Time in Archaeology : Time Perspectivism Revisited [1 ed.] 9781607817925, 9780874809299

Citation preview

Time in Archaeology

Time in Archaeology: Time Perspectivism Revisited

Edited by

Simon Holdaway and LuAnn Wandsnider

The University of Utah Press Salt Lake City

© 2008 by The University of Utah Press. All rights reserved. The Defiance House Man colophon is a registered trademark of the University of Utah Press. It is based on a four-foot-tall, Ancient Puebloan pictograph (late PIII) near Glen Canyon, Utah. 12  11  10  09  08    1  2  3  4  5 Library of Congress Cataloging-in-Publication Data Time in archaeology : time perspectivism revisited / edited by Simon Holdaway and LuAnn Wandsnider.    p.  cm.   Includes bibliographical references and index.   isbn 978-0-87480-929-9 (cloth : alk. paper)  1. Archaeology— Methodology—Philosophy.  2. Time.  3. Time perspective. 4. Archaeology and history.  I. Holdaway, Simon.  II. Wandsnider, LuAnn.   CC75.7.T558 2008   930.101—dc22 2008005214 Printed and bound by Sheridan Books, Inc., Ann Arbor, Michigan Interior printed on recycled paper with 50% post-consumer content.

I think metaphysics is good if it improves everyday life; otherwise forget it. Robert M. Pirsig, Zen and the Art of Motorcycle Maintenance

Contents

List of Figures   ix List of Tables   xi Preface and Acknowledgments   xiii 1. Time in Archaeology: An Introduction   1 Simon Holdaway and LuAnn Wandsnider 2. Time Perspectivism: Origins and Consequences   13 Geoff Bailey 3. Time Perspectivism and the Interpretive Potential of Palimpsests: Theoretical and Methodological Considerations of Assemblage Formation History and Contemporaneity   31 Alan P. Sullivan III 4. Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation   46 Michael J. Shott 5. Time-Averaged Deposits and Multitemporal Processes in the Wyoming Basin, Intermontane North America: A Preliminary Consideration of Land Tenure in Terms of Occupation Frequency and Integration   61 LuAnn Wandsnider 6. Investigating Persistent Places in the Northern Great Plains, Central North Dakota   94 Mathew A. Dooley 7. Assemblage Accumulation as a Time-Dependent Process in the Arid Zone of Western New South Wales, Australia   110 Simon Holdaway, Patricia Fanning, and Ed Rhodes 8. Time Averaging and the Structure of Late Pleistocene Archaeological Deposits in Southwest Tasmania   134 Nicola Stern vii

9. Time Perspectivism and the Structure of Archaeological Records: A Case Study   149 Josara de Lange 10. No Time like the Present   161 Philip J. Arnold III 11. Paradigms and Metaphysics, or “Is this the End of Archaeology as we Know It?”   170 Tim Murray References Cited   181 Index   203

viii

Figures

3.1. Location of the Upper Basin Archaeological Research Project study area in northern Arizona   34 3.2. Plans and profiles of four burned structures at Site 17, which has been tree-ring dated to ad 1054–1064   35 3.3. View of a portion of Structure 4 at Site 17   36 3.4. Histograms of structure-specific tree-ring dates from Site 17   37 3.5. Plan of Structure 1 at Site 17 showing abutments of two parallel walls, placement of five ceramic vessels in and on the lip of a deep thermal feature, and locations of several other floor-contact artifacts   38 3.6. Bar chart of the assemblage composition of surface cumulative palimpsests for 242 masonry structures in the Upper ­Basin   39 3.7. Bar chart of mean artifact density of surface cumulative palimpsests for 242 masonry structures in the Upper Basin   40 3.8. Box-and-whisker plot showing the median value, range of 50 percent of the cases, and outlier values for artifact density, ceramic artifact density, and lithic artifact density of surface cumulative palimpsests associated with 242 masonry structures in the Upper Basin   41 3.9. Bar chart of lithic assemblage composition at Site 17 illustrating differences in variability between spatial palimpsests that arise

when either a context-specific or a site-wide denominator is used to calculate percentages of the various artifact classes   41 3.10. Plan of Structure 3 at Site 17 showing arrangement of artifacts on the floor at time of abandonment   42 3.11. Plan of Structure 4 at Site 17 showing arrangement of artifacts on the floor at time of abandonment   44 4.1. Possible reduction relationships between Lower Paleolithic tool types   50 4.2. Log10 tool counts versus log10 assemblage size for scatters and patches   53 4.3. Log10 tool proportions versus log10 assemblage size for scatters and patches   54 4.4. “SHE” plot of Lower Okote Member data resembling Hayek and Buzas’s log series model   58 5.1. Wyoming Basin study sites   66 5.2. Pit structure dimensions by period and location   77 5.3. Fire-cracked rock versus debitage densities by location   77 5.4. Mean hearth area versus hearth density by location   79 5.5. Mean hearth area versus external hearth density by (a) estimated minimum number of occupation events, (b) surface stability, (c) pit structures and site structure, and (d) seasonal indicators   82 5.6. Mean hearth area versus external hearth density by (a) relative debitage density ix

List of Figures







z-score, (b) relative fire-cracked rock density z-score, (c) relative ground stone density z-score, and (d) relative minimum number of individuals density z-score   84 5.7. Mean hearth area versus external hearth density by (a) the sum of relative firecracked rock (FCR), ground stone, and minimum number of individuals (MNI) density z-scores; and (b) sum of chipped stone, FCR, ground stone, and MNI density z-scores   86 5.8. Occupation frequency and integration   87 5.9. Mean hearth area versus external hearth density by water permanence   92 6.1. Location of study area   98 6.2. Frequency diagram showing the difference between siltation and lichening for individual features   101 6.3. Geographic information system operations used to quantify persistence   103 6.4. Persistence index   104 6.5. Comparison of the number of culturalbearing­excavation levels by derived persistence category for a sample of 57 1-×-1-m test units   106 6.6. The position of late features with respect to the position of early features   107 6.7. Comparison of persistence and the location of early features   108 7.1. Western New South Wales, showing locations mentioned in the text   112 7.2. The deflated remains of a heat-retainer hearth   113 7.3. Fowlers Creek hearth age estimates   118 7.4. Fowlers Creek hearth locations with age estimates   120

7.5. Nundooka hearth age estimates   121 7.6. Mulga Dam hearth age estimates   121 7.7. Hearth age estimates for the last 1,000 years from Fowlers Gap   122 7.8. Raw material proportions calculated in three ways for each of the sampling locations   125 7.9. Flake-to-core ratios   126 7.10. Complete, noncortical flake-to-cortical flake ratio for each raw material class   126 7.11. Noncortical core-to-cortical core ratio for each raw material class   128 7.12. Ratio of complete flakes to complete tools, by raw material class   128 7.13. Tool type proportions for each raw material class   129 7.14. The ratio of complete scraper surface area to complete flake surface area   130 8.1. Mackintosh shelter and Nunamira Cave, located in southwest Tasmania, ­Australia   136 8.2. Mackintosh shelter, showing area excavated   137 8.3. Nunamira Cave, showing area excavated   138 8.4. Archaeological deposit at Mackintosh shelter   139 8.5. Stratigraphic sequence at Nunamira Cave   143 9.1. Klithi and other sites mentioned in the text in their regional setting   151 9.2. The layout of excavation trenches within the Klithi rockshelter   152

x

Tables

4.1. Tool Counts in Original and Combined Tool Classes in Lower Okote Member Assemblages   52 5.1. Model of Place History and Taphochronometric Indicators   63 5.2. Processes by Length of Term over Which They Are Manifested   65 5.3. Wyoming Basin Cultural Chronology   67 5.4. Wyoming Basin Excavated Components by Basin Location   68 5.5. Deposit Interpretation by Site, Strata, and Components   71 5.6. Components by Surface Stability and Basin Position   70 5.7. Minimum Component Grain   75 5.8. Taphochronometric Indicators Derived from Assemblage Information, by Component   78 5.9. Summary of Occupation History by Site and Time Period   89

5.10. Components by Occupation History and Surface Activity   91 5.11. Multiple Radiocarbon Dated Components by Number of Radiocarbon Events and Nature of Integration   91 7.1. Optically Stimulated Luminescence Age Determinations on Quartz Extracted from Sedimentary Units Underlying Surface Artifact Scatters at Three Locations at Fowlers Gap in Western New South Wales, Australia   117 7.2. Interpretation of Bayes Factor Model 2 versus Model 1   119 8.1. Radiocarbon Determinations for the Archaeology-Bearing Deposits at the Mackintosh Shelter   142 9.1. Three Possible Scenarios for the Depositional History of a Hypothetical Assemblage Spanning 500 Years and Containing the Remains of Both Adult Animals and Young and Old Animals   155

xi

Preface and Acknowledgments

Merriam-Webster Online (accessed September 1, 2007) identifies 14 different meanings for time as a noun. (In contrast, culture, the noun, is listed with five different meanings.) Thus, for the time being, we students of time must devote considerable time to establishing the timing of past events, determining the elapse of time, and learning about past times. In archaeology, time is used to convey a similarly wide range of meanings, with common usage in just a couple of senses, as testified to by the titles of recent books on the topic: The Archaeology of Time; It’s about Time; Time and Archaeology; Time, Culture and Identity; Time, Tradition, and Society in Greek Archaeology; Time, Process, and Structured Transformation in Archaeology; Measuring Time with Artifacts; Picking the Lock of Time; and Time, Trees, and Prehistory. The point of departure for this volume lies in the post-tumultuous­ times of processual archaeology, in the 1970s and 1980s, when Geoff Bailey, Lewis Binford, David Clarke, Robert Dunnell, Robert Foley, and Michael Schiffer, among others, initiated a deconstruction of time as used in archaeology. This deconstruction, part of which has come to be known as “time perspectivism,” was closely tied to what was at that time an investigation into the dimly understood properties of archaeological phenomena. In this volume, we attempt to further this deconstruction. Time in Archaeology was originally convened as an electronic symposium held at the Milwaukee Society of American Archaeology meetings in

2003. Early versions of the chapters in this volume were posted on a Web site before the meetings, and participants as well as the audience spent the allocated meeting time discussing ideas put forward in the papers and time perspectivism as well as time and archaeology in general. The papers were subsequently revised as a result of discussion at the meetings and then again following the outcome of the referee process. The result is a tightly focused group of essays that provide a historical background to the development of the idea of time perspectivism as well as a range of case studies that illustrate where scholars have taken the ideas. We owe a special vote of thanks to Anne Ramenofsky and Julie Stein, who were instrumental in helping this book get off the ground. An anonymous reviewer also provided helpful comments on the volume. Intellectually our interest in time in archaeology was fostered by Tim Murray, at La Trobe University, and Lewis Binford, then at the University of New Mexico. Murray kindly agreed to write a summary chapter for the volume, and Binford was actively involved in the session at the Mil­waukee meetings. Geoff Bailey provided much encouragement that helped to ensure this book was completed. Dorothy Brown helped with initial proofing of the volume, and some of the illustrations were ­redrawn by Peter Quinn and Seline McNamee. John Herbert and the staff at the University of Utah Press ensured that production of the volume ran smoothly.

xiii

1

Time in Archaeology: An Introduction Simon Holdaway and LuAnn Wandsnider University of Auckland and University of Nebraska–Lincoln

Lifeway reconstruction is listed as one of the objectives of “World Prehistory,” the ubiquitous course taught in universities and colleges the world over (e.g., Fagan 1995:8). It complements well the other subdisciplines of anthropology, at least for beginning anthropology students, offering them a familiar approach to foreign material: if cultural anthropologists study the behavior of present-day (or at least near-to-present-day) peoples, then archaeologists may be expected to deal with peoples’ behavior from the past. Certainly, some archaeologists study the past aided by textual records, and some cultural anthropologists are interested in past historical experience. But this overlap only enhances the perceived integration of approaches. The clear message is that archaeology is about doing the ethnography of the past. The problem is that our cultural anthropology colleagues have changed the way they do ethnography. The postmodernist critique has laid bare the fictive nature of the objective anthropological experience. Ethnographies tell a story from a particular point of view that is only one of a range of understandings of why things happen. What, then, is the status of the archaeologists’ lifeways reconstruction? To some, particularly the more radical members of the postprocessual archaeology of the 1980s, all archaeological reconstruction was seen as theory dependent and therefore subjective. Lifeway reconstruction, therefore, was held to reflect as much about the society from whence the archaeologist originated as it reflected a reality experienced

by people in the past. And from the late twentieth century, the indigenous voice can be added. No longer do archaeologists have a monopoly on explaining what went on in the past. There are competing views and multiple lifeway reconstructions. As archaeologists, we are being openly challenged to defend the veracity of our reconstructions (e.g., Bender 2002). The postprocessual critique has been well rehearsed in a variety of monographs and edited ­essay collections, and we do not intend to add this volume to the stack. Rather, the authors collected herein wish to address the question of meaning in the past from a different tack, one that we develop by taking inspiration from articles written in the early 1980s by Bailey, Binford, and others grouped here under the term time perspectivism. As Bailey defines the term in chapter 2, time perspectivism treats all archaeological material records as palimpsests and asserts that there is a relationship between the scale at which such records can be resolved and the types of research questions they can be used to answer. That different explanations of the past are possible depending on the temporal scale at which past human behavior is viewed is hardly a new point or one that has been cast aside since Bailey and Binford published their seminal papers (e.g., Ramenofsky 1998). Other theoretical approaches such as historical ecology (e.g., Balée 1998) and ­Annaliste treatments (e.g., Bintliff, ed. 1991; Knapp 1992) have insisted on multiscalar views of the past. 1

Simon Holdaway and LuAnn Wandsnider What sets time perspectivism apart from other approaches, however, is the insistence on readings of the archaeological record as a unique historical data set on which to base multiple scales of explanation. It is the rise of formation studies over the last 30 years that has provided the means to view the archaeological record in this way. The authors of this volume seek explanations of the past that conform to our understanding of how the archaeological record was formed while at the same time dealing with deposits as palimpsests and seeking explanations that are scale dependent.

To be sure, archaeologists have kept up with and adopted many of the advances in social theory. But in seeking to make archaeology relevant, they have in many instances left the archaeological record behind. Archaeological explanation is often treated as just another form of social explanation, the difficulty of which should immediately be obvious to someone standing in front of a midden, eroding house wall, or deflated hearth. The danger archaeologists face is that in failing to emphasize the archaeological nature of our perspective on the past and our perspective on explanation, archaeology will fast become an irrelevance (van der Leeuw and Redman 2002). Why should indigenous people or anyone else consider our explanations as valid in their own terms, if we couch them in contemporary social theory while failing to convey that the archaeological basis for explanation is quite different from that provided by contemporary “human time” (Stein 1993) observation? Of course there are exceptions to the blanket criticisms made in the paragraphs above. Some archaeologists have considered the relationship between the formation of the archaeological record and the nature of archaeological explanation. The authors in this volume draw inspiration from a series of essays by Geoff Bailey (1981, 1983, 1987) published in the early 1980s and a series of essays published by Lewis Binford (1977a, 1978a, 1980, 1981a, 1981b, 1982, 1983a) during the same time period. Bailey (2007, this volume [ch. 2]) has provided his own introduction to the genesis of his ideas. Similarly, Murray pursues the intellectual trajectories of time perspectivism in chapter 11, using the term introduced by Bailey. Murray (1999a) has also recently written on Binford and time within the context of the “Pompeii premise” debate with Michael Schiffer. Despite the lingering interest in time perspectivism by Murray and others, it must be said that both Bailey and Binford ultimately failed to provide programmatic statements that inspired a new body of research, something that Bailey addresses in this volume. What their work lacked was a clear method for implementing the theoretical insights they developed. There are therefore two objectives for this book: to demonstrate that the problems

The Archaeological Record If archaeological explanations are to be taken seriously, on a par with, rather than replacing, other kinds of explanations of the past, then we need to be clear how our explanations are derived from the archaeological record. Archaeologists have spent a great deal of effort searching the theoretical literature to learn what drives humans to act the way they do. Much of this searching has ranged across the social sciences, often delving into studies conducted over the short term, using observational scales rarely exceeding the lifetime. To what extent do these studies actually engage the content of the archaeological record? The need to show that archaeology may be used to study the same types of phenomena as those studied by social scientists when dealing with contemporary peoples seems to have largely overtaken the need to answer this question. As Yoffee and Sherrat (1993) comment, archaeology alone among the social sciences has failed to build its own social theory. The contemporary social theory of other disciplines, first seen as a source for explanatory inspiration, has in some cases become a prescription for how archaeology should be undertaken. Shennan (1989), when retrospectively reviewing the impact of Binford and Binford’s New Perspectives in Archeology (1968) and Clarke’s Analytical Archaeology (1968), makes it clear that this charge is not unique to postprocessual archaeology. New Archaeology’s initial interest in culture process rapidly gave way to interests in social, ecological, economic, and ideological processes, isolating what to Clarke was unique about archaeology. 2

Time in Archaeology: An Introduction identified in the early literature have not gone away and to illustrate, through a series of case studies presented in the chapters that follow, a set of methods that can be applied to overcome these problems and thereby reinstate time perspectivism in the agenda of archaeological theoretical discourse. In this introduction, we review time perspectivism and provide a brief intellectual history of time in archaeology, indicating why we have brought together a group of authors to talk about their ideas for an archaeological concept of time derived from formation studies of the archaeological record. In so doing, we provide an introduction to the chapters that make up this volume, as well as illustrating in a little more depth our reaction to the topic of lifeways with which we opened.

resent individual items accumulated at the time of manufacture, construction, and initial use, they are also reflective of the reuse and redeposition of artifacts as well as the reoccupation of places by a variety of peoples for a variety of purposes. Features, for example, show the accumulation of instances of refurbishment, destruction, and reconstruction (M. E. Smith 1989), whereas artifacts may be reduced through wear or resharpening, acquiring traces that reflect their use-life histories (sensu Sullivan 1978). The studies that emphasize the significance of these processes for interpreting archaeological materials have largely developed since Bailey’s, Binford’s, and Foley’s seminal essays. These studies have allowed a new generation of archaeologists to develop ­methods that allow the application of time perspectivist ideas. Most archaeologists would accept the importance of site and artifact reuse, but although ethnoarchaeological studies are widely directed at investigating the manufacture, use, and abandonment of artifacts and features, when such use-lives are considered at all, they are, rather ironically, often synthesized to construct long-term conditions that show little or no temporal change. The various strategies—mobility, technological, settlement, organizational, behavioral—that archaeologists increasingly turn to as explanatory devices (Hegmon 2003; Holdaway and Wandsnider 2006) are typical examples. These strategies invariably take stability over some span of time as a given. The great time depth offered by archaeology is often vaunted (e.g., Hodder 2001, introducing Mithen 2001 and Meskell 2001). Apart from an extension or refinement of the chronometry of human prehistory, however, the outline of significant events in general archaeology has changed little over the decades since the radiocarbon “revolution” (Dunnell 1982). Much explanation continues to be, in Dunnell’s terms, proximate and functional and, therefore, timeless. Despite a wider range of tools with which to assess the palimpsest-like nature of the archaeological record that Bailey and Binford discussed, little progress has been made in understanding how the life histories of the artifacts and features that form such a palimpsest might influence the nature of archaeological inference and therefore the

Time Perspectivism Time perspectivism was formulated around the idea that observations made at different temporal scales differentially make different processes apparent. Applied to the archaeological record, time perspectivism provides an alternative to the view that the vagaries of preservation provide for only an incomplete account of past (“human time” [Stein 1993]) behavior (Bailey 2007, this volume). Bailey, Binford, and Foley (1981a, 1981b, 1981c) independently developed variations on this idea at much the same time. It was clear to all three scholars that archaeological deposits in the main represent the remains of repeated events and therefore offer the opportunity of studying processes operating at temporal scales longer than an event (e.g., Bailey 1981; Binford 1981a; Foley 1981a). But what was less clear was how this observation could be applied to archaeological remains. Both Bailey (1983) and Binford (1977a) were interested in what they termed methodological (or conceptual) uniformitarianism. If stone artifacts or animal carcasses are reduced in nonrandom ways and can be shown to have clear material signatures identifiable over the short term, and if a methodological uniformitarianism based on the observations can be sustained, then archaeological relationships can be interpreted (see also DeBoer and Lathrap 1979). This was clear enough in the late 1970s and 1980s. But while archaeological deposits rep3

Simon Holdaway and LuAnn Wandsnider u­ ltimate goals of an archaeological interpretation of historical processes.

erties of the archaeological record (the way various time-dependent processes are responsible for artifact deposition) and ultimately a much greater understanding of the kinds of questions that may be asked of this record and the kinds of explanations of the human past it supports. The temporality of deposits formed a key aspect in Clarke’s (1973) well-known exposition of the New Archaeology in Britain, a statement picked up by Sullivan (1978, 1995a) in the United States. Equally important was an early essay by Isaac (1972) wherein he suggested that the long time depth represented in the Paleolithic record might require a different type of explanation than the culture histories being written for more recent periods. Binford (1981a), Foley (1981c:8–9), and others followed these leads, arguing that patterns in artifact densities are a product of repetitive behaviors maintained over long time periods reflecting stable configurations of humans, artifacts, and the land surface and emphasizing the taphonomic nature of archaeological deposits through studies of the various ways in which objects accumulate. For Foley (1981b:​173), all archaeological deposits are palimpsests that vary only in the scale at which they may be interpreted. His (1981c:​180) off-site approach was directed at providing spatial rather than chronological information relating to past behavior, with the aim of understanding long-term land use in relation to resource distribution. Binford’s theoretical interests were directed slightly differently. In the now famous article introducing foragers and collectors, he (1980) related the development of palimpsest deposits with different histories to different types of mobility among hunter-gatherers. Specialized versus generalized palimpsests were discussed with reference to the Mask site (Binford 1978a) as well as Nunamiut seasonal camps (Binford 1978b). In addition, Binford’s understanding of the archaeological record is clear from two further concepts. First, he (1978a, 1978b, 1980) discussed the temporal significance of geomorphological processes of archaeological site formation by using the concept of the temporal grain of deposits, the degree to which behavioral events might be resolved within a deposit. Second, he (1981b) introduced the concept of historical

The Taphonomic Metaphysic Paynter (2002), correctly in our view, argues that, whatever the original goals of the New Archaeology, its application quickly descended into a synchronic, functionalist interpretation of the archaeological record. Nevertheless, there remain several developments in the archaeological literature beginning in the early 1970s that either directly or indirectly addressed the status of the archaeological record as a medium through which to develop historical explanations. We believe that, viewed with the hindsight of history, these studies, though never forming a recognizable alternative to the processual and postprocessual bodies of literature, are sufficiently coherent to be labeled the taphonomic or formational metaphysic and provide a methodological door through which time perspectivism can be approached. This metaphysic began to cohere, we suggest, with publication of Michael Schiffer’s (1972) seminal article on archaeological and systemic context. Here, Schiffer effectively promoted the study of the archaeological record and, with his colleagues, subsequently went on to develop behavioral archaeology (Reid et al. 1975). The behavioral archaeologists extended archaeological research to explain the full breadth of relationships between human behavior and material culture in all times and places (Schiffer 1995:ix). Although Schiffer was criticized for attempting lifeway reconstruction (Binford 1981a), albeit from a perspective that offered a detailed consideration of nonfunctional sources of variation (Murray 1999a), behavioral archaeology emphasized Cartesian time in a way that had few precedents. Cartesian views of time assume that objects have both a position in space and a trajectory through time independent of other objects. Behavioral archaeologists reflected this view by studying artifact life histories and describing the complex ways artifacts moved back and forth from Schiffer’s systematic and archaeological contexts through time (e.g., DeBoer 1974). From this came a much better understanding of the temporal prop4

