Introduction to Bases in Banach Spaces

Similar documents
NOTES ON FRAMES. Damir Bakić University of Zagreb. June 6, 2017

Spectral theory for compact operators on Banach spaces

Real Analysis Notes. Thomas Goller

Math 328 Course Notes

Exercise Solutions to Functional Analysis

CLASSES OF STRICTLY SINGULAR OPERATORS AND THEIR PRODUCTS

Boundedly complete weak-cauchy basic sequences in Banach spaces with the PCP

Homework If the inverse T 1 of a closed linear operator exists, show that T 1 is a closed linear operator.

Eigenvalues and Eigenfunctions of the Laplacian

CLASSES OF STRICTLY SINGULAR OPERATORS AND THEIR PRODUCTS

Part V. 17 Introduction: What are measures and why measurable sets. Lebesgue Integration Theory

Part III. 10 Topological Space Basics. Topological Spaces

Eberlein-Šmulian theorem and some of its applications

Problem Set 6: Solutions Math 201A: Fall a n x n,

Filters in Analysis and Topology

Functional Analysis HW #5

FUNCTIONAL ANALYSIS LECTURE NOTES: WEAK AND WEAK* CONVERGENCE

FUNCTIONAL ANALYSIS-NORMED SPACE

FUNCTION BASES FOR TOPOLOGICAL VECTOR SPACES. Yılmaz Yılmaz

(convex combination!). Use convexity of f and multiply by the common denominator to get. Interchanging the role of x and y, we obtain that f is ( 2M ε

3 Orthogonality and Fourier series

A PROPERTY OF STRICTLY SINGULAR 1-1 OPERATORS

THE ALTERNATIVE DUNFORD-PETTIS PROPERTY FOR SUBSPACES OF THE COMPACT OPERATORS

16 1 Basic Facts from Functional Analysis and Banach Lattices

On Riesz-Fischer sequences and lower frame bounds

Bases in Banach spaces

08a. Operators on Hilbert spaces. 1. Boundedness, continuity, operator norms

MATH 117 LECTURE NOTES

RESTRICTED UNIFORM BOUNDEDNESS IN BANACH SPACES

Trace Class Operators and Lidskii s Theorem

FUNCTIONAL ANALYSIS CHRISTIAN REMLING

Finite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product

BANACH SPACES WITHOUT MINIMAL SUBSPACES - EXAMPLES

Overview of normed linear spaces

Normed Vector Spaces and Double Duals

Smith theory. Andrew Putman. Abstract

Your first day at work MATH 806 (Fall 2015)

MAT 578 FUNCTIONAL ANALYSIS EXERCISES

Hilbert Spaces. Contents

arxiv: v1 [math.fa] 14 May 2008

THE DAUGAVETIAN INDEX OF A BANACH SPACE 1. INTRODUCTION

INFINITE ASYMPTOTIC GAMES

Zweier I-Convergent Sequence Spaces

I teach myself... Hilbert spaces

DEFINABILITY UNDER DUALITY. 1. Introduction

MAT 544 Problem Set 2 Solutions

Weak-Star Convergence of Convex Sets

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability...

Spectral theory for linear operators on L 1 or C(K) spaces

NOTES (1) FOR MATH 375, FALL 2012

Notions such as convergent sequence and Cauchy sequence make sense for any metric space. Convergent Sequences are Cauchy

4 Hilbert spaces. The proof of the Hilbert basis theorem is not mathematics, it is theology. Camille Jordan

1 Topology Definition of a topology Basis (Base) of a topology The subspace topology & the product topology on X Y 3

CHAPTER 8: EXPLORING R

Your first day at work MATH 806 (Fall 2015)

Real Analysis, 2nd Edition, G.B.Folland Elements of Functional Analysis

5 Compact linear operators

MA103 Introduction to Abstract Mathematics Second part, Analysis and Algebra

Review of Linear Algebra

4 Linear operators and linear functionals

Functional Analysis F3/F4/NVP (2005) Homework assignment 3

Math General Topology Fall 2012 Homework 11 Solutions

Subsequences of frames

4 Countability axioms

W if p = 0; ; W ) if p 1. p times

1 Functional Analysis

Fredholm Theory. April 25, 2018

Hilbert Spaces. Hilbert space is a vector space with some extra structure. We start with formal (axiomatic) definition of a vector space.

