Weak amenability of weighted measure algebras and their second duals

M. J. Mehdipour and A. Rejali

arXiv: 2210.04075v1 · PDF · 8 October 2022
Categories: math.FA
Licence: CC BY 4.0 The site shows the text.

Abstract

In this paper, we study the weak amenability of weighted measure algebras and prove that M(G, \omega) is weakly amenable if and only if G is discrete and every bounded quasi-additive function is inner. We also study the weak amenability of L^1(G, \omega)^{**} and M(G, \omega)^{**} and show that the weak amenability of theses Banach algebras are equivalent to finiteness of G. This gives an answer to the question concerning weak amenability of L^1(G, \omega)^{**} and M(G, \omega)^{**}.

Items 28

Proposition

Proposition 2.1

GCSS

Let \omega_i be a weight function on locally compact group G_i for i=1, 2. Then

M(G_1, \omega_1)\hat{\otimes} M(G_2, \omega_2)=M(G_1\times G_2, \omega_1\otimes\omega_2).

Used by Proof

Proof of Proposition 2.1

Let \eta_i\in M(G_i, \omega_i), for i=1, 2. Then for every f\in C_0(G_1\times G_2), we have

\langle\eta_1\otimes\eta_2, f\rangle=\int_{G_1}\int_{G_2}f(x, y)\;d\eta_1(x)d\eta_2(y).

It is easy to prove that

\eta_1\otimes\eta_2\in C_0(G_1\times G_2, 1/\omega_1\otimes\omega_2)^*= M(G_1\times G_2, \omega_1\otimes\omega_2).

Conversely, let \eta\in M(G_1\times G_2, \omega_1\otimes\omega_2). In view of Theorem Lusin’s theorem, there exists sequences (f_n) and (g_n) in the unit ball C_c(G_1, 1/\omega_1) and C_c(G_2, 1/\omega_2) with compact support, respectively, such that for almost every where x\in G_1 and y\in G_2

f_n(x)\rightarrow 1\quad\hbox{and}\quad g_n(y)\rightarrow 1

as n\rightarrow\infty. We define the functionals \eta_1 and \eta_2 by

\eta_1(f)=\lim_n\eta(f\otimes g_n)\quad\hbox{and}\quad \eta_2(g)=\lim_n\eta(f_n\otimes g)

for all f\in C_0(G_1, 1/\omega_1) and g\in C_0(G_2, 1/\omega_2). Then \eta_1\in M(G_1, \omega_1), \eta_2\in M(G, \omega_2). In fact,

|\eta_1(f)|\leq\|\eta\|\|f\|_{\infty, 1/\omega}\quad\hbox{and}\quad|\eta_2(g)|\leq\|\eta\|\|g\|_{\infty, 1/\omega}.

On the other hand,

\begin{aligned} \eta_1\otimes\eta_2(f\otimes g)&=\lim_n\eta(f\otimes g_n)\eta(f_n\otimes g)\\ &=\lim_n\int_{G_1\times G_2} f(x)f_n(x)g(y)g_n(y)\;d\eta(x, y). \end{aligned}

Since \eta is bounded, it follows from Lebesgue dominated convergence theorem that \|f_n\otimes g_n\|_{\infty, 1/\omega}\leq 1 and 1\in L^1(\eta). Furthermore, f_n\otimes g_n(x, y)\rightarrow 1 for every x\in G_1, y\in G_2. For every f\in C_0(G_1, 1/\omega_1) and g\in C_0(G_2, 1/\omega_2)

\eta_1\otimes\eta_2(f\otimes g)=\int_{G_1\times G_2}f\otimes g(x, y)\;d\eta(x, y)=\eta(f\otimes g).

It follows that for every h\in C_0(G_1, 1/\omega_1)\otimes C_0(G_2, 1/\omega_2)

\eta_1\otimes\eta_2(h)=\eta(h)

and so for every h\in C_0(G_1\times G_2, 1/\omega_1\otimes\omega_2)

\eta_1\otimes\eta_2(h)=\eta(h).

