Definition

A family is coherent, under a correspondence structure, to the degree that its members agree with one another where the structure compares them. The comparisons can be:

Coherence is internal to the family. It can be exact or graded: by a smallest Laplacian eigenvalue, by a holonomy’s distance from the identity, or by direction (CLAIM-495). It is always relative to the correspondence structure, so a claim of coherence should name the maps (TERM-093).

What it is not

  • Fidelity. Fidelity is a relation between a rendering and a source, relative to a receiver and a task. The record’s senses of it are the Active fidelity TERMs, TERM-034 among them. A constant rendering is perfectly coherent and has no fidelity at all, and two faithful routes can disagree (CLAIM-496).
  • Integration. Compatible descriptions can belong to parts that do not interact. CASE-179’s “coherence without integration” arm, and CLAIM-474’s separation of compatibility from dynamics, keep the two apart.
  • Knowledge or correctness. A family can become coherent by changing how its members are expressed (THEORY-1088), or by a record admitting only the disagreements that fit (CLAIM-448). Coherence counts as evidence of correctness only where the truth is known to be coherent, as in a synchronization benchmark built flat (THEORY-1158).

Adopted

Adopted by the owner on 2026-10-29. Claims that use “coherent” or “consistent” in this sense may name it in uses. A narrower rival that restricted coherence to gluing would leave agreement between routes and around loops without a word.

Note of 2026-10-31: coexisting meanings are not yet incoherence

Bencherki, Brummans and Vézy (LIT-1902, p. 1358) set “Coexisting meanings of the same action” against Weick’s sensemaking as the reduction of equivocality. THEORY-1278’s Connections reads this as “incoherence that is not failure”. By this term’s definition it is not yet incoherence at all. Coherence is agreement under a stated correspondence structure, and the paper states none under which the meanings would have to agree. CLAIM-265 applies: disagreeing local descriptions are not thereby incompatible. The dispute with Weick is therefore about the correspondence structure, not about the meanings. Weick’s organizations impose one, since members must act on a shared reading. The paper’s organizations impose none. The question the paper leaves open is when an organization has such a structure: when coexisting meanings must be made to agree, and when they can simply coexist. No status, title or warrant changes.