Applications of Descriptive Set Theory in the Geometry of Banach spaces

ON UNCONDITIONALLY SATURATED BANACH SPACES. 1. Introduction

Spaces of continuous functions

Optimization and Optimal Control in Banach Spaces

Differentiability of Convex Functions on a Banach Space with Smooth Bump Function 1

2.2 Annihilators, complemented subspaces

L p Functions. Given a measure space (X, µ) and a real number p [1, ), recall that the L p -norm of a measurable function f : X R is defined by

CHAPTER I THE RIESZ REPRESENTATION THEOREM

Math 61CM - Solutions to homework 6

7 Complete metric spaces and function spaces

PROBLEMS. (b) (Polarization Identity) Show that in any inner product space

Spectral Theory, with an Introduction to Operator Means. William L. Green

On the topology of pointwise convergence on the boundaries of L 1 -preduals. Warren B. Moors

FUNCTIONAL ANALYSIS LECTURE NOTES: COMPACT SETS AND FINITE-DIMENSIONAL SPACES. 1. Compact Sets

Measures and Measure Spaces

Chapter 8. P-adic numbers. 8.1 Absolute values

David Hilbert was old and partly deaf in the nineteen thirties. Yet being a diligent

SOLUTIONS TO SOME PROBLEMS

Summer Jump-Start Program for Analysis, 2012 Song-Ying Li

1.2 Fundamental Theorems of Functional Analysis

Rolle s Theorem for Polynomials of Degree Four in a Hilbert Space 1

Exercises to Applied Functional Analysis

3. The Sheaf of Regular Functions

A local Ramsey theory for block sequences

NORMS ON SPACE OF MATRICES

A VERY BRIEF REVIEW OF MEASURE THEORY

CONTENTS. 4 Hausdorff Measure Introduction The Cantor Set Rectifiable Curves Cantor Set-Like Objects...

Winter School on Galois Theory Luxembourg, February INTRODUCTION TO PROFINITE GROUPS Luis Ribes Carleton University, Ottawa, Canada

Footnotes to Linear Algebra (MA 540 fall 2013), T. Goodwillie, Bases

MATH 426, TOPOLOGY. p 1.

Chapter 3: Baire category and open mapping theorems

Transcription:

Introduction to Bases in Banach Spaces Matt Daws June 5, 2005 Abstract We introduce the notion of Schauder bases in Banach spaces, aiming to be able to give a statement of, and make sense of, the Gowers Dichotomy Theorem. Contents 1 Introduction 1 2 Schauder bases 2 3 Basic sequences 4 4 Unconditional bases 6 5 The Gowers Dichotomy Theorem 9 1 Introduction When working with finite dimensional vector spaces, it is often convenient to take a basis, and then work with the co-ordinate system which this basis gives us. In an infinite dimensional vector space, the axiom of choice shows that we can still find a basis: however, because it is now infinite, its use is often less. This is especially true in a Banach space, because such an algebraic basis takes no account of the extra structure induced by the norm. This leads to the notion of a Schauder basis. Let E be a Banach space and let (e n ) be a sequence of vectors in E. Then (e n ) is a Schauder basis (or, from now on, simply a basis) if every x E admits an expansion of the form x = for some unique sequence of scalars (x n ). x n e n, Note. As someone interested in algebraic questions, I always work over the complex numbers. However, it is common (for simplicity) to work with real scalars in Banach space theory. Often, a similar result for complex numbers follows easily, sometimes we have to work somewhat harder, and occasionally, the result is true only for real scalars. We shall try to give proofs which work with either choice of scalar, but occasionally the reader should assume the result is only certainly known for real scalars. 1

