1:
2: \documentclass[11pt,reqno]{amsart}
3: \usepackage{amssymb,amscd,amsbsy}
4: \usepackage{amsthm}
5: \input{classstart}
6: \pagestyle{plain}
7: %\date{\mbox{}}
8:
9:
10:
11: \theoremstyle{remark}
12: \newtheorem{rem}[thm]{Remark}
13: \newtheorem*{rem*}{Remark}
14:
15:
16:
17:
18: %\numberwithin{equation}{section}
19:
20: \begin{document}
21:
22: \newcommand{\vse}{\vspace{.2in}}
23: \numberwithin{equation}{section}
24:
25: \title{\bf On S. Mazur's problems 8 and 88 \\from the Scottish Book}
26: \author{V.V. Peller}
27: \thanks{The author is partially supported by NSF grant DMS 0200712}
28: \maketitle
29:
30:
31: \begin{abstract}
32: The paper discusses Problems 8 and 88 posed by Stanis\pl aw Mazur in
33: the Scottish Book \cite{SB}. It turns out that negative solutions to both
34: problems are immediate consequences of the results of \S 5 of \cite{P1}.
35: We discuss here some quantitative aspects of Problems 8 and 88 and give answers to open problems
36: discussed in a recent paper [PS] in connection with Problem 88.
37: \end{abstract}
38:
39: %\setcounter{section}{0}
40: \section{\bf Introduction}
41: \setcounter{equation}{0}
42:
43: \
44:
45: We are going to discuss in this paper Problems 8 and 88 posed by Stanis\pl aw Mazur in the Scottish Book
46: \cite{SB}. Problem 88 asks whether a Hankel matrix in the injective tensor product
47: $\ell^1\check{\otimes}\ell^1$ of two spaces $\ell^1$
48: must have finite sum of the moduli of its matrix entries. Problem 8 asks whether for an arbitrary sequence
49: $\{z_n\}_{n\ge0}$ in the space $c$ of converging sequences
50: there exist sequences $\{x_n\}_{n\ge0}$ and $\{y_n\}_{n\ge0}$
51: in the space $c$ such that
52: $$
53: z_n=\frac1{n+1}\sum_{k=0}^nx_ky_{n-k},\quad n\ge0.
54: $$
55: We give precise statements of the problems and all necessary definitions later.
56:
57: It turned out that both problems have negative solutions. Independently, solutions were obtained by
58: Kwapie\'n and Pe\pl czy\'nski \cite{KP} and Eggermont and Leung \cite{EL}. In a recent paper by
59: Pe\pl czy\'nski and Sukochev \cite{PS} in Section 6 certain quantitative results related to negative
60: solutions of Problems 8 and 88 are obtained and certain open problems are raised.
61:
62: It turns out, however, that the results of Section 5 of my earlier paper \cite{P1} immediately
63: imply negative solutions to Problems 8 and 88. Moreover, Section 5 of \cite{P1} contains much stronger results.
64: In particular, a complete description of the Hankel matrices\footnote{Note that Hankel
65: matrices and Hankel operators play an important role in many areas of mathematics and applications, see
66: \cite{P2}.} in the injective tensor product of two
67: spaces $\ell ^1$ is obtained in \cite{P1} in terms of the Besov space $B^1_{\be,1}$.
68: Unfortunately, I was not aware about the Problems 8 and 88 when I wrote the paper \cite{P1}.
69:
70: In Sections 3 and 4 of this paper we explain why the results of \cite{P1} immediately imply negative
71: solutions to Problems 8 and 88 and we give a solution to the problems raised in \cite{PS}.
72:
73: In \S 2 we collect necessary information on tensor products and Besov spaces.
