Semicommutative Hardy Spaces

• Semicommutative Hardy spaces are function spaces combining classical geometric structures with operator-valued elements from von Neumann algebras.
• They encompass diverse models, including projection-based, martingale, and operator-valued Hardy–BMO theories, each employing distinct square function and spectral projection techniques.
• Their robust framework ensures stability under interpolation and duality, facilitating applications in noncommutative harmonic analysis and Calderón–Zygmund theory.

Searching arXiv for the cited papers and closely related semicommutative Hardy-space work. Semicommutative Hardy spaces are Hardy-type function spaces in which the underlying geometry remains classical while the function values lie in a noncommutative algebra, typically a von Neumann algebra with trace. The basic ambient object is a semicommutative algebra of the form $L_\infty(X)\,\bar\otimes\,M$ or $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$, where the $L_\infty$-factor is commutative and the $M$- or $\mathcal M$-factor may be genuinely noncommutative. In the recent literature, the term encompasses several distinct but related theories: projection-defined Hardy spaces driven by interpolation and singular integrals, row/column and conditioned martingale Hardy spaces attached to filtrations, and operator-valued Hardy–BMO spaces on spaces of homogeneous type defined through Lusin area integrals (Moyart, 26 Apr 2026, Junge et al., 2013, Fan et al., 2023). This suggests that “semicommutative Hardy space” is best understood as a family of operator-valued Hardy theories rather than a single canonical construction.

1. Ambient framework and basic meaning

In the semicommutative setting of singular-integral and interpolation theory, one starts with a semifinite von Neumann algebra $M$ equipped with a normal semifinite faithful trace $\tau$, a measure space $X$, and the tensor-product algebra

$N:=L_\infty(X)\,\bar\otimes\, M,$

equipped with the tensor-product trace $\sigma$. Elements of $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$0 are essentially bounded $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$1-valued functions on $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$2, and $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$3 is the corresponding noncommutative $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$4-space. The framework is called semicommutative because the $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$5-factor is commutative while the $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$6-factor may be genuinely noncommutative (Moyart, 26 Apr 2026).

In the martingale setting, the same structural idea appears with

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$7

equipped with a filtration $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$8 coming from a classical filtration on $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$9 or from mixed classical/operator-valued filtrations. Here semicommutative Hardy spaces arise as special cases of Hardy spaces over finite von Neumann algebras with trace-preserving conditional expectations (Junge et al., 2013).

On spaces of homogeneous type, the ambient semicommutative algebra is

$L_\infty$0

where $L_\infty$1 is a space of homogeneous type in the sense of Coifman–Weiss and $L_\infty$2 is a von Neumann algebra equipped with a normal semifinite faithful trace $L_\infty$3. In that theory, geometry is classical, algebraic size and duality are noncommutative, and the Hardy and BMO spaces come in column, row, and mixture forms (Fan et al., 2023).

A recurrent clarification is that semicommutativity does not mean one is working with the rotation or noncommutative torus in Rieffel’s sense. For the torus model in the projection-based theory, the algebra is

$L_\infty$4

so the spatial variable is the classical torus and only the values are operator-valued (Moyart, 26 Apr 2026).

2. Principal models of semicommutative Hardy spaces

The recent literature supports at least three principal models.

Framework	Ambient algebra	Defining mechanism
Projection-based Hardy spaces	$L_\infty$5	Weak closure of $L_\infty$6
Martingale Hardy spaces	$L_\infty$7	Row/column/conditioned square functions over filtrations
Operator-valued Hardy spaces on homogeneous type	$L_\infty$8	Lusin area functions from a Calderón reproducing family

In the projection-based theory, if $L_\infty$9 is a self-adjoint projection on $M$0 and $M$1 is an exact interpolation space for $M$2, the Hardy space is defined by

$M$3

as the $M$4-weak closure in $M$5 of

$M$6

For $M$7, one writes

$M$8

and by definition

$M$9

These are abstract Hardy spaces associated with $\mathcal M$0. The construction is explicitly not based on square functions, maximal functions, row/column structures, conditioned martingale norms, or subdiagonal algebras as the primary model (Moyart, 26 Apr 2026).

