Arveson's Hyperrigidity Conjecture

• Arveson's Hyperrigidity Conjecture is a proposal asserting that an operator system is hyperrigid if all its irreducible *-representations serve as boundary representations with the unique extension property.
• The framework unifies key concepts including maximal dilations, the noncommutative Choquet boundary, and the C*-envelope, linking abstract dilation theory to concrete algebraic structures.
• Recent developments illustrate precise criteria in C*-correspondences and commutative models, while counterexamples demonstrate that boundary information alone may not guarantee hyperrigidity.

Arveson’s Hyperrigidity Conjecture is the proposal that, for a separable operator system $S$ generating a C$^*$-algebra $C^*(S)$, hyperrigidity should be determined at the level of irreducible representations: if every irreducible $*$-representation of $C^*(S)$ is a boundary representation for $S$, then $S$ should be hyperrigid. The conjecture places the unique extension property, maximal dilations, the noncommutative Choquet boundary, and the C$^*$-envelope into a single framework. Subsequent work has both deepened this framework in major classes—especially C$^*$-correspondences, function systems, and geometric quotient constructions—and shown that the conjecture is false in full generality, including type I and finite-dimensional settings (Bilich et al., 2024, Scherer, 14 Feb 2025, Dessi et al., 20 Mar 2025).

1. Hyperrigidity, unique extension, and maximality

Let $S\subseteq C^*(S)$ be an operator system. Arveson’s asymptotic definition says that $^*$0 is hyperrigid if, for every faithful $^*$1-representation $^*$2 and every sequence of u.c.p. maps $^*$3 with

$^*$4

one has

$^*$5

In the representation-theoretic formulation used throughout the later literature, hyperrigidity is equivalent to the statement that every $^*$6-representation of $^*$7 has the unique extension property with respect to $^*$8 (Thompson, 2023).

The unique extension property means that if a u.c.p. map agrees with a representation on $^*$9, then it must equal that representation on all of $C^*(S)$0. In dilation-theoretic language, this is equivalent to maximality: a u.c.p. map or representation is maximal when every dilation is trivial. For operator systems and, more generally, for generating sets, maximality, the u.c.p. unique extension property, and the c.c.p. unique extension property coincide under the hypotheses isolated in recent refinements of Arveson’s criterion (Dessi et al., 20 Mar 2025).

This equivalence is the basic structural mechanism behind the conjecture. Hyperrigidity is a global statement about all representations, whereas the conjecture asks whether it is enough to test maximality only on irreducible ones.

2. Boundary representations, the noncommutative Choquet boundary, and the C$C^*(S)$1-envelope

A boundary representation for $C^*(S)$2 is an irreducible $C^*(S)$3-representation $C^*(S)$4 of $C^*(S)$5 such that $C^*(S)$6 has the unique extension property. Boundary representations form the noncommutative Choquet boundary, and their kernels determine the Shilov ideal. Quotienting by that ideal yields the C$C^*(S)$7-envelope.

In the classical commutative theory, the Choquet boundary governs Korovkin-type approximation. Arveson’s conjecture was the noncommutative analogue of this principle: if the noncommutative Choquet boundary is as large as possible—equivalently, if every irreducible representation is boundary—then $C^*(S)$8 should be hyperrigid. One implication is immediate: hyperrigidity forces every representation, hence every irreducible representation, to have the unique extension property (Thompson, 2023).

For commutative function systems, Davidson–Kennedy reduce the conjectural mechanism to a comparison between the classical Choquet order on measures and a dilation-theoretic order adapted to the unique extension property (Davidson et al., 2016). This reduction shows that, even in the commutative case, hyperrigidity is not merely a statement about extreme points of state spaces; it is a statement about the way dilation maximality propagates from pure objects to arbitrary representations.

A central theme in subsequent work is that the C$C^*(S)$9-envelope records boundary information, but hyperrigidity requires more than envelope minimality. This distinction becomes decisive in the later counterexamples.

3. Approximate, local, and order-theoretic reformulations

Several later developments reframed the conjecture without changing its core content. One direction replaces exact uniqueness by approximate uniqueness. The approximate unique extension property (AUEP) requires that if a u.c.p. extension agrees with a representation on the generating operator space, then it is approximately unitarily equivalent to that representation. AUEP is stable under approximate unitary equivalence, unlike the exact UEP, and in the separable setting it yields an “approximate hyperrigidity ladder”: $*$0 In the postliminal case, the approximate and exact formulations partially collapse (Thompson, 2023).

A second direction is local. Using characteristic sequences, one obtains a localized version of hyperrigidity: if a state $*$1 admits a characteristic sequence $*$2 in the operator system, then any u.c.p. extension $*$3 of a representation $*$4 satisfies

$*$5

for every $*$6. This is a state-by-state rigidity statement, rather than a global one, and it generalizes Arveson’s commutative local theorem to arbitrary operator systems (Clouâtre, 2017).

A third direction is order-theoretic. Clouâtre’s work on unperforated pairs and weak expectation type relaxations gives state-based criteria ensuring unique extension for broader classes of representations (Clouâtre, 2017). Clouâtre and Saikia then reinterpret the conjecture via the dilation order on the state space and show that dilation maximal states form a norm-closed face, encoded by a boundary projection $*$7. Under the hypothesis that every pure state is dilation-maximal, hyperrigidity becomes equivalent to a topological regularity property of this boundary projection—namely, that $*$8 is closed, or equivalently an infimum of open projections (Clouâtre et al., 2023). This suggests that the conjecture can be read as a noncommutative topological regularity problem in the bidual.

