# GetWiki

*Banach space*

ARTICLE SUBJECTS

being →

database →

ethics →

fiction →

history →

internet →

language →

linux →

logic →

method →

news →

policy →

purpose →

religion →

science →

software →

truth →

unix →

wiki →

ARTICLE TYPES

essay →

feed →

help →

system →

wiki →

ARTICLE ORIGINS

critical →

forked →

imported →

original →

Banach space

[ temporary import ]

**please note:**

- the content below is remote from Wikipedia

- it has been imported raw for GetWiki

**Banach space**(pronounced {{IPA-pl|Ëˆbanax|}}) is a complete normed vector space. Thus, a Banach space is a vector space with a metric that allows the computation of vector length and distance between vectors and is complete in the sense that a Cauchy sequence of vectors always converges to a well defined limit that is within the space.Banach spaces are named after the Polish mathematician Stefan Banach, who introduced this concept and studied it systematically in 1920â€“1922 along with Hans Hahn and Eduard Helly.{{harvnb|Bourbaki|1987|loc=V.86}} Banach spaces originally grew out of the study of function spaces by Hilbert, FrÃ©chet, and Riesz earlier in the century. Banach spaces play a central role in functional analysis. In other areas of analysis, the spaces under study are often Banach spaces.

## Definition

A Banach space is a vector space {{mvar|X}} over the field {{math|**R**}} of real numbers, or over the field {{math|

**C**}} of complex numbers, which is equipped with a norm and which is complete with respect to that norm, that is to say, for every Cauchy sequence {{math|{

*xn*} }} in {{mvar|X}}, there exists an element {{mvar|x}} in {{mvar|X}} such that

lim_{ntoinfty}x_n=x,

or equivalently:
lim_{ntoinfty}left |x_n - x right |_X = 0.

The vector space structure allows one to relate the behavior of Cauchy sequences to that of converging series of vectors. A normed space {{mvar|X}} is a Banach space if and only if each absolutely convergent series in {{mvar|X}} converges,see Theorem 1.3.9, p. 20 in {{harvtxt|Megginson|1998}}.
sum_{n=1}^{infty} |v_n|_X < infty quad text{implies that} quad sum_{n=1}^{infty} v_n text{converges in} X.

Completeness of a normed space is preserved if the given norm is replaced by an equivalent one.All norms on a finite-dimensional vector space are equivalent. Every finite-dimensional normed space over {{math|**R**}} or {{math|

**C**}} is a Banach space.see Corollary 1.4.18, p. 32 in {{harvtxt|Megginson|1998}}.

## General theory

### Linear operators, isomorphisms

If {{mvar|X}} and {{mvar|Y}} are normed spaces over the same ground field {{math|**K**}}, the set of all continuous {{math|

**K**}}-linear maps {{math|

*T*:

*X*â†’

*Y*}} is denoted by {{math|

*B*(

*X*,

*Y*)}}. In infinite-dimensional spaces, not all linear maps are continuous. A linear mapping from a normed space {{mvar|X}} to another normed space is continuous if and only if it is bounded on the closed unit ball of {{mvar|X}}. Thus, the vector space {{math|

*B*(

*X*,

*Y*)}} can be given the operator norm

|T| = sup left {|Tx|_Y mid xin X, |x|_Xle 1 right }.

For {{mvar|Y}} a Banach space, the space {{math|*B*(

*X*,

*Y*)}} is a Banach space with respect to this norm.If {{mvar|X}} is a Banach space, the space {{math|

*B*(

*X*) {{=}}

*B*(

*X*,

*X*)}} forms a unital Banach algebra; the multiplication operation is given by the composition of linear maps.If {{mvar|X}} and {{mvar|Y}} are normed spaces, they are

**isomorphic normed spaces**if there exists a linear bijection {{math|

*T*:

*X*â†’

*Y*}} such that {{mvar|T}} and its inverse {{math|

*T*âˆ’1}} are continuous. If one of the two spaces {{mvar|X}} or {{mvar|Y}} is complete (or reflexive, separable, etc.) then so is the other space. Two normed spaces {{mvar|X}} and {{mvar|Y}} are

**isometrically isomorphic**if in addition, {{mvar|T}} is an isometry, i.e., {{math|{{!!}}

*T*(

*x*){{!!}} {{=}} {{!!}}

*x*{{!!}}}} for every {{mvar|x}} in {{mvar|X}}. The Banachâ€“Mazur distance {{math|

*d*(

*X*,

*Y*)}} between two isomorphic but not isometric spaces {{mvar|X}} and {{mvar|Y}} gives a measure of how much the two spaces {{mvar|X}} and {{mvar|Y}} differ.

### Basic notions

Every normed space {{mvar|X}} can be isometrically embedded in a Banach space. More precisely, there is a Banach space {{mvar|Y}} and an isometric mapping {{math|*T*:

*X*â†’

*Y*}} such that {{math|

*T*(

*X*)}} is dense in {{mvar|Y}}. If {{mvar|Z}} is another Banach space such that there is an isometric isomorphism from {{mvar|X}} onto a dense subset of {{mvar|Z}}, then {{mvar|Z}} is isometrically isomorphic to {{mvar|Y}}.This Banach space {{mvar|Y}} is the

**completion**of the normed space {{mvar|X}}. The underlying metric space for {{mvar|Y}} is the same as the metric completion of {{mvar|X}}, with the vector space operations extended from {{mvar|X}} to {{mvar|Y}}. The completion of {{mvar|X}} is often denoted by widehat X.The cartesian product {{math|

*X*Ã—

*Y*}} of two normed spaces is not canonically equipped with a norm. However, several equivalent norms are commonly used,see {{harvtxt|Banach|1932}}, p. 182. such as

|(x, y)|_1 = |x| + |y|, qquad |(x, y)|_infty = max (|x|, |y|)

and give rise to isomorphic normed spaces. In this sense, the product {{math|*X*Ã—

*Y*}} (or the direct sum {{math|

*X*âŠ•

*Y*}}) is complete if and only if the two factors are complete.If {{mvar|M}} is a closed linear subspace of a normed space {{mvar|X}}, there is a natural norm on the

**quotient space**{{math|

*X*â€‰/â€‰

*M*}},

| x + M| = inflimits_{m in M} |x+m|.

The quotient {{math|*X*â€‰/â€‰

*M*}} is a Banach space when {{mvar|X}} is complete.see pp. 17â€“19 in {{harvtxt|Carothers|2005}}. The

**quotient map**from {{mvar|X}} onto {{math|

*X*â€‰/â€‰

*M*}}, sending {{mvar|x}} in {{mvar|X}} to its class {{math|

*x*+

*M*}}, is linear, onto and has norm {{math|1}}, except when {{math|

*M*{{=}}

*X*}}, in which case the quotient is the null space.The closed linear subspace {{mvar|M}} of {{mvar|X}} is said to be a

**complemented subspace**of {{mvar|X}} if {{mvar|M}} is the range of a bounded linear projection {{mvar|P}} from {{mvar|X}} onto {{mvar|M}}. In this case, the space {{mvar|X}} is isomorphic to the direct sum of {{mvar|M}} and {{math|Ker(

*P*)}}, the kernel of the projection {{mvar|P}}.Suppose that {{mvar|X}} and {{mvar|Y}} are Banach spaces and that {{math|

*T*âˆˆ

*B*(

*X*,

*Y*)}}. There exists a

**canonical factorization**of {{mvar|T}} as

T = T_1 circ pi, T : X overset{pi}{longrightarrow} X / operatorname{Ker}(T) overset{T_1}{longrightarrow} Y

where the first map {{mvar|Ï€}} is the quotient map, and the second map {{math|*T*1}} sends every class {{math|

*x*+ Ker(

*T*)}} in the quotient to the image {{math|

*T*(

*x*)}} in {{mvar|Y}}. This is well defined because all elements in the same class have the same image. The mapping {{math|

*T*1}} is a linear bijection from {{math|

*X*â€‰/â€‰Ker(

*T*)}} onto the range {{math|

*T*(

*X*)}}, whose inverse need not be bounded.

### Classical spaces

Basic examplessee {{harvtxt|Banach|1932}}, pp. 11-12. of Banach spaces include: the {{math|*Lp*}} spaces and their special cases, the sequence spaces {{math|â„“

*p*}} that consist of scalar sequences indexed by {{math|

**N**}}; among them, the space {{math|â„“1}} of absolutely summable sequences and the space {{math|â„“2}} of square summable sequences; the space {{math|

*c*0}} of sequences tending to zero and the space {{math|â„“âˆž}} of bounded sequences; the space {{math|

*C*(

*K*)}} of continuous scalar functions on a compact Hausdorff space {{mvar|K}}, equipped with the max norm,

|f|_{C(K)} = max { |f(x)| : x in K }, quad f in C(K).

