As day one of the First International Conference on Quantitative History drew to a close, the chairman, Konrad H. Jarausch of the University of Missouri, whispered an apology. He was sorry the discussion had wandered so far from the agenda, Jarausch said, but since the 50 conferees were "grown-ups," they had the right to talk about anything they wanted.
What they had wanted to talk about, much of the time, were all the criticisms lodged against them recently by their fellow historians. Harvard's Bernard Bailyn, in his December address to the American Historical Association, had questioned the value of "vision-limiting, technical problem-solving." Princeton's Lawrence K. Stone, himself a pioneer in the use of quantitative techniques, had called for a retreat from the computer toward methods "which will shed more light with less trouble." Others had railed against "quanto-maniacs" and "statistical junkies."
Cal Tech's Morgan Kousser urged the conference not to treat the challenge lightly. "Historians," he warned, "are becoming--to paraphrase John Dos Passos--two nations."
If so, the nation that met at the Smithsonian Castle last week is certainly the less well-known of the two. Not all the practitioners of quantitative history even agree that it should be viewed as a separate domain. But they do agree that, beginning in the early '60s, the invention of the computer led to a profound revision of traditional ideas about immigration, slavery, dissident movements and a host of other stimulating historical subjects. And they agree that the profession as a whole--including, to quote Kousser, "some people with substantial investments in history as it used to be done"--remains suspicious.
A few of the quantifiers seized their get-together as an occasion to respond to attacks on their own particular work. The University of Pennsylvania's Theodore Hershberg--whose massive Philadelphia Social History Project has thus far consumed 12 years and $3 million in federal money, while generating 100 papers by an interdisciplinary team of 45 researchers--had been singled out (by Princeton's Stone) as the "Tyrannosaurus rex" of his profession. But Hershberg countered that the traditional historian, working alone with only a few traditional tools, is like a "19th-century artisan" obstinately resisting the advantages of the assembly line.
Stanley L. Engerman of the University of Rochester, introduced to the conference as "the worst-dressed man in economic history," recalled the controversy that swirled around the 1974 slavery study "Time on the Cross," which he coauthored. The book, drawing on computer analysis of census, probate and plantation records, concluded that slaves were less brutally treated than earlier scholars had assumed, and disputed the theory that slavery became unprofitable in the years before the Civil War. Many of the critics professed to challenge the data and methodology, said Engerman, but their fundamental objections were ideological or emotional.
"The major fight is really not about methods but about what are the questions to be pursued," said Engerman, who came to the conference in a bulky crew-neck sweater and a sports jacket of slightly wrinkled tweed.
Howard University's T.J. Davis protested that Engerman could not dismiss all criticisms of "Time on the Cross" so easily. There had been charges that the book, in estimating the frequency of slave whippings, "did not properly calculate the population at risk," said Davis, and surely this was a disagreement over methods. Engerman declined to concede the point.
Kousser said he would gladly submit "Time on the Cross" or the Philadelphia Social History Project to a cost/benefit comparison with such traditional history ventures as the collected papers of the American presidents. And in what a later speaker described as the "so's your old man" argument, Kousser contended that traditional historians are no less error prone than quantifiers. The difference, he said, is that the traditionalists' claims usually have to be taken on faith, while the quantifiers' can at least be verified or disproven. The authors of "Time on the Cross" had voluntarily turned their raw data over to some of their fiercest critics, Kousser pointed out. Obscuring History?
By the second day, much of the fraternal conflict seemed to have spent itself, and the conference, to chairman Jarausch's relief, became more businesslike--although no less energetic. Louis Galambos of Johns Hopkins questioned a tendency to "apply more powerful tools to smaller and smaller subjects." But Louise Tilly of the University of Michigan replied that there are times when a smaller unit of study can yield larger insights into human behavior.
There was near unanimity on the need for more skeptical criticism of statistical sources. Belgian economic historian Herman Van Der Wee noted that the official definition of "peasant" may change over time. In one census, the word might refer only to the head of the household, while in the next census it might include other members of the family as well. Obviously, a historian must be aware of these differences, but as France's Patrice Bourdelais commented--in five minutes of French that an interpreter mysteriously boiled down to a single sentence--source criticism is "no fun."
During their coffee and lunch breaks, several of the conferees explained a major, and controversial, advantage of quantitative history--the ability to "work around" weak or incomplete data. Voting records are usually totaled at the county or provincial level, for example, but quantifiers have gradually evolved techniques for making "ecological inferences" about the voting habits of different ethnic and class groups within the electorate.
Some traditional historians--and some quantifiers, too--argue that the methods have become too sophisticated. Princeton's Theodore Rabb warned his fellow conferees against a temptation to say, "I'm going to make it as obscure and as difficult as possible, even to my own colleagues." Serious historians must try to reach a broader public, said Rabb, "or the Antonia Frasers will get in there first."
"Physics hasn't followed that road," Kousser retorted, "and it has, perhaps even by its very arcaneness, increased its funding."
English historian Roderick Floud recalled predictions in the '60s that all historians would soon be embarking on a "painful process of retooling in the struggle to avoid technological obsolescence." But 20 years later, as Hershberg observed, "There are still very many historians who have not felt it necessary to acquire the quantitative skills even to understand the quantitative work of their colleagues, let alone to do it themselves." You can still get a PhD in history with no statistical training, Hershberg lamented. Enriching, Not Exhausting
The United States and the rest of the world were represented about equally at the conference, but not all the foreigners were entirely pleased with the flow of the discussion. One Latin American participant complained that his problems were far more basic than those he had heard about here. In his country, "if you use quantification to analyze the distribution of income and you show that the minister of economics is lying, you'll be thrown in jail."
Of the communist nations, East Germany, Romania and the Soviet Union were represented, and the two Russian participants not only took a vigorous part in the discussions but--during a coffee break--experimented eagerly with a Texas Instruments pocket computer demonstrated by Richard Jensen of the Newberry Library of Chicago. Modern quantitative history had its beginnings in the Soviet Union 25 years ago, Jensen explained, and to this day the "fancy methods" of the Russians sometimes leave their American brethren "dumbfounded." But small computers have yet to appear on the Soviet scene, so academics must battle for time-sharing access to big computers, even for relatively simple calculations.
The Russians presented their papers late Friday to a packed house in the Woodrow Wilson Center library. (The attendance was unusual, one American noted. "Normally the attrition rate is much higher. Most everybody finds the bars by the afternoon of the second day.") I. Kovalchenko--with Juhan Kakh translating--said he didn't understand the pessimistic talk he had heard here about quantitative history's status and future prospects. "The renaissance speaks for itself," he said. "I don't see any other way to make real good international comparisons than only quantitative methods . . . because when we don't use them we only have emotions and then we don't understand each other." But quantitative methods "can just enrich our knowledge of the past," said Kovalchenko, "not exhaust it."
In the same vein, Floud said the quantifiers, instead of claiming "special virtue," should learn to "inter-weave quantitative material with other types of historical evidence which we have tended to dismiss as impressionistic or unscientific."
The urge to merge seemed to inspire others as well, and Cornell's Joel Silbey had something like the last word. Earlier, several participants had lashed out against J.H. Hexter's "Doing History," a major anti-quantitative work which argues that historians should be interested in the "why" of history--why, for example, the Giants came from 13 1/2 games back in August to win the 1951 National League pennant. At the conference, Hexter's book had been referred to more than once by the derisive name "Doing Baseball." But Silbey insisted on taking Hexter's argument more seriously.
"I think it should be part of the record," he said, "that the reason the Giants won in 1951 was that they were quantitatively and qualitatively better than the Dodgers.