The Half-Life of Facts

by Samuel Arbesman

  So Granovetter wasn’t quite right. Ultimately it’s the medium-strength ties that are the most important. They are that happy medium between ties that are too weak to spread anything and those too strong to be found in anything but socially (and informationally) inbred groups.

  These are the types of ties that allow knowledge to spread, facts to disseminate, sometimes even errors to propagate. The people you trust a little bit but aren’t your closest friends, your work friends, or something a bit more than strangers but less than a good buddy: These people provide the ties that are the most important in allowing something to spread far and wide. They connect different enough social circles that the fact can infect a new group, but they are also strong enough to provide a good chance of spreading it.

  We see hints of this when it comes to spreading the printing press and Gutenberg’s bundle of innovations. The key individuals in spreading new facts were other Germans in different towns. These individuals, who had the necessary knowledge and skills (and often some sort of connection to Gutenberg), were able to bring these innovations to new communities.

  How facts spread and reach each of us is intimately tied to how we are connected to one another, and network science can provide us with a guide to understanding how this works. But just because a fact spreads doesn’t necessarily mean that it’s right. Just as quickly as truth can spread, so can facts that are wrong.

  Errors are especially pernicious facts. Is there anything that science can say about how errors spread and persist in a population? To understand that we have to look at some of these errors, starting with one that I fell victim to many years ago: the celebrated case of the brontosaurus.

  • • •

  AS a child, I was a student of dinosaurs. When you’re a six-year-old boy, this is not really a choice—it’s some sort of biological imperative. I was well versed in the different types of sauropods and theropods, the dietary habits of these ancient giants, and even the recently popularized theory that they were warm-blooded rather than the slow, dim-witted, and cold-blooded dinosaurs known to previous generations.

  We had a wonderful activity in my kindergarten class in which each of the eighth-grade students were paired with a kindergartner. They were assigned to interview us, to discuss our interests, hopes, and dreams, and then to write a storybook for us based on all of these findings. The subject matter for my book was never in any doubt. Dr. Sam Arbesman, paleontologist, was given the honor of being sent to South America to investigate rumors that a live dinosaur was there, and to capture it.

  But despite this obsession with dinosaurs, there was one fact that I got wrong. It’s entirely basic, and yet no one had told me that I was incorrect: the name brontosaurus. The four-legged saurischian, with the long neck and tiny head, is one of the iconic creatures of my youth. In fact, it was the very species of dinosaur that my fictional self was sent to South America to capture. And yet its name is actually apatosaurus.

  • • •

  BY the 1860s, dinosaur-fossil hunting was in full swing. Darwin had published On the Origin of Species the previous decade, and decades before that the argument that these fossils were the bones of creatures that perished in the deluge had been discarded. They were now bona fide monsters that had lived unbelievably long ago, and they were part of the pageantry of interconnected life explained by evolution.

  Bestriding this wave of scientific discoveries were two American paleontologists: Edward Cope and Othniel Marsh. While entwined by history, they were highly unlike each other. Cope had little formal training in paleontology, and only received an honorary master’s at the same time he began a position at Haverford College. Marsh, on the other hand, had a doctorate from the University of Heidelberg, was a professor at Yale University, and was the curator of the Peabody Museum of Natural History.

  Beginning in 1863, Marsh and Cope began something of a competition, though that term is not quite accurate. The phrase used by some—the Great American Dinosaur Rush—also doesn’t do it justice. This was no friendly rivalry over who might discover more new species. If you consider the depths of bribery, theft, and even ideological dispute to which both paleontologists sank, it is certainly more appropriate to call this conflict by its more common name: the Bone Wars.

