What is actually aggregated
A page is only one support. The system also collects titles, sections, structured data, links, outside citations, repeated phrasings, language versions, and sometimes off-site environment.
The computed entity is therefore not a faithful copy of one page. It is a reconstruction from multiple traces whose status and freshness are not equal.
How a role ends up emerging
When a name repeatedly appears next to a service, a product, an expertise, or an organisation, the system tends to connect those signals. Through repeated proximity it may produce an implicit role: founder, expert, spokesperson, author, brand, or category.
That emergent role is not always explicitly published. It may result from an accumulation of micro-signals strong enough to become a stable reading.
Why the entity exceeds the page
A governed entity is not limited to one document. It moves across pages, languages, traces, and sometimes multiple domains. That is why local page quality can coexist with a distorted global identity.
A site may look clean at the editorial level while remaining fragile at the entity level: role confusion, surface fusion, scope expansion, or displaced authority.
The governance consequence
Governing pages alone leaves the aggregation problem intact. Governing the entity means declaring permitted relations, role boundaries, negations, and canonical dependencies across surfaces.
The passage from page to entity is therefore not a technical detail. It is the zone where a system either becomes able to produce a defensible reading or falls back to a merely plausible synthesis.
Links and continuity
- Topic: Interpretation phenomena — Vocabulary for recurrent drifts, collisions, and reductions.
- A site that ranks well but is poorly understood — When discoverability does not stabilise the computed entity.
- Governed identity graph — Declare the relations, roles, and boundaries that should resist synthesis.