11 January 2008

Indian Population 1492: John D. Daniels

Praise and Expectations

At first glance John D. Daniels, “The Indian Population of North America in 1492,” William and Mary Quartery (April 1992), 298-320, is an impressive article that deserves praise as “[t]he best single review of all the literature on Indian population numbers” (Schweikart and Allen, A Patriot’s History, 9). It offers a staggering list of citations, boiling down many articles and books to three or four sentence summaries of general method to identify “eleven methods,” one “expedient,” and “three schools of thought”. He states, “conflict among them culminated in a major controversy that involved far more than numbers” (298).

These three schools of thought developed chronologically from low to higher estimates. Bottom up dominated 1910-1955 and produced “estimates in the 500,000 to 2,500,000 range” (310). Area modeling “came to prevail from the mid-1950s to the mid-1980s” (312), with estimates in the range “2,500,000 to 7,000,000” (310). The top down school originated with a 1966 article by Henry Dobyns, and offers estimates from “7,000,000 to 18,000,000” (310); it “gradually displaced area modeling, achieving dominance by the mid-1980s” (314).

The most influential members of bottom up school are James Mooney and Alfred Kroeber. They estimate each tribe individually, and add all estimates to achieve their totals. Area modeling is the name employed by Daniels for the work of Sherburne Cook, Harold Driver, and others that develop ecological and cultural models of areas; such factors as population density and resource availability form the base for their estimates. The top down method was initiated by Henry Dobyns who employed epidemic correction, carrying capacity, multiple multiple, and depopulation ratio multiplied by actual census data of nadir population. Although others have arrived at lower estimates, while his continued to climb, many have adopted his methods.

The crux of the problem, as Daniels sees it, stems from “the inadequacy of historical evidence” because reports by colonial Europeans “are invariably defective” (298). In order to compensate for gaps and issues of credibility in the documentary record, sophisticated methods of estimation have been developed. Daniels concludes:

Despite all the discussion of rules of evidence and logic, no one has attempted tribe-by-tribe analysis, taking all direct evidence seriously and using simple inference only when absolutely essential. In order to create a reliable estimate of the North American population in 1492, scholars will either have to find new evidence or develop better methods of handling evidence.
Daniels, “Indian Population,” 320.

From such assertions, we might expect some detailed analysis of representative primary sources to illustrate how they are “defective”. We also might expect further assessment of how particular methodologies have employed such “direct evidence,” and some explanation regarding when inference is essential. It goes without saying that Daniels’ own use of evidence should prove exemplary.


These expectations produce disappointment.

Daniels does not cite or analyze a single primary source. Rather, he assesses the interpretation of primary sources in secondary literature through that literature alone. In contrast, much of the literature reviewed by Daniels works through and argues about a wealth of such sources. Even a look through that literature reveals telling inconsistencies in his elaboration of historiography.

He identifies “report discounting” and “guesstimates” among the methods employed by Mooney, Kroeber, and others of the “bottom up” school. Daniels notes regarding guesstimates, “critics can assess reliability only when the guesser explains the reasoning behind the estimate” (304-305). Report discounting calls for comparable scrutiny, but instead Daniels shifts the emphasis: “Some scholars have scathingly rejected this procedure, defending the value of primary sources and accusing the discounters of cultural and racial bias” (306). Later he highlights cultural concerns:

[Critics of Mooney and Kroeber] alleged that prejudice and a desire to make Indians appear insignificant led to deliberate discounting of evidence of large early native populations. …
Some critics began to wonder if Henry Dobyns and his followers were just as biased in favor of the Indians as Mooney and Kroeber supposedly had been against them.
Daniels, “Indian Population,” 317-318.

He wears the mask of objectivity through careful attribution to others of this binary:

anti-Indian = low estimate / pro-Indian = high estimate

However, he reveals his face by defending Mooney and Kroeber against attacks he feels are unjustified: “since they had spent much of their lives documenting the surviving Indian cultures, they probably would have regarded the charge of cultural bias as absurd” (318). After cultural issues came into academic debates:

The second issue involves the principles of evidence and reasoning. The bottom-up approach is essentially historical; it admits only direct primary evidence and uses only simple deductive logic involving evaluation and comparison of reported data. The school rejects all forms of inference except by simple analogy.
Daniels, “Indian Population,” 318.

