Sidebar: Pseudoscience to the Rescue


The use of pseudoscience, that is studies, methodologies, or theories that pretend to use the scientific method and that look like a real scientific effort, but are actually not based on any science at all, is more common than regularly admitted.

Usually the pseudoscience moniker is applied to beliefs like astrology or creationism, or to renegade authors such as Immanuel Velikovsky, the author of Worlds in Collision, or Michael A. Cremo, author of Forbidden Archaeology. But in truth beliefs such as astrology or creationism are not similar to science, nor are authors such as Velikovsky or Cremo scientists. In most cases–notable exceptions are the creationists–there is not even the pretence that these are sciences or these authors scientists. In addition, neither these beliefs nor these authors are published in scientific journals or quoted in scientific studies, nor do they have anything to do with the scientific community.

True pseudoscience are those studies or theories that circulate among the scientific community and are proposed by scientists to the public as real science, but are based on deeply flawed or non-existent methodologies, often for the purpose of promoting a particular belief. Examples of pseudoscience can be found in most fields, but are especially common in social sciences like economics, psychology, and anthropology.

One hundred years ago it was difficult to distinguish between pseudoscience and science, in part because the scientific method itself was under development, but also because of pervasive religious and social prejudices that science had not yet escaped from. Discredited fields like craniometry are now infamous for their role in classifying people. Yet the use of dubious methods, flawed studies, and the suppression of evidence, the hallmark of Aleš Hrdlička’s efforts to try to bolster the Bering Strait Theory, have continued to this very day. An unfortunate example of this has been the promotion by archaeologists and geneticists of Joseph Greenberg’s “Three-Migration hypothesis.”


Greenburg, an influential and pioneering linguist at Stanford University, was well known for his work on language universals and his regrouping of African languages. He had long been impatient over the time and effort it normally took to group language families together; “by comparing languages two at a time and in great depth they will arrive at the true system–in another 50 to 100 years.” Greenberg developed a system called multilateral comparison to try to find relationships between languages faster than the time-consuming and labor-intensive system then largely in use, known as the comparative method.

His system worked reasonably well on African languages, where Greenberg, in the words of linguist Benji Wald, “established order where there was prejudice and chaos, and a grateful set of Africanists adopted his labels, fully aware that they were problematic.” But his attempt in 1971 to reclassify the languages of the South Pacific, known as “The Indo-Pacific hypothesis,” was not successful and was rejected by the linguists who study that region.Greenberg then turned his attention and his methods to the languages of the Americas. As he recognized in an article co-written with his student Merritt Ruhlen in 1992, the great language diversity of the Americas was a problem for the Bering Strait Theory.

The number of [language families] reached about 60 in North America and about 100 in South America, far greater than the number in the Old World, where, for example, Africa has but four. These estimates are puzzling, because taxonomic diversity normally increases with time. Yet most archaeologists have long agreed that human settlement in the Old World substantially predates that in the new. The current consensus is that modern humans . . . did not reach the Americas until 12,000 to 20,000 years ago. How could the American languages have diversified to such a great extent?

Greenberg’s solution to this dilemma was simply to eliminate the diversity. He grouped all of the existing language families in the Americas into just three large family stocks. He also proposed that the first of these stocks (which he called “Amerind”) migrated from Asia 12,000 years ago, exactly the same as the Clovis First Theory proposed, a second (which he called “Na-Dene”) arrived 4,000 years ago or so, and the last (Eskimo-Aleut) in historical times.

Attempts to reclassify Indian languages into super stocks were by no means new, in 1919 Paul Radin proposed that all of the languages in North America could be grouped into two super stocks, but he offered no proof for this and his idea was discarded. In 1921, Edward Sapir grouped the North American languages into 6 super stocks, some of which have since been accepted. But Greenberg’s sweeping new changes, in particular his “Amerind” super-super stock, was almost completely rejected after it was proposed in 1987 in his book, Language in the Americas. As Robert L. Rankin wrote in his review of Greenberg’s work:

The author, a generalist, wishing maximally to clarify vast stretches of history for the non-specialist, has gotten wrong the detail necessary to justify his claims or has used methods that rob him of credibility among his peers. Greenberg’s lack of acceptance of classificatory principles established in historical linguistics over the past 75 to 150 years, plus his cavalier treatment of data and sources and his lack of familiarity with most of the language families with which he deals have produced a deeply flawed book.

Greenberg made it difficult for other linguists to try to duplicate his efforts, an important scientific concept known as “reproducibility,” by not publishing his data, because it “would have added greatly to the length and the cost of the work.”

To make matters worse, Greenberg was sloppy. The Andean specialist Willem Adelaar called Greenberg’s work “riddled with errors” to such an extent that the “number of erroneous forms probably exceeds that of the correct forms.” The influential cognitivist and professor of linguistics at the University of California Santa Barbara, Wallace Chafe, criticized Greenberg’s methodology as “vague” and his book “a random collection of chance resemblances.” Lyle Campbell, author of the standard work in this field, American Indian Languages: the Historical Linguistics of Native America, was withering:

Greenberg compared arbitrary segments of words, equated words with very different meanings, misidentified many languages, failed to analyze the morphology of some words and falsely analyzed that of others, neglected regular sound correspondences, failed to eliminate loanwords, and misrepresented well-established findings.

