When Many Pages Agree, Check Whether They Are One Page

Repetition feels like corroboration and is usually just circulation.

Our instinct for agreement is old and it does not cope well with copying at scale. When six accounts of an event agree, that historically meant six people saw it. Now it may mean one account was written and five outlets rewrote it, or one summary was produced and everything downstream paraphrased that summary. The agreement is perfect precisely because there is only one observation involved, and the more thoroughly a claim has been copied the more overwhelming the apparent consensus becomes. Volume and independence are entirely different things and only one of them is evidence.

Test for it directly. Do the accounts share unusual phrasing, the same example, the same odd number, the same slightly awkward turn of expression? Independent sources describing the same reality rarely converge on identical wording. When you see wording travelling intact from page to page you are watching a single source propagate, and you should treat the whole cluster as one item rather than many. Then go and find that item, because it is the only thing in the cluster with any evidential weight at all.

This is where the exercise sometimes turns uncomfortable. A claim that is everywhere, in confident terms, occasionally traces back to something surprisingly thin: a passing remark, an estimate offered as a rough illustration, a small piece of work whose limits were clearly stated by the people who did it and stripped away by everybody afterwards. That discovery is not a failure of the search. It is the most valuable possible result, because you now know something about the claim that none of the people repeating it know.