$L_p+L_\infty$ and $L_p\cap L_\infty$ are not isomorphic for all $1\le p<\infty,$ $p\ne 2$
S.V. Astashkin, L. Maligranda

TL;DR
This paper proves that for all p between 1 and infinity, except 2, the nonseparable spaces L_p+L_infinity and L_p∩L_infinity are not isomorphic, highlighting unique structural properties of these Banach spaces.
Contribution
It demonstrates the non-isomorphism of L_p+L_infinity and L_p∩L_infinity for p ≠ 2, and analyzes the presence of complemented subspaces within these spaces.
Findings
L_p∩L_infinity does not contain a complemented subspace isomorphic to L_p for p ≠ 2
L_p∩L_infinity contains a complemented subspace isomorphic to l_2 if and only if p=2
The isomorphism problem for L_2+L_infinity and L_2∩L_infinity remains open
Abstract
Isomorphic classification of symmetric spaces is an important problem related to the study of symmetric structures in arbitrary Banach spaces. This research was initiated in the seminal work of Johnson, Maurey, Schechtman and Tzafriri (JMST, 1979). Somewhat later it was extended by Kalton to lattice structures (1993). In particular, in JMST (see also Lindenstrauss-Tzafriri book [1979, Section 2.f]) it was shown that the space for (resp. for ) is isomorphic to . A detailed investigation of various properties of separable sums and intersections of -spaces (i.e., with ) was undertaken by Dilworth in the papers from 1988 and 1990. In contrast to that, we focus here on the problem if the nonseparable spaces and , , are isomorphic or not. We prove that these…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topicsadvanced mathematical theories
** and are not isomorphic
for all , **
Sergei V. Astashkin and Lech Maligranda Research partially supported by the Ministry of Education and Science of the Russian Federation.
Abstract
We prove the result stated in the title. It comes as a consequence of the fact that the space , , , does not contain a complemented subspace isomorphic to . In particular, as a subproduct, we show that contains a complemented subspace isomorphic to if and only if .
00footnotetext: 2010 Mathematics Subject Classification: 46E30, 46B20, 46B4200footnotetext: Key words and phrases: symmetric spaces, isomorphic spaces, complemented subspaces
1 Preliminaries and main result
Isomorphic classification of symmetric spaces is an important problem related to the study of symmetric structures in arbitrary Banach spaces. This research was initiated in the seminal work of Johnson, Maurey, Schechtman and Tzafriri [9]. Somewhat later it was extended by Kalton to lattice structures [10].
In particular, in [9] (see also [12, Section 2.f]) it was shown that the space for (resp. for ) is isomorphic to . A detailed investigation of various properties of separable sums and intersections of -spaces (i.e., with ) was undertaken by Dilworth in the papers [5] and [6]. In contrast to that, we focus here on the problem if the nonseparable spaces and , , are isomorphic or not.
In this paper we use the standard notation from the theory of symmetric spaces (cf. [3], [11] and [12]). For the space consists of all sums of -integrable and bounded measurable functions on with the norm defined by
[TABLE]
The consists of all bounded -integrable functions on with the norm
[TABLE]
Both and for all are non-separable Banach spaces (cf. [11, p. 79] for ). The norm in satisfies the following sharp estimates
[TABLE]
(cf. [4, p. 109], [13, p. 176] and with details in [14, Theorem 1]) – see also [3, pp. 74-75] and [11, p. 78], where we can find a proof of (1) in the case when , that is,
[TABLE]
Here, denotes the decreasing rearrangement of , that is,
[TABLE]
(if is a measurable set, then is its Lebesgue measure). Note that every measurable function and its decreasing rearrangement are equimeasurable, that is,
[TABLE]
for all .
Denote by and , the closure of in and in , respectively. Clearly, . Note that
[TABLE]
and
[TABLE]
i.e., , is a dual space (cf. [11, pp. 79-80] and [3, pp. 76-77]). Also, and , are dual spaces because
[TABLE]
where .
Now, we state the main result of this paper.
THEOREM 1**.**
For every , the spaces and are not isomorphic.
Clearly, the space contains the complemented subspace (L_{p}+L_{\infty})_{{\big{|}}[0,1]} isomorphic to for every . As a bounded projection we can take the operator because
[TABLE]
In the next two sections we show that for does not contain a complemented subspace isomorphic to , which gives our claim. At the same time, note that , , contains a subspace isomorphic to and hence a subspace isomorphic to .