Therefore, \eta_1\otimes\eta_2=\eta.$

Proposition

Proposition 2.2

9O1B

Let G be a locally compact group and \omega be a weight function on G. If p\in D(G, \omega), then there exists a unique bounded derivation D\in {\cal Z}( M(G, \omega), M(G, \omega)^*) such that p(x, y)=\langle D(\delta_x), \delta_y\rangle for all x, y\in G, where \delta_\cdot is the Diract measure at \cdot.

Proof of Proposition 2.2

Let p\in D(G, \omega). Then \Gamma(D_1)=p for some D_1\in {\cal Z}( L^1(G, \omega), L^\infty(G, 1/\omega)). By Proposition 2.1.6 [r], there exists D_2\in {\cal Z}( M(G, \omega), L^\infty(G, 1/\omega)) such that D_2 is strict-weak^* continuous and D_2|_{L^1(G, \omega)}=D_1. Hence for every f\in L^1(G, \omega),

\begin{aligned} \langle D_2(\delta_x), f\rangle&=\lim \langle D_1(e_\alpha\ast\delta_x), f\rangle\\ &=\lim \int_G\int_G p(x, y) (e_\alpha\ast\delta_x)(z) f(z)\;dzdy\\ &=\int_G\int_G p(x, y) e_\alpha(zx^{-1}) f(y)\;dzdy. \end{aligned}

On the other hand, there exists a linear functional T_1:L^1(G\times G,\omega^\otimes )\rightarrow{\Bbb C} such that

\langle T_1, f\otimes g\rangle=\langle D_1(f), g\rangle

for all f, g\in L^1(G, \omega). Since L^1(G\times G, \omega^\otimes ) is a closed ideal in M(G\times G, \omega^\otimes ), it follows that T_1 has a strict continuous extension, say T_2:M(G, \omega)\hat{\otimes}M(G, \omega)\rightarrow{\Bbb C}. Define D: M(G, \omega)\rightarrow M(G, \omega)^* by

\langle D(\mu), \nu\rangle=\langle T_2, \mu\otimes\nu\rangle

for all \mu, \nu\in M(G, \omega). If (e_\alpha) is a bounded approximate identity of L^1(G, \omega), then for every x\in G, e_\alpha\ast\delta_x\rightarrow\delta_x in the strict topology. So

T_2(e_\alpha\ast\delta_x\otimes e_\alpha\ast\delta_y)\rightarrow T_2(\delta_x\otimes\delta_y).

Therefore,

\begin{aligned} \langle D(\delta_x), \delta_y\rangle&=\lim \langle T_2(e_\alpha\ast\delta_x\otimes e_\alpha\ast\delta_y)\\ &=\lim\langle D_2(e_\alpha\ast\delta_x),e_\alpha\ast\delta_y\rangle=p(x, y), \end{aligned}

as claimed.$

Proposition

Proposition 2.3

4678

Let G be a locally compact group and \omega be a weight function on G. Then the following statements hold.

(i) The function \Gamma: {\cal Z}( M(G, \omega), M(G, \omega)^*)\rightarrow GD(G, \omega) is an isometric isomorphism. Furthermore, \Gamma({\cal I}_{nn}(M(G, \omega),M(G, \omega)^*))= GI(G, \omega).

(ii) If D\in {\cal Z}( M(G, \omega), M(G, \omega)^*), then for every \mu\in M(G, \omega) there exists F=(F_{\mu\otimes\nu})_\nu\in GD(G, \omega) such that D(\mu)=(p_{\mu, \nu})_\nu and p_{\mu, \nu}(y)=\int_GF_{\mu\otimes\nu}(x, y)\;d\mu(x) for almost every where y\in G.

