At the time of this writing, the most recent book to examine Sand Creek is a collection of essays entitled Battles and Massacres on the Southwest Frontier. This book undeniably places the engagement in the massacre category. One of the contributors, Ari Kelman, examines the historical angle and makes a reasonably good case that it was indeed a massacre. Still, it is not difficult to notice that he had a proclivity to use the testimony to prove a point.1
For instance, in his use of Silas Soule’s court testimony, Kelman references Soule’s statement that he saw soldiers with scalps, but omits Soule’s later statement that he did not see any soldiers scalping Indians. Kelman quotes Sam Tappan’s statement that the barbarism of Sand Creek culminated in Soule’s assassination. However, he does not identify Tappan or note that he was the president of the military inquiry. Instead, he describes that important player only as “one federal official looking into Sand Creek.” Tappan’s attitude and his position are germane to the investigation. Kelman also went on to note that after Soule’s death, he was “unavailable to provide further testimony.” This is obviously true. What was left unsaid was that his death did not impact the information he could have shared because Soule completed his testimony a month before he was killed.2
This use of select pieces of testimony could be a conscious attempt to steer readers toward the conclusion favored by Kelman. Or, it could be just a matter of space limitation that would not allow for a complete discussion of relevant matters. More likely, however, this is the work of the subconscious mind, implicit memory, confirmation bias, and other psychological factors going about their stealthy business—all of which will be examined in more detail in Section Three.
In addition to Kelman’s historical assessment of Sand Creek, there is an archaeological perspective on Sand Creek by Doug Scott, a superb historian and archeologist. Scott has investigated and written on many different events in attempts to discover, among other things, the accuracy of the relationships of the historical record to the physical record. His work on the Little Big Horn, for example, is ground-breaking. As far as he his concerned, Sand Creek “unequivocally qualifies as a massacre.”3
Scott makes many good arguments, but there are other points to also consider. He discovered that there was a striking absence of firearm calibers likely to have been used by the Cheyenne or Arapaho. “Few bullets in trade gun calibers or firearm types known to have been issued as annuity items are present,” he observed, “thus supporting earlier conclusions that the Indians offered little resistance to the soldiers’ attack.”4
On its face, this sounds reasonable. But what if the absence of firearm evidence in the village and on the retreat route has another explanation—like many of the Indians used bows and arrows instead of other weapons? By examining the soldier casualties in the appendix, we can see that of the 76 casualties, 29 were attributed to a specific weapon. Of those 29, 16 were from arrows and 13 were from bullets. Perhaps a valid explanation for the dearth of firearm evidence was not entirely that the Indians put up little resistance, but that in more than half the cases, they inflicted damage with arrows, and not firearms. This also demolishes the contention some have made that most of the soldier casualties were self-inflicted via “friendly fire.” It is also important to note that both sides are susceptible to “friendly fire.” It is therefore reasonable to conclude that if some of the soldiers shot themselves, so did some of the Indians.
Scott also observed that, “Among the contemporary and survivor accounts of the fight, from both sides, there is little disagreement that the affair was very one-sided,” and that among the accounts “there is general agreement that the Colorado soldiers attacked a protected camp.” The Sand Creek affair may well have been very one-sided and the camp “protected,” but the eyewitness accounts in no way agreed on these points. If anything, one is struck after a careful reading of the records sampled above by the great disagreement over these very assertions.5
Scott also observes that one can discern the difference between a battle and a massacre by the numbers engaged, and by the wounded-to-killed ratios. It makes sense that in a massacre, the number of killed would exceed the number of wounded. So how many Indians were killed and wounded at Sand Creek? As already demonstrated, the estimates in the number of Indians killed ranges from about 70 to as many as 600, with most estimates falling in the 150 range. How many Indians were wounded? It appears as though estimates on that issue do not exist, or at least they have never been found and published. We can estimate the number of Indians killed, but we have no idea of the number of Indians wounded. How, then, can we draw conclusions about the ratios of wounded to killed and whether that indicates a battle or a massacre?
We will have more success in using the soldiers as an example. Studies of conventional warfare demonstrate that the wounded-to-killed ratio can be from two-to-one (two wounded for every man killed) to as many as five-to-one (five wounded to every man killed). In recent conflicts, the number has been estimated as high as 13-to-one, with the high modern ratios mostly the result of expeditious battlefield evacuation, expert medical treatment, and body armor. More wounded survive today than they did in the nineteenth century.6 Scott suggests that in conventional warfare, the wounded-to-killed ratio should be about five-to-one. We don’t know the Indian warfare wounded and killed rates, but we can be certain they were not be anything near that ratio. For example, Billy Breakenridge and Morse Coffin estimated the numbers of soldiers marching out of Fort Lyon to be from 500 to 650. Let’s reasonably assume the number was 600. We must then subtract Soule’s and Cramer’s companies from the mix because we know they did not actively take part in any hard fighting. That knocks out roughly another 100 men, leaving about 500 soldiers who were actively engaged in fighting at Sand Creek.
