- A Census-Based Count of the Civil War Dead
Counting the Civil War Dead
For more than five decades, Civil War History has served as the leading venue for scholarly publications on the Civil War era. Even in light of this impressive run, the editors of Civil War History feel that the following contribution by J. David Hacker of Binghamton University, SUNY, stands among the most consequential pieces ever to appear in this journal’s pages. Hacker, a specialist in quantitative methods, has utilized recently released microdata samples from nineteenth-century censuses to examine one of the archetypal “facts” about the Civil War—the oft-cited total of 620,000 plus deaths. Through a comparison of male survival rates between 1860 and 1870 with male survival rates in surrounding censuses, Hacker finds the traditional statistic understates the number of actual Civil War deaths by approximately 20 percent. In his estimation, the most probable number of deaths attributable to the Civil War is 752,000, although the upper bounds of his data set point to as many as 851,000 deaths.
As an exercise in the recalculation of a statistic, “Counting the Civil War Dead” might be regarded by skeptics either as a form of what Thomas Kuhn [End Page 307] described as “normal science” or as a misleading evocation of numeracy that belies the constructed nature of statistics. Such readings, we believe, miss the mark. “Counting the Civil War Dead” does more than modify a hoary bit of Civil War trivia; instead, it implicitly asks us to consider several questions that lie at the heart of the modern historical enterprise. How do “facts” emerge and become accepted by the profession writ large? How does the inevitably limited nature of historical evidence constrain our thinking about the past—and can we ever transcend these limits? Simply put, can we ever count the Civil War dead?
As readers will soon discover, the practical answer is no. The use of the most sophisticated tools of quantitative analysis can certainly overturn what was once accepted wisdom, but, in the final analysis, they can only provide us with a probabilistic range of excess male deaths during the 1860s.1 In a very real sense, however, fixating upon a precise number obscures the actual meaning of the numbers, as scholars such as William Blair, David Blight, Jane Turner Censer, Drew Gilpin Faust, Barbara Gannon, Caroline Janney, Stuart McConnell, and John Neff have clearly established the central roles occupied by loss and trauma in postbellum America. By placing the Civil War’s enormous death toll at the center of the postwar world, this generation of scholarship forces us to stop and reconsider the war’s meaning for period Americans. And since, as Hacker implies, the majority of the uncounted dead were likely southerners (thanks to deficiencies in Confederate record-keeping and the troubled postwar condition of the south), the “ghosts of the Confederacy” now seem more numerous and persistent than ever. In terms of the scale of the carnage, Richmond in 1865 was Paris in 1918.
Thus, what you are about to read takes us to “the frontiers of historical imagination” (to borrow a phrase from Kerwin Klein) and serves as a reminder that for all we know about the Civil War, there is still plenty that we do not—and can never—know. To prepare us for this journey, James McPherson has graciously provided this commentary on Hacker’s work. [End Page 308]
Commentary on “A Census-Based Count of the Civil War Dead” by James M. McPherson
I have been waiting more than twenty-five years for an article like this one. As I was working on my two syntheses of the Civil War era, Ordeal by Fire and Battle Cry of Freedom, I became increasingly aware that the standard estimate of 258,000 Confederate war dead was a significant undercount. Many Confederate records were lost or incomplete, especially for the last—and bloodiest—year of the war. The number of disease-related deaths of Confederate soldiers was clearly underreported. There were no reported Confederate noncombat deaths from “miscellaneous” causes—accidents, drownings, causes not stated, et cetera—compared with nearly twenty...