,
For example, if
then
The element appears only once because sets do not record multiplicity.
Union differs from intersection, written , which contains only elements belonging to both sets. It also differs from the Cartesian product , whose elements are ordered pairs rather than the individual elements of and .
Unions of families of sets
Union applies to any set-indexed family, not just to two sets. If is a family of sets indexed by a set , its union is
If is itself a set whose elements are sets, the equivalent notation is
For example,
This operation removes one level of set membership. It does not recursively unpack every set that occurs as an element. For instance,
Here , rather than , is an element of the union.
Empty and infinite unions
The union of an empty family is the empty set:
There is no set in the family from which an element could enter the union. This is distinct from a family containing one empty set, although its union is also empty:
Infinite unions can produce sets substantially larger than any individual member. For example,
Using interval notation in the real numbers,
Every real number strictly between and lies in at least one interval, but neither endpoint belongs to any of them.
Algebraic properties
Union satisfies several basic identities.
| Property | Identity |
|---|---|
| Commutativity | |
| Associativity | |
| [[idempotence | Idempotence]] |
| Identity | |
| Absorption |
Associativity makes expressions such as unambiguous without parentheses.
Union and intersection distribute over one another:
Intersection with a fixed set also distributes over an arbitrary union:
Relative to a fixed universal set , De Morgan’s laws give
Thus, an element lies outside a union exactly when it lies outside every set in the family.
These identities can be proved by comparing membership conditions on the two sides. For example, membership in means belonging to and to at least one of or , which is equivalent to membership in .
Inclusion and least upper bounds
Union is closely related to the subset relation. Both input sets are contained in their union:
Moreover, if and , then
Consequently, is the smallest set containing both and .
More generally, is the least upper bound of the family under inclusion. Within the power set , ordered by the partial order , union therefore serves as the join operation. Together with intersection and complementation, it forms part of the standard Boolean algebra of subsets of .
Union is also monotone:
Cardinality and disjoint unions
For finite sets, the cardinality of a union satisfies
The subtraction corrects for elements counted once in each set. This is the two-set case of the inclusion–exclusion principle.
If the sets are disjoint, meaning , then
For a finite pairwise disjoint family,
An ordinary union identifies elements shared by different sets. A disjoint union instead retains their origins by attaching labels. One construction is
An element common to and then contributes two distinct tagged elements.
In standard set theory with the axiom of choice, a countable union of countable sets is countable. This statement is not generally provable in set theory without some form of choice.
Uses in mathematics
Probability and measure
In probability, the union of two events represents the event that at least one occurs:
For a countable family of events, the union bound states
Equality holds for pairwise disjoint events by countable additivity.
A sigma-algebra is closed under countable unions, not necessarily arbitrary unions. This restriction is central to measure theory: an arbitrary union of measurable sets need not be measurable.
Topology
In topology, arbitrary unions of open sets are open. Finite unions of closed sets are closed, but infinite unions need not be. For example, each interval
is closed in , while
is not closed.
Algebraic structures
The union of structured subsets need not preserve their structure. For example, the union of the two coordinate axes in is not a linear subspace: it contains and , but not their sum .
By contrast, the union of a nonempty family of subspaces totally ordered by inclusion is a subspace. Any two vectors in the union belong together to one member of the family, allowing closure under vector addition and scalar multiplication to be checked there.
Set-theoretic foundations
In Zermelo–Fraenkel set theory, the axiom of union ensures that for every set , there is a set containing exactly the elements that belong to some member of . Binary union follows by first forming the pair and then taking its union:
The requirement that be a set matters. Union notation does not imply that every unrestricted collection has a union that is itself a set. For example, the collection of all sets is not a set in this theory.