  Marsh and Cope were initially collaborators, but they soon had a falling-out. The proximate cause seems to be Marsh’s pointing out an error in the reconstruction of a swimming creature known as the elasmosaurus, a sort of Loch Ness monster. This, coupled with Marsh’s payment—behind Cope’s back—of diggers to divert all future fossil finds in their New Jersey area to him, cemented the beginnings of the war. The conflict was exacerbated by the fact that Marsh was a staunch Darwinist, while Cope adhered to an older theory known as neo-Lamarckism. Eventually the two became fierce rivals, resorting to nearly any stratagem in order to describe more dinosaur species.

  In the midst of all of this, Othniel Marsh published his discovery of the brontosaurus in 1879. Its full taxonomic name is Brontosaurus excelsus, essentially meaning in Latin “most sublime thunder lizard.” Two years earlier, in 1877, he also submitted a paper entitled “Notice of New Dinosaurian Reptiles from the Jurassic Formation,” in which he described a slightly smaller dinosaur (it was clearly a juvenile, or child, dino) that he called apatosaurus. This name means “deceptive lizard,” due to Marsh’s observation that its bones looked similar to those of another species. Marsh, in these descriptions, even managed to get a dig in against Cope, noting that his findings related to these species were superior to Cope’s, whose “[c]onclusions based on such work8 will naturally be received with distrust by anatomists.”

  The brontosaurus discovery went on to be supplemented with a complete skeleton, beautiful to behold and the harbinger of its fame in popular culture. The apatosaurus, on the other hand, languished as a tiny collection of bones that included not much more than a pelvic bone and a shoulder blade.

  Marsh and Cope continued their respective breakneck paces of dinosaur discovery, lashing out at each other over and over. Cope even went so far as to purchase a controlling interest in the distinguished scientific journal American Naturalist in order to make it easier for him to publish his discoveries. Despite their vitriol and animosity, they actually didn’t fight any more about the brontosaurus.

  But in 1903, an error was found by the paleontologist Elmer Riggs. This time it was Marsh who had gotten something wrong. While Cope didn’t have the satisfaction of knowing this (both had already died several years earlier), it was a rather large error. Riggs argued that the brontosaurus was in fact simply a version of the apatosaurus. Due to the error, the brontosaurus no longer formally existed. Since it had been discovered after the apatosaurus, the apatosaurus name received precedence. And while the name brontosaurus was much more impressive, the title apatosaurus was now the correct one due to being first.

  Despite this problem, many paleontologists persisted in using brontosaurus. Why confuse the public when it was a rather minor issue?

  Then, in 1978, two other paleontologists, J. S. McIntosh and David Berman, noticed something even bigger that was amiss: The original brontosaurus had been graced with the wrong head! It had the head of another large, plant-eating dinosaur. They recognized that a skull, misidentified as belonging to a different species, in fact belonged to the apatosaurus. After this discovery was made they realized it was time to set this error straight; scientists began to agitate for a switch to the name apatosaurus.

  But this did nothing to diminish the appeal of the name brontosaurus. This dinosaur was already out of the bag. The brontosaurus was featured in popular books of all types, including those from which I gained much of my childhood dinosaur expertise. The United States Postal Service even included it as one of four dinosaur stamps in 1989, nearly a century after the discovery of the misclassification and two decades after the beginning of the discontent in the paleontological community.
I only learned of the misclassification in the early 1990s, through Stephen Jay Gould’s essay “Bully for Brontosaurus,” when he argued that the postal service did the right thing, even though the name was technically incorrect.

  Since then, apatosaurus has been gaining currency, although rather slowly, as seen here in a Google Ngram:

  But is this always how erroneous facts persist? Sadly, it seems that this is often the case, that there are many examples of errors that have stuck around for far longer than they should have.

  Figure 7. The number of uses of word brontosaurus (black) versus apatosaurus (gray) over time. Data courtesy of Google Books Ngrams and the Cultural Observatory.

  IN chapter 9 we’ll examine why people refuse to change their knowledge, or at least neglect to update their mental databases of facts. But how does a novel fact, even a wrong one, spread and persist in the population? Are there regularities to how errors spread?