According to the Banachâ€“Mazur theorem, every Banach space is isometrically isomorphic to a subspace of some {{math|*C*(

*K*)}}.see {{harvtxt|Banach|1932}}, Th. 9 p. 185. For every separable Banach space {{mvar|X}}, there is a closed subspace {{mvar|M}} of {{math|â„“1}} such that {{math|

*X*â‰… â„“1/

*M*}}.see Theorem 6.1, p. 55 in {{harvtxt|Carothers|2005}}Any Hilbert space serves as an example of a Banach space. A Hilbert space {{mvar|H}} on {{math|

**K**{{=}}

**R**,

**C**}} is complete for a norm of the form

|x|_H = sqrt{langle x, x rangle},

where
langle cdot, cdot rangle : H times H to mathbf{K}

is the inner product, linear in its first argument that satisfies the following:
begin{align}

forall x, y in H: quad langle y, x rangle &= overline{langle x, y rangle}, forall x in H: quad langle x, x rangle &ge 0, langle x,x rangle = 0 Leftrightarrow x &= 0.end{align}For example, the space {{math|*L*2}} is a Hilbert space.The Hardy spaces, the Sobolev spaces are examples of Banach spaces that are related to {{math|

*Lp*}} spaces and have additional structure. They are important in different branches of analysis, Harmonic analysis and Partial differential equations among others.

### Banach algebras

A**Banach algebra**is a Banach space {{mvar|A}} over {{math|

**K**{{=}}

**R**}} or {{math|

**C**}}, together with a structure of algebra over {{math|

**K**}}, such that the product map

*A*Ã—

*A*âˆ‹ {{math|(

*a*,

*b*) â†¦

*ab*âˆˆ

*A*}} is continuous. An equivalent norm on {{mvar|A}} can be found so that {{math|{{!!}}

*ab*{{!!}} â‰¤ {{!!}}

*a*{{!!}} {{!!}}

*b*{{!!}}}} for all {{math|

*a*,

*b*âˆˆ

*A*}}.

#### Examples

- The Banach space {{math|
*C*(*K*)}}, with the pointwise product, is a Banach algebra. - The disk algebra {{math|
*A*(**D**)}} consists of functions holomorphic in the open unit disk {{math|**D**âŠ‚**C**}} and continuous on its closure: {{math|{{overline|**D**}}}}. Equipped with the max norm on {{math|{{overline|**D**}}}}, the disk algebra {{math|*A*(**D**)}} is a closed subalgebra of {{math|*C*({{overline|**D**}})}}. - The Wiener algebra {{math|
*A*(**T**)}} is the algebra of functions on the unit circle {{math|**T**}} with absolutely convergent Fourier series. Via the map associating a function on {{math|**T**}} to the sequence of its Fourier coefficients, this algebra is isomorphic to the Banach algebra {{math|â„“1(**Z**)}}, where the product is the convolution of sequences. - For every Banach space {{mvar|X}}, the space {{math|
*B*(*X*)}} of bounded linear operators on {{mvar|X}}, with the composition of maps as product, is a Banach algebra. - A C-algebra is a complex Banach algebra {{mvar|A}} with an antilinear involution {{math|
*a*â†¦*a*âˆ—}} such that {{math|{{!!}}*a*âˆ—*a*{{!!}} {{=}} {{!!}}*a*{{!!}}2}}. The space {{math|*B*(*H*)}} of bounded linear operators on a Hilbert space {{mvar|H}} is a fundamental example of C-algebra. The Gelfandâ€“Naimark theorem states that every C-algebra is isometrically isomorphic to a C-subalgebra of some {{math|*B*(*H*)}}. The space {{math|*C*(*K*)}} of complex continuous functions on a compact Hausdorff space {{mvar|K}} is an example of commutative C-algebra, where the involution associates to every function {{math|*f*}} its complex conjugate {{math|{{overline|*f*}}}}.

### Dual space

If {{mvar|X}} is a normed space and {{math|**K**}} the underlying field (either the real or the complex numbers), the

**continuous dual space**is the space of continuous linear maps from {{mvar|X}} into {{math|

**K**}}, or

**continuous linear functionals**. The notation for the continuous dual is {{math|

*X*â€‰â€² {{=}}

*B*(

*X*,

**K**)}} in this article.Several books about functional analysis use the notation {{math|

*X*âˆ—}} for the continuous dual, for example {{harvtxt|Carothers|2005}}, {{harvtxt|Lindenstrauss|Tzafriri|1977}}, {{harvtxt|Megginson|1998}}, {{harvtxt|Ryan|2002}}, {{harvtxt|Wojtaszczyk|1991}}. Since {{math|

**K**}} is a Banach space (using the absolute value as norm), the dual {{math|

*X*â€‰â€²}} is a Banach space, for every normed space {{mvar|X}}.The main tool for proving the existence of continuous linear functionals is the Hahnâ€“Banach theorem.

**Hahnâ€“Banach theorem.**Let {{mvar|X}} be a vector space over the field {{math|

**K**{{=}}

**R**,

**C**}}. Let further * {{math|

*Y*âŠ†

*X*}} be a linear subspace, * {{math|

*p*:

*X*â†’

**R**}} be a sublinear function and * {{math|

*f*:

*Y*â†’

**K**}} be a linear functional so that {{math|Re(

*f*(

*y*)) â‰¤

*p*(

*y*)}} for all {{mvar|y}} in {{mvar|Y}}. Then, there exists a linear functional {{math|

*F*:

*X*â†’

**K**}} so that

F|_Y=f, quad text{and} quad forall xin X, operatorname{Re}(F(x))leq p(x).

*f*}} on {{mvar|X}} such that

f(x) = |x|_X, quad | f |_{X'} le 1.

When {{mvar|x}} is not equal to the {{math|**0**}} vector, the functional {{math|

*f*}} must have norm one, and is called a

**norming functional**for {{mvar|x}}.The Hahnâ€“Banach separation theorem states that two disjoint non-empty convex sets in a real Banach space, one of them open, can be separated by a closed affine hyperplane. The open convex set lies strictly on one side of the hyperplane, the second convex set lies on the other side but may touch the hyperplane.see also Theorem 2.2.26, p. 179 in {{harvtxt|Megginson|1998}}A subset {{mvar|S}} in a Banach space {{mvar|X}} is

**total**if the linear span of {{mvar|S}} is dense in {{mvar|X}}. The subset {{mvar|S}} is total in {{mvar|X}} if and only if the only continuous linear functional that vanishes on {{mvar|S}} is the {{math|

**0**}} functional: this equivalence follows from the Hahnâ€“Banach theorem.If {{mvar|X}} is the direct sum of two closed linear subspaces {{mvar|M}} and {{mvar|N}}, then the dual {{math|

*X*â€‰â€²}} of {{mvar|X}} is isomorphic to the direct sum of the duals of {{mvar|M}} and {{mvar|N}}.see p. 19 in {{harvtxt|Carothers|2005}}. If {{mvar|M}} is a closed linear subspace in {{mvar|X}}, one can associate the

*orthogonal of*{{mvar|M}} in the dual,

M^perp = left { x' in X' : x'(m) = 0, forall m in M right }.

The orthogonal {{math|*M*âŠ¥}} is a closed linear subspace of the dual. The dual of {{mvar|M}} is isometrically isomorphic to {{math|

*X*â€‰â€²â€‰/â€‰

*M*âŠ¥}}. The dual of {{math|

*X*â€‰/â€‰

*M*}} is isometrically isomorphic to {{math|

*M*âŠ¥}}.Theorems 1.10.16, 1.10.17 pp.94â€“95 in {{harvtxt|Megginson|1998}}The dual of a separable Banach space need not be separable, but:

**Theorem.**Theorem 1.12.11, p. 112 in {{harvtxt|Megginson|1998}} Let {{mvar|X}} be a normed space. If {{math|

*X*â€‰â€²}} is separable, then {{mvar|X}} is separable.

*X*â€‰â€²}} is separable, the above criterion for totality can be used for proving the existence of a countable total subset in {{mvar|X}}.