Time in Archaeology: An Introduction i­ ntegrity to describe the similarities and differences in the conditions that led to the formation of an archaeological deposit. These studies culminated in what might be described as a landscape perspective using the “rocks with eyes” analogy (Binford 1983a). For Binford (1982, 1983a), archaeological deposits result from the actions of many generations of individuals, all of whom abandoned artifacts and features as epiphenomena of a collective long-term behavioral system in which they were involved, the definition of which forms the goal of archaeology. Extensions of this landscape perspective by other authors quickly followed, with the introduction of the term place use histories to describe the differing sequences of deposition and different geographic locations (Camilli 1983, 1988; Camilli et al. 1988; Sullivan 1992a; Wandsnider 1998). Dewar and McBride (1992) discussed remnant settlement patterns and introduced the concepts of spatial contiguity and temporal continuity to discuss place occupation through time. Kelly (1988), in following these ideas, uniquely included geological criteria such as surface deflation and stability in his attempts to describe depositional history. It was Stern (1993, 1994a), however, who most clearly articulated the relationship among depositional history, time averaging as described in the geological (paleontological) literature, and observations on time perspectivism offered by Bailey (1981, 1983, 1987; see also Bailey 2007; Murray 1997, 1999a, 2002). For Stern, the archaeological record is to be seen as a time-averaged material sample of the remains of past human activity. Like many paleontological deposits, time-averaged archaeological deposits are formed over prolonged periods of time such that items (artifacts or fossil organisms) found within a single deposit may originate from a variety of different cultural systems (or habitats). In applying these concepts, Stern elaborated on Binford’s notion of historical integrity, noting the timeaveraged­nature of fluvial deposits from Koobi Fora that incorporated materials derived from a variety of landscape features with different temporalities. Because of the temporal complexity of the deposits, Stern argued that many of the high-resolution behavioral, that is, “lifeways,” interpretations of ar-

chaeological deposits with hominid remains and artifacts are inappropriate. Not only must the temporality of the interpretation be matched to the temporality of the deposit, but time-averaged deposits do not represent an average in time (see de Lange, this volume; Stern, this volume). It is not a question of trying to match a single behavioral scenario to explain the formation of an archaeological deposit. Rather, time-averaged deposits are better thought of as the summation of materials derived from a variety of behaviors and contexts. Thus, lifeway reconstructions cannot be made as though materials derive from an “average” of behavior. In treating the archaeological record as a timeaveraged sample composed of items that do not necessarily share a common depositional history, Stern espoused a view of the archaeological record close to that advocated by geoarchaeologists (as well as some others). DeBoer (1983), for instance, drew a parallel between archaeology and paleontological taphonomic studies, arguing that the complexity of formation processes would disallow an isomorphism between the archaeological and systemic contexts. Stein (1987) proposed that archaeological deposits should be seen as an aggregate of sedimentary particles, each particle having its own particular history. According to this view, the record is an accumulation of separate particle histories, only some of which are the result of a single process. Dunnell (1992) commented favorably on this approach while promoting his evolutionary view of archaeology and drawing a contrast between functional ecological explanations and historical explanations of how things come to be (Dunnell 1980). For Dunnell (1982), materialist approaches to historical causation offer the only hope for a truly evolutionary archaeology. In what he described as a time-like reality, things are always in the process of becoming, and relations between observations are dependent on both time and place. Thus, the goal of evolutionary archaeology continues to be not an understanding of the archaeological record as a series of essentially timeless behavioral scenarios but, rather, to seek cause in the selection of attributes from a pool of continuous variation. For many, whereas the explanatory framework of 5

Simon Holdaway and LuAnn Wandsnider e­ volutionary archaeology seems overly restrictive (but see Shennan 2002), a materialist rather than essentialist metaphysic has much appeal. Ramenofsky (1998), for instance, has discussed the need to match the temporal scale at which a research question is pitched with the data observed at an appropriate scale. She argues that a materialist view is most consistent with such a multiscale approach. As referenced above, those who follow the Braudelian Annales scheme see virtue in searching for conjunctures, the interaction of processes that occur over the short term (involving individuals and events), the medium term (periods spanning socioeconomic and demographic cycles), and the long term (environmental changes [Bintliff 2004]). Smith (1992), in discussing the Annales approach, links the Braudelian scheme to Butzer’s (1982) configuration of ecology and systems theory, thereby overcoming Braudel’s static view of the environment. Like Braudel, Smith sees the interaction of processes operating at these different scales as critical for interpreting the past. For Smith, however, the distinction between the essentialist and materialist metaphysic discussed above is something of a red herring. Smith argues that periodization is unavoidable because it is not possible to study continuous change. Therefore Smith feels that chronology poses methodological rather than theoretical problems and is best addressed by refining chronology into as small a set of units as possible, the equivalent of Braudel’s short-term events. On its own the identification of short-scale events in the archaeological record is not problematic. Lucas (2005:48), for instance, argues persuasively that these are manifest in the deposition of single artifacts and burials. But where both Lucas and Smith fail is in their inability to demonstrate how the time-averaged nature of the archaeological record is to be addressed when these events are combined into assemblages. Whereas short-term events may be easily defined, the medium- and long-​term patterns are only manifest at the assemblage level, defined according to Stern (1994a, this volume) by the minimum chronological resolution of the deposit as a whole. It is not at all clear how these patterns are formed from the individual events that accumulate to form the assemblage and therefore

how they may be interpreted using conventional social theory. The problem that both Smith and Lucas face is made clear by Blake’s (2003) analysis of Byzantine-era reuse of Sicily’s prehistoric rockcut tombs. Rather than giving an explanation for the reuse of these sites based on an attempt by the ­Byzantine-era people to identify with a prehistoric past (and hence the medium to long term), Blake argues that “this phenomenon of reuse resulted from a fortuitous convergence of the older site’s familiar presence and new pan-​Mediterranean­ cultural currents” (2003:​218). In this case, documentary evidence is sufficient to demonstrate the absence of a link between the short and medium to long term; however, the situation is not always this clear. The temptation is always to stretch the linkages and construct a narrative account that forges links between temporal scales without sufficiently demonstrating their presence or, indeed, as Blake was able to do, their absence.

Anthropology, Archaeology, and Time Anthropologists are interested in the temporality implied in telling time. “Going slow” does not just mean working less quickly but has a definite social implication as well (Munn 1992). The classic time questions in anthropology refer to the social conceptualization of time (Gell 1992). Recent essays edited by Karlsson (2001) and by Gingrich and colleagues (2002), for instance, discuss a variety of approaches to time. Artelius (2001) considers Viking conceptualizations of time, relating these to Viking resistance to Christianity and its foreign temporal concepts. Damm (2001) comments on the disjuncture between the timeless traditional stories of the Bugakhwe of Botswana and the new, historical stories they now recognize as being necessary to argue claims to land. Schiefflelin (2002) provides examples on linguistic usages that deliver temporal signals, and Paynter (2002) investigates time in terms of the various narratives told about a particular place. Bender (2002) discusses landscapes as a multivocal entity associated with different views of the historical past. A few archaeologists have taken their lead from studies like these and investigated aspects of time 6

Time in Archaeology: An Introduction that are specific to certain cultures and worldviews (e.g., chapters in Murray, ed. 1999; see also Lucas 2005 and van Dyke and Alcock 2003 on the archaeology of memory). Bradley (1998, 2005) has developed the archaeology of time in a slightly different direction focusing on monuments. Using a distinction made by Rowlands (1993) between memories in the form of monuments as inscriptions versus memories that are incorporated through the use of monuments (i.e., through practice), he (1998:90) argues that despite the acknowledged imprecision of archaeological chronologies, regularities exist in the archaeological record because societies maintained rituals over long periods of time. More recently Bradley (2005:chapter 7) has developed this idea using Renfrew’s (2001) claim that the major change in human society occurred not with the advent of modern humans but with the beginnings of sedentary life. For Bradley, the shift from huntinggathering to the sedentary exploitation of domestic plants and animals marked a fundamental change in the nature of the archaeological record, one that is manifest both in the nature of ritual practice and in the creation of more temporally stable patterns in the archaeological record compared with the record created by hunter-gatherers. Whether or not the record of hunter-gatherers is truly different from that of those with a domesticated economy is of course open to debate, but from a time perspectivist position the worry is that the apparent stability and relative ease of interpretation of the record of sedentary peoples are an artifact of the presence of long-lived structures and the temptation to interpret activity in relation to these structures in synchronic terms. In much the same way that the hunt for disjuncture in the Annales approach promotes seamless transtemporal narrative (e.g., Bintliff 2004) rather than interrogation of such (à la Blake 2003), it is very tempting to look, for instance, for activity zones within a house structure that has a longevity measured in decades or more. Patterns that appear are, of course, the outcome of palimpsest deposits and therefore little different from the medium-term accumulation of artifacts deposited by hunter-gatherers on a surface. Thus, easily apparent pattern may be as much the result of the presence of interpretable structures as an out-

growth of a change in the nature of ritual behavior. Nevertheless, Bradley’s observations help explain why the examples discussed in the chapters here are largely those of hunter-gatherers rather than more sedentary peoples. Because hunter-gatherer groups by and large lack permanent structures, the palimpsest nature of the archaeological record they create is much more apparent and much harder to gloss using synchronic narrative. Therefore, archaeologists, including a number of authors in this volume, have to some degree been forced to look for alternative types of explanation, particularly forms that are less reliant on contemporary social theory. In a now classic article, Ingold (1993) discussed the relationship between what he termed the “taskscape” and the landscape. The taskscape consists of an interconnected set of tasks, and the landscape is an array of related features. But rather than keep these as separate entities, Ingold argued that with the understanding of the landscape as fundamentally temporal comes the notion that human activities must be seen as nested within “the wider pattern of activity for all animal life, which in turn nests within the pattern of activity for all so-called living things, which nests within the life-process of the world” (1993:​164). This is clearly a call for the type of multiscalar explanation advocated by Ramenofsky (1998) and Smith (1992), among others. Such explanation, however, has not so far been informed by the “formational metaphysic” discussed above. Multiscalar, materialist explanation provides a unique archaeological view on human history that is most effective when combined with a sophisticated understanding of the time-averaged nature of the archaeological record and interpretations derived from the last 30 years of formational studies. As Murray (2004) indicates, too often archaeologists have taken the easy “out” and, in the face of a palimpsest record, effectively ignored time and sought a synchronic, functional explanation that denies a historical past. The problem is not the lack of formation studies; although there are no doubt more such studies that need to be undertaken, the message is clear enough that there exist no simple relationships between the archaeological record and the nature of behavior that created this record. Rather, the 7

Simon Holdaway and LuAnn Wandsnider p­ roblem lies in the ease with which the message can be dismissed in the search for “higher” social ex­planation (read “lifeway reconstruction,” even if couched in the guise of, for example, an evolutionary behavioral strategy or the manipulation of agents). Formation studies are often acknowledged as important, but they are used as a way of removing the “noise” from patterns that can then be directly assessed with behavioral or social theory. Why are archaeologists so disinterested in time, and why have they resisted the implications of the formational metaphysic? Paynter (2002) sources the problem to a processual archaeology that was not concerned with chronology but only with documenting change from one steady state to the next. Dunnell (1980) makes much the same criticism, noting that the historical explanation favored by Binford (1962) at the dawn of the New Archaeology quickly came to be replaced by a synchronic functionalism. Murray (1987, 1993, 1997), however, places the problem much earlier, shortly after the foundation of archaeology as a discipline in the mid-nineteenth century. Here, he argues, the founding fathers of archaeology were faced with the need to develop an explanation for humanity’s past, greatly extended in time. They chose contemporary ethnographic analogy and, with it, essentially timeless synchronic explanation, a temporal scale of interpretation that modern archaeology has been unable to shed. He argues that rather than accept a consensus position on archaeological theory, it is time that archaeologists acknowledge the antiquity of their own explanatory framework and consider a range of alternatives. Most archaeologists consider studies of the archaeological record to be important but tend to separate these studies into a methodological category, separate from theoretical concerns (e.g., Hegmon 2003). From the earliest days of the New Archaeology, it was not the archaeological record that was thought to limit the types of questions that could be addressed but, rather, the ingenuity of the archaeologist in deriving ways to interpret this record. Thus, though the New Archaeology, and particularly behavioral archaeology, led to a concern with how the record formed, with rare exceptions (e.g., Plog 1973, 1974), the nature of the record was

seen as a methodological rather than a theoretical challenge. Not everyone views the record in this way. Some archaeologists, particularly ­geoarchaeologists, like their colleagues in geomorphology, paleontology, and ecology, accept that the nature of deposits, archaeological or otherwise, influences the analytical and therefore interpretative scales at which the past is viewed. The case is most clearly stated among ­paleontologists, where the long time spans and range of taphonomic processes combine to limit the degree to which fossil deposits may be interpreted as the ancient remnants of biological communities (see de Lange, this volume). Paleontological deposits are time averaged; they combine within a single unit materials that derive from a number of potentially unrelated events. The minimum chronological unit used to interpret these events does not represent the time span of the events themselves but, rather, that of the unit in which they are deposited. In many cases this time span will extend well beyond that typically associated with observations of organisms existing together within an ecosystem. In the same way that paleontologists have questioned the relationship between fossil populations and ecological communities, archaeologists need to ask whether behavioral observations derived from short-term ethnographic observations relate to the long-term and palimpsest nature of most archaeological records. Numerous authors in this volume deal with deflated records where the question of time averaging is made much more obvious through the lack of stratigraphy in the conventional sense. With buried deposits, it may be easier to maintain the fiction that material from the same layer was deposited at the same time or at least as a result of the same occupation. That this is only very rarely so is increasingly apparent as a result of geoarchaeological studies that critique not only the simple equation of stratigraphic changes with changes in occupation type but even the cultural association of stratigraphy at all (e.g., Stein 2000). Palimpsests, rather than living floors, characterize the archaeological record of all times and places. If an archaeological deposit results from the activities of many different peoples who undertook 8

Time in Archaeology: An Introduction different types of actions that produced artifacts through time, and moreover occupied a location for variable amounts of time, sometimes leaving the place completely, what then do the patterns apparent in the material culture left by these peoples mean in a behavioral sense? As Michael Schiffer (1972) noted so long ago, items found associated were not necessarily used together. This must introduce an element of difficulty into functional assessments, yet this problem has received relatively little discussion (but see Wandsnider 1996). Some archaeologists undertake a search for living floors, sites where the conditions of burial mean that artifacts were deposited within a relatively short span of time, in the hope of discovering a record where Time’s Arrow can be effectively ignored (surely the ultimate irony for an archaeologist). Taken to an extreme, one might imagine a living floor representing such a short span of time that no artifacts were deposited and no structures formed at all, although even such a site, in effect unrecognizable archaeologically, would still not be immune from the influence of Time’s Arrow (Bailey 2007). For patterns to appear in the archaeological record artifacts need to accumulate. Therefore the passage of time and the formation of palimpsests as a consequence are the very processes that make the record interpretable. Without the accumulation of palimpsest deposits, too few events will have occurred to form an archaeological record. Living floors and functionally associated tool kits presuppose artifacts manufactured, used, and discarded as the result of a single event. The artifacts in such scenarios are manufactured to fulfill a single need. Yet such a synchronic view underplays the results of research that indicates that all artifacts are to some degree the products of complex use-life histories (DeBoer 1974; DeBoer and Lathrap 1979). This is most clearly seen when the form of artifacts changes as a result of use. The wellknown tool resharpening studies of Middle Paleolithic scrapers, for instance, indicate that certain tool forms, classified on their morphological differences into distinct types, were in fact the result of repeated reworking of the tool edge (Dibble 1984, 1987). Thus, assemblage composition, measured as the proportion of different tool forms, each with

a different life history, reflects not the functional utility of a group of tools used at one particular instant but the outcome of combining many such scenarios through time. The pattern resides in the palimpsest, not in the functional instant.

Time Perspectivism and This Volume The chapters in this volume provide a variety of approaches that seek to exploit the palimpsest nature of the archaeological record in its various forms. Bailey’s chapter sets the scene by discussing the history of the development of time perspectivism and the nature of palimpsests. Subsequent chapters deal first with palimpsests of artifacts and features found within sites and then with the temporal significance of artifact use-life histories. A series of chapters relates use-life histories to their geoarchaeological situations within a landscape context, and a further series uses the same concepts to consider assemblage formation. Two final chapters deal first with the relationship between ethnoarchaeology and time and, second, provide an overview of time perspectivism in relation to the present volume. Sullivan, in chapter 3, discusses the types of behavioral information that may be inferred as a result of feature abandonment, but rather than seek an interpretation based on a simple dichotomy between planned and unplanned abandonment, diachronic and synchronic (the equivalent of a living floor) assemblage formation, he accepts that all assemblages take time to form. What may appear to be the simple consequence of different modes of abandonment in his house assemblages becomes, when the assemblages are viewed as accumulations with temporal depth, the result of distinctly different seasonal poses practiced by the people who deposited these assemblages. His investigation relies on the elaboration of the “trace” concept first introduced by Sullivan in 1978. Shott (ch. 4) picks up on the notion of artifact use-life histories to investigate the temporality of the earliest East African sites. His goal is to resolve a debate that pits a time perspectivist view of assemblage composition against one based on synchronic function. Like Olivier (1999), he is able to resolve the impasse by viewing artifacts and 9

Simon Holdaway and LuAnn Wandsnider a­ ssemblages not as the static outcome from manufacturing events but, rather, as the accumulation of a series of activities that continuously form and reform the archaeological record. As in Sullivan’s piece, significantly different inferences are possible if it is accepted that the archaeological record represents an accumulation of material through time, and explanations are sought that are compatible with this accumulative nature. Although use-life history may be most apparent in portable artifacts, particularly those manufactured in materials like stone where the reductive nature of the technology means that much of the sequence of reuse episodes is retained on the artifact, it is no less important in the morphology of other artifact forms and materials. Thus, Olivier (1999) is able to show how use-life histories of a variety of artifact forms found as grave goods in Hallstatt burials inform on the complex temporality of a site type once identified as a “closed find,” the structural equivalent of a living floor. Despite the burials superficially appearing to be the outcome of a burial event, Olivier is able to show that the temporality of the site is reflected in the different histories of grave goods, some from distant sources, some from much more restricted geographic locations. Even the structure itself shows a history of use beyond manufacture for a single event. Evidently, the burial chamber was opened at a later time, with material added and perhaps some removed. The elegance of Olivier’s analysis comes from interpreting the sites not as the result of temporally discrete events but, rather, as the outcome of a series of actions undertaken through time. This alternative reading of the temporality of material things forms the basis for three of the chapters in this volume that turn the problem posed by surface archaeological sites on its head. The living floor ideal is rejected, and the “problem” posed by the conflation of artifacts from multiple events is turned into an advantage. All three chapters exploit the life histories of artifact types as a way of drawing inferences about the temporality of the record, albeit from widely differing geographic locations and therefore different technologies. Wandsnider (ch. 5) uses a variety of portable artifact types as well as structures and geoarchaeo-

logical approaches to the time periods over which deposits were formed to assess the temporality of occupation across the Wyoming Basin of intermontane North America. Not all sites represent the same temporality; some are the product of quite short occupation duration and limited reuse of features, whereas other locations saw more intense reuse over longer periods. If the impact of the period over which surfaces were available to accumulate material is brought into the mix, it is possible to assess the span of time during which occupation occurred and the degree to which features were reutilized. Developing instruments that permit analysis of the formational complexity of palimpsests offers a landscape archaeology that is not based on analogies drawn from a synchronic settlement pattern and not based solely in timeless system time (departing from Ebert 1992). Relying on geographic information system manipulations, Dooley (ch. 6) proposes a battery of such instruments to help assess the temporality of medium- and long-term occupation in the northern North American Great Plains. He is well aware that the chronological resolution of his surface assemblage is not fine enough to put forward conventional synchronic behavioral interpretations, so he has modified his goal to document instead longterm landscape evolution as a way of addressing how the humanly created environment attracted or deflected past occupation. Holdaway, Fanning, and Rhodes (ch. 7) first discuss the temporality of the landscape, noting that in western New South Wales, Australia, land surfaces, and therefore the archaeological record they preserve, may differ considerably in age. A “dots on maps” approach to the identification of settlement patterns, wherein all sites and assemblages are treated as part of a coherent whole, is not applicable. Rather, like Wandsnider (ch. 5), they suggest that assemblage formation must be assessed in relation to the geomorphological history of the surfaces on which the deposits are found. Artifact assemblages in turn are not the functional remains of moments in time but reflect time accumulations over the time periods during which sedimentary deposits formed. Assemblage comparison shows patterns produced by varying temporal histories of 10

Time in Archaeology: An Introduction deposition that are not interpretable using conventional functional sets of inferences. Three further chapters use geoarchaeological concepts combined with considerations of time averaging to look at various aspects of assemblage formation. Stern in chapter 8 makes a detailed case for an alternative form of analysis to one based on the definition of a set of sequential, stratigraphically defined assemblages at different sites. Based on a careful geoarchaeological analysis of the formation of sediments in Pleistocene caves from Tasmania, she argues for the definition of the equivalent of the minimum archaeological stratigraphic unit used in the analysis of a paleolandscape (Stern 1993). The crucial question for analysis then becomes the differences in assemblage composition that accrue as a result of different rates and durations of accumulation, rather than interpretations based on assumptions about the functional equivalence of activities conducted at different sites. De Lange (ch. 9) reviews the results of Bailey’s Klithi research considering the nature of inferences that are drawn from lithic and faunal materials. Her critique centers on the relationship between the temporality of deposits containing artifact and faunal assemblages and the temporality implied by the types of analyses performed. At Klithi, a mismatch is apparent in the time-averaged nature of the deposits versus short-term explanations framed in terms of lithic reduction sequences and prey selection options. De Lange’s solution is to suggest that archaeologists follow more closely the approach adopted by paleobiologists, where the time-averaged nature of the record is used as the basis for picking which analytical techniques are selected and therefore what forms of inference can be made. One of the outcomes of an essentially synchronic view of the archaeological record is that we also tend to look for causation in synchronic terms, yet at any moment, action is a product of processes beginning in the past and continuing into the future (Bailey 2007). This is as true for ethnography as it is for archaeology, a point emphasized by ­Arnold in chapter 10. Ethnoarchaeological descriptions should not be seen as timeless vignettes but, rather, as having as much relevance to the study of change as archaeological data. As Arnold

states, both ethnographers and archaeologists effectively view their data at one particular time, the intersection of the observer and the observed, and processes that affect these data have operated, and will continue to operate, both before and after this point in time. Ethnoarchaeologists are particularly well suited to study some types of change—Arnold discusses supplanting/supplementing and intensification/extensification—when the temporality of the observations can be matched to the temporality of the explanation. A final chapter in the volume, by Tim Murray, provides an epilogue extending the background to the previous essays by identifying some of the questions that remain to be answered as well as the themes that are central to all of the chapters. Murray is interested in the relationships among time, the empirical, and the theoretical in archaeology. He argues that despite the range of what at face value appear to be fundamentally different theoretical approaches to contemporary archaeology, nearly all maintain the archaeology as anthropology metaphysic. Time perspectivism emanating from Bailey’s 1980s essays has clearly failed to spark the archaeological imagination and promulgate an alternative to mainstream archaeological theory. Murray searches for reasons why time perspectivism should hold such a peripheral status in relation to the disciplinary cultural norms.

