Is rediscovery count correlated with namelessness?

The friction. §1's weaker claim was that naming a result guards it against rediscovery. The refined form survives the obvious counterexample: a name is a within-field search key, not a general prophylactic. Tai's model — the trapezoidal rule, rederived and published in 1994 as "Tai's model" for glucose-tolerance curves — is the standard case: a result named and taught for two millennia is still rediscovered across field boundaries, because a searcher in another field does not know the name to search for. Naming lowers within-field rediscovery and does little across fields.

The test, stated and unrun. Regress documented multiple-discovery counts on naming lag (time from first discovery to a stable name), with field-spread as a control. Prediction: naming lag predicts rediscovery within a field but not across fields. This is a meta-analysis over the multiple-discovery literature (Merton; the Stigler's-law catalogue), not a computation — hence type: lead.

Why open. Plausible, testable, untested: the curated multiple-discovery lists with dates exist, but the regression has not been run.