#### Weak topologies

The**weak topology**on a Banach space {{mvar|X}} is the coarsest topology on {{mvar|X}} for which all elements {{math|

*x*â€‰â€²}} in the continuous dual space {{math|

*X*â€‰â€²}} are continuous. The norm topology is therefore finer than the weak topology. It follows from the Hahnâ€“Banach separation theorem that the weak topology is Hausdorff, and that a norm-closed convex subset of a Banach space is also weakly closed.Theorem 2.5.16, p. 216 in {{harvtxt|Megginson|1998}}. A norm-continuous linear map between two Banach spaces {{mvar|X}} and {{mvar|Y}} is also

**weakly continuous**, i.e., continuous from the weak topology of {{mvar|X}} to that of {{mvar|Y}}.see II.A.8, p. 29 in {{harvtxt|Wojtaszczyk|1991}}If {{mvar|X}} is infinite-dimensional, there exist linear maps which are not continuous. The space {{math|

*X*âˆ—}} of all linear maps from {{mvar|X}} to the underlying field {{math|

**K**}} (this space {{math|

*X*âˆ—}} is called the algebraic dual space, to distinguish it from {{math|

*X*â€‰â€²}}) also induces a topology on {{mvar|X}} which is finer than the weak topology, and much less used in functional analysis.On a dual space {{math|

*X*â€‰â€²}}, there is a topology weaker than the weak topology of {{math|

*X*â€‰â€²}}, called

**weak* topology**. It is the coarsest topology on {{math|

*X*â€‰â€²}} for which all evaluation maps {{math|

*x*â€²â€‰âˆˆâ€‰

*X*â€‰â€² â†’

*x*â€²(

*x*),

*x*â€‰âˆˆâ€‰

*X*}}, are continuous. Its importance comes from the Banachâ€“Alaoglu theorem.

**Banachâ€“Alaoglu Theorem.**Let {{mvar|X}} be a normed vector space. Then the closed unit ball {{math|

*B*â€‰â€² {{=}} {

*x*â€² âˆˆ

*X*â€‰â€² : {{!!}}

*x*â€²{{!!}} â‰¤ 1} }} of the dual space is compact in the weak* topology.

*B*â€‰â€²}} of the dual is a metrizable compact in the weak* topology.see Theorem 2.6.23, p. 231 in {{harvtxt|Megginson|1998}}.

#### Examples of dual spaces

The dual of {{math|*c*0}} is isometrically isomorphic to {{math|â„“1}}: for every bounded linear functional {{math|â€‰

*f*â€‰}} on {{math|

*c*0}}, there is a unique element {{math|

*y*{{=}} {

*yn*} âˆˆ â„“1}} such that

f(x) = sum_{n in mathbf{N}} x_n y_n, qquad x = {x_n} in c_0, text{and} |f|_{(c_0)'} = |y|_{ell_1}.

The dual of {{math|â„“1}} is isometrically isomorphic to {{math|â„“âˆž}}. The dual of {{math|*Lp*([0, 1])}} is isometrically isomorphic to {{math|

*Lq*([0, 1])}} when {{math|1 â‰¤

*p*< âˆž}} and {{math|â€‰{{sfrac|1|

*p*}} + {{sfrac|1|

*q*}} {{=}} 1}}.For every vector {{mvar|y}} in a Hilbert space {{mvar|H}}, the mapping

x in H to f_y(x) = langle x, y rangle

defines a continuous linear functional {{math|â€‰*fy*â€‰}} on {{mvar|H}}. The Riesz representation theorem states that every continuous linear functional on {{mvar|H}} is of the form {{math|â€‰

*fy*â€‰}} for a uniquely defined vector {{mvar|y}} in {{mvar|H}}. The mapping {{math|

*y*âˆˆ

*H*â†’ â€‰

*fy*â€‰}} is an antilinear isometric bijection from {{mvar|H}} onto its dual {{math|

*H*â€‰â€²}}. When the scalars are real, this map is an isometric isomorphism.When {{mvar|K}} is a compact Hausdorff topological space, the dual {{math|

*M*(

*K*)}} of {{math|

*C*(

*K*)}} is the space of Radon measures in the sense of Bourbaki.see N. Bourbaki, (2004), "Integration I", Springer Verlag, {{ISBN|3-540-41129-1}}. The subset {{math|

*P*(

*K*)}} of {{math|

*M*(

*K*)}} consisting of non-negative measures of mass 1 (probability measures) is a convex w*-closed subset of the unit ball of {{math|

*M*(

*K*)}}. The extreme points of {{math|

*P*(

*K*)}} are the Dirac measures on {{mvar|K}}. The set of Dirac measures on {{mvar|K}}, equipped with the w*-topology, is homeomorphic to {{mvar|K}}.

**Banachâ€“Stone Theorem.**If {{mvar|K}} and {{mvar|L}} are compact Hausdorff spaces and if {{math|

*C*(

*K*)}} and {{math|

*C*(

*L*)}} are isometrically isomorphic, then the topological spaces {{mvar|K}} and {{mvar|L}} are homeomorphic.see also {{harvtxt|Banach|1932}}, p. 170 for metrizable {{mvar|K}} and {{mvar|L}}.

*C*(

*K*)}} and {{math|

*C*(

*L*)}} is {{math|< 2}}. The theorem is no longer true when the distance is {{math|{{=}} 2}}.JOURNAL, H. B., Cohen, A bound-two isomorphism between {{math,

*C*(

*X*), Banach spaces |journal=Proc. Amer. Math. Soc. |volume=50 |year=1975 |pages=215â€“217 |doi=10.1090/s0002-9939-1975-0380379-5}}In the commutative Banach algebra {{math|

*C*(

*K*)}}, the maximal ideals are precisely kernels of Dirac mesures on {{mvar|K}},

I_x = ker delta_x = {f in C(K) : f(x) = 0}, quad x in K.

More generally, by the Gelfandâ€“Mazur theorem, the maximal ideals of a unital commutative Banach algebra can be identified with its charactersâ€”not merely as sets but as topological spaces: the former with the hull-kernel topology and the latter with the w*-topology. In this identification, the maximal ideal space can be viewed as a w*-compact subset of the unit ball in the dual {{math|*A*â€‰â€²}}.

**Theorem.**If {{mvar|K}} is a compact Hausdorff space, then the maximal ideal space {{math|Îž}} of the Banach algebra {{math|

*C*(

*K*)}} is homeomorphic to {{mvar|K}}.JOURNAL, Eilenberg, Samuel, Banach Space Methods in Topology, Annals of Mathematics, 1942, 43, 3, 568, 10.2307/1968812,

*C*(

*K*)}} for some compact Hausdorff space {{mvar|K}}. However, this statement holds if one places {{math|

*C*(

*K*)}} in the smaller category of commutative C*-algebras. Gelfand's representation theorem for commutative C*-algebras states that every commutative unital

*C**-algebra {{mvar|A}} is isometrically isomorphic to a {{math|

*C*(

*K*)}} space.See for example BOOK, W., Arveson, 1976, An Invitation to C*-Algebra, Springer-Verlag, 0-387-90176-0, The Hausdorff compact space {{mvar|K}} here is again the maximal ideal space, also called the spectrum of {{mvar|A}} in the C*-algebra context.

#### Bidual

If {{mvar|X}} is a normed space, the (continuous) dual {{math|*X*â€‰â€²â€²}} of the dual {{math|

*X*â€‰â€²}} is called

**bidual**, or

**second dual**of {{mvar|X}}. For every normed space {{mvar|X}}, there is a natural map,

begin{cases} F_X : X to X'' F_X(x) (f) = f(x) & forall x in X, forall f in X'end{cases}

This defines {{math|*FX*(

*x*)}} as a continuous linear functional on {{math|

*X*â€‰â€²}}, i.e., an element of {{math|

*X*â€‰â€²â€²}}. The map {{math|

*FX*:

*x*â†’

*FX*(

*x*)}} is a linear map from {{mvar|X}} to {{math|

*X*â€‰â€²â€²}}. As a consequence of the existence of a norming functional {{math|

*f*}} for every {{mvar|x}} in {{mvar|X}}, this map {{math|

*FX*}} is isometric, thus injective.For example, the dual of {{math|

*X*{{=}}

*c*0}} is identified with {{math|â„“1}}, and the dual of {{math|â„“1}} is identified with {{math|â„“âˆž}}, the space of bounded scalar sequences. Under these identifications, {{math|

*FX*}} is the inclusion map from {{math|

*c*0}} to {{math|â„“âˆž}}. It is indeed isometric, but not onto.If {{math|

*FX*}} is surjective, then the normed space {{mvar|X}} is called

**reflexive**(see below). Being the dual of a normed space, the bidual {{math|

*X*â€‰â€²â€²}} is complete, therefore, every reflexive normed space is a Banach space.Using the isometric embedding {{math|

*FX*}}, it is customary to consider a normed space {{mvar|X}} as a subset of its bidual. When {{mvar|X}} is a Banach space, it is viewed as a closed linear subspace of {{math|

*X*â€‰â€²â€²}}. If {{mvar|X}} is not reflexive, the unit ball of {{mvar|X}} is a proper subset of the unit ball of {{math|

*X*â€‰â€²â€²}}. The Goldstine theorem states that the unit ball of a normed space is weakly*-dense in the unit ball of the bidual. In other words, for every {{math|

*x*â€‰â€²â€²}} in the bidual, there exists a net {{math|{

*xj*} }} in {{mvar|X}} so that

sup_j |x_j| le |x

The net may be replaced by a weakly*-convergent sequence when the dual {{math|*|, x*(f) = lim_j f(x_j), quad f in X'.*X*â€‰â€²}} is separable. On the other hand, elements of the bidual of {{math|â„“1}} that are not in {{math|â„“1}} cannot be weak*-limit of

*sequences*in {{math|â„“1}}, since {{math|â„“1}} is weakly sequentially complete.

