Minimal index and dimension for inclusions of von Neumann algebras with finite-dimensional centers

Luca Giorgetti

Minimal index and dimension for inclusions of von Neumann algebras with finite-dimensional centers

T0 review · 0 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read For connected inclusions of von Neumann algebras with finite-dimensional centers, the minimal index equals the squared $\ell^2$-norm of a finite matrix $D$ of subfactor index square roots.

desk verdict A transparent conference announcement of the Giorgetti–Longo results; no new proofs, but a clean formula and interesting open problems. read the letter →

arxiv 1908.09121 v1 pith:RFUR5BPO submitted 2019-08-24 math.OA math-phmath.CTmath.MP

Luca Giorgetti This is my paper · ORCID

classification math.OAmath-phmath.CTmath.MP MSC 46L1046L37

keywords minimalindexvonNeumannalgebrasconditionalexpectationsmatrixdimensionfinite-dimensionalcentersPerron-Frobeniustheorysubfactors2-C*-categories

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes that the minimal index of a connected inclusion of von Neumann algebras with finite-dimensional centers is not an elusive variational quantity: it is exactly the squared $\ell^2$-norm of a finite matrix $D$, called the matrix dimension, whose entries are square roots of minimal indices of subfactors obtained by cutting with minimal central projections. Because the matrix is finite and irreducible, Perron-Frobenius theory yields a unique minimizing conditional expectation and canonical states on the centers. This matters because it extends index theory beyond factors and traces to algebras with a finite classical part, where the index of a conditional expectation must be minimized over many possible expectations. It also connects the minimal index to quantum information, where the same matrix plays the role of a dimension for a quantum channel.

What carries the argument

The load-bearing object is the matrix dimension $D$, the $m\times n$ matrix with entries $d_{ij}=[M_{ij}:N_{ij}]_0^{1/2}$ for $p_iq_j\neq 0$ and $0$ otherwise. Its role is to turn the global problem of minimizing $\|\mathrm{Ind}(E)\|$ over conditional expectations into a finite spectral problem: connectedness makes $DD^t$ and $D^tD$ irreducible nonnegative matrices, so Perron-Frobenius theory supplies unique strictly positive eigenvectors $\sqrt{\mu}$ and $\sqrt{\nu}$ for the eigenvalue $d^2=\|D\|_{\ell^2}^2$. These eigenvectors determine the weights of the unique minimal expectation and define canonical left and right states on the centers, giving a sphericality characterization of minimality.

What would settle it

Take a connected inclusion with $Z(N)\cong\mathbb{C}^2$ and $Z(M)\cong\mathbb{C}^2$, compute $D$ from the minimal indices of the four subfactors, and numerically minimize $\|\mathrm{Ind}(E)\|$ over all conditional expectations $E:M\to N$; any value strictly below $\|D\|_{\ell^2}^2$ would disprove the claimed equality, as would the appearance of two distinct minimizers with that value.

Watch

Extended reading notes

Core claim

The central claim is that for a unital inclusion $N\subset M$ with finite index and finite-dimensional centers, connected in the sense that $Z(N)\cap Z(M)=\mathbb{C}1$, the minimal index is captured by a finite matrix. Writing $p_1,\dots,p_m$ and $q_1,\dots,q_n$ for the minimal central projections of $M$ and $N$, and setting $d_{ij}=[M_{ij}:N_{ij}]_0^{1/2}$ when $p_iq_j\neq 0$ and $0$ otherwise, one has $[M:N]_0=\|D\|_{\ell^2}^2$. The minimizing conditional expectation is then unique and is assembled from the unique minimal expectations of the subfactors $N_{ij}\subset M_{ij}$, weighted by the Perron-Frobenius eigenvectors of $D^tD$ and $DD^t$; the weights are $\lambda^{E_0}_{ij}=d_{ij}\mu_i^{1/2}\nu_j^{1/2}/d$, where $\mu$ and $\nu$ are the normalized positive eigenvectors and $d=\|D\|_{\ell^2}$.

Load-bearing premise

The load-bearing premise is that the centers of $N$ and $M$ are finite-dimensional, so the matrix dimension is a finite matrix and the Perron-Frobenius argument is finite-dimensional; without this assumption the definition and the conclusion are no longer available.