Used by Proof

Proof of Proposition 2.3

Let D\in {\cal Z}( M(G, \omega), M(G, \omega)^*). Then D\in{\cal B}(M(G, \omega), M(G, \omega)^*). Putting A=B= M(G, \omega) in the definition of \Gamma, we have

F:=\Gamma(D)\in (M(G, \omega)\hat{\otimes} M(G, \omega))^*=GL(G\times G, 1/\omega^\otimes )

and

\begin{aligned} \langle D(\mu), \nu\rangle&=\langle F, \mu\otimes\nu\rangle\\ &=\int_G F_{\mu\otimes\nu}(x, y)\; d(\mu\otimes\nu)(x, y)\\ &=\int_G\int_G F_{\mu\otimes\nu}(x, y)\; d\mu(x)d\nu(y). \end{aligned}

On the other hand, if P=(p_\mu)_{\mu\in M(G, \omega)}, then

\begin{aligned} \langle\hbox{ad}_P(\mu), \nu\rangle&=\langle P\cdot\mu-\mu\cdot P, \nu\rangle\\ &=\langle P, \mu\ast\nu\rangle-\langle P, \nu\ast\mu\rangle\\ &=\int_Gp_{\mu\ast\nu}\;d(\mu\ast\nu)-\int_Gp_{\nu\ast\mu}\; d(\nu\ast\mu)\\ &=\int_G\int_G(p_{\mu\ast\nu}(xy)-p_{\nu\ast\mu}(yx))\;d\mu(x)d\nu(y). \end{aligned}

Now, by the argument used in the proof of Theorem 2.3 in [mr12], it can be shown that the statement (i) holds. For (ii), assume that D\in {\cal Z}( M(G, \omega), M(G, \omega)^*) and \mu\in M(G, \omega). Then

D(\mu)\in M(G, \omega)^*=GL(G, 1/\omega).

Thus D(\mu)=(p_{\mu, \nu})_{\nu} for some (p_{\mu, \nu})_{\nu}\in GL(G, 1/\omega). Hence for every \nu\in M(G, \omega), we have

\langle D(\mu), \nu\rangle=\int_G p_{\mu, \nu}\; d\nu.

This together with (Proof) s shows that

p_{\mu, \nu}(y)=\int_GF_{\mu\otimes\nu}(x, y)\;d\mu(x)

for almost every where y\in G.$

Theorem

Theorem 2.4

4AHS

Let G be a locally compact group and \omega be a weight function on G. Then the following assertions are equivalent.

(a) M(G, \omega) is weakly amenable.

(b) For every D\in {\cal Z}( M(G, \omega), M(G, \omega)^*) there exists P=(p_\mu)_{\mu\in M(G, \omega)} such that \langle D(\mu), \nu\rangle=\int_G\int_G(p_{\mu\ast\nu}(xy)-p_{\nu\ast\mu}(yx))\;d\mu(x)d\nu(y) for all \mu, \nu\in M(G, \omega).

(c) Every generalized quasi-additive function is inner.

(d) M(G) is weakly amenable and D(G, \omega)=I(G, \omega).

(e) G is discrete and every non-inner quasi-additive function in L^\infty(G, 1/\omega) is unbounded.

Used by Proof, Proof

Proof of Theorem 2.4

The implications (a)\Rightarrow(b)\Rightarrow(c)\Rightarrow(a) follow from Proposition 2.3. By Theorem 1.2 in [dgh] the implication (d)\Leftrightarrow(e) holds. Also, the implication (e)\Rightarrow(a) follows from Corollary 2.5 in [mr12]. For (a)\Rightarrow(e), let M(G, \omega) be weakly amenable and \varphi be a character of M(G). If d is a continuous point derivation at \varphi on M(G), then d|_{M(G, \omega)} is a continuous point derivation of M(G, \omega) at \varphi|_{M(G, \omega)}. Hence d is zero on M(G, \omega). Since M(G, \omega) is dense in M(G), we have d=0 on M(G) which is implies G is discrete. Apply Theorem 2.4 in [mr12] to conclude that D(G, \omega)=I(G, \omega).$

Uses Proposition 2.3

Corollary

Corollary 2.5

V5D5

Let G be a locally compact group. Then the following assertions are equivalent.