When linguists did test his system, it did not work. Alexis Manaster Ramer, in a paper in the International Journal of American Linguistics, while “trying to employ Greenberg’s own methodology,” tested Greenberg’s classification of the Tonkawa language as a subset of Hokan and then a subset of Amerind:

Greenberg’s classification of this language should not be accepted. Since this case seems to be one of the most obvious test cases available to us, we end up with a new and rather telling argument against taking for granted the validity of the Greenberg classification of the languages of the Americas.

But an even larger problem, according to Donald A. Ringe, Jr. in “The Mathematics of ‘Amerind,’” was that the methodology was so loose that mathematical tests found:

The similarities Greenberg has adduced as evidence for the genetic unity of 'Amerind' fall within the range to be expected by chance alone, and concluded that Greenberg's method of 'multilateral comparison' is utterly unreliable, as well-informed specialists have long claimed.

Indeed, using his methods, linguists found they could classify Finnish as an American Indian language. Greenberg’s hypothesis simply did not meet the test of a true work of science. As the Concise Encyclopedia of Languages of the World summed it up, “In short, it is with good reason that Amerind has been rejected.”

Yet despite the universal rejection by linguists, paleoanthropologists and geneticists wildly cheered Greenberg’s hypothesis, using it to show that linguistics had finally come around to support the Bering Strait Theory. As William Croft wrote in his obituary of Greenberg: “Another controversial aspect of Greenberg’s Amerind hypothesis was the support it received from physical anthropology and from genetics.”

Over 80% of all genetic studies on American Indian origins have cited Greenberg’s hypothesis since it was published, most of them using his flawed classifications in classifying Indians, and thus leading to skewed genetic reports. The anthropologist E. James Dixon, in his work Bones, Boats & Bison, referenced Greenberg when saying that, “linguistics and biological anthropology, demonstrate that ancestors of living Native Americans most likely came to the Americas from northeastern Asia.” Archaeologist Thomas D. Dillehay, in his book, The Settlement of the Americas, wrote that, “Greenberg’s model has been the most dominant linguistic interpretation of the peopling of the Americas.”

Lauding Greenberg’s theories, The New York Times dubbed him “that rare breed of academic, a synthesizer who derives patterns from the work of many specialists, an exercise the specialists do not always welcome.” The few times that paleoanthropologists did acknowledge that Greenberg’s hypothesis might be controversial, they did so in a way to make it appear as if it was simply academic squabbling. Luca Cavalli-Sforza, a population geneticist at Stanford, said the linguists “have attacked Greenberg cruelly, and I think frankly there is some jealousy behind it.”

But Greenberg himself admitted that the hypothesis was based, not so much on linguistics, but upon “archeological considerations,” in particular, the Clovis First Theory, which he said had “wide acceptance.” Greenberg collaborated with the physical anthropologist Christy G. Turner II (whose study of dental patterns was met with suspicion and is now largely discredited) and the geneticist Stephen L. Zegura (whose early genetic findings have sincebeen overturned), and together they promoted the Three Migration hypothesis, stating that their work independently backed each other up. But many believed they did more than that, as they themselves wrote in 1986 in “The Settlement of the Americas,”

If the investigator in one field is aware of the conclusions proposed in another, he or she may be influenced by this knowledge in developing a theory.

And Greenberg never hid the fact, which was made clear when his data was reviewed, that most of his classification attempts preceded his system, in other words, he used his system to prove a preconceived outcome.

Ives Goddard, the curator and senior linguist emeritus of the Department of Anthropology of the National Museum of Natural History at the Smithsonian Institution, the linguistic editor of the monumental, Handbook of North American Indians, and generally considered to be one of the most prominent figures in the study of historical linguistics, stated in 1994 in “The History and Classification of American Indian Languages,”

We are not aware of a single specialist working on American Indian historical linguistics who thinks that Greenberg has established the validity of his postulated Amerind phylum.

Despite the complete lack of scientific validity, archaeologists and geneticists continued to use Greenberg’s classifications, leading a group of linguists–led by Campbell and Goddard–to write to the American Journal of Human Genetics in 2004 to condemn the widespread use of Greenberg’s work in genetic studies about Indian origins. Yet these objections were ignored.

Greenberg was a devotee of the Bering Strait Theory and he created a body of work to prove it. Although it had no scientific merit and was completely rejected by its own scientific field, his hypothesis was still widely promoted by archaeologists and geneticists because it upheld the prevailing Clovis First dogma. Had it not supported the Bering Strait Theory, it would have been dismissed without controversy like many other theories, including his Indo-Pacific hypothesis twenty-five years previously.

But this was a hypothesis to good to be true, and even knowing full well it was worthless, the archaeologists and geneticists adopted it, much to Greenberg’s delight and to the outrage of most linguists. It was only twenty years later, after Clovis First bit the dust and the Three Migration hypothesis was proven incompatible with genetic evidence that his theory began to be abandoned. But by then the damage to science was done.

As Jason Eshleman, Ripan Malhi, and David Glenn Smith observed in their 2003 article in Evolutionary Biology, that even though Greenberg’s “linguistic divisions themselves have not held up to persistent scrutiny. Nonetheless, the model has strongly influenced designs for research on Native American population genetics.” And has undoubtedly made much of this genetic research useless.

Pseudoscience is used to endorse dogmas and ideologies, not scientific theories. The blatant misuse and promotion of Greenberg’s discredited work by archaeologists and geneticists was pseudoscience in its purest form.