Emic versus Etic Records

Euroamerican documents such as newspapers, military reports, and history books frequently misrepresent armed conflicts. Because of the high stakes involved, each side tends to demonize the enemy, minimize its own failures, and exaggerate its own successes. The saying “the first casualty of war is truth” is often repeated in journalism schools; its more general version is that the winners get to write the history books.¹ Today, many people wonder why 23 Congressional Medals of Honor were awarded to US soldiers for the Wounded Knee massacre, but at the time many people viewed the event as a hard-won victory against a fierce enemy. In an article for the Nebraska State Journal about the burial of US soldiers killed at Wounded Knee, reporter William Fitch Kelley (1971:206) wrote: “Gallant soldiers, you fought the foe most nobly; you wavered not in the hour of danger, when the treacherous Indian, without warning, shot you down upon Wounded Knee.” Kelley’s stories omitted the fact that these noble men had mostly died from their own side’s “circular firing squad” at a conflict in which they took the lives of an estimated 200 unarmed women, children, and infants. That might still be the story in the history books, had not photographs such as that in figure 4.1 and the accounts of Lakota survivors collected by Charles and Mary Eastman (1945) and James Mooney (1896) eventually come to light.

Figure 4.1. Wounded Knee battlefield: Big Foot’s camp three weeks after the Wounded Knee Massacre. (Digital file from original item. Digital ID: ppmsca 15849 Library of Congress Prints and Photographs Division Washington, DC 20540 USA http://www.loc.gov/rr/print/.)

This is just one example of why it is important to seek out multiple accounts. The winners might write history, but they typically do not write the whole, unvarnished truth.

Kinds of Plains Indian Records

Warriors created painted hides and ledger-book drawings to publicize individual deeds or coups. These Contact-era works have been a key to interpreting late-period Indian rock art throughout the Great Plains (Keyser 1979, 1987b, 1996, 2000; McCleary 2008b; Rodee 1965; Sundstrom 1990:316–321; 2004:99–113; Sundstrom and Keyser 1998). Because ledger-book drawings, hide paintings, and Biographic rock art together form a single narrative art tradition, interpretations from art in one medium can be applied to another. In the words of Father Pierre-Jean DeSmet, “They have . . . still more remarkable modes of communicating thought. The large figures displayed on their buffalo robes are hieroglyphics, as easily understood by an intelligent Indian as written words are by ourselves; and they often contain the narrative of some important event” (Chittendon and Richardson 1905:681). Those creating these records retained the same or similar pictographic conventions for showing the protagonist, enemy/victim, capture of weapons or horses, and the like as the media shifted from painted hides to pencil drawings on paper (Greene 1985) (figure 4.2).

Figure 4.2. Pre-horse-era hide painting from Musée del Homme, Paris, (left) and two pages from Running Antelope’s pictographic autobiography (right).

Like the ledger-book drawings and hide paintings (Afton et al. 1997; Berlo 1996; Greene 1985; Keyser 2000; Mallery 1972; Maurer 1992; Szabo 1994), the vast majority of the narrative rock art in the northern Plains depicts battle scenes, horse raids, or other warrior activities. The art tradition includes few scenes of everyday activities or ceremonies, although some artists created such drawings on paper to sell to non-Indians during the early reservation period (Berlo 1996:18, 35; Ewers et al. 1985:8–10; Maurer 1992; Szabo 1994:27). Many petroglyphs can be interpreted within the larger narrative warrior-art tradition of the northern Plains, including hide paintings and ledger-book drawings, which are in turn interpreted based on the recorded statements of the people who created them (Keyser 1996, 2000; Mallery 1886, 1972; McCleary 2008b).

Winter counts are lists of year-names representing the most significant events in the life of the individual or band for each of the years included in the count (Calloway 1996:31–33; Greene and Thornton 2007; Howard 1960, 1976; Mallery 1886, 1972; McCoy 1983; Sundstrom 1997, 2003, 2006). Originally, these lists comprised pictographs, but later many were recorded as Lakota, Dakota, or English text. The vast majority of winter counts available for study are from the Lakota, but a few Yanktonai Dakota, Mandan, Kiowa, and Blackfoot winter counts also exist. Each year’s name was something like a headline for a specific event, the details of which the winter count keeper had committed to memory, and that event in turn evoked everything else important that had happened that year (figure 4.3).

