Content uploaded by Taher Yazdanpanah
Author content
All content in this area was uploaded by Taher Yazdanpanah
Content may be subject to copyright.
Czechoslovak Mathematical Journal
Alireza Medghalchi; Taher Yazdanpanah
Problems concerning n-weak amenability of a Banach algebra
Czechoslovak Mathematical Journal, Vol. 55 (2005), No. 4, 863--876
Persistent URL: http://dml.cz/dmlcz/128029
Terms of use:
© Institute of Mathematics AS CR, 2005
Institute of Mathematics of the Academy of Sciences of the Czech Republic provides access to
digitized documents strictly for personal use. Each copy of any part of this document must contain
these Terms of use.
This paper has been digitized, optimized for electronic delivery and stamped
with digital signature within the project DML-CZ: The Czech Digital Mathematics
Library http://project.dml.cz
Czechoslovak Mathematical Journal, 55 (130) (2005), 863–876
PROBLEMS CONCERNING n-WEAK AMENABILITY
OF A BANACH ALGEBRA
, Teheran, and , Boushehr
(Received October 24, 2002)
Abstract. In this paper we extend the notion of n-weak amenability of a Banach algebra A
when n ∈ . Technical calculations show that when A is Arens regular or an ideal in A
∗∗
,
then A
∗
is an A
(2n)
-module and this idea leads to a number of interesting results on
Banach algebras. We then extend the concept of n-weak amenability to n ∈
.
Keywords: Banach algebra, weakly amenable, Arens regular, n-weakly amenable
MSC 2000 : 46H20, 46H40
1. Introduction
Let A be a Banach algebra, X a Banach A -bimodule. Then we denote by X
∗
the
topological dual space of X; the value of x
∗
∈ X
∗
at x ∈ X is denoted by hx, x
∗
i.
We recall that X
∗
is a Banach A -bimodule under the actions
hx, ax
∗
i = hxa, x
∗
i, hx, x
∗
ai = hax, x
∗
i (a ∈ A , x ∈ X, x
∗
∈ X
∗
).
A derivation D : A −→ X is a (bounded) linear map such that
D(ab) = D(a)b + aD(b) (a, b ∈ A ).
For each x ∈ X, δ
x
(a) = ax − xa is a derivation, which is called inner. The first
cohomology group H
1
(A , X) is the quotient of the space of derivations by the in-
ner derivations, and in many situations triviality of this space is of considerable
importance. In particular, A is called contractible if H
1
(A , X) = {0} for every
Banach A -bimodule X, A is called amenable if H
1
(A , X
∗
) = {0} for every Ba-
nach A -bimodule X, A is called n-weakly amenable if H
1
(A , A
(n)
) = {0}, and
863
weakly amenable if A is 1-weakly amenable. For the theory of amenable and weakly
amenable Banach algebras see [1], [2], [4], [6], [8] and [9] for example.
Let A be a Banach algebra. Given a
∗
∈ A
∗
and F ∈ A
∗∗
, then F a
∗
and a
∗
F are
defined in A
∗
by the formulae
ha, F a
∗
i = ha
∗
a, F i, ha, a
∗
F i = haa
∗
, F i (a ∈ A ).
Next, for F, G ∈ A
∗∗
, F G and F 4 G are defined in A
∗∗
by the formulae
ha
∗
, F Gi = hGa
∗
, F i, ha
∗
, F 4 Gi = ha
∗
F, Gi (a
∗
∈ A
∗
).
Then A
∗∗
is a Banach algebra with respect to either of the products and 4. These
products are called the first and second Arens products on A
∗∗
, respectively. The
algebra A is called Arens regular if the two products and 4 coincide. For the
general theory of Arens products, see [5] and [10], for example.
Let A be a Banach algebra, n ∈ ∪ {0} and let P
n
: A
(n)
−→ A
(n+2)
be the
natural embedding, i.e., hϕ
n+1
, P
n
ϕ
n
i = hϕ
n
, ϕ
n+1
i (ϕ
n
∈ A
(n)
, ϕ
n+1
∈ A
(n+1)
),
where A
(0)
= A and A
(n)
is the nth dual of A . We shall require the following
standard properties of the Arens products. Suppose (a
α
) and (b
β
) are nets in A
with P
0
a
α
−→ F and P
0
b
β
−→ G in (A
∗∗
, σ), where σ = σ(A
∗∗
, A
∗
) is the weak
∗
topology on A
∗∗
. Then F G = lim
α
lim
β
P
0
(a
α
b
β
) and F 4 G = lim
β
lim
α
P
0
(a
α
b
β
)
in (A
∗∗
, σ). Also, for a ∈ A and F ∈ A
∗∗
, we have P
0
(a) 4 F = P
0
(a) F and
F 4 P
0
(a) = F P
0
(a).
By easy calculations we can obtain the following properties of the P
n
maps.
Lemma 1.1. Let m ∈
and n ∈ ∪ {0}. Then
(i) P
∗∗
n
P
n
= P
n+2
P
n
;
(ii) P
∗
n
P
n+1
= id;
(iii) P
(2m+1)
n
P
n+2m+1
. . . P
n+3
P
n+1
= P
n+2m−1
. . . P
n+3
P
n+1
;
(iv) P
(2m)
n
P
n+2m−2
= P
n+2m
P
(2m−2)
n
.
Lemma 1.2. Let A be a Banach algebra, n ∈
and let D : A −→ A
(n)
be a
derivation. Then P
∗
n−1
P
∗
n+1
. . . P
∗
n+2m−3
D
(2m)
P
2m−2
P
2m−4
. . . P
0
= D (m ∈ ).