4. C$*$9-correspondences, tensor algebras, and structural criteria

One of the strongest positive results concerns tensor algebras of C$C^*(S)$0-correspondences. If $C^*(S)$1 is a non-degenerate C$C^*(S)$2-correspondence over $C^*(S)$3, with tensor algebra $C^*(S)$4, Cuntz–Pimsner algebra $C^*(S)$5, and Katsura ideal $C^*(S)$6, then hyperrigidity is completely characterized by the action of $C^*(S)$7. Specifically,

$C^*(S)$8

The same condition is equivalent to hyperrigidity of the associated selfadjoint operator space $C^*(S)$9, and the paper establishing this equivalence also sharpens the relation between hyperrigidity, maximality, and unique extension for general generating sets (Dessi et al., 20 Mar 2025).

This resolves a question left open by earlier work of Katsoulis–Ramsey and unifies it with Kim’s criterion for the selfadjoint operator space. In this class, hyperrigidity is not a subtle asymptotic phenomenon; it is controlled by the concrete module-theoretic condition $S$0. The result is therefore a particularly clear realization of Arveson’s original philosophy.

A complementary development is Bilich’s characterization of maximal representations of correspondences. There maximality is equivalent to full Cuntz–Pimsner covariance, expressed by

$S$1

where $S$2 is the support projection of the left action in the bidual. This gives an explicit description of the noncommutative Choquet boundary for correspondence algebras and leads to new counterexamples to the conjecture. In particular, for correspondences the boundary can equal the entire unitary dual while hyperrigidity still fails; nevertheless, the conjecture remains valid for proper correspondences and for correspondences arising from topological graphs (Bilich, 2024).

5. Commutative models, convex geometry, and essential normality

The commutative case remains a major testing ground. For compact convex sets $S$3, the function system $S$4 of continuous affine functions is hyperrigid in $S$5. This gives a broad positive verification of the conjectural picture in dimension two and yields, as an operator-theoretic corollary, that the weak and strong operator topologies coincide on the set of normal operators whose spectrum lies in $S$6 (Scherer, 2024).

A further geometric extension shows that if $S$7 is a compact spectrahedron with closed extreme point set, then $S$8 is hyperrigid in $S$9 (Scherer, 22 Jan 2026). Since all currently known general counterexamples are noncommutative, these spectrahedral and planar convex results suggest that commutative function systems continue to exhibit a more classical Choquet-type rigidity mechanism.

Another major family comes from quotient algebras of complete Nevanlinna–Pick spaces. For the norm-closed quotient algebra $S$0 associated with a weak-$S$1 closed ideal $S$2, hyperrigidity is equivalent to two conditions: essential normality of the compressed multiplier tuple $S$3, and failure of complete isometry of the Gelfand transform. Thus, in this setting hyperrigidity is detectable through essential normality, extending the Kennedy–Shalit picture from Drury–Arveson space to general maximal regular unitarily invariant complete Nevanlinna–Pick spaces (Clouâtre et al., 2019).

The compressed $S$4-shift provides the prototype of this interaction. Kennedy and Shalit proved that, for the compression $S$5 of the $S$6-shift to the orthogonal complement of a homogeneous ideal, essential normality is equivalent to hyperrigidity of $S$7 as a generating set of the Toeplitz algebra $S$8 (Kennedy et al., 2013). This does not identify essential normality with hyperrigidity in complete generality, but it shows that in several natural commutative and nearly commutative models the two notions are tightly coupled.

6. Counterexamples, obstructions, and present status

The conjecture is false in general. Bilich and Dor-On constructed a counterexample with a type I C$S$9-algebra generated by a single operator: the noncommutative Choquet boundary is the entire spectrum, yet the generating operator algebra is not hyperrigid (Bilich et al., 2024). Scherer then produced a finite-dimensional counterexample: an operator system generated by four operators that is not hyperrigid although all restrictions of irreducible representations have the unique extension property (Scherer, 14 Feb 2025). These results show that the implication

$^*$0

fails even in highly concrete settings.

Subsequent work identified a new obstruction. Given an operator system $^*$1 and an ideal $^*$2, one may enlarge $^*$3 to an operator system $^*$4. Hyperrigidity of $^*$5 then forces an additional rigidity requirement: every $^*$6-representation of $^*$7 annihilating $^*$8 must admit a unique contractive completely positive extension from $^*$9 to $^*$0. The obstruction is encoded by the mutual orthogonality of the atomic projection of $^*$1 and the support projection of $^*$2, and this mechanism recovers the Bilich–Dor-On example when $^*$3 is the compact ideal (Clouâtre, 23 Sep 2025).

The present status is therefore bifurcated. On the one hand, the original conjecture is refuted as a universal principle. On the other hand, the program it initiated remains structurally productive: in tensor algebras of non-degenerate C$^*$4-correspondences, in compact convex function systems in $^*$5, in compact spectrahedra with closed extreme boundary, and in several essential-normality regimes, hyperrigidity admits precise positive criteria (Dessi et al., 20 Mar 2025, Scherer, 2024, Scherer, 22 Jan 2026). A persistent misconception is that full noncommutative Choquet boundary should automatically force hyperrigidity; the counterexamples show that boundary information alone can be insufficient. A plausible implication is that the enduring content of Arveson’s conjecture now lies less in its original universal form than in the search for exact structural hypotheses under which boundary representations do control all dilations.