GLM-BD ยท Schema 1.4 ยท Documentation

When Two Governance Layers Declare Each Other

GLM-BD does not establish interoperability. GLM-BD makes interoperability assessable through independently readable boundary declarations.

This page explains one specific and important case: two governance layers that each publish a GLM-BD manifest declaring composability with the other. This is materially more informative than a single declaration — and it still establishes nothing about runtime behaviour. Both halves of that sentence matter.

1. One declaration versus two

A GLM-BD manifest published by one governance layer describes how that layer sees its own boundary: its layer_type, its position on the timing axis, what it claims (authoritative_layer_claim), what it explicitly does not claim (explicit_non_claims), how consumers must not interpret it (forbidden_interpretations), and which layer types and manifests it declares itself composable with (composition).

When that manifest lists another layer under composable_with_types or composable_with_manifests, the declaration is unilateral. It tells you how Layer A describes its composability with Layer B. It tells you nothing about how Layer B describes the same thing — or whether Layer B describes it at all.

A reciprocal declaration exists when Layer B independently publishes, at its own domain, a manifest that declares composability with Layer A in return.

2. What reciprocal declarations add

GLM-BD still does not establish that the two layers interoperate. What changes is the amount and structure of information available for independent assessment.

A third party can now retrieve two declarations, each published by a different party at a different domain, and compare them — instead of relying on one system's description, an architecture diagram, or a marketing statement.

The comparison can ask, for example:

If the declarations align, that alignment is useful information. If they conflict, the conflict is equally useful — it exposes a boundary disagreement that would otherwise remain implicit.

3. What they still do not prove

A reciprocal declaration is evidence about declarations. It is not evidence about behaviour. Two mutually declared manifests do not prove:

Reciprocal declaration ≠ proven interoperability.
Reciprocal declaration = independently comparable evidence about the composability each side declares.

4. What a third party can compare

Because each manifest is published at its own /.well-known/governance-layer-manifest.json, attribution follows publication: the declaration at Layer A's domain is Layer A's, the declaration at Layer B's domain is Layer B's. That is the only attribution GLM-BD provides — there is no cryptographic signature and no counter-signature by the other side.

An integrator, auditor, reviewer or customer can therefore evaluate the two declarations independently, at any time, and repeat the evaluation whenever either manifest changes.

5. Positive alignment

If the two declarations are coherent under the Analyzer's comparison rules, the result is useful evidence of declared alignment within the scope those declarations represent.

Read it as
"The examined declarations do not expose an incompatibility on the signals the Analyzer examines."
Do not read it as
"These systems are interoperable."

6. Negative alignment

If the declarations do not fully align — for example, one side lists the other's layer type as forbidden, only one side declares the other, or both occupy the same timing position — the result is not a failure of GLM-BD. Exposing a disagreement or incomplete alignment that would otherwise stay hidden is part of the value of making boundaries machine-readable.

A negative result does not establish that the systems can never interoperate. It establishes that the declarations, as currently published, do not provide a coherent basis for that composition.

7. What the GLM-BD Analyzer actually establishes

The Analyzer (Compatibility Match tab) compares each pair of manifests on six declared signals: whether the layer types differ; whether each lists the other's layer_type under composable_with_types; whether each references the other's manifest URL under composable_with_manifests (only when both URLs are provided; the match is by exact URL or hostname); whether either manifest's layer_type appears in the other's forbidden_interpretations; whether the timing positions differ; whether the schema_version is the same.

It reports each pair as declared compatible, partially declared, or not declared compatible, with the individual signals visible in the detail panel.

It does not compare the wording of claims and non-claims, controlling_test, validity windows, manifest_version or digests, and it does not observe any runtime. Everything runs in the browser on the pasted text. A "declared compatible" result is a statement about the two declarations — nothing more.

8. Forbidden interpretations

The following are examples of interpretations that may be explicitly excluded by GLM-BD declarations and that must not be inferred from reciprocal listing alone: execution_authorization, authority_transfer, post_bind_authority_validation, interoperability_certification, and implied_composability. A manifest may declare additional forbidden interpretations. Mutual listing never lifts a limit declared by either side.

The Organization Index is unaffected by any of this. It discovers layers; it does not compose them.

9. Practical example (synthetic)

Layer A

layer_type: boundary
timing_axis.position: at-bind
composable_with_types: ["closure"]
composable_with_manifests: ["https://layer-b.example/.well-known/governance-layer-manifest.json"]
forbidden_interpretations: ["execution_authorization", "authority_transfer"]

Layer B

layer_type: closure
timing_axis.position: post-bind
composable_with_types: ["boundary"]
composable_with_manifests: ["https://layer-a.example/.well-known/governance-layer-manifest.json"]
forbidden_interpretations: ["authority_transfer", "post_bind_authority_validation"]

Fed to the Analyzer, this pair yields: distinct types · mutual type composability · mutual manifest reference · no forbidden-type conflict · distinct timing positions · same schema version → Declared compatible.

Declared alignment — established
Both manifests, independently published, describe a composability that does not contradict on the examined signals.
Runtime interoperability — not established
Whether an actual A→B handoff occurs, succeeds, or preserves authority in a given event is outside GLM-BD and must be assessed by other means.

10. Boundary statement

Boundaries make interoperability assessable.They do not make interoperability true.
← Back to the GLM-BD Analyzer