The Identification Decision Tree

Identification is best approached as a structured sequence of narrowing questions rather than a single lookup, because most pieces present several independent lines of evidence — movement, case, dial, markings — that need to be weighed together rather than any one being treated as automatically decisive. A workable decision tree starts broad and narrows: first, what general category and era is this (pocket watch, wristwatch, or clock; roughly what period, judged from case style, escapement type and construction)? Second, what does the movement itself state directly — a signature, a calibre or serial number, jewel count, adjustment markings?

Third, does the case corroborate or contradict the movement's evidence — do hallmarks (if present) date to a period consistent with the movement, does the case type match what that maker or grade normally used, and are there signs of a case swap such as mismatched winding-stem alignment or a seconds bit in the wrong position for the case style? Fourth, what do published references (calibre databases, serial-number registers where a manufacturer's records survive, marks and makers' directories) say about the specific serial range, calibre, or signature observed, and does more than one independent reference agree?

Distinguishing supporting evidence from merely consistent evidence is the crux of sound identification reasoning. A serial number falling within a documented range a manufacturer assigned to a specific model is supporting evidence for that identification. A case style that was 'commonly used' across many makers and grades in a given period is, at best, consistent with an attribution and does nothing to distinguish it from other equally plausible attributions — treating merely consistent evidence as if it were confirming evidence is one of the most common reasoning errors in amateur identification work, and is worth naming explicitly so it can be caught.

Good identification practice also means actively looking for disconfirming evidence, not just accumulating support: if a proposed attribution requires ignoring a marking, a case feature, or a serial number that does not fit, that is a signal to reconsider the attribution rather than to discount the inconvenient detail. Multiple independent lines of evidence pointing the same way (movement signature, serial range, case style and known production history all agreeing) constitute a genuinely strong identification; a single piece of evidence, however striking, generally does not.

Finally, and importantly, a rigorous identification process sometimes correctly concludes 'not confidently identifiable from available evidence', and that outcome should be recorded and stated plainly rather than resolved by picking the most attractive available guess. This academy's own editorial statuses — verified, partially verified, needs research, historical attribution, estimated — exist for exactly this reason, and applying that same honest calibration to a personal identification effort is the single most important habit this lesson teaches.