Semantic Authority Mesh

An engine that reads meaning places every page of a site as a point, and the site as the centre of those points. A site whose pages sit close to that centre, each using the same defined terms, linked where their meanings meet and cited by other sites on the subject, reads to the machine as one body of knowledge. The semantic authority mesh is that structure, built on purpose and measured as one: the pages and how far each sits from the centre, the terms they share, the links between them and the citations from outside. Authority is what the engines conclude; the mesh is what they read to conclude it.

Where the Term Comes From

Noo–Studio coined the term on 10 October 2026. That day no earlier use of the exact phrase was found on the web, in Google’s search suggestions, or in the arXiv, Zenodo and OpenAlex indexes of research. It joins two phrases already in use, semantic authority and semantic mesh, and names the structure behind the first.

It descends from the topical mesh, Laurent Bourrelly’s method of linking a site’s pages by meaning (topicalmesh.com), and it names the structure behind answer gravity, the pull such a site has on AI answers. It is the middle of three terms the studio uses as one set: semantic token optimisation for the words, the mesh for the structure they build, and semantic entanglement optimisation for the field. The semantic gravity gradient, coined on 12 October 2026, is the slope all three measure, and the mesh measures it at the size of a site.

The paper of the same name, in preparation, measures it on real sites.

Why It Matters

Google keeps the geometry. Its internal documentation of March 2024 stores, for each site, a siteFocusScore, “how much a site is focused on one topic”, and a siteRadius, “the measure of how far page_embeddings deviate from the site_embedding”, beside embeddings of the site and of each page (the documentation). The same documents hold an “LLM-based effort estimation for article pages” (the documentation): in 2024 a language model was already scoring pages beside the map of where they sit.

For citations, Google’s news ranking looks at “how original reporting … is cited by other publishers to understand how a publication is influential and authoritative on a topic” (Google Search Central, 2023). Since then, Google’s AI Overviews and AI Mode may answer one question with many related searches across subtopics (Google Search Central), so a site that covers the subtopics of its subject, close to its centre, can be met by more of them. The fields show what is kept.

How much they count, then or now, is not published.

What It Is Not

A topical mesh is drawn by an editor and held by links. A semantic authority mesh is read the way the engines read it, as distances in a model’s space, and it holds only while every page earns its place with first-hand content. Topical authority is the common name for the result; the mesh is the structure that earns it.

Pages added only to widen it are what Google’s spam policies call scaled content abuse (Google Search Central).

How the Studio Uses It

A System is a main site with its satellites on one engine. Each satellite teaches one subject in its own defined terms, close to its own centre, and each term is defined once and linked back to the main site. The paper sets out the measure: every page embedded with an open model, the site’s centre found, and each page’s distance from it set against what Search Console reports for that page.

How to Check It on Your Own Site

1. List your pages by subject. Does each subject have one page that defines it, and do the others link to it?
2. Read your ten newest pages. How many of them sit on the subject your firm is known for?
3. Search for your subject and note which sites are cited beside you. Those are your neighbours in meaning.