Conclusion Perhaps it is not so surprising that archaeology, the discipline that above all others should be concerned with time, is in effect so unconcerned with anything to do with time except chronometry. Despite much discussion about the need to consider long-scale temporal processes, as the chapters in this volume illustrate, taking time seriously and thinking about the archaeological record are not straightforward issues. Interpreting what happened in the past as a reflection of familiar scales of behavior is the easy out. Such explanations seem so much clearer because they are easy to comprehend, but in many cases, probably the majority, these explanations fail because they lack a connection between the temporal scale at which the record can be ­analyzed and the temporal scale implicit in the 11

Simon Holdaway and LuAnn Wandsnider explanation. That this is no trivial failure is underlined by the virtual ubiquity of palimpsests, in all their various forms, in the archaeological record. Archaeologists may be able to address a wide range of research questions, but not all these questions will be applicable to the entire archaeological record. In addition, the types of explanations gener-

ated will in many cases depart from those familiar examples based on short-term lifeways. There are two ways to react to this situation: a retreat into conventional archaeological explanation or, as authors attempt here, the development of alternate ways to describe an archaeologically based past.

12

2

Time Perspectivism: Origins and Consequences Geoff Bailey University of York

Time perspectivism is an idea that was born of the intellectual ferment in the archaeology of the late 1960s and 1970s and specifically the challenge posed in Cambridge during that period by the rapid succession and overlap of a very diverse array of “older” and “newer” archaeologies. The purpose of this chapter is threefold: to summarize what I mean by time perspectivism, drawing on another recently written article (Bailey 2007); to consider the historical context in which the ideas arose as a way of illuminating their wider meaning and significance; and to consider some of the difficulties that have inhibited their acceptance and their practical implementation, as well as the consequences that must follow from fully embracing a time perspective view of the world. The historical section offers a personal view of the events that unfolded over 30 years ago and in which I was involved.

ples. This corresponds to the notion of time as process (Bailey 1983:​168) or what I (2007:201) have recently described as the substantive definition of time perspectivism, a definition in terms of how the world “out there” is supposed to work. The second foundation was the emphasis on the structure of the archaeological record, especially the palimpsest nature of material data, the variable but generally coarse resolution that usually accompanies it, and how that affects what we can or cannot know about the past. In common with others writing at the same time (notably Binford 1981a; Foley 1981a), I saw in this apparent loss of resolution an opportunity to focus on a different scale of phenomena not accessible to the student of recent and present-day events and processes, rather than a limitation. I (2007:202) have described this as the methodological definition of time perspectivism, a definition in terms of how we study the past and how the questions we ask and the way we go about answering them are conditioned by our time scales of observation. The original conception of palimpsest is worth some further elaboration in the words of the 1981 publications:

What Is Time Perspectivism? In its original formulation (Bailey 1981), time perspectivism was based on two principal ideas. The first was an emphasis on differences of scale, especially differences of time scale, and how such differences affect our understanding of events and processes—​whether past or present ones. The idea here is that changes in the time scale at which we make observations change what we see and that varying time scales bring into focus different variables and processes that are not visible, or not so easily visible, at other time scales, thus requiring different sorts of concepts and explanatory princi-

A sample of archaeological ...data often represents a palimpsest of activities ranging over a period of at least a hundred years to several thousand or more. It refers not to the activities of individuals, or even individual societies, but to larger aggregates of behaviour, ­reflecting 13

Geoff Bailey a­ verage tendencies . . .over long periods of time​ . . .​with a coarseness of resolution . . . further accentuated by the margins of error inherent in radiometric dating . . . . [Such a] record refers, by definition, to long term trends. While this may seem a severe limitation, it also offers an opportunity to focus on a different scale of behaviour and imposes a requirement to adjust one’s conceptual framework accordingly. [Bailey 1981:​ 109–110]

ception of the archaeological record contrasts with the more widely held view that the materials recovered by archaeologists are inherently incomplete and second-rate data, data that stand, rather inadequately, for something else—human behaviors, actions, and thoughts as we are used to describing those entities in our day-to-day lives and in the records of social anthropologists, sociologists, and historians. On this conventional view, progressive loss of evidence and loss of resolution, especially as one goes further back in time, result in progresSimilarly, Binford refers to sively worse data, by definition inadequate to answer the questions we really want to ask about the a massive palimpsest of derivatives from many past or in need of conversion into some other inseparate episodes . . . a different order of real- tellectual currency by means of clever theories or ity, the patterned structure of which represents clever techniques before they can be used to say not a simple accumulation of little events ... anything useful or interesting. not a poor or distorted manifestation of ethThese ideas about differences of scale and the nographic reality, but most likely a structured varying resolution of the archaeological record reconsequence of the operation of a level of orga- main the two most important and fundamental nization difficult, if not impossible, for an eth- ideas in time perspective thinking, though they nographer to observe directly. [1981a:197] have been variously expanded, modified, refined, and critiqued subsequently. In particular, they emFor Foley, archaeological data relate body what are really two distinct aspects of time scale: time depth or time span, longer or shorter as primarily to long term, gross characteris- the case may be, essentially the sense in which time tics . . .​the prolonged accumulation of repeated scale was used in my 1981 article; and time resoluevents . . . ​a blurring of the spatial patterns and ...​ tion, coarser or finer according to the available data richer but less resolved pattern . . .​the accumu- ing methods and the temporal resolution and size lated residue of long periods of time . . . relating of the samples of material available for study, the to those aspects of behaviour that are main- meaning implied in the use of the term palimptained over longer periods, and . . . [which] may sest. In a general sense these two different aspects of in fact be serving the useful purpose of filtering time scale are closely correlated in archaeology. Deout the ephemeral . . . . Long term trends may be tailed and accurate dating methods and chronoloof greater significance to the prehistorian than gies are typically associated with large samples of the understanding of a few short events . . . . Thus data resulting in high-resolution records that both archaeologists must perforce deal with accumu- permit and encourage a focus on shorter spans of lated, palimpsest residues of prehistoric behav- time. Conversely, as we move further back in time, iour. . .​[and this] significantly affects the scale so the dating methods become less accurate; the of analysis and interpretation. [1981a:1–16] margins of error, wider; and the time span within which data have to be aggregated to form meanI have quoted at length from these early essays ingful samples, larger. Nevertheless, these different not least because all three of us writing in 1981 ap- meanings of scale—size and resolution—should be pear to have published similar views independently recognized as distinct in order to avoid confusion. but also because nearly 30 years on it is pertinent to This correlation between high-resolution reask how far any of us has been able to realize that cords and short time spans, and between coarseoriginal vision, a point I return to later. This con- resolution­ records and long time spans, is not 14

Time Perspectivism: Origins and Consequences absolute. There are exceptions, or examples that are claimed as exceptions, and we should take the evidence as we find it. Nevertheless, the correlation is typical and may indeed have a basis in the physical laws of our universe and the inevitable and progressive decay and loss of material with the increased passage of time. Moreover, this contrast between higher-resolution shorter–time depth records, on the one hand, and lower-resolution longer–time depth records, on the other, also tends to correspond to whether we are looking at more recent or more distant parts of the record. This introduces another meaning implicit in time perspectivism, and identified in the 1981 essay, the notion that what we can see and understand of the world changes according to whether we are closer in time to the phenomena we are observing or more distant. Closeness in time allows observation of higher-definition­ detail but within a narrower field of view; remoteness in time results in loss of local definition but the potential to observe a bigger pattern (cf. Renfrew 1981). This meaning corresponds to the way in which we use the concept of perspective when interpreting spatial phenomena and to what I (2007:202) have described as the strict definition of time perspectivism. As in the spatial dimension so in the temporal dimension, increasing the distance (in space or time) between the observer and what is observed not only creates distortions that require correction but also places particulars in a wider perspective that can introduce new understandings and perception of new relationships. The article that followed (Bailey 1983) expanded on these themes and developed the idea of differences in the way different observers conceive of time, whether from cognitive, conceptual, psychological, cultural, or cosmological differences— time as representation rather than time as process (Bailey 1983:​169), or what I (2007:202) have called the subjective definition of time perspectivism, how different observers both present and past, including archaeologists, have incorporated the time dimension into their experience of the world and developed varying “time structures” (Bailey 1983:186). This idea drew on anthropological ideas about time, which have been the topic of an expanding literature (summarized in James and Mills 2005

and references therein) and a fertile and popular source of inspiration for archaeological studies that have sought to use material culture to throw light on past people’s sense of time (Bradley 2002; Clark 1992; Lucas 2005). The 1983 article is notable also for framing the discussion about time scales in terms of a contrast between “environmentalist” (ecological and environmental) schools of thought in archaeology and “internalist” ones (social and symbolic) and relating the former to long time scales and the latter to short time scales. This has caused immense difficulties for subsequent commentators. Most have assumed, either from an incomplete reading of the article or from deducing my views from my reputation as a practicing paleoeconomist without reference to the words of the text, that I was advocating such a polarity (Hull 2005 is a notable exception). This problem of misunderstanding has undoubtedly been a major reason why many have rejected time perspectivism or viewed it with suspicion, and I shall return to this point later. In fact, I was advocating almost exactly the opposite. While such a polarity might help to explain why archaeologists with more or less environmentalist or internalist interests tended to gravitate to different time scales of observation and hence to different parts of the archaeological record, we needed to move beyond a simple temporal categorization of environmental and internalist processes. Both might operate on both longer and shorter time scales, but their relative influence and the nature of their interaction might differ depending on the time scale of observation, and these differences were matters that needed to be investigated rather than assumed a priori: “If environmental factors can have an impact over short time spans [as well as long ones], the question naturally arises as to the impact of social and psychological factors over long time spans ...a very poorly explored area” (Bailey 1983:​182). I went on to consider the question of interaction between scales in terms of two notions, supported by a range of examples. “Hierarchical causation” refers to circumstances where the difference between scales is so large that the variables at each scale appear to have no relationship to each other, other than as boundary conditions, 15

Geoff Bailey and the processes under investigation at each scale and the causes considered appropriate in framing explanations are essentially independent of each other. “Interactive causation” refers to overlapping scales where the variables interact, resulting in mutual transformation. Many subsequent examples have been explored, though mostly of the interactive nature and in relation to shorter time spans and more recent periods, as in archaeological work inspired by the Annales tradition in history (Bintliff, ed. 1991; Knapp, ed. 1992) or in the sociologically inspired debate over the roles of individual agency and structure in the British Neolithic (Barrett 1994; Bradley 1993; Harding 2005; Hodder 1999). Also notable in the 1983 article is the absence of the terms time perspectivism and time perspective, but I reintroduced these terms in the 1987 article, which was a short summary of existing themes along with some additional consideration of both the virtues of time perspectivism and the potential objections to it, issues also discussed at some length in the conclusion to the 1981 essay.1 The final article in the sequence (Bailey 2007) returns to many of the existing themes and seeks to clarify much of the earlier thinking and to respond to criticisms of the earlier work. In this article I also explore the issue of implementation through a more detailed consideration of the concept of palimpsest and address the issue of what sorts of substantive longer-term processes might be illuminated by a ­focus on palimpsests, drawing on examples from the Klithi fieldwork project that had occupied much of the intervening period. This article offers a definition of palimpsest as “a superimposition of successive activities, the material traces of which are partially destroyed or reworked because of the process of superimposition” (Bailey 2007:203), and goes on to analyze different types of palimpsests defined by differences of temporal and spatial scale and resolution. I conclude that palimpsests are a universal phenomenon that we can never escape (cf. Olivier 2001); that there are no such phenomena as isolated events or moments in time, or none that is knowable from the archaeological record, because of the durational properties of the material world; and that through an analysis of the varying properties of different types of pa-

limpsests we might find a different and empirically better grounded way of writing about human history in deep time. This article also revisits the fundamental issue of the relationship between present and past and introduces the term durational present, defined as “the envelope of time within which phenomena of interest are accessible to study, and beyond which they appear to recede from view” (Bailey 2007:216), in order to give formal expression to the essentially arbitrary nature of the boundary between “the present” and “the past” and to highlight the ways in which the time span of the durational present varies according to the interests of the observers and their preferred techniques of observation. This notion of where the boundary lies between the present and the past has been a prominent theme in all these essays and a key to understanding the theoretical basis of time perspectivism through a critical evaluation of the concept of uniformitarianism, which received explicit treatment in all three of the 1980s essays and particularly extensive treatment in the 1983 article. In its most general form, uniformitarianism, a concept introduced by geologists in the nineteenth century and adopted in Darwinian evolutionary biology, is based on a belief in the uniformity of the world, an essential basis for developing an understanding of geological and biological history that does not depend on the arbitrary intervention of a divine power. In this sense uniformitarianism is contrasted with supernaturalism and is hardly a source of controversy in modern intellectual discourse except perhaps for Creationists. In its more specific form, uniformitarianism entails a belief that events in the past should be explained in terms of processes observable in the present and hence that the present is the key to interpretation of the past. This, however, poses much greater difficulties, and both geologists and some biologists have abandoned this position on the grounds that it fails to allow for past events that have no present-day analogue or for larger-scale processes with longer temporal rhythms that are literally not visible in the present or visible only with the use of very sensitive instruments, as, for example, in the use of laser beams to measure tectonic 16

Time Perspectivism: Origins and Consequences

Historical Context

plate motions. Gould (1965) specifically labeled this variant of uniformitarianism “substantive uniformitarianism,” a belief in the uniformity and similarity of past and present processes. This he distinguished from methodological uniformitarianism, a concept that retains the notion of a past that is knowable through empirical investigation without resort to supernatural intervention but allows both for greater variability of behavior and variations of scale and hence for the discovery of new phenomena that we could not otherwise have learned about from our knowledge of processes that we can directly observe in the present. Gould argued that both terms should be scrapped, the former because it is demonstrably incorrect and the latter because it is simply a statement of scientific method by another name. Ironically, archaeologists, who deal with the most variable type of data, the activities of humans, have been slowest to abandon the substantive position. Some have made a virtue of it, notably Trigger (1970), who claimed that archaeology is a historical discipline concerned with particular phenomena in the past, an idiographic discipline in his terms, and that general theory is the province of social scientists dealing with present-day phenomena. The 1983 article forcefully rejected that claim on the grounds that it presupposes where the boundary lies between present and past, closes off a priori any exploration of longer-term processes without investigating whether or not they may exist, and denies to archaeology the status of a fully autonomous discipline capable of generating new theoretical knowledge. To deny archaeology that opportunity seems to condemn it to the status of a second-rate discipline that attempts to study with imperfect data the past tense of phenomena that are better studied in the present and to the production of knowledge that at best does no more than add to the stock of particular instances of general phenomena that are already known about (see also Arnold, this volume). Most disciplines in the natural and human sciences have descriptive, historical, and theoretical elements and grapple in their own way with the relationship between ideas and observations and between the particular and the general. Why should archaeology be any different?

Cambridge in the late 1960s and early 1970s presented an extraordinarily stimulating and sometimes bewildering collection of different but strongly held views about archaeology, contrasting personalities, and a rapidly shifting landscape of intellectual leaders, cliques, affiliations, and fashions, all of this enlivened by frequent visitors from outside. Those of us who were undergraduates in the now defunct first-year course in archaeology and anthropology also learned something about social anthropology from Meyer Fortes, Jack Goody, and Edmund Leach, enough to appreciate that relationships between the two disciplines and their proponents are far from straightforward. The archaeologies on offer included the traditional and essentially “cultural-historical” and descriptive (but highly contrasting) approaches of the senior generation—​Grahame Clark, John Coles, Glyn Daniel, and Charles McBurney; the varied and sometimes mutually contradictory expressions of “new” archaeology espoused in the paleoeconomy of Eric Higgs and the analytical archaeology of David Clarke; and the early development of postprocessual archaeology by Ian Hodder. It is impossible to do justice to the varied viewpoints and intellectual crosscurrents of this period or offer more than a very brief insight here into the genesis of my own views. Others, myself included, who were involved in the events of that period have written elsewhere about some of the key figures as well as their own involvement, which provide alternative viewpoints (see in particular Bailey 1999; Clark 1989; Fletcher 1999; Gamble 1999a; Gowlett 1999; RowleyConwy 1999). My own pathway into time perspectivism stems from a variety of sources. I had already been out to Greece, before starting my university undergraduate career, on one of Higgs’s field expeditions in 1967 and continued as a student of Paleolithic archaeology with Charles McBurney and of paleoeconomy with Eric Higgs during the final period of Grahame Clark’s tenure as Disney Professor.2 All three believed in the possibility of seeing in the long term a pattern and an order of reality not available on shorter time scales. More than 50 years ago, Clark, in his inaugural address as Disney Professor, 17

Geoff Bailey had prophetically and eloquently foreshadowed later thinking thus:

nature of the archaeological record. For the paleo­ economic school that rapidly grew up around Higgs in the late 1960s and early 1970s, the belief A word may usefully be said at this point about that the archaeological record is necessarily a reperspective. It is only natural that those who ap- cord of what had survived and therefore of what proach prehistory from the literary standpoint had worked “in the long run” became a mantra, but should tend to grade its various phases accord- there was already some debate about what counted ing to the degree in which they are illuminated as “long” and how to deal with the short-run “failby the light of historical records, be this never ures.” so feeble or so fitful. To such, the final moments Another powerful stimulus for the young at that of prehistory occupy the foreground and so time was David Clarke, whose Analytical Archaeoltake on an exaggerated scale, whereas the great ogy published in 1968 marked the call to arms of the formative stages of human history are dimin- New Archaeology in Britain and who was beginished by perspective and fade into the distant ning to attract his own group of students, mostly horizon. Such an unhistorical way of view- younger than the paleoeconomists. Clarke had no ing prehistory backwards, against the unfold- proper lecturing position in the department until ing of evolutionary processes and the impact of 1975, but as a tutorial fellow of his college, Peterevents, is justified on the plea, which I believe to house, appointed in 1966 under Grahame Clark’s be specious, that it is a case of proceeding from patronage, he gave undergraduate lectures and suthe known to the unknown. [1954:9–10]3 pervisions (critical discussions of student essays) to small groups of undergraduate students (the norCharles McBurney offered similar sentiments in mal mode of education in the Cambridge system). his introduction to the monumental publication of In some respects, of course, Clarke’s emphasis on his excavations at the cave site of Haua Fteah: artifact taxonomy and sociocultural processes and Higgs’s emphasis on bones and seeds and ecologiIt is only just beginning to be realised that his- cal processes set them poles apart. But Clarke, like tory itself, and a fortiori the study of living com- Higgs, was very emphatic that archaeology needed munities, are incapable of throwing light on ...​ to develop its own concepts and theories to deal processes of long-term cultural accretion and with the peculiarities of its own data, to avoid the decay. . . . These are accessible only to a discipline meaningless accumulation of data, and to give up dealing with a different order of time-unit, and the pretence that the archaeological record could have been opened to archaeology mainly since be interpreted in terms of a faded history of peoples the advent of viable methods of time-measure- and civilizations. As Clarke put it: “The data studments . . . [w]hat is certain is that we are seeing ied is so inherently unlike that of other disciplines for the first time important aspects of group that archaeology must erect its own systematic apbehaviour in their proper perspective . . .with a proach or perish as a separate study” (1968:20). new class of data which will ultimately require a Again, “although archaeological data was genernew series of hypotheses. [1967:15] ated by people and societies the peculiar nature of the archaeological record has erased their precise For Eric Higgs, the archaeological record for all identification. Consequently, archaeology is a disits inadequacies could “iron out the fluctuations cipline with its own peculiar data, its own frame of of short periods and perceive the long-term trends reference and its own conceptual entities and prowhich are lost in the short-term multitudinous cesses” (Clarke 1968:41). variables of the terminal millennia” (1968:617). Clarke had also recognized that there was a In addition, both McBurney and Higgs, at problem of differences of scale between different that time the directors of active Paleolithic field sorts of archaeological entities that were usually igprograms, repeatedly emphasized the palimpsest nored. That received expression in his development 18

Time Perspectivism: Origins and Consequences of a hierarchical framework for artifact taxonomies and also led to an explicit recognition of the far-reaching consequences of radiometric dating for archaeology, the scale dependence of different sorts of concepts and explanations, and the need to transform thinking to cope with larger spatio­ temporal scales and novel behavior patterns not encompassed by social theories applied to modern societies (see Clarke 1973). However, Clarke had scarcely begun to explore these ideas before the developing creative tension between his own thinking and that of his contemporaries was stopped in its tracks in June 1976 by his unexpectedly early death at the age of 38. Three months later, on the very eve of his retirement, Higgs died at the age of 67 after a long illness. Clarke and Higgs were both highly influential teachers and had tended to attract the largest number of research students. Their near-simultaneous­ departure represented a traumatic rupture in the intellectual life of the time, left more than a dozen research students and a much wider circle of students and colleagues bereft of their intellectual guidance and stimulation, and in my case removed the prospect as a recently appointed lecturer in the department of working with David Clarke as a colleague rather than as a student. I had already been appointed in early summer 1976 as Higgs’s successor to begin my lecturing career the following October. But it took another year before Ian Hodder, who had been an undergraduate at the Institute of Archaeology in London before coming to Cambridge as a research student to work with David Clarke in 1971, and then briefly a lecturer at Leeds University, arrived to fill the vacant post, a move that marked the beginning of his conversion from the positivist pattern seeker of spatial archaeology (Hodder and Orton 1976) to founder of the postprocessual movement. A new group of research students rapidly formed around the new ideas and, as had been true of their paleoeconomic predecessors, formed the shock troops to advance the cause, adopting an evangelical style dismissive or disregarding of those outside the charmed circle of true believers. Paleoeconomy, which had openly treated the small scale of individual action and interaction as largely 19

an epiphenomenon, bound to be overridden in the longer run by considerations of economic viability, was dismissed as a functionalist paradigm. Indeed, the equally extreme insistence of the first generation of postprocessualists on the priority of the individual and the small scale, and the epiphenomenal nature of the long term, was perhaps an only too predictable reaction to the extremes of paleoeconomy. Colin Renfrew, still a distant voice in Southampton, was invited to give a seminar in Cambridge and was duly taken apart for the error of his processualist ways. Lewis Binford was given much the same treatment soon afterward in a celebrated confrontation in the South Lecture Room of the Cambridge faculty building. It was precisely the desire to sidestep this rather unproductive factionalism between such apparently irreconcilable opposites, and to come to terms with the tidal wave of new and mutually contradictory theories, both ecological and sociological, that was beginning to flood archaeological interpretation in the 1970s, that provided the added spur to thinking about time perspectivism. Initial reactions came mainly from informal conversations in Cambridge with colleagues, visitors, and research students and varied enormously. Some people were interested and understood that time was a conceptual issue that had been largely taken for granted and needed exploring, even if they did not agree with my particular view of the issues. One or two understood what I was driving at and shared my conviction that differences of time scale are fundamental to the development of archaeological theory. Roland Fletcher was an early ally, because of our common intellectual roots as student contemporaries and as students of both Clarke and Higgs, and had already begun to develop his distinctive approach to the analysis of scale ­differences in the built environment (see Fletcher 1977, 1995). Tim Murray immediately saw the point on his first visit to Cambridge in 1981 and went on to develop the ideas during the 1990s (see Murray 1992, 1993, 1997, 1999a, 2001; Murray, ed. 1999). Many thought time perspectivism was incomprehensible or at best irrelevant and too abstract to have much bearing on the real business of exploring new theories, collecting new data, or developing new techniques.