The spaces and are not isomorphic since contains a complemented subspace isomorphic to and is a prime space (this follows from the Lindenstrauss and Pełczyński results – see [1, Theorems 5.6.5 and 4.3.10]). Similarly, the spaces and are not isomorphic because of contains a complemented subspace isomorphic to (take, for instance, the span of the sequence in ).
If is a sequence from a Banach space , by we denote its closed linear span in . As usual, the Rademacher functions on are defined as follows: .
2 does not contain a complemented subspace isomorphic to
Our proof of Theorem 1 in the case will be based on an application of the Hagler-Stegall theorem proved in [8] (see Theorem 1). To state it we need the following definition.
The space , , is the Banach space of all sequences , , such that
[TABLE]
THEOREM 2** (Hagler-Stegall).**
Let be a Banach space. Then its dual contains a complemented subspace isomorphic to if and only if contains a subspace isomorphic to .
Note that (L_{1}+L_{\infty}^{0})_{{\big{|}}[0,1]}=L_{1}[0,1], and hence contains a complemented copy of , and so of . Moreover, its subspace
[TABLE]
is isomorphic to and so, by Sobczyk theorem (cf. [1, Theorem 2.5.8]), is complemented in the separable space . Therefore, the latter space contains uniformly complemented copies of . However, we have
THEOREM 3**.**
The space does not contain any subspace isomorphic to the space .
Proof.
On the contrary, assume that contains a subspace isomorphic to . Let , form the sequence from equivalent to the unit vector basis of . This means that there is a constant such that for all
[TABLE]
In particular, for any , every subset and all , we have
[TABLE]
and for all the sequence is equivalent in to the unit vector basis of , i.e., for all
[TABLE]
Moreover, we can assume that for all , i.e.,
[TABLE]
Firstly, we show that for every there is such that for all and any with we have
[TABLE]
Indeed, assuming the contrary, for some we can find , such that
[TABLE]
Denoting , for all we have
[TABLE]
Moreover, by the Fubini theorem, Khintchine’s inequality in (cf. [12, pp. 50-51] or [17]) and Minkowski inequality, we obtain
[TABLE]
Therefore, for each there are signs such that
[TABLE]
Combining this with (8) we obtain that
[TABLE]
Since as , the latter inequality contradicts (4). Thus, (7) is proved.
Now, we claim that for all and
[TABLE]
where depending on is taken from (7).
Indeed, otherwise, we can find and , card such that for every there is with
[TABLE]
Setting , we see that . Moreover, by the definition of and ,
[TABLE]
which is impossible because of (7).
Now, we construct a special sequence of pairwise disjoint functions, which is equivalent in to the unit vector basis in . By (7), for arbitrary there is such that for all
[TABLE]
Therefore, taking , we can find satisfying
[TABLE]
and, by (9), such that from with it follows that
[TABLE]
Moreover, recalling (2) we have as . Therefore, since and any measurable function is equimeasurable with its decreasing rearrangement, there exists such that . Then, setting , we have
[TABLE]
Next, by (7), for arbitrary there is such that for all and
[TABLE]
Let be such that . Then, by the preceding inequality and (9), there is such that for all we have
[TABLE]
and from with , it follows that
[TABLE]
Note that (10) implies , whence
[TABLE]
As above, by (2), there is such that . Thus, putting , we have
[TABLE]
Continuing this process, for any , by (7), we can find such that for all and it holds
[TABLE]
So, again, applying (9) and taking we find such that
[TABLE]
and
[TABLE]
whenever . This implies that , and so
[TABLE]
Choosing so that and setting , we obtain
[TABLE]
As a result, we get the increasing sequences of natural numbers, and the sequence of pairwise disjoint functions from such that
[TABLE]
where . Noting that the sequence of positive reals can be chosen in such a way that the numbers would be arbitrarily small, we can assume, by the principle of small perturbations (cf. [1, Theorem 1.3.10]) and by inequalities (5), that is equivalent in to the unit vector basis of . Moreover, by construction, for all and we have
[TABLE]
Let be arbitrary. Since are disjoint functions, then
[TABLE]
where are disjoint sets from such that . Clearly, for a fixed we have
[TABLE]
[TABLE]
So, assuming that , we obtain
[TABLE]
Since as , the latter inequality contradicts the fact that is equivalent in to the unit vector basis of . The proof is complete. ∎
Remark 1**.**
Since the space , , is of Rademacher cotype , the result of Theorem 3 can be generalized as follows: For every the space does not contain any subspace isomorphic to the space .
