Can B(l^p) ever be amenable?

Abstract: It is known that

c a l B (e l l^{p})

is not amenable for

p = 1, 2, i n f t y

, but whether or not

c a l B (e l l^{p})

is amenable for

p i n (1, i n f t y) s e t m i n u s 2

is an open problem. We show that, if

c a l B (e l l^{p})

is amenable for

p i n (1, i n f t y)

, then so are

e l l^{i} n f t y (c a l B (e l l^{p}))

and

e l l^{i} n f t y (c a l K (e l l^{p}))

. Moreover, if

e l l^{i} n f t y (c a l K (e l l^{p}))

is amenable so is

e l l^{i} n f t y (m a t h b b I, c a l K (E))

for any index set

m a t h b b I

and for any infinite-dimensional

{c a l L}^{p}

-space

E

; in particular, if

e l l^{i} n f t y (c a l K (e l l^{p}))

is amenable for

p i n (1, i n f t y)

, then so is

e l l^{i} n f t y (c a l K (e l l^{p} o p l u s e l l^{2}))

. We show that

e l l^{i} n f t y (c a l K (e l l^{p} o p l u s e l l^{2}))

is not amenable for

p = 1, i n f t y

, but also that our methods fail us if

p i n (1, i n f t y)

. Finally, for

p i n (1, 2)

and a free ultrafilter

c a l U

over

p o s i n t s

, we exhibit a closed left ideal of

(c a l K (e l l^{p}))_{c a l U}

lacking a right approximate identity, but enjoying a certain, very weak complementation property.

Recommendations

Cited in

(7)

This page was built for publication: Can B(l^p) ever be amenable?

Can B(lp) ever be amenable?