Figure 4.3. The Thin Elk-Steamboat Winter Count, Buechel Lakota Museum, St. Francis Mission, St. Francis, South Dakota. Drawing by author.

Most of the scholarly literature on winter counts treats them as items of material culture, rather than sources of historical data (e.g., Maurer 1992). Few scholars have brought winter counts and other indigenous documents to bear as primary sources on historic research. One exception is George Hyde, who used Lakota winter counts in his accounts of Oglala history (Hyde 1957, 1961). James Howard’s study of Yankton ethnohistory also treats winter counts as historical documents (Howard 1976). Ron McCoy analyzed winter counts as a source for pre-1800 Lakota history (McCoy 1983). Other studies using winter counts have focused not on history per se, but on epidemics, migrations, and astronomy (Chamberlain 1984; Henning 1982; Sundstrom 1997).

The neglect of these sources has several explanations. The first is either a lack of awareness that they exist or a lack of understanding of how to interpret them. As mnemonics for more detailed oral narratives, the pictures themselves provide limited information. With winter counts, the picture refers to an event, which in turn refers to all the important events of that year; thus, the picture itself is only a tag for the historical events linked to it. Drawings on hide or paper provide a more complete narrative of a war event or deed, but the researcher must learn how to interpret the narrative from the pictures (e.g., Greene 1985).

A second problem is distrust of the accuracy of the indigenous accounts. This springs from Western culture’s privileging of the written word and from an ethnocentric view that oral history is necessarily less accurate than written history (Lowie 1915). Because our culture does not train us to hear and memorize narratives with accuracy, we assume that this skill is either difficult to acquire or not present in other cultures. But even literate cultures use rote memorization to implant information such as the Koran, Bible verses, addition and multiplication tables, nursery rhymes, and poems. Clearly, people can learn to recite long narratives with near-perfect accuracy. A related, third problem is researchers’ assumption that such records are too biased to be of value. While it is true that the system of recording and publicizing war deeds so central to Plains Indian life was likely to omit or minimize unsuccessful expeditions, it also demanded detail and accuracy in the records and their interpretation. A warrior who falsely claimed credit for a coup faced ridicule and punishment. The omission of defeats is less a problem in the winter counts, in which unsuccessful military actions are as likely to be recorded as successful ones. A fourth impediment to use of indigenous documents is simply one of cultural chauvinism, whereby researchers privilege the familiar forms of European documents over the less familiar and more esoteric non-European forms or lump together all forms of oral tradition as mythological (DeMallie and Parks 2001:1062; Goldenweiser 1915; Wolf 1997).

While indigenous records have limitations, pictographic narratives from the contact and late precontact era on the Great Plains, properly interpreted, can fill in the blanks left by archaeological data and can help to correct the biases inherent in the history that archaeological studies produce for this time and place.

Weaponry and Tactics

Scenes of conflict are rare in early Plains Indian rock art. Although rock art is notoriously difficult to date, best current data indicate that so-called warrior art is limited to the last 1,500 or so years before European contact. The earliest conflict-related Plains rock art consists of pictures of shields and shield-bearing warriors. The weapons complex here consists of large body-shields and bows, spears, antler-tine pikes, and clubs (Greer and Greer, chapter 2, this volume; Keyser, chapter 3, this volume). Most of this early warrior art focuses on showing the details of shields, weapons, and other accoutrements, including items indicating membership in warrior societies, such as the bow-spear. Early warrior art that shows action often depicts the shield-bearing warrior spearing or counting coup on a hapless individual lacking shield or weapons. The observation that women as well as men are depicted as such unarmed victims suggests that, as in later times, warriors could gain status by slaying any enemy, not just men or other warriors (Keyser et al. 2006). Some women did go to war in later times (Ewers 1994; Greer and Greer, chapter 2, this volume; Keyser et al. 2006), but the extent to which this happened before the contact era is unknown. It is frequently impossible to tell the sex of warriors in the early warrior art, because the warrior’s body is concealed behind the large shield; those that do show sex are male. An exception is a rock art panel depicting a woman—or a person in a dress, at least—engaged in combat with a warrior on horseback (McCleary 2008b:266, figure 78).