Proof of Theorem 1 for .
By the Hagler-Stegall theorem 2, Theorem 3 and the fact that , we obtain that (in contrast to ) the space does not contain a complemented subspace isomorphic to , which gives our claim. ∎
There is a natural question (cf. also [2, p. 28]) if the space is isomorphic to a dual space? Our guess is that not, but we don’t have a proof. Of course, the answer “not” would imply immediately the result of Theorem 1 for .
Problem 1**.**
Is the space isomorphic to a dual space?
3 for does not contain a complemented subspace isomorphic to
The well-known Raynaud’s result (cf. [15, Theorem 4]) presents the conditions under which a separable symmetric space (on or on ) contains a complemented subspace isomorphic to . The following theorem can be regarded as its extension to a special class of nonseparable spaces.
THEOREM 4**.**
Let . Then the space contains a complemented subspace isomorphic to the space if and only if .
Proof.
If , then clearly the sequence is equivalent in to the unit vector basis of and spans a complemented subspace.
Let us prove necessity. On the contrary, let be a sequence equivalent in to the unit vector basis of so that is a complemented subspace of .
Firstly, let us show that there is not such that for all
[TABLE]
Indeed, the latter equivalence implies
[TABLE]
Since , we see that the sequence spans in both spaces and the same infinite-dimensional space. However, by the well-known Grothendieck’s theorem (cf. [7, Theorem 1]; see also [16, p. 117]) it is impossible. As a result, we can find a sequence , such that for every
[TABLE]
Hence, (convergence in Lebesgue measure ) on any interval . Since spans , then passing to a subsequence if it is necessary (and keeping the same notation), we can assume that weakly in . Therefore, combining the Bessaga-Pełczyński Selection Principle (cf. [1, Theorem 1.3.10]) and the principle of small perturbations (cf. [1, Theorem 1.3.10]), we can select a further subsequence, which is equivalent to the sequence in (and so to the unit vector basis in ) and which spans a complemented subspace in . Let it be denoted still by . Now, we will select a special subsequence from , which is equivalent to a sequence of functions whose supports intersect only over some subset of with Lebesgue measure at most .
Let be an arbitrary (by now) decreasing sequence of positive reals, . Since on , there is such that
[TABLE]
Moreover, the fact that as allows us to find , for which
[TABLE]
Clearly, from (14) it follows that
[TABLE]
Denoting
[TABLE]
and
[TABLE]
from (13), (14) and (15) we have
[TABLE]
Further, since on , there exists such that
[TABLE]
Again, using the fact that as , we can choose in such a way that
[TABLE]
and also
[TABLE]
From (17), obviously, it follows that
[TABLE]
Setting
[TABLE]
and
[TABLE]
by (16), (17) and (19), we get
[TABLE]
Let’s do one more step. Since on , there exists such that
[TABLE]
As above, we can choose with the properties
[TABLE]
[TABLE]
and
[TABLE]
From (21) we infer that
[TABLE]
Finally, putting
[TABLE]
and
[TABLE]
by (20), (21) and (24), we have
[TABLE]
Continuing in the same way, we get the increasing sequences of natural numbers , the sequences of sets and the sequence of functions
[TABLE]
(where ), satisfying the properties
[TABLE]
[TABLE]
[TABLE]
In particular, by the last inequality, choosing sufficiently small , , and applying once more the principle of small perturbations [1, Theorem 1.3.10], we may assume that the sequence is equivalent to (and so to the unit vector basis of ) and the subspace is complemented in . Thus, for some and all ,
[TABLE]
Now, denote
[TABLE]
[TABLE]
Setting and applying (25) and (26), we have
[TABLE]
whenever are sufficiently small. Putting
[TABLE]
and recalling the definition of , we infer that
[TABLE]
Note that , whence (27) can be rewritten as follows
[TABLE]
Moreover, the subspace is also complemented in and, by (28), we have
[TABLE]
Now, suppose that . Then passing to a subsequence (and keeping the same notation), by (29), we obtain
[TABLE]
Since , are pairwise disjoint, we have
[TABLE]
Firstly, let us assume that . If , then selecting a further subsequence (and again keeping notation), we obtain the inequality
[TABLE]
which contradicts the right-hand estimate in (31). So, , and then from (32) for some subsequence of (we still keep notation) we have
[TABLE]
and now the left-hand side of (31) fails. Thus, if , inequality (31) does not hold.