Geoff Bailey Many more thought it was absolutely wrong, especially those interested in some form of social archaeology, which was then gaining momentum from a variety of sources, not least from the ­arrival in Cambridge of Colin Renfrew in 1981. For them, what I was articulating seemed no more than a ­rear­g uard defense of paleoeconomy, an advocacy of economic or environmental determinism by another name and a justification for the disregard of social organization, individual agency, or beliefs as powerful motivating forces in human life (or at any rate an unwarranted pessimism about the possibility of investigating such factors with archaeological data), relevant at best in the short run but not in the long-term archaeological record. An even worse error from the postprocessual point of view was my apparent failure to appreciate the inherent subjectivities of knowledge and perception, whether our own as archaeologists, often unrecognized as such, or those of the past people we study. The 1981 and 1983 essays were duly interpreted as an attack on the postprocessualists and as an attempt to desocialize or dehumanize the past (see, e.g., Moore 1981; Shanks and Tilley 1987; Tilley 1981a). Whether the former charge is correct is open to interpretation, and the latter charge is demonstrably wrong. That reaction is understandable given my interests in paleoeconomy and the language in which the early ideas were expressed, but it is absolutely wrong, being based on a misunderstanding of some of my published statements (see Bailey 2007:200). The 1983 article has given particular difficulties as noted earlier (Knapp [1992:12] refers to all three essays as “somewhat recondite”). The fact that I work with subsistence data and the physical landscape setting of archaeological sites, rather than material culture as conventionally defined, is irrelevant. These are as much imbued with information about human action and initiative, sociality, and meaning, and as challenging to investigate from the point of view of their human significance, as any other sort of material data. My own position is and has always been that time perspectivism is not “antisocial” or “asocial” at all. Rather, my objection is that many of the social explanations preferred by archaeologists have all too often involved an inappropriate translation

from the very different contexts, scales, and types of observation employed by social anthropologists and sociologists to an archaeological context where the theories are immune to empirical challenge. Often this has seemed to involve little more than the use of empty rhetoric, a search for legitimation from outside the discipline of archaeology, or an exercise in intellectual one-upmanship and sometimes, so it seemed, a never-ending competition to see who could find the most arcane source of higher authority, as yet unread by any other archaeologist, from within the Franco-Prussian axis of European social philosophy (cf. Gellner 1985). It also results in the delusion that archaeological data need to be translated into a different and more readable form before they can be interpreted, and for those outside the discipline it reinforces the status of archaeology as a second-class intellectual discipline that works with intrinsically poor quality data that can never sustain worthwhile or verifiable conclusions. Having worked in a university environment that included at various times such figures as Jack Goody, Edmund Leach, Ernest Gellner, Anthony Giddens, and Tim Ingold, names that are often quoted as ultimate authorities by archaeologists, and having frequently sat around committee tables with many of them arguing about the allocation of scarce resources, I am certain that the generality of social anthropologists and sociologists, though mildly flattered that the archaeological community wishes to use their theories and defer to their authority, are quite unimpressed with the results, not least because archaeologists usually lack access to participant-observations that would provide both a deeper understanding of social anthropologists’ theories and an independent evaluation of explanations derived from them.4 As for objections to postprocessualism, this is not the time or place for an evaluation or a weighing in the balance of the positives and negatives (but see Murray 1999a). Postprocessualism has in any case evolved and diversified to the extent that it is probably mistaken to regard it as representative of a homogeneous body of ideas or beliefs. My principal difficulties in this context are with the persistent unwillingness of many postprocessualists to engage with the problem of scale differences except 20

Time Perspectivism: Origins and Consequences

Difficulties and Consequences

over a very narrow spectrum; the tendency to collapse all interpretation to the level of subjective, individual experience; and the frequent adoption of a somewhat Orwellian rhetoric that while all opinions are equal, some, so it seems, are always going to be more equal than others. It is symptomatic of the problem that Gavin Lucas (2005), in his stimulating and anthropologically well-informed treatment of the archaeology of time, largely ignores the problems posed by differences of time scale, mainly, it seems, on the grounds of their political incorrectness (Bailey 2006), and that the most popular aspect of time in archaeology over the past decade has been the attempt to reconstruct past people’s subjective experience of time—the “past in the past” of Richard Bradley (2002), even though the empirical basis for such reconstructions is open to question (Bailey 2007:219). My original motivations in developing the idea of time perspectivism were thus primarily to create an intellectual space in which diversity of opinion and interpretation could coexist without abandoning the aspiration to systematic empirical evaluation inherited from processual trends in archaeology and to show how quite different and apparently irreconcilable ideas about human existence arise in large part from different time perspectives. A second motivation was to work at the development of archaeology as an independent discipline grounded in the distinctive properties of its own data and creating its own ideas and theories, rather than one that constantly looks over its shoulder for theoretical inspiration and approval from other disciplines, seeking legitimation through deference to authority or intellectual fashion. Nevertheless, it is true that the development of the postprocessual program has been a continuing point of provocation and point of reference for the development of my own thinking, as well as a source of resistance to time perspective ideas. It could hardly have been otherwise, given that Ian Hodder and I were colleagues in the same department for 19 years and that I along with other staff members had as one of our various duties the task of reading the progress reports on all of Ian’s research students, beginning with Henrietta Moore and Chris Tilley.

Why has it taken so long for time perspective ideas to develop, other than from disciplinary inertia and the slowly turning spiral of intellectual fashion? I suggest that three problems are paramount: the problem of the individual, the problem of narrative, and the problem of empirical implementation.

The Problem of the Individual

The first and perhaps most formidable obstacle to an understanding of time perspectivism is a belief in the importance of the individual. A necessary consequence of time perspective ideas and the palimpsest nature of the archaeological record is that individuals and even recognizable social groupings very quickly disappear from view, and we are left with what Benjamin has described as “shadowy organizational themes or clusters of ideas” (1985:223). In fact, we may not even have any coherent largescale entities at all but, rather, the accumulated fragments of many different such entities that we are tempted to stitch together to form a pseudoentity that never existed (Holdaway et al., this volume). We may suspect that individuals and individual societies are hidden away in the palimpsests of the long-term record, and indeed are responsible for some of its patterning, even if we cannot see the evidence of the individuals or even the individual societies, any more than we can lay bare the specific genes that contribute to the paleontological record of biological evolution. Yet, however hard we try to leave our individual selves out of the equation, a story without individuals somehow seems incomplete, abnormal, and dehumanized. Various strategies have been employed to try and put the individual back into the archaeological picture. One strategy of course is to ignore the problem and to continue to explain patterning in the archaeological record in the traditional way, in terms of individuals, societies, and historical movements, as those terms might be used by sociologists or historians. But this is scarcely sustainable when the mismatch between such concepts and the scale and resolution of the archaeological record, and the time-averaged nature of so much archaeological material, has become so obvious. 21

Geoff Bailey Another strategy is to focus on the occasional windows into deep time (Hodder 1999) or evidence of real-time experiences (Lucas 2005) that provide a flash of insight into small-scale phenomena: the Austrian Ice Man caught and buried in a snowstorm complete with his clothing and tool kit, the Boxgrove knapping floor and horse butchery site, and other so-called moments in time. The difficulty here is that these examples, even if they are genuine “moments in time,” are absolutely rare and tell us little about how we are to deal with the vast bulk of the remaining record of archaeology. A third strategy is to incorporate some generic concept of individual agency, but the difficulties of such a concept let alone their implementation have been well rehearsed (Hodder 1999:136–137; Johnson 1989). Yet another strategy is to focus on the biography of individual artifacts (Gosden and Marshall 1999; Lucas 2005). The biographical analogy is itself telling, and although such studies may illuminate in interesting ways the structure of the archaeological record, they sidestep the more challenging problem of how different artifacts are to be related to each other (Bailey 2007:218). Another possibility is to people archaeological texts with imaginary observers, who provide an individual commentary on what it might have been like to stand in front of the bison frieze at Lascaux or huddle round the campfire on the Siberian steppe waiting for the mammoth (Mithen 2000). Some of these strategies are more convincing than others, and one sometimes feels that it would be better to let the imagination have full play and write convincing fiction! Let me be absolutely clear about this. I am not opposed in principle to looking for individuality or viewing the world through the lens of an individual’s eye or to any of the strategies outlined above, many of which have the potential to offer new insight and understanding and, if nothing else, a source of entertainment and communication for a wider audience. My point is that this is only one source of inspiration, one point of view, and one that should not become an obsession in disregard of others and especially not a tool with which to dismiss other points of view. Many, of course, see in the appeal to larger-scale

processes the threat of determinism and an unpalatable notion of individuals so overwhelmed by larger-scale phenomena over which they have no control that their role as conscious agents seems quite irrelevant. It is not clear who or what is threatened by such a notion, and such fears in my view are in any case quite groundless. The archaeology of the large scale provides an opportunity for us to enlarge our understanding of our humanity rather than a compulsion to reduce it, a “megascope” with which to view ourselves from a different point of view. There may well be some very interesting questions to be asked about how individual actions and small-scale social interactions generate pattern at a larger scale (cf. Hopkinson and White 2005), as well as some equally interesting and important questions to be asked about how those larger-scale patterns and processes then react back on behavior and human thought and action at the smaller scale. There is every reason to suppose that larger-scale processes—conceptual, social, political, biological, environmental, and climatic as the case may be— were at work beneath the surface of everyday events in the world of our distant predecessors, just as they are today, some of them working over very long time scales well beyond human lifetimes or living memory and probably well beyond the conscious recognition of the individuals who lived through them. Archaeology provides access to a wider range of scales than any other discipline concerned with human affairs, and especially those at the largerscale end of the spectrum, and insight into how and to what extent those processes intermesh with each other and with individual lives. Rather than ignoring the larger scale on the grounds that to do otherwise is to succumb to determinism, we might do better to embrace the archaeology of the large scale in order to learn in what ways such processes are continuing to affect the world we live in, lest we suffer the very fate that opponents of determinism most fear, that we will indeed be determined by forces that we neglected to investigate and which are therefore beyond our control. Much of what has happened in the development of archaeological thinking over the past three decades can, I think, be seen as an attempt to deal with the two-sided problem of studying the past 22

Time Perspectivism: Origins and Consequences revealed by time perspectivism in its strict sense, the problem of distorted perspective, on the one side, and the interest in exploring relationships at a larger scale, on the other. The first problem has been tackled by trying to work out how the world appeared from the individual perspective of someone living in, say, the Bronze Age or the Upper Paleolithic period and showing that their world was characterized by the same sort of cognitive and social complexity as our own is, through a process of “normalization” (Bailey 2007:218, see also Murray 1993, 2001, this volume). That has been the dominant focus of attention and source of inspiration for many of the most prominent trends in theoretical archaeology in the past 30 years and, indeed, a prominent theme in archaeological reconstruction long before that. My own interest in time perspectivism has always been much more in the second aspect. My view remains not only that this latter approach is better suited to the archaeological palimpsests from the remoter past but that this larger-scale perspective may actually tell us something new about ourselves that we cannot otherwise gain from any other source of knowledge. Conversely, the attempt to reconstruct small-scale phenomena and the world as interpreted and acted on by the prehistoric individual, for all that it may help to correct the tendency to simplify the past because of a distorted time perspective, must ultimately fail in many cases, except perhaps as an exercise in literary imagination, not because prehistoric people did not think and act in the complex ways that we would like to believe but because in all too many cases we simply do not have access to the variety and resolution of data that would enable us to investigate such a fine level of detail (Murray 1997). It would be rather surprising if prehistoric people did not experience the same sorts of small-scale phenomena that we are familiar with when we take a close-up view of ourselves and our contemporaries. It would be equally surprising if they did not have lives enriched and no doubt at times oppressed by sources of social convention, myth, and symbolism. But that does not require us to attempt to try and reconstruct everything in the remoter past that might conceivably have taken place. The

attempt to reconstruct past worlds as they might have been perceived and experienced by past individuals should not be dismissed where the material record offers opportunities for developing such an approach. But to assert that this is the only worthwhile mode of interpretation, and that any attempt to explore other layers of meaning at a larger scale dehumanizes the past, is at best narrow-minded. It is rather as if one were to complain that to study the vegetational history of, say, the Pliocene period without first demonstrating that Pliocene plants obtained their primary source of energy from photosynthesis is to “de-botanize” the study of ancient plant life or that to ignore processes of fluvial sedimentation and erosion when studying the geological history of the Scottish highlands is a threat to the autonomy and sense of identity of sedimentary particles.

The Problem of Narrative

The problem of narrative follows from the problem of the individual. Indeed, Lucas (2005) refers to real time, the individual experience of lived time, also as narrative time. As a historical discipline, we usually expect the primary output of archaeological research to be some sort of history, a story with a narrative structure. It may be the history of a particular place, a particular time, or a particular people, or it may be a history of the world. It may be the history of a particular artifact, a history of ideas, or a history of archaeologists. The structure may be couched in terms of a linear story with a beginning, a middle, and an end; it may be multilinear or even nonlinear; and it is often cumulative, involving some measure of progress or at any rate of development and increasing complexity over time. Or it may be a narrative of cyclical change or even of underlying continuity. It is frequently a story of ultimate origins, of humanity, of language, of agriculture, or of civilization. It may be one or many of these things, but in every case we expect a story line that says something about trajectory, preferably a trajectory that explains how we have come to be as we are as individuals in the world we presently inhabit. With time perspectivism, however, we have a problem. The first is the problem of scale. If 23

Geoff Bailey d­ ifferent processes and phenomena become apparent at different scales of observation, there can be no ­single unified history of the world or even of some restricted place or period, only a multiscalar history written from many different points of view. How are we to narrate such a multiscalar history? The answer to that question is further complicated by the palimpsest problem. Consider the longstanding conventions of world prehistory. If we examine the general textbook approach to “World Prehistory,” the themes and the general structure of the narrative have scarcely changed from the first edition of Grahame Clark’s World Prehistory published in 1961 to the eighth edition of Brian Fagan’s People of the Earth (2001) published 40 years later. The key developments remain the origins of humans, the origins of modern humans and the symbolic explosion that accompanied their development and geographical expansion, the origins and spread of agriculture, and the origins of urban civilizations, to which we might add the origins of empires and the origins of the modern world system (if we are prepared to regard later “historical” periods as falling within the purview of a prehistoric perspective, as many think we should if we are not to erect an entirely artificial barrier between prehistory and history). To be sure, the volume of data and the number and geographical range of examples have increased, and some of the dates have been pushed back in time. The narrative has also become a more richly branching multilinear narrative and occasionally a nonlinear one, rather than a simple linear progression, and one that has become subdivided and indeed fragmented into many regional prehistories, as more geographical regions have been brought into the picture and themes like art, language, religion, and economy have become more fully substantiated. There are also more varieties of explanation on offer for how these great transformations were brought about. But the basic structure and the main themes and highlights have changed remarkably little, as has the basic conception of what is there to be explained. What we have is a predominantly linear narrative, marked by episodic bursts of development. It is also a narrative that conveys a large measure of progression with time from simpler to more complex modes of orga-

nization and from slower to more rapid sequences of change, punctuated by “revolutions” that mark the ladder of human progress. Little has changed since Dunnell (1982) remarked on this phenomenon over 20 years ago. In contrast, one of the most striking features of archaeological thinking and practice over the past 40 years, at least in the estimation of archaeologists themselves, is the huge changes and upheavals that have occurred in archaeological theory and archaeological method. Think of the great “paradigm shifts” associated with the New Archaeology, environmental archaeology, anthropological archaeology, ecological archaeology, processual archaeology, postprocessual archaeology, cognitive archaeology—I could go on. Or consider the huge range of new methods introduced by scientific developments, new methods of radiometric and cosmogenic dating, chemical and physical methods of characterizing materials, statistical and computing techniques, spatial analyses, biomolecular techniques—​again the list could go on. Trunk loads of books have been produced on these themes and continue to be produced—witness the bookstalls at any large international conference. Add to that the vastly increased quantity of data that has been produced as a result of ongoing fieldwork and especially as a result of rescue excavations in many parts of the world. The fact that what has been supposed to be the primary output of our discipline, the writing of large-scale history in a continental and global perspective, has changed so little in its basic structure amid an ongoing tumult of changing theoretical orientations and a torrent of new technical developments is a singular paradox. It is almost as if for much of the past few decades archaeologists have been living in two parallel universes, between which there has been very little communication and between which there appears to be a growing disjunction that must surely require some resolution before too long. A time perspective examination of this paradox raises two questions about the conventional structure of world prehistory. The first is the degree of correspondence or otherwise between the narrative as interpreted and the material record. It is 24

Time Perspectivism: Origins and Consequences a­ xiomatic that the material record of prehistory on the global scale, as at every other scale of investigation, is a palimpsest, a huge palimpsest that encompasses the whole surface of the earth. Moreover, there are good reasons to suppose that large segments of this palimpsest are true palimpsests (Bailey 2007:​203), in which the previous record has been wiped out and replaced by more recent material, or at any rate cumulative palimpsests (Bailey 2007:​204) in which there has been substantial agerelated loss of material. Thus the apparent growth of detail and complexity as the record of human history unfolds over time might have much to do with the selective loss of information and loss of resolution as we attempt to move back through the successive layers of this global palimpsest. Moreover, it is clear that as with smaller-scale palimpsests, what are being sought after and highlighted in the narrative of world prehistory are the highest-resolution data sets, the hot spots of the record that provide the best detail of pattern and resolution of dating. This raises the suspicion that emphasis on the high-resolution episodes is often achieved only at the cost of disregarding the great bulk of the other available data. Clearly these are matters that require investigation, but analysis of very large-scale palimpsests is just as likely to subvert conventional wisdom as analysis of smallerscale ones. A second question has to do with the impact of time perspectivism in the strict sense of the term. The emphasis on “origins” research and the reality of “revolutions” are seriously called into question if we take account of the distortion of perspective that comes with tracing phenomena back in time until they disappear from view over a temporal horizon that lies beyond the interests of the observers or their preferred techniques of observation. That distortion of time perspective, like its equivalent in the spatial dimension, is bound to make events that are closer in time appear more detailed and more complex than those that are further away. This poses the challenge that the changes that we see as we move forward in time through the narrative of world prehistory may be no more than the result of successive increases in the scale and resolution of our powers of observation, rather than inherent changes in the

phenomena themselves. So-called revolutions tend on closer examination to reveal multiple strands of ongoing change as well as continuities that cross the supposed borderline. The notion of an agricultural revolution was deconstructed in that way over 30 years ago, and it is perhaps no surprise that the removal of that episode as a revolution has simply displaced the search, for those who want to mark the trajectory of human progress, to other parts of the record, such as the symbolic revolution of anatomically modern ­humans or the secondary products revolution of the European Bronze Age (Gamble 2007). Just what a fully worked-out time perspective account of world prehistory might look like, or whether such an enterprise is possible, remains unclear. One can certainly take a global perspective, but that may be something quite different. For Grahame Clark the advent of radiocarbon dating was the key that made world prehistory possible by enabling events in disparate geographical regions to be brought into some chronological relationship. Now that we have a vastly increased number of radio­carbon dates from all over the world, added to by many new dating methods of varying range and precision, it may be that in a paradoxical sort of way, independently derived dating has now become an obstacle to further developments in understanding. By continuing to focus on issues of sequence and correlation and to emphasize the possibility of ever more refined chronologies that are necessarily always one step beyond our current technical abilities, absolute dating has rather diverted attention away from the analysis of things in themselves and especially away from the differential temporalities that are inherent in different sorts of palimpsests. Improved chronologies are constantly being sought as the indispensable prerequisite to do something else. But what is that “something else,” and what degree of chronological resolution is required to study it? One suspects that what is being sought is the holy grail of a universal high-resolution chronology that will enable us to write about the history of the Lower Paleolithic period in the same detail as the history of the medieval period or the twentieth century, something that a time perspective view of the world suggests is 25

Geoff Bailey a physical impossibility in the universe in which we happen to live. I suspect that large-scale syntheses of prehistory will increasingly move away from the attempt to achieve a sort of large-scale history of everything, made possible and comprehensible only by being compressed and distorted into a linear narrative. What I suspect we will see in their place are histories that are more thematic in their approach and a conception of time in history not so much as a linear pathway but as something more akin to a sphere that surrounds us on all sides and into which we can reach in any direction, over any distance and at any scale, according to the themes that we wish to explore.

remain many ambiguities in the interpretation of the Klithi data, as de Lange (this volume) observes, ambiguities in the relationship between the structure of the data and the particular interpretations preferred by different contributors to that project, particularly in the interpretation of the on-site data of stone artifacts and animal bones. This, perhaps, is no surprise given the number of participants in that project and their heterogeneous prior expectations and approaches to interpretation. Many of the examples of time perspectivism emerged most obviously at the larger scale of the Klithi site in its wider landscape setting rather than at the intrasite scale of analysis (Bailey et al. 1997; Bailey 2007; Galanidou 1997a, 1997b). Tim Murray’s influence in Australia also strongly influenced field projects carried out in the 1980s and 1990s by his colleagues in the La Trobe department. Nicola Stern’s (1993, 1994a, this volume) work at Koobi Fora combined time perspective thinking with the strong tradition of paleontologically inspired taphonomic analysis developed in the investigation of early sites of the African Rift Valley and highlighted the time-averaged­ nature of archaeological deposits and the impossibility of applying an ethnographic scale of reconstruction to their interpretation. Jim Allen and Richard Cosgrove grappled with similar issues of how to reconcile fragments of individual action captured in archaeological palimpsests with largescale low-resolution data in the Southern Forests Archaeological Project of southwest Tasmania (Cosgrove 1995; Cosgrove and Allen 1996). Here they devised strategies to deal with late Pleistocene cave sequences with an immense abundance of data but very low temporal resolution, concluding that the only satisfactory approach to investigation was to search for variability in the totality of the available archaeological data against an independently derived framework of environmental and climatic change, shorn of all ethnographic preconceptions. This work is of particular interest on a continent where a rich hunter-gatherer ethnographic record has often been extolled as a virtue for archaeological interpretation. The Southern Forests project highlighted the incongruity of an ethnographic approach to interpretation, not only because of the