. It is enough to show that P
∗
n+(2m−3)
D
(2m)
P
2m−2
= D
(2m−2)
for all
m ∈ . For ϕ ∈ A
(2m−2)
and ψ ∈ A
(n+2m−3)
we have
hψ, P
∗
n+2m−3
D
(2m)
P
2m−2
(ϕ)i = hD
(2m−1)
P
n+2m−3
(ψ), P
2m−2
(ϕ)i
hϕ, D
(2m−1)
P
n+2m−1
(ψ)i = hψ, D
(2m−2)
(ϕ)i,
and so P
∗
n+(2m−3)
D
(2m)
P
2m−2
= D
(2m−2)
.
864
2. When A
(m)
is an A
(2n)
-module?
Let A be a Banach algebra. Clearly A
(4)
is a Banach algebra with four Arens
products. We denote these algebras by (A
4
, ) = ((A
∗∗
, )
∗∗
, ), (A
4
, 4) =
((A
∗∗
, 4)
∗∗
, ), (A
4
, 4) = ((A
∗∗
, )
∗∗
, 4), (A
4
, 44) = ((A
∗∗
, 4)
∗∗
, 4). For
a ∈ A and ϕ ∈ A
(4)
it is easy to check that
P
2
P
0
(a) ϕ = P
2
P
0
(a) 4 ϕ = P
2
P
0
(a) 4 ϕ = P
2
P
0
(a) 44 ϕ,
ϕ P
2
P
0
(a) = ϕ 4 P
2
P
0
(a) = ϕ 4 P
2
P
0
(a) = ϕ 44 P
2
P
0
(a).
Let A be a Banach algebra and n ∈ . Consider the maps (a
∗
, ϕ
2n
) 7→ a
∗
· ϕ
2n
and
(a
∗
, ϕ
2n
) 7→ ϕ
2n
· a
∗
from A
∗
× A
(2n)
into A
∗
defined by
ha, a
∗
· ϕ
2n
i = hP
2n−3
. . . P
3
P
1
(aa
∗
), ϕ
2n
i,
ha, ϕ
2n
· a
∗
i = hP
2n−3
. . . P
3
P
1
(a
∗
a), ϕ
2n
i (a ∈ A ).
Then a
∗
· ϕ
2n
= a
∗
P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
2n
) and ϕ
2n
· a
∗
= P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
2n
)a
∗
.
Clearly these maps are continuous and bilinear. Note that with respect to these
actions A
∗
is not necessarily a Banach A
(2n)
-module. By dualizing these actions
we obtain continuous bilinear maps from A
(m)
× A
(2n)
into A
(m)
for every m ∈
.
For example, for F ∈ A
∗∗
and ϕ
2n
∈ A
(2n)
we have
ha
∗
, F · ϕ
2n
i = hϕ
2n
· a
∗
, F i
= hP
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
2n
)a
∗
, F i
= ha
∗
, F P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
2n
)i (a
∗
∈ A
∗
),
and so F ·ϕ
2n
= F P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
2n
). Similarly, ϕ
2n
·F = P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
2n
)4
F . From now on we regard these actions as A
(2n)
-actions on A
(m)
induced from A
∗
.
Now consider the maps (F, ϕ
2n
) 7→ F ·ϕ
2n
and (F, ϕ
2n
) 7→ ϕ
2n
·F from A
∗∗
×A
(2n)
into A
∗∗
defined by
ha
∗
, F · ϕ
2n
i = hP
2n−3
. . . P
3
P
1
(a
∗
F ), ϕ
2n
i,
ha
∗
, ϕ
2n
· F i = hP
2n−3
. . . P
3
P
1
(F a
∗
), ϕ
2n
i (a
∗
∈ A
∗
).
Clearly these are continuous bilinear maps, F · ϕ
2n
= F 4 P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
2n
) and
similarly ϕ
2n
· F = P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
2n
) F . Note that these actions are different
from the actions induced from A
∗
. Again by dualizing these actions we have con-
tinuous bilinear maps from A
(m)
× A
(2n)
into A
(m)
for every m > 2. So we have
A
(2n)
-actions on A
(m)
(m > 2) induced from A
∗∗
.
865
Let A be a Banach algebra and let n, k ∈ be such that n > 2k. Set B =
(A
(2k)
, ·), where · is one of the 2
k
Arens products on A
(2k)
. Then B is a Banach
algebra and B
∗
is a Banach B-module. By a similar argument we have continuous
bilinear maps from B
∗
× A
(2n)
into B
∗
and from B
∗∗
× A
(2n)
into B
∗∗
. Therefore
for every m > 2k + 1 we have A
(2n)
-actions on A
(m)
induced from B
∗
and for every
m > 2k + 2 we have A
(2n)
-actions on A
(m)
induced from B
∗∗
.
Proposition 2.1. Let A be an Arens regular Banach algebra and n ∈ . Then
A
∗
is a Banach A
(2n)
-bimodule with actions induced from A
∗
and any of Arens
products on A
(2n)
. In particular, A
(m)
is a Banach A
(2n)
-bimodule by actions
induced from A
∗
.