### Banach's theorems

Here are the main general results about Banach spaces that go back to the time of Banach's book ({{harvtxt|Banach|1932}}) and are related to the Baire category theorem. According to this theorem, a complete metric space (such as a Banach space, a FrÃ©chet space or an F-space) cannot be equal to a union of countably many closed subsets with empty interiors. Therefore, a Banach space cannot be the union of countably many closed subspaces, unless it is already equal to one of them; a Banach space with a countable Hamel basis is finite-dimensional.**Banachâ€“Steinhaus Theorem.**Let {{mvar|X}} be a Banach space and {{mvar|Y}} be a normed vector space. Suppose that {{mvar|F}} is a collection of continuous linear operators from {{mvar|X}} to {{mvar|Y}}. The uniform boundedness principle states that if for all {{mvar|x}} in {{mvar|X}} we have {{math|sup

*T*âˆˆ

*F*{{!!}}

*T*(

*x*){{!!}}

*Y*< âˆž}}, then {{math|sup

*T*âˆˆ

*F*{{!!}}

*T*{{!!}}

*Y*< âˆž}}.

**0**}} in {{mvar|X}} such that all {{mvar|T}} in {{mvar|F}} are uniformly bounded on {{mvar|U}},

sup_{T in F} sup_{x in U} ; |T(x)|_Y < infty.

**The Open Mapping Theorem.**Let {{mvar|X}} and {{mvar|Y}} be Banach spaces and {{math|

*T*:

*X*â†’

*Y*}} be a surjective continuous linear operator, then {{mvar|T}} is an open map.

**Corollary.**Every one-to-one bounded linear operator from a Banach space onto a Banach space is an isomorphism.

**The First Isomorphism Theorem for Banach spaces.**Suppose that {{mvar|X}} and {{mvar|Y}} are Banach spaces and that {{math|

*T*âˆˆ

*B*(

*X*,

*Y*)}}. Suppose further that the range of {{mvar|T}} is closed in {{mvar|Y}}. Then {{math|

*X*/â€‰Ker(

*T*)}} is isomorphic to {{math|

*T*(

*X*)}}.

*Banach isomorphism theorem*and of the canonical factorization of bounded linear maps.

**Corollary.**If a Banach space {{mvar|X}} is the internal direct sum of closed subspaces {{math|

*M*1, ...,

*Mn*}}, then {{mvar|X}} is isomorphic to {{math|

*M*1 âŠ• ... âŠ•

*Mn*}}.

*M*1 âŠ• ... âŠ•

*Mn*}} onto {{mvar|X}} sending {{math|(

*m*1, ...,

*mn*)}} to the sum {{math|

*m*1 + ... +

*mn*}}.

**The Closed Graph Theorem.**Let {{math|

*T*:

*X*â†’

*Y*}} be a linear mapping between Banach spaces. The graph of {{mvar|T}} is closed in {{math|

*X*Ã—

*Y*}} if and only if {{mvar|T}} is continuous.

### Reflexivity

The normed space {{mvar|X}} is called**reflexive**when the natural map

begin{cases} F_X : X to X'' F_X(x) (f) = f(x) & forall x in X, forall f in X'end{cases}

is surjective. Reflexive normed spaces are Banach spaces.
**Theorem.**If {{mvar|X}} is a reflexive Banach space, every closed subspace of {{mvar|X}} and every quotient space of {{mvar|X}} are reflexive.

**Theorem.**If {{mvar|X}} is a Banach space, then {{mvar|X}} is reflexive if and only if {{math|

*X*â€‰â€²}} is reflexive.

**Corollary.**Let {{mvar|X}} be a reflexive Banach space. Then {{mvar|X}} is separable if and only if {{math|

*X*â€‰â€²}} is separable.

*Y*â€‰â€²}} of a Banach space {{mvar|Y}} is separable, then {{mvar|Y}} is separable. If {{mvar|X}} is reflexive and separable, then the dual of {{math|

*X*â€‰â€²}} is separable, so {{math|

*X*â€‰â€²}} is separable.

**Theorem.**Suppose that {{math|

*X*1, ...,

*Xn*}} are normed spaces and that {{math|

*X*{{=}}

*X*1 âŠ• ... âŠ•

*Xn*}}. Then {{mvar|X}} is reflexive if and only if each {{math|

*Xj*}} is reflexive.

*Lp*}} spaces are reflexive when {{math|1 <

*p*< âˆž}}. More generally, uniformly convex spaces are reflexive, by the Milmanâ€“Pettis theorem. The spaces {{math|

*c*0, â„“1,

*L*1([0, 1]),

*C*([0, 1])}} are not reflexive. In these examples of non-reflexive spaces {{mvar|X}}, the bidual {{math|

*X*â€‰â€²â€²}} is "much larger" than {{mvar|X}}. Namely, under the natural isometric embedding of {{mvar|X}} into {{math|

*X*â€‰â€²â€²}} given by the Hahnâ€“Banach theorem, the quotient {{math|

*X*â€‰â€²â€²â€‰/â€‰

*X*}} is infinite-dimensional, and even nonseparable. However, Robert C. James has constructed an exampleJOURNAL, R. C. James, A non-reflexive Banach space isometric with its second conjugate space, Proc. Natl. Acad. Sci. U.S.A., 37, 174â€“177, 1951, 10.1073/pnas.37.3.174, 1063327, 16588998, 1951PNAS...37..174J, of a non-reflexive space, usually called "

*the James space*" and denoted by

*J*,see {{harvtxt|Lindenstrauss|Tzafriri|1977}}, p. 25. such that the quotient {{math|

*J*â€‰â€²â€²â€‰/â€‰

*J*}} is one-dimensional. Furthermore, this space {{mvar|J}} is isometrically isomorphic to its bidual.

**Theorem.**A Banach space {{mvar|X}} is reflexive if and only if its unit ball is compact in the weak topology.

**R**}}, every continuous linear functional {{math|

*f*}} in {{math|

*X*â€‰â€²}} attains its maximum {{math|{{!!}}

*f*{{!!}}}} on the unit ball of {{mvar|X}}. The following theorem of Robert C. James provides a converse statement.

**James' Theorem.**For a Banach space the following two properties are equivalent: * {{mvar|X}} is reflexive. * for all {{math|

*f*}} in {{math|

*X*â€‰â€²}} there exists {{mvar|x}} in {{mvar|X}} with {{math|{{!!}}

*x*{{!!}} â‰¤ 1}}, so that {{math|

*f*(

*x*) {{=}} {{!!}}

*f*{{!!}}.}}

*norm-attaining*. However, the Bishopâ€“Phelps theoremJOURNAL, bishop, See E., Phelps, R., 1961, A proof that every Banach space is subreflexive, Bull. Amer. Math. Soc., 67, 97â€“98, 10.1090/s0002-9904-1961-10514-4, states that norm-attaining functionals are norm dense in the dual {{math|

*X*â€‰â€²}} of {{mvar|X}}.

### Weak convergences of sequences

A sequence {{math|{*xn*} }} in a Banach space {{mvar|X}} is

**weakly convergent**to a vector {{math|

*x*âˆˆ

*X*}} if {{math|

*f*(

*xn*)}} converges to {{math|

*f*(

*x*)}} for every continuous linear functional {{math|

*f*}} in the dual {{math|

*X*â€‰â€²}}. The sequence {{math|{

*xn*} }} is a

**weakly Cauchy sequence**if {{math|

*f*(

*xn*)}} converges to a scalar limit {{math|

*L*(

*f*)}}, for every {{math|

*f*}} in {{math|

*X*â€‰â€²}}. A sequence {{math|{

*fn*} }} in the dual {{math|

*X*â€‰â€²}} is

**weakly* convergent**to a functional {{math|

*f*âˆˆ

*X*â€‰â€²}} if {{math|

*fn*(

*x*)}} converges to {{math|

*f*(

*x*)}} for every {{mvar|x}} in {{mvar|X}}. Weakly Cauchy sequences, weakly convergent and weakly* convergent sequences are norm bounded, as a consequence of the Banachâ€“Steinhaus theorem.When the sequence {{math|{

*xn*} }} in {{mvar|X}} is a weakly Cauchy sequence, the limit {{mvar|L}} above defines a bounded linear functional on the dual {{math|

*X*â€‰â€²}}, i.e., an element {{mvar|L}} of the bidual of {{mvar|X}}, and {{mvar|L}} is the limit of {{math|{

*xn*} }} in the weak*-topology of the bidual. The Banach space {{mvar|X}} is

**weakly sequentially complete**if every weakly Cauchy sequence is weakly convergent in {{mvar|X}}. It follows from the preceding discussion that reflexive spaces are weakly sequentially complete.

