Volume 10, No. 2 - March 2000
More in this Issue

"Who is buried in Grant's tomb?" That was the bonus question Groucho Marx used to put to unfortunate contestants on his 1950s quiz show, You Bet Your Life. It sounds like a giveaway, but beware: Questions of this form can be treacherous. Consider: Who discovered Bayes's theorem? Who discovered Giffen's paradox? Who discovered the Pythagorean theorem? Who discovered America? If your answers were, respectively, Bayes, Giffen, Pythagoras, and Amerigo Vespucci, then no box of Snickers for you.

The practice of naming things after people (real or mythical) who are associated with them is called eponymy. There are eponymous words, like "guillotine," "bowdlerize," and "sadism." There are eponymous place-names, like Pennsylvania and the PelopÛnnisos. And there are eponymous compound expressions, like "Copernican system" and "Halley's comet." When such expressions occur in the sciences, the presumption is that the thing designated was discovered by the scientist whose name is affixed to it. That presumption is nearly always false.

If you think I exaggerate, you are obviously not familiar with Stigler's Law of Eponymy. This law, which in its simplest form states that "no scientific discovery is named after its original discoverer," was so dubbed by Stephen Stigler in his recent book Statistics on the Table (Harvard). An immodest act of nomenclature? Not really. If Stigler's law is true, its very name implies that Stigler himself did not discover it. By explaining that the credit belongs instead to the great sociologist of science Robert K. Merton, Stigler not only wins marks for humility; he makes the law to which he has lent his name self-confirming.

What explains the truth of Stigler's law? One might start with Merton's famous hypothesis that "all scientific discoveries are in principle 'multiples.'" Perhaps, for some reason, a discovery invariably gets named for the wrong one of its multiple discoverers.

But Stigler's law is more interesting than that. Take the Pythagorean theorem. Pythagoras was not one of its discoverers. The theorem was known before him and proved after him; moreover, he may even have been unaware of its geometric import. Such radical misnamings abound. In checking out Stigler's suggestion that Giffen's paradox ("demand for some goods increases with price") was undreamed of by its eponym, the economist Robert Giffen, I chanced upon an encyclopedia entry for Sir Thomas Gresham, the sixteenth-century Englishman after whom Gresham's law ("bad money drives out good") is named. "It was thought that Gresham was the first to state the principle," the entry reads, "but it has been shown that it was stated long before his time and that he did not even formulate it."

Such eponymic blunders might be the exception rather than the rule if historians of science were in charge of labeling scientific discoveries. But they are not; it's practicing scientists who make the decisions, and for all their vaunted vigor, most of them have no historical expertise. And as Stigler observes, "names are rarely given and never generally accepted unless the namer...is remote in time or place (or both) from the scientist being honored." That is to ensure the appearance of impartiality. After all, having a theorem or a comet named for you confers something like intellectual immortality; such an honor must be perceived by the community of scientists as being based on merit, not on national affiliation or personal friendship or political pressures.

Given that "eponyms are only awarded after long time lags or at great distances, and then only by active (and frequently not historically well informed) scientists with more interest in recognizing general merit than an isolated achievement," Stigler concludes, "it should not then come as a surprise that most eponyms are inaccurately assigned, and it is even possible (as I have boldly claimed) that all widely accepted eponyms are, strictly speaking, wrong."

He proceeds to show the great power of his eponymous law by applying it to a particular case: the formula for the probability distribution known as the bell curve. Since this is also called Gaussian distribution, one can infer from Stigler's law that Gauss was not its discoverer. Sure enough, in an 1809 book, Gauss cites Laplace in connection with the distribution, and in fact Laplace did touch on it as early as 1774. But the distribution is also sometimes called the Laplace or Laplace-Gauss distribution, so it can further be inferred from Stigler's law that Laplace was not its discoverer either! Indeed, current scholarship traces its origin to a 1733 publication by Abraham De Moivre.

Oddly, I have found that Stigler's law is valid even for pseudo-eponyms. Take "crap." People often claim that this word eponymously derives from Thomas Crapper, the celebrated Victorian inventor of the flush toilet. But this etymology is spurious: The word "crap" in its excremental sense entered Middle English from Old French. Nevertheless, the mere fact that Crapper is folkishly linked to "crap" suggests, by Stigler's law, that he was not the original inventor of the flush toilet. And, mirabile dictu, that turns out to be the case: The flush toilet was designed by Sir John Harington in the court of Elizabeth I.

I could give more examples, but lunchtime approaches, and I am looking forward to eating something that I feel quite certain was not invented by the fourth earl of Sandwich.

In addition to "Hypotheses," Jim Holt writes a regular column for the Wall Street Journal covering books on science and philosophy. His book Worlds Within Worlds: How the Infitesimal Revolutionized Thought will be published in the fall by Four Walls/Eight Windows.