Editorial extensions

If this is right

For any connected finite-index inclusion with finite-dimensional centers, the minimal index can be read off from finitely many subfactor indices: sum the squares of the entries of $D$; no infinite search over conditional expectations is needed.
The minimal conditional expectation is unique and is built from the minimal expectations of the subfactors with Perron-Frobenius weights, so the structure of the optimizer is known once $D$ is known.
The possible values of the minimal index are quantized: either $4\cos^2(\pi/k)$ for some integer $k\ge 3$, or at least $4$.
For multi-matrix inclusions the matrix dimension coincides with the classical inclusion matrix, the inclusion is always extremal, and super-extremality is equivalent to $\Lambda^t h=d^2 k$, yielding index $\|h\|^2/\|k\|^2$; every positive integer occurs this way.
The scalar dimension is only submultiplicative in general, while the matrix dimension is exactly multiplicative and additive, so passing to matrices restores the clean composition laws familiar from subfactors.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the same matrix-dimension idea extends to atomic infinite-dimensional centers, the minimal index would plausibly be the norm of an infinite matrix, but the paper's observation that minimizers need not be unique suggests a canonical choice would require extra structure.
In the quantum-information reading, the Perron-Frobenius vectors define canonical states on the classical centers; these states could serve as reference weights for sector-wise versions of energy or dimension bounds for channels, refining bounds that use a single trace.
The equality of the matrix dimension with the inclusion matrix for multi-matrix inclusions suggests that the minimal index is the natural operator-algebraic dimension of an inclusion channel; checking whether the Perron-Frobenius vector of $D^tD$ always coincides with the Markov trace vector would sharpen the relation between minimal and trace theories.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Nothing in this note is formally new—all main theorems are quoted from the author's joint paper with Longo—but as a conference announcement it is honest, well-organized, and useful. The headline result, that for connected inclusions with finite index and finite-dimensional centers the minimal index equals the squared l2-norm of the matrix D of square roots of subfactor minimal indices, is a genuine unification of the multi-matrix and connected type II_1 cases. The Perron–Frobenius description of the unique minimal expectation and the associated left/right/spherical states is the real technical heart, and the note lays it out clearly. I also like the open problems, especially 4.1 and 4.3 on infinite-dimensional centers, and the connection to rigid 2-C*-categories.

The soft spots are exactly what the author discloses: no proofs appear here, so the correctness of Theorem 2.2 rests entirely on [GL19]. That is not circular reasoning—the note says it is reporting on that paper—but it does mean this text is not self-contained. A reader cannot verify the central equality without going to GL19. There is also a small trap in the notation: ||D||_{l2} means the operator norm, not the Frobenius norm. The stress-test note already flags this; the paper's own formulas (the Perron–Frobenius equations in Theorem 2.3) force that reading, so I think the text is correct, but someone skimming could misread.

The paper is heavy on self-citation, but that is the nature of an announcement. It does not hide the reliance. I would accept it for peer review in a proceedings or short-note format; a serious referee should check that GL19 indeed contains the stated results and that the open problems are not answered elsewhere. As a contribution to the literature it is minor—the results are already published—but the synthesis has value, and the Landauer connection is a nice hook.

Referee Report

0 major / 6 minor

Summary. This note announces results, proved in the companion paper [GL19], on the minimal index of unital inclusions N ⊂ M of von Neumann algebras with finite-dimensional centers. For a connected inclusion of finite index, Theorem 2.2 states that the minimal index [M : N]_0 equals the squared l2-norm of the m × n matrix D whose entries are the square roots of the minimal indices of the reduced subfactors N p_i q_j ⊂ q_j M p_i q_j. Theorem 2.3 describes the unique minimal expectation in terms of Perron–Frobenius eigenvectors of D^t D and DD^t, yielding canonical left, right, and spherical states. The paper further discusses extremal and super-extremal inclusions of multi-matrix algebras, derives that super-extremality forces the index to be the ratio of algebraic dimensions, and lists open problems about infinite-dimensional centers and 2-C*-categorical formulations.

