2.  Albert Einstein’s Conjecture  Sir Arthur Eddington’s Test  Popper’s Account

4.  In 1915, Albert Einstein publishes his General Theory of Relativity (GTR).  His Special Theory of Relativity (STR) had already been published in 1905.

5.  STR is based on two postulates. (Roughly…)  The laws of physics are invariant with respect to frame of reference.  The speed of light is constant in all inertial frames of reference.

6.  STR was special—it did not address gravity.  GTR was intended to unite STR with Newton’s law of universal gravitation.

7.  GTR, like STR, was considered philosophically compelling because it unified separate theories and observations in a simple, elegant way.  But neither had been empirically tested.

8.  Many people thought GTR (which included the idea that “time is relative”) was a wild idea, on par with Marxism and Freudianism.  But the distinction, as Karl Popper sees it, is that GTR is falsifiable, for GTR predicted that the path of light would be measurably curved by massive objects.

9.  Another way of looking at it: GTR was informative in the sense that it ruled out many possible observations, saying in effect that we will never observe certain things x, y, z.  GTR is thus a bold conjecture since if we ever do observe x, y, or z, then we can refute it.  As Godfrey-Smith and others have put it, GTR “sticks its neck out”.  GTR is precise. It is easier to refute a precise theory than a vague one.

11.  GTR predicts that light paths (or the space through which light propagates) are curved by massive objects.  In those days, we had no easy way to test this.  Predicted curvature in most cases was too small to measure.  And in the one case in which it would be large enough to measure, there was a problem.

12.  The Sun is large enough to measurably curve light from distant stars.  However, the Sun is also very bright. We cannot normally see or otherwise detect stars in daylight.

13.  The one thing massive and local enough to test GTR is also the thing that keeps us from testing it!  If only we could blot out the sun’s light in a way that allowed us to look at distant stars which appeared to be next to it.

14.  The one thing massive and local enough to test GTR is also the thing that keeps us from testing it!  If only we could blot out the sun’s light in a way that allowed us to look at distant stars which appeared to be next to it.  Fortunately, we have a moon. And it sometimes passes in front of the sun, making it dark enough to detect the starlight coming from distant stars.

15.  GTR is falsifiable because it predicts that during an eclipse, the paths of starlight which pass near the Sun will be detectably curved.  Thus it will appear to us that the stars are in different locations.

16.  GTR is falsifiable because it predicts that during an eclipse, the paths of starlight which pass near the Sun will be detectably curved.  Thus it will appear to us that the stars are in different locations.  If this does not occur, GTR is refuted.

19.  In 1919, Eddington goes to the island of Principe to critically test GTR by measuring the apparent positions of stars during the eclipse…

20.  GTR passes the test!

22.  GTR passes the test.  Does this mean GTR was confirmed?

23.  GTR passes the test.  Does this mean GTR was confirmed?  Not according to Karl Popper.  GTR survived an attempted refutation, which means it might be true.

24.  Indeed, any theory which predicted the stars would appear just where they were was falsified.  But this does not mean we have confirmed GTR.

25.  GTR has survived an attempted refutation.

26.  Its distinction lies in its being falsifiable, bold/risky, having been critically tested in an attempted refutation, and having survived that attempted refutation.