A Kuhn loss is:
a success, empirical or theoretical, of a prior theory – or paradigm as Kuhn would have preferred – that does not carry over to the theory or paradigm that replaced it. [Midwinter and Janssen, see below.]
Kuhn introduced the concept in The Structure of Scientific Revolutions (p. 99–100, 3rd ed.). The examples he gives there come from the phlogiston theory, which explained:
- Why metals are so much more alike than their ores: All metals contain phlogiston.
- Why burning something in a closed flask reduces the pressure of the air: The phlogiston released by combustion “spoils” the elasticity of the air, just as fire “spoils” the elasticity of a steel spring.
I have also run into two examples from Cartesian vortex theory, which explained:
- Why all the planets revolve around the sun in the same direction: They are all carried by a common vortex.
- Why there is gravity: The vortex “pushes” bodies “downward”, due to differential impact of the vortex stream (see Aiton’s book The Vortex Theory of Planetary Motions for more details).
Finally, the article by Midwinter and Janssen, "Kuhn Losses Regained: Van Vleck from Spectra to Susceptibilities" gives a lengthy account of a technical example in the theory of electric and magnetic susceptibilities. Briefly, there is a constant C in that theory. According to classical physics, C = ⅓. The old quantum theory (1913–1925) gave much larger values, the exact value depending on the model and the way quantum conditions were imposed. In the new quantum theory (i.e., modern quantum mechanics), the experimentally correct value of ⅓ is restored.
Since posting this, another near-example has occurred to me: Kepler's 1609 prediction of the sun's rotation, confirmed shortly by the discovery of sunspots. Kepler's prediction stemmed from his theory of a whirlpool-like force from the sun sweeping the planets around; in Newton's system, the rotation of the sun is just an unexplained fact. But this example doesn't quite qualify, since Kepler's whirlpool force was never widely accepted; also, the prediction was not fully successful. (Kepler predicted the wrong period and the wrong axis of rotation.)
I should mention that the concept of Kuhn loss remains controversial, but these examples seem reasonable at first glace. Note that a Kuhn loss can later be regained with a yet newer theory. Electron theory, for example, explains the similarities of metals.
Taxonomy. Kuhn loss comes in several varieties. Both explanations and predictions can be lost. Most examples concern lost explanations. Subsequent history can take several courses. (a) Changing standards of scientific explanation can turn the old issue into a non-problem. For example, Newton's gravity was eventually accepted as fact, not needing a mechanical cause. (b) A yet newer theory can provide a new explanation, as in the metal example. (Here, the electron theory of metals does not refute Lavoisier's "chemical revolution"; on the other hand, the new quantum theory replaced its immediate predecessor. So two versions of this case.) (c) The orphaned phenomenon (deprived of its old explanation) can be treated as an anomaly, something still requiring resolution. Example: geocentric theories explained the lack of stellar parallax, and for almost 300 years, detecting it was an outstanding problem. When Bessel finally detected it in 1838, heliocentricity had long been accepted.
To answer a question in a comment, then:
Doesn't this also mean that they must have been "rediscovered" during a later theory/paradigm shift, since they are defined in terms of previous successes? Or do all things that were believed to be successes in the previous paradigm but later discarded count, even though they were due to e.g. biased observation methods or other misconceptions?
All these count as Kuhn loss. The C-constant example is of type (b); the stellar parallax example was of type (c) until 1838, transitioning then to type (b); the metal example transitioned from type (a) to type (b).
Last clarification: you don't have to accept Kuhn's relativism to be interested in Kuhn loss.
What other examples of Kuhn loss are there?