Significance. If the quoted results are correct, they give a complete and conceptually clean description of the minimal index and the minimizing conditional expectation for a broad class of non-factorial inclusions, reducing the problem to Perron–Frobenius theory for a finite matrix of subfactor indices. The construction is parameter-free and the formula is directly testable in examples, and it generalizes the known multi-matrix and connected finite-inclusion results. The main caveat, which the paper itself discloses, is that all central theorems are quoted from [GL19] and no proofs appear here; the note therefore functions as a research announcement rather than a self-contained proof. I found no internal inconsistency: the Perron–Frobenius equations in Theorem 2.3 are mutually consistent, and they force \|D\|_{l2} to be read as the operator norm rather than the Frobenius norm. I also credit the paper for giving precise open problems and for being transparent about the division of labor with [GL19].

minor comments (6)

[§2, Theorem 2.2] The notation \|D\|_{l2} is not explicitly defined; please state that it is the operator (spectral) norm of the matrix D acting on l2, not the Frobenius norm. Without this clarification the central identity is easy to misread: for D = [[1,1],[0,1]], the Frobenius norm squared is 3, whereas the Perron–Frobenius equations in Theorem 2.3 force the value d^2 ≈ 2.618.
[§3, before Definition 3.1] The expression Ind(Eτ) = [M : N]1 in the paragraph before Definition 3.1 uses an undefined notation [M : N]_1; please clarify whether this is the Jones index [M : N] as used elsewhere in the paper.
[§1, last paragraph] The multiplicativity and additivity of the matrix dimension D_α for general quantum channels are asserted before D_α has been defined; please add a forward reference to the definition in Theorem 2.2 and a precise citation to [GL19] or [Lon18] for these properties.
[§2, first paragraph] The sentence 'it equals +∞ if M is way bigger than N' is informal and potentially misleading, since proper inclusions can have finite index; please rephrase.
[§3, Proposition 3.3] The claim that every positive integer occurs as the index of a super-extremal multi-matrix inclusion would benefit from a brief explanation or a precise pointer to the construction in [GL19], as the present note gives no example.
[Abstract and Introduction] Please proofread for line-break artifacts such as 'a nd' and 'subfa ctor'; they do not affect the mathematics.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the main theorems are quoted from the companion paper [GL19] rather than derived here, but they are not definitionally equivalent to their inputs.

full rationale

The paper is a research announcement: Theorem 2.2 and Theorem 2.3 are stated with the attribution '[GL19]' and no proof is reproduced. This is an omitted-proof / self-citation limitation, explicitly disclosed in the abstract ('we report on our analysis, contained in [GL19]'), but it is not circularity. The matrix dimension D is defined from the square roots of the minimal indices of the corner subfactors ([M_ij:N_ij]_0^{1/2}), and the asserted equality [M:N]_0 = ||D||^2_l2 is a nontrivial Perron-Frobenius statement about the operator norm of that matrix; neither side is defined in terms of the other. The formula for the minimal expectation in Theorem 2.3 uses the Perron-Frobenius eigenvectors of D, and the weighted additivity identity d = sum d_ij nu_j^{1/2} mu_i^{1/2} is a consequence of the eigenvalue equations, not a fitted parameter. The multi-matrix identification D = Lambda in Theorem 3.2 is a consistency check with the Bratteli matrix, not a renaming that imports the conclusion. The reliance on [GL19] means the note is not self-contained, but that is a completeness matter; no load-bearing step reduces the main claim to its own inputs.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The note's central theorems are taken from the authors' companion paper [GL19] and are assumed without proof, along with standard background from index theory and Perron-Frobenius theory. The matrix dimension is a definition, not a fitted parameter or a postulated entity.

assumptions (4)

domain assumption Theorems 2.2 and 2.3 of [GL19]: minimal index is the squared l2-norm of D and the PF characterization of E0
The note states these as theorems but provides no proof, citing the companion paper [GL19].
domain assumption Kosaki index theory and uniqueness of minimal expectation for connected inclusions (Havet, Teruya)
Used to define d_ij and E0^ij; cited from [Kos86], [Hia88], [Lon89], [Hav90], [Ter92].
standard math Perron-Frobenius theorem for irreducible nonnegative matrices
Used in Theorem 2.3 to produce the unique positive eigenvectors defining the left and right states and the minimizing weights. Connectedness ensures irreducibility of D D^t and D^t D.
domain assumption Connectedness can be assumed without loss of generality via direct-sum decomposition
Stated in Section 2; reduces the general finite-dimensional-center case to connected inclusions.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Minimal index and dimension for inclusions of von Neumann algebras with finite-dimensional centers." pith.science (2026). https://pith.science/paper/RFUR5BPO