The Problem of Implementation

In some ways this is the most difficult problem of all, for even if we can see beyond all the obstacles and distractions outlined above, there remains the question of how we are to put a time perspective approach into practice. How are we to set about unraveling the structure of the archaeological record, the varying resolutions of different palimpsests, and their congruence or otherwise with different sorts of processes and different sorts of substantive questions about human life and human history? What are the longer-term processes that are supposed to come into focus at larger time scales, and what is the relationship between different scales of observation and the balance of advantage between “microscopic” and “macroscopic” (Bailey 2007:210) approaches to the material record? These are not questions that will admit of a quick answer, and the long interval between Bailey 1981 and Bailey 2007 is symptomatic. Much of that time interval was for me taken up with a large-scale field project, the Klithi Project, a major undertaking that lasted without loss of momentum for 18 years from the very first visit to the Ioannina Museum in 1979 to the final publication in 1997 (Bailey, ed. 1997). That experience, however, provided a test bed for experimenting with the investigation of palimpsests, and from it emerged examples of time perspectivism and an understanding of longer-term processes that I have drawn on in more recent discussions (see Bailey 2007). There 26

Time Perspectivism: Origins and Consequences a­ lmost total lack of overlap in scale and quality of the archaeological and ethnographic records but also because the late Pleistocene occupation of southern Tasmania took place in a glaciated landscape with no modern environmental analogue, let alone a behavioral one. In fact, a great deal of other research has been going on in the past 30 years that can be seen to have been moving toward a more critical exploration of the structure of the archaeological record. This includes Schiffer’s (1976, 1987) focus on formation processes, the large literature on taphonomy (Lyman 1994), and Binford’s (1983b) articulation of middle-range theory and the importance of actualistic studies (studies of the relationship between behavior and materials in the present). Some of this research was perhaps motivated by a desire to clean up the record and remove sources of potential bias, so that one could then proceed to interpretation of the humans that lie behind the materials according to existing interpretive conventions. Much of it was often viewed, and indeed can be criticized, as an exercise in “spoiler” arguments, the demonstra­ tion of what we cannot do with archaeological data rather than of what we can do and thus as the development of methodological tools in isolation from theories and interpretations, methodologies in search of a substantive problem worthy of attention. The same charge has been laid against the analysis of palimpsests and the focus on the timeaveraged nature of many archaeological deposits (cf. Stern 1993 and comments, 1994a, this volume). Others who have traveled the same road seem to have faced similar difficulties. Clarke (1973), who saw the problem very clearly, did not of course live to pursue it. Binford (1981a) and Foley (1981a), publishing in the same year as my first time perspective essay, likewise seem to have had variable success subsequently in realizing their original vision of investigating a different order of reality potentially revealed by archaeological palimpsests. The fact that they, like me, saw the most useful theoretical tools for investigating what was revealed by long-term palimpsests as lying in various sorts of ecological, adaptational, or economic thinking has been a huge barrier for that great swathe of archaeological opinion that wants to do social or

symbolic archaeology and suspects anything else of being a sort of crude materialism or environmental determinism. Foley has moved far away from the investigation of archaeological palimpsests and into the study of human evolution, a really large-scale, longterm phenomenon, where there is also a reasonable degree of congruence between evolutionary theory and the analysis and interpretation of paleobiological data (e.g., Foley 1995). Binford, in concentrating on the methodology of palimpsests and the development of middle-range theory, seems to have moved away from conceptions of long-term processes to concentrate on an ethnographic scale of interpretation, albeit one that emphasizes the system rather than the individual (see Binford 1983b). His emphasis on using actualistic studies to develop techniques for converting the statics of the archaeological record into the dynamics of behavior also seems at times to be reverting to the view that the palimpsest material record is a sort of second-rate data set in need of some keys that can convert it into something else that is more easily interpretable. This seems paradoxical given the assertion of his 1981 article quoted earlier. Yet his more recent work suggests that his view of ethnography is quite distinctive, a “frame of reference” for investigating past phenomena, that is, a body of secure knowledge (source-side knowledge in Binford’s terms) that can be used as a tool of investigation to provide new knowledge about a less well-known set of phenomena (subject-side phenomena [see Binford 2001a]). This is an approach that S. J. Gould would have recognized as a perfectly respectable application of methodological uniformitarianism, a means of acquiring new and empirically wellgrounded knowledge. My only question is how far source-side knowledge derived from short-time span ethnographic data, albeit viewed on a global scale, can be applied to long-time span phenomena in the more distant archaeological past. If archaeological data are robust enough to provide an independent test of predictions derived from analysis of ethnographic data, they should be robust enough to provide a source side of knowledge in their own right or subject-side data that can be illuminated by other frames of reference that are more congruent 27

Geoff Bailey in scale and resolution with the archaeological record under investigation. One very powerful frame of reference in Binford’s sense is the geological and geomorphological history of the earth’s surface, which provides the physical landscape setting within which people have made their living and within which archaeological data are embedded. Another frame of reference is the behavior of the plants and animals that have occupied those landscapes. It is significant that the field investigations informed by a time perspective approach and referred to earlier, the Klithi Project in Greece, the Southern Forests Archaeological Project in Tasmania, and the work of Stern in Africa and of Holdaway in Australia, have made extensive use of both of these frames of reference. As Cosgrove and Allen (1996) put it, a paleoecological framework is intended to provide an independently derived record that can be used as a methodological tool for investigating variability in the archaeological data and generating testable hypotheses, rather than an interpretive tool that imposes deterministic explanations. It is no surprise that archaeologists have ­often sought common ground with earth scientists in addressing issues of time, that the understanding of time scale embraced by earth scientists has been a significant source of inspiration for my own ideas, or that an understanding of geomorphological context is central to the work of others in this volume. On all but the very shortest time scales, the physical landscape is immensely unstable, whether from climatically driven processes of erosion, sedimentation, and sea level change or from geophysical processes that generate earthquakes, volcanic eruptions, and earth movements. If we pose the question of how such physical instabilities have affected the course of human history, we might assume that they have been essentially disruptive and destructive, a more or less frequent interruption of cultural processes of human development, and possibly even a determinant of them. Closer examination shows that prehistoric people, so far from being determined by such geological instabilities, turned them, and the characteristic landscapes of topographic complexity that result from such instability, to their advantage (Bailey et al. 1993; King

et al. 1997; Sturdy et al. 1997) and that this positive engagement with geological instability may have played a powerful role in our early evolution (Bailey et al. 1994; Bailey et al. 2000; King and Bailey 2006). In our research on the landscapes of Epirus, where we were able to talk to the people who live in such landscapes in the modern setting, they literally did not see the eroded surfaces, landslides, and other hazards that so impress visiting geologists and archaeologists or did not see them as relevant (Green 1997, 2005). They had long ago developed highly flexible economic and social structures that turned the features of a geologically unstable environment to positive effect while minimizing the hazards, in what might almost be called a process of coevolution between human activities and geological change (Bailey 1997c, 2007). The reality is that time perspective thinking is not easy. It offers no quick fix, no simple code for transforming the incomprehensible into the familiar, no ready-made map for reinterpreting the past, no fully developed set of new principles. It cannot even in its present state of development be described as a new theory in the sense of a fully worked out and coherent system of ideas and principles that account for the way in which the world has come to be the way it is. Nor does it deal with “people,” “culture,” or “behavior” in the sense in which any of those concepts might be used in everyday usage or in the anthropology or sociology of contemporary and historically recent societies. Taken to its logical conclusion, time perspectivism requires us to abandon most of the conventional foundations of existing archaeological thinking, indeed of conventional thinking more generally about the human condition. Those who embrace time perspectivism will find themselves in an alien intellectual landscape where most of the familiar landmarks and signposts are missing. Such a landscape is full of liberating possibilities for new exploration and the discovery of new knowledge, but it requires us to work out the “maps” for ourselves. Nobody else can help us. Some may find the effort of abandoning so much inherited intellectual baggage too much. Others may find the prospect of moving into unknown territory too disturbing and turn the signposts around so that they point back 28

Time Perspectivism: Origins and Consequences to the familiar world that we already know. Those who cannot accept this starting point for inquiry will find time perspectivism quite incomprehensible or quite irrelevant, and it is clear that during the past two decades many have.

the ultimate outcomes are yet to emerge with any clear definition. Methodologically there is a need to expand the analysis of palimpsests into every domain of material phenomena and at every scale, ranging inward, microscopically, and outward, macroscopically, to establish where the limits of resolution and comparison are in different circumstances and to clarify just what is there in terms of the varying structural, spatial, and temporal properties of different sorts of records. This is a task that can and should be applied to every type of material record ranging from art and ceramics, as among the most malleable and obvious products of human imagination at one extreme, to mollusk shells and sediments, as among the most environmentally constrained at the other, and on every scale from the individual object at one extreme to the global palimpsest of world prehistory at the other. Substantively there is a very clear need to explore and clarify what sorts of theories and questions work best with the structural properties of different sorts of material records at different scales of observation. This will not be an easy task, but it is one that needs to be conducted in concert with the analysis of palimpsests. The sources of inspiration for that task may come from many different directions, including induction from the structural properties of the records themselves and the problems they pose, deduction from models and theories drawn from elsewhere, combinations of different frames of reference, challenges to conventional assumptions and practices by the identification and exploration of alternatives, thoughtful and creative engagement with the empirical record, and above all the simple asking of questions free from the dictates of moral righteousness. A proper spirit of open scientific inquiry is one that should encourage the asking of different questions but should also be demanding in the standards of rigor and evidence that are applied in the search for answers—and above all should be open to the possibility of error.

Conclusion In standing back from the development of time perspective ideas over the past 30 years, two factors stand out above all others as the key to understanding the slow development of ideas, the requirement to abandon familiar territory, and the need to work out a new way of thinking from the beginning. Those of us who first peered across the frontier into this new and alien landscape initially drew back again, because we simply were not sure how to proceed. Progress in developing the ideas has been slow and is likely to remain so, with many halts, diversions, wrong turnings, and false trails. If this sounds daunting, we should remember one of the most important lessons of archaeological survey, the “sampling paradox” of the field archaeologist working in new territory, that usually we do not find anything until we know what we are looking for, and we do not know what we are looking for or even where to start looking until we find something. Our methods of inquiry and our understanding of what we are trying to investigate need to proceed together and interact with each other, and this process is likely to require many trials and errors and an engagement with archaeological material that, given the complexity, cost, and scale of modern fieldwork, must of necessity be a long-term one. Anyone hoping for a quick solution to archaeology’s problems will find no encouragement here. Above all it will require us to abandon a dependent relationship on sociological or ethnographic disciplines, even at the risk of overreacting against these popular sources of knowledge lest they entice us back into the old ways (cf. Sullivan, this volume), but not to ignore them so much as ultimately to better engage with them in a two-way process of communication. The other chapters in this volume show that many are now prepared to travel this route and to explore in new ways the structure of their archaeological records and how they should be interpreted, even if

Acknowledgments I am grateful to Simon Holdaway and LuAnn Wandsnider for their original invitation to talk in a session about time at the meeting of the Society 29

Geoff Bailey for American Archaeology in Milwaukee in 2003 and for their subsequent encouragement and comments and the comments of all those who took part in the original session and commented on earlier version of this chapter. I am also grateful to two anonymous assessors whose comments have helped me to clarify my own thinking on time and to Mark Edmonds for a critical reading of the text.

time perspectivism differs, I believe, from other relativist approaches is that it provides a means for understanding why different points of view differ, because of their basis in different time perspectives, rather than by recourse merely to differences of personal preference or social context. 2. Grahame Clark retired in 1974 but continued to publish and to be intellectually active for another 20 years. Glyn Daniel succeeded to the Disney Chair in that year, though both David Clarke and Colin Renfrew had been strongly tipped on that occasion. Colin Renfrew became Disney Professor following Glyn Daniel’s retirement in 1981. 3. Grahame Clark’s inaugural lecture contains a number of pointed attacks on his contemporaries, none of whom is mentioned by name but among which one can easily discern Gordon Childe, Glyn Daniel, Christopher Hawkes, and Stuart Piggott as particular targets. Hawkes in particular had advocated views almost exactly the reverse of those expressed here by Clark. It is clear that the tensions between generalization enforced on, or made possible in, earlier periods of prehistory by lack of data and the finer detail possible in text-aided archaeological studies of protohistoric and historic periods were a source of ongoing debate throughout much of twentieth-century British archaeological thinking (see Evans 1999). 4. Anthony Giddens was invited on an occasion in about 1980 to give a lecture to the Archaeology Department, shortly after the publication of his Central Problems in Social Theory (1979), in which he gave a characteristically lucid exposition of structuration theory, describing his theory as a logical theory and culminating with the statement that the Industrial Revolution represented for him the major discontinuity in human affairs. When it was pointed out to him that most of his audience were archaeologists working on periods long before the Industrial Revolution, which called into question the logic of translating principles that worked in a postindustrial world across such a discontinuity, he pleaded the usual excuse of a prior engagement and took his leave. It was not a problem that he had thought about or had any interest in. It was not his problem. It is, however, our problem as archaeologists, and it is up to us to find a solution. No one else can do it for us.

Notes 1. At first I was not much enthused by the term time perspectivism, which sounds like a rather cumbersome piece of potentially unnecessary and intellectually pretentious jargon, and had considered using the term time relativism (see Bailey 1981:113n2). However, I wished to avoid the connotations of relativism, in particular the self-contradictory view then gaining ground within archaeology, and one that has remained popular ever since, that “all truth is relative” (including of course that statement), an extreme version of the so-called theory dependence of observations (perhaps better described as the theory dependence of questions), which has too often been misused as an excuse for intellectual laziness and armchair theorizing and for a misguided belief that scientific method can be dispensed with and that evidence does not matter. Perspectivism, so I thought, had no precedents, and I eventually decided for that reason alone that the term was appropriate to characterize the distinctiveness of intellectual discourse required in an archaeology that aspires to the status of an autonomous discipline with its own type of data and concepts. I also rather liked the concept of time perspective, if not the noun perspectivism, and the optical metaphors implied by perspective, which seem to me exactly appropriate to the ideas under discussion. In fact there is a precedent for the use of the term perspectivism in the philosophy of Ortega y Gasset and some of the writings of Nietzsche, where it refers to different systems of beliefs, the relative validity of which cannot be established by recourse to any independent criterion (Flew 1983). That, of course, is exactly the point of time perspectivism, that there is no one time perspective that is more valid than any other. There may be a hierarchy of time scales but not a hierarchy of significance or validity. Where

30

3

Time Perspectivism and the Interpretive Potential of Palimpsests: Theoretical and Methodological Considerations of Assemblage Formation History and Contemporaneity Alan P. Sullivan III University of Cincinnati

Time is a nuisance—it is always in short supply, is virtually impossible to control or manage, and is invariably misspent by youthful foolhardiness. Cross-culturally, ruminations about the implications of the passage of time preoccupy humans seemingly endlessly (Munn 1992). In a timeless overview, Heirich (1964) enumerates the many ways that the concept of time itself has been manipulated to achieve different outcomes (see also Bender 2002; Bradley 2003), citing as one famous example Marx’s use of time as a resource in the determination of surplus value. That time reckoning has been implicated in various theories of cultural evolution has a long tradition in British and American archaeology, principally as a scale for sequencing the appearance of different attributes of human societies (Ramenofsky 1998; Trigger 2006). Mid-twentieth-century philosophers of science (e.g., Carnap 1966:78–85; Gardner 1976) wrestled with two problematic attributes of time that continue to have profound consequences for contemporary archaeological inquiry: (1) time is anisotropic, or asymmetrically unidirectional (e.g., Ear­man 1967; Grun­baum 1969), and (2) establishment of causal relations between successive events is challenging on logical and empirical grounds (e.g., Maxwell 1969; Rosenberg 1975; also Binford 1998:1). The first attribute, which refers to the irreversibility

of Time’s Arrow, is problematic because archaeologists endeavor to investigate unobservable ancient cultural dynamics by analyzing observable properties of the contemporary archaeological record (Fritz 1972) and hence boldly endeavor to transpose Time’s Arrow (DeBoer 1983). The second attribute, which concerns the uncertainty of causal connections between behavior and the content of the archaeological record, is problematic, as well, because similarities among archaeological phenomena are often the equifinal results of different sources of variation (Salmon 1992; Sullivan 2007). In this chapter, I focus on the degree to which, within the framework of time perspectivism (Bailey 2007), various conceptions of palimpsests affect archaeologists’ ability to infer ancient behavior (i.e., to reverse Time’s Arrow) and to ascertain the factors that influence assemblage formation (i.e., to reckon causality). The usefulness of several time perspectivism concepts is illustrated with analyses of regional survey data and a catastrophically abandoned, mid-eleventh-century settlement located near the Grand Canyon in north-central Arizona.

Epistemological Foundations In exploring the “effect of time concepts on archaeological interpretation” (Bailey 1983:165), it is worth considering the prospect that the ­apprehension of 31

Alan P. Sullivan III archaeological time is space based (e.g., Moberg 1971:​557). That is, in the absence of demonstrable anthropogenic changes in the physical and spatial characteristics of matter, we would have no archaeological phenomena to investigate (Sullivan 2008). In other words, the very perception of the archaeological record, and the historical events that are bound up in its content, is premised on applying criteria that distinguish human-caused anomalies in geological settings (Kvamme 2006). It is largely because humans have developed an incomparable ability to move matter and alter its properties that archaeologists can even presume to study ancient behavior (Ascher and Ascher 1965; Binford 1983b:​19). Fortunately, the by-products of these rearrangements create spatial and geophysical traces that implicate time. Considerations of the theoretical bases of time reckoning are certainly not without precedent in American archaeology (Nash and Dean 2000). In the context of time perspectivism, however, two American contributions to space-based conceptions of time, which have consequences for understanding the origins of the archaeological record, stand out. In 1971, Richard A. Krause and Robert M. Thorne published a provocatively titled ­essay, “Toward a Theory of Archaeological Things.” Not widely cited, surprisingly, this dense article illustrates a method for determining time relations among anthropogenic traces in the archaeological record (cf. Lucas 2005:39). For instance, postholes must be dug before posts are inserted into them (an example of a “noncommutative” relation), but it matters little which posthole is dug first (an example of a “commutative” relation). From an analysis of commutative and noncommutative relations of traces, Krause and Thorne argue, formation histories of archaeological phenomena can be developed (cf. Harris 1979). Extending Krause and Thorne’s logic, David R. Wilcox, in a series of studies involving prehistoric puebloan ruins (Wilcox 1975, 1982), nonpuebloan pithouse villages (Wilcox et al. 1981), and macroregional systems (Wilcox 1996:242) in the American Southwest, proposes an approach to isolate the remains of different social groups using the concepts of “living surfaces” and modifications

(“episodes”) to them (see especially Wilcox et al. 1981:​152–153). Building on these developments, I have explored the theoretical implications of the trace concept, which refers to a demonstrable “alteration in the physical properties of an object (or the relations between objects) or a surface (or the relations between surfaces)” (1978:194). Another useful concept is that of trace load, which refers to the accumulation of traces that pertain to a specific archaeological phenomenon (e.g., the number of artifacts that repose on the occupation floor of a structure). Importantly, the integration of these concepts creates an “ex post facto time perspective” (Wilcox 1975:​127), whereby the establishment of sets of commutative and noncommutative relations provides essential data for inferring the formation histories and establishing the contemporaneity of archaeological phenomena.

Time Perspectivism, Refuse Types, and the Archaeological Record With the aforementioned concepts in mind, consider the time perspectivism basis of primary refuse, secondary refuse, and de facto refuse (Schiffer 1972). In the original definitions of these terms, as well as in subsequent (minor) revisions (Schiffer 1976, 1987), the critical distinction among them is space, that is, whether the locations of use, discard, and abandonment change during an artifact’s—not a deposit’s—life history (see also DeBoer 1983:22). The slippage in the unit of analysis—from artifact to deposit—is responsible, in my view, for the exchanges between Binford (1981a) and Schiffer (1985) regarding the “structure” of the archaeological record (see also Lucas 2005:​34–35; Murray 1999a). Whereas Schiffer (1985) declares that the archaeological record is a distortion of systemiccontext­ dynamics involving material culture, Binford contends that it is “most likely a structured consequence of the operation of a level of organization difficult, if not impossible, for an ethnographer to observe directly” (1981a:​197–198). These largely incompatible views about the nature of the archaeological record make appeals to claims of knowledge that are essentially unverifiable (cf.

32

Time Perspectivism and the Interpretive Potential of Palimpsests Kosso 2001:​55). After all, which characteristics of an artifact or a deposit differentiate a “distortion” from a nondistortion of human behavior, and which distinguish a “structured” from a nonstructured consequence of organization? Adoption of space-based time concepts avoids these conceptual conundrums, particularly if the archaeological record is viewed as the contemporary expression of the cumulative aggregation of traces, some of which are attributable to human actions and some, not (Clarke 1973). One task of the archaeologist, therefore, is to parse traces according to the circumstances that produced them using the logic of commutative and noncommutative relations (e.g., Chapman and Gaydarska 2007).

Bailey (2007:​204–207) refers to as cumulative and spatial palimpsests where multiple and discontinuous episodes of artifact deposition resulted in the formation of extensive scatters of material (largely fragmentary) directly on the ground’s surface (see also Sullivan et al. 2003). For these cases, the units of analysis were areas defined by variation in artifact-density clines, and the trace loads were numbers of artifacts per unit area. Hence, variation among artifact-scatter palimpsests is attributable to differences in the number of occupations, often by different groups, although the duration of each episode, though undoubtedly short, is not ascertainable.

