1 Definition and axioms
An inner product is a rule that assigns a scalar to an ordered pair of vectors in a vector space, while encoding geometric information such as size, angle, and orthogonality. It generalizes the ordinary dot product from Euclidean space and provides the algebraic foundation for norms and projections.
1.1 Vector spaces with inner products
An inner product is defined on a vector space over the real or complex numbers. A vector space equipped with such a rule is called an inner product space. In the real case, the scalar output is real; in the complex case, it is complex and must be handled with conjugation in one argument. The choice of scalar field affects the exact form of the axioms but not the overall geometric interpretation.
1.2 Linearity
The inner product is linear in one of its arguments, usually the first in real spaces and one specified argument in complex spaces by convention. This means it respects addition and scalar multiplication in that slot. Linearity makes the inner product compatible with the structure of the vector space and allows many algebraic manipulations used throughout linear algebra.
1.3 Symmetry and conjugate symmetry
In a real inner product space, swapping the two vectors leaves the value unchanged. In a complex inner product space, the corresponding property is conjugate symmetry: interchanging the vectors produces the complex conjugate of the original value. This condition ensures that the scalar obtained from a vector with itself is always real.
1.4 Positive definiteness
The inner product of a vector with itself is never negative, and it is zero only for the zero vector. This requirement distinguishes genuine geometric length from arbitrary bilinear or sesquilinear forms. Positive definiteness is what permits the definition of a norm derived from the inner product.
1.5 Sesquilinearity
In complex settings, an inner product is often described as sesquilinear, meaning linear in one argument and conjugate linear in the other. This property is the complex analogue of bilinearity and is essential for preserving positivity. It also aligns the formalism with applications in analysis and quantum theory.
2 Basic examples
Inner products appear in many familiar algebraic and analytic settings. The most important examples come from coordinate vectors, functions, and matrices, where the same abstract axioms are realized in different concrete forms.
2.1 Euclidean dot product
The standard inner product on real coordinate space is the dot product. For vectors with finitely many components, it is the sum of pairwise products of corresponding entries. This example underlies elementary geometry and is the model from which the general theory is abstracted.
2.2 Complex inner products
For complex vectors, the standard inner product includes complex conjugation in one factor. This ensures that the value of a vector with itself is real and nonnegative. Complex inner products are central in areas where amplitudes and phases matter, such as signal processing and quantum mechanics.
2.3 Function spaces
Functions can also be treated like vectors, and inner products can be defined by combining their values across a domain. This viewpoint is fundamental in analysis, where the geometry of function spaces often mirrors finite-dimensional vector geometry.
2.3.1 Integrals as inner products
A common example is the integral of the product of two functions over an interval or region. When the functions are square-integrable, the integral defines an inner product that measures how closely they align. This construction is widely used in approximation theory and Fourier analysis.
2.3.2 Weighted inner products
In some settings, the integrand is multiplied by a positive weight function. The weight changes how different parts of the domain contribute to the result. Weighted inner products are useful when certain regions or variables should count more heavily than others.
2.4 Matrix spaces
Matrices may be viewed as vectors in a higher-dimensional space, and an inner product can be defined using the sum of products of corresponding entries. One common choice is the Frobenius-type inner product. This is useful in numerical computation, statistics, and the study of matrix approximations.
3 Geometric interpretation
Inner products encode geometric relations in algebraic form. They allow one to measure lengths, compare directions, and detect orthogonality without leaving the language of vectors.
3.1 Length and norm
The inner product of a vector with itself determines its length through the associated norm. This norm generalizes the notion of magnitude from Euclidean geometry. As a result, vector spaces with inner products inherit a natural way to measure how large a vector is.
3.2 Angle between vectors
When an inner product is available, the angle between nonzero vectors can be defined by a cosine relation. This extends the familiar formula from ordinary geometry. The concept is especially meaningful in real inner product spaces, where it directly reflects directional similarity.
3.3 Orthogonality
Two vectors are orthogonal when their inner product is zero. This is the algebraic counterpart of perpendicularity. Orthogonality is a powerful organizing principle because it often simplifies computations and decomposition of vector spaces.
3.4 Pythagorean theorem
If two vectors are orthogonal, the norm of their sum satisfies a Pythagorean relation. The length of the resulting vector is obtained from the square root of the sum of the squared lengths. This theorem extends a familiar geometric fact to abstract inner product spaces.
4 Fundamental identities
Several key inequalities and identities follow from the axioms of an inner product. These results are central because they connect algebraic definitions with geometric consequences and support the theory of norms and convergence.
4.1 Cauchy–Schwarz inequality
The Cauchy–Schwarz inequality bounds the absolute value of an inner product by the product of the two norms. It is one of the most important results in the subject. Many later theorems depend on it, including the triangle inequality and estimates for projections.
4.2 Triangle inequality
The norm induced by an inner product satisfies the triangle inequality. In geometric terms, the length of a sum does not exceed the sum of the lengths. This property is essential for interpreting the norm as a distance-like measure.
4.3 Parallelogram law
The parallelogram law relates the norms of two vectors and the norms of their sum and difference. It characterizes norms that arise from inner products. In this way, it provides a bridge between metric geometry and linear structure.