Example 1.1. Let E = l p for 1 p <, or E = c 0. Note that some authors write l p, l p, or l p instead. For n 1, let e n E be the sequence which is 0 except with a 1 in the nth position. Then (e n ) is a Schauder basis for E, called the standard unit-vector basis of E. Example 1.2. Let H be a separable Hilbert space with an orthonormal basis (e n ). Then (e n ) is a basis (in fact, an unconditional basis: see later) for H. 2 Schauder bases We first give some general results on Schauder bases. Let E be a Banach space with a basis (e n ). Clearly, we are free to multiply each e n by a non-zero scalar without changing the basis property. Hence we may assume that (and will henceforth do so) (e n ) is a normalised basis, that is, e n = 1 for each n. Then, for x E, we can uniquely write x = x n e n, and hence define x 0 = sup N 1 N. x n e n This is well-defined, as N x ne n x as N, and every convergent sequence in a Banach space is clearly bounded. We also easily see that x x 0 for each x E, and that x 0 = 0 if and only if x = 0. Theorem 2.1. Let E be a Banach space with a basis (e n ), and let 0 be as defined above. Then and 0 are equivalent norms on E. Proof. Let F be the normed space which is E together with the norm 0 (it is easily checked that F is indeed a normed-space). Let ι : F E be the formal inclusion map. Then ι is norm-decreasing, and is a bijection. Suppose that F is Banach space, that is, F is complete. Then the Open Mapping Theorem implies that ι has a continuous inverse, which implies that 0 is equivalent to. So, we wish to show that F is complete. Firstly, given x = x ne n, and m 1, we notice that m x m = x m e m e m 1 = e m 1 x n e n m 1 x n e n ( m e m 1 m 1 ) x n e n + x n e n 2 e m 1 x 0, by the definition of 0. Now let (x n ) be a Cauchy-sequence in F, and let x n = x n,m e m (n 1). Then, for ɛ > 0, there exists N ɛ so that for r, s N ɛ, x r x s 0 = sup N 1 N (x r,m x s,m )e m < ɛ. By the calculation above, we hence see that for each m 1, the sequence of scalars (x n,m ) is a Cauchy-sequence, with limit y m say. We now show that y = y m e m 2

converges in F. As 0, this implies that the sum converges in E, and hence that y exists, as E is complete. Towards this end, let n N ɛ, so that for each N 1, we have that N N (y m x n,m )e m = lim (x N r,m x n,m )e m = lim (x r,m x n,m )e m ɛ, r r by the definition of N ɛ. Notice that this shows that y is indeed the limit of (x n ), supposing of course that y exists. Then x n,me m converges in E, so there exists M ɛ such that if M ɛ r < s, then s m=r x n,me m < ɛ. Thus we have that s m=r y m e m 0 = sup t r min(t,s) m=r m=r y m e m ( t t ) sup (y m x n,m )e m + x n,m e m r t s m=r ( t r 1 ) < sup (y m x n,m )e m + (y m x n,m )e m + ɛ 3ɛ, r t s so that m y me m does converge in F. We can now give a finite characterisation of a basis. Theorem 2.2. Let E be Banach space, and let (e n ) be a sequence in E. Then (e n ) is a basis for E if and only if: 1. each e n is non-zero; 2. the linear space of (e n ) is dense in E; 3. there exists a constant K such that for every sequence of scalars (x n ), and each N < M, we have that N M. x n e n K x n e n Proof. Suppose that (e n ) is a basis, and form the norm 0, as above, so that by the above theorem, there exists K such that x x 0 K x (x E). Clearly we have condition (1) by the uniqueness of the expansion of 0 = 0e n; condition (2) is trivial. Then, for (3), let (y n ) be the sequence of scalars defined by y n = x n for n M, and y n = 0 otherwise. Then we see that N N 0 M, x n e n = y n e n y n e n K y n e n = K x n e n as required. Conversely, let F be the linear span of (e n ), a dense subspace of E. For each n 1, define a linear map P n : F F by ( N ) P n x k e k = n x k e k (N n, 3 N ) x k e k F.

By condition (3), we see that P n K. It is then clear that P n is a bounded projection on F, and that P n thus extends to a bounded projection on E. For each n 1, define a linear functional e n on F by the formula e n(x)e n = P n (x) P n 1 (x) (x F ), where P 0 = 0. Then e n is well-defined upon F (by condition (1)) and e n 2K, so that e n extends by continuity to a bounded linear functional on E. It is clear that P n (x) = n e k(x)e k (x F ), so by continuity, this formula holds for x E as well. Now let x E. We claim that x = e n(x)e n. Let (x n ) be a sequence in F which converges to x. Then, for ɛ > 0, let M 1 be such that x x M < ɛ, and let N 1 be sufficiently large so that P N (x M ) = x M, which we can do, as x M F. Then we see that N e n(x)e n x = P N (x) x Finally, we note that if P N (x) P N (x M ) + P N (x M ) x M + X M x K x x M + x M x < (K + 1)ɛ. x = x n e n = e n(x)e n, then applying e N, we conclude that x N = e N (x) for each N, so that such an expansion is unique. The smallest constant which can arise in (3) is the basis constant of the basis (e n ). Theorem 2.1 shows that we can always renorm E to give a basis on 1. Such bases are called monotonic. It is immediate that a Banach space with a basis is separable, so that, for example, l does not have a basis. Most common separable Banach spaces do have bases, although they are often non-obvious to find. For a while, it was thought that all separable Banach spaces would have a basis, but Enflo produced a counter-example in [Enflo, 1973]. There is, however, a weaker notion where we do have a positive result. 3 Basic sequences Let E be a Banach space, and let (x n ) a sequence in E such that (x n ) is a basis for its closed linear span. Then (x n ) is a basic sequence in E. It is easy to show that every Banach space has a basic sequence; that is, every Banach space E contains a closed, infinite-dimensional subspace F with a basis. 4