Some early warrior art includes depictions of severed heads or scalps, indicating that scalp-taking was practiced at this time; however, whether the practice had a religious function as in later times cannot be construed from either the rock art or other archaeological remains (figure 4.4). The first definite reference to scalp-taking in the winter counts is for the year 1749, although one can speculate that reports of small, wide-ranging war parties first mentioned in 1711 also refer to scalp-taking expeditions.

Figure 4.4. Rock art and ledger drawing of severed heads or scalps. Upper left and upper right, rock art, North Cave Hills, South Dakota; lower left, rock art, southern Black Hills, South Dakota; lower right, Amos Bad Heart Bull drawing of Lakota Victory Dance, date unknown (pre-1918 book of drawings), showing women with scalps attached to coup sticks (Amos Bad Heart Bull Ledger, Plate 39b, No. 85. View the complete book at plainsledgerart.org).

The early days of equestrian warfare are poorly recorded in the winter counts. This is because most winter counts available for study today do not extend back before the early 1800s: too late to record initial encounters with horses, which must have taken place before 1700. A few winter counts are older (McCoy 1983), but it appears that the extent of any given winter count was limited, as the earliest years were forgotten or generalized and dropped off the winter counts over time. For example, the winter count of Battiste Good generalizes early history into one pictograph for every 70-year period (Mallery 1972:287–328). This series of legendary events records early encounters with horses for the periods ad 1141–1210 and 1421–1490, both of which are obviously not historically accurate. The same winter count refers to using horses for bison hunting sometime between 1631 and 1700, which is reasonable for the northern Great Plains. At 1700, the Battiste Good winter count begins a year-by-year record of Lakota history, with horses referred to routinely throughout the subsequent record. We do not know when northern Plains warriors began to employ horses in battle, but the Battiste Good winter count records equestrian warfare for 1714 and 1715, and horse raids for 1708, 1709, 1717, and 1718.

Turning instead to rock art, the earliest pictures of equestrian warfare in the Great Plains show warriors carrying shields atop horses with leather armor (figure 4.5; also see Greer and Greer, chapter 2, this volume, and Keyser, chapter 3, this volume). These warriors carry lances and are most often shown attacking unarmed and unmounted enemies. Horse armor had a short tenure on the Plains, probably because it impeded the great advantage of horse warfare, the speed and agility of a well-trained mount. After the larger shield and horse armor were cast off, the weapons of choice were the bow and arrow and the lance, with a much smaller shield, if any. Rock art and ledger art of this later period of equestrian warfare typically show mounted warriors overcoming pedestrians, including women (figure 4.6). On the northern Great Plains, guns are first mentioned in the winter counts later than horses, circa 1800. According to the winter counts, the bow and arrow retained its place as the primary combat weapon until the mid-1800s and continued in use well after guns were widely available.

Figure 4.5. Rock art depicting horse armor. Upper left, Bruner Ranch, eastern Montana; upper right, North Cave Hills, South Dakota; lower, Wilson Creek, Kansas.

Figure 4.6. Warrior on branded (US Cavalry?) horse counting coup on pedestrian warrior, North Cave Hills, South Dakota. Drawing by James D. Keyser.

Type, Scale, and Geographic Range of Armed Conflict

Regarding types of conflict, the indigenous sources considered here are consistent in indicating that the vast majority of armed conflict during the contact era was limited in scale. Because winter counts list only one event per year, battles that involved large numbers of combatants, prolonged engagements, or mass casualties should be well represented. Nevertheless, a sample of Yanktonai Dakota, Sicanju Lakota, Oglala Lakota, northern Lakota, and Peigan Blackfoot winter counts² mentions small war parties and opportunistic attacks twice as often as large battles (Appendix 4.A). This sample included 26 mentions of large-scale battles, meaning battles involving more than 30 warriors per side. By comparison, the sample included 29 mentions each of small war parties (a few to as many as 30 warriors) and opportunistic attacks on hunters, wood gatherers, eagle trappers, women gathering prairie turnips, families camping away from the main group, and the like. Another 43 events recorded in the winter counts are individual battle casualties, of which most if not all took place in skirmishes between enemy war parties. Only 22 multiple-casualty events appear in the winter counts. Of these, 12 report between two and 12 deaths, and presumably refer to battles between war parties, as opposed to attacks on villages. Events involving more than 30 deaths or simply recorded as “many killed” include three attacks on Middle Missouri earthlodge villages and one on a Skiri Pawnee village. Two others refer to the annihilation of entire bands, but it is not clear whether this took place at a village or encampment or in some other circumstance. Four of the 10 high-casualty (more than 30 killed) events were losses suffered by forces attacking encampments or settlements. In these cases, an enemy attack on a camp or village was suppressed at high cost to the attacking force. Three others refer to fights against non-Indian forces: the Villasur fight, the Grattan fight, and the Fetterman fight.