A Palimpsest of Palimpsests

Palimpsests in Archaeological Interpretation

In contrast to artifact-scatter palimpsests, which are often consequences of “successional use” (BinSince its introduction (Binford 1981a), the term pa- ford 1981a:200), abandoned structures afford oplimpsest has achieved prominence and centrality in portunities to explore the interpretive potential the conceptual tool kit of time perspectivism (Bai- of different types of palimpsests within a singleley 1983). In common practice, a palimpsest “re- duration deposit. For example, consider the case fers to a superposition of successive activities, the of Site 17 (Sullivan 1986), located just south of material traces of which are partially destroyed Grand Canyon National Park in northern Arizona or reworked because of the process of superposi- (Figure 3.1). This comparatively small settlement, tion” (Bailey 2007:203). Of the palimpsests Bai- which consisted of a single, semisubterranean maley (2007:​203–208) distinguishes, I focus on four sonry structure and three near-surface, low-walled types​—​true, cumulative, spatial, and temporal​—​ wooden structures (Figure 3.2), had been abanwhose empirical consequences interdigitate with doned abruptly because of a catastrophic conflamy view of the archaeological record and entail gration, perhaps attributable to a passing wildfire the concepts introduced earlier (consequently, I (Roos et al. in press), that consumed it. Burned exclude “palimpsests of meaning” [Bailey 2007:​ beams and architectural debris had collapsed di207–​208]). The basic proposition is that, by under- rectly upon all structure floor surfaces and their arstanding the extent to which surfaces and matter are tifact arrays, thereby sealing them (Figure 3.3). contemporary, archaeologists can rationally adjust Of 33 cutting dates determined from the dentheir scale of inquiry and make it relevant for ad- drochronological analysis of 170 samples, 32 clusdressing different questions about the nature of the ter between ad 1054 and 1064. Thus, not only is cultural past. A consideration of palimpsests, then, the duration of occupation well established, but provides a means to understand how different as- so are the “build dates” for the settlement’s strucpects of assemblage formation pertain to establish- tures and features (Figure 3.4). That is, based on the ing contemporaneity and vanquishing equifinality. presumption that the earliest cutting dates within For example, my previous investigations of pa- modes mark construction events, Structure 4 was limpsests have focused on inferring the formation built in 1054, a pottery-production facility (Feahistory of lithic artifact scatters (Sullivan 1992a) tures 8 and 38 in Figure 3.2 [Sullivan 1988]) was and sherd and lithic artifact scatters (Sullivan installed in 1055 (not shown in Figure 3.4; see Sul1995b). In these studies, I was dealing with what livan 1986:​140–141, 351), Structures 2 and 3 were

33

Alan P. Sullivan III

Figure 3.1.  Location of the Upper Basin Archaeological Research Project (UBARP) study area in northern Arizona.

erected in 1056, and Structure 1 was constructed in 1057. With this intrasite chronology for one excavated settlement, as well as with systematically ­recorded surface assemblage data from 242 other single-component ruins in the Upper Basin, I intend to demonstrate that even short-term, continuously occupied settlements are palimpsests of palimpsests, that different types of palimpsests register different assemblage formation histories, and that assorted palimpsests vary in their inferential potential.

by this building event persisted, however, as the structure continued to be used until ad 1064, during which time a considerable trace load of features and artifacts built up in the remodeled space. In contrast, Structure 2 appears to have fallen into disuse after ad 1061; consequently, no evidence is available attesting to the nature of any activities, other than two vessel fragments and one unifacially retouched tool deposited sometime between ad 1056 and ad 1061. Hence, the ambiguous commutative relations expressed by the low trace load in Structure 2, particularly in comparison with that of Site 17 as a True Palimpsest Structure 1, illustrate that true palimpsests, within True palimpsests, according to Bailey (2007:203–​ the same site, vary considerably in their interpre204), are those where traces of earlier behavior have tive potential regarding assemblage formation and been eradicated by the activities of later occupa- surface contemporaneity. tions. At Site 17, we have two examples of true paSite-Specific and Regional limpsests. First, the bonding-abutting relationships Cumulative Palimpsests of walls (Wilcox 1982) revealed in Figure 3.5 show that when Structure 1 was rebuilt in ad 1061, any A cumulative palimpsest is one that materialevidence of space and surface use prior to ad 1061 izes as traces accumulate and become comminwas obliterated when partitioning walls were con- gled from “successive periods of deposition” (Baistructed. The noncommutative relations expressed ley 2007:​204–​205). The 247 lithic artifacts (flaked 34

Figure 3.2.  Plans and profiles of four burned structures at Site 17 (AZ I:1:17 [ASM]), which has been tree-ring dated to ad 1054–1064. MN is the abbreviation for Magnetic North.

Alan P. Sullivan III

Figure 3.3.  View of a portion of Structure 4 at Site 17, in the initial stages of excavation, showing burned roof beams resting directly on the structure’s floor and floor-contact artifacts, such as the metate in the upper right.

stone [n  = 245] and ground stone [n = 2]) and 5.28 kg of ­ceramics (n = 780 sherds) collected from 736 m2 of Site 17’s surface can be considered a ­cumulative ­palimpsest, therefore, because these objects amassed at different times during 10 years of perennial occupation. It is unclear, however, whether there is any interpretive value in categorizing this material as “trash” or a type of refuse. Abundant lithic artifacts recovered from exterior surface contexts, for example, imply that they likely were discarded where they were produced and possibly used, which would cause this cumulative palimpsest and its associated artifacts to be regarded either as primary refuse or as de facto refuse when accumulation terminated abruptly when the settlement was consumed by fire.

These refuse-type ambiguities notwithstanding, the interpretive potential of this cumulative palimpsest, and of cumulative palimpsests generally, takes on enhanced meaning at the regional scale of analysis. For instance, the composition of Site 17’s surface ceramic assemblage (75.9 percent) and lithic assemblages, as measured by four general categories (debitage = 22.0 percent, tools = 1.9 percent, metates = .00 percent, and manos = .20 percent [cf. Diehl 1998]), is typical when compared with surface cumulative palimpsest assemblages associated with 242 masonry structures in the ­surrounding area (Figure 3.6; mean ceramics = 70.3 percent; mean debitage = 28.0 percent, mean tools = .80 percent, mean metates = .30 percent, and mean ­manos = .10 percent). Moreover, several ­significant 36

Time Perspectivism and the Interpretive Potential of Palimpsests

Figure 3.4.  Histograms of structure-specific tree-ring dates from Site 17. Construction events are inferred when cutting dates first occur in modes.

o­ bservations can be made about the assemblage content of these palimpsests: (1) flaked-stone tools and ground-stone artifacts are uncommon regardless of the number of rooms per structure; (2) cumu­ lative palimpsest assemblages associated with structures with three or more rooms are more variable than those associated with one-room and two-room structures; and (3) one-room and two-room structures have significantly higher percentages of debitage (Mann-Whitney U = 1,858, Z = –3.046, p = .002), whereas structures with three or more rooms have significantly higher percentages of ­ceramics (Mann-Whitney U = 1,920, Z = 2.866, p = .004).

These patterns based on categorical data carry over largely when the artifact density, which is a space-based measure of variation in trace loads, of these surface cumulative palimpsests is examined (Figure 3.7). That is, the mean densities of ­ceramics and debitage are far more variable for cumulative palimpsests associated with structures with three or more rooms than they are for one-room and two-room structures, and mean ceramic density is significantly higher for the larger structures (MannWhitney U = 1,898, Z = –2.93, p = .003). Such patterned differences imply that (1) lithic artifacts accumulate in cumulative palimpsests at a steady 37

Figure 3.5.  Plan of Structure 1 at Site 17 showing abutments of two parallel walls (left of “Hearth”), which are interpreted as evidence of a rebuilding event in ad 1061; placement of five ceramic vessels in and on the lip of a deep thermal feature (“Hearth”); and locations of several other floor-contact artifacts. MN is the abbreviation for Magnetic North.

Time Perspectivism and the Interpretive Potential of Palimpsests

Figure 3.6.  Bar chart of the assemblage composition of surface cumulative palimpsests for 242 masonry structures in the Upper Basin. The percentages are based on the total number of artifacts recorded in 5-mdiameter enumeration units that were placed in the densest concentration of each surface cumulative palimpsest. The percentage values for each class were averaged across all 242 structures; error bars are the 95 percent confidence intervals for each artifact class, broken down by the number of rooms per structure.

Site 17 as a Spatial Palimpsest

rate regardless of structure size; (2) ceramic artifact breakage and sherd accumulation intensify in exterior areas of structures with three or more rooms (perhaps as “founder” vessel use-lives expire at roughly the same time [Tani and Longacre 1999]); and (3) although the elapsed occupation spans of the larger three- to six-room structures may be longer than the smaller one- to two-room structures, both classes of structures are still accompanied by significant variation in duration (as measured either by overall density or ceramic density; Figure 3.8). At the least, it seems that surface cumulative palimpsest assemblage differences, measured categorically or in terms of density, quantify regional variation in occupation duration, which is an aspect of ancient behavior that cannot be determined from other types of palimpsests.

A spatial palimpsest arises when artifact deposition occurs at (roughly) the same time in different locations within the same settlement (Bailey 2007:​ 207). A comparison of the lithic artifacts recovered from Site 17’s surface and structure floors provides a compelling example of some of the methodological challenges entailed in understanding the factors that influence the formation of assemblages recovered from different contexts. For instance, whether lithic assemblage composition (as measured by percentages of debitage, tools, metates, and manos) is calculated with a context-specific denominator (n = 247 and n = 33, for surface and structure floors, respectively) or with a site-wide denominator (n = 280), the percentages of surface lithics remain relatively stable in contrast to those for structure floors,

39

Alan P. Sullivan III

Figure 3.7.  Bar chart of mean artifact density of surface cumulative palimpsests for 242 masonry structures in the Upper Basin. The values are based on the total number of artifacts recorded in 5-m-diameter enumeration units that were placed in the densest concentration of each surface cumulative palimpsest. The density values for each class were averaged across all 242 structures; error bars are the 95 percent confidence intervals for each artifact class, broken down by the number of rooms per structure.

which vary dramatically (Figure 3.9). This degree of statistical volatility implies that surface cumulative palimpsest assemblages register long-term consequences of human activities, floor-contact assemblages measure short-term technological configurations, and there is little correspondence between surface and subsurface lithic assemblage composition, particularly with respect to groundstone artifacts (Downum and Brown 1998).

faces and ­artifacts, were merged into a single deposit (yielding a time-of-abandonment artifact assemblage of 399 pieces of debitage, 43 tools, 3 cores, 5 metates, 14 manos, and 32.06 kg of ceramics). In a catastrophically abandoned settlement, such “Pompeii-like” situations are often thought to provide snapshots of ancient life, particularly the organization of daily activities (Diehl 1998). Conventionally, variation in the trace loads of floorcontact artifact assemblages has been attributed to functional differences (e.g., Ciolek-Torrello 1984), to abandonment behavior (Diehl 1998), or to combinations of such factors (Lightfoot 1993). Because the structure floors at Site 17 had been sealed by the catastrophic fire that swept the settlement, it is unlikely that postabandonment scavenging affected the assemblage composition of any floor ­array (Tomka 1993).

Site 17 as a Temporal Palimpsest According to Bailey, a temporal palimpsest is “an ­assemblage of materials and objects that form part of the same deposit but are of different ages and ‘life’ spans” (2007:207). When Site 17 was destroyed by fire, all the traces and trace loads of each structure, such as the floor-contact artifact arrays, exterior features, and previously abandoned sur40

Figure 3.8.  Box-and-whisker plot showing the median value (line inside the box), range of 50 percent of the cases (represented by the length of the box), and outlier values (represented by circles and asterisks) for artifact density, ceramic artifact density, and lithic artifact density of surface cumulative palimpsests associated with 242 masonry structures in the Upper Basin.

Figure 3.9.  Bar chart of lithic assemblage composition at Site 17 illustrating differences in variability between spatial palimpsests (surface versus structure floors) that arise when either a context-specific or a site-wide denominator is used to calculate percentages of the various artifact classes.

Figure 3.10.  Plan of Structure 3 at Site 17 showing arrangement of artifacts on the floor at time of abandonment. MN is the abbreviation for Magnetic North.

Time Perspectivism and the Interpretive Potential of Palimpsests With these understandings, Structure 1 (Figure 3.5) could be interpreted at time of abandonment as a “habitation” room where cooking occurred ­because smashed ceramic vessels were found re­ posing in or on the lip of a deep thermal feature; otherwise, the floor was virtually devoid of artifacts. At time of abandonment, Structure 2 already had fallen into disuse and, being a true palimpsest, provides no data regarding the nature of activities that once may have been conducted there. Structure 3 (Figure 3.10) could be considered another type of “habitation” room (Schiffer 1985:21) at time of abandonment in view of the comparatively large amount of available “free” floor space, a complete “pitcher” (VN81), and a complete jar (VN66); however, no “cooking” or other form of processing appears to have transpired because all floor-​contact­ ceramic artifacts, including three fragments (VN64, VN70, and VN71), appear to have been abandoned in “cached” mode or provisional storage around the periphery of the floor (Deal 2008). Finally, Structure 4 would be regarded at time of abandonment as a “storage and processing” room in view of the comparatively large numbers of complete artifacts representing all major artifact classes (Figure 3.11). Moreover, because of the lack of redundancy among floor-contact artifact arrays at time of abandonment, these surfaces and their ­variable trace loads would be considered economically and organizationally complementary (Lightfoot 1994:​119–120). That is, they were intended to secure and maintain the livelihood of the social unit that was responsible for creating the surfaces and introducing objects upon them. That said, the time perspectivism concepts introduced above provide an alternative view of how these floor-contact artifact array differences arose and what their variability means. In Structure 1, for example, the placement of whole ceramic vessels is feature dependent and kinetic-activity specific (cooking). That is, the floor, the floor-contact artifacts, and the most recent thermal feature in Structure 1 are spatially contemporaneous and behaviorally related. In contrast, the arrangement of complete and fragmentary ceramic artifacts in Structure 3 (where no ground-stone or flaked-stone artifacts were 43

r­ ecovered) is feature independent and artifact specific (storage). Actually, this floor-contact assemblage appears to have formed as a coincidental aggregation of artifacts (I would venture that the fragmentary ceramics, in particular, had been consolidated there in anticipation of being transported and used off-site [see Sullivan 1992b]). Hence, these artifacts and the surface upon which they reposed, as well as the shallow thermal feature, are spatially contemporaneous but behaviorally unrelated. The arrangement of artifacts in Structure 4 is feature independent and either kinetic-activity specific (processing) or artifact specific (storage). That is, the cluster of ground-stone artifacts and ceramic vessels in the western half of the structure and the surface upon which they rested all seem to be behaviorally interrelated but are unrelated to the shallow thermal feature. In contrast, the artifacts in the eastern half of the structure are behaviorally unrelated to each other and to any activities that may have occurred elsewhere on the floor or in or around the thermal feature (i.e., they appear to be cached or in provisional discard). Hence, the floor-contact artifacts in Structure 4, although spatially contemporaneous with respect to a common occupation surface, represent an “internal” spatial palimpsest. Based on these patterns, I infer that, at time of abandonment, each floor-contact artifact array at Site 17 had encoded a different if not unique formation history that emerged in response to short-term situational factors. If so, this time perspectivism account of spatial variation in arrangements of matter illustrates that care must be exercised in inferring “room function” by “use averaging” artifact classes. For the most part, object arrays on structure floors at Site 17 materialized independently of the attributes of the structures themselves, were not “designed in” when the structures were built, and did not become “locked in” until the settlement was catastrophically destroyed.

Conclusions From the viewpoint of time perspectivism, site-specific and regional data from the Upper Basin support the idea that at least two space-based clocks track variation in the accumulation of matter. The

Figure 3.11.  Plan of Structure 4 at Site 17 showing arrangement of artifacts on the floor at time of abandonment. MN is the abbreviation for Magnetic North.

Time Perspectivism and the Interpretive Potential of Palimpsests first clock, represented by cumulative palimpsest and spatial palimpsest surface assemblages, records aggregate trace loads that accrue season after season, year after year. In contrast, the second clock, represented by the temporal palimpsest of floor-contact artifact arrays, records short-term variation in trace loads that is caused by the situational deployment of artifacts upon surfaces where, with each reconfiguration, the trace clock is reset. Together, these two clocks measure how different causal factors influence assemblage formation; affect the contemporaneity of surfaces, features, and artifacts; and facilitate the development of inferences regarding the spatial (and, hence, temporal) consequences of human behavior and technological organization. Time perspectivism, particularly its thesis that it is essential to develop appropriate concepts and units of analysis for different problems, compels archaeologists to rethink the basic premises of archaeological inquiry—in short, it is a long overdue call for clarifying and updating our epistemology. Anyone familiar with David L. Clarke’s (1968) approach to archaeology is aware of the centrality of these issues that, lamentably, often are set aside in the rush to make archaeology more anthropological or have it resonate more convincingly with other social sciences (Dunnell 1982). Surveying the landscape of theoretical development since “Archaeology: The Loss of Innocence” (Clarke 1973) was published, it is clear that we are only incrementally closer to realizing aspects of Clarke’s ambitious

agenda. To accelerate the process, my principal objective in this study has been to illustrate the usefulness of a set of time perspectivism concepts for inferring the formation histories of archaeological phenomena by means of space-based units of analysis. Such concepts are key elements of an independent archaeological theory that allows us to pursue the investigation of the cultural past without relying on problematic assertions about the meaning of refuse types, the usability of artifacts, intentionality, or the structure (or lack thereof ) of the archaeological record.

Acknowledgments The Upper Basin Archaeological Research Project has been supported by grants from the University Research Council (University of Cincinnati), the C. P. Taft Research Center (University of Cincinnati), the McMicken College of Arts and Sciences (University of Cincinnati), the Center for Field Research (Earthwatch), the National Center for Preservation Technology and Training (National Park Service), and the U.S. Department of Agriculture Forest Service (Kaibab National Forest). I thank John A. Hanson, Kaibab National Forest archaeologist (now retired), for his long-term intellectual and administrative support of the Upper Basin Archaeological Research Project. I also thank Dr. Sissel Schroeder, Department of Anthropology, University of Wisconsin, for helping me at crucial stages with various aspects of this study.

45

4

Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation Michael J. Shott University of Akron

“Time perspectivism” is an awkward expression for an important archaeological problem. It denotes the view that different parts of the material record formed at different rates and therefore over different time scales (Bailey 1983). Time perspectivism implies that the archaeological record is not commensurate with the ethnographic record or its customary time scales. If formation theory concerns how the record is formed (Shott 1998), then time perspectivism is one of its chief corollaries, suggesting ways to calibrate formation processes to archaeology’s vast time spans.

tional causes—​“early man ...doing different things to different degrees at given places and times” (Kleindienst 1961:​44)—or, alternatively, normative or iconic causes, that is, the view, questioned by Isaac, Harris, and Kroll, that “each clearly distinguishable assemblage of artifact forms must represent a distinct culture” (1997:262). Eventually, some archaeologists reinterpreted the debate about assemblage variation by altering its terms and rejecting premises shared by the protagonists (Dibble and Rolland 1992; Shott 2003). They stopped thinking of assemblages as categorical things, whether ethnic calling cards or tool kits. Instead, assemblages develop at places where variable, complex relationships between tool classes, relationships that crosscut the boundaries erected between assemblage “types,” were played out. Assemblages were reimagined not as things—“solid bricks”—but as contexts constantly in a state of becoming but never arriving, in fixed, essential form. This status implicates time in both assemblage formation and variation among assemblages. Accumulation span is therefore a source of assemblage variation.

The Nature of Assemblages Many archaeologists regard assemblages as fixed entities in concept—Clarke’s “discrete and solid bricks” (1978:​35)—that possess intrinsic meaning. By the presence of particular tool classes (“index fossils”) or the proportional abundance of classes, assemblages represent the norms or icons of identity-conscious groups or the kinds and proportions of activities conducted using their constituent tools. Bordes’s Mousterian facies exemplified the former view; Binford’s tool kits, the latter. The same assemblages had different meanings, depending on the view. The Mousterian argument occasioned by this clash of views flared briefly and then languished unresolved for years before passing into oblivion (Shott 1996). The same argument persisted, the same resolution evaded, in many other settings. For instance, interpretations of East African hominid assemblage variation stressed func-

Assemblage Boundaries in Practice Like the concept itself, empirical assemblages must be defined. Often, we consider whatever we find at a “site” as an assemblage. In stratified deposits, obviously, vertical boundaries define assemblages, as though natural limits give legitimacy naturally. But such boundaries may be as apparent as real; the 46

Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation distribution of tools across strata depends on the changing relationship between the pace of geological and archaeological deposition. However defined in cultural terms, distinct assemblages may be conflated in one stratum or a single one arbitrarily subdivided between strata. Many assemblages lack natural horizontal boundaries, so archaeologists must draw them. Often we draw boundaries according to artifact distribution or density, whether defined rigorously or not. This can yield problematic results, because boundaries are not always legitimate, but once defined, they are nearly always treated as such (Shott 2004). Always, though, we draw boundaries, a practice that seems so fundamental as not to bear serious thought. Boundaries can be useful depending on context and purpose but are not always. As against fixed boundaries, archaeologists might define assemblages at different scales and examine how their properties vary with scale. Shifting boundaries require us to abandon the “site” concept, at least for some purposes. Besides the agony of its definition, site bears needless ontological baggage (Dunnell 1992). It is a construct imposed, often with synchronic ethnographic inspiration, on a record that is continuous, if variable, across the landscape. Archaeology has come gradually to this realization; it remains to reconceive assemblages and their boundaries free of the strictures of “site.” Among other things, the reimagining means that assemblage properties are not fixed but, in fact, scale dependent, a quality observed in empirical data (Ebert 1992; Holdaway et al. 1998).

tions, and indirectly in derived measures such as richness (number of classes, not tools), evenness (range or variation in proportions across categories), and heterogeneity (a composite measure of richness and evenness). Composition apportions specimens into constituent classes. In faunal and paleontological assemblages, species or higher taxa form analytical classes. Biological taxonomy has its own ontological problems but largely establishes the boundaries of constructed analytical units in contemporary life that possess integrity; giraffes are not tigers, nor can they become them. But depending on how we define them, stone tools may violate the assumption of class integrity. In use, tools experience resharpening to maintain their edges. Extensive use entails extensive reduction, which trivially reduces size but also changes tool form. To the extent that typology is based on form, reduction can change a tool’s typological status. Reduction effects are well documented ethnographically (e.g., Hayden 1977, where large flakes passed through progressive reduction stages that correspond to defined classes such as denticulates). Typological effects of reduction also are documented in Middle Paleolithic industries (Dibble 1995), widely in prehistoric Australian assemblages (Hiscock and Attenbrow 2003), and in North American bifaces (e.g., Andrefsky 2006; Shott and Ballenger 2007), among other examples. I (2005) have discussed reduction effects in general. The integrity of stone tool classes must be demonstrated, not merely identified with apparent modes in size and form. In considering how assemblages formed via the interaction of tool classes made, their varying degrees of use, and the uses to which they were put (Ammerman and Feldman 1974), we begin to appreciate the complexities of time averaging via assemblage formation. As a simple illustration, imagine one person who uses five classes of tools, one of each at a time; that the classes are used the same amount; that time passes in unit increments; and that class longevities differ: the first having a uselife of one time unit, the second, of two time units, and so on. The person occupies a series of places for varying lengths of time, doing exactly the same thing and using the same classes of tools in the same

Assemblage Properties However defined or at least circumscribed, assemblages possess certain intrinsic properties. These include size, the number of objects they contain, and composition, the proportions of objects in various classes. Analysis to follow suggests that time is a third intrinsic assemblage property. Fundamental as they are, size and composition are not easily measured. Proper size quantification—determining the number of original wholes from combinations of broken and intact tools—is a serious problem, not a simple matter (Shott 2000). Composition is measured directly in counts, in relative propor47

Michael J. Shott frequencies at each. But a place occupied for one time unit will witness the discard of one specimen of the shortest-lived class. A second place occupied for two time units will witness the discard of two specimens of the shortest-lived class and one of the next shortest-lived one. Tool-using behavior is constant, yet assemblage size is three times larger over an occupation span twice as long. Carrying through to a place occupied for five time units, assemblage measures such as richness correlate with assemblage size in the absence of any variation in the behavior that underlies the tool discard and, hence, assemblage formation (Shott 1997:213, Figure 5). Instead, variation is the joint product of differing class uselives and accumulation spans of places. This is time averaging. The example is hypothetical, but !Kung ethnographic assemblages in which tool use-lives, occupation span, and behavior are known show the same effect (Shott 1989). There, the composition of small assemblages accumulated over short periods was not correlated with the near-constant behavior that produced them because occupation spans were much smaller than the average use-lives of the tools used and discarded. Instead, assemblage composition followed a size-dependent trend. Only when assemblages are pooled does a correlation emerge between the frequencies of tools and toolusing behavior, because pooling artificially creates, as it were, an accumulation span that exceeds tool use-lives. Similar effects occur in North American Great Lakes Paleoindian assemblages, where end scraper proportion correlates positively with uselife and fluted biface proportion does not (Shott 1997:216–217). Here too some combination of use rate and use-life determines patterns. Time averaging has a great implication for assemblage analysis because it suggests that assemblage size and composition are not independent but correlated. That is, composition can vary with assemblage size. Correlation owes much to assemblage formation processes, including the relationship between tool use rates and use-lives, on the one hand, and occupation span—a measure of time​ —​or assemblage size, on the other. Deterministic simulations show that composition changes as occupation span and assemblage size increase, except 48

in the unlikely event that use-lives are identical between tool classes (Dibble and Rolland 1992:Figure 1.3; Shott 1997:213–215). Kleindienst (1961:40) has hinted at size dependence in Paleolithic assemblage composition in East Africa. Patterning in archaeological data that range widely in time and space suggests the same effect (Shott 1997, 2003). Composition often changes as assemblage size increases. Size effects like these on assemblage composition are not an irritating complication to be eliminated before interpretation begins. Instead, they possess intrinsic value because they register the kind and strength of cultural or technological organization; the nature of local group formation and affiliation rules; or the play of formation processes in the interaction of occupation span, activity, and material culture (Shott 1997, 2003). Thus, “size dependency, always considered an obstacle, is actually a missing requirement for successful ...analysis” (Hayek and Buzas 1997:378). Size effects mediated by time must be explained, not explained away.