74:
75:
76: \
77:
78: \section{\bf Preliminaries}
79: \setcounter{equation}{0}
80:
81: \
82:
83: {\bf 1. Projective and injective tensor products.} We define the {\it projective tensor product}
84: $\ell^\be\hat{\otimes}\ell^\be$ as the space of matrices $\{q_{jk}\}_{j,k\ge0}$ of the form
85: \bay
86: \label{lbe}
87: q_{jk}=\sum_{n\ge0}a^{(n)}_jb^{(n)}_k,
88: \ey
89: where the $a^{(n)}=\{a^{(n)}_j\}_{j\ge0}$ and $b^{(n)}=\{b^{(n)}_j\}_{j\ge0}$ are sequences in
90: $\ell^\be$ such that
91: \bay
92: \label{fin}
93: \sum_{n\ge0}\|a^{(n)}\|_{\ell^\be}\|b^{(n)}\|_{\ell^\be}<\be.
94: \ey
95: The norm of the matrix $\{q_{jk}\}_{j,k\ge0}$ in $\ell^\be\hat{\otimes}\ell^\be$ is defined
96: as the infimum of the left-hand side of \rf{fin} over all sequences $a^{(n)}=\{a^{(n)}_j\}_{j\ge0}$ and
97: $b^{(n)}=\{b^{(n)}_j\}_{j\ge0}$ satisfying \rf{lbe}.
98:
99: Similarly, one can define the projective tensor products $c\hat{\otimes}c$ and $c_0\hat{\otimes}c_0$, where
100: $c$ is the subspace of $\ell^\be$ that consists of the converging sequences and $c_0$ is the subspace of $c$
101: that consists of the sequences with zero limit.
102:
103: We consider the space $V^2$ that is a kind of a ``weak completion'' of $\ell^\be\hat{\otimes}\ell^\be$. $V^2$
104: consists of the matrices $Q=\{q_{jk}\}_{j,k\ge0}$ for which
105: $$
106: \sup_{n>0}\|P_nQ\|_{\ell^\be\hat{\otimes}\ell^\be}<\be,
107: $$
108: where the projections $P_n$ are defined by
109: $$
110: (P_nQ)_{jk}=\left\{\begin{array}{ll}q_{jk},&j\le n,~k\le n,\\[.2cm]0,&\mbox{otherwise}.\end{array}\right.
111: $$
112: Note that
113: $c\hat\otimes c\subset \ell^\infty\hat\otimes\ell^\infty\subset V^2$.
114:
115: The {\it injective tensor product} $\ell^1\check{\otimes}\ell^1$ of two spaces $\ell^1$ is, by definition, the
116: space of matrices $Q=\{q_{jk}\}_{j,k\ge0}$ such that
117: $$
118: \|Q\|_{\ell^1\check{\otimes}\ell^1}=\sup\left|\sum^N_{j,k=0}q_{jk}x_jy_k\right|<\be,
119: $$
120: where the supremum is taken over all sequences $\{x_j\}_{j\ge0}$ and $\{y_k\}_{k\ge0}$ in the unit ball of
121: $\ell^\be$ and over all positive integers $N$. The space $\ell^1\check{\otimes}\ell^1$ can naturally be
122: identified with the space of bounded linear operators from $c_0$ to $\ell^1$ (note that every bounded operator
123: from $c_0$ to $\ell^1$ is compact).
124:
125: \medskip
126:
127: {\bf2. Besov spaces.} In this paper we consider only Besov spaces of functions analytic in the unit disk $\dd$.
128: Besov spaces $B_{p,q}^s$ admit many equivalent descriptions. We give a definition
129: in terms of dyadic Fourier expansions. We define the polynomials $W_n$, $n\ge0$, as follows. If $n\ge1$, then
130: $\widehat W_n(2^n)=1$, $\widehat W_n(k)=0$ for $k\not\in(2^{n-1},2^{n+1})$, and $\widehat W_n$ is a linear
131: function on $[2^{n-1},2^n]$ and on $[2^n,2^{n+1}]$. We put $W_0(z)=1+z$. It is easy to see that
132: $$
133: \|W_n\|_{L^1}\le\frac32,\quad n\ge0,
134: $$
135: and
136: $$
137: f=\sum_{n\ge0}f*W_n
138: $$
139: for an arbitrary analytic function $f$ in $\dd$.
