Linear form - WikiHQ

In mathematics, a linear form (also known as a linear functional, a one-form, or a covector) is a linear map from a vector space to its field of scalars (often, the real numbers or the complex numbers).

If is a vector space over a field , the set of all linear functionals from to is itself a vector space over with addition and scalar multiplication defined pointwise. This space is called the dual space of , or sometimes the algebraic dual space, when a topological dual space is also considered. It is often denoted , or, when the field is understood, <math>V^*</math>; other notations are also used, such as <math>V'</math>, <math>V^{\#}</math> or <math>V^{\vee}.</math> It is, however, affine-linear.

Visualization

thumb|200px|Geometric interpretation of a 1-form α as a stack of [[hyperplanes of constant value, each corresponding to those vectors that α maps to a given scalar value shown next to it along with the "sense" of increase. The zero plane is through the origin.]]

In finite dimensions, a linear functional can be visualized in terms of its level sets, the sets of vectors which map to a given value. In three dimensions, the level sets of a linear functional are a family of mutually parallel planes; in higher dimensions, they are parallel hyperplanes. This method of visualizing linear functionals is sometimes introduced in general relativity texts, such as Gravitation by .

Applications

Application to quadrature

If <math>x_0, \ldots, x_n</math> are <math>n + 1</math> distinct points in , then the linear functionals <math>\operatorname{ev}_{x_i} : f \mapsto f\left(x_i\right)</math> defined above form a basis of the dual space of , the space of polynomials of degree <math>\leq n.</math> The integration functional is also a linear functional on , and so can be expressed as a linear combination of these basis elements. In symbols, there are coefficients <math>a_0, \ldots, a_n</math> for which

for all <math>f \in P_n.</math> This forms the foundation of the theory of numerical quadrature.

In quantum mechanics

Linear functionals are particularly important in quantum mechanics. Quantum mechanical systems are represented by Hilbert spaces, which are anti–isomorphic to their own dual spaces. A state of a quantum mechanical system can be identified with a linear functional. For more information see bra–ket notation.

Distributions

In the theory of generalized functions, certain kinds of generalized functions called distributions can be realized as linear functionals on spaces of test functions.

Dual vectors and bilinear forms

thumb|400px|Linear functionals (1-forms) α, β and their sum σ and vectors u, v, w, in [[three-dimensional space|3d Euclidean space. The number of (1-form) hyperplanes intersected by a vector equals the inner product.]]

Every non-degenerate bilinear form on a finite-dimensional vector space <math>V</math> induces an isomorphism <math>V\rightarrow V^*:v\mapsto v^*</math> such that

<math display="block"> v^*(w) := \langle v, w\rangle \quad \forall w \in V ,</math>

where the bilinear form on <math>V</math> is denoted <math>\langle \,\cdot\, , \,\cdot\, \rangle</math> (for instance, in Euclidean space, <math>\langle v, w \rangle = v \cdot w</math> is the dot product of <math>v</math> and <math>w</math>).

The inverse isomorphism is <math>V^*\rightarrow V:v^*\mapsto v</math>, where <math>v</math> is the unique element of <math>V</math> such that

<math display="block"> \langle v, w\rangle = v^*(w)</math>

for all <math>w \in V</math>.

The above defined vector <math>v^*\in V^*</math> is said to be the dual vector of <math>v \in V</math>.

In an infinite dimensional Hilbert space, analogous results hold by the Riesz representation theorem. There is a mapping <math>V\mapsto V^*</math> from <math>V</math> into its <math>V^*</math>.

Relationship to bases

Basis of the dual space

Let the vector space have a basis <math>\mathbf{e}_1, \mathbf{e}_2,\dots,\mathbf{e}_n</math>, not necessarily orthogonal. Then the dual space <math>V^*</math> has a basis <math>\tilde{\omega}^1,\tilde{\omega}^2,\dots,\tilde{\omega}^n</math> called the dual basis defined by the special property that

<math display="block"> \tilde{\omega}^i (\mathbf e_j) = \begin{cases}

1 &\text{if}\ i = j\\

0 &\text{if}\ i \neq j.