**Theorem.**see III.C.14, p. 140 in {{harvtxt|Wojtaszczyk|1991}}. For every measure {{mvar|Î¼}}, the space {{math|

*L*1(

*Î¼*)}} is weakly sequentially complete.

**0**}} vector. The unit vector basis of {{math|â„“

*p*, 1 <

*p*< âˆž}}, or of {{math|

*c*0}}, is another example of a

**weakly null sequence**, i.e., a sequence that converges weakly to {{math|

**0**}}. For every weakly null sequence in a Banach space, there exists a sequence of convex combinations of vectors from the given sequence that is norm-converging to {{math|

**0**}}.see Corollary 2, p. 11 in {{harvtxt|Diestel|1984}}.The unit vector basis of {{math|â„“1}} is not weakly Cauchy. Weakly Cauchy sequences in {{math|â„“1}} are weakly convergent, since {{math|

*L*1}}-spaces are weakly sequentially complete. Actually, weakly convergent sequences in {{math|â„“1}} are norm convergent.see p. 85 in {{harvtxt|Diestel|1984}}. This means that {{math|â„“1}} satisfies Schur's property.

#### Results involving the {{math|â„“1}} basis

Weakly Cauchy sequences and the {{math|â„“1}} basis are the opposite cases of the dichotomy established in the following deep result of H. P. Rosenthal.JOURNAL, Rosenthal, Haskell P, 1974, A characterization of Banach spaces containing â„“1, Proc. Natl. Acad. Sci. U.S.A., 71, 2411â€“2413, 10.1073/pnas.71.6.2411, math.FA/9210205v1, 1974PNAS...71.2411R, Rosenthal's proof is for real scalars. The complex version of the result is due to L. Dor, in JOURNAL, Dor, Leonard E, 1975, On sequences spanning a complex â„“1 space, Proc. Amer. Math. Soc., 47, 515â€“516, 10.1090/s0002-9939-1975-0358308-x,**Theorem.**see p. 201 in {{harvtxt|Diestel|1984}}. Let {{math|{

*xn*} }} be a bounded sequence in a Banach space. Either {{math|{

*xn*} }} has a weakly Cauchy subsequence, or it admits a subsequence equivalent to the standard unit vector basis of {{math|â„“1}}.

**Theorem.**{{citation|last1 = Odell| first1 = Edward W. | last2 = Rosenthal | first2 = Haskell P. | title = A double-dual characterization of separable Banach spaces containing â„“1 | journal = Israel J. Math. | volume = 20 | year = 1975 |pages = 375â€“384 | doi=10.1007/bf02760341}}. Let {{mvar|X}} be a separable Banach space. The following are equivalent: *The space {{mvar|X}} does not contain a closed subspace isomorphic to {{math|â„“1}}. *Every element of the bidual {{math|

*X*â€‰â€²â€²}} is the weak*-limit of a sequence {{math|{

*xn*} }} in {{mvar|X}}.

*B*â€‰â€²â€²}} of {{math|

*X*â€‰â€²â€²}} is weak*-limit of a net in the unit ball of {{mvar|X}}. When {{mvar|X}} does not contain {{math|â„“1}}, every element of {{math|

*B*â€‰â€²â€²}} is weak*-limit of a

*sequence*in the unit ball of {{mvar|X}}.Odell and Rosenthal, Sublemma p. 378 and Remark p. 379.When the Banach space {{mvar|X}} is separable, the unit ball of the dual {{math|

*X*â€‰â€²}}, equipped with the weak*-topology, is a metrizable compact space {{mvar|K}}, and every element {{math|

*x*â€‰â€²â€²}} in the bidual {{math|

*X*â€‰â€²â€²}} defines a bounded function on {{mvar|K}}:

x' in K mapsto x

This function is continuous for the compact topology of {{mvar|K}} if and only if {{math|*(x'), quad left |x*(x') right | le left |x'' right |.*x*â€‰â€²â€²}} is actually in {{mvar|X}}, considered as subset of {{math|

*X*â€‰â€²â€²}}. Assume in addition for the rest of the paragraph that {{mvar|X}} does not contain {{math|â„“1}}. By the preceding result of Odell and Rosenthal, the function {{math|

*x*â€‰â€²â€²}} is the pointwise limit on {{mvar|K}} of a sequence {{math|{

*xn*} âŠ‚

*X*}} of continuous functions on {{mvar|K}}, it is therefore a first Baire class function on {{mvar|K}}. The unit ball of the bidual is a pointwise compact subset of the first Baire class on {{mvar|K}}.for more on pointwise compact subsets of the Baire class, see {{citation|last1=Bourgain |first1=Jean |author1-link=Jean Bourgain |last2=Fremlin |first2=D. H. |last3=Talagrand |first3=Michel |title=Pointwise Compact Sets of Baire-Measurable Functions|journal=Am. J. Math. |volume=100 |year=1978 |pages=845â€“886 |jstor=2373913 |doi=10.2307/2373913}}.

#### Sequences, weak and weak* compactness

When {{mvar|X}} is separable, the unit ball of the dual is weak*-compact by Banachâ€“Alaoglu and metrizable for the weak* topology, hence every bounded sequence in the dual has weakly* convergent subsequences. This applies to separable reflexive spaces, but more is true in this case, as stated below.The weak topology of a Banach space {{mvar|X}} is metrizable if and only if {{mvar|X}} is finite-dimensional.see Proposition 2.5.14, p. 215 in {{harvtxt|Megginson|1998}}. If the dual {{math|*X*â€‰â€²}} is separable, the weak topology of the unit ball of {{mvar|X}} is metrizable. This applies in particular to separable reflexive Banach spaces. Although the weak topology of the unit ball is not metrizable in general, one can characterize weak compactness using sequences.

**Eberleinâ€“Å mulian theorem**.see for example p. 49, II.C.3 in {{harvtxt|Wojtaszczyk|1991}}. A set {{mvar|A}} in a Banach space is relatively weakly compact if and only if every sequence {{math|{

*an*} }} in {{mvar|A}} has a weakly convergent subsequence.

## Schauder bases

A**Schauder basis**in a Banach space {{mvar|X}} is a sequence {{math|{

*en*}

*n*â€‰â‰¥â€‰0}} of vectors in

*X*with the property that for every vector {{mvar|x}} in {{mvar|X}}, there exist

*uniquely*defined scalars {{math|{

*xn*}

*n*â€‰â‰¥â€‰0}} depending on {{mvar|x}}, such that

x = sum_{n=0}^{infty} x_n e_n, quad textit{i.e.,} quad x = lim_n P_n(x), P_n(x) := sum_{k=0}^n x_k e_k.

Banach spaces with a Schauder basis are necessarily separable, because the countable set of finite linear combinations with rational coefficients (say) is dense.It follows from the Banachâ€“Steinhaus theorem that the linear mappings {{math|{*Pn*} }}are uniformly bounded by some constant {{mvar|C}}. Let {{math|{

*e*{{su|p=âˆ—|b=

*n*}}} }} denote the coordinate functionals which assign to every {{mvar|x}} in {{mvar|X}} the coordinate {{math|

*xn*}} of {{mvar|x}} in the above expansion. They are called

**biorthogonal functionals**. When the basis vectors have norm {{math|1}}, the coordinate functionals {{math|{

*e*{{su|p=âˆ—|b=

*n*}}} }} have norm {{math| â‰¤ 2

*C*}} in the dual of {{mvar|X}}.Most classical separable spaces have explicit bases. The Haar system {{math|{

*hn*} }} is a basis for {{math|

*Lp*([0, 1]), 1 â‰¤

*p*< âˆž}}. The trigonometric system is a basis in {{math|

*Lp*(

**T**)}} when {{math|1 <

*p*< âˆž}}. The Schauder system is a basis in the space {{math|

*C*([0, 1])}}.see {{harvtxt|Lindenstrauss|Tzafriri|1977}} p. 3. The question of whether the disk algebra {{math|

*A*(

**D**)}} has a basisthe question appears p. 238, Â§3 in Banach's book, {{harvtxt|Banach|1932}}. remained open for more than forty years, until BoÄkarev showed in 1974 that {{math|