140:
141: For $1\le p\le\be$, $1\le q\le\be$, and $s\in\R$, we define the Besov space $B^s_{p,q}$ as the space of
142: analytic functions in $\dd$ satisfying
143: \bay
144: \label{bes}
145: f\in B^s_{p,q}\quad\Longleftrightarrow\quad
146: \{2^{ns}\|f*W_n\|_{L^p}\}_{n\ge0}\in\ell^q.
147: \ey
148: If $q=\be$, the space $B^s_{p,q}$ is nonseparable. We denote by $b^s_{p,\be}$ the closure of the set of
149: polynomials in $B^s_{p,\be}$. It is easy to verify that
150: $$
151: f\in b^s_{p,\be}\quad\Longleftrightarrow\quad
152: \{2^{ns}\|f*W_n\|_{L^p}\}_{n\ge0}\in c_0.
153: $$
154:
155: Besov spaces admit many other descriptions (see \cite{Pe} and \cite{P2}).
156:
157: \
158:
159: \section{\bf Problem 8}
160: \setcounter{equation}{0}
161:
162: \
163:
164: To state Mazur's Problem 8 of the Scottish Book \cite{SB}, consider the bilinear form
165: $\B$ on $c\times c$ defined by
166: $$
167: \B\big(\{x_n\}_{n\ge0},\{y_n\}_{n\ge0}\big)=\{z_n\}_{n\ge0},
168: $$
169: where
170: $$
171: z_n=\frac1{n+1}\sum_{k=0}^nx_ky_{n-k},\quad n\ge0,
172: $$
173: and $c$ is the space of sequences that have a limit at infinity.
174:
175: It is easy to see that $B$ maps $c\times c$ into $c$. {\it S. Mazur asked in Problem 8 whether
176: $\B$ maps $c\times c$ onto $c$}.
177:
178: As mentioned in the Introduction, a negative solution to problem 8 follows immediately from
179: the results of \S 5 of \cite{P1}. To state Theorem 5.1 of \cite{P1}, we define the operator $\A$ on the space of
180: matrices. Let $Q=\{q_{jk}\}_{j,k\ge0}$. Then $\A Q$ is the sequence defined by
181: $$
182: \A Q=\{z_n\}_{n\ge0},\quad\mbox{where}\quad z_n=\frac1{n+1}\sum_{j+k=n}q_{jk}.
183: $$
184:
185: \medskip
186:
187: {\bf Theorem 5.1 of [P1].} {\it$\A$ maps the space $V^2$ onto the space of Fourier coefficients of the
188: Besov space $B^0_{1,\be}$.}
189:
190: \medskip
191:
192: Recall that the space $V^2$ is defined in the introduction. In particular, it follows from Theorem 5.1 of [P1]
193: that
194: $$
195: \A(c\hat{\otimes}c)\subset\A(\ell^\be\hat{\otimes}\ell^\be)\subset\A(V^2)\subset
196: \left\{\{\hat f(n)\}_{n\ge0}:~f\in B^0_{1,\be}\right\},
197: $$
198: and so
199: $$
200: \B(c\times c)\subset\left\{\{\hat f(n)\}_{n\ge0}:~f\in B^0_{1,\be}\right\}.
201: $$
202: It is easy to see that
203: $$
204: c\not\subset\left\{\{\hat f(n)\}_{n\ge0}:~f\in B^0_{1,\be}\right\}.
205: $$
206: Indeed, if $f\in B^0_{1,\be}$, then it follows immediately from \rf{bes} and from \cite{R}, \S 8.6 that
207: $$
208: \sup_{n\ge0}\sum_{k=0}^n\big|\hat f(2^n+2^k)\big|^2<\be.
209: $$
210: This gives a negative solution to Problem 8.
211:
212: In fact, Theorem 5.1 of \cite{P1} allows one to describe $\A(c\hat{\otimes}c)$. First, let us
213: observe that Theorem 5.1 of \cite{P1} easily implies the following description of $\A(c_0\hat{\otimes}c_0)$.
214:
215: \begin{thm}
216: \label{c0}
217: $$
218: \A(c_0\hat{\otimes}c_0)=\left\{\{\hat f(n)\}_{n\ge0}:~f\in b^0_{1,\be}\right\}.