In the martingale theory for continuous filtrations, the discrete column and row Hardy norms are

$\mathcal M$1

with conditioned norms

$\mathcal M$2

and diagonal norm

$\mathcal M$3

For continuous filtrations, these norms are replaced by partition limits and ultrafilter constructions based on the brackets

$\mathcal M$4

The corresponding spaces are denoted $\mathcal M$5 (Junge et al., 2013).

In the operator-valued Hardy–BMO theory on spaces of homogeneous type, Hardy spaces are defined through Lusin area functions. For an $\mathcal M$6-valued simple function $\mathcal M$7, with cones

$\mathcal M$8

the column and row area functions are

$\mathcal M$9

$M$0

The norms are

$M$1

and the mixture space is defined by

$M$2

At $M$3,

$M$4

(Fan et al., 2023).

3. Spectral and analytic Hardy spaces on the torus

The torus provides the main concrete model for projection-defined semicommutative Hardy spaces. Let

$M$5

With Fourier basis $M$6, where

$M$7

the $M$8-valued Fourier coefficients are defined by

$M$9

Given $\tau$0, the spectral projection is

$\tau$1

and the associated Hardy-type subspace is

$\tau$2

For exact interpolation spaces $\tau$3 with order-continuous norm, one has

$\tau$4

and $\tau$5 is dense in the appropriate norm or $\tau$6-topology. Thus the abstract Hardy space associated with a spectral projection is exactly the Fourier-support-defined subspace one expects (Moyart, 26 Apr 2026).

For $\tau$7, the analytic Hardy space on the torus is defined as the subspace $\tau$8 consisting of all $\tau$9 such that

$X$0

Equivalently, Fourier coefficients vanish on the “negative” spectrum indexed by $X$1 in the convention of the paper. In dimension $X$2, this is the standard operator-valued Hardy space of analytic $X$3-valued functions on the circle (Moyart, 26 Apr 2026).

For $X$4 and $X$5, the spectral projection $X$6 is the Riesz projection, whose kernel is

$X$7

or equivalently

$X$8

This kernel is singular on $X$9, so the general semicommutative singular-integral theory applies to analytic operator-valued Hardy spaces on the circle (Moyart, 26 Apr 2026).

A distinct but related commutative model appears in the theory of symmetric norms. For a continuous rotationally symmetric norm $N:=L_\infty(X)\,\bar\otimes\, M,$0, one defines

$N:=L_\infty(X)\,\bar\otimes\, M,$1

and the Fourier characterization is

$N:=L_\infty(X)\,\bar\otimes\, M,$2

Also,

$N:=L_\infty(X)\,\bar\otimes\, M,$3

The paper does not define a noncommutative Hardy space $N:=L_\infty(X)\,\bar\otimes\, M,$4 for a subdiagonal algebra, but it presents this theory as a blueprint for such an extension (Chen, 2014).

4. Duality, interpolation, and $N:=L_\infty(X)\,\bar\otimes\, M,$5-closedness

A major theme of semicommutative Hardy-space theory is that Hardy subspaces behave stably under interpolation and admit BMO-type duals.

In the projection-based theory, the organizing tool is the Peetre $N:=L_\infty(X)\,\bar\otimes\, M,$6-functional

$N:=L_\infty(X)\,\bar\otimes\, M,$7

together with the notions of $N:=L_\infty(X)\,\bar\otimes\, M,$8-closedness and quasi-complementation. If $N:=L_\infty(X)\,\bar\otimes\, M,$9 is a Jones-type projection on $\sigma$0, then for every $\sigma$1,

$\sigma$2

is quasi-complemented in

$\sigma$3

with constant depending only on $\sigma$4. Consequently the Hardy couple is $\sigma$5-closed in the ambient $\sigma$6-couple. If $\sigma$7 is a $\sigma$8-parameter such that

$\sigma$9

has order-continuous norm, then

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$00

with equivalent norms; in particular,

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$01

For noncommutative $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$02-spaces, Holmstedt’s formula

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$03

underlies the real interpolation results (Moyart, 26 Apr 2026).

