“The Associations of Normal Subjects” and “The Association Method,” Collected Works 2 (Experimental Researches), the latter at p. 439 — though note the test failed as forensic lie detection and its modern descendants have weak individual-level predictive validity, so treat it as a lens on yourself rather than an instrument for reading anyone else.
Before the archetypes, Jung ran a stopwatch. In the word-association work he and Riklin published from 1904, the method was blunt: read a stimulus word, start a one-fifth-second stopwatch on its accented syllable, stop it when the subject speaks, write the number down. The payoff case is a hospital theft, where Jung mixed about twenty-three words drawn from the specific details of the crime — cupboard, key, banknote, fur, to hide — among twice as many indifferent ones, and then compared each nurse’s mean latency on critical versus indifferent words. Nurse A came in at 10.0 on indifferent words and 16.0 on critical ones (the figures are in fifths of a second, so 2.0s against 3.2s); the other two moved by 1 and 3. A also led on false reproductions in the follow-up memory test, 19 against 9 and 12. Jung wrote that “the greatest suspicion falls on A,” and A confessed that evening. The usable idea is not lie detection — it is that the material you have not metabolized announces itself as latency, and hesitation is a measurement you can take on yourself.
TRY THIS TODAY Pick the ten-minute window after your next piece of real feedback — a code review, a product decision that went against you, a clinician call that did not go well. Write the six words that actually appeared in it. Then, next to each, write your first reaction without editing it, and mark the ones where you paused, re-read the word, or produced a sentence about someone else instead of an answer. Those marks are the list. You are not looking for what you think about the feedback; you are looking for where the needle stuck.