Academic journal article
By Kay, Marvin; Sabo, George, III
Southeastern Archaeology , Vol. 25, No. 1
A fundamental metaphor of Southeastern Indian belief was expressed by the structural relationship death:winter. Archaeological data and documentary accounts illuminate this relationship and demonstrate a widespread cognitive lore about the marking of the seasons by use of the sun and possibly other celestial bodies. The apparent timing of pre-contact mortuary ceremonialism with the winter solstice sunset consistent with Caddoan winter death imagery reflects a local manifestation of the North American Orpheus myth.
Guidance to the meanings of prehistoric rituals may come from ethnohistoric accounts of indigenous beliefs and practices. We suggest the many parallels in eastern North American mythology provide a symbolic coherence transcending ethnic and language group affiliation. For reasons of geographic proximity and long-term cultural stability (for insightful evaluations see Roper  and Story [1990:320-323, 346-356]), we believe that elements derived from Caddoan ethnohistories of the central and southern Plains may be used to guide our understanding of prehistoric mortuary rituals in the western Ozark highlands and adjacent portions of the southern Plains. These ethnohistoric details satisfy generally accepted criteria (Wood 1990:89-90; see also Echo-Hawk 2000) for believability: they were likely to have been of "little consequence" to the authors or "noncontroversial matters" for which early explorers would have had little to gain by reporting them; they also appear to be independently confirmed in ethnohistoric accounts at different times and places.
We are specifically interested in the symbolic identification of death with winter that occurs across regions occupied by precontact Caddoans and their neighbors. The Caddoan study area is, as Griffin (1967) noted long ago, the location of renowned mound complexes of Spiro, Harlan, and Norman, and geographically it is on the western fringe of late prehistoric Southeastern developments. But in our opinion and that of many others, the study area and its multiplicity of agrarian settlement forms ought to be central to understanding the evolution of the Mississippian world.
The structural relation death:winter is metaphorically represented in postcontact Caddoan symbols (Hatcher 1927a, 1927b; Swanton 1942; Weltfish 1965) and in ethnohistories for many other eastern North American groups (e.g., Gruber 1971; Hall 1997; Robinson et al. 1985). These associations are consistent with the archaeological data and further amplify the meaning of prehistoric mortuary or related rituals and their timing. The material referents of the rituals are amenable to archaeological analysis, too (Brown 1996, 1997, 2003; Goldstein 2000; Hall 2000; Knight 1986; Knight et al. 2001; Schnell et al. 1981; Sears 1961; Story 1990:340-342; Story 1998; Watson 2000). The ritual practices, as Rogers (1991) predicted, redefined society along new lines of non-kin-based authority while being structured by categories and principles borrowed from kinship idioms (Sabo 1998:171). They were vital to the evolution of ranked societies in southeastern North America and have general ethnographic parallels throughout this region (Knight 1986, 1989, 1990).
We suggest that the death:winter symbolism provides a useful interpretive framework for understanding the functions of prehistoric, mound-related rectangular structures now found from northeast Texas to southwest Missouri. These structures differed from contemporary Mississippian constructions in the adjacent Midwest in having had hall-like, extended entryways. Entryway posts often were set in trenches. Insofar as may be judged from archaeological details of construction and ethnographic accounts (Swanton 1942), these buildings mostly had a single entryway and lacked windows. Daylight illumination of the interior most likely came from the entryway. The buildings were common in the Arkansas River drainage and western Ozark highland (for overviews, see Brown 1984, 1996; O'Brien and Wood 1998:268-271; Vehik 1993) but also occurred in the Sanders phase (Bruseth letter to Kay 1995; Bruseth and Perttula 1981; Bruseth et al. …