1 Definition and basic idea

The orthogonal complement of a subspace is the collection of all vectors that have zero inner product with every vector in that subspace. It captures the notion of directions that are perpendicular, not merely to a single vector, but to an entire linear set of vectors.

1.1 Inner product spaces

An inner product space is a vector space equipped with a rule that assigns a scalar to each pair of vectors, usually written as an inner product. This rule measures angles and lengths in a general algebraic setting. Orthogonal complement is defined in such spaces and depends on the chosen inner product.

1.2 Orthogonality to a subspace

A vector is orthogonal to a subspace if it is orthogonal to each vector in that subspace. The orthogonal complement of a subspace consists exactly of those vectors. This makes the construction useful for describing all directions that are perpendicular to a given set of constraints.

1.3 Notation and terminology

If \(W\) is a subspace, its orthogonal complement is often written as \(W^\perp\). The symbol indicates that the new set is built from vectors orthogonal to \(W\). The term complement here does not mean set-theoretic complement, but a related subspace determined by perpendicularity.

2 Examples

Concrete examples help show how orthogonal complements arise in familiar coordinate spaces and how they depend on the subspace being considered.

2.1 Orthogonal complement in Euclidean space

In \(\mathbb{R}^3\), the orthogonal complement of a plane through the origin is a line through the origin perpendicular to that plane. Likewise, the orthogonal complement of a line through the origin is the plane consisting of all vectors perpendicular to that line.

2.2 Orthogonal complement of the zero subspace

The orthogonal complement of the zero subspace contains every vector, since every vector has inner product zero with the zero vector. Therefore, the orthogonal complement of \(\{0\}\) is the entire space.

2.3 Orthogonal complement of the whole space

The orthogonal complement of the entire space is just the zero subspace. A vector orthogonal to every vector in a space must have zero inner product with itself, which forces it to be the zero vector in an inner product space.

2.4 Coordinate examples

In \(\mathbb{R}^2\), if a subspace is spanned by \((1,1)\), then its orthogonal complement is spanned by \((1,-1)\). In \(\mathbb{R}^3\), if a subspace is spanned by \((1,0,0)\) and \((0,1,0)\), then the orthogonal complement is spanned by \((0,0,1)\). Such examples are found by solving linear equations expressing orthogonality.

3 Basic properties

The orthogonal complement has several structural properties that make it a central tool in linear algebra.

3.1 Subspace property

The orthogonal complement of any subspace is itself a subspace. It contains the zero vector, is closed under addition, and is closed under scalar multiplication. This follows directly from the linearity of the inner product in each argument.

3.2 Inclusion-reversing behavior

If one subspace is contained in another, then their orthogonal complements are reversed in order. Specifically, if \(U \subseteq W\), then \(W^\perp \subseteq U^\perp\). A larger subspace imposes more orthogonality conditions, so fewer vectors satisfy them.

3.3 Relation to dimension

In finite-dimensional inner product spaces, the dimension of a subspace and the dimension of its orthogonal complement add up to the dimension of the whole space. This provides a precise numerical link between a subspace and the directions perpendicular to it.

3.4 Double orthogonal complement

Taking the orthogonal complement twice brings one back to the original subspace under common hypotheses, though the exact statement depends on dimension and closure properties.

3.4.1 Finite-dimensional case

For a subspace of a finite-dimensional inner product space, the orthogonal complement of the orthogonal complement equals the original subspace. This identity reflects the fact that finite-dimensional subspaces are completely determined by the vectors orthogonal to them.

3.4.2 Infinite-dimensional case

In infinite-dimensional settings, the double orthogonal complement need not equal the original subspace unless additional conditions are met. For example, in Hilbert spaces, the double orthogonal complement of a subspace equals its closure. This distinction makes closure an important issue in analysis.

4 Computation in finite-dimensional spaces

In practice, orthogonal complements are often computed using matrices, systems of equations, or basis methods.

4.1 Using matrices

If a subspace is described by vectors arranged as columns of a matrix, finding the orthogonal complement often reduces to solving a homogeneous linear system. The vectors in the orthogonal complement satisfy equations expressing zero inner product with each generating vector.

4.2 Null space interpretation

A subspace’s orthogonal complement can be interpreted as the null space of a suitable matrix. If the rows of a matrix span the given subspace, then vectors in the orthogonal complement are exactly those annihilated by that matrix. This gives a direct computational route.

4.3 Basis and spanning set methods

To compute an orthogonal complement, one may start with a basis for the subspace and write the orthogonality conditions against each basis vector. Solving the resulting linear system yields a basis for the orthogonal complement. A spanning set can be replaced by any basis, since orthogonality to the spanning set implies orthogonality to the whole subspace.

4.4 Orthogonal complement from row and column spaces

For a matrix, the orthogonal complement of the row space is the null space. Similarly, the orthogonal complement of the column space of a matrix is related to the left null space of that matrix. These relationships are central in matrix theory and connect subspace geometry to linear systems.

5 Geometric interpretation

Orthogonal complements have a clear geometric meaning: they identify all directions that lie at right angles to a given subspace.

5.1 Perpendicular directions

In low-dimensional spaces, orthogonal complements are easy to visualize. A line’s orthogonal complement in a plane is another line perpendicular to it. In higher dimensions, the complement may have dimension greater than one, representing an entire family of perpendicular directions.

