Evidence for identity names and motion

Bounded review, 4 October 2026. This records evidence for three pending choices, not approval or a full-system audit.

History and inspected examples

• Read the complete current decisions/brand-identity-review.md and the original 16 August dictation in docs/reference/2026-08/2026-08-16/brand-identity-retail-round-notes.md. The name and extraction request are approved; that is not approval of every extracted value.
• Read the complete source archive/reference/og-bt-docs-research-ref/__REF/database-ref/brand-schema/brandIdentity.csv. Brand Book describes strategy/vision/values, Style Guide describes visual usage, and Digital Guidelines describes application scope. A blanket title-based alias hides those differences.
• Rendered and visually inspected pages 16 and 17 of archive/reference/brand-guidelines-ref/github/GitHub Brand Guidelines 2025.pdf: both are titled Key Art 01, with the central composition on different backgrounds. Text extraction produced no useful matches, so the visual pages are the evidence. No claim to have reviewed all 73 pages in this batch.
• Read the reusable and client-specific sections of docs/reference/motion-discovery/MOTION-DISCOVERY-FRAMEWORK.md. The reader page uses only the general conceptual distinction. Client-specific material is not reproduced.
• Read OpenAI's public guidelines and IBM's animation overview. IBM's own navigation includes Animation; the earlier universal-sounding premise that motion must sit below Digital is not a sound basis for the decision. No animation interval was classified or certified.
• Two initial IBM/Carbon URLs were inaccessible. A focused search located the official IBM overview above; inaccessible pages support no claim.

Selected live schema and records

Exact queries and raw results are in docs/reference/2026-10/2026-10-04/identity-names-and-motion-live-evidence.json.

The information_schema.columns probe covered schema_enums, taxonomy_alias and entity_motion_identity. A subsequent exact-name registry lookup for brand_guidelines, brand_book, style_guide, digital_guidelines, key_art and key_visual returned no rows. The selected alias lookup also returned no rows. This does not establish that equivalent labels, documents or classification data are absent elsewhere.

The probed alias relation exposes an enum name, raw and canonical names, status and supporting metadata. It does not expose a content-context condition among those columns. A conditional recommendation must not be implemented as an unconditional string replacement.

Local consumers and risks to verify later

• backend/services/post_processor.py, _load_aliases and _resolve_soft_enum: the local resolver uses an approved per-vocabulary string map after exact matching and before fuzzy matching. It does not inspect document purpose inside that function. This is an affected mechanism to audit, not proof that every content-type path calls it.
• backend/routers/pipeline_scene_classify.py, motion-identity write block: local code links selected brand/typography/logo moments to an entity based on family or element. That observation link must not be read as proof of a brand's declared motion policy. Schema, consumer semantics, provenance and corrections still need the comprehensive trace.
• Shared headings affect discovery and query interpretation. They do not by themselves authorize extra content records, duplicated identity systems or a new physical hierarchy.

No database, prompt, API, schedule or classification changes were made for this packet. No deployed equivalence, full permissions audit or end-to-end correction guarantee is claimed.

Choice mapping and remaining coverage

B-01/guideline-names and B-01/key-art split the original bundled aliases question. B-04/motion-heading asks only about the identity heading. Their parent identifiers and original questions remain in the existing inventory. B-02, the remaining B-03 definitions, B-05, additional piece kinds and the exact motion vocabulary are not closed by this packet.