\end{cases} </math>

Or, more succinctly,

<math display="block"> \tilde{\omega}^i (\mathbf e_j) = \delta_{ij} </math>

where <math>\delta_{ij}</math> is the Kronecker delta. Here the superscripts of the basis functionals are not exponents but are instead contravariant indices.

A linear functional <math>\tilde{u}</math> belonging to the dual space <math>\tilde{V}</math> can be expressed as a linear combination of basis functionals, with coefficients ("components") ,

<math display="block">\tilde{u} = \sum_{i=1}^n u_i \, \tilde{\omega}^i. </math>

Then, applying the functional <math>\tilde{u}</math> to a basis vector <math>\mathbf{e}_j</math> yields

<math display="block">\tilde{u}(\mathbf e_j) = \sum_{i=1}^n \left(u_i \, \tilde{\omega}^i\right) \mathbf e_j = \sum_i u_i \left[\tilde{\omega}^i \left(\mathbf e_j\right)\right] </math>

due to linearity of scalar multiples of functionals and pointwise linearity of sums of functionals. Then

<math display="block">\begin{align}

\tilde{u}({\mathbf e}_j)

&= \sum_i u_i \left[\tilde{\omega}^i \left({\mathbf e}_j\right)\right] \\&

= \sum_i u_i {\delta}_{ij} \\

&= u_j.

\end{align}</math>

So each component of a linear functional can be extracted by applying the functional to the corresponding basis vector.

The dual basis and inner product

When the space carries an inner product, then it is possible to write explicitly a formula for the dual basis of a given basis. Let have (not necessarily orthogonal) basis <math>\mathbf{e}_1,\dots, \mathbf{e}_n.</math> In three dimensions (), the dual basis can be written explicitly

<math display="block"> \tilde{\omega}^i(\mathbf{v}) = \frac{1}{2} \left\langle \frac { \sum_{j=1}^3\sum_{k=1}^3\varepsilon^{ijk} \, (\mathbf e_j \times \mathbf e_k)} {\mathbf e_1 \cdot \mathbf e_2 \times \mathbf e_3} , \mathbf{v} \right\rangle ,</math>

for <math>i = 1, 2, 3,</math> where ε is the Levi-Civita symbol and <math>\langle \cdot , \cdot \rangle</math> the inner product (or dot product) on .

In higher dimensions, this generalizes as follows

<math display="block"> \tilde{\omega}^i(\mathbf{v}) = \left\langle \frac{\sum_{1 \le i_2 < i_3 < \dots < i_n \le n} \varepsilon^{ii_2\dots i_n}(\star \mathbf{e}_{i_2} \wedge \cdots \wedge \mathbf{e}_{i_n})}{\star(\mathbf{e}_1\wedge\cdots\wedge\mathbf{e}_n)}, \mathbf{v} \right\rangle ,</math>

where <math>\star</math> is the Hodge star operator.

Over a ring

Modules over a ring are generalizations of vector spaces, which removes the restriction that coefficients belong to a field. Given a module over a ring , a linear form on is a linear map from to , where the latter is considered as a module over itself. The space of linear forms is always denoted , whether is a field or not. It is a right module if is a left module.

The existence of "enough" linear forms on a module is equivalent to projectivity.

Change of field

Suppose that <math>X</math> is a vector space over <math>\Complex.</math> Restricting scalar multiplication to <math>\R</math> gives rise to a real vector space <math>X_{\R}</math> called the of <math>X.</math>

Any vector space <math>X</math> over <math>\Complex</math> is also a vector space over <math>\R,</math> endowed with a complex structure; that is, there exists a real vector subspace <math>X_{\R}</math> such that we can (formally) write <math>X = X_{\R} \oplus X_{\R}i</math> as <math>\R</math>-vector spaces.