Daniels’s Apologia for Discounting

Daniels seizes several opportunities to commend the work of the “bottom up” school because “[t]hey normally used whatever direct historical evidence was available” (316). On the other hand, “area modelers had loosened the rules of evidence and logic” (319). Such “loosening” opened the door for Dobyns’s “radical departure at which the work of others had only hinted” (314).

This “radical departure” was first proposed in “Estimating Aboriginal Population: An Appraisal of Techniques with a New Hemispheric Estimate,” Current Anthropology (Sept. 1966), 395-416. Inasmuch as Dobyns raises significant methodological points against Mooney, Kroeber, and those that popularized their estimates, it is worth looking more closely at his arguments. He mentions as a case in point a figure of eight million Indians perishing in the Caribbean and Peru that was put forth in 1685 by the Marques de Varina to dispute the contentions of others that mining operations were the principle cause of Indian depopulation in Spanish colonies. Dobyns admits that such figures might have been inflated, but cautions against the wholesale discounting of the “bottom up” school.

To make such an assumption does not entitle the analyst to discard those figures; it does require him to seek corroborative or negating evidence. Rosenblat, Kroeber and their imitators have not done so. Their characteristic methodology has included depreciation of all historical population figures. They deprecate the departure of historical witnesses from the "truth" for motives they intuitively impute, but which uniformly led said witnesses to overestimate, in their opinion, aboriginal populations. They ignore the fact that eyewitnesses, whatever their biases, at least observed population trends which the modern analyst can never witness.
Dobyns, “Estimating Aboriginal Population,” 398.

Nowhere in “The Indian Population of North America” does John Daniels discuss how the scholarly methods that he names reveal efforts to corroborate or negate the evidence contained in colonial accounts. Yet, such assessment of historical documents is at the heart of the conception of historiography that he assumes.

New England Case Study

Henry Dobyns asserts in Their Number Become Thinned (1983) and other works that the smallpox epidemic that afflicted Mexico during its conquest by Hernán Cortés may have spread as a continent wide pandemic. He then draws upon a letter from Roger Williams to John Winthrop to offer documentary evidence for sixteenth century epidemics among the Narragansett (Dobyns, 318-319). Dean R. Snow and Kim M. Lanphear dispute his contention in several works, including “European Contact and Indian Depopulation in the Northeast: The Timing of the First Epidemics,” Ethnohistory (Winter 1988), 15-33. They review available Spanish sources, stating that “evidence suggests that epidemics did not spread northward from Mexico” (17). Then they turn toward the Northeast and offer a careful examination of the Williams letter and Dobyns use of it. They find other documents contradict Williams, or at least Dobyns's reading of him.

They contend that the first epidemic to hit New England arrived in 1616. It may have been the bubonic plague, or it may have been yellow fever or trichinosis. In any case, they find wanting the evidence that it reached the interior, although it clearly devastated the coast from Maine to Massachusetts. There is no dispute that an epidemic struck the coast of New England before the English established their colony there. There is considerable debate regarding its extent—epidemic or pandemic—and whether it was the first.

Daniels’ assertion—citing earlier work by Dean R. Snow—thus seems terribly misleading:

[According to some] major epidemics of Old World diseases preceded European explorers by a few years or decades and killed a significant percentage of the native population. … Some authorities maintain, for example, that a serious epidemic hit the Northeast before the first English settlers arrived.
Daniels, “Indian Population,” 306.

Although it is not cited by Snow and Lanphear, Mourt's Relation corroborates their recognition that an epidemic swept the coast of New England in 1616. The document was written principally by Edward Winslow and published in London in 1622. It is one of the key primary sources for the history of the Pilgrims in what later came to be called New England. The entire text with modernized spellings is accessible online at The Plymouth Colony Archive Project.

