Two.
That is the practical floor beneath Wikipedia's general notability guideline, even though the guideline text never states it as a number. It asks only for 'significant coverage in multiple reliable sources that are independent of the subject' — plural, undefined, left for reviewing editors to work out case by case. In practice, over two decades of deletion discussions, that word has settled into a floor: one substantial, independent source establishes that a subject exists. It takes a second before an editor will call it a pattern worth an encyclopedia entry.
A profile can run three thousand words, carry a staff byline, sit behind no paywall, and still not clear the guideline on its own. Length measures how much one publication decided to say. It says nothing about whether any other publication, working independently, reached the same judgment — that the subject was worth writing about. A single outlet's attention is a data point; an encyclopedia entry needs a trend.
The test reviewing editors actually run is blunt: strip the article down to its two best sources and ask what is left. If the lead paragraph — birth, career, the fact that makes the subject notable — depends on a single citation, the article still has work ahead of it, whatever else is stacked underneath it. Ten routine listings, four passing mentions and one gala photo caption do not add a second source; they add decoration around the first.
A second citation only counts if it reflects an independent editorial decision. A wire story picked up by forty regional papers is one judgment, repeated by copy-paste; an interview transcript run under two different mastheads is the same interview twice. Reviewing editors read past the byline count to ask whether two separate newsrooms actually looked at the subject and decided, on their own, that a story was warranted.
Multiple sources are generally expected... In some cases, multiple newspaper articles or television programs from a single publisher or media group may be treated as a single source for the purposes of establishing notability.
None of this makes three sources better than two, or twenty better than three, once the floor is cleared; the guideline weighs depth and independence, not a running count. But nothing below that floor gets weighed at all. A file with one excellent profile is not a weak case — it is not yet a case. The second footnote is usually the difference between a draft that waits and an article that stands.