219: $$
220: \end{thm}
221:
222: Recall that $b^0_{1,\be}$ is the closure of the polynomials in $B^0_{1,\be}$ (see \S 2). Theorem \ref{c0},
223: in turn, easily implies the following description of $\A(c\hat{\otimes}c)$.
224:
225: \begin{thm}
226: \label{c}
227: $$
228: \A(c\hat{\otimes}c)=\left\{\{\hat f(n)+d\}_{n\ge0}:~f\in b^0_{1,\be},~d\in\C\right\}.
229: $$
230: \end{thm}
231:
232: \
233:
234: \section{\bf Problem 88}
235: \setcounter{equation}{0}
236:
237: \
238:
239: Recall that in Problem 88 of \cite{SB} {\it S. Mazur asked whether a Hankel matrix
240: $\{\g_{j+k}\}_{j,k\ge0}$ in the injective tensor product $\ell^1\check{\otimes}\ell^1$
241: must satisfy the condition}:
242: $$
243: \sum_{k=0}^\be(1+k)|\g_k|<\be,
244: $$
245: i.e., whether the sum of the moduli of the matrix entries must be finite.
246:
247: As mentioned in the Introduction, a negative solution to Problem 88 follows immediately from the results
248: of \S 5 of \cite{P1}. A complete description of Hankel matrices in $\ell^1\check{\otimes}\ell^1$
249: is given by Theorem 5.2 of \cite{P1}:
250:
251: \medskip
252:
253: {\bf Theorem 5.2 of [P1].} {\it A Hankel matrix $\{\g_{j+k}\}_{j,k\ge0}$ belongs to
254: $\ell^1\check{\otimes}\ell^1$ if and only if the function $f$ defined by
255: $$
256: f(z)=\sum_{n\ge0}\g_n z^n
257: $$
258: belongs to the Besov class $B^1_{\be,1}$.}
259:
260: \medskip
261:
262: Let us obtain the best possible estimate on the moduli of the matrix entries of Hankel matrices in
263: $\ell^1\check{\otimes}\ell^1$.
264:
265: Since $\|f*W_n\|_{L^2}\le\|f\|_{L^2}\|W_n\|_{L^1}\le 3/2\|f\|_{L^2}$,
266: it follows easily from \rf{bes} that if $f\in B^1_{\be1}$, then
267: \bay
268: \label{mod}
269: \sum_{n=0}^\infty2^n\left(\sum_{k=2^n}^{2^{n+1}-1}|\hat f(k)|^2\right)^{1/2}<\infty.
270: \ey
271: Let us show that this is the best possible estimate for the moduli of the Fourier coefficients
272: of functions in $B^1_{\be1}$. To show this, we are going to use a version of the
273: de Leeuw--Katznelson--Kahane theorem. It was proved in \cite{dLKK} that if $\{\b_n\}_{n\in\Z}$
274: is a sequence of nonnegative numbers in $\ell^2(\Z)$, then there exists a continuous function $f$ on $\T$ such
275: that
276: $$
277: |\hat f(n)|\ge\b_n,\quad n\in\Z.
278: $$
279: We refer the reader to \cite{K1}, \cite{K2}, and \cite{N} for refinements of the
280: de Leeuw--Katznelson--Kahane theorem and different proofs. We need the following version of the
281: de Leeuw--Katznelson--Kahane theorem:
282:
283: \begin{lem}
284: \label{kis}
285: There is a positive number $K$ such that for arbitrary nonnegative numbers
286: $\b_0,\b_1,\cdots,\b_m$, there
287: exists a polynomial $f$ of degree $m$ such that
288: $$
289: |\hat f(j)|\ge\b_j,\quad 0\le j\le n,\quad\mbox{and}\quad\|f\|_{L^\be}\le
290: K\left(\sum_{j=0}^n\b_j^2\right)^{1/2}.
291: $$
292: \end{lem}
293:
294: Lemma \ref{kis} follows easily from the results of \cite{K2}.