Lemma 3.1. Let E be an infinite-dimensional Banach space, let F be a finite-dimensional subspace of E, and let ɛ > 0. Then there exists x E such that x = 1 and for all y F and all scalars a. y (1 + ɛ) y + ax, Proof. We may suppose that ɛ < 1. As the unit ball of F is compact, there is a finite set {y 1,..., y n } in F such that y k = 1 (1 k n), min 1 k n y y k < ɛ/2 (y F, y = 1). Pick y1,..., yn norm-one vectors in E, the dual of E, such that yk (y k) = 1 for each k. Then there exists x E with x = 1 and yk (x) = 0 for each k. For any norm-one y F, pick y k such that y k y < ɛ/2. For a scalar a, we have that as required. y + ax y k + ax y y k > y k + ax ɛ/2 y k(y k + ax) ɛ/2 = 1 ɛ/2 (1 + ɛ) 1, Theorem 3.2. Every Banach space E contains a basic sequence. Proof. We use induction to pick a sequence of norm-one vectors (x n ) such that condition (3) of Theorem 2.2 always holds, with K = 2 say (the proof works for any K > 1). Suppose we have chosen x 1,..., x n and ɛ > 0 such that m n, a k x k (2 ɛ) a k x k for any m n and any scalars (a k ) n. Note that we can clearly do this for n = 1. We now try to find x n+1. We need to ensure that x n+1 = 1 and that for some ɛ 0 > 0, we have that m n+1, a k x k (2 ɛ0 ) a k x k for any m n and any scalars (a k ) n+1. Let F n be the linear span of x 1,..., x n, a finite-dimensional subspace of E. Use the above lemma to find a norm-one vector x n+1 such that y (1 + δ) y + a n+1 x n+1 for each y F n and each scalar a n+1, where δ > 0 is chosen so that (2 ɛ)(1 + δ) = 2 ɛ/2, that is, δ = ɛ/2(2 ɛ). Then, for a sequence of scalars (a k ) n+1, let y = n a kx k F n, so that for m n, we see that m n a k x k (2 ɛ) a k x k = (2 ɛ) y (2 ɛ)(1 + δ) y + an+1 x n+1 as required. n+1, = (2 ɛ/2) a k x k 5

4 Unconditional bases A basis for a Banach space is not, in many cases, of a huge amount of use, as the convergence properties are rather weak. A more useful notion is that of an uncondtional basis. A series (x n ) in a Banach space E is said to sum unconditionally if, for each permutation σ on N, the sum x σ(n) converges, and converges to the same limit (although, see below, this is automatic), independently of σ. Recall that in a finite-dimensional Banach space (or just in R or C) this notion is equivalent to x n <. This is not true in the infinite-dimensional case (indeed, it characterises finite-dimensional Banach spaces). Then a basis (e n ) for a Banach space E is an unconditional basis if, for each x E, there exists a unique expansion of the form x = x n e n, where the sum converges unconditionally. Firstly, we need to explore some properties of unconditional convergence in Banach spaces. Proposition 4.1. Let E be a Banach space, and let (x n ) be a series in E. Then the following are equivalent: 1. the sum x n is unconditional; 2. for each permutation σ of N, the sum x σ(n) converges in E; 3. for each ɛ > 0, there exists a finite subset A ɛ N such that if B N is finite with A ɛ B =, then n B x n < ɛ. Proof. Clearly (1) implies (2). To show that (2) implies (3), suppose towards a contradiction that (3) does not hold, so that we can find ɛ > 0 such that for each finite A N, there exists a finite B A N with A B A = and n B A x n ɛ. So let B 1 = B {1}, then let A 2 = {1, 2,..., max B 1 }, and let B 2 = B A2. Continue in this manner to find a sequence of finite sets (B n ) such that max B n < min B n+1 for each n 1. We can easily construct a permutation σ of N such that each B n is the image under σ of some interval, say B n = {σ(r n ), σ(r n + 1),..., σ(s n )}, so that ɛ x k k B n = s n k=r n x σ(k) (n 1). However, this implies that k x σ(k) does not converge, giving the required contradiction. Now suppose that (3) holds, and that σ is a permutation of N. For ɛ > 0, choose A ɛ N and let N ɛ > 0 be such that if n N ɛ, then σ(n) A ɛ, σ 1 (n) A ɛ, and n A ɛ. Then, for N ɛ r < s, let B = {σ(n) : r n s} and let C = {n : r n s}, so that B and C are finite, and B A ɛ = C A ɛ =. Thus we have that s x σ(n) = x n < ɛ, n B s x n = x n < ɛ, n C 6