(a) M(G, \omega) is weakly amenable.

(b) M(G, \omega) is cyclically weakly amenable.

(c) M(G, \omega) is point amenable.

(d) G is discrete and every non-inner quasi-additive function in L^\infty(G,1/\omega) is unbounded.

Used by Proof, Proof

Corollary

Corollary 2.6

TPCT

Let \omega and \omega_0 be weight functions on a locally compact group G. Then the following statements hold.

(i) If \omega\leq m\omega_0 for some m>0, M(G, \omega_0) is weakly amenable and I(G, \omega_0)=D(G, \omega_0) , then M(G, \omega) is weakly amenable.

(ii) If \omega and \omega_0 are equivalent, then weak amenability of M(G, \omega) is equivalent to weak amenability of M(G, \omega_0).

(iii) M(G, \omega^\prime) is weakly amenable if and only if M(G, \omega) is weakly amenable.

(iv) If M(G, \omega^*) is weakly amenable and I(G, \omega^*)=D(G, \omega^*), then M(G, \omega) is weakly amenable.

(v) If G is Abelian, then M(G, \omega^*) is weakly amenable if and only if M(\frak{D}, \omega^\otimes) is weakly amenable, where \frak{D}:=\{(x, x^{-1}): x\in G\}.

Corollary

Corollary 2.7

A0RQ

Let \omega be weight function on locally compact group G. Then the following statements hold.

(i) If \phi: G\rightarrow G is a continuous group epimorphism, M(G, \overleftarrow{\omega}) is weakly amenable and \frak{S}(I(G, \omega))=I(G, \overleftarrow{\omega}), then M(G, \omega) is weakly amenable.

(ii) If G is Abelian and M(G, \tilde{\omega}) is weakly amenable, then M(H, \omega|_H) is weakly amenable, where H is a subgroup of G.

Corollary

Corollary 2.8

OLEF

Let \omega_i be a weight function on a locally compact group G_i, for i=1, 2. Then the following assertions are equivalent.

(a) M(G_1, \omega_1)\hat{\otimes}M(G_2, \omega_2) is weakly amenable.

(b) M(G_1, \omega_1)\hat{\otimes}M(G_2, \omega_2) is cyclically weakly amenable.

(d) M(G_i, \omega_i) is weakly amenable and G_i is discrete, for i=1,2.

Used by Proof

Proof of Corollary 2.8

Let M(G_1, \omega_1)\hat{\otimes}M(G_2, \omega_2) be point amenable. Since M(G_i, \omega_i) is unital, for i=1, 2, from Proposition 2.1 we infer that then M(G_1\times G_2, \omega_1\otimes\omega_2) is point amenable. By Theorem 2.4, G_1\times G_2 is discrete. It follows that G_i is discrete, for i=1,2. Hence M(G_i, \omega_i)=\ell^1(G_i, \omega_i) and so

\ell^1(G_1, \omega_1)\hat{\otimes}\ell^1(G_1, \omega_1))=M(G_1, \omega_1)\hat{\otimes}M(G_2, \omega_2)

is weakly amenable. In view of Corollary 4.8 in [mr12], \ell^1(G_i, \omega_i) is weakly amenable. So (c) implies (d).

Let M(G_i, \omega_i) is weakly amenable and G_i is discrete, for i=1,2. By Corollary 2.5, M(G_i, \omega_i) is point amenable, for i=1,2. It follows from Theorem 4.1 in [mr5] and Proposition 2.1 that

M(G_1, \omega_1)\hat{\otimes}M(G_2, \omega_2)=M(G_1\times G_2, \omega_1\otimes\omega_2)

is point amenable. Again, apply Corollary 2.5 to conclude that M(G_1, \omega_1)\hat{\otimes}M(G_2, \omega_2) is weakly amenable. That is, (d) implies (a).$

Uses Proposition 2.1, Theorem 2.4, Corollary 2.5

Corollary

Corollary 2.9

542U

Let \omega_i be a weight function on a locally compact discrete group G_i, for i=1, 2. Then the following assertions are equivalent.