*A*(

**D**)}} admits a basis constructed from the Franklin system.see S. V. BoÄkarev, "Existence of a basis in the space of functions analytic in the disc, and some properties of Franklin's system". (Russian) Mat. Sb. (N.S.) 95(137) (1974), 3â€“18, 159.Since every vector {{mvar|x}} in a Banach space {{mvar|X}} with a basis is the limit of {{math|

*Pn*(

*x*)}}, with {{math|

*Pn*}} of finite rank and uniformly bounded, the space {{mvar|X}} satisfies the bounded approximation property. The first example by Enflo of a space failing the approximation property was at the same time the first example of a separable Banach space without a Schauder basis.see JOURNAL, Enflo, P., 1973, A counterexample to the approximation property in Banach spaces,weblink PDF, Acta Math., 130, 309â€“317, 10.1007/bf02392270, Robert C. James characterized reflexivity in Banach spaces with a basis: the space {{mvar|X}} with a Schauder basis is reflexive if and only if the basis is both shrinking and boundedly complete.see R.C. James, "Bases and reflexivity of Banach spaces". Ann. of Math. (2) 52, (1950). 518â€“527. See also {{harvtxt|Lindenstrauss|Tzafriri|1977}} p. 9. In this case, the biorthogonal functionals form a basis of the dual of {{mvar|X}}.

## Tensor product

(File:Tensor-diagramB.jpg|thumb)Let {{mvar|X}} and {{mvar|Y}} be two {{math|**K**}}-vector spaces. The tensor product {{math|

*X*âŠ—

*Y*}} of {{mvar|X}} and {{mvar|Y}} is a {{math|

**K**}}-vector space {{mvar|Z}} with a bilinear mapping {{math|

*T*:

*X*Ã—

*Y*â†’

*Z*}} which has the following universal property:

If {{math|

The image under {{mvar|T}} of a couple {{math|(*T*1 :*X*Ã—*Y*â†’*Z*1}} is any bilinear mapping into a {{math|**K**}}-vector space {{math|*Z*1}}, then there exists a unique linear mapping {{math|*f*:*Z*â†’*Z*1}} such that {{math|*T*1 {{=}}*f*âˆ˜*T*}}.*x*,

*y*)}} in {{math|

*X*Ã—

*Y*}} is denoted by {{math|

*x*âŠ—

*y*}}, and called a

**simple tensor**. Every element {{mvar|z}} in {{math|

*X*âŠ—

*Y*}} is a finite sum of such simple tensors.There are various norms that can be placed on the tensor product of the underlying vector spaces, amongst others the projective cross norm and injective cross norm introduced by A. Grothendieck in 1955.see A. Grothendieck, "Produits tensoriels topologiques et espaces nuclÃ©aires". Mem. Amer. Math. Soc. 1955 (1955), no. 16, 140 pp., and A. Grothendieck, "RÃ©sumÃ© de la thÃ©orie mÃ©trique des produits tensoriels topologiques". Bol. Soc. Mat. SÃ£o Paulo 8 1953 1â€“79.In general, the tensor product of complete spaces is not complete again. When working with Banach spaces, it is customary to say that the

**projective tensor product**see chap. 2, p. 15 in {{harvtxt|Ryan|2002}}. of two Banach spaces {{mvar|X}} and {{mvar|Y}} is the

*completion*X widehat{otimes}_pi Y of the algebraic tensor product {{math|

*X*âŠ—

*Y*}} equipped with the projective tensor norm, and similarly for the

**injective tensor product**see chap. 3, p. 45 in {{harvtxt|Ryan|2002}}. X widehat{otimes}_varepsilon Y. Grothendieck proved in particular thatsee Example. 2.19, p. 29, and pp. 49â€“50 in {{harvtxt|Ryan|2002}}.

begin{align}

C(K) widehat{otimes}_varepsilon Y &simeq C(K, Y), L^1([0, 1]) widehat{otimes}_pi Y &simeq L^1([0, 1], Y),end{align}where {{mvar|K}} is a compact Hausdorff space, {{math|*C*(

*K*,

*Y*)}} the Banach space of continuous functions from {{mvar|K}} to {{mvar|Y}} and {{math|

*L*1([0, 1],

*Y*)}} the space of Bochner-measurable and integrable functions from {{math|[0, 1]}} to {{mvar|Y}}, and where the isomorphisms are isometric. The two isomorphisms above are the respective extensions of the map sending the tensor {{math|

*f*â€‰âŠ—â€‰

*y*}} to the vector-valued function {{math|

*s*âˆˆ

*K*â†’

*f*(

*s*)

*y*âˆˆ

*Y*}}.

### Tensor products and the approximation property

Let {{mvar|X}} be a Banach space. The tensor product X' widehat otimes_varepsilon X is identified isometrically with the closure in {{math|*B*(

*X*)}} of the set of finite rank operators. When {{mvar|X}} has the approximation property, this closure coincides with the space of compact operators on {{mvar|X}}.For every Banach space {{mvar|Y}}, there is a natural norm {{math|1}} linear map

Y widehatotimes_pi X to Y widehatotimes_varepsilon X

obtained by extending the identity map of the algebraic tensor product. Grothendieck related the approximation problem to the question of whether this map is one-to-one when {{mvar|Y}} is the dual of {{mvar|X}}. Precisely, for every Banach space {{mvar|X}}, the map
X' widehat otimes_pi X longrightarrow X' widehat otimes_varepsilon X

is one-to-one if and only if {{mvar|X}} has the approximation property.see Proposition 4.6, p. 74 in {{harvtxt|Ryan|2002}}.Grothendieck conjectured that X widehat{otimes}_pi Y and X widehat{otimes}_varepsilon Y must be different whenever {{mvar|X}} and {{mvar|Y}} are infinite-dimensional Banach spaces. This was disproved by Gilles Pisier in 1983.see Pisier, Gilles (1983), "Counterexamples to a conjecture of Grothendieck", Acta Math. **151**:181â€“208. Pisier constructed an infinite-dimensional Banach space {{mvar|X}} such that X widehat{otimes}_pi X and X widehat{otimes}_varepsilon X are equal. Furthermore, just as Enflo's example, this space {{mvar|X}} is a "hand-made" space that fails to have the approximation property. On the other hand, Szankowski proved that the classical space {{math|

*B*(â„“2)}} does not have the approximation property.see Szankowski, Andrzej (1981), "{{math|

*B*(

*H*)}} does not have the approximation property", Acta Math.

**147**: 89â€“108. Ryan claims that this result is due to Per Enflo, p. 74 in {{harvtxt|Ryan|2002}}.

## Some classification results

### Characterizations of Hilbert space among Banach spaces

A necessary and sufficient condition for the norm of a Banach space {{mvar|X}} to be associated to an inner product is the parallelogram identity:
forall x, y in X : qquad |x+y|^2 + |x-y|^2 = 2 left (|x|^2 + |y|^2 right).

It follows, for example, that the Lebesgue space {{math|*Lp*([0, 1])}} is a Hilbert space only when {{math|

*p*{{=}} 2}}. If this identity is satisfied, the associated inner product is given by the polarization identity. In the case of real scalars, this gives:

langle x, yrangle = tfrac{1}{4} left (|x+y|^2 - |x-y|^2 right ).

For complex scalars, defining the inner product so as to be {{math|**C**}}-linear in {{mvar|x}}, antilinear in {{mvar|y}}, the polarization identity gives:

langle x,yrangle = tfrac{1}{4} left( |x+y|^2 - |x-y|^2 + i left (|x+iy|^2 - |x-iy|^2 right )right).

To see that the parallelogram law is sufficient, one observes in the real case that {{math|}} is symmetric, and in the complex case, that it satisfies the Hermitian symmetry property and {{math| {{=}} *i*}}. The parallelogram law implies that {{math|}} is additive in {{mvar|x}}. It follows that it is linear over the rationals, thus linear by continuity.Several characterizations of spaces isomorphic (rather than isometric) to Hilbert spaces are available. The parallelogram law can be extended to more than two vectors, and weakened by the introduction of a two-sided inequality with a constant {{math|

*c*â‰¥ 1}}: KwapieÅ„ proved that if

c^{-2} sum_{k=1}^n left |x_k right |^2 le operatorname{Ave}_{pm} left | sum_{k=1}^n pm x_k right |^2 le c^2 sum_{k=1}^n left |x_k right |^2

for every integer {{mvar|n}} and all families of vectors {{math|{*x*1, ...,

*xn*} âŠ‚

*X*}}, then the Banach space {{mvar|X}} is isomorphic to a Hilbert space.see KwapieÅ„, S. (1970), "A linear topological characterization of inner-product spaces", Studia Math.