Historians have had a penchant for undercounting the soldier casualties at Sand Creek. According to Scott, 10 soldiers were killed and 38 wounded.7 Based on a thorough assessment of the official reports, newspapers, letters, and other documents, the number appears to be 24 killed and 52 wounded. This is a big difference. Scott claims, for example, that “The percentages of soldier dead (1.4 percent) and wounded (6 percent) are far more typical of conventional war, even in a one-sided battle, thus reinforcing by scientific methods the interpretation of the Sand Creek event as a war crime and massacre.”8
However, it appears as though there were about 500 soldiers engaged in the fighting and not 700, and 76 total casualties instead of 48. If so, then five percent of the soldiers engaged were killed and ten percent were wounded. That’s about a two-to-one ratio. If a five-to-one (wounded-to-killed) ratio is the norm, then it appears that the soldiers were the ones who were massacred at Sand Creek.
It might also be better to use Civil War-era statistics than comparing Sand Creek fighting with battle in Afghanistan in the twenty-first century. Trevor Depuy, who devoted years to military studies about numbers and war, stated that during the Civil War, which for the most part was a conventional fight, the North lost 2.1 men killed for every 100 engaged.9 The Civil War-era troopers at Sand Creek lost five killed for every 100 engaged. Does this mean they were massacred?
What all of this illustrates is that a historian can reach just about any conclusion and support it with statistics. How you use them is important. As historians, we can’t fire hundreds of bullets into the side of a barn and then draw a bull’s eye around the spot with the most bullet holes to show the accuracy of the men pulling the trigger. For a valid test, the bull’s eye would have be painted onto the side of the barn before pulling a trigger.10
If someone believes a massacre took place at Sand Creek, it is not difficult to set parameters accordingly, sift through the evidence, and find proof of a massacre. The same is true if you believe a battle was fought there. The task of trying to determine the truth from the preponderance of the evidence, scrupulously collected and studied with the best of intentions, cannot be divorced from one’s predilections. Historians and archaeologists are trapped to some degree by forces we cannot control, and often we do not realize these forces are influencing our research and writing.
In the end, it seems as though no matter how hard we try, we cannot extricate ourselves from the battle vs. massacre quagmire. And therein lies the ultimate conundrum, because the question itself is fallacious to begin with. Asking whether Sand Creek was a battle or massacre is a false dichotomous question. As historian David Hackett Fischer notes, the dichotomous question is divided into two parts—mutually exclusive with no overlap, no opening in the middle, and nothing possible beyond either end. According to Fischer, it is rare that any two historical terms can be so related unless one is specifically defined as the negation of the other. The law of the excluded middle, he continues, might demand obedience in formal logic, but in history it just doesn’t cut it. The dichotomous historical question is false because it is “constructed so that it demands a choice between two answers which are in fact not exclusive or not exhaustive.”11
Fischer makes his point by listing a number of book titles, including Napoleon III: Enlightened Statesman or Proto-Fascist?; The Abolitionists: Reformers or Fanatics?; Plato: Totalitarian or Democrat?; The Robber Barons: Pirates or Pioneers?; The Medieval Mind: Faith or Reason?; Martin Luther: Reformer or Revolutionary?; The Dred Scott Decision: Law or Politics?; What is History—Fact or Fancy?
These, and more like them, drive Fischer to distraction. According to Fischer, they are unsatisfactory, grossly anachronistic, very shallow, imprecise, and ambiguous. They are evidence of a “faulty pedagogical practice” that encourages “simple-minded moralizing.” He sympathizes with the poor students faced with such false dichotomies. Does the student try to demonstrate that dichotomous terms can coexist, demonstrate that there is a third possibility, or repudiate one or the other or both propositions? Fischer believes the questions serve only to shackle the student to false options, and the best response is to try to indicate the serious problem of structural deficiencies in the question as it is framed.12
So Sand Creek could have been a conventional battle, a massacre, both, or something in between. As we have seen, even those who were there did not agree. It is important to be mindful of the challenges facing historians who are trying to explain complicated events.
1 Richard K. Wetherington and Frances Levine, eds., Battles and Massacres on the Southwestern Frontier: Historical and Archaeological Perspectives (Norman, OK, 2014).
2 Ari Kelman, “What’s in a Name? The Fight to Call Sand Creek a Battle or a Massacre,” in Wetherington and Levine, ed., Battles and Massacres on the Southwestern Frontier, 123.
3 Douglas D. Scott, “Reassessing the Meaning of Artifact Patterning,” in Wetherington and Levine, eds., Battles and Massacres on the Southwestern Frontier, 147.
4 Scott, “Artifact Patterning,” 144.
5 Scott, “Artifact Patterning,” 147-48.
6 “Statistical Summary of America’s Major Wars,” www.civilwarhome.com/warstats.htm (accessed July 13, 2014); “Costs of War,” http://costsofwar.org/article/us-and-alliedwounded accessed July 13, 2014; “Mortality associated with use of weapons in armed conflicts, wartime atrocities, and civilian mass shootings: literature review,” www.bmi.com/contents/319/7207/407, accessed July 13, 2014.
7 Scott, “Artifact Patterning,” 135.
8 Ibid., 149.
9 Colonel Trevor N. Depuy, The Evolution of Weapons and Warfare (New York, 1980), 171.
10 Gary Smith, “How to Lie with Statistics,” Skeptic, Vol. 19, no. 4 (2014), 47.
11 David Hackett Fischer, Historians’ Fallacies: Toward a Logic of Historical Thought (New York, 1970), 9-10.
12 Ibid., 11-12.