Canonical Tag

A canonical tag is a link element in the head of a page, rel="canonical", that names the one URL the publisher considers the preferred address for that content when the same content can be reached at more than one. It tells a search engine which copy to index and which address to credit.

What the relation is defined to do

RFC 6596 defines the canonical link relation. The target URL must hold content that duplicates, or is a superset of, the content on the page that declares it, and a publisher who declares it should expect a search engine to index only the target, consolidate properties such as link popularity onto it, and display it as the representative address. The target may be the page itself, may sit on a different host, and may be a view-all page that several component pages point at. A page should not point at a page that holds less than it does, because the extra content is then discarded.

A strong signal, not an instruction

Google’s canonicalization documentation ranks the ways a preference can be stated. A redirect and a rel="canonical" annotation are both strong signals; inclusion in a sitemap is a weak one; the signals stack. The engine still makes the final choice, and will override a tag that contradicts everything else it sees. The commonest self-inflicted failures are pointing every page at the home page, and pointing the parts of a multi-page piece at part one instead of at a view-all page or at themselves.

Canonical and republished content

When an article is republished on another platform, the copy carries a canonical pointing back to the original, so that the links and the standing the copy earns are credited to the page it came from. Whether an AI assistant honors the tag when it picks a URL to cite is not documented by any of them. What is documented is the search side, and an assistant that searches before it answers inherits whichever URL the index settled on.

A common question

Will an AI assistant honor a canonical tag when it picks which URL to cite?

None of the assistants document what they do with the tag, so there is no published answer. What is documented is the search side: a search engine treats the annotation as a strong signal about which copy to index and which address to credit, and it makes the final choice. An assistant that runs a search before writing its reply takes on whichever URL the index settled on, so the tag acts on citation indirectly, through the index, rather than by instructing the assistant.

Sources

The relation: RFC 6596. Signal strength and best practice: Google’s canonicalization documentation. No figures on this page.