**38**:277â€“278. Here, {{math|AveÂ±}} denotes the average over the {{math|2

*n*}} possible choices of signs {{math|Â±1}}. In the same article, KwapieÅ„ proved that the validity of a Banach-valued Parseval's theorem for the Fourier transform characterizes Banach spaces isomorphic to Hilbert spaces.Lindenstrauss and Tzafriri proved that a Banach space in which every closed linear subspace is complemented (that is, is the range of a bounded linear projection) is isomorphic to a Hilbert space.see Lindenstrauss, J. and Tzafriri, L. (1971), "On the complemented subspaces problem", Israel J. Math.

**9**:263â€“269. The proof rests upon Dvoretzky's theorem about Euclidean sections of high-dimensional centrally symmetric convex bodies. In other words, Dvoretzky's theorem states that for every integer {{mvar|n}}, any finite-dimensional normed space, with dimension sufficiently large compared to {{mvar|n}}, contains subspaces nearly isometric to the {{mvar|n}}-dimensional Euclidean space.The next result gives the solution of the so-called

*homogeneous space problem*. An infinite-dimensional Banach space {{mvar|X}} is said to be

**homogeneous**if it is isomorphic to all its infinite-dimensional closed subspaces. A Banach space isomorphic to {{math|â„“2}} is homogeneous, and Banach asked for the converse.see p. 245 in {{harvtxt|Banach|1932}}. The homogeneity property is called "propriÃ©tÃ© (15)" there. Banach writes: "on ne connaÃ®t aucun exemple d'espace Ã une infinitÃ© de dimensions qui, sans Ãªtre isomorphe avec (

*L*2), possÃ¨de la propriÃ©tÃ© (15)".

**Theorem.**Gowers, W. T. (1996), "A new dichotomy for Banach spaces", Geom. Funct. Anal.

**6**:1083â€“1093. A Banach space isomorphic to all its infinite-dimensional closed subspaces is isomorphic to a separable Hilbert space.

**hereditarily indecomposable**when no subspace of it can be isomorphic to the direct sum of two infinite-dimensional Banach spaces. The Gowers dichotomy theorem asserts that every infinite-dimensional Banach space {{mvar|X}} contains, either a subspace {{mvar|Y}} with unconditional basis, or a hereditarily indecomposable subspace {{mvar|Z}}, and in particular, {{mvar|Z}} is not isomorphic to its closed hyperplanes.see JOURNAL, Gowers, W. T., 1994, A solution to Banach's hyperplane problem, Bull. London Math. Soc., 26, 523â€“530, 10.1112/blms/26.6.523, If {{mvar|X}} is homogeneous, it must therefore have an unconditional basis. It follows then from the partial solution obtained by Komorowski and Tomczakâ€“Jaegermann, for spaces with an unconditional basis,see JOURNAL, Komorowski, Ryszard A., Tomczak-Jaegermann, Nicole, 1995, Banach spaces without local unconditional structure, Israel J. Math., 89, 205â€“226, math/9306211, 10.1007/bf02808201, and also JOURNAL, Komorowski, Ryszard A., Tomczak-Jaegermann, Nicole, 1998, Erratum to: Banach spaces without local unconditional structure, Israel J. Math., 105, 85â€“92, math/9607205, 10.1007/bf02780323, that {{mvar|X}} is isomorphic to {{math|â„“2}}.

### Spaces of continuous functions

When two compact Hausdorff spaces {{math|*K*1}} and {{math|

*K*2}} are homeomorphic, the Banach spaces {{math|

*C*(

*K*1)}} and {{math|

*C*(

*K*2)}} are isometric. Conversely, when {{math|

*K*1}} is not homeomorphic to {{math|

*K*2}}, the (multiplicative) Banachâ€“Mazur distance between {{math|

*C*(

*K*1)}} and {{math|

*C*(

*K*2)}} must be greater than or equal to {{math|2}}, see above the results by Amir and Cambern. Although uncountable compact metric spaces can have different homeomorphy types, one has the following result due to Milutin:Milyutin, AlekseÄ A. (1966), "Isomorphism of the spaces of continuous functions over compact sets of the cardinality of the continuum". (Russian) Teor. FunkciÄ Funkcional. Anal. i PriloÅ¾en. Vyp.

**2**:150â€“156.

**Theorem.**Milutin. See also Rosenthal, Haskell P., "The Banach spaces C(K)" in Handbook of the geometry of Banach spaces, Vol. 2, 1547â€“1602, North-Holland, Amsterdam, 2003. Let {{mvar|K}} be an uncountable compact metric space. Then {{math|

*C*(

*K*)}} is isomorphic to {{math|

*C*([0, 1])}}.

langle 1, alpha rangle = { gamma : 1 le gamma le alpha}

equipped with the order topology, where {{math|*Î±*}} is a countably infinite ordinal.One can take {{math|

*Î±*{{=}}

*Ï‰ Î²n*}}, where {{math|

*Î²*+ 1}} is the Cantorâ€“Bendixson rank of {{mvar|K}}, and {{math|

*n*> 0}} is the finite number of points in the

*Î²*-th derived set {{math|

*K*(

*Î²*)}} of {{mvar|K}}. See Mazurkiewicz, Stefan; SierpiÅ„ski, WacÅ‚aw (1920), "Contribution Ã la topologie des ensembles dÃ©nombrables", Fundamenta Mathematicae 1: 17â€“27. The Banach space {{math|

*C*(

*K*)}} is then isometric to {{math|

*C*()}}. When {{math|

*Î±*,

*Î²*}} are two countably infinite ordinals, and assuming {{math|

*Î±*â‰¤

*Î²*}}, the spaces {{math|

*C*()}} and {{math|

*C*()}} are isomorphic if and only if {{math|

*Î²*<

*Î±Ï‰*}}.Bessaga, CzesÅ‚aw; PeÅ‚czyÅ„ski, Aleksander (1960), "Spaces of continuous functions. IV. On isomorphical classification of spaces of continuous functions", Studia Math.

**19**:53â€“62.For example, the Banach spaces

C(langle 1, omegarangle), C(langle 1, omega^{omega} rangle), C(langle 1, omega^{omega^2}rangle), C(langle 1, omega^{omega^3} rangle), cdots, C(langle 1, omega^{omega^omega} rangle), cdots

are mutually non-isomorphic.## Examples

A glossary of symbols:- {{math|
**K**{{=}}**R**,**C**}}; - {{mvar|X}} is a compact Hausdorff space;
- {{mvar|I}} is a closed and bounded interval {{math|[
*a*,*b*]}}; - {{math|
*p*,*q*}} are real numbers with {{math|1 <*p*,*q*< âˆž}} so that {{math| {{sfrac|1|*p*}} + {{sfrac|1|*q*}} {{=}} 1.}} - {{math|Î£}} is a Ïƒ-algebra of sets;
- {{math|Îž}} is an algebra of sets (for spaces only requiring finite additivity, such as the ba space);
- {{mvar|Î¼}} is a measure with variation {{math|{{!}}
*Î¼*{{!}}}}.

Classical Banach spaces |

**K**

*n**}}*Kn}} | {{yes}} | {{yes}} | _2 = left(sum_{i=1}^n | ^2right)^{frac{1}{2}} | Euclidean space |

*! {{math|â„“{{su|p=**n*|b=*p*}}}}{{math | p=n | q}}}} > | x | x_i | | |

*! {{math|â„“{{su|p=**n*|b=âˆž}}}}â„“{{su | n>b=1}}}} | {{yes}} | {{yes}} | _infty = maxnolimits_{1le ile n} |

*! {{math|â„“**p*}}{{math | q}} > | x | x_i | | |

*! {{math|â„“1}}*{{math | x | x_i | | |

*! {{math|â„“âˆž}}*ba space | ba}} | {{no}} | {{no}} | _infty = supnolimits_i |

*! {{mvar|c}}*â„“1}} | {{no}} | {{no}} | _infty = supnolimits_i |

*! {{math|**c*0}}â„“1}} | {{no}} | {{no}} | _infty = supnolimits_i | Isomorphic but not isometric to {{mvar|c}}. |

*! {{math|bv}}*â„“âˆž}} | {{no}} | {{yes}} | _{bv} = | + sum_{i=1}^infty | Isometrically isomorphic to {{math|â„“1}}. |

*! {{math|bv0}}*â„“âˆž}} | {{no}} | {{yes}} | _{bv_0} = sum_{i=1}^infty | Isometrically isomorphic to {{math|â„“1}}. |

*! {{math|bs}}*ba space>{{math | > | x | sum_{i=1}^nx_iright | â„“âˆž}}. |

*! {{math|cs}}*â„“1}} | {{no}} | {{no}} | _{bs} = supnolimits_nleft | Isometrically isomorphic to c space | c}}. |

