The Archaeology of Houses and Households in the Native Southeast

The Archaeology of Houses and Households in the Native Southeast

by Benjamin A. Steere


View All Available Formats & Editions
Use Express Shipping for guaranteed delivery by December 24 
Qualifying orders placed by 11:59 a.m. ET on Dec. 21 can still be sent to Manhattan, NY addresses. Details

Product Details

ISBN-13: 9780817319496
Publisher: University of Alabama Press
Publication date: 04/11/2017
Series: Archaeology of the American South: New Directions and Perspectives Series
Pages: 232
Product dimensions: 6.10(w) x 9.00(h) x 1.00(d)

About the Author

Benjamin A. Steere is an assistant professor of anthropology at Western Carolina University.

Read an Excerpt

The Archaeology of Houses and Households in the Native Southeast

By Benjamin A. Steere

The University of Alabama Press

Copyright © 2017 University of Alabama Press
All rights reserved.
ISBN: 978-0-8173-1949-6


Patterns of Architectural Variability in the Native Southeast

Most architectural studies begin at a fine spatial and temporal scale, usually at the level of a single house or site, and then move out for comparisons with other areas. In contrast, in this study, I start out at a larger scale and zoom in. I examine variation in structures grouped by chronological periods used by other archaeologists in broadscale syntheses in the Southeast: Middle Woodland (ca. 200 B.C. to A.D. 400), Late Woodland (ca. A.D. 400 to 1000), Early Mississippian (ca. A.D. 1000 to 1200), Middle Mississippian (ca. A.D 1200 to 1350), Late Mississippian (ca. A.D. 1350 to 1550), and Historic Indian (ca. A.D. 1550 to 1800) (following Anderson and Mainfort 2002; Anderson and Sassaman 2012; Bense 1994; Hally and Mainfort 2004). I focus on a single architectural variable (or one small set of related variables) at a time, exploring synchronic variation and diachronic change. This strategy is effective for identifying major patterns of architectural variation over time and gives additional context for analyses and comparisons at finer scales. Some of these trends and patterns have been identified in previous research and syntheses but others are not as well documented or understood.

In this chapter and the chapters that follow, I use the functional classes of structures outlined in the introduction to group and compare structures. The term structure refers to all types of buildings, regardless of function. Domestic structures are those that appear to have served primarily as dwellings in domestic contexts. These are the structures we usually define as "houses." Nondomestic structures are mostly large public structures used for gatherings and ceremonies, like town houses and earth lodges, but also some smaller special-purpose buildings, like sweat houses. They can be identified by their size and shape and also their special locations in settlements. Storage structures were primarily used for storing maize. They are generally small and clearly associated with one or more domestic structures in a household cluster. In early Spanish accounts these are called "barbacoas." The remaining structures fall into the other category, and only make up about 4 percent of the total.

Comparing structures over space and time in the Southeast produces evidence for at least four major architectural trends. First, architectural investment — the quantity of material and labor embodied in structures — generally increases over time and peaks in the Late Mississippian period. This is not especially surprising, given our general understanding of increasing sedentism in the Southeast after the adoption of intensive maize cultivation around A.D. 800 (Anderson and Mainfort 2002:18). There are, however, interesting deviations from this general trend. For example, there are Late Mississippian domestic structures in the uplands of the Oconee River valley that seem far more ephemeral than contemporary domestic structures elsewhere (Hatch 1995) and Middle Woodland domestic structures in middle Tennessee that seem unusually robust and permanent for purportedly semisedentary households (Faulkner 1988, 2002).

Second, given the radical changes in settlement patterns and community organization from the Woodland to the Mississippian to the Historic Indian period, there is more continuity in building traditions than might be expected. The average size of domestic structures changes over time, but the size difference between nondomestic structures and domestic structures remains similar throughout the chronological sequence. Certain architectural traits, such as the use of four central corner posts and semisubterranean basins, have a deep history.

In terms of broad spatial patterns of variation, changes in structure shape and size during the Late Woodland to Early Mississippian transition have a similar trajectory in the American Bottom and west-central Alabama. Something different happens to the east, in the Southern Appalachians. By the Early Mississippian period, structures look very similar in both areas. This regional variation has important implications for differences in household organization.

Finally, across the entire study area, changes in domestic structure size, burial practices, storage practices, and the interior segmentation of domestic structures from the Woodland to the Mississippian period suggest that households become increasingly autonomous over time and that relatively small extended or nuclear family households emerge as the basal unit of social organization. The timing and nature of this change plays out differently across the study area during the Mississippian period.

This first set of variables includes prominent architectural traits that would have been highly visible from the exterior of a structure: shape, size, orientation, and style of entryway. Given their high visibility, much of the variation observed in these traits may relate to their role in symbolic communication (see Blanton 1994:11–12). The size and shape of structures may reflect the size, activities, social position, and cosmologies of the social groups that inhabit them. The orientation of structures and their style of entryway may be determined by very prosaic causes (e.g., domestic structures in a linear settlement arranged parallel to a ridge), but they can also reflect shared norms of house building rooted in broader belief systems (Cunningham 1973; Hodder 1984).