Real versus complex linear functionals

Every linear functional on <math>X</math> is complex-valued while every linear functional on <math>X_{\R}</math> is real-valued. If <math>\dim X \neq 0</math> then a linear functional on either one of <math>X</math> or <math>X_{\R}</math> is non-trivial (meaning not identically <math>0</math>) if and only if it is surjective (because if <math>\varphi(x) \neq 0</math> then for any scalar <math>s,</math> <math>\varphi\left((s/\varphi(x)) x\right) = s</math>), where the image of a linear functional on <math>X</math> is <math>\C</math> while the image of a linear functional on <math>X_{\R}</math> is <math>\R.</math>

Consequently, the only function on <math>X</math> that is both a linear functional on <math>X</math> and a linear function on <math>X_{\R}</math> is the trivial functional; in other words, <math>X^{\#} \cap X_{\R}^{\#} = \{ 0 \},</math> where <math>\,{\cdot}^{\#}</math> denotes the space's algebraic dual space.

However, every <math>\Complex</math>-linear functional on <math>X</math> is an <math>\R</math>-linear (meaning that it is additive and homogeneous over <math>\R</math>), but unless it is identically <math>0,</math> it is not an <math>\R</math>-linear on <math>X</math> because its range (which is <math>\Complex</math>) is 2-dimensional over <math>\R.</math> Conversely, a non-zero <math>\R</math>-linear functional has range too small to be a <math>\Complex</math>-linear functional as well.

Real and imaginary parts

If <math>\varphi \in X^{\#}</math> then denote its real part by <math>\varphi_{\R} := \operatorname{Re} \varphi</math> and its imaginary part by <math>\varphi_i := \operatorname{Im} \varphi.</math>

Then <math>\varphi_{\R} : X \to \R</math> and <math>\varphi_i : X \to \R</math> are linear functionals on <math>X_{\R}</math> and <math>\varphi = \varphi_{\R} + i \varphi_i.</math>

The fact that <math>z = \operatorname{Re} z - i \operatorname{Re} (i z) = \operatorname{Im} (i z) + i \operatorname{Im} z</math> for all <math>z \in \Complex</math> implies that for all <math>x \in X,</math>

<math display=block>\begin{alignat}{4}\varphi(x) &= \varphi_{\R}(x) - i \varphi_{\R}(i x) \\

&= \varphi_i(i x) + i \varphi_i(x)\\

\end{alignat}</math>

and consequently, that <math>\varphi_i(x) = - \varphi_{\R}(i x)</math> and <math>\varphi_{\R}(x) = \varphi_i(ix).</math>

The assignment <math>\varphi \mapsto \varphi_{\R}</math> defines a bijective <math>\R</math>-linear operator <math>X^{\#} \to X_{\R}^{\#}</math> whose inverse is the map <math>L_{\bull} : X_{\R}^{\#} \to X^{\#}</math> defined by the assignment <math>g \mapsto L_g</math> that sends <math>g : X_{\R} \to \R</math> to the linear functional <math>L_g : X \to \Complex</math> defined by

The real part of <math>L_g</math> is <math>g</math> and the bijection <math>L_{\bull} : X_{\R}^{\#} \to X^{\#}</math> is an <math>\R</math>-linear operator, meaning that <math>L_{g+h} = L_g + L_h</math> and <math>L_{rg} = r L_g</math> for all <math>r \in \R</math> and <math>g, h \in X_\R^{\#}.</math>

Similarly for the imaginary part, the assignment <math>\varphi \mapsto \varphi_i</math> induces an <math>\R</math>-linear bijection <math>X^{\#} \to X_{\R}^{\#}</math> whose inverse is the map <math>X_{\R}^{\#} \to X^{\#}</math> defined by sending <math>I \in X_{\R}^{\#}</math> to the linear functional on <math>X</math> defined by <math>x \mapsto I(i x) + i I(x).</math>

This relationship was discovered by Henry Löwig in 1934 (although it is usually credited to F. Murray), and can be generalized to arbitrary finite extensions of a field in the natural way. It has many important consequences, some of which will now be described.