For the torus Hardy spaces in dimension $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$04, the recovered Pisier–Xu result states that

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$05

is quasi-complemented in

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$06

with a universal constant, and for every $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$07,

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$08

with equivalent norms, where $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$09, and the constants depend only on $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$10. The same framework yields complex interpolation: $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$11 (Moyart, 26 Apr 2026).

In the continuous-filtration martingale theory, one has the Hardy/Lebesgue equivalence

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$12

where

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$13

More explicitly,

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$14

and

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$15

The paper also proves

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$16

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$17

and interpolation identities

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$18

(Junge et al., 2013).

In the operator-valued Hardy theory on spaces of homogeneous type, the main structural theorem states for $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$19: $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$20

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$21

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$22

and

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$23

Interpolation takes the form

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$24

as well as the real interpolation identity

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$25

(Fan et al., 2023).

5. Calderón–Zygmund structure and geometric extensions

A decisive recent development is the transplantation of Bourgain’s method for $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$26-closedness to the semicommutative setting. In that framework, a bounded operator $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$27 on $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$28 is of Calderón–Zygmund type if for every $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$29 and $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$30, there exist

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$31

and a projection $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$32 such that

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$33

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$34

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$35

This is the semicommutative substitute for the classical good/bad decomposition plus localization by a “good” projection. Such operators are weak type $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$36, and if a projection $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$37 satisfies the technical assumptions and is of Calderón–Zygmund type, then

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$38

is quasi-complemented in

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$39

This is the semicommutative analogue of Bourgain’s key $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$40-closedness step (Moyart, 26 Apr 2026).

The technical engine is a semicommutative Calderón–Zygmund decomposition on

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$41

for $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$42 Ahlfors $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$43-regular and $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$44 semifinite. Starting from Cuculescu projections

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$45

and setting

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$46

one decomposes

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$47

with

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$48

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$49

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$50

The presence of the diagonal part $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$51 and off-diagonal part $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$52 reflects genuinely noncommutative interactions. The associated projection

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$53

encodes the bad region. The key estimates are

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$54

and for singular-kernel operators $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$55,

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$56

Combining these yields

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$57

which is exactly what Bourgain’s method requires (Moyart, 26 Apr 2026).

The same paper shows that if $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$58 is a bounded operator on $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$59 given by a scalar singular kernel $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$60 on an Ahlfors-regular space $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$61, satisfying the size estimate

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$62

and an $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$63-Hörmander regularity condition, then $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$64 is of Calderón–Zygmund type with constants independent of $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$65. Consequently, a projection $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$66 on $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$67 that satisfies the technical assumptions and is given by a singular kernel on $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$68 is of Jones-type. Hence the associated semicommutative Hardy spaces are $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$69-closed and form a real interpolation scale (Moyart, 26 Apr 2026).

A parallel geometric extension is provided by the operator-valued Hardy–BMO theory on spaces of homogeneous type. There the standing assumption is a Calderón reproducing formula

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$70

with kernels $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$71 satisfying size, Hölder regularity in both variables, and cancellation: $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$72 A key point is that the theory works under doubling only and does not assume reverse doubling. Under these assumptions it establishes $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$73–BMO duality, atomic decomposition of $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$74, interpolation between Hardy spaces and BMO spaces, and $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$75-boundedness of operator-valued Calderón–Zygmund operators (Fan et al., 2023).