which shows that x = n x n and y = n x σ(n) exist. Furthermore, let K = {1 n N ɛ : σ(n) > N ɛ } and L = {σ(n) : 1 n N ɛ, σ(n) > N ɛ }, so that L A ɛ =. Also, if n K, then m = σ(n) > N ɛ so that σ 1 (m) = n A ɛ, that is, K A ɛ =. Then we see that N ɛ N ɛ x n x σ(n) = x n < 2ɛ, n K x n n L which easily shows that x y 4ɛ. Thus we see that (1) holds. Notice that condition (3) shows that if n x n converges unconditionally, then given any A N, the sum n A x n converges unconditionally. Indeed, we can say more. Proposition 4.2. Let E be a Banach space, and let (x n ) be a series in E. If (x n ) sums unconditionally, then for each sequence (z n ) in {+1, 1}, the sum n z nx n converges uniformly, independently of the choice of (z n ). Furthermore, the following are equivalent: 1. the sum x n is unconditional; 2. for any bounded sequence of scalars (z n ), the sum z nx n converges in E. Proof. We use condition (3) from the above proposition to characterise when n x n converges unconditionally. So suppose that x = n x n converges unconditionally. By the above comment, for each A N, we may define P A (x) = n A x n. Then suppose that (z n ) is a sequence in {+1, 1}, so that if A = {n : z n = 1}, then clearly n z nx n = P A (x) P N\A (x) converges. Suppose that there exists ɛ > 0 such that for each r 1, there exists a sequence (a n ) in {+1, 1} and s > r such that s a nx n ɛ. Then we find can sequences (A n ) and (B n ) of finite subsets of N such that A n B n = A n A m = B n B m = for all n m, and k A n x k k B n x k ɛ. Let σ and τ be permutations of N such that each A n occurs as the image under σ of an interval, and each B n occurs as the image under τ of an interval. As n x σ(n) and n x τ(n) converge, we again conclude that for n sufficiently large, k A n x k < ɛ/3 and k B n x k < ɛ/3, which is a contradiction. This shows that n z nx n converges uniformly, independently of the choice of the (z n ). Suppose now we are working over the real numbers, and let (z n ) be a bounded sequence in R. For 1 r < s, pick µ in the dual space to E with µ = 1 and s z nµ(x n ) = N z nx n. For each n, let a n = 1 if µ(x n ) 0 or a n = 1 otherwise, so that s z n x n s z n µ(x n ) = s s. z n a n µ(x n ) (z n ) a n x n By the above, we know that we can make s a nx n arbitrarily small, independently of the choice of (a n ). Thus we conclude that n z nx n converges. We have already essentially shown that (2) implies (1), just by taking sequences (z n ) in {+1, 1}. We are now in a position to prove a version of Theorem 2.2 for unconditional bases: this is left as an exercise to the reader. Instead, we prove the hard parts of such a theorem. 7