295:
296: \begin{thm}
297: \label{LKK}
298: Let $\{\a_k\}_{k\ge0}$ be a sequence of nonnegaitive numbers such that
299: \bay
300: \label{maj}
301: \sum_{n=0}^\infty2^n\left(\sum_{k=2^n}^{2^{n+1}-1}\a_k^2\right)^{1/2}<\infty.
302: \ey
303: Then there exists a function $\f$ in the space $B_{\be1}^1$ such that
304: $|\hat\f(k)|\ge \a_k$ for $k\ge0$.
305: \end{thm}
306:
307: \Pf By Lemma \ref{kis}, there exists $K>0$ and a sequence of polynomials $f_n$, $n\ge0$, such that
308: $$
309: f_0(z)=\hat f_0(0)+\hat f_0(1)z,\quad f_n(z)=\sum_{k=2^n}^{2^{n+1}-1}\hat f_n(k)z^k,
310: \quad\mbox{for}\quad n\ge1,
311: $$
312: $$
313: |\hat f_0(k)|\ge\a_k,\quad\mbox{for}\quad k=0,\,1,\quad|\hat f_n(k)|\ge\a_k\quad\mbox{for}
314: \quad n\ge1,~~2^n\le k\le2^{n+1}-1,
315: $$
316: and
317: $$
318: \|f_0\|_{L^\be}\le K(\a_0^2+\a_1^2)^{1/2},
319: \quad\|f_n\|_{L^\be}\le K\left(\sum_{k=2^n}^{2^{n+1}-1}\a_k^2\right)^{1/2}
320: \quad\mbox{for}\quad n\ge1.
321: $$
322:
323: We can define now the function $\f$ by
324: $$
325: \f=\sum_{n\ge0}f_n.
326: $$
327: Obviously, $|\hat\f(k)|\ge \a_k$ for $k\ge0$. Let us show that $\f\in B_{\be1}^1$. We have
328: \begin{align*}
329: \qquad\sum_{n\ge1}2^n\|\f*W_n\|_{L^\be}&=\sum_{n\ge1}2^n\|\big(f_{n-1}+f_n+f_{n+1}\big)*W_n\|_{L^\be}\\[.2cm]
330: &\le\sum_{n\ge1}2^n\|\big(f_{n-1}+f_n+f_{n+1}\big)\|_{L^\be}\|W_n\|_{L^1}\\[.2cm]
331: &\le3\sum_{n\ge1}2^n\|f_n\|_{L^\be}\|W_n\|_{L^1}
332: \le3\cdot\frac32\sum_{n\ge1}2^n\|f_n\|_{L^\be}\\[.2cm]
333: &\le\frac92K
334: \sum_{n=0}^\infty2^n\left(\sum_{k=2^n}^{2^{n+1}-1}\a_k^2\right)^{1/2}<\be.\qquad\bl
335: \end{align*}
336:
337: In \cite{PS} the following problem was considered. Let $\Psi$ be the function on $(0,\be)$ defined by
338: $$
339: \Psi(t)=\left\{\begin{array}{ll}\frac32t-1,&0<t\le2,\\[.2cm]
340: t,& t>2.\end{array}\right.
341: $$
342: Let $\{\g_{j+k}\}_{j,k\ge0}$ be a Hankel matrix. The following result was proved
343: in \cite{PS} (Theorem 6.7):
344:
345: (i) {\em if $\b<\Psi(t)$, then}
346: $$
347: \sum_{k\ge0}|\g_k|^t(1+k)^\b<\be\quad\mbox{whenever}\quad\{\g_{j+k}\}_{j,k\ge0}\in\ell^1\check{\otimes}\ell^1;
348: $$
349:
350: (ii) {\em if $\b>\Psi(t)$, then}
351: $$
352: \sum_{k\ge0}|\g_k|^t(1+k)^\b=\be\quad\mbox{for some}\quad\{\g_{j+k}\}_{j,k\ge0}\in\ell^1\check{\otimes}\ell^1;
353: $$
354:
355: (iii) {\em if $\b=\Psi(t)$ and $\frac43\le t<\be$, then}
356: $$
357: \sum_{k\ge0}|\g_k|^t(1+k)^\b<\be\quad\mbox{whenever}\quad\{\g_{j+k}\}_{j,k\ge0}\in\ell^1\check{\otimes}\ell^1.