Two excerpts:

So we light on a path, but saw no house, and followed a great way into the woods. At length we found where corn had been set, but not that year. Anon we found a great burying place, one part whereof was encompassed with a large palisade, like a churchyard, with young spires for or five yards long, set as close one by another as they could, two or three feet in the ground. Within it was full of graves, some bigger and some less; some were also paled about, and others had like an Indian house made over them, but not matted. Those graves were more sumptuous than those at Cornhill, yet we digged none of them up, but only viewed them, and went our way. Without the palisade were graves also, but not so costly. From this place we went and found more corn-ground, but not of this year. As we ranged we light on four or five Indian houses, which had been lately dwelt in, but they were uncovered, and had no mats about them, else they were like those we found at Cornhill but had not been so lately dwelt in. There was nothing left but two or three pieces of old mats, and a little sedge. Also, a little further we found two baskets full of parched acorns hid in the ground, which we supposed had been corn when we began to dig the same; we cast earth thereon again and went our way. All this while we saw no people.
Mourt’s Relation

About noon we met again about our public business, but we had scarce been an hour together, but Samoset came again, and Tisquantum, the only native of Patuxet, where we now inhabit, who was one of the twenty captives that by Hunt were carried away, and had been in England.
Mourt’s Relation

The Pilgrim records are clear that Tisquantum (Squanto) was the only survivor of his home village and that abandoned villages were present throughout the region. Although there are several possible explanations for abandoned villages, disease was certainly one cause prior to English settlement in the Northeast.


Daniels’ seminal article is nearly comprehensive, but terribly superficial in its review. Its convoluted scheme to organize a mass of scholarship gets cited, but not embraced in subsequent scholarship. His footnotes offer a useful point of entry into the field of pre-Columbian demography, but an alphabetized bibliography might serve just as well. His analysis offers a few useful questions, but no credible answers. The entire article seems to be a thinly disguised defense of the lowest estimates—the lack of credibility of which is the starting point for all scholarship in the field.

A far better review of the literature is available in Russell Thornton, “Aboriginal North American Population and Rates, ca. A.D. 1500-1900,” Current Anthropology (April 1997), 310-315. Thornton’s article, as Daniels, is not without an agenda. Thornton’s concern emphasizes the interplay of disease with other aspects of colonialism to produce long-term population decline for the indigenous population of the Americas from initial contact through the nineteenth century. His 1997 summary is broader in scope and less comprehensive than Daniels, but its summary of the themes and contentious issues in the scholarship presents a far more credible overview. The failure of Schweikart and Allen, A Patriot’s History, to mention Thornton while they praise Daniels is but one indication that their claims rest upon shoddy scholarship.

Daniels’s article might appear to be the source for the statement: “One recent scholar, examining the major assessments of the numbers, points to at least nine different measurement methods, including the time-worn favorite, guesstimates” (Schweikart and Allen, 8). This sentence neither rests upon Daniels’ identification of eleven methods, nor upon his observation that count multiple “has been used more than any other” (305), but it otherwise bears a strong similarity to his thesis. In any case, we can apply Daniels’s scheme to some of the numbers in A Patriot’s History. Daniels explains that the method he calls depopulation ratio (DR) assumes the significance of disease, so scholars start from a nadir population that is usually reasonably certain; then they multiply. Ratios have ranged “between 5:1 and 25:1 and nadirs for North America from 1870 to 1930” (308).

Schweikart and Allen both assert and conclude a depopulation ratio (DR) near 1:2. They state that 56 million deaths, as some have alleged, requires an aboriginal population of 100 million (56% depopulation). Then by rounding Ubelaker’s 1,894,350 down to 1.8 million, they state 800,000 died (50% depopulation). I’ve already noted in “Depopulation: Ubelaker’s Low Estimate” that Ubelaker’s own figures include nearly 1.4 million dead (72% depopulation). It would seem that Schweikart and Allen have reversed Dobyns’s most controversial process, applying a ratio of 1:2 to the lowest available estimate after rounding it down. Such numbers are dishonest.


Greg Stripes said...

Has there been any research in how the introduction of horses, and their adoption by some tribes may have further the spread of disease, in that they may have increased contact between native peoples?

James Stripes said...

Thanks for the good question, Greg. Maybe. I'll let you know when I find it. Henry Dobyns has an article that I have in my "to read" stack regarding transmission of disease at trading centers. Horses would have brought folks to trading centers from further away.

It seems commonsense that an increase in mobility would increase the spread of illness, except when the carrier is rendered immobile by the disease, as is generally the case with smallpox.

On the other hand, many studies highlight ways that large urban centers were particularly prone to the spread of some illnesses--influenza for instance. I'll be saying more about some of these studies in the coming weeks.

Steve Sailer said...

Thank you for this most informative summary.

  © Blogger templates The Professional Template by 2008

Back to TOP