. When n = 1, one can immediately see that A
∗
is a left Banach
(A
∗∗
, )-module and a right Banach (A
∗∗
, 4)-module. Since A is Arens regular,
A
∗
is a left and right Banach A
∗∗
-module. For a ∈ A , a
∗
∈ A and F, G ∈ A
∗∗
we
have
ha, (F a
∗
)Gi = h(aF )a
∗
, Gi = ha
∗
, G (aF )i
= ha
∗
G, P
0
(a) F i = hF (a
∗
G), P
0
(a)i
= ha, F (a
∗
G)i,
and so (F a
∗
)G = F (a
∗
G). Hence A
∗
is a Banach A
∗∗
-bimodule. Now suppose the
result has been proved for n. We may assume that A
(2n+2)
= ((A
(2n)
)
∗∗
, ). Let
a ∈ A , a
∗
∈ A
∗
, ϕ, ψ ∈ A
(2n+2)
and let (ϕ
α
) , (ψ
β
) be nets in A
(2n)
such that
P
2n
(ϕ
α
) −→ ϕ and P
2n
(ψ
β
) −→ ψ in the weak
∗
topology. Then
ha, a
∗
· (ϕ ψ)i = lim
α
lim
β
hP
2n−3
. . . P
3
P
1
(aa
∗
), ϕ
α
ψ
β
i
= lim
α
lim
β
ha, a
∗
· (ϕ
α
ψ
β
)i
= lim
α
lim
β
ha, (a
∗
· ϕ
α
) · ψ
β
i
= lim
α
lim
β
ha, a
∗
P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
α
)P
∗
1
P
∗
3
. . . P
∗
2n−3
(ψ
β
)i
= lim
α
lim
β
hP
2n−3
. . . P
3
P
1
(aa
∗
P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
α
)), ψ
β
i
= lim
α
haa
∗
P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
α
), P
∗
1
P
∗
3
. . . P
∗
2n−1
(ψ)i
= lim
α
hP
∗
1
. . . P
∗
2n−1
(ψ)aa
∗
, P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
α
)i
= ha, a
∗
P
∗
1
. . . P
∗
2n−1
(ϕ)P
∗
1
. . . P
∗
2n−1
(ψ)i
= ha, (a
∗
· ϕ) · ψi,
866
and so a
∗
· (ϕ ψ) = (a
∗
· ϕ) · ψ. Similarly (ϕ ψ) · a
∗
= ϕ · (ψ · a
∗
). On the other
hand,
(ϕ · a
∗
) · ψ = (P
∗
1
. . . P
∗
2n−1
(ϕ)a
∗
)P
∗
1
. . . P
∗
2n−1
(ψ)
= P
∗
1
. . . P
∗
2n−1
(ϕ)(a
∗
P
∗
1
. . . P
∗
2n−1
(ψ))
= ϕ · (a
∗
· ψ).
Hence A
∗
is a Banach A
(2n+2)
-bimodule. So we are done by induction.
Proposition 2.2. Let A be an Arens regular Banach algebra and n ∈ . Then
with any of the Arens products on A
(2n)
, the A
(2n)
-actions on A
∗∗
induced from A
∗
and A
∗∗
coincide. In particular, A
∗∗
is a Banach A
(2n)
-bimodule with any of these
actions.
is straightforward.
Let A be a Banach algebra and m, n ∈ . Then A
(2m)
is a Banach algebra
with one of the 2
m
Arens products. We recall that every closed subalgebra of an
Arens regular Banach algebra is Arens regular. In particular, when A
(2m)
is Arens
regular for a Banach algebra A and m ∈ , then A
∗∗
, A
(4)
, . . . , A
(2m−2)
are Arens
regular, and these algebras have only one Arens product. The following proposition
is a generalization of Proposition 2.1 and Proposition 2.2.
Proposition 2.3. Let A be a Banach algebra and let n, m ∈ be such
that n > 2m. If A
(2m)
is Arens regular, then A
(2m+1)
and A
(2m+2)
are Banach
A
(2n)
-bimodules with actions induced from A
(2m+1)
. Moreover, the A
(2n)
-actions
on A
(2m+2)
induced from A
(2m+1)
and A
(2m+2)
coincide.
Definition 2.4. Let A be a Banach algebra. A is called completely Arens
regular, if for every n ∈ , A
(2n)
is Arens regular.
It is well known that every C
∗
-algebra is Arens regular and the second dual of a
C
∗
-algebra is a C
∗
-algebra. Therefore, every C
∗
-algebra is completely Arens regular.
Proposition 2.5. Let A be a completely Arens regular Banach algebra. Then
A
(m)
is a Banach A
(2n)
-module with actions induced A
(m)
.
. A direct consequence of Proposition 2.3.
867
Lemma 2.6. Let A be a Banach algebra and P
0
(A ) a left (right) ideal in A
∗∗
.
Then A
∗
is a Banach (A
∗∗
, )-module((A
∗∗
, 4))-module).
. For a
∗
∈ A
∗
, a ∈ A and F, G ∈ A
∗∗
we have
ha, a
∗
(F G)i = hG(aa
∗
), F i = hG 4 P
0
(a)a
∗
, F i
= ha
∗
, F 4 G 4 P
0
(a)i = h(a
∗
F )G, P
0
(a)i
= ha, (a
∗
F )Gi
and
ha, F (a
∗
G)i = ha
∗
G P
0
(a), F i = ha
∗
, G P
0
(a) F i
= hF (a
∗
G), P
0
(a)i = ha, F (a
∗
G)i.
Therefore A
∗
is a Banach (A
∗∗
, )-module.
Lemma 2.7. Let A be a Banach algebra. Then P
0
(A ) is an ideal in A
∗∗
with
any of the Arens products if and only if P
2
P
0
(A ) is an ideal in A
(4)
with any of the
Arens products.