358: $$
359:
360: In \cite{PS} the problem is raised to find out whether
361: $$
362: \sum_{k\ge0}|\g_k|^t(1+k)^{\Psi(t)}
363: $$
364: has to be finite for $t\in\big(0,\frac43\big)$ whenever
365: $\{\g_{j+k}\}_{j,k\ge0}\in\ell^1\check{\otimes}\ell^1$.
366:
367: It is easy to deduce Theorem 6.7 of \cite{PS} from \rf{mod} and above Theorem \ref{LKK}. Moreover,
368: using \rf{mod} and Theorem \ref{LKK}, we can solve the problem posed in \cite{PS} and settle the case
369: $t\in\big(0,\frac43\big)$.
370:
371: \begin{thm}
372: \label{re}
373: If $1\le t<\frac43$, then
374: $$
375: \sum_{k\ge0}|\g_k|^t(1+k)^{3t/2-1}<\be
376: \quad\mbox{whenever}\quad\{\g_{j+k}\}_{j,k\ge0}\in\ell^1\check{\otimes}\ell^1.
377: $$
378: If $0<t<1$, then
379: $$
380: \sum_{k\ge0}|\g_k|^t(1+k)^{3t/2-1}=\be
381: \quad\mbox{for some}\quad\{\g_{j+k}\}_{j,k\ge0}\in\ell^1\check{\otimes}\ell^1.
382: $$
383: \end{thm}
384:
385: \Pf Suppose that $1\le t<2$. By H\"older's inequality, we have
386: \begin{align*}
387: \sum_{k\ge1}|\g_k|^t(1+k)^{3t/2-1}&\le\,
388: \const\sum_{n\ge0}2^{n(3t/2-1)}\sum_{k=2^n}^{2^{n+1}}|\g_k|^t\\[.2cm]
389: &\le\sum_{n\ge0}2^{\frac{3n}{2}t}2^{-n}\left(\sum_{k=2^n}^{2^{n+1}}|\g_k|^2\right)^{t/2}
390: 2^{n(1-t/2)}\\[.2cm]
391: &=\sum_{n\ge0}2^{nt}\left(\sum_{k=2^n}^{2^{n+1}}|\g_k|^2\right)^{t/2}.
392: \end{align*}
393: Since $t\ge1$, the $\ell^t$ norm of a sequence does not exceed its $\ell^1$ norm, and so
394: $$
395: \sum_{n\ge0}2^{nt}\left(\sum_{k=2^n}^{2^{n+1}}|\g_k|^2\right)^{t/2}\le
396: \left(\sum_{n\ge0}2^n\left(\sum_{k=2^n}^{2^{n+1}}|\g_k|^2\right)^{1/2}\right)^t.
397: $$
398: The result follows now from Theorem 5.2 of \cite{P1} and \rf{mod}.
399:
400: Suppose now that $0<t<1$. It follows from Theorem \ref{LKK} that it suffices to find a sequence
401: $\{\a_k\}_{k\ge0}$ of nonnegative numbers that satisfies \rf{maj} and such that
402: $$
403: \sum_{k\ge0}\a_k^t(1+k)^{3t/2-1}=\be.
404: $$
405: Let $\{\d_n\}_{n\ge0}$ be a sequence of positive numbers such that $\{2^{3n/2}\d_n\}_{n\ge0}\in\ell^1$, but
406: \lb$\{2^{3n/2}\d_n\}_{n\ge0}\not\in\ell^t$.
407:
408: Put
409: $$
410: \a_0=0 \quad\mbox{and}\quad\a_k=\d_n\quad\mbox{if}\quad2^n\le k\le2^{n+1}-1.
411: $$
412: We have
413: $$
414: \sum_{n\ge0}2^n\left(\sum_{k=2^n}^{2^{n+1}-1}\a_k^2\right)^{1/2}=
415: \sum_{n\ge0}2^{3n/2}\d_n<\be.