Interaction of Assemblage Formation and Landscape In contemplating sets of roughly coeval assemblages, we commonly assume that they are linked in time or behavior. We thus accommodate assemblages to our preconceptions. That is, we assume that they all formed part of the same settlement or behavioral system. Then we consult ethnography or common sense to identify idealized site types amid the assemblages. Undeniably, ethnographic sources describe people doing different things at different places; hunter-gatherers and others used home bases, scouting places, and kill and other sites across their landscapes. But rates of ethnographic occupation or use are not equivalent to rates at which archaeological assemblages accumulate (Dewar and McBride 1992) any more than, say, little Johnny digging up a bucket of sand at the beach one fine summer day causes coastline regression on Cape Cod. However tidy land use and settlement may be in ethnographic context, itself an arguable proposition, prehistoric settlement was manifestly complex over longer periods. The sets of artifacts that we define as assemblages often accumulated in

Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation many visits to a place, not one, by groups of varying size, composition, and purpose, whose tool-using activity differed between visits. No doubt some artifacts even circulated between places. Yet we treat assemblages as the products of single brief occupations by one group of fixed size and composition, which used tools in one set of activities. Archaeological assemblage variation is often much more complex than ethnographic models accommodate.

along with several rare classes. This scheme remains in common use with modification for ­local circumstances and research questions (e.g., Isaac et al. 1997:​Table 6.1, 267, where two subsequent schemes are assimilated into Leakey’s original and spheroids and subspheroids are omitted as rare at Koobi Fora). Whatever the sources of variation in artifact size and form, Leakey’s “system greatly facilitates comparison of the artefactual material from Koobi Fora Koobi Fora Lower Okote with that from Olduvai, which is by far the largest Member Assemblages and most important body of data concerning stone We must disentangle the separate effects of short- artefacts” (Isaac et al. 1997:264) in the East African term activity variation and long-term assemblage Lower Paleolithic. Isaac and colleagues (1997:267, formation. This issue arises in many contexts, Tables 6.1–6.2) retain Leakey’s scheme while proamong them a recent example that concerns the posing a higher-order distinction between “flaked meaning of variation in Paleolithic assemblages. pieces” and “detached pieces.” The former encomThe question is simple: Does variation relate to dif- passes practically all tool classes; the latter, mostly ferent activities conducted at different places or to debris or debitage. As legitimate as the distinction how assemblages formed over varying but long in- is on technological grounds, it obscures patterning tervals, especially to the joint agencies of accumula- among tools; useful for some purposes, it is irreltion and assemblage size? evant to the study of proportional differences beThe archaeological example is from Stern’s tween tool classes, however they are defined. (1993) study of assemblage variation in the Lower Originally, Lower Paleolithic classes were reOkote Member (LOM) at Koobi Fora in Kenya. garded as distinct formal and functional categories: Maximum age of LOM is circa 1.64 ± .04 million if a tool was fashioned as, for example, a chopper, years; thickness and assumed sedimentation rate it was used and discarded without undergoing subsuggest that it accumulated over about 90,000 stantial change in its size or form during use; and years (Stern 1993:​209–210). Broadly, LOM as- choppers were used for some purpose(s), discoids, semblages are of Lower Paleolithic age and charac- for qualitatively different ones, and so on. Like giter. But the Lower Paleolithic was itself variable in raffes and tigers, classes were essential units that both time and space. Compared with earlier assem- possessed formal and functional integrity. Paleoblages in the Omo Turkana Basin, LOM assem- lithic archaeologists know now that these classes blages contain “new, distinctive tool types” whose partition a joint metric–technological continuum deposition was more structured and therefore clus- determined by raw material, by industrial variatered across the landscape and not so tethered to or tion, and significantly, by degree of reduction of governed by local abundance of stone (Potts 2001:​ original, unmodified cobbles (Cormack 1996; Isaac 141, Figure 3). et al. 1997:​264–265; Ludwig and Harris 1998; Potts 1991; Sahnouni et al. 1997; Toth 1985). Typology and the Legitimacy of Classes In the reduction perspective, therefore, some African Lower Paleolithic typology defines tool Lower Paleolithic classes emerge as arbitrary segclasses by combinations of artifact size, form, and ments of a reduction continuum of cobbles to spent industrial character (Isaac et al. 1997:262; Klein­ tools and cores: “These forms can often grade into dienst 1961:38; Leakey 1971). Leakey (1971:4–8) one another during reduction” (Toth 1985:​107; see defines the following basic classes: chopper, biface also Isaac et al. 1997:​271–272). Reduction schemes and protobiface, polyhedron, discoid, spheroid differ in details, but Potts (1991:Figure 1) and Sahand subspheroid, battered nodule, and scrapers, nouni et al. (1997:​Figure 3) agree that ­spheroids 49

Michael J. Shott

Figure 4.1.  Possible reduction relationships between Lower Paleolithic tool types.

might simply be exhausted cores that passed through intermediate stages as “choppers” and “polyhedrons” (see also Ludwig and Harris 1998:​ 99). Cor­mack (1996:​43–​44) has linked polyhedrons, subspheroids, and spheroids in a reduction sequence. Some archaeologists demur on grounds of raw material; polyhedrons and spheroids may sometimes be fashioned from different toolstones, thus forming separate reduction sequences ( Jones 1994:​276; Kimura 1999; Willoughby 1985:​56). Still, a growing consensus argues for subsuming them in a single sequence. Figure 4.1 synthesizes proposed reduction relationships among Lower Paleolithic classes. We should not distinguish tool classes unless their differences are demonstrated. Obviously, it would be unfair to criticize earlier studies for fol-

50

lowing typological conventions; criticism is intended to improve future practice, not to belittle the past. Nevertheless, the Lower Paleolithic typology used in East Africa essentially subdivides reduction continua in separate classes of core and flake tools. If faceted spheroids are reduced choppers, then at least for some purposes it may be mistaken to distinguish them. Instead, all might be placed in a larger class of core tools. Such treatment does not deny possible functional or other variation during reduction, that slightly reduced cores qua “choppers” were not used differently than heavily reduced cores qua “polyhedrons.” Nor, conversely, does the recognition of different classes establish their functional significance. Whatever the function(s) of cores and cobble tools during reduction, traditionally defined classes linked in

Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation a ­reduction continuum can be treated as a single larger class for analytical purposes, which has the added virtue of revealing unsuspected relationships between reduction and assemblage composition. Such analytical units are more nearly valid, like species and other units established by biological taxonomy.

artifacts were deposited in patches as the occupations or sites of archaeological common sense, but others were deposited sparsely across the landscape outside of patches. Over sufficient intervals, these independent discard events formed diffuse scatters. Patches are “sites,” and scatters are the background continuous distribution of artifacts. Patches stand out because distributions seem denser there and Spatial Units of Observation probably are. Thus, patches seem to be places that Like assemblage, “site” is a primitive concept more differ qualitatively from the background scatter of often taken for granted than demonstrated. Its on- artifacts, the archaeological equivalent of distinct tological status increasingly exercises archaeolo- ethnographic sites. gists in North America (e.g., Dunnell 1992; Ebert In the LOM, however, Stern found no such dif1992) and elsewhere (e.g., Thompson 2004), as we ferences between the categories, concluding that see in this volume, where complex artifact distribu- “patches ...are simply denser scatters” (1993:214). tions are documented in the hunter-gatherer record If Stern is right, the implications are far-reaching of Australia and North America. But East African for our understanding of hominid behavior and hominid studies furnish among the best empirical land use and for the ontological status of patches; data for continuous artifact distribution on scales like “sites,” they may be constructions. If so, patches far larger than “site” can accommodate (Blumen- are not qualitatively distinct places, vignettes that schine and Masao 1991; Isaac 1981; Potts et al. 1999; played out over hours or days, but places where Rogers and Harris 1992; cf. Rogers 1996). There, scatters accumulated over longer intervals than artifacts are distributed widely and continuously they did elsewhere. Juell and Edwards (1994) disacross the landscape; nevertheless, distributions agree, seeing patches as different from scatters in vary greatly in local density (e.g., Potts et al. 1999:​ composition, not just size. To them patches as sites 775; Rogers and Harris 1992:42), such that some are distinct places, revealed entities. The dispute areas seem all diffuse distribution, others are con- obviously is important, its resolution in significant centrations distinguished to varying degrees from part a matter of assemblage interpretation. background distribution, and “sites” are largely arbitrary in definition because conjoined arti- Analysis facts can cross site boundaries (Blumenschine and In Stern’s own data, simple statistical measures of Masao 1991:​452) and because artifact density suffi- association support Juell and Edward’s view that cient to define sites in some places qualifies only as scatters and patches differ in composition. They background distribution elsewhere (Blumenschine are different things. But the developing revision of and Masao 1991:456; Rogers and Harris 1992:43). Lower Paleolithic tool typology justifies a new assay of the evidence. Table 4.1 shows data as reported by Sites are constructions, not revealed entities. Stern started from Isaac’s (1981) “scatters and Juell and Edwards (1994:167, Table 1), themselves patches” trope to compare LOM assemblages. Isaac derived from Stern’s article (Isaac et al. [1997:Aphas legitimately treated artifacts as fundamental pendix 6A] report slightly different counts for sevparticles or units of observation. Sites must be con- eral classes, probably because they used a slightly structed, but artifacts might be regarded as funda- different typology). Defined tool classes largely mental units that require only observation (even duplicate standard East African Lower Paleolithic though the status of some objects as artifacts itself typology (whether manuports were deliberately is a construction that rests on inference [e.g., de la transported to locations or are natural occurrences Torre and Mora 2005]). Artifacts exist indepen- in Lower Paleolithic assemblages [de la Torre and dently of archaeologists and their concepts. Some Mora 2005] is irrelevant; all Table 4.1 classes are

51

Michael J. Shott Table 4.1.  Tool Counts in Original and Combined Tool Classes in Lower Okote Member Assemblages Assemblage Scatters Patch FxJj20M FxJj20E FxJj20S FxJj20AB FxJj37

C

D

P

Sc

Sf

Class Mc

Mf

O

Σ

1

2

2

4

7

32

50

100

20

2

217

145

70

10 18 4 4 8

34 16 7 9 7

11 9 4 2 7

9 3 1 5 1

18 11 1 5 8

4 3 6 0 14

0 0 0 0 0

0 0 0 0 0

86 60 23 25 45

68 49 22 20 37

18 11 1 5 8

Note: C = chopper, D = discoid, P = polyhedron, Sc = core scraper, Sf = flake scraper, Mc = miscellaneous cobble, Mf = miscellaneous flake tool, O = other, 1 = C + D + P + Sc + Mc, 2= Sf + Mf. Source: Juell and Edwards 1994:Table 1.

chipped stone). Several patches were distinguished and treated separately, but scatters were combined for analysis. Indeed, if scatters are a continuous background distribution, it would be difficult if not impossible to parse them. Thus, comparison is between all scatters combined and the several patches jointly and separately.

pers and discoids ( Juell and Edwards 1994:168). Yet if these and other core tools are arbitrary segments of a ­cobble-​reduction continuum, not distinct functional classes, then differences owe not to past behavior but, instead, to the effects of formation processes and core reduction. If patches accumulated in repeated occupations of places, then more cobbles should have been deposited, and those cobbles might have become more reduced as people repeatedly visited patches, reusing and retouching the tools. Indeed, the presence of cobbles might have drawn people to the places. With such artifact circulation, assemblage size would increase, and some tools present in assemblages would become progressively more reduced (“curated” sensu Shott 1996). More core tools would pass from forms recognizable as choppers through forms recognizable as discoids and so on toward polyhedrons and spheroids, at the same time that new flakes were struck from cobbles. If spheroids are reduced polyhedrons are reduced choppers and so forth, then it is legitimate to combine Juell and Edwards’s counts into larger categories of core tools (choppers, discoids, polyhedrons, core scrapers, and miscellaneous cobbles) and flake tools (flake scrapers and miscellaneous flakes). Table 4.1 shows counts of separately combined core and flake tools, omitting Juell and Edwards’s class “Other” both for its lack of definition and because n = 2. Combining classes has the twin virtues of justification on technological grounds and reducing the number of cells whose expected

Patterns of Association between Scatters and Patches The most common tool class in scatters is miscellaneous cobbles, whose name suggests a default identification. Its frequency differs significantly in two of the five pair-wise comparisons of patches to scatter ( Juell and Edwards 1994:Table 2). Patches thus seem different from scatters. Yet χ2, often used to gauge significance of differences in counts, exag­ gerates differences when counts are high overall and some cell counts are low (Drennan 1996:195). It is not a particularly robust measure of difference. With fairly high counts and 16 cells in χ2 ­tables, significant results like Juell and Edwards’s are no surprise. Yet the frequency of flake scrapers, the second most common class in scatters, does not differ in any pair-wise comparisons of scatters and patches. Among the three most common classes— cobbles, flake scrapers, and core scrapers—scatters and patches differ in only four of 15 pair-wise comparisons despite Table 4.1’s high cell counts. In this perspective, scatters and patches seem not so different. Differences are greater between scatters and patches in uncommon classes like chop52

Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation

Figure 4.2.  Log10 tool counts versus log10 assemblage size for scatters and patches.

Patterns in Tool Counts and Proportions between Scatters and Patches By tool class, Juell and Edwards also plot log10 class counts against log10 assemblage size. Including scatters with patches, they (1994:Figure 1) find significant positive correlations between assemblage size and tool count for three classes: core scrapers, flake scrapers, and (somewhat ambiguously) polyhedrons. Size dependence is evident; assemblage size, not composition, seems to distinguish scatters from some patches in some class counts. Assemblage differences in these classes seem continuous, not categorical, and so consistent with the view that scatters and patches differ by degree, not kind. But Juell and Edwards correctly note that scatters and patches do not obey the same size-dependent trend in other classes. As Juell and Edwards do for six classes, Figure 4.2 plots log10 core and flake tool counts against log10 assemblage size for scatters and all patches. It also shows the least-squares regression (LSR) line and 95 percent confidence limits. In these data, size dependence is clear (for core tools: r = .99, p = .00;

values are less than five. In combined classes, differences persist between scatters and five of six patches, although χ2 and p values are much lower. However, there are no differences between scatters and patch FxJj20AB (χ2 = 1.67, p = .20). Kerrich’s (1957:​138) test of significance in class proportions yields similar results; combined scatters differ in combined core–tool proportions from all patches except FxJj20AB. Most patches still seem different from scatters, as Juell and Edwards conclude, but less so than they believed.

Patterns of Association between Patches Whether or not scatters and patches are different things, patches need not be similar among themselves. In Table 4.1 data, five of 10 pair-wise comparisons between patches differ at p = .05, and seven of 10 differ at p = .10. In pooled classes, no pairs differ at p = .05 and only two differ at p = .10, with two others nearly reaching that significance level. Whatever their relationship to scatters, it is at least questionable to treat different patches as examples of the same assemblage type. 53

Michael J. Shott

Figure 4.3.  Log10 tool proportions versus log10 assemblage size for scatters and patches.

for flake tools: r = .94, p = .005); all assemblages obey one pattern, and all cases fall within confidence limits. For both classes, the upper-right point marks scatters. In one sense, the pattern is banal. After all, core tools are most abundant in all assemblages, so their counts must increase as assemblage size increases; it is their rising frequency that mostly creates larger assemblages. For both classes, however, the plot reveals continuous variation as a function of time, not just among patches but also between scatters and patches. In this view, scatters and patches do not differ in kind, only in degree as a matter of time or accumulation span. Counts measure assemblage size, and proportions are one measure of assemblage composition. Together, core and flake tools account for all artifacts, so their proportions must sum to one in each assemblage and thus reduce to binomial probabilities p and q where p + q = 1. Figure 4.3 shows the joint size dependence of core and flake tool proportions and their separate LSR lines and 95 percent confidence limits. As in counts, patterning in proportions is clear (for core tools: r = –.84, p = .04;

for flake tools: r = .84, p = .04; the binomial constraint explains the identical r and p values but opposite signs). Again, all cases fall within confidence limits. Core tools make up most of each assemblage and so contribute greatly to assemblage size; little wonder that core tool proportion covaries with size. This constraint may seem to create, not reveal, patterns in the data. Yet pooling is legitimate on archaeological grounds and incidental on statistical ones; it was not done to create a desired result. The most important point is that as assemblage size increases, flake tools grow not just absolutely but also proportionally more abundant. Perhaps assemblage size is a crude measure of cumulative occupation intensity or span (Surovell 2003). As more people visited places and used stone tools, perhaps they increasingly struck flakes from cobbles previously accumulated there, whether or not they also imported new cobbles. In this way, the rate of use, discard, and accumulation of flake tools would be higher than that of core tools. Juell and Edwards (1994:170) themselves note apparent size dependence in counts of some core 54

Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation tool classes but not in others. Such partial size dependence becomes nearly complete when combined core tools are considered. In this perspective, the independence of assemblage size and some class counts obscures clear size dependence in the larger, arguably more legitimate, combined class. It is not merely a question of size dependence in general within assemblages but also of which particular tool classes vary in count or proportion with size. As above, Juell and Edwards show that choppers and discoids, which occupy intermediate segments of the core tool reduction continuum (Figure 4.1), were especially abundant in patches. As people revisited the places that gradually became patches, perhaps they not only used and reused core tools already accumulated in them but also imported new ones. Gradually, proportions of less reduced core tools like choppers rose against proportions of more reduced ones like discoids. Elsewhere, Oldowan assemblages show similar evidence of circulation (e.g., too few flakes to account for scars on cores) as to be “nodes in a dynamic system of transport of materials by hominids over the landscape” (Potts 1991:​163). Perhaps other cobbles, previously reduced slightly and discarded, were also reused as tools or as cores to produce flake tools. One core, even in a “Mode 1” technology, can yield many flakes. Proportion of flakes then would rise at a higher rate than proportion of cobbles. Formation processes and reduction seem to account for much of the LOM evidence. Thus, LOM assemblage size and composition are related, not independent. Patches differ from scatters by degree, not kind. The continuous variation in assemblage size and variation is parsed, somewhat arbitrarily, by the categories “scatter” and “patch.”

differ culturally from Developed Oldowan ones because the respective assemblages differ in composition, corroborated by stratigraphic ordering (Leakey 1971:​Table 1). Developed Oldowan arose from Oldowan, as its name implies, and thus was descended from but different from its predecessor in cultural and industrial character. Oldowan and Developed Oldowan are a time series of cultures, different things distinguished in substantial part by the different compositions of their assemblages. Using Leakey’s own tool classes, however, analysis of assemblage richness, evenness, and heterogeneity shows that Developed Oldowan assemblages are Oldowan ones possessing more tools and therefore more classes. The classes that distinguish Oldowan and Developed Oldowan are rare, presumably because they were rarely made and used, or they survived so long in use that they were rarely discarded. Rare classes are likeliest to occur in larger assemblages, because the places where these accumulate were occupied longer or more often in the aggregate than were other places. The distinction between Oldowan and Developed Oldowan may not be cultural or chronological—perhaps not industrial at all in the broad sense but one of time averaging mediated by assemblage size. Elsewhere in East Africa, the Paleolithic archaeological record reveals the continuous distribution that Stern found in the LOM. For instance, Olduvai Bed II stone tools are distributed as scatters and patches. Cross-mends between “sites” and nonsites suggest that the former lack occupational integrity and instead were “defined in part arbitrarily” (Blumenschine and Masao 1991:452) in a continuous regional artifact distribution. One stratigraphic unit at Olorgesailie well accommodated the scatterand-patch model. In extensive regional excavation as well, patches yielded artifact densities roughly five times higher than did scatters (Potts et al. 1999:​ 765). Even if patches there register single occupations of a place, perhaps one-fifth of the artifacts found in them were not discarded in those occupations. Yet lower stratigraphic units at Olorgesailie displayed much more clustered artifact distributions that suggest a discrete archaeological record consisting of sites and empty or sparsely populated areas between them (Potts et al. 1999:775). Rogers

Other Cases Other cases also illustrate the covariation and continuous patterning of assemblage size and composition that cast doubt on the validity of discrete assemblage types. Leakey (1971:258), for instance, attributes variation in Olduvai assemblage composition to functional differences at places called living floors and kill sites but in other cases to cultural succession over time. Oldowan assemblages 55