(a) \ell^1(G_1, \omega_1)\hat{\otimes}\ell^1(G_2, \omega_2) is weakly amenable.

(b) \ell^1(G_1, \omega_1)\hat{\otimes}\ell^1(G_2, \omega_2) is cyclically weakly amenable.

(d) \ell^1(G_i, \omega_i) is weakly amenable and G_i is discrete, for i=1,2.

Corollary

Corollary 2.10

QFWM

Let G be a locally compact group. Then the following assertions are equivalent.

(a) M_*(G, \omega)^* is weakly amenable.

(b) M_*(G, \omega)^* is cyclically weakly amenable.

(c) M_*(G, \omega)^* is point amenable.

(d) G is discrete and every non-inner quasi-additive function in L^\infty(G, 1/\omega) is unbounded.

Used by Proof

Proof of Corollary 2.10

Let M_*(G, \omega)^* be point amenable. Since M(G, \omega) is a direct summand of M_*(G, \omega)^*, by Theorem 3.7 in [mr5], M(G, \omega) is point amenable. Hence G is discrete and every non-inner quasi-additive function in L^\infty(G, 1/\omega) is unbounded. Thus (c) implies (d). It is easy to see that if G discrete, then

M_*(G, \omega)*=\ell^1(G, \omega)=M(G, \omega).

It follows that (d) implies (a).$

Corollary

Corollary 2.11

HX00

Let G be a locally compact group. Then the following assertions are equivalent.

(a) L_0^\infty(G, 1/\omega)^* is weakly amenable.

(b)L_0^\infty(G, 1/\omega)^* is cyclically weakly amenable.

(c) L_0^\infty(G, 1/\omega)^* is point amenable.

(d) G is discrete and every non-inner quasi-additive function in L^\infty(G, 1/\omega) is unbounded.

Corollary

Corollary 2.12

GNBM

Let \omega_i be a weight function on a locally compact group G_i, for i=1, 2. Then the following assertions are equivalent.

(a) M_*(G_1, \omega_1)^* and M_*(G_2, \omega_2)^* are weakly amenable.

(b) L_0^\infty(G, 1/\omega)^* and L_0^\infty(G, 1/\omega)^* are weakly amenable.

(c) M_*(G_1, \omega_1)^*\hat{\otimes}M_*(G_2, \omega_2)^* is weakly amenable and G_i is discrete, for i=1,2.

(d) L_0^\infty(G, 1/\omega)^*\hat{\otimes}L_0^\infty(G, 1/\omega)^* is weakly amenable and G_i is discrete, for i=1,2.

Proof of Corollary 2.12

Assume that M_*(G_1, \omega_1)^* and M_*(G_2, \omega_2)^* are weakly amenable. By Corollary 2.10, G_i is discrete and M(G_i, \omega_i)=M_*(G_i, \omega_i)^* is weakly amenable, for i=1,2. It follows from Corollary 2.8 that

\begin{aligned} M_*(G_1, \omega_1)^*\hat{\otimes}M_*(G_2, \omega_2)^*=M(G_1, \omega_1)\hat{\otimes}M(G_2, \omega_2) \end{aligned}

is weakly amenable. So (a) implies (c).

Let M_*(G_1, \omega_1)^*\hat{\otimes}M_*(G_2, \omega_2)^* be weakly amenable and G_i is discrete, for i=1,2. This implies that M(G_1, \omega_1)\hat{\otimes}M(G_2, \omega_2) is weakly amenable. Thus M_*(G_i, \omega_i)^*=M(G_i, \omega_i) is weakly amenable. Hence (c) implies (a). Similarly, (b) and (d) are equivalent.$

Uses Corollary 2.10, Corollary 2.8

Corollary

Corollary 2.13

HMMI

Let \hbox{WAP}(L^1(G, \omega))^* or \hbox{LUC}(G, \omega)^* be 0-point amenable. Then G is discrete.