Properties and relationships

Suppose <math>\varphi : X \to \Complex</math> is a linear functional on <math>X</math> with real part <math>\varphi_{\R} := \operatorname{Re} \varphi</math> and imaginary part <math>\varphi_i := \operatorname{Im} \varphi.</math>

Then <math>\varphi = 0</math> if and only if <math>\varphi_{\R} = 0</math> if and only if <math>\varphi_i = 0.</math>

Assume that <math>X</math> is a topological vector space. Then <math>\varphi</math> is continuous if and only if its real part <math>\varphi_{\R}</math> is continuous, if and only if <math>\varphi</math>'s imaginary part <math>\varphi_i</math> is continuous. That is, either all three of <math>\varphi, \varphi_{\R},</math> and <math>\varphi_i</math> are continuous or none are continuous. This remains true if the word "continuous" is replaced with the word "bounded". In particular, <math>\varphi \in X^{\prime}</math> if and only if <math>\varphi_{\R} \in X_{\R}^{\prime}</math> where the prime denotes the space's continuous dual space.

Let <math>B \subseteq X.</math> If <math>u B \subseteq B</math> for all scalars <math>u \in \Complex</math> of unit length (meaning <math>|u| = 1</math>) then <math display=block>\sup_{b \in B} |\varphi(b)| = \sup_{b \in B} \left|\varphi_{\R}(b)\right|.</math>

Similarly, if <math>\varphi_i := \operatorname{Im} \varphi : X \to \R</math> denotes the complex part of <math>\varphi</math> then <math>i B \subseteq B</math> implies <math display=block>\sup_{b \in B} \left|\varphi_{\R}(b)\right| = \sup_{b \in B} \left|\varphi_i(b)\right|.</math>

If <math>X</math> is a normed space with norm <math>\|\cdot\|</math> and if <math>B = \{x \in X : \| x \| \leq 1\}</math> is the closed unit ball then the supremums above are the operator norms (defined in the usual way) of <math>\varphi, \varphi_{\R},</math> and <math>\varphi_i</math> so that <math display=block>\|\varphi\| = \left\|\varphi_{\R}\right\| = \left\|\varphi_i \right\|.</math>

This conclusion extends to the analogous statement for polars of balanced sets in general topological vector spaces.

If <math>X</math> is a complex Hilbert space with a (complex) inner product <math>\langle \,\cdot\,| \,\cdot\, \rangle</math> that is antilinear in its first coordinate (and linear in the second) then <math>X_{\R}</math> becomes a real Hilbert space when endowed with the real part of <math>\langle \,\cdot\,| \,\cdot\, \rangle.</math> Explicitly, this real inner product on <math>X_{\R}</math> is defined by <math>\langle x | y \rangle_{\R} := \operatorname{Re} \langle x | y \rangle</math> for all <math>x, y \in X</math> and it induces the same norm on <math>X</math> as <math>\langle \,\cdot\,| \,\cdot\, \rangle</math> because <math>\sqrt{\langle x | x \rangle_{\R = \sqrt{\langle x | x \rangle}</math> for all vectors <math>x.</math> Applying the Riesz representation theorem to <math>\varphi \in X^{\prime}</math> (resp. to <math>\varphi_{\R} \in X_{\R}^{\prime}</math>) guarantees the existence of a unique vector <math>f_{\varphi} \in X</math> (resp. <math>f_{\varphi_{\R \in X_{\R}</math>) such that <math>\varphi(x) = \left\langle f_{\varphi} | \, x \right\rangle</math> (resp. <math>\varphi_{\R}(x) = \left\langle f_{\varphi_{\R | \, x \right\rangle_{\R}</math>) for all vectors <math>x.</math> The theorem also guarantees that <math>\left\|f_{\varphi}\right\| = \|\varphi\|_{X^{\prime</math> and <math>\left\|f_{\varphi_{\R\right\| = \left\|\varphi_{\R}\right\|_{X_{\R}^{\prime.</math> It is readily verified that <math>f_{\varphi} = f_{\varphi_{\R.</math> Now <math>\left\|f_{\varphi}\right\| = \left\|f_{\varphi_{\R\right\|</math> and the previous equalities imply that <math>\|\varphi\|_{X^{\prime = \left\|\varphi_{\R}\right\|_{X_{\R}^{\prime,</math> which is the same conclusion that was reached above.

