No plagiarism required. The same page often exists with and without www, with and without a trailing slash, with tracking parameters from campaigns, in http and https. To a search engine each variant is a separate address, and links pointing at four variants build a quarter of the authority in four places instead of full authority in one.
A one-line tag in the page head declares the official address. Search engines consolidate the variants' signals onto it, and your rankings stop leaking into copies of yourself. Every generated page on this site declares an absolute canonical automatically, because a convention enforced by the build system never gets forgotten on page 61.
Stale canonicals that point at an old domain after a migration, pages canonicalizing to the wrong sibling, or every page pointing at the homepage. Wrong canonicals are worse than none, because you are formally instructing Google to credit the wrong page. After any migration or rebuild, auditing canonicals is step one, not an afterthought.
Yes, because duplication comes from URLs, not content volume. The tag costs nothing when generated automatically and prevents silent signal-splitting.
Redirect when the duplicate should not exist at all; canonical when both addresses must keep working, like tracking-parameter variants of a landing page.
Yes. Pointing pages at the wrong target tells search engines to index the wrong thing, and they often comply. It is one of the quietest ways a site loses positions after a redesign.
Everything in this glossary, we build and operate for real businesses. Thirty minutes maps it to yours.