Theorem 4.3. Let E be a Banach space, and let (e n ) be an unconditional basis for E. For x = n x ne n, define x 0 = Then 0 is an equivalent norm on E. sup (z n) 1. z n x n e n Proof. As (e n ) is a basis for E, we may assume that (e n ) is actually a monotone, normalised basis. Firstly, note that the above proposition shows that 0 is indeed welldefined. This follows as for ɛ > 0, there exists N ɛ, independent of (z n ), such that if N ɛ r < s, then s z nx n e n ɛ (z n ). Thus we have that n z nx n e n ( (z n ) ɛ + Nɛ x n e n ). We proceed as in the proof of the preceeding proposition. For x = n x ne n, define x 1 = sup (z n) {+1, 1} N, z n x n e n again, this is well-defined. If we can show that 1 is an equivalent norm on E, then the argument used above shows that 0 is also an equivalent norm on E. Let Λ be the collection of sequences in {+1, 1}, which is a group under pointwise multiplication. For each z = (z n ) Λ, let T z : E E be the linear map ( ) T z x n e n = z n x n e n. Suppose that (x n ) is a sequence in E tending to 0, that x n = m x n,me m for each n, and that T z (x n ) m y me m as n. That is, lim n (z m x n,m y m )e m = 0. As (e n ) is a monotone basis, we may apply the co-ordinate functional e m to conclude that lim z mx n,m y m = 0 (m 1). n However, as x n,m = e m(x n ) 0 as n, we see that y m = 0 for each m, so that y = 0. The Closed Graph Theorem then implies that T z is bounded. The above proof shows that, for a fixed x E, the family {T z (x) : z Λ} is bounded. It follows from the Principle of Uniform Boundedness that the family of operators {T z : z Λ} is bounded. However, this is simply the statement that 1 is equivalent to, as required. Example 4.4. In c, the closed subspace of l consisting of convergent sequences, let f n = (0,..., 0, 1, 1,...) c, where the 1 appears in the nth position. Then, for x = (x n ) c, let y 1 = x 1, and y n = x n x n 1 for n 2, so that N y n f n = ( y 1, y 1 + y 2,..., y 1 + y 2 + + y n,... ) = ( x 1, x 2,..., x n, x n,... ), which tends to x as N. That is, (f n ) is a basis for c, called the summing basis. 8

However, we notice that N f n = N, N ( 1) n f n = 1, so that the above theorem implies that (f n ) is a conditional basis. Let E be a Banach space with an unconditional basis (e n ). Then we may renorm E so that there is a norm-decreasing algebra homomorphism l B(E), from the Banach algbera l under pointwise multiplication to the algebra of operators on E. In particular, for A N, define an operator P A : E E by P A ( n x ne n ) = n A x ne n. Then {P A : A N} is a bounded family of projections on E; by renorming, we may suppose that each P A is actually norm-decreasing. 5 The Gowers Dichotomy Theorem Let E be a Banach space, and let P be a bounded projection on E. Then let F be the kernal of P, a closed subspace of E, and let G be the image of P, which is equal to the kernal of I P, and is hence a closed subspace. It is trivial to see that E = F G. Conversely, if E = F G for some closed subspaces F and G, then for each x E, we can write x = f + g for unique f F and g G. It is easy to see that the map x g defines a bounded projection on E with kernel F and image G. If F is a finite-dimensional subspace of E, then by taking a finite basis of F, we can easily define a bounded projection E F. Hence E always admits some projections: however, need there exist a projection with an infinite-dimensional image? An infinite-dimensional Banach space E is indecomposable if whenever we can write E as the direct sum of two closed subspaces, E = F G, then one of F or G is finitedimensional. That is, the only bounded projections on E are the trivial ones. A Banach space E is hereditarily indecomposable if, furthermore, every infinite-dimensional closed subspace F of E is indecomposable. That such Banach spaces exist is somewhat surprising, but in [GM, 1993], Gowers and Maurey constructed an example of such a space. Notice that by the comments at the end of the last section, a hereditarily indecomposable Banach space E cannot contain an unconditional basic sequence, or we would have an infinite dimensional closed subspace F which would admit many non-trivial bounded projections. Indeed, this was the problem which Gowers and Maurey were trying to find a counter-example to. Hence we see that being hereditarily indecomposable and having an unconditional basic sequence are somehow the opposite of each other. Remarkably, this statement can be made precise, in the Gowers Dichotomy Theorem. Theorem 5.1. Let E be a Banach space. Then there exists a subspace F of E such that either F is hereditarily indecomposable, or F has an unconditional basis. Proof. See [Gowers, 1996] for a discussion, and [Gowers, 2002] for the full proof. References [Enflo, 1973] Enflo, P., A counterexample to the approximation problem in Banach spaces, Acta Math. 130 (1973) 309 317. [Gowers, 1996] Gowers, W.T., A new dichotomy for Banach spaces, Geom. Funct. Anal. 6 (1996) 1083 1093. 9

[Gowers, 2002] Gowers, W. T., An infinite Ramsey theorem and some Banach-space dichotomies, Ann. of Math. 156 (2002) 797 833. [GM, 1993] Gowers, W. T., Maurey, B., The unconditional basic sequence problem, J. Amer. Math. Soc. 6 (1993) 851 874. 10