5.2 Projection onto a subspace

The orthogonal complement is the part of a vector space that is removed when projecting onto a subspace. A vector can be split into a component inside the subspace and a component orthogonal to it. Orthogonal projection uses this decomposition to find the closest point in the subspace.

5.3 Decomposition of vectors

Every vector in a suitable inner product space can often be decomposed into a sum of a vector in a subspace and a vector in its orthogonal complement. This is one of the most useful features of the concept.

5.3.1 Orthogonal decomposition

An orthogonal decomposition expresses a vector as \(v = u + w\), where \(u\) lies in the subspace and \(w\) lies in the orthogonal complement. The two components are perpendicular, so their interaction is especially simple.

5.3.2 Orthogonal direct sum

When a space is written as an orthogonal direct sum of two subspaces, each vector has a unique decomposition into a sum of vectors from those subspaces. This structure combines algebraic direct sum with geometric perpendicularity.

6 Connections with linear algebra

Orthogonal complements are closely tied to classical linear algebra, especially the study of equations, ranks, and matrix subspaces.

6.1 Systems of linear equations

A system of linear equations can be viewed as a search for vectors orthogonal to certain coefficient vectors. The solution set often emerges as an orthogonal complement or is related to one through matrix formulations. This perspective unifies geometry and algebra in the study of linear constraints.

6.2 Rank-nullity relationships

The dimensions of a subspace and its orthogonal complement fit into rank-nullity-type relations. For matrices, the sizes of the row space, column space, null space, and left null space are linked through rank. Orthogonal complement language clarifies why these dimensions balance.

6.3 Fundamental subspaces of a matrix

The four fundamental subspaces of a matrix provide a standard framework in which orthogonal complements appear naturally.

6.3.1 Row space and null space

The null space is the orthogonal complement of the row space. Vectors in the null space satisfy all equations encoded by the rows, so they are perpendicular to every row vector.

6.3.2 Column space and left null space

The left null space is the orthogonal complement of the column space. It consists of vectors orthogonal to every column of the matrix. This duality is essential in understanding solvability and matrix structure.

Several related ideas are often discussed alongside orthogonal complements because they rely on the same geometric framework.

7.1 Orthogonal projection

Orthogonal projection is the process of mapping a vector to the nearest point in a subspace. The error vector lies in the orthogonal complement. This makes projection computations closely tied to the structure of complements.

7.2 Orthogonal basis

An orthogonal basis is a basis whose vectors are pairwise orthogonal. Such bases simplify the calculation of complements, projections, and coordinates. They are often produced from arbitrary bases by orthogonalization procedures.

7.3 Orthonormal complements

An orthonormal complement is a basis for the orthogonal complement consisting of vectors of unit length and mutual orthogonality. Orthonormal sets provide especially convenient coordinates and simplify numerical work.

7.4 Orthogonal direct complements

A subspace is an orthogonal direct complement of another if together they form the whole space and each is orthogonal to the other. This is stronger than merely being a direct complement, because it includes perpendicularity as part of the decomposition.

8 Applications

Orthogonal complements appear in many areas where constraints, approximation, and decomposition are important.

8.1 Least squares approximation

In least squares problems, one seeks the best approximation to an inconsistent system of linear equations. The error vector is orthogonal to the space of allowable fits, which leads to the normal equations. Orthogonal complements provide the conceptual basis for this method.

8.2 Signal processing

In signal processing, orthogonal complements help separate desired components from noise or interference. Subspaces may represent signal models, while orthogonal directions capture residual error or unwanted components. This viewpoint supports filtering and data compression.

8.3 Numerical linear algebra

Computational methods for solving linear systems and eigenvalue problems often rely on orthogonal decompositions. Orthogonal complements improve stability and make algorithms more robust. They also play a role in QR factorization and related procedures.

8.4 Functional analysis contexts

In functional analysis, orthogonal complements help describe closed subspaces in Hilbert spaces and support decomposition theorems. They are used in studying bounded linear operators, variational problems, and approximation in infinite-dimensional settings. Their behavior depends on closure, which distinguishes them from the finite-dimensional case.

</INTERNAL_LINK_CANDIDATES> Inner product space (a vector space with a scalar-valued inner product) Subspace (a subset closed under vector addition and scalar multiplication) Orthogonality (the condition of having zero inner product) Orthogonal projection (the closest-point projection onto a subspace) Direct sum (a decomposition of a space into complementary subspaces) Orthogonal direct sum (a direct sum with perpendicular subspaces) Null space (the set of vectors mapped to zero by a linear transformation) Row space (the span of the rows of a matrix) Column space (the span of the columns of a matrix) Left null space (vectors orthogonal to all columns of a matrix) Least squares (an approximation method minimizing squared error) QR factorization (a matrix decomposition using orthogonal factors) Hilbert space (a complete inner product space) Closure (the smallest closed set containing a given set) Projection theorem (a theorem guaranteeing orthogonal decomposition in Hilbert spaces) Orthogonal basis (a basis of pairwise perpendicular vectors) Orthonormal basis (an orthogonal basis of unit vectors) Rank-nullity theorem (the dimension relation linking rank and nullity) Homogeneous linear system (a linear system with zero right-hand side) Normal equations (equations characterizing least-squares solutions)