The binomial theorem is an identity in algebra that expresses a nonnegative integer power of a sum of two terms as a finite sum of products. Its coefficients are binomial coefficients, which count the ways of selecting objects from a finite collection. The theorem connects polynomial algebra with combinatorics and probability. A generalized form expresses noninteger powers as infinite series, subject to convergence conditions.
Statement
For every nonnegative integer ,
where
Here denotes the factorial of , with . The coefficient , read “ choose ,” counts the -element subsets of an -element set.
The expansion has indexed terms. In each term, the exponents of and add to : the exponent of decreases from to , while that of increases from to . Thus the right-hand side is a homogeneous polynomial of total degree .
For example,
and
Replacing with gives
The identity is valid for real numbers, complex numbers, and, more generally, commuting elements of a ring with identity. The case is interpreted as the constant identity , with zeroth powers understood algebraically.
Why the coefficients appear
The product
is expanded by choosing either or from each factor and multiplying the choices.
A term results when is chosen from exactly of the factors. There are ways to choose those factors. Since multiplication is commutative, all such choices give the same monomial, so their contributions combine into
This counting argument both proves the theorem and explains why the coefficients are nonnegative integers, even though their factorial formula is written as a quotient.
For instance, in , the term occurs once for every choice of two factors from which to take . The six choices produce the coefficient .
Pascal’s triangle and induction
The binomial coefficients satisfy Pascal’s identity:
with coefficients outside the range taken to be zero. They form the rows of Pascal’s triangle:
Each interior entry is the sum of the two entries immediately above it. The row beginning with supplies the coefficients of .
Pascal’s identity also gives a proof by mathematical induction. The result holds for . If it holds for , multiplying the expansion by and collecting the coefficient of gives
The resulting expression is the expansion for .
The coefficients are symmetric:
Combinatorially, selecting objects is equivalent to specifying the objects left unselected. Algebraically, the symmetry corresponds to interchanging and .
Consequences and applications
Coefficient identities
Substituting yields
This counts all subsets of an -element set: every element is either included or excluded.
Substituting and gives, for ,
Consequently, a nonempty finite set has equally many subsets of even and odd cardinality.
Taking the derivative of
and evaluating at gives, for ,
Further differentiation produces identities involving higher products such as .
Probability
For independent trials, each having success probability , the binomial distribution assigns probability
to exactly successes.
The factor is the probability of one specified sequence containing successes, using statistical independence. The binomial coefficient counts the possible placements of those successes.
The binomial theorem verifies that these probabilities sum to one:
Number theory
If is a prime number, then
Therefore, for integers and ,
In modular arithmetic, all intermediate terms disappear modulo . The same identity holds exactly in a commutative ring of characteristic , where adding the multiplicative identity to itself times gives zero.
Approximation
For a fixed nonnegative integer , the exact expansion gives
When is sufficiently small relative to the required accuracy, retaining only the first few terms gives a useful approximation. The quality of that approximation depends on both and ; a small alone does not justify neglecting all higher terms when is large.
Generalized binomial theorem
For an arbitrary real or complex exponent , define
The generalized binomial theorem states that
For complex , the left-hand side is defined using the analytic branch of on the disk that equals zero at :
This specifies the branch unambiguously within that disk.
If is a nonnegative integer, the coefficients eventually vanish, recovering the finite theorem. Otherwise, the expansion is an infinite power series with radius of convergence . Convergence on the boundary requires separate analysis and is not guaranteed by the stated theorem.
For example,
which is a geometric series, and
For a positive integer ,
so
This expansion is useful in counting problems and generating functions.
Relation to Taylor series
The generalized expansion is the Taylor series of about zero. Repeated differentiation gives
so the Taylor coefficient is precisely .
For noninteger exponents, truncation after the quadratic term gives
as , with fixed. The remainder is expressed here using big-O notation.
Extensions and limitations
The multinomial theorem extends the finite expansion to sums of more than two terms:
Its coefficients count the assignments of labeled positions to categories with specified category sizes.
Commutativity is essential to the ordinary binomial formula. For elements and in a possibly noncommutative algebra,
This becomes only when . In particular, the usual theorem applies to square matrices of the same size when they commute, but not to arbitrary pairs of matrices. Without commutativity, expansion requires retaining the distinct ordered products.
The finite and generalized theorems also have different logical status: the former is an algebraic identity requiring no convergence argument, whereas the latter is an equality involving an infinite series whose domain and interpretation must be specified.
Historical development
The finite binomial theorem developed through several mathematical traditions. Geometric identities for squares and related products appear in ancient Greek mathematics, including Euclid’s *Elements*. Arrays and rules corresponding to binomial coefficients were studied in Indian, Islamic, and Chinese mathematics before their later systematic treatment in Europe.
Chinese work associated with Jia Xian in the eleventh century and Yang Hui in the thirteenth century included triangular arrangements of coefficients. In seventeenth-century Europe, Blaise Pascal gave an influential systematic treatment of the arithmetic triangle and its combinatorial properties. The modern name “Pascal’s triangle” therefore does not imply that Pascal was its earliest discoverer.
Isaac Newton developed the generalized binomial expansion for fractional and negative exponents in the 1660s. This extension made binomial expansions important tools for infinite-series calculations, including the evaluation of quadratures and the development of calculus. Later analysis supplied rigorous treatments of convergence and of complex powers.