Michael J. Shott (1996:​106) reports similar results in his own LOM excavations. At Maastricht-Belvédère in the Netherlands, Roebroeks (1988; Roebroeks et al. 1992) excavated several Middle Paleolithic patches or sites but also a large area in which no site was identified. Most tools and debris found there could not be linked by raw material, refitting, or context and so could not be shown to be associated in time or use. In effect, this nonsite scatter was a large surface that many people traversed at many times for many reasons, carrying and dropping some artifacts at intervals. The scatter was a “veil of stones” laid over the landscape, background noise upon it. In East African terms, it was a time-averaged scatter. To Roebroeks and colleagues “sites” were patches imposed over scatters such that “at least some of the artefacts excavated in the patches have nothing to do whatsoever with . . .the majority of the finds from these patches” (1992:11). Similarly, English Lower Paleolithic assemblages are not discrete moments in time but “gravel deposits with a minimum chronological duration of c. 70–100,000 years. The behavioural information contained within depositional units of that duration would inevitably be subject to time averaging and blurring” (Hosfield 2001:90). The size, composition, and location of these assemblages do not reveal short-term activity but, properly interpreted, patterns of behavior at time-space scales that ethnography cannot accommodate. Thus, Hosfield (2001:​92–93, Figure 11) identifies spatial patterning in hand axe accumulations well above the “site” scale as well as chronological patterning in human population size. Traditionally, variation in the composition of Old World Middle Paleolithic assemblages was interpreted in iconic or social terms, the stratigraphic successions of Périgord rockshelters, for instance, registering the comings and goings of distinct ­identity-​conscious social groups. But much presumed assemblage difference is illusory, and much of the rest is owed to prosaic but important factors such as stone supply (e.g., Dibble and Rolland 1992). Equally as important, the composition of many such assemblages is dependent on their size (Shott 2003). Whatever social or functional dif-

ferences played out in these assemblages, they are background noise to the play of reduction effects and assemblage formation over long spans. Like LOM ones, Middle Paleolithic assemblages are classic time-averaged deposits. Patches or sites often are complex palimpsests, not discrete moments from the past with spatial and compositional integrity as assemblages. These examples span a considerable range of time and space. Others cited above document time averaging in recent archaeological (Holdaway et al. 2004) and even ethnographic assemblages (Shott 1989, 1997), the latter of which accumulated in very short spans. Examples do not prove that LOM patches are dense scatters. Yet that conclusion is consistent with LOM evidence and conforms to the larger pattern found in other archaeological evidence. This is time averaging, and it is pervasive in but not peculiar to Paleolithic assemblages and time scales. Time averaging is nearly a constant property of assemblages, although one that itself varies both in causes and the patterns they create. Causes include the number of tool classes and their integrity. Giraffes and tigers are integral classes because one cannot become the other. No less is true of some tool classes, but if choppers are polyhedrons are spheroids depending on the amount and pattern of reduction that specimens experienced, then those classes lack integrity. Their liminal quality affects composition as accumulation span and assemblage size increase because, with time, specimens pass from one class to another. Causes include not just use-life but also its “relativities” (Orton 1993:179), the range of variation in use-life among tool classes. As little research as has been done on use-life and its determinants, there is practically none on such relativities, and this remains an urgent necessity. Finally, causes include scale of and variation in occupation spans, rates of reoccupation of places, and the complex patterns of behavior and land use characteristic of ancient cultures. Patterns created in time averaging vary in the strength and slope of the relationship between assemblage size and composition, as well as the location of asymptotes above which size varies independently of composition. We should not be surprised that ethnographic assemblages accumulating in brief spans (e.g., Shott 56

Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation tic. In this version H = –Σ ( pi ln[ pi ]), where pi is the proportion of tools in class i for all n classes. Archaeologists are fond of condemning H as derived, abstract, and size dependent. H certainly is derived and abstract, because the same value can be calculated for sets of archaeological tool classes and gumballs in a machine. It is size dependent— its value depending on sample size—in some but by no means all cases (Shott 1997, 2003), so its size dependence is a variable not a constant property. But S and E are equally derived, abstract, and certainly size dependent (Hayek and Buzas 1997; Shott 1997) without receiving the same condemnation. S, for instance, is a crude measure that says no more about assemblage composition than does the number of taxpayers about the distribution of wealth. S cannot distinguish between two assemblages of 100 tools each where one has 10 specimens of each class and the other has 91 of one class and one each of the remaining nine. In both, S = 10. If size dependence is not just an artifact of assemblage measures, then it must have causes. Among them are assemblage formation processes, particularly the relationship between tool use rates and use-lives, on the one hand, and occupation span or assemblage size, on the other. As noted above, deterministic simulations show clearly that composition changes as occupation span and assemblage size increase, except in the unlikely event that use-lives are identical between tool classes. Hayek and Buzas (1997:379) suggest new ways to interpret assemblage variation. They (1997:380) parse H into separate richness and evenness components as H = lnS + lnE and call the joint study of variation in the three measures “SHE” analysis. lnE is always negative and often varies inversely with lnS. Essentially, then, H can be deconstructed to reveal the separate effects of S and E. Perhaps both or perhaps only one covaries with assemblage size, explaining H’s size dependence via that component only. Different scenarios for the joint behavior of S, H, and E implicate different theoretical models governing assemblage size and composition. S and H can increase together while E remains constant, they can increase at different rates as E decreases, or S and E can increase as H remains constant (Hayek and Buzas 1997:​318–383). The models that govern

1989) and archaeological ones in long spans (e.g., the LOM and other examples cited here) all display the interdependence of size and composition that time averaging causes. But they vary a great deal among themselves in the strength and degree of the patterns created. That variation demands ­explanation in archaeological, ethnographic, or other terms.

Meanings of Assemblage Variation Covariation of assemblage size and composition may seem esoteric, even nihilistic in its complexity. After all, if assemblage composition does not reveal either activity patterns across the landscape or self-conscious social identity, what does it reveal? The lack of an obvious equivalent to the variation we either imagine or see plainly in ethnographic evidence, unfolding over time scales familiar to our own thoughts and experience, disquiets many archaeologists. It suggests the absence of meaning, random noise that drowns out patterning in assemblage variation. But recognition of assemblage formation and time averaging as one of its expressions is not a counsel of despair that the archaeological record has no coherent story to tell. On the contrary, it improves understanding of how assemblages formed and prevents spurious interpretations that are wrong no matter how much they seem to be right. Archaeological assemblage patterning has meaning, but not in ethnographic terms. One way to reveal it is by an example involving common assemblage measures such as richness (number of classes present without regard to each one’s proportions), evenness (similarity or not in class proportions), and heterogeneity, a joint measure of richness and evenness. Hayek and Buzas (1997) ­label richness “S,” evenness “E,” and heterogeneity “H.” Richness is measured simply. Evenness is measured variously, for instance, as the standard deviation of class proportions or the ratio of observed to maximum heterogeneity; Hayek and Buzas (1997:​379) measure it as E = e H/S, where e is the base of Naperian logarithms. Heterogeneity is also measured variously but most often as the finite population estimate of the Shannon-Weaver information statis57

Michael J. Shott

Figure 4.4.  “SHE” plot of Lower Okote Member data resembling Hayek and Buzas’s (1997:Figure 14.1) log series model. E = evenness, H = heterogeneity, S = richness.

assemblage variation are revealed by SHE analysis and give that variation meaning that ethnographic analogy cannot provide. Because archaeologists interpret LOM assemblage variation in terms of Table 4.1’s traditional tool classes, no matter their problematic nature, I calculate S, H, and E for LOM assemblages using those classes. Figure 4.4 shows results in the format of Hayek and Buzas’s (1997:Figure 14.1) SHE analysis. As they have done, both lnE and the ratio lnE/lnS are rescaled by one order of magnitude to bring out size-dependent patterning. Such joint plots reveal the scenario that governs each set of assemblages. Among LOM assemblages, lnE varies inversely with assemblage size, whereas S and H change little. By Hayek and Buzas’s reasoning, this pattern matches the log series model where one “strongly hierarchical” (May 1975:83) factor governs assemblage composition. In this case, that factor can be interpreted as assemblage size, itself the product of varying occupation spans and rates of reoccupation of different places in the landscape. Some places are more persistent (Schlanger 1992) than others, but much the same behavior is prac-

ticed everywhere. LOM assemblage variation is not the result of different activities or behavior. In Wandsnider’s (this volume, Table 5.2) terms, it owes instead to intermediate or long processes, be they cyclical or linear. Archaeologists are accustomed to interpreting assemblage composition and perhaps size in ethnographic terms, as corresponding to one or another settlement type. That is an intuitive, satisfying form of interpretation. My interpretation, like others in this volume, is less intuitive, more abstract, and probably less satisfying to most archaeologists. Instead of a set of “sites,” some of them hunting locales and others homes bases, analysis suggests a log series pattern that encompasses all assemblages rather than parsing them into categories like “home base.” Many assemblages can no longer be interpreted as social icons or ethnographic camps. But the artifact distributions that we define as assemblages nevertheless reveal long-term patterns in land use and perhaps activity and social identity. They might reveal demographic oscillations over time spans that ethnography and history could never perceive but which might reveal properties of ­human 58

Lower Paleolithic Industries, Time, and the Meaning of Assemblage Variation nificance of assemblage differences between scatters and patches is not her chief point and that Juell and Edwards’s behavioral interpretation of those differences cannot be reconciled to the nature of the archaeological record and its accumulation span. Even if LOM patches accumulated over the very short intervals required for this interpretation, it seems improbable that the assemblages would differ substantially. Instead, detailed models linking activity to the kind, number, and distribution of tools discarded show equifinality, different activities yielding similar assemblages (Blumenschine and Peters 1998:595, Table 3). Viewed in reduction perspective, core and flake tool counts and proportions are determined as much by assemblage size as inferred activity variation. Many patches indeed are dense scatters. Here is good evidence for the relationship between assemblage size and composition and between typology and analytical results. Patches and scatters are not distinct things but, rather, conceptual poles of joint continua of artifact density and distribution. Because many assemblages conform to the same pattern, it seems as dubious to consider scatters and patches categorically distinct as it would be to classify children and adults as separate species. Archaeologists order evidence commensurate with short-term ethnographic time scales. But assemblages have accumulated over spans much longer than ethnography describes. We must study and compare assemblages in their own terms, not force them into ethnographic pigeonholes using crude measures (e.g., number of artifacts, number of hunting vs. domestic artifacts). It may seem unsatisfactory, even banal, to merely conclude that patches and scatters vary on a continuum or that they are density modes in a landscape of scatters. But this conclusion reveals land use patterns and identifies occupation modes that in turn beg an interesting question: Why concentrate at these points in the landscape? Only general theory, not formation theory, can answer this question. Attending to formation theory is no distraction or substitute for explanation but, in fact, a necessary precondition to explanation in cultural terms. “Time averaging,” which captures the relationship among assemblage size, composition, and

p­ opulations manifested only in archaeological time (Holdaway et al. 2004; Hosfield 2001) and understandable only in terms of theories of the truly long term. As new tool classes are invented, they change the relationship among classes, functions, and uselives that may register in changing rates or patterns of size dependence in assemblage composition. If, at least for some purposes, we abandon “site,” we can study how spatial scale affects the relationship between assemblage size and composition and patterns of association between artifact classes (Ebert 1992). All of these possibilities inform about the past, just differently than the ethnographic tableaux archaeologists typically devise. In the LOM, patches are not qualitatively different places from scatters. They simply are places that people visited for longer spans or more often than others or some combination of the two. Why? Only when we understand the causes of assemblage variation can we even contemplate the question, let alone answer it. Whatever the answer, it will not involve short-term ethnographic vignettes of activity or identity, because many assemblages do not accumulate in the short term. This realization does not mean that archaeology should abandon anthropology—on the contrary, judicious use of ethnographic data remains valuable—merely that it should not be bound to ethnographic time scales and habits of thought. “Archaeology is anthropology or it is nothing” is a manifesto familiar to generations of archaeologists. True enough. Archaeology is the anthropology of past cultures but not just that, nor is the ethnographic record reproduced in the material one. We need a new manifesto to complement, not replace, the old one: “Archaeology is time perspectivism or it is nothing” because the past is expressed only through a material record whose many time scales of formation must be parsed, not reduced to a simplistic reproduction of ethnography.

Conclusion Both sides in the Lower Okote Member debate are right, but in different ways. Juell and Edwards (1994) are right that scatters and patches differ in some ways and that Stern (1993:214) says as much. Stern (1994b:170–171) is right to argue that the sig59

Michael J. Shott a­ ccumulation span, lies at the center of this view. As span increases, so does size, except when tools are recycled. Unless all classes have identical use-lives, composition changes systematically with increases in the other two quantities. Number of classes and range of variation in their respective use-lives, both culturally determined, fix the rate and pattern of composition’s covariation with size and, by extension, span. As above, size and composition are intrinsic assemblage properties, but they are incomprehensible without span or time. Assemblages

cannot be understood without knowing the time span over which they accumulated. Formation theory and time perspectivism reveal time as an intrinsic dimension of assemblages, not a distracting complication to their analysis.

Acknowledgments My thanks go to Simon Holdaway, Nicola Stern, and LuAnn Wandsnider for comments about the manuscript. None of them necessarily agrees with my conclusions. I am responsible for any errors.

60

5

Time-Averaged Deposits and Multitemporal Processes in the Wyoming Basin, Intermontane North America: A Preliminary Consideration of Land Tenure in Terms of Occupation Frequency and Integration LuAnn Wandsnider University of Nebraska–Lincoln

Archaeological time perspectivism encompasses the notion that archaeological deposits are formed through the operation of processes occurring at a variety of tempos over the short, medium, and long term (Bailey 1981, 1983, 1987, 2007, this volume). The processes involved may be behavioral, social, formational, organizational, or evolutionary, to name a few. Through their operation, material consequences may be immediate, lagged, or follow after some threshold is breached. Moreover, interaction may occur among and between different processes, depending on whether they operate at approximately the same scale (Bailey 1983; Fletcher 1995). A corollary of the first statement is that different archaeological deposits, by virtue of their different temporal structures (Kirch 2005:414; Murray 2004), have potentially captured processes operating at different tempos. This corollary recognizes that modern (and ancient) surfaces are temporal mosaics (Bettis and Mandel 2002), with some local surfaces rather ancient and others more recent. It also recognizes that different landforms in close or far proximity to sediment sources will be differentially active, with consequences for the deposition and sealing of archaeological remains. For example,

a surface that has been stable over a millennium has the potential to receive the fallout from a variety of processes operating at different tempos. On the other hand, on a very dynamic land surface, fasttempo cultural processes, such as frequently occurring occupation events, may be well represented, whereas slower processes, for example, rarely (oncein-a-lifetime) occurring ritual events, may be more incompletely sampled. Here, I explore the temporal structure and interpretive potential of deposits from throughout the Wyoming Basin (intermontane North America) that have been well documented through compliance archaeology. I assume that all of these deposits likely represent cumulative and spatial palimpsests (sensu Bailey 2007:204–207) manifesting different degrees of integration (Holdaway and Wandsnider 2006; see below). My goal is exploratory and follows from Murray’s (1997; see also Olivier 2001; Bailey, this volume) observations that archaeology as a discipline still seeks the means to interpret social process over the medium to long term from the convoluted human/natural phenomenon that is the archaeological record. This exploration represents an exercise in pattern recognition as alluded to by Clarke (1973) and Binford (1977b) when they note 61

LuAnn Wandsnider that theory building must follow the development of an understanding of archaeological subject matter. But, of course, we also recognize that the units we choose to describe archaeological phenomenon prefigure the kinds of interpretations that can be entertained (Ramenofsky and Steffen 1998; Wylie 1989). To escape this methodological conundrum requires that simultaneous efforts be made along the paths of theory building, pattern recognition, and unit formation and that the hermeneutic spiral be completed, with an eternal dialogue among and between theory building, unit construction, data collection, and interpretation evaluation (Hodder 1999). A consideration of these deposits, with coarse—and very coarse—temporal grain (see below), opens the door to inferences about short-, medium-, and ­longer-term processes of wider anthropological interest, for example, the development of different land tenure systems. Translating this larger goal into concrete objectives, I first elaborate on the properties of the temporal structure of archaeological deposits and then discuss the potential processes that may contribute toward the character of Wyoming Basin archaeological assemblages. Next, I introduce the Wyoming Basin study area and then move to an analysis of the temporal structure of Wyoming Basin components especially focusing on occupation frequency and integration. Discussion follows on aspects of land tenure.

From the paleontological literature come other concepts. Time averaging refers to “fossilif­ erous units that represent extended periods of time and mixing of organisms from different habitats” (Behrensmeyer 1982:213); that is, the material consequences of behaviors from many agents are integrated over the time span during which the sedimentary envelope accumulated (Stern 1993, 1994a, 1995). As ecologists O’Neill and King (1998) and paleontologists Behrensmeyer and Schindel (1983) note, the degree and nature of time averaging set the grain or resolution of a deposit and, hence, the kinds of generating processes that can be interpreted. Paleontologists have identified two other parameters describing overall deposit temporal structure. Behrensmeyer (1982) and Behrensmeyer and Schindel (1983) refer to the scope or total temporal time span represented in a deposit, as well as depositional gaps, owed to erosion or nondeposition, within a deposit. All of these parameters together— grain (influenced by time averaging), scope or span, and gaps—determine the kinds of processes and their tempos that are accessible through analysis of archaeological deposits, as emphasized by paleontologists and archaeologists (Bailey 2007, this volume; Murray 1997, 1999a, 2004; Stern 1993, 1994a). Finally, Holdaway and I (2006) have recently called attention to the degree to which materials are integrated between occupation events, that is, remains from succeeding occupations are mapped onto or acknowledge remains from preceding occupations (see also Wandsnider 1992). For ­example, Bamforth, Becker, and Hudson (2005) find evidence that succeeding occupants of the Paleoindian Allen site in western Nebraska situated hearths and middens with respect to previously constructed features. Paleontologists and archaeologists commonly rely on various standard chronometric tools to measure scope, gap, and grain. But within the resolution of these tools, other taphochronometric tools, which rely on the accumulation of traces by artifacts, features, and spaces over time (sensu Sullivan 1978, this volume), situationally permit a finergrain temporal structure to be approached. For

The Temporal Structure of Archaeological Deposits Over the last 20 years, archaeologists have come to recognize and understand various properties of the temporal structure of deposits (Murray 2004). Binford (1978a; see also Ferring 1986) introduced the notion of temporal grain to describe deposits and explicitly recognizes this character as a convolution of the tempo of sedimentation and behavioral or cultural processes; coarse-grained deposits represent the accumulation of many, likely irresolvable cultural depositional events, whereas fine-grained deposits may preserve the remains of a short sequence of behavioral events (see also O’Neill and King 1998:7 and Schindel 1982 on resolution and microstratigraphic acuity in paleontology). 62

Time-Averaged Deposits and Multitemporal Processes Table 5.1.  Model of Place History and Taphochronometric Indicators Occupation Events Grain or Span One

Few

Many

Short

Medium Long Short Medium Long Short

Medium Long

Indicators Local tool source: high primary debitage/tertiary debitage ratio Little site structure Thermal features: charcoal and oxidation well preserved Thermal features: charcoal stains, little oxidation (Sharrock 1966) Amorphous thermal features Mean hearth area increases (Yellen 1977) Low fire-cracked rock (FCR)/thermal feature ratio Unknown Simple site structure Pit structures and other facilities (if anticipated reuse [Chatters 1987:343–346; Smith and McNees 1999]) Many hearth types (Yellen 1977) High standard deviation of hearth area High artifact/feature ratio High FCR/thermal feature ratio Multiple modes of bone weathering (Behrensmeyer 1978) Complex site structure (Binford 1978a; O’Connell 1987; Wandsnider 1996) High thermal feature density High proportion of thermal features recycled into middens Overdeveloped anthropogenic A horizon (Eckerle and Hobey 1999) Low artifact/feature ratio

Note: With multiple events, artifact, core, and incomplete tool densities increase (Binford 1977a; Camilli 1988) and FCR fragment size decreases (Chatters 1987:345).

example, Holliday, Johnson, and Stafford (1999) consider the weathering profiles of faunal materials at Plainview and Firstview (American Great Plains) to argue for the relative contemporaneity (plus or minus 15 years, the temporal resolution of bone weathering) of these materials. Varien and Mills (1997) and Varien and Ortman (2005) consider the accumulation of sherds to estimate relative occupation spans of Puebloan sites in the American Southwest. Bamforth and colleagues (2005) consider accumulations of middens and hearths on an aggrading surface that was repeatedly visited at the Allen site. When dealing with aggradational and deflational deposits (e.g., Holdaway et al., this volume; Kelly 1988), archaeologists commonly employ such taphochronometric observations to strengthen interpretations of temporal grain finer than that accessible using standard chronometric tools, although often in an ad hoc manner. (For fluvial deposits, likely accumulating from multiple temporal planes, the use of taphochronomet-

ric tools to approach fine-grained interpretations is likely not valid, as discussed by Stern [1993, 1994a, 1995] in her analysis of Koobi Fora deposits.) In what follows, I rely on a number of such tools (Table 5.1), each of them variously based in theory (e.g., radiocarbon dating [Newtonian physics]), empirical observation (e.g., repeatedly cleaned out hearths may grow), and intuition (e.g., the accumulation of features on a land surface). Because of space constraints, I will employ these tools without further substantiation, acknowledging that, in fact, such is critical. Indeed, many of these tools remain plausible, middle-range assertions. Only Yellen’s (1977) observations on hearth “creep” and Behrens­ meyer’s (1978) observations on bone weathering have had some empirical scrutiny, but even that has been limited. Of course, available technology (e.g., the bow and arrow came into use here about 1800 bp; pemmican production, between 5000 and 3000 bp [Reeves 1990]), the nature of activities (e.g., retooling in anticipation of a major ­community hunt 63

LuAnn Wandsnider as argued by Reher and Frison [1980] for the Vore site), season of occupation (with greater numbers of thermal features perhaps constructed in winter months), and other similar factors come into play when interpreting indicator values (as per Binford’s [1980] comments on Yellen 1977). That is, these taphochronometric indicators are ambiguous, yielding nonspecific interpretations (but denying some interpretations [Wandsnider 2004]). Also, they very likely perform contextually, that is, differentially well in some contexts but poorly in others (Wandsnider 2004). For this reason, context-rich comparative (as opposed to simple diagnostic) analysis, undertaken here, is essential. From notions of temporal structure come important implications for interpretation. Paleontologists (Behrensmeyer 1982; Behrensmeyer and Schindel 1983; Stern 1993, 1994a, 1995) contend that with fine-resolution geological deposits, one can begin to interpret fast-tempo processes, such as evolutionary changes in rapidly reproducing populations; fast-tempo processes, however, cannot be directly interpreted using coarse-grain geological deposits (also known in the spectral or sampling literature as the Nyquist effect). Thus, Stern (