Proposition

Proposition 2.14

J2QS

Let G be a locally compact group. If \frak{A} is cyclically amenable, then every element of CD(G, \omega) is inner.

Proof of Proposition 2.14

Let M(G, \omega) be cyclically amenable. Since L^1(G, \omega) is a direct summand of M(G, \omega), by Theorem 3.7 in [mr5], the Banach algebra L^1(G, \omega) is cyclically amenable. It follows from Theorem 5.6 in[mr12] that every element of CD(G, \omega) is inner. For the other cases, we only need to recall that

\frak{A}=M(G, \omega)\oplus\frak{B}

for some closed subspace \frak{B} of \frak{A}.$

Theorem

Theorem 3.1

1HFI

Let G be a locally compact group. Then the following assertion are equivalent.

(a) L^1(G, \omega)^{**} is weakly amenable.

(b) L^1(G, \omega)^{**} is cyclically weakly amenable.

(c) L^1(G, \omega)^{**} is point amenable.

(d) G is finite.

Used by Proof

Proof of Theorem 3.1

Let \iota: L^1(G, \omega)\rightarrow L^1(G) be the inclusion map. Since L^1(G, \omega) is dense in L^1(G), \iota is a continuous homomorphism with dense range. So \iota^{**}: L^1(G,\omega)^{**}\rightarrow L^1(G)^{**} is epimorphism. Hence if L^1(G, \omega)^{**} is point amenable, then by Theorem 2.1 in [mr5] the Banach algebra L^1(G)^{**} is point amenable. It follows that every continuous point derivation of L^1(G)^{**} at the discrete augmentation character is zero. From Theorem 11.17 in [dls] infer that G is finite. So (c)\Rightarrow(d). The implications (a)\Rightarrow(b)\Rightarrow(c) follows from Theorem 4.1 in [mr4]. $

Corollary

Corollary 3.2

TMZB

Let G be a locally compact group. Then the following assertion are equivalent.

(a) M(G, \omega)^{**} is weakly amenable.

(b) M(G, \omega)^{**} is cyclic amenable.

(c) M(G, \omega)^{**} is point amenable.

(d) G is finite.

Proof of Corollary 3.2

Let M(G, \omega)^{**} is point amenable. By Proposition 5.2 in [mr5], the Banach algebra M(G, \omega) is point amenable. In view of Corollary 2.5, G is discrete. Hence L^1(G, \omega)^{**} is weakly amenable. Now, apply Theorem 3.1.$

Uses Corollary 2.5, Theorem 3.1

Theorem

Theorem 3.3

9HB0

Let G be a connected locally compact group. If either G_d is amenable or G is an [IN]-group, then the following assertions are equivalent.

(a) L^1(G, \omega)^{**} is weakly amenable.

(b) M(G, \omega) is weakly amenable.

(c) G=\{e\}.

Proof of Theorem 3.3

Let L^1(G, \omega)^{**} be weakly amenable. Since

L^1(G, \omega)^{**}=M(G, \omega)\oplus C_0(G, \omega)^\perp

and C_0(G, \omega)^\perp is an ideal in L^1(G, \omega)^{**} , we have M(G, \omega) is weakly amenable. So (a)\Rightarrow(b). Let’s show that (b)\Rightarrow(c). To this end, let M(G, \omega) be weakly amenable. It follows from Theorem 2.4 that G discrete and M(G) is weakly amenable. If G_d is amenable, then from Theorem 3.3 in [ll] we infer that G=\{e\}. If G is an [IN]-group, then by Theorem 3.4 in [ll], G is compact. Since G is also discrete, it follows that G is finite. Hence G_d is amenable. Thus G=\{e\}. So (b)\Rightarrow(c). The implication (c)\Rightarrow(a) is clear. $

Uses Theorem 2.4

Source and licence

The text on this page is from arXiv:2210.04075v1 by M. J. Mehdipour and A. Rejali, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/2210/2210.04075.jsonl.