. Let P
0
(A ) be an ideal in A
∗∗
. For a ∈ A and ϕ ∈ A
(4)
, one can
immediately see that
P
2
P
0
(a) ϕ = P
2
(P
0
(a) P
∗
1
(ϕ)) and ϕ P
2
P
0
(a) = P
2
(P
∗
1
(ϕ) P
0
(a)).
Therefore P
2
P
0
(A ) is an ideal in A
(4)
with any of the Arens products on A
(4)
.
Conversely, let P
2
P
0
(A ) be an ideal in A
(4)
. Take a ∈ A and F ∈ A
∗∗
. It is easy
to see that P
2
(P
0
(a)F ) = P
2
P
0
(a)P
2
(F ) ∈ P
2
P
0
(A ). Hence P
0
(a)F ∈ P
0
(A )
and similarly F P
0
(a) ∈ P
0
(A ). Therefore P
0
(A ) is an ideal in A
∗∗
.
Proposition 2.8. Let A be a Banach algebra and P
0
(A ) an ideal in A
∗∗
. Then
A
∗
is a Banach A
(2n)
-module with any of the Arens products on A
(2n)
(n ∈ ).
. When n = 1 the result is true by Lemma 2.6. Now suppose, inductively,
the result has been proved for n−1. We may assume that A
(2n+2)
= ((A
(2n)
)
∗∗
, ).
Let a ∈ A , a
∗
∈ A
∗
and ϕ, ψ ∈ A
(2n+2)
, and let (ϕ
α
), (ψ
β
) be nets in A
(2n)
such
that P
2n
(ϕ
α
) −→ ϕ and P
2n
(ψ
β
) −→ ψ in the weak
∗
topology. Now we have
ha, a
∗
· (ϕ ψ)i = hP
2n−1
. . . P
3
P
1
(aa
∗
), ϕ ψi
= lim
α
lim
β
ha
∗
, P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
α
) 4 P
∗
1
P
∗
3
. . . P
∗
2n−3
(ψ
β
) 4 P
0
(a)i
= lim
α
haa
∗
P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
α
), P
∗
1
P
∗
3
. . . P
∗
2n−1
(ψ)i
868
= lim
α
ha
∗
, P
∗
1
P
∗
3
. . . P
∗
2n−3
(ϕ
α
) P
∗
1
P
∗
3
. . . P
∗
2n−1
(ψ) P
0
(a)i
= haa
∗
, P
∗
1
P
∗
3
. . . P
∗
2n−1
(ϕ) P
∗
1
P
∗
3
. . . P
∗
2n−1
(ψ)i
= ha, (a
∗
· ϕ) · ψi,
and so a
∗
· (ϕ ψ) = (a
∗
· ϕ) · ψ. Similarly (ϕ ψ) · a
∗
= ϕ · (ψ · a
∗
). Since A
∗
is a
Banach A
∗∗
-module,
(ϕ · a
∗
) · ψ = (P
∗
1
P
∗
3
. . . P
∗
2n−1
(ϕ)a
∗
)P
∗
1
P
∗
3
. . . P
∗
2n−1
(ψ)
= P
∗
1
P
∗
3
. . . P
∗
2n−1
(ϕ)(a
∗
P
∗
1
P
∗
3
. . . P
∗
2n−1
(ψ))
= ϕ · (a
∗
· ψ).
Consequently, A
∗
is a Banach A
(2n)
-module.
Proposition 2.9. Let A be a Banach algebra. Then P
2
((A
∗∗
, )) is a left (right,
two-sided) ideal in (A
(4)
, ) if and only if P
∗
2
is an A
(4)
-module homomorphism
between left (right, two-sided) Banach A
(4)
-modules.
. Let P
2
((A
∗∗
, )) be a left ideal in (A
(4)
, ). For F ∈ A
∗∗
, ϕ
4
∈ A
(4)
and ϕ
5
∈ A
(5)
we have
hF, P
∗
2
(ϕ
4
ϕ
5
)i = hP
2
(F ), ϕ
4
ϕ
5
i = hP
2
(F ) ϕ
4
, ϕ
5
i
= hP
∗
2
(ϕ
5
), P
2
(F ) ϕ
4
i = hF, ϕ
4
P
∗
2
(ϕ
5
)i.
Hence P
∗
2
is an A
(4)
-module homomorphism between left Banach A
(4)
-modules.
Conversely, for F ∈ A
∗∗
, ϕ
4
∈ A
(4)
it is easy to see that
P
2
(F ) ϕ
4
= P
2
(P
∗
1
(P
2
(F ) ϕ
4
)),
so P
2
((A
∗∗
, )) is a left ideal in (A
(4)
, ).
3. N -weak amenability for N ∈ Z
Lemma 3.1. Let A be a Banach algebra and D : A −→ A
∗
a derivation. Then
(i) D
∗∗
: (A
∗∗
, ) −→ (A
∗∗
)
∗
is satisfied in
D
∗∗
(F G) = D
∗∗
(F )G + P
∗∗
0
(F )D
∗∗
(G) (F, G ∈ A
∗∗
),
(ii) D
∗∗
: (A
∗∗
, 4) −→ (A
∗∗
)
∗
is satisfied in
D
∗∗
(F 4 G) = D
∗∗
(F )P
∗∗
0
(G) + F D
∗∗
(G) (F, G ∈ A
∗∗
).