In infinite dimensions

Below, all vector spaces are over either the real numbers <math>\R</math> or the complex numbers <math>\Complex.</math>

If <math>V</math> is a topological vector space, the space of continuous linear functionals — the — is often simply called the dual space. If <math>V</math> is a Banach space, then so is its (continuous) dual. To distinguish the ordinary dual space from the continuous dual space, the former is sometimes called the . In finite dimensions, every linear functional is continuous, so the continuous dual is the same as the algebraic dual, but in infinite dimensions the continuous dual is a proper subspace of the algebraic dual.

A linear functional on a (not necessarily locally convex) topological vector space is continuous if and only if there exists a continuous seminorm on such that <math>|f| \leq p.</math>

Characterizing closed subspaces

Continuous linear functionals have nice properties for analysis: a linear functional is continuous if and only if its kernel is closed, and a non-trivial continuous linear functional is an open map, even if the (topological) vector space is not complete.

Hyperplanes and maximal subspaces

A vector subspace <math>M</math> of <math>X</math> is called maximal if <math>M \subsetneq X</math> (meaning <math>M \subseteq X</math> and <math>M \neq X</math>) and does not exist a vector subspace <math>N</math> of <math>X</math> such that <math>M \subsetneq N \subsetneq X.</math> A vector subspace <math>M</math> of <math>X</math> is maximal if and only if it is the kernel of some non-trivial linear functional on <math>X</math> (that is, <math>M = \ker f</math> for some linear functional <math>f</math> on <math>X</math> that is not identically ). An affine hyperplane in <math>X</math> is a translate of a maximal vector subspace. By linearity, a subset <math>H</math> of <math>X</math> is a affine hyperplane if and only if there exists some non-trivial linear functional <math>f</math> on <math>X</math> such that <math>H = f^{-1}(1) = \{ x \in X : f(x) = 1 \}.</math>

If <math>f</math> is a linear functional and <math>s \neq 0</math> is a scalar then <math>f^{-1}(s) = s \left(f^{-1}(1)\right) = \left(\frac{1}{s} f\right)^{-1}(1).</math> This equality can be used to relate different level sets of <math>f.</math> Moreover, if <math>f \neq 0</math> then the kernel of <math>f</math> can be reconstructed from the affine hyperplane <math>H := f^{-1}(1)</math> by <math>\ker f = H - H.</math>

Relationships between multiple linear functionals

Any two linear functionals with the same kernel are proportional (i.e. scalar multiples of each other).

This fact can be generalized to the following theorem.

If is a non-trivial linear functional on with kernel , <math>x \in X</math> satisfies <math>f(x) = 1,</math> and is a balanced subset of , then <math>N \cap (x + U) = \varnothing</math> if and only if <math>|f(u)| < 1</math> for all <math>u \in U.</math>

Hahn–Banach theorem

Any (algebraic) linear functional on a vector subspace can be extended to the whole space; for example, the evaluation functionals described above can be extended to the vector space of polynomials on all of <math>\R.</math> However, this extension cannot always be done while keeping the linear functional continuous. The Hahn–Banach family of theorems gives conditions under which this extension can be done. For example,

Equicontinuity of families of linear functionals

Let be a topological vector space (TVS) with continuous dual space <math>X'.</math>

For any subset of <math>X',</math> the following are equivalent:

is equicontinuous;
is contained in the polar of some neighborhood of <math>0</math> in ;
the (pre)polar of is a neighborhood of <math>0</math> in ;

If is an equicontinuous subset of <math>X'</math> then the following sets are also equicontinuous:

the weak-* closure, the balanced hull, the convex hull, and the convex balanced hull.

Moreover, Alaoglu's theorem implies that the weak-* closure of an equicontinuous subset of <math>X'</math> is weak-* compact (and thus that every equicontinuous subset weak-* relatively compact).

Notes

Footnotes

Proofs

References

Bibliography