While the scope of warfare was limited and many conflicts unplanned, the Lakota winter counts also show that the geographic range of war parties was wide and gradually shifted toward the west over the period covered by these records. Conflicts sometimes broke out between widely separated groups, such as the Ho-Chunk and Lakota, or the Lakota and Utes (figure 4.7). Because these groups were not fighting for commonly claimed territory, it would appear that such conflicts were the result of chance encounters between war parties. At the same time, the locations of battles recorded in the Lakota winter counts indicate a shift toward the west (figure 4.8). This presumably reflects the increased power and territory of the Lakota alliance as it expanded westward across the Missouri and into the Black Hills, Powder River, and Little Missouri country.

Figure 4.7. Enemies mentioned in Dakota/Lakota winter counts, 1685–1875; specific incidents (black) and timespan (gray). Ho-Chunk, ??; Mand, Mandan; Shosh, Shoshone.

Figure 4.8. Geographic locations of battle mentioned in Dakota/Lakota winter counts.

Motives for Warfare

As noted, scalp-taking was an important motive for small-scale armed conflict in the contact era and the late precontact Great Plains. Acquiring a scalp to assist the soul of a deceased comrade or relative along the way to the afterlife was a regular part of mourning and is best understood in this religious context, rather than as trophy-taking (Sundstrom 2015). Such events most often occur in the winter counts not as records of the scalp raid itself, but as records of the ceremony in which the deceased person’s loved ones impelled his comrades to pledge themselves to getting a scalp when the season for battle next came around. In the 200 conflict-related events in the winter counts, scalping is explicitly mentioned only four times, although the pictographic winter counts frequently depict a forehead wound on warriors killed in battle.

Horse raids are mentioned only four times in the sample of winter counts used in this study. This apparently is a result of underreporting, because horse raids are shown frequently in ledger-book drawings and in rock art. Unless a new breed of horse, an especially fine warhorse or racehorse, or a very large herd was taken, horse-raiding events were likely overshadowed in the winter counts by more significant or unusual events. By contrast, records of horse raids are common in northern Plains rock art (figure 4.9).

Figure 4.9. Three horse-raid scenes in rock art from the North Cave Hills, South Dakota: (left) remnant of horse with tether rope; (center) horse raider with quirt in back hand; (right) capture of a horse carrying a travois.

Counting coup was undoubtedly a strong motivator for offensive warfare. This is less evident in the winter counts than in the innumerable ledger-book drawings and rock art panels depicting individual coups. But even in the winter counts, which were not created explicitly to record or publicize coups, 12 of 200 conflict events recount unusual or especially impressive coups. Only a small percentage of coups would have been recorded as year names in winter counts, because they record only the most memorable or important event of each year.

Winter counts indicate that several attacks on hunting parties, as well as many of the large battles, were attempts to take over or defend good buffalo-hunting grounds. A Blackfoot winter count mentions a large battle with the Salish over a bison hunting territory, while another mentions Blackfoot fighting Kutenai who came to hunt bison in the Alberta foothills. The latter may have been during a time when the herds were decreasing, but the former event in 1811 is not likely to reflect reduction of the herds. The numerous clashes between Lakota and Pawnee, and Lakota and Crow, were motivated in large part by Lakota attempts to take over bison grounds on the North Platte and in the Powder River Basin (McGinnis 1990:109–128). An event recorded in Lakota winter counts not included in the sample used here records a fierce battle with a group of Metís who had come south to hunt bison in territory claimed by the northern Lakota tribes.