869
. (i) Let F, G ∈ A
∗∗
and let (a
α
), (b
β
) be nets in A such that
P
0
(a
α
) −→ F and P
0
(b
β
) −→ G in the weak
∗
topology. We have
hH, D
∗∗
(F G)i = hD
∗
(H), F Gi
= lim
α
lim
β
hD(a
α
)b
β
+ a
α
D(b
β
), Hi
= lim
α
lim
β
hb
β
, HD(a
α
) + D
∗
(Ha
α
)i
= lim
α
ha
α
, D
∗
(G H) + P
∗
0
(D
∗∗
(G)H)i
= hH, D
∗∗
(F )G + P
∗∗
0
(F )D
∗∗
(G)i
and so D
∗∗
(F G) = D
∗∗
(F )G + P
∗∗
0
(F )D
∗∗
(G).
(ii) The proof is similar to (i).
Corollary 3.2. Let A be a Banach algebra and D : A −→ A
∗
a derivation.
Then
(i) D
∗∗
: (A
∗∗
, ) −→ (A
∗∗
)
∗
is a derivation if and only if P
∗∗
0
(F )D
∗∗
(G) =
F D
∗∗
(G) for F, G ∈ A
∗∗
;
(ii) D
∗∗
: (A
∗∗
, 4) −→ (A
∗∗
)
∗
is a derivation if and only if D
∗∗
(F )P
∗∗
0
(G) =
D
∗∗
(F )G for F, G ∈ A
∗∗
.
Definition 3.3. Let A be a Banach algebra m, n ∈ , and 1 6 m < 2n. The Ba-
nach algebra A
(2n)
is called (−m)-weakly amenable, if A
(2n−m)
is a Banach A
(2n)
-
bimodule with actions induced from A
(2n−m)
and H
1
(A
(2n)
, A
(2n−m)
) = {0}.
Theorem 3.4. Let A be a Banach algebra and P
0
(A ) a left (right) ideal in A
∗∗
.
If (A
∗∗
, )((A
∗∗
, 4)) is (−1)-weakly amenable, then A is weakly amenable.
. By Lemma 2.6, A
∗
is a Banach (A
∗∗
, )-module. Let D : A −→ A
∗
be a derivation. Put d = P
∗
0
D
∗∗
: (A
∗∗
, ) −→ A
∗
. For F, G ∈ A
∗∗
, a ∈ A we
have
ha, P
∗
0
(D
∗∗
(F )G)i = hG P
0
(a), D
∗∗
(F )i = hd(F ), G 4 P
0
(a)i = ha, d(F )Gi
and
ha, P
∗
0
(P
∗∗
0
(F )D
∗∗
(G))i = hP
∗
0
(D
∗∗
(G)P
0
(a)), F i = hd(G)P
0
(a), F i = ha, F d(G)i.
Therefore, by Lemma 3.1, d is a derivation. Since H
1
((A
∗∗
, ), A
∗
) = {0}, there
exists a
∗
∈ A
∗
such that d = δ
a
∗
. Using Lemma 1.2 we obtain
aa
∗
− a
∗
a = P
0
(a)a
∗
− a
∗
P
0
(a) = dP
0
(a)
= P
∗
0
D
∗∗
P
0
(a) = D(a) (a ∈ A ).
Hence D = δ
a
∗
is an inner derivation.
870
Theorem 3.5. Let A be a Banach algebra. If P
0
(A ) is an ideal in A
∗∗
and the
Banach algebra A
(2n)
(n ∈ ) with one of 2
n
Arens products is (−2n + 1)-weakly
amenable, then A is weakly amenable.
. Let D : A −→ A
∗
be a derivation. We claim that
d
n
= P
∗
0
P
∗∗∗
0
. . . P
(2n−1)
0
D
(2n)
: A
(2n)
−→ A
∗
is a derivation. By Proposition 3.4, the result is true for n = 1. Now suppose,
inductively, that the result has been proved for n. We may suppose that A
(2n+2)
=
((A
(2n)
)
∗∗
, ). For a ∈ A , a
∗
∈ A
∗
, ϕ, ψ ∈ A
(2n+2)
, let (ϕ
α
) and (ψ
β
) be nets
in A
(2n)
such that P
2n
(ϕ
α
) −→ ϕ and P
2n
(ψ
β
) −→ ψ in the weak
∗
topology. Then
we have
ha, d
n+1
(ϕ ψ)i = lim
α
lim
β
ha, d
n
(ϕ
α
)ψ
β
+ ϕ
α
d
n
(ψ
β
)i
= lim
α
lim
β
hP
2n−3
. . . P
3
P
1
(ad
n
(ϕ
α
)), ψ
β
i
+ lim
α
lim
β
hψ
β
, D
(2n+1)
P
(2n)
0
. . . P
∗∗
0
(aP
∗
1
. . . P
∗
2n−3
(ϕ
α
))i
= lim
α
had
n
(ϕ
α
), P
∗
1
. . . P
∗
2n−1
(ψ)i
+ lim
α
hd
n+1
(ψ)a, P
∗
1
. . . P
∗
2n−3
(ϕ
α
)i
= ha, d
n+1
(ϕ) · ψ + ϕ · d
n+1
(ψ)i,
so d
n+1
is a derivation. Since H
1
(A
(2n)
, A
∗
) = {0}, there exists a
∗
∈ A
∗
such that
d
n
= δ
a
∗
. Using Lemma 1.1 (iv) and Lemma 1.2, we conclude that
aa
∗
− a
∗
a = P
2n−2
. . . P
2
P
0
(a) · a
∗
− a
∗
· P
2n−2
. . . P
2
P
0
(a)
= d
n
P
2n−2
. . . P
2
P
0
(a) = D(a) (a ∈ A ).