6. Martingale, atomic, and generalized-norm perspectives

The continuous-filtration theory supplies the semicommutative martingale analogue of classical Hardy/BMO structure without pathwise stopping-time methods. The continuous bracket is defined through an ultrafilter: $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$76 and

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$77

A second candidate norm is

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$78

and the paper proves

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$79

with equivalent norms. The same holds for conditioned Hardy spaces: $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$80 A structural fact important in semicommutative applications is that the norms are intrinsic up to equivalence, even though the bracket elements themselves are introduced through ultrafilters (Junge et al., 2013).

The same martingale theory establishes continuous Davis decompositions. For $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$81,

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$82

with equivalent norms, and for $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$83,

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$84

It also proves the conditioned/global Burkholder–Rosenthal theorem

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$85

where

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$86

When specialized to semicommutative algebras $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$87, these become continuous operator-valued semicommutative Hardy, BMO, conditioned, and diagonal decompositions (Junge et al., 2013).

The operator-valued Hardy theory on spaces of homogeneous type adds an atomic endpoint description. An $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$88-atom $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$89 is supported in a ball $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$90, satisfies

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$91

and has cancellation

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$92

The atomic Hardy space $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$93 consists of all

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$94

with $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$95, and the theorem

$L_\infty(\Omega)\,\bar\otimes\,\mathcal M$96

holds for $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$97. This gives a semicommutative $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$98 atomic theory in rough geometric settings (Fan et al., 2023).

Finally, the symmetric-norm framework indicates one route beyond $L_\infty(\Omega)\,\bar\otimes\,\mathcal M$99-based Hardy spaces. For $L_\infty$00, the spaces $L_\infty$01 and $L_\infty$02 generalize $L_\infty$03 and $L_\infty$04. In the von Neumann algebra setting, one defines $L_\infty$05 by completing $L_\infty$06 in the induced symmetric gauge norm. The corrected reflexivity criterion is

$L_\infty$07

Moreover, if $L_\infty$08 is continuous and $L_\infty$09 is strongly continuous, then

$L_\infty$10

The paper does not build semicommutative Hardy spaces in this generalized-norm setting, but it explicitly presents duality and module-map representation theorems as foundational tools for such an extension (Chen, 2014).

7. Scope, distinctions, and recurring misconceptions

A central distinction is that semicommutative Hardy spaces are not exhausted by one definition. The projection-based spaces $L_\infty$11, the martingale spaces $L_\infty$12, and the Lusin-area-function spaces $L_\infty$13 are all semicommutative, but they are built from different primitives: spectral projections and interpolation, filtrations and square functions, or Calderón reproducing families and area integrals (Moyart, 26 Apr 2026, Junge et al., 2013, Fan et al., 2023).

Another recurring clarification concerns the role of row and column structures. In the projection-based Bourgain framework, the Hardy spaces are not introduced via row/column norms, conditioned Hardy spaces, square function characterizations, or maximal function characterizations. By contrast, in the martingale and homogeneous-type theories, row and column spaces are fundamental, and the row/column distinction disappears only when the value algebra is commutative (Moyart, 26 Apr 2026, Junge et al., 2013).

The torus terminology also requires care. The “semicommutative torus” in the interpolation and Bourgain theory means $L_\infty$14, namely the classical torus with values in a von Neumann algebra, not the rotation/noncommutative torus in Rieffel’s sense (Moyart, 26 Apr 2026).

Taken together, the available theories show that semicommutative Hardy spaces form a broad operator-valued Hardy paradigm. In one direction, singular-kernel projections on $L_\infty$15 produce $L_\infty$16-closed Hardy scales and recover analytic torus Hardy spaces via the Riesz projection. In another, continuous filtrations produce row, column, conditioned, and diagonal Hardy/BMO theories for operator-valued martingales. In a third, Calderón reproducing formulas on spaces of homogeneous type yield a full Hardy–BMO theory with atomic decomposition, interpolation, and Calderón–Zygmund applications. A plausible implication is that semicommutativity functions less as a single definition than as a unifying structural principle: classical geometry or probability on the outside, noncommutative $L_\infty$17- and duality structure on the inside.