I grouped structures in the database into five major shape classes: circular, oval-shaped, rectangular, square, and irregular. A small number of houses were T-shaped or keyhole-shaped. This categorization obscures some finer details of structure form but captures most of the variation. From a broad view, changes in shape track with breaks between the major chronological periods. There is a general shift from circular domestic structures in the Middle and Late Woodland to rectangular domestic structures in the Late Woodland and Early Mississippian, to square domestic structures and rectangular storage structures in the Late Mississippian, to circular and rectangular domestic structures in the Historic Indian period. There are deviations from this pattern at smaller scales. Table 1.1 records the shape of all the structures in the database, sorted by time period and functional class.

Circular structures occur in all time periods. There are some circular domestic structures in all periods except the Early Mississippian, but there are circular nondomestic structures in Early Mississippian times, including possible earthlodges and sweat houses. This shape is most common during the Middle Woodland period. Over 65 percent of Middle Woodland domestic structures in the database are circular. Only about 7 percent of all Early, Middle, and Late Mississippian structures are circular. They make up a low proportion of structures, mostly nondomestic, in the Mississippian components at Hiwassee Island, Toqua, Macon Plateau, Jewell, Bessemer, Martin Farm, and Cahokia, but are common at Town Creek, Rucker's Bottom, several sites in the Oconee River valley in piedmont Georgia, and possibly at Brasstown Valley, where the Late Woodland to Early Mississippian chronology is not entirely clear. During the Historic Indian period at Cherokee sites, circular-to-octagonal domestic structures, often called "winter houses," are usually paired with rectangular domestic structures called "summer houses."

Twenty-three oval-shaped structures (2 percent) were recorded at 13 sites dating to the Woodland, Early Mississippian, and Historic Indian period. Over half of these houses are from Owl Hollow phase sites (n=6) in Tennessee and the Late Woodland occupation at Brasstown Valley in northeast Georgia (n=8).

Rectangular structures (n=595, 47 percent) were the most common of all, appearing in all time periods, but most frequently in the Late Woodland, Mississippian, and Historic Indian periods. Approximately 80 percent of the Late Woodland and 60 percent of all the Early Mississippian structures are rectangular. Nearly half of the Historic Indian structures in the study are rectangular. Rectangular domestic structures are not as common during the Middle and Late Mississippian, but rectangular storage buildings commonly occur at Late Mississippian sites. Only four Middle Woodland structures are rectangular, and these all come from the Yearwood site, which may have been a special ceremonial center (Butler 1979).

Square structures (n=358, 28 percent) occur in all periods but are especially common in the Middle and Late Mississippian periods. Domestic structures in villages and domestic and nondomestic structures on mound summits are often square with rounded corners. Only seven square structures were recorded from Woodland period sites; one square structure of unknown function with rounded corners in the premound midden at Garden Creek, three square domestic structures at Rivermoore, and three roughly square domestic structures at the Yearwood site. Several square Connestee phase structures with rounded corners and a single support post, nearly identical to square house at Garden Creek, have recently been uncovered at the Iotla site at the Macon County Airport near Franklin, North Carolina (Tasha Benyshek, personal communication 2010).

Only four of the Historic Indian period structures were square; two large nondomestic public structures at Toqua, and single domestic structures at Yuchi Town and Tukabatchee. Three Cherokee domestic structures at the Townsend site described as octagonal closely resemble Late Mississippian domestic structures that are square with rounded or truncated corners.

The remaining structures (n=119, 9 percent) did not fit into these four broad shape categories. Most of these structures (n=97) have an undefined shape due to poor preservation or incomplete excavation. However, 22 structures had complex or unusual shapes, such as a single keyhole-shaped domestic structure at Kolomoki, two T-shaped, nondomestic structures at Cahokia Tract 15A, and semicircular structures at Yearwood. These structures do not share a common function (i.e., they are not all "ceremonial structures") and have to be understood within the context of individual sites.

In an early and influential attempt to link variation in house shape with variation in social and economic organization, Flannery (1972:22–44) used archaeological data from early villages in the Near East and Mesoamerica and cross-cultural data (especially from African herders and horticulturists) to argue that circular dwellings correlate with partially nomadic groups with extended, possibly patrilocal and polygamous families and shared storage. Rectangular structures, on the other hand, correlate with more sedentary societies composed of nuclear, possibly monogamous families with private storage facilities.

It is tempting to argue that a similar process plays out in the Southeast. There is a shift from round to rectangular structures, and increasing sedentism with the adoption of intensive maize agriculture after A.D. 800 (Anderson and Mainfort 2002:18). However, this model has been criticized as overly simple, and a recent study of early sedentary communities on the Anatolian plateau shows a poor fit between house shape and degree of sedentism (Steadman 2004:519–520). Flannery himself recently revised this model, placing less emphasis on the importance of house shape as a correlate of social and economic behavior, and more on the importance on a broader set of changes in house form that are related to the privatization of storage (2002:421). Similarly for the Southeast, structure shape is probably better understood as part of a suite of architectural changes related to changes in social organization and domestic production and consumption.