Hence D = δ
a
∗
is inner.
Lemma 3.6. Let A be a Banach algebra, n ∈
and let D : A −→ A
(2n)
be a
derivation. Then for every F, G ∈ A
∗∗
(i) D
∗∗
: (A
∗∗
, ) −→ ((A
(2n)
)
∗∗
, ) holds in
D
∗∗
(F G) = D
∗∗
(F ) P
∗∗
2n−2
. . . P
∗∗
2
P
∗∗
0
(G) + P
∗∗
2n−2
. . . P
∗∗
2
P
∗∗
0
(F ) D
∗∗
(G).
(ii) D
∗∗
: (A
∗∗
, 4) −→ ((A
(2n)
)
∗∗
, 4) holds in
D
∗∗
(F 4G) = D
∗∗
(F )4P
∗∗
2n−2
. . . P
∗∗
2
P
∗∗
0
(G)+P
∗∗
2n−2
. . . P
∗∗
2
P
∗∗
0
(F )4D
∗∗
(G).
is straightforward.
871
Proposition 3.7. Let A be a Banach algebra, n ∈ and let D : A −→ A
(2n)
be a derivation. If A
(2n)
is Arens regular and
D
∗∗
(A
∗∗
)A
(2n+1)
∪ A
(2n+1)
D
∗∗
(A
∗∗
) ⊆ P
2n−1
. . . P
3
P
1
(A
∗
),
then D
∗∗
: A
∗∗
−→ (A
(2n)
)
∗∗
is a derivation.
. Since A
(2n)
is Arens regular, A is Arens regular. For ϕ
2n+1
∈ A
(2n+1)
,
F, G ∈ A
∗∗
, there exists a
∗
∈ A
∗
such that
ϕ
2n+1
· D
∗∗
(F ) = P
2n−1
. . . P
1
(a
∗
).
By Lemma 1.1 we have
hϕ
2n+1
, D
∗∗
(F ) P
∗∗
2n−2
. . . P
∗∗
0
(G)i = hP
2n−1
. . . P
1
(a
∗
), P
∗∗
2n−2
. . . P
∗∗
0
(G)i
= hP
2n−2
. . . P
2
(G), ϕ
2n+1
D
∗∗
(F )i
= hϕ
2n+1
, D
∗∗
(F )Gi.
Similarly, P
∗∗
2n−2
. . . P
∗∗
0
(F ) D
∗∗
(G) = F D
∗∗
(G). Hence D
∗∗
is a derivation by
Lemma 3.6.
Lemma 3.8. Let A be a Banach algebra and D : A −→ A
(2n+1)
(n ∈
) a
derivation. Then D
∗∗
: (A
∗∗
, ) −→ ((A
(2n)
)
∗∗
, )
∗
is valid in
D
∗∗
(F G) = D
∗∗
(F )P
∗∗
2n−2
. . . P
∗∗
0
(G) + P
∗∗
2n
. . . P
∗∗
0
(F )D
∗∗
(G) (F, G ∈ A
∗∗
).
is straightforward.
Proposition 3.9. Let A be a Banach algebra and let D : A −→ A
(2n+1)
(n ∈ ) be a derivation. If
D
∗∗
(A
∗∗
) · A
(2n+2)
∪ A
(2n+2)
· D
∗∗
(A
∗∗
) ⊆ P
2n+1
. . . P
1
(A
∗
),
then D
∗∗
: (A
∗∗
, ) −→ (A
∗∗
)
(2n+1)
is a derivation.
. By Lemma 3.8, it is clear.
872
Lemma 3.10. Let A be a Banach algebra and D : A −→ A
∗
a derivation.
Then for every ϕ and ψ in A
(4)
(i) D
(4)
: (A
(4)
, ) −→ (A
(4)
, )
∗
holds in D
(4)
(ϕψ) = D
(4)
(ϕ)ψ+P
(4)
0
(ϕ)
D
(4)
(ψ);
(ii) D
(4)
: (A
(4)
, 44) −→ (A
(4)
, 44)
∗
holds in D
(4)
(ϕ44ψ) = D
(4)
(ϕ)P
(4)
0
(ψ)+
ϕD
(4)
(ψ).
. (i) Let ξ, ϕ, ψ ∈ A
(4)
and let (F
α
), (G
β
) be nets in A
∗∗
such that
P
2
(F
α
) −→ ϕ and P
2
(G
β
) −→ ψ in the weak
∗
topology. By Lemma 3.1 we have
hξ, D
(4)
(ϕ ψ)i = lim
α
lim
β
hD
∗∗
(F
α
G
β
), ξi
= lim
α
lim
β
hD
∗∗
(F
α
)G
β
+ P
∗∗
0
(F
α
)D
∗∗
(G
β
), ξi
= lim
α
hξD
∗∗
(F
α
) + D
(3)
(ξ P
∗∗
0
(F
α
)), ψi
= hD
(3)
(ψ ξ) + P
(3)
0
(D
(4)
(ψ)ξ), ϕi
= hξ, D
(4)
(ϕ)ψ + P
(4)
0
(ϕ)D
(4)
(ψ)i.
(ii) The proof is similar to (i).
Proposition 3.11. Let A be a Banach algebra and D : A −→ A
∗
a derivation.
(i) If D
(4)
(A
(4)
) · A
(4)
⊆ P
3
P
1
(A
∗
), then D
(4)
: (A
(4)
, ) −→ (A
(4)
)
∗
is a
derivation.