@misc{pith2026190809121,
  author       = {Pith},
  title        = {Pith review of: Minimal index and dimension for inclusions of von Neumann algebras with finite-dimensional centers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RFUR5BPO}},
  note         = {Machine review of arXiv:1908.09121}
}

abstract

The notion of index for inclusions of von Neumann algebras goes back to a seminal work of Jones on subfactors of type ${I\!I}_1$. In the absence of a trace, one can still define the index of a conditional expectation associated to a subfactor and look for expectations that minimize the index. This value is called the minimal index of the subfactor. We report on our analysis, contained in [GL19], of the minimal index for inclusions of arbitrary von Neumann algebras (not necessarily finite, nor factorial) with finite-dimensional centers. Our results generalize some aspects of the Jones index for multi-matrix inclusions (finite direct sums of matrix algebras), e.g., the minimal index always equals the squared norm of a matrix, that we call \emph{matrix dimension}, as it is the case for multi-matrices with respect to the Bratteli inclusion matrix. We also mention how the theory of minimal index can be formulated in the purely algebraic context of rigid 2-$C^*$-categories.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 28 canonical work pages

[1]

C. H. Bennett. Notes on L andauer's principle, reversible computation and M axwell's demon. Studies in History and Philosophy of Modern Physics , 34:501--510, 2003

work page 2003
[2]

Brothier, M

A. Brothier, M. Hartglass, and D. Penneys. Rigid C^* -tensor categories of bimodules over interpolated free group factors. J. Math. Phys. , 53:123525 (43 pages), 2012

work page 2012
[3]

Bischoff, Y

M. Bischoff, Y. Kawahigashi, R. Longo, and K.-H. Rehren. Tensor categories and endomorphisms of von N eumann algebras. With applications to quantum field theory . Springer Briefs in Mathematical Physics , Vol. 3. Springer, Cham, 2015

work page 2015
[4]

Bratteli

O. Bratteli. Inductive limits of finite dimensional C^ -algebras. Trans. Amer. Math. Soc. , 171:195--234, 1972

work page 1972
[5]

M. Burns. Subfactors, planar algebras and rotations . PhD thesis, University of California at Berkeley, 2003. Available at http://arxiv.org/abs/1111.1362v1

work page Pith review arXiv 2003
[6]

Etingof, S

P. Etingof, S. Gelaki, D. Nikshych, and V. Ostrik. Tensor categories . Mathematical Surveys and Monographs , Vol. 205. American Mathematical Society, Providence, RI, 2015

work page 2015
[7]

Fidaleo and T

F. Fidaleo and T. Isola. Minimal expectations for inclusions with atomic centres. Internat. J. Math. , 7:307--327, 1996

work page 1996
[8]

F. M. Goodman, P. de la Harpe, and V. F. R. Jones. Coxeter graphs and towers of algebras . Mathematical Sciences Research Institute Publications , Vol. 14. Springer-Verlag, New York, 1989

work page 1989

Show all 28 references

[9]

Giorgetti and R

L. Giorgetti and R. Longo. Minimal index and dimension for 2- C^* -categories with finite-dimensional centers. Comm. Math. Phys. , 370:719--757, 2019

2019
[10]

P. Ghez, R. Lima, and J. E. Roberts. W^ -categories. Pacific J. Math. , 120:79--109, 1985

1985
[11]

Giorgetti and W

L. Giorgetti and W. Yuan. Realization of rigid C^* -tensor categories via T omita bimodules. J. Operator Theory , 81:433--479, 2019

2019
[12]

J.-F. Havet. Esp \'e rance conditionnelle minimale. J. Operator Theory , 24:33--55, 1990

1990
[13]

F. Hiai. Minimizing indices of conditional expectations onto a subfactor. Publ. Res. Inst. Math. Sci. , 24:673--678, 1988

1988
[14]

