Last night I found myself circling an old philosophical fault line that turns out to explain a great deal of modern arguing: English asks one word, “knowledge,” to do the work that Ancient Greek split cleanly between two.
The Greeks had gnosis and episteme. Gnosis is knowing by direct contact — the knowledge of having been somewhere, felt something, watched an event unfold in front of you. It’s personal, immediate, and doesn’t need to be argued for, because it isn’t a claim about the world so much as a report of an experience. Episteme is a different animal: knowledge built and checked, tested against other people’s observations, revised when it fails, and only trusted once it survives that process. Its root, epistamai, literally means “to stand upon” — knowledge as a platform sturdy enough to hold weight, not a private impression.
There’s a neat irony buried in the etymology. The English verb “know” isn’t a loan from episteme’s family at all — it descends, alongside gnosis, from the same ancient root, gno-, that also gives Latin its gnoscere and French its connaître. In other words, when English speakers say “I know,” the word itself is already leaning toward that personal, first-hand sense: knowing as acquaintance, as something felt or witnessed. We inherited the personal half of the pair and lost its systematic twin as a separate word.
This isn’t just a curiosity for etymologists. Most European languages kept the split alive as two ordinary verbs: French distinguishes connaître (to be acquainted with a person, place, or experience) from savoir (to know a fact or how to do something); German does the same with kennen and wissen. A French or German speaker is grammatically nudged, every time they open their mouth, to specify which kind of knowing they mean. English speakers get no such nudge. We say “I know” for both “I met her once” and “I have read the peer-reviewed literature,” and the sentence gives no sign of which one is doing the work.
There is a third Greek word worth borrowing alongside these two: doxa, meaning opinion or received belief — not tested, not personally witnessed, just widely held. Doxa is where a lot of secondhand certainty lives: things repeated so often they feel like knowledge without being either kind of real knowing. A great deal of what circulates as “everybody knows” is actually opinion dressed up as epistemic fact.
That gap is where a lot of science misinformation quietly moves in. “I know vaccines caused it, I watched it happen the next day” is personal testimony — a real, often painful, firsthand account. “Vaccines cause it,” stated flatly as a fact about the world, is an epistemic claim — a statement that needs to survive contact with everyone else’s cases, not just one. English lets the first sentence slide into the second without anyone, including the speaker, noticing the swap. The same slippage runs the other way too: “I know someone who smoked forty a day and lived to ninety-five” is offered as though a single personal anecdote could overturn an epistemic conclusion built from population-level data, when the two were never competing on the same terms. It shows up again in “I know it’s cold today, so global warming isn’t real” (a local observation dressed as a global conclusion), in diet and supplement testimonials, and in the confident certainty that pattern-matched intuition sometimes produces even when no real pattern is there.
None of this means personal testimony is the lesser kind of knowing, or that experience should be waved away. Epistemic knowledge starts from testimony — science begins with someone noticing something and wanting to check whether it holds up beyond their own case. The problem isn’t that people trust their experience; it’s that English gives them no built-in way to flag the moment they’ve stopped reporting an experience and started making a claim about how the world works generally.
So here is a habit worth borrowing, even without the Greek vocabulary in the room: before a claim about cause and effect gets asserted as settled, ask which kind of knowing is actually on the table. A useful pair of sentence starters does the job without sounding academic: “What I experienced was…” and “What the evidence shows is…” Used as a genuine two-step rather than a gotcha, this does something both sides of a disagreement usually want: it lets a personal account stand as real and respected on its own terms, while keeping it from being smuggled in, unexamined, as proof of a general fact.
It’s a small linguistic repair for a word English broke a long time ago. But naming the seam is often enough to stop people talking past each other — not because either side is wrong to trust what they’ve seen, but because “I know” was always doing two jobs, and it helps to say which one you mean.