(ii) If A
(4)
· D
(4)
(A
(4)
) ⊆ P
3
P
1
(A
∗
), then D
(4)
: (A
(4)
, 44) −→ (A
(4)
)
∗
is a
derivation.
. (i) Let ξ, ϕ, ψ ∈ A
(4)
, there exists a
∗
∈ A
∗
such that D
(4)
(ϕ) · ξ =
P
3
P
1
(a
∗
). By Lemma 1.1 (iii) we have
hξ, P
(4)
0
(ϕ)D
(4)
(ψ)i = hP
(3)
0
P
3
P
1
(a
∗
), ϕi = hϕ, D
(4)
(ψ)ξi = hξ, ϕD
(4)
(ψ)i.
Therefore D
(4)
is a derivation by Lemma 3.10.
(ii) The proof is similar to (i).
We recall that an operator T : X −→ Y between Banach spaces is weakly compact
if and only if T
∗∗
X
∗∗
⊂ Y (considered as a subspace of Y
∗∗
) if and only if T
∗
is
weakly compact.
873
Lemma 3.12. Let A be a Banach algebra and D : A −→ A
∗
a weakly compact
operator. Then D
(2n)
(A
(2n)
) ⊆ P
2n−1
. . . P
3
P
1
(A
∗
) (n ∈ ).
. When n = 1, clearly the result is true. Now suppose, inductively, that
the result has been proved for n. Let ϕ, ξ ∈ A
(2n+2)
and let (ϕ
α
) be a net in A
(2n)
such that P
2n
(ϕ
α
) −→ ϕ in the weak
∗
topology. Then
hξ, D
(2n+2)
(ϕ)i = lim
α
hD
(2n)
(ϕ
α
), ξi = lim
α
hP
∗
2n−1
(ξ), D
(2n)
(ϕ
α
)i
= hD
(2n−1)
P
∗
2n−1
(ξ), P
∗
2n−1
(ϕ)i = hP
∗
2n−1
(ξ), D
(2n)
P
∗
2n−1
(ϕ)i
= hD
(2n)
P
∗
2n−1
(ϕ), ξi = hξ, P
2n+1
D
(2n)
P
∗
2n−1
(ϕ)i.
Consequently, D
(2n+2)
(ϕ) = P
2n+1
D
(2n)
P
∗
2n−1
(ϕ) ⊆ P
2n+1
. . . P
3
P
1
(A
∗
).
Dales, Rodrigues-Palacios and Velasco in [3] proved the following theorem.
Theorem 3.13. Let A be an Arens regular Banach algebra and D : A −→ A
∗
a weakly compact derivation. Then D
(∗∗)
: A
(∗∗)
−→ (A
∗∗
)
∗
is a derivation.
Now we have the same result for A
(4)
.
Theorem 3.14. Let A be an Arens regular Banach algebra and D : A −→
A
∗
a weakly compact derivation. Then D
(4)
: (A
(4)
, ) −→ (A
(4)
)
∗
and D
(4)
:
(A
(4)
, 44) −→ (A
(4)
)
∗
are derivations.
. Let ξ, ϕ, ψ ∈ A
(4)
and (F
α
), (G
β
), (H
γ
) be nets in A
∗∗
such that
P
2
(F
α
) −→ ϕ, P
2
(G
β
) −→ ψ and P
2
(H
γ
) −→ ξ in the weak
∗
topology, let a
∗
∈ A
∗
and let a
∗
α
be a net in A
∗
such that P
1
(a
∗
α
) = D
∗∗
(F
α
) and P
1
(a
∗
) = D
∗∗
P
∗
1
(ϕ). We
have
hξ, D
(4)
(ϕ)ψi = hψ ξ, D
(4)
(ϕ)i = lim
α
lim
β
lim
γ
ha
∗
α
, G
β
H
γ
i
= lim
α
lim
β
ha
∗
α
G
β
, P
∗
1
(ξ)i = lim
α
hP
∗
1
(ψ) P
∗
1
(ξ), D
∗∗
P
∗∗
1
(ϕ)i
= hξ, P
3
(P
1
(a
∗
)P
∗
1
(ψ))i = hξ, P
3
P
1
(a
∗
P
∗
1
(ψ))i,
and so D
(4)
(A
(4)
)A
(4)
⊆ P
3
P
1
(A
∗
) and by Proposition 3.11, D
(4)
: (A
(4)
, ) −→
(A
(4)
)
∗
is a derivation. The other part is similar.
874
Corollary 3.15. Let A be an Arens regular Banach algebra such that
(A
(4)
, ) or (A
(4)
, 44) is weakly amenable and each derivation from A to A
∗
is
weakly compact. Then A is weakly amenable.
. Let D : A −→ A
∗
be a derivation. We may suppose that (A
(4)
, ) is
weakly amenable. By Theorem 3.14, D
(4)
: (A
(4)
, ) −→ (A
(4)
)
∗
is a derivation.
So there exists ϕ
5
∈ (A
(4)
)
∗
such that D
(4)
= δ
ϕ
5
. Set a
∗
= P
∗
0
P
∗
2
(ϕ
5
). Then by
Lemma 1.2 we have
aa
∗
− a
∗
a = P
∗
0
P
∗
2
(P
2
P
0
(a)ϕ
5
− ϕ
5
P
2
P
0
(a))
= P
∗
0
P
∗
2
D
(4)
P
2
P
0
(a) = D(a) (a ∈ A ).
Therefore D = δ
a
∗
is inner. Hence A is weakly amenable.