Luc Bouchet-Bert (1999) proposed that some warrior art was made to warn away enemy intruders. This idea is given credence by ethnographic accounts of war parties leaving behind pictographic notices when they killed enemies (Grinnell 1926:31–34; Hans 1981:50; Hyde 1968:54; King 1880:84), and by the observation that warrior art appears to be relatively uncommon in areas held as neutral zones, such as the Black Hills, Pipestone (Minnesota), and the Sweetgrass Hills. By depicting well-armed, powerful warriors and scenes of victory, warrior art that was placed permanently within the landscape in the form of rock art served notice to any trespassers. If correct, this use of rock art supports the concept of warriors defending hunting territories. The winter counts, as well as other oral histories, clearly demonstrate that warriors moved within vast territories, occasionally encountering and often confronting their counterparts from tribes living hundreds of miles away. This degree of mobility very likely created opportunities to exploit new territories and increased the conflict between groups as core territories shifted and expanded or contracted. The warrior set could keep their own villages and camps safe and scout out promising hunting grounds by engaging in such long-distance expeditions. In this way, warriors could effectively defend and expand their territories while protecting their villages from direct attack.

Finally, the Lakota and US Cavalry attack on the Arikara village known archaeologically as the Leavenworth site (Krause 1972) is recorded in the winter counts primarily as a plunder event. The invading Lakota force raided the Arikara corn fields and corn caches, but not much fighting ensued because the Arikara abandoned the village when they saw the combined force approaching. The year name is “much spoiled corn” or “much dried corn,” in reference to the raid on the Arikara fields and cache pits. It seems unlikely that the Lakota were induced to join the US force by the possibility of plundering the corn stores, but this is mentioned here as a possible secondary motive for attacking the horticultural village.

While precontact and early Contact-era rock art depicts both men and women as victims of enemy attack, a trend emerges in the later records of taking women and children captive rather than killing them (Keyser et al. 2006). The Yanktonai winter count records that in an attack on the Nuptadi Mandan village in 1780 many women and children were taken captive to boost the population of the Yanktonai. This presents a striking contrast to the Crow Creek and Larson archaeological sites, at which large numbers of women and children were among those killed in attacks on earthlodge villages. By the late 1700s, epidemic disease introduced from Europe had taken a heavy toll on native populations; thus, enemy women were increasingly seen as more valuable as potential wives and mothers than as sources of scalps. Although the attacks on the Larson site and Nuptadi were only nine years apart, the attack on Nuptadi corresponded with, or closely followed, a very widespread smallpox epidemic that may have both weakened the village’s defenses and made its potential captives an attractive target for attackers. In 1838, a Skiri Pawnee war party attacked a Lakota family and carried off a young girl to be sacrificed in the Morning Star Ceremony, which they were performing in hopes of abating an especially deadly smallpox epidemic. A Lakota war party avenged her death soon thereafter, but brought the deadly virus back with them to their own camps.

It is not clear whether warfare on the northern Great Plains was motivated by a slave trade. The winter counts mention captive women and children, but not slaves for trade. Pictographic records such as rock art and ledger-book drawings clearly show capture of enemy women (Keyser, chapter 3, this volume), but these lack clear evidence of taking captives as chattel, as opposed to adopting women and children into the captors’ group (figure 4.10). This contrasts with some ethnographic accounts of Plains Indians trading captive women to other groups (Greer and Greer, chapter 2, this volume).

Figure 4.10. This rock art panel from the North Cave Hills, South Dakota, shows a warrior touching a woman’s genitals, indicating that he is claiming her as a wife.

Conclusion

In conclusion, records made by Plains Indians reveal several things about warfare. First, they show that motives behind warfare were many and overlapping, but that principal among them were scalp-taking, acquiring war honors, and defending or expanding hunting territories. Horse raiding was an activity that might be accomplished without fighting; however, many such raids precipitated defensive or retaliatory action from the enemy. Both acquiring the horses themselves and gaining the honor of bravely capturing them belong in the list of motivations for warfare. Most hostile encounters were either between wandering groups of warriors or between war parties and civilians caught off guard while away from their camps or villages. Large-scale attacks on villages that led to mass casualties, such as those recorded archaeologically at the Crow Creek and Larson village sites, were relatively rare and not always successful. Finally, tactics and motives changed over time as horses and European-manufactured weapons became available and as population declines due to waves of imported epidemic disease promoted taking women and children as captives to be adopted into the captor society.