Volume 11, No. 2—March 2001  
Table of contents for this issue  
 
HYPOTHESES

Eclipse of Reason
The false confirmation of Albert Einstein
By Jim Holt

REVOLUTON IN SCIENCE/NEWTONIAN IDEAS OVERTHROWN announced the front page of the London Times on November 7, 1919. The occasion for this headline was the reported confirmation of Einstein's general theory of relativity. Photographs of a total eclipse taken by a British expedition to the South Atlantic had borne out Einstein's prediction that the gravitational field of the sun bends starlight. The event was seen as a triumph not just for relativity but for science itself. Under closer scrutiny, however, it looks more like a case of high-minded fraud.

Light bending was the most dramatic effect predicted by relativity, and a solar eclipse was the only chance to observe it. During an eclipse, the stars around the sun become visible; if the sun's gravity deflects their light, then their apparent positions will shift a bit. (The shift can be measured by comparing a photograph of the stars during the eclipse with one taken when the sun is elsewhere in the sky.) Einstein had come up with two different figures for the expected shift. In 1911, he calculated that it would be 0.875 arc second (3,600 arc seconds = 1 degree). Four years later, using a more subtle relativistic argument, he doubled the prediction to 1.75 arc seconds.

In fact, the bending of light by gravity was not unique to Einstein's theory. It could also be derived from Newtonian axioms. More than a century earlier, a Bavarian astronomer named Johann Georg von Soldner had done just that. The deflection figure for starlight skimming the sun that Soldner arrived at was 0.875 arc second—precisely the same as Einstein's 1911 calculation. Thus, adjudicating between Newton and Einstein was not a matter of distinguishing no deflection from some deflection; it was a matter of distinguishing a minuscule deflection of 1.75 arc seconds from an even smaller one of 0.875 arc second—tricky to do even in the best of observational circumstances. (Atmospheric turbulence adds to the measurement difficulty, as do the imperfections of the telescope and photographic plates. Clouds can be fatal.)

The 1919 eclipse expedition was led by Sir Arthur Eddington, the Plumian Professor at Cambridge. A Quaker and a pacifist, Eddington had been exempted from military service in World War I because of his important journey. Mindful of how previous efforts to observe the deflection of starlight had been foiled by bad weather, he split the expedition into two parties: One would photograph the eclipse from northeastern Brazil, the other (of which he was a part) from an island off the coast of West Africa. As Clifford M. Will explains in Was Einstein Right? (1993), three possible outcomes were envisaged by Eddington: "No deflection would show that light was not affected by gravity, a half-deflection would confirm Newton, and a full-deflection would confirm Einstein."

The night of the eclipse, May 29, the expedition took three sets of photographs, two from Brazil (using different telescopes) and one from Africa. Of the sets from Brazil, the better of the two yielded a deflection figure of 1.98 arc seconds, well above the Einstein prediction; in fact, the chance of getting such a large measurement by random error if Einstein's theory were true would only be about one in ten. The less good set from Brazil yielded a deflection figure of 0.86 arc second—almost exactly the Newtonian value. The poorest of the three sets of eclipse photographs were those taken by Eddington's team off the coast of Africa, where cloudy conditions resulted in only two usable plates with five blurry dumbbell-shaped star images on each. The deflection calculated from these plates was about 1.61 arc seconds, somewhat below the Einstein prediction.

So, having obtained these equivocal sets of eclipse data, what did Eddington report to the Royal Society five months later? First he threw out the middle-quality set, which favored Newton. Then he averaged the figure from the most accurate set—which by itself cast grave doubt on Einstein's theory—with that from the least accurate set. And, mirabile dictu, the result was a near bull's-eye for general relativity. "It is a gift of fate that I have been allowed to experience this," a grateful Einstein wrote.

Eddington, who had long been Einstein's champion in England, was not unaware of his own bias. He later confessed to a colleague that if it had been up to him he "would not have planned the expeditions since he was fully convinced of the truth of the general theory of relativity." But he had an extra-scientific motive for massaging the eclipse data, and a rather decent one at that. As the philosophers of science John Earman and Clark Glymour have pointed out, Eddington—like Einstein—was a passionate opponent of nationalist prejudice. A British verification of Einstein's theory, he hoped, would help dispel the jingoism of World War I and bring about a rapprochement between British and German scientists.

Although Eddington convinced the world that general relativity was true, it was not until after World War II that astronomers obtained really solid evidence for it. Yet the myth of the 1919 expedition lives on. For example, in Michael White and John Gribbin's Einstein: A Life in Science (1996), we read that the deflections measured by the British team "were exactly in line with the predictions made by Einstein's theory." In science as in war, history is written by the victors.

JIM HOLT

Fill out the form below and receive a free trial issue of Lingua Franca. If you like what you see, you'll pay only $19.95 (55% off the cover price) for a full year!
Name:
Address:
City:
State:
Zip:
Email:

Our monthly dissertation feature for Contentville brings to light the paper trail left by the good and the great, the famous and the infamous.

Learn what you most need to know about most every topic from our regular Barnes & Noble column.

If you have problems accessing or using any area of this site, please contact us at web@linguafranca.com.

Copyright © 2001 Lingua Franca, Inc. All rights reserved.
Sponsored by Seven Bridges Press