Proposition 3.16. Let A be a Banach algebra, D : A −→ A
∗
a derivation
and A
(2n)
= ((. . . ((A
∗∗
, )
∗∗
, ) . . .)
∗∗
, ) (n ∈
). Then
(i) D
(2n)
: A
(2n)
−→ (A
(2n)
)
∗
holds in
D
(2n)
(ϕ ψ) = D
(2n)
(ϕ)ψ + P
(2n)
0
(ϕ)D
(2n)
(ψ) (ϕ, ψ ∈ A
(2n)
).
(ii) If D
(2n)
(A
(2n)
) · A
(2n)
⊆ P
2n−1
. . . P
3
P
1
(A
∗
), then D
(2n)
is a derivation.
(iii) If A
(2n−2)
is Arens regular and D is weakly compact, then D
(2n)
is a derivation.
Corollary 3.17. Let A be a completely regular Banach algebra such that
A
(2n)
is weakly amenable for some n ∈ , and each derivation from A to A
∗
is
weakly compact. Then A is weakly amenable.
Lemma 3.18. Let A be an Arens regular Banach algebra such that (A
(4)
, )
or (A
(4)
, 44) is (−2)-weakly amenable. Then A is 2-weakly amenable.
. Let D : A −→ A
∗∗
be a derivation, and let (A
(4)
, ) be (−2)-weakly
amenable. Set d = P
∗
1
D
∗∗
P
∗
1
: (A
(4)
, ) −→ A
∗∗
. For a
∗
∈ A
∗
, ϕ, ψ ∈ A
(4)
let
(F
α
), (G
β
) be nets in A
∗∗
such that P
2
(F
α
) −→ ϕ and P
2
(G
β
) −→ ψ in the weak
∗
topology. Then
ha
∗
, d(ϕ ψ)i = hP
1
D
∗
P
1
(a
∗
), ϕ ψi
= lim
α
lim
β
hP
1
(a
∗
), D
∗∗
(F
α
G
β
)i
= lim
α
lim
β
ha
∗
, P
∗
1
D
∗∗
(F
α
)G
β
+ F
α
P
∗
1
D
∗∗
(G
β
)i
= lim
α
ha
∗
P
∗
1
D
∗∗
(F
α
), P
∗
1
(ψ)i + hD
∗
P
1
(a
∗
F
α
), P
∗
1
(ψ)i
= hP
∗
1
(ψ)a
∗
, d(ϕ)i + hd(ψ)a
∗
, P
∗
1
(ϕ)i
= ha
∗
, d(ϕ) · ψ + ϕ · d(ψ)i.
875
Therefore d is a derivation. Since H
1
(A
(4)
, A
∗∗
) = {0}, there exists F ∈ A
∗∗
such
that d = δ
F
. It is easy to see that D = δ
F
. So A is 2-weakly amenable.
Proposition 3.19. Let A be an Arens regular Banach algebra such that
A
(2n+2)
(n ∈ ) with one of Arens products is (−2n)-weakly amenable. Then
A is 2-weakly amenable.
. Let D : A −→ A
∗∗
be a derivation. By Lemma 3.18 and by in-
duction, d = P
∗
1
D
∗∗
P
∗
1
P
∗
3
. . . P
∗
2n−1
: A
(2n+2)
−→ A
∗∗
is a derivation. Since
H
1
(A
(2n+2)
, A
∗∗
) = {0}, there exists F ∈ A
∗∗
such that d = δ
F
. It is easy to see
that D = δ
F
is inner.
Acknowledgement. We would like to thank the referee for carefully reading the
paper and giving some interesting and fruitful suggestions.
References
[1] W. G. Bade, P. G. Curtis and H. G. Dales: Amenability and weak amenability for Beurl-
ing and Lipschitz algebra. Proc. London Math. Soc. 55 (1987), 359–377.
[2] H. G. Dales, F. Ghahramanim and N. Gronbaek: Derivations into iterated duals of Ba-
nach algebras. Studia Math. 128 (1998), 19–54.
[3] H. G. Dales, A. Rodriguez-Palacios and M. V. Valasco: The second transpose of a deriva-
tion. J. London Math. Soc. 64 (2001), 707–721.
[4] M. Despic and F. Ghahramani: Weak amenability of group algebras of locally compact
groups. Canad. Math. Bull. 37 (1994), 165–167.
[5] J. Duncan and Hosseiniun: The second dual of a Banach algebra. Proc. Roy. Soc.
Edinburgh 84A (1978), 309–325.
[6] N. Gronbaek: Weak amenability of group algebras. Bull. London Math. Soc. 23 (1991),
231–284.
[7] U. Haagerup: All nuclear C
∗
-algebras are amenable. Invent. Math. 74 (1983), 305–319.
[8] B. E. Johnson: Cohomology in Banach Algebras. Mem. Amer. Math. Soc. 127 (1972).
[9] B. E. Johnson: Weak amenability of group algebras. Bull. Lodon Math. Soc. 23 (1991),
281–284.
[10] T. W. Palmer: Banach Algebra, the General Theory of ∗-algebra. Vol. 1: Algebra and
Banach Algebras. Cambridge University Press, Cambridge, 1994.
Authors’ address:
, Faculty of Mathematical Science, Teacher Train-
ing University, 599, Taleghani Avenue, Tehran, 15614, Iran, e-mail: medghalchi@saba.tmu.
ac.ir; , Department of Mathematics, Persian Gulf University, 75168
Boushehr, Iran, e-mail: yazdanpanah@pgu.ac.ir.
876