4.4 Polarization identity
The polarization identity reconstructs the inner product from the norm in spaces where the relevant axioms hold. It shows that the norm contains enough information to recover the full inner product structure. This identity is especially useful in theoretical investigations and uniqueness arguments.
5 Derived constructions
Once an inner product is given, several additional constructions become available. These tools are among the main reasons inner product spaces are so useful in mathematics and its applications.
5.1 Norms induced by inner products
Every inner product determines a norm by taking the square root of the self-inner product of a vector. This induced norm measures length in a way consistent with the geometry encoded by the inner product. Not every norm comes from an inner product, but those that do have special structural properties.
5.2 Metric and distance
The induced norm yields a distance function between vectors by measuring the norm of their difference. This turns the vector space into a metric space. Distances defined in this way support notions of convergence, continuity, and approximation.
5.3 Projections
An orthogonal projection maps a vector onto a subspace so that the error is orthogonal to that subspace. Projections are central in approximation theory and least squares methods. They provide the best approximation in the least-distance sense when the appropriate conditions hold.
5.4 Orthogonal complements
The orthogonal complement of a subspace consists of all vectors orthogonal to every vector in that subspace. This construction helps decompose spaces into mutually perpendicular parts. It is a standard tool for solving linear equations and understanding subspace structure.
5.5 Orthonormal bases
An orthonormal basis is a basis whose vectors are mutually orthogonal and each of unit length. In such a basis, coordinates and inner products become especially simple. Orthonormal bases are invaluable for computation, decomposition, and theoretical clarity.
6 Inner product spaces
Inner product spaces vary greatly depending on dimension and completeness. Finite-dimensional cases are often more concrete, while infinite-dimensional settings require greater analytical care.
6.1 Finite-dimensional inner product spaces
In finite dimensions, all inner product spaces behave in a strongly controlled way. Many properties can be expressed using coordinates and matrices, and every subspace has an orthonormal basis. This setting is the natural home of many standard linear algebra methods.
6.2 Infinite-dimensional inner product spaces
In infinite-dimensional spaces, familiar geometric ideas remain valid, but new phenomena arise. Sequences of vectors may converge or fail to converge in ways that do not appear in finite dimensions. Such spaces are central in analysis, especially when studying functions and operators.
6.3 Hilbert spaces
A Hilbert space is an inner product space that is complete with respect to the norm it induces. Completeness means that every Cauchy sequence converges within the space. Hilbert spaces form one of the most important frameworks in modern analysis.
6.4 Completeness
Completeness guarantees that limits of well-behaved sequences are not lost outside the space. This property is crucial for solving equations, minimizing errors, and developing spectral theory. Without completeness, many analytic techniques become harder to apply.
7 Linear operators
Inner products interact naturally with linear maps. This interaction leads to a rich operator theory in which geometry and algebra reinforce each other.
7.1 Adjoint operators
The adjoint of a linear operator is defined through the inner product relation that moves the operator from one argument to the other. It is the analog of a transpose in real spaces and a conjugate transpose in complex spaces. Adjoint operators are fundamental in both theory and computation.
7.2 Self-adjoint operators
An operator is self-adjoint when it equals its own adjoint. Such operators often have especially orderly spectral behavior and frequently correspond to real-valued measurements or symmetric transformations. They play a major role in analysis and physics.
7.3 Unitary operators
A unitary operator preserves inner products, and therefore lengths and angles. It is the complex analogue of an orthogonal transformation. Unitary operators represent structure-preserving changes of coordinates and are important in many branches of mathematics.
7.4 Orthogonal projections
Orthogonal projections are linear operators that map each vector to the nearest point in a subspace. They are idempotent and self-adjoint in the standard setting. These operators appear in decomposition theorems, approximation methods, and numerical algorithms.
8 Applications
Inner products have wide-ranging applications because they convert geometric intuition into algebraic machinery. They are used whenever notions of similarity, orthogonality, and best approximation are needed.
8.1 Least squares approximation
In least squares problems, one seeks the vector or function that best fits data by minimizing the squared error. Inner products provide the framework for formulating and solving these problems. The solution is often characterized by orthogonality of the residual to the approximation space.
8.2 Fourier series
Fourier series expand functions into sums of orthogonal basis functions, typically trigonometric ones. Inner products determine the coefficients of the expansion. This method is a cornerstone of harmonic analysis and signal representation.
8.3 Gram–Schmidt process
The Gram–Schmidt process converts a linearly independent set of vectors into an orthonormal set spanning the same subspace. It relies on repeated orthogonal projection and subtraction. The procedure is widely used in theoretical derivations and practical computations.
8.4 Quantum mechanics notation
In quantum mechanics, states are represented using vectors in complex inner product spaces. Inner products encode transition amplitudes and probabilities, while orthogonality corresponds to distinguishable states. The formalism is usually expressed with bra-ket notation, which is adapted to the geometry of Hilbert spaces.
8.5 Numerical linear algebra
Inner products are central to many algorithms for solving systems, eigenvalue problems, and approximation tasks. They are used to measure residuals, construct orthogonal bases, and stabilize computations. Their role is especially visible in iterative methods and matrix decompositions.