  One of the strangest examples of the spread of error is related to Popeye the Sailor. Popeye, with his odd accent and improbable forearms, used spinach to great effect, a sort of anti-Kryptonite. It gave him his strength, and perhaps his distinctive speaking style. But why did Popeye eat so much spinach? What was the reason for his obsession with such a strange food?

  The truth begins more than fifty years earlier. Back in 1870, Erich von Wolf, a German chemist, examined the amount of iron within spinach, among many other green vegetables. In recording his findings, von Wolf accidentally misplaced a decimal point when transcribing data from his notebook, changing the iron content in spinach by an order of magnitude. While there are actually only 3.5 milligrams of iron in a 100-gram serving of spinach, the accepted fact became 35 milligrams. To put this in perspective, if the calculation were correct each 100-gram serving would be like eating a small piece of a paper clip.

  Once this incorrect number was printed, spinach’s nutritional value became legendary. So when Popeye was created, studio executives recommended he eat spinach for his strength, due to its vaunted health properties. Apparently Popeye helped increase American consumption of spinach by a third!

  This error was eventually corrected in 1937, when someone rechecked the numbers. But the damage had been done. It spread and spread, and only recently has gone by the wayside, no doubt helped by Popeye’s relative obscurity today. But the error was so widespread9 that the British Medical Journal published an article discussing this spinach incident in 1981, trying its best to finally debunk the issue.

  Ultimately, the reason these errors spread is because it’s a lot easier to spread the first thing you find, or the fact that sounds correct, than to delve deeply into the literature in search of the correct fact.

  Michael Mauboussin, the chief investment strategist of Legg Mason Global Asset Management, in an article about fact-checking, relates his own experience with this sort of error propagation. While working on his book10 Think Twice, he came across an equation in a book about statistics that calculated the value of wines from the Bordeaux region. The problem was, when Mauboussin tried it, it didn’t work. It turned out that his source was riddled with errors, ranging from one number that was ten times too small to another that contained a rounding error, completely changing the meaning of the equation. Only when Mauboussin tracked down the original scientific paper did he find the correct version.

  There are many examples where a small error, despite being corrected later, has spread through a population. If you want to spend days poring over persistent errors that have spread far and wide, Snopes.com is a great font for these bits of information. Or even look at Wikipedia. In his delightfully nerdy Web comic xkcd, author Randall Munroe wishes for a world11 in which schoolchildren read the Wikipedia page on common misconceptions weekly, in order to learn truth as well as skepticism. Both of these sites are full of urban legends, false facts, and misconceptions that have become prevalent.

  One good rule of thumb when examining how errors propagate over time is to look for a simple phrase: contrary to popular belief. While the phrase is a favorite of writers with a love for the counterintuitive point (and this author is not immune to this), it’s also a clear indication that a bit of knowledge has spread far and wide despite being inaccurate. The antidote to this false fact, which of course the writer is about to tell you, has yet to penetrate the popular consciousness. And this phrase is by no means new. I have found instances of it12 in books and magazines from the nineteenth century debunking false facts about lunar phases, medical knowledge, and even the heredity of genius.

  There are even examples where such misinformation has been spread purposefully, albeit sometimes with a wink rather than with malicious intent. But since we often don’t track our sources, this can have a rather problematic effect. For example: I have a book on my shelf entitled Dictionary of Theories. Perusing it one day, I came across a curious entry:

  Dynamics of an asteroid13 (1809) Astronomy Initiated by C F Gauss (1777–1855), but reputed to have had its outstanding exposition in an elusive textbook by James Moriarty (c. 1840–1891), with later contributions by other mathematicians including K Weierstrass (1815–1897) and J E Littlewood (1885–1977).

  The motion of an asteroid, which is now generally understood as a minor planet, is that of a body of negligible gravitational attraction in the gravitational field of two massive bodies, just like a spacecraft under the influence of the Earth and the Moon.

