"I dispute in particular," said M. Dupin in The Purioined Letter, "the reason educed by mathematical study." Thinkers naturally espouse their own talents, ignoring or even evading the unfamiliar, the unorthodox and the unknown. Dupin excelled at a peculiarly unsystematic form of detective work--hence his own aversion to the mathematical. Historians, the most sleuth-like of social scientists, held off the mathematical tide longer than scholars in many fields, although few historians today question the importance of the heavily statistical works of history produced in the last few decades.

The study of history has its equivalent of Dupin the relaxed thinker puffing on his meerschaum, scoffing at the scurrying police as they collect their clues. Worried because "the nineteenth-century pre-eminence of history in the sphere of intellect no longer obtains," intellectual and musical historian Jacques Barzun (University Professor at Columbia, author of Darwin, Marx. Wagner) has undertaken to incite resistance to modern modes of history. In Clio and the Doctors: Psycho History Quanto-History, and History (University of Chicago Press) he cites the depths of the problem he and some other older historians see: The historical sense in modern populations is feeble or nonexistent, as Ortega pointed out, even though the mania for keeping records, building archives, and celebrating trivial anniversaries is rampant. Indeed it is probably the decline of a true sense of history that encourages those pseudo-historical manifestations.

However manic the score-keepers of society have become though, there is no reason we should suppose, as Barzun does, that proliferation of information necessarily leads to the sterilization of history. Barzun seeks the victory of historical artist over historical statistician, without considering the possibility that someone might be both. Yet Stephen Thernstrom's heavily statistical studies are as sensitive to the unquantifiable as any previous works on social mobility. Richard Sennett's book on nineteenth-century family life in Chicago (Families Against the City) is as audacious and speculative, though not as wide-ranging, as anything Barzun has written--but, unlike Barzun, Sennett presents carefully examined statistics to support his conclusions. Statistical gaffs of the sort Barzun describes, such as a study (a half-century old) of "geniuses" that studied all the people in Who's Who, are problems of historical thinking, and not record-keeping.

Few historians are likely to take Barzun's objections seriously, except perhaps as a warning that they should be careful no matter what sort of data they employ. Luckily, not all older historians are so full of methodological reaction. Oscar Handlin, not the most liberal man at Harvard, is the mentor of both Sennett and Thernstrom, and he managed to pass on to the younger historians the critical historical sense that served him so well in less quantitative works. The two cultures of humanism and science have probably come closer to merging in history than in any other field.

More dissatisfaction with quantitative history is found among undergraduates than among threatened historians. Supposedly so well trained at least on an elementary level by SMSG and the new math, many of the most able students at Harvard fear or hate thinking with numbers so much that they cannot evaluate some of the most important works of our time. Part of the problem may be that, at this school, anything mathematical smacks too much of the dreaded pre-med courses. Worse, "thoughtful" courses in the social sciences often cover the feeblest exclusively verbal epigones of Mars but not far more important works by thinkers like Paul Baran, who are articulate both with words and numbers. Too often, students are taught that they should think thoroughly before looking at facts.

At the same time, those mathematical works that do find their way into discussions in or out of the classroom are being accepted too uncritically. The massive reevaluative scholarship on slavery, summarized in Time on the Cross: The Economics of American Negro Slavery by Robert William Fogel (who will teach at Harvard beginning next year) and Stanley Engerman, is now the single best-known work of quantitative history. The book is published in two volumes--the first presents the conclusions, while the second, more technical volume explains how the authors got there. Much of the second volume cannot be understood without advanced mathematical training--but much of it can be, and reading the first without the second allows little basis for critical evaluation. Yet few undergraduates seem aware of the second volume, and of the half dozen or so Harvard courses that required Time on the Cross last term, only one considered volume two.

Quite a few History and Literature majors were nonplussed last year by the selection of Stephen. Thernstrom as the speaker at the annual History and Lit dinner. His brand of history was not theirs. Yet the distinction they saw was not one of historical philosophy or bias or method of analysis, but merely that Thernstrom employed certain facts from history that, being numbers, required some basic mathematical treatment before being placed in a larger frame of analysis. The real differences among historians--those we should tune our minds to discover--do not concern whether data is mathematical or not but rather the ways in which the data is used to form broader historical generalizations.

Unfortunately, the authors of the new quantitative history themselves often deter this broader evaluation of their work by their exaggerated claims to objectivity. "Success in this operation required, no less than in the operating room of a modern hospital, the adroit use of professional skills in a cool, detached manner." Fogel and Engerman write in their second volume. But it was not a detached analysis that told Fogel and Engerman how many whippings constitute harsh treatment of slaves, or how much confidence slaveholders had that the system would endure. So long as a researcher confines himself to recompiling old records, his work can indeed be devoid of historical context and bias. But as soon as he compares his results to those of others, as soon as he draws any historical conclusions, he loses his claim to special objectivity and his work becomes another work of history, whether good or bad.

Time on the Cross is an ambitious enough work to contain a good deal of bad history. Its lapses would be easier to excuse if Fogel and Engerman were not so insistent on the revolutionary nature of their method. Often they exaggerate the myopia of earlier historians, in order to make their own conclusions seem more extraordinary. Their work is not, even in conception, the comprehensive evaluation the authors believe it to be. It is a work with new insights and with new speculation; more of the truth could be discovered if people spent more time thinking about the history, less about the technique.

Time on the Cross is heavily influenced, for example, by many of the conceptual biases of neo-classical economics, which are even more out of touch with the early nineteenth century than they are with the modern world. In the same way, many earlier historians of slavery were influenced by racism. These are the real problems. The muse Clio, to whom Barzun appeals, should be more tolerant of methods than either Barzun or Fogel, but far more attentive to preconceptions--aware still that history never embraces more than a small part of reality, and coupling whatever means of reason with whatever means of observation, to allow a fuller embrace.