*! {{math|**B*(*X*, Îž)}}ba space | ba(Îž)}} | {{no}} | {{no}} | _B = supnolimits_{xin X} |

*! {{math|**C*(*X*)}}ba space>{{math | X)}} > | x | f(x) | | |

*! {{math|ba(Îž)}}*mu | mu | | |

*! {{math|ca(Î£)}}*mu | mu | ba(Î£)}}. |

*! {{math|rca(Î£)}}*mu | mu | ca(Î£)}}. |

*! {{math|**Lp*(*Î¼*)}}Lq(Î¼)}} | {{yes}} | {{yes}} | _p = left (int | ^p,dmuright)^{frac{1}{p}} |

*! {{math|**L*1(*Î¼*)}}Lâˆž(Î¼)}} | {{no}} | {{yes}} | _1 = int | ,dmu | The dual is {{math | Lâˆž(Î¼)}} if {{mvar>Î¼}} is Ïƒ-finite measure | Ïƒ}}-finite. |

*! {{math|BV(**I*)}}f | Vf (I)}} is the total variation of {{math| f }} |

*! {{math|NBV(**I*)}}f | NBV(I)}} consists of {{math|BV(I)}} functions such that limnolimits_{xto a^+}f(x)=0 |

*! {{math|AC(**I*)}}K + Lâˆž(I)}} | {{no}} | {{yes}} | _{BV} = V_f(I) + limnolimits_{xto a^+}f(x) | Isomorphic to the Sobolev space {{math|W 1,1(I)}}. |

*! {{math|**Cn*([*a*,*b*])}}Ba space | rca([a,b])}} | {{no}} | {{no}} | = sum_{i=0}^n supnolimits_{xin [a,b]} left | Isomorphic to {{math|Rn âŠ• C([a,b])}}, essentially by Taylor's theorem. |

*{{clear}}*

## Derivatives

Several concepts of a derivative may be defined on a Banach space. See the articles on the FrÃ©chet derivative and the GÃ¢teaux derivative for details. The FrÃ©chet derivative allows for an extension of the concept of a directional derivative to Banach spaces. The GÃ¢teaux derivative allows for an extension of a directional derivative to locally convex topological vector spaces. FrÃ©chet differentiability is a stronger condition than GÃ¢teaux differentiability. The quasi-derivative is another generalization of directional derivative that implies a stronger condition than GÃ¢teaux differentiability, but a weaker condition than FrÃ©chet differentiability.## Generalizations

Several important spaces in functional analysis, for instance the space of all infinitely often differentiable functions**R**â†’**R**, or the space of all distributions on**R**, are complete but are not normed vector spaces and hence not Banach spaces. In FrÃ©chet spaces one still has a complete metric, while LF-spaces are complete uniform vector spaces arising as limits of FrÃ©chet spaces.## See also

## Notes

{{Reflist|30em}}## References

- {{citation|first=Stefan|last=Banach|authorlink=Stefan Banach|url=http://kielich.amu.edu.pl/Stefan_Banach/pdf/teoria-operacji-fr/banach-teorie-des-operations-lineaires.pdf|title=ThÃ©orie des opÃ©rations linÃ©aires|publication-place=Warszawa|publisher=Subwencji Funduszu Kultury Narodowej|year=1932|series=Monografie Matematyczne|volume=1|zbl=0005.20901}}.
- {hide}citation|author=Beauzamy, Bernard|title=Introduction to Banach Spaces and their Geometryorigyear=1982|edition=Second revised|publisher=North-Holland

- {{citation|first=Nicolas|last=Bourbaki|authorlink=Nicolas Bourbaki|title=Topological vector spaces|series=Elements of mathematics|publisher= Springer-Verlag|publication-place=Berlin|year=1987|isbn=978-3-540-13627-9}}.
- {{citation

| last = Carothers | first = Neal L.

| title = A short course on Banach space theory

| series = London Mathematical Society Student Texts

| volume = 64

| publisher = Cambridge University Press

| location = Cambridge

| year = 2005

| pages = xii+184

| isbn = 0-521-84283-2

}}. | title = A short course on Banach space theory

| series = London Mathematical Society Student Texts

| volume = 64

| publisher = Cambridge University Press

| location = Cambridge

| year = 2005

| pages = xii+184

| isbn = 0-521-84283-2

- {hide}citation

| last = Diestel

| first = Joseph

| title = Sequences and series in Banach spaces

| series = Graduate Texts in Mathematics

| volume = 92

| publisher = Springer-Verlag

| location = New York

| year = 1984

| pages = xii+261

| isbn = 0-387-90859-5

{edih}. | first = Joseph

| title = Sequences and series in Banach spaces

| series = Graduate Texts in Mathematics

| volume = 92

| publisher = Springer-Verlag

| location = New York

| year = 1984

| pages = xii+261

| isbn = 0-387-90859-5

- {{Citation

| last1=Dunford | first1=Nelson

| last2=Schwartz | first2=Jacob T. with the assistance of W. G. Bade and R. G. Bartle

| title=Linear Operators. I. General Theory

| publisher=Interscience Publishers, Inc.

| location = New York

| series = Pure and Applied Mathematics

| volume = 7

| mr=0117523

| year=1958}}

| last2=Schwartz | first2=Jacob T. with the assistance of W. G. Bade and R. G. Bartle

| title=Linear Operators. I. General Theory

| publisher=Interscience Publishers, Inc.

| location = New York

| series = Pure and Applied Mathematics

| volume = 7

| mr=0117523

| year=1958}}

- {hide}citation

| last1=Lindenstrauss | first1=Joram |author1-link = Joram Lindenstrauss

| last2=Tzafriri | first2=Lior

| isbn = 3-540-08072-4

| location = Berlin

| publisher = Springer-Verlag

| series = Ergebnisse der Mathematik und ihrer Grenzgebiete

| title=Classical Banach Spaces I, Sequence Spaces

| volume = 92

| year=1977{edih}.

| last2=Tzafriri | first2=Lior

| isbn = 3-540-08072-4

| location = Berlin

| publisher = Springer-Verlag

| series = Ergebnisse der Mathematik und ihrer Grenzgebiete

| title=Classical Banach Spaces I, Sequence Spaces

| volume = 92

| year=1977{edih}.

- {{citation

| last = Megginson | first = Robert E. | authorlink = Robert Megginson

| title = An introduction to Banach space theory

| series = Graduate Texts in Mathematics

| volume = 183

| publisher = Springer-Verlag

| location = New York

| year = 1998

| pages = xx+596

| isbn = 0-387-98431-3

}}. | title = An introduction to Banach space theory

| series = Graduate Texts in Mathematics

| volume = 183

| publisher = Springer-Verlag

| location = New York

| year = 1998

| pages = xx+596

| isbn = 0-387-98431-3

- {{citation

| last = Ryan |first = Raymond A.

| year = 2002

| title = Introduction to Tensor Products of Banach Spaces

| publisher = Springer-Verlag

| series = Springer Monographs in Mathematics

| location = London

| isbn = 1-85233-437-1

| pages = xiv+225

}}. | year = 2002

| title = Introduction to Tensor Products of Banach Spaces

| publisher = Springer-Verlag

| series = Springer Monographs in Mathematics

| location = London

| isbn = 1-85233-437-1

| pages = xiv+225

- {hide}citation

| last=Wojtaszczyk | first= PrzemysÅ‚aw

| title = Banach spaces for analysts

| series = Cambridge Studies in Advanced Mathematics

| volume = 25

| publisher = Cambridge University Press,

| location = Cambridge

| year= 1991

| pages = xiv+382

| ISBN = 0-521-35618-0

{edih}.| title = Banach spaces for analysts

| series = Cambridge Studies in Advanced Mathematics

| volume = 25

| publisher = Cambridge University Press,

| location = Cambridge

| year= 1991

| pages = xiv+382

| ISBN = 0-521-35618-0

## External links

- {{springer|title=Banach space|id=p/b015190}}
- {{MathWorld|BanachSpace|Banach Space}}

**- content above as imported from Wikipedia**

- "

- time: 11:37am EDT - Mon, Jul 23 2018

- "

__Banach space__" does not exist on GetWiki (yet)- time: 11:37am EDT - Mon, Jul 23 2018

[ this remote article is provided by Wikipedia ]

*
*LATEST EDITS [ see all ]

GETWIKI 09 MAY 2016

GETWIKI 18 OCT 2015

GETWIKI 20 AUG 2014

GETWIKI 19 AUG 2014

GETWIKI 18 AUG 2014

© 2018 M.R.M. PARROTT | ALL RIGHTS RESERVED