Let . Clearly, from (32) it follows that
[TABLE]
and so the left-hand side estimate in (31) cannot be true. Thus, (31) fails for every , and as a result we get
[TABLE]
Now, if , then, as above, , and we come (for some subsequence of ) to inequality (33). Clearly,
[TABLE]
and from (33) and (29) it follows that for some and all we have
[TABLE]
Therefore, the subspace is isomorphic in to .
We show that the last claim holds also in the case . On the contrary, assume that the left-hand side of (36) fails (note that the opposite side of (36) follows from (29)). In other words, assume that there is a sequence , such that for all and
[TABLE]
Then, by (35), we have as . Therefore, since
[TABLE]
we have as . Combining this together with (34), we obtain
[TABLE]
and so the left-hand estimate in (29) does not hold. This contradiction shows that (36) is valid for every . Thus, the subspace is complemented in and isomorphic to . As an immediate consequence of that, we infer that is a complemented subspace of the space , where . Since by (30) , it follows that is isometric to . As a result we come to a contradiction, because does not contain any complemented reflexive subspace (cf. [1, Theorem 5.6.5]). ∎
Proof of Theorem 1 for .
Clearly, if then (and hence ) contains a complemented copy of (for instance, the span of the Rademacher sequence). Therefore, by applying Theorem 4, we complete the proof. ∎
Note that if is a symmetric space on , then contains a complemented space isomorphic to since
[TABLE]
where . In fact, for , using estimate (4.2) from [11, p. 91], we obtain
[TABLE]
So, an inspection of the proofs of Theorems 4 and 1 (in the case when ) shows that the following more general result is true.
THEOREM 5**.**
Suppose is a separable symmetric space on satisfying either the upper -estimate for or lower -estimate for . Then the space does not contain any complemented subspace isomorphic to .
If, in addition, the space contains a complemented subspace isomorphic to , then the spaces and are not isomorphic.
4 Concluding remarks related to the spaces and
We do not know whether the spaces and are isomorphic or not.
Problem 2**.**
Are the spaces and isomorphic?
We end up the paper with the following remarks related to the above problem.
Remark 2**.**
The predual spaces and for and , respectively, are not isomorphic.
In fact, is a separable dual space since (cf. [5, Proposition 2(a)]). Therefore, the space cannot be embedded in this space (cf. [1, p. 147]) but has a complemented subspace isomorphic to , which completes our observation.
Remark 3**.**
Either of the spaces and is isomorphic to a (uncomplemented) subspace of , and hence is isomorphic to a subspace of and vice versa.
To see this, for instance, for , it is sufficient to take arbitrary dense sequence of the unit ball of the space , say, , and to set
[TABLE]
It is easy to see that this mapping defines an isometrical embedding of into .
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] F. Albiac, N. J. Kalton, Topics in Banach Space Theory , Springer-Verlag, New York 2006.
- 2[2] S. V. Astashkin, K. Leśnik and L. Maligranda, Isomorphic structure of Cesàro and Tandori spaces , to appear. Preprint of 33 pages submitted on 10 December 2015 at ar Xiv:1512.03336
- 3[3] C. Bennett and R. Sharpley, Interpolation of Operators , Academic Press, Boston 1988.
- 4[4] J. Bergh and J. Löfström, Interpolation Spaces. An Introduction , Springer-Verlag, Berlin-New York 1976.
- 5[5] S. J. Dilworth, Intersection of Lebesgue spaces L 1 subscript 𝐿 1 L_{1} and L 2 subscript 𝐿 2 L_{2} , Proc. Amer. Math. Soc. 103 (1988), no. 4, 1185–1188.
- 6[6] S. J. Dilworth, A scale of linear spaces related to the L p subscript 𝐿 𝑝 L_{p} scale , Illinois J. Math. 34 (1990), no. 1, 140–158.
- 7[7] A. Grothendieck, Sur certains sous-espaces vectoriels de L p superscript 𝐿 𝑝 L^{p} , Canadian J. Math. 6, (1954), 158–160.
- 8[8] J. Hagler and Ch. Stegall, Banach spaces whose duals contain complemented subspaces isomorphic to C ( [ 0 , 1 ] ) ∗ 𝐶 superscript 0 1 C([0,1])^{*} , J. Funct. Anal. 13 (1973), 233–251.
