Statement

Every is a . Indeed, if π:AB(Hπ)\pi:A\to\mathcal B(H_\pi) is an , liminality gives

π(A)=K(Hπ),\pi(A)=K(H_\pi),

whereas the type I condition asks only that

K(Hπ)π(A).K(H_\pi)\subseteq\pi(A).

Equality therefore implies the required containment for every irreducible representation. In the older terminology, every CCR algebra is GCR, or every liminal algebra is postliminal. No separability or unitality hypothesis is needed for this implication. The proof is purely definitional.

Why the converse fails

The containment may be proper. If HH is infinite-dimensional, the unitization K(H)+CIHK(H)+\mathbb C I_H is type I but not liminal: its defining contains K(H)K(H), yet its image also contains the identity. Thus type I is strictly weaker than liminal Dixmier, §4.2.

Position in the hierarchy

The implication places liminal algebras inside the type I class while keeping their stronger kernel rigidity. For a liminal algebra, every irreducible quotient is represented exactly by . A general type I algebra may instead be assembled from liminal or continuous-trace subquotients through an ordinal ideal series.

References
  1. Jacques Dixmier, C-Algebras, North-Holland Mathematical Library 15, North-Holland, 1977. Publisher record. Relevant: §4.2 on liminal algebras and Chapter 4 on postliminal algebras.
  2. Gert K. Pedersen, C-Algebras and Their Automorphism Groups, 2nd ed., Academic Press, 2018. DOI record. Relevant: Chapter 6 on the CCR and GCR hierarchy.