Structure size has probably received more attention in archaeological studies of households than any other architectural variable in the database, for practical and theoretical reasons. Variation in house size in a community is highly visible, broadcasting strong signals about individual and household status (Wilk 1983; Blanton 1994). House size is also easily recorded, and even the earliest and most cursory excavation records in the Southeast include planview maps of houses drawn to scale. I was able to record the floor area of 1,054 of the 1,258 structures in the database. Only structures with incomplete floor plans were not measured.

The reasons for variation in house size are complex, but archaeologists should not jettison house size as a useful variable for understanding households. A few recent studies point to the value in comparing the size of domestic structures within settlements. In his study of households at Tract 15A at Cahokia, Pauketat (1998:135–136) uses domestic structure size to argue for a shift from larger courtyard groups to smaller households as the basal unit of social organization. At Moundville during the Early Moundville I and Late Moundville I–Early Moundville II phases, Wilson (2008:87–92) compares house size within multihousehold residential groups. The distribution of domestic structures sizes are quite similar, suggesting that there are not major status differences between the residential groups. Hally's examination of Barnett phase domestic structures at the King site suggests that house size expands and contracts along with household size over the course of the domestic cycle (Hally and Kelly 1998; Hally 2008:271–279). Large, rebuilt domestic structures may represent the founding households at King, and their size may be a function of larger household size and special status in the community.

Despite the obvious differences in Woodland, Mississippian, and Historic Indian societies, there is value in treating all the structures in the database as a single population and examining the distribution of structures size. This broad view provides a background for comparison of domestic and nondomestic structures at finer scales and provides upper and lower size limits. The median size of all structures in the database is 28.4 m2 (n=1,054). The smallest is a 1.2 m2 circular Qualla-Lamar phase storage building at Brasstown Valley and the largest is a 300 m2 non-domestic Early to Middle Mississippian mound summit structure at Kincaid. The distribution of structure size is continuous and positively skewed; 75 percent of the structures measure 46.2 m2 or less and 95 percent fell below 103.7 m2.

Middle Woodland period domestic structures range in size from 7.1 m2 to 146.6 m2, with a median floor area of 40 m2 (n=66) (Figure 1.1). All but 7 domestic structures fall into a normal distribution centered on 40 m2 and ranging from 7 to 80 m2. It is notable that the seven largest domestic structures are represented by McFarland or Owl Hollow phase structures from the Yearwood, Duncan Tract, Banks III, and Banks V sites. The largest of these is a well-defined structure from the Banks V site with two large interior hearths and four central support posts, but several others in this larger size category are large circular domestic structures with few interior posts or features.

Overall, Late Woodland domestic structures range from 3 to 94.7 m2 with a median size of 8.4 m2 (n=101) (Figure 1.2). In contrast to the Middle Woodland period, the distribution of house size is positively skewed. All but one of the structures fall into a range of 3 to 48.9 m2. The single outlier is a large, nondomestic Teal phase building at Town Creek. An unusual keyhole-shaped domestic structure with a basin excavated at Kolomoki measures 7.5 m2.

There are notable regional differences in structure size during the Late Woodland period. In the American Bottom during the Late Woodland/Emergent Mississippian period, at Late Woodland sites in west-central Alabama, and in the Cairo Lowlands of Missouri, settlements are organized into clusters of closely spaced, small, rectangular domestic structures arranged around a central open area. These domestic structures range in size from 3.0 to 16.4 m2. They are so small that it is hard to imagine them inhabited by any group larger than a small nuclear household.

In northern Georgia, Late Woodland domestic structures at Brasstown Valley, Summerour, and Rivermoore are larger and more widely spaced, ranging in size from 17.8 to 48.9 m2. The spatial arrangement of these structures is similar to the arrangement of some Middle Woodland domestic structures at Hickory Log and McFarland. They are large compared to the small rectangular structures from sites to the west and are spaced farther apart in their settlements.

Keeping in mind the warning that house size is not always a direct reflection of household size, the architectural data seem to indicate two different forms of household organization. Late Woodland groups in the western part of the study area may be organized into courtyard groups made up of small, nuclear households. These courtyard groups may be the basic unit of social organization. In the eastern part of the study area, larger household groups, perhaps extended families, may occupy single structures, and these groups may be the basal social units.


Excerpted from The Archaeology of Houses and Households in the Native Southeast by Benjamin A. Steere. Copyright © 2017 University of Alabama Press. Excerpted by permission of The University of Alabama Press.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

Table of Contents

List of Illustrations ix

Acknowledgments xiii

Introduction 1

1 Patterns of Architectural Variability in the Native Southeast 16

2 Environmental Factors in Architectural Variation 64

3 Household Composition and Economics 85

4 Houses and Architectural Symbolism 111

5 Houses, Status, and Settlement 138

6 Conclusion: A Macroregional Perspective on Architectural Variation in the Native Southeast 176

Appendix: Description of the Architectural Variables 185

References Cited 191

Index 213

Customer Reviews

Most Helpful Customer Reviews

See All Customer Reviews