Hayashi and S

T. Hayashi and S. Yamagami. Amenable tensor categories and their realizations as AFD bimodules. J. Funct. Anal. , 172:19--75, 2000

2000
[15]

Jolissaint

P. Jolissaint. Index for pairs of finite von N eumann algebras. Pacific J. Math. , 146:43--70, 1990

1990
[16]

Jolissaint

P. Jolissaint. Indice d'esp \'e rances conditionnelles et alg \`e bres de von N eumann finies. Math. Scand. , 68:221--246, 1991

1991
[17]

V. F. R. Jones. Index for subfactors . Invent. Math. , 72:1--25, 1983

1983
[18]

H. Kosaki. Extension of J ones' theory on index to arbitrary factors . J. Funct. Anal. , 66:123--140, 1986

1986
[19]

Landauer

R. Landauer. Irreversibility and heat generation in the computing process. IBM Journal of Research and Development , 5:183--191, 1961

1961
[20]

R. Longo. Index of subfactors and statistics of quantum fields. I . Comm. Math. Phys. , 126:217--247, 1989

1989
[21]

R. Longo. On L andauer's principle and bound for infinite systems. Comm. Math. Phys. , 363:531--560, 2018

2018
[22]

Longo and J

R. Longo and J. E. Roberts. A theory of dimension . K -Theory , 11:103--159, 1997

1997
[23]

Ohya and D

M. Ohya and D. Petz. Quantum entropy and its use . Texts and Monographs in Physics. Springer-Verlag, Berlin, 1993

1993
[24]

Pimsner and S

M. Pimsner and S. Popa. Entropy and index for subfactors . Ann. Sci. Ecole Norm. Sup , 19:57--106, 1986

1986
[25]

Pimsner and S

M. Pimsner and S. Popa. Finite-dimensional approximation of pairs of algebras and obstructions for the index. J. Funct. Anal. , 98:270--291, 1991

1991
[26]

M. B. Plenio and V. Vitelli. The physics of forgetting: L andauer's erasure principle and information theory. Contemp. Phys. , 42:25--60, 2001

2001
[27]

T. Teruya. Index for von N eumann algebras with finite-dimensional centers. Publ. Res. Inst. Math. Sci. , 28:437--453, 1992

1992
[28]

Yamagami

S. Yamagami. C^ -tensor categories and free product bimodules. J. Funct. Anal. , 197:323--346, 2003

2003

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

C. H. Bennett. Notes on L andauer's principle, reversible computation and M axwell's demon. Studies in History and Philosophy of Modern Physics , 34:501--510, 2003

work page 2003

[2] [2]

Brothier, M

A. Brothier, M. Hartglass, and D. Penneys. Rigid C^* -tensor categories of bimodules over interpolated free group factors. J. Math. Phys. , 53:123525 (43 pages), 2012

work page 2012

[3] [3]

Bischoff, Y

M. Bischoff, Y. Kawahigashi, R. Longo, and K.-H. Rehren. Tensor categories and endomorphisms of von N eumann algebras. With applications to quantum field theory . Springer Briefs in Mathematical Physics , Vol. 3. Springer, Cham, 2015

work page 2015

[4] [4]

Bratteli

O. Bratteli. Inductive limits of finite dimensional C^ -algebras. Trans. Amer. Math. Soc. , 171:195--234, 1972

work page 1972

[5] [5]

M. Burns. Subfactors, planar algebras and rotations . PhD thesis, University of California at Berkeley, 2003. Available at http://arxiv.org/abs/1111.1362v1

work page Pith review arXiv 2003

[6] [6]

Etingof, S

P. Etingof, S. Gelaki, D. Nikshych, and V. Ostrik. Tensor categories . Mathematical Surveys and Monographs , Vol. 205. American Mathematical Society, Providence, RI, 2015

work page 2015

[7] [7]

Fidaleo and T

F. Fidaleo and T. Isola. Minimal expectations for inclusions with atomic centres. Internat. J. Math. , 7:307--327, 1996

work page 1996

[8] [8]

F. M. Goodman, P. de la Harpe, and V. F. R. Jones. Coxeter graphs and towers of algebras . Mathematical Sciences Research Institute Publications , Vol. 14. Springer-Verlag, New York, 1989

work page 1989