  J F Bowers, “James Moriarty: A Forgotten Mathematician,” New Scientist, 124 (1989). Parts 1696–7, 17–19.

  While I had never heard of this theory, I wasn’t terribly surprised, as the dictionary contains many obscure ideas. But I did recognize the reference, and that surprised me. That’s because the citation was to a monograph by none other than James Moriarty, the arch nemesis of Sherlock Holmes.

  As brilliant as he might have been, Moriarty never existed. Yet here he was, in a fictional reference that had somehow jumped into the real world. Searching a bit further, I tracked down the article referenced in14 New Scientist, and discovered that it was a tongue-in-cheek analysis, by John Bowers of the School of Mathematics at the University of Leeds, of the mathematical contributions of Professor Moriarty, including his analysis of how gravity operates on asteroids.

  But the citations to Moriarty’s work15 didn’t end there. I even found a dissertation by Kristian Kennaway, a physics doctoral student at the University of Southern California, that cites one of the other celebrated bits of research by Moriarty: a treatise on the binomial theorem, published in none other than the Bohemian Journal of Counting.

  While this is no doubt a fun bit that Kennaway inserted to see if his committee members were paying attention (my aunt placed a banana bread recipe into her master’s thesis and no one noticed), Kennaway actually cites Moriarty’s mathematical work to explain something, providing the justification for a mathematical concept using the work of a fictional character.

  Thus far, I haven’t seen many other examples of this, and I doubt such bizarre overlaps of fiction with reality have propagated far. But it does give one pause.

  Bad information can spread fast. And a first-mover advantage in information often has a pernicious effect. Whatever fact first appears in print, whether true or not, is very difficult to dislodge. Sara Lippincott, a former fact-checker for The New Yorker, has made this explicit. These errors “will live on and on, . . .16 deceiving researcher after researcher through the ages, all of whom will make new errors on the strength of the original errors, and so on into an explosion of errata.” This is strong stuff. These errors become ever present and extremely difficult to correct. It’s like trying to gather dandelion seeds once they have been blown to the wind.

  I myself was a victim when I actually propagated the myth that a frog, if boiled slowly, will not jump out of a pot. I mentioned this in passing in the Boston Globe, using it to explain how people don’t notice factual change if
it happens slowly. I was taken to task soon after by James Fallows,17 of The Atlantic, who has worked hard to remove this falsehood from the population; in fact, the frog only remains in the pot if it’s brain-dead.

  Can we understand in any rigorous way how these sorts of falsehoods continue to propagate? Happily, there is scientific research that delves into how they spread. But that science requires us first to take a little detour to examine some very old typos in ancient manuscripts, their surprising relationship to genetics, and how both of these fields deal with error.

  • • •

  WE can look to the children’s game of telephone to understand how facts can be corrupted and spread: The children sit in a circle, and one person begins by whispering a phrase or sentence to the child next to them. This person whispers to their neighbor, who in turn does the same, continuing until the person who completes the circle, the last one to hear the sentence, says aloud what they heard. This is then compared to what the first person initially said, often with hilarious results. Of course, sometimes this is because there’s someone malicious somewhere along the line—the kid who delights in replacing every verb with fart, for example. But in general, the sentence decays without any malice or intent. It simply gets changed because hearing a whispered sentence doesn’t provide great fidelity. It’s what information scientists would refer to as a noisy channel. When information is passed from one person to another it has the potential to become inaccurate unless there are a whole host of error-checking mechanisms.

  A clear case of this that we can actually measure and study quantitatively can be found in the world of old texts. Surprisingly, understanding the errors in these manuscripts is actually quite similar to understanding genetics. This may sound a bit odd. What do handwritten manuscripts from the medieval period or earlier have to do with genetics? On the surface, nothing: One is a distinguished part of the humanities and the other a hard experimental science. However, while those who study each of these fields have very little to do with one another, it turns out that there is a great deal of symmetry. It mainly comes down to mutation.