Applying Mill's Methods
This video presents the same text shown beside it, spoken and on screen. It adds nothing the text does not say.
Applying Mill's methods means narrowing a field of candidate causes by agreement, eliminating survivors by difference, then corroborating what remains by concomitant variation.
Six diners, four dishes, two of them ill afterwards. Build the table. The two who fell ill both ate the salad and the fish; one had the soup, the other did not; neither had cake. Agreement looks for what every case of the effect shares, which leaves two candidates — salad and fish. Now difference. Find a pair alike in everything except one factor. One well diner ate the salad and skipped the fish, and stayed well. That single difference eliminates the salad and leaves the fish. Now concomitant variation. Of the two who fell ill, the one who took two servings of fish was severely ill and the one who took a single serving was mildly ill. Effect tracks dose, which corroborates rather than proves. State the verdict as the methods actually license it: fish is the leading candidate cause, supported by three converging lines of evidence, not established beyond question.
The methods eliminate candidates that were put on the list; they cannot rank one that never made it. A contaminated shared water glass appears in no column and survives every test by being absent from the table, which is why the verdict names a candidate and reports a confidence rather than closing the case.
Builds on
Unlocks
- Nothing yet depends on this.