Two sourced facts, placed side by side, can add up to a claim that neither source actually makes — and Wikipedia has a rule for exactly that gap. Editors call it synthesis, and violating it is one of the quieter ways an otherwise well-cited article gets flagged, rewritten, or dragged to a deletion discussion.
The policy against original research (WP:NOR) covers the obvious case — inventing a fact with no source at all. Its subtler clause, WP:SYNTH, covers something practitioners trip over far more often: combining verifiable material from separate sources to imply a conclusion that none of those sources states. Each fact can be true. Each source can be reliable. The problem sits in the seam between them, where an editor's own argument fills a gap the sources left open.
A profile once paired two accurate facts: an industry report noting the sector had grown forty percent over three years, and a press release noting the company's revenue had roughly doubled in the same period. Placed in adjacent sentences under a single heading, 'Market position,' the pairing read as a claim the article never stated outright — that the company had captured a disproportionate share of that growth. No source made that claim. An editor removed the section within a week, citing WP:SYNTH, not because either sentence was false, but because their proximity manufactured a conclusion nobody had written.
Do not combine material from multiple sources to reach or imply a conclusion that is not explicitly stated by any of the sources.
Context still belongs in an article. A biography can note a subject's graduation year and, in a separate sentence, a company's founding date, without implying that one caused the other. The forbidden move is the bridge: a sentence, a heading, or even a paragraph break engineered so the reader draws a connection the sources themselves never drew.
For anyone drafting or reviewing a page, truth alone does not settle the question. What settles it is whether a reader who checked every citation would find that exact conclusion stated somewhere among them. A conclusion that exists only in the arrangement of the sentences does not belong in the article, however defensible it might be off the page.
The rule cuts in both directions, which is easy to forget when the implied conclusion happens to flatter a client. An assembled claim that a founder 'pioneered' a category, built from two unrelated early-mover facts, fails the same test as an assembled claim that a company is 'struggling,' built from two unrelated setbacks. Wikipedia asks only whether the words on the page already existed in that shape somewhere else. If they did not, connecting the dots was never neutral — it was authorship, without a byline.