416: $$
417: However,
418: \begin{align*}
419: \qquad\sum_{k\ge0}\a_k^t(1+k)^{3t/2-1}&\ge\const\sum_{n\ge0}2^{n(3t/2-1)}\sum_{k=2^n}^{2^{n+1}}\a_k^t\\[.2cm]
420: &=\const\sum_{n\ge0}2^{n(3t/2-1)}2^n\d_n^t=\const\sum_{n\ge0}2^{3nt/2}\d_n^t=\be.\quad\bl
421: \end{align*}
422:
423: \
424:
425: \begin{thebibliography}{99}
426:
427: \bibitem[dLKK]{dLKK} {\sc K. de Leeuw, Y. Katznelson}, and {\sc J.-P. Kahane},
428: {\em Sur les coefficients de Fourier des fonctions continues},
429: C. R. Acad. Sci. Paris S\'er. A-B {\bf285} (1977), A1001--A1003.
430:
431: \bibitem[EL]{EL} {\sc P.P.B Eggermont} and {\sc Y.J. Leung}, {\em On a factorization problem for convergent
432: sequences and on Hankel forms in bounded sequences}, Proc. Amer. Math. Soc. {\bf96} (1986), 269--274.
433:
434: \bibitem[K1]{K1} {\sc S.V. Kislyakov}, {\em Fourier coefficients of boundary values of functions that are
435: analytic in the disc and bidisc} (Russian). In: Spectral theory of functions and operators, II. Trudy Mat.
436: Inst. Steklov. {\bf155} (1981), 77--94.
437:
438: \bibitem[K2]{K2} {\sc S.V. Kislyakov}, {\em Fourier coefficients of continuous functions and a class of
439: multipliers}, Ann. Inst. Fourier (Grenoble) {\bf38} (1988), 147--183.
440:
441: \bibitem[KP]{KP} {\sc S. Kwapie\'n} and {\sc A. Pe\pl czy\'nski}, {\em On two problems of S. Mazur from the
442: Scottish Book}, Lecture at the Colloquium dedicated to the memory of Stanis\pl aw Mazur, Warsaw Univ. (1985)
443: (unpublished).
444:
445: \bibitem[N]{N} {\sc F.L. Nazarov}, {\em The Bang solution of the coefficient problem} (Russian), Algebra i
446: Analiz {\bf9} (1997), 272--287;
447:
448: English Translation: St. Petersburg Math. J. {\bf9} (1998), 407--419.
449:
450: \bibitem[Pe]{Pe} {\sc J. Peetre},
451: {\em New thoughts on Besov spaces}, Duke Univ. Press., Durham, NC, 1976.
452:
453: \bibitem[PS]{PS} {\sc A. Pe\pl czy\'nski} and {\sc F. Sukochev}, {\em Some remarks on Toeplitz multipliers and
454: Hankel matrices}, Studia Math. {\bf175} (2006), 175--204.
455:
456: \bibitem[P1]{P1} {\sc V.V. Peller}, {\em Estimates of functions of power bounded operators
457: on Hilbert spaces}, J. Operator Theory {\bf7} (1982), 341--372.
458:
459: \bibitem[P2]{P2} {\sc V.V. Peller}, {\em Hankel operators and their applications,}
460: Springer-Verlag, New York, 2003.
461:
462: \bibitem[R]{R} {\sc W. Rudin}, {\em Fourier analysis on groups}, Wiley Classics
463: Library, Wiley-Interscience Publication. John Wiley \& Sons, Inc., New York, 1990.
464:
465: \bibitem[SB]{SB} {\em The Scottish Book}, R.D. Mauldin (ed.), Birkh\"auser, Boston, MA, 1979.
466:
467: \end{thebibliography}
468:
469: \
470:
471: \noindent
472: \begin{tabular}{p{8cm}p{14cm}}
473: Department of Mathematics \\
474: Michigan State University \\
475: East Lansing, Michigan 48824\\
476: USA\\
477: email: peller@math.msu.edu
478: \end{tabular}
479:
480:
481:
482:
483:
484: \end{document}
485:
486:
487:
488:
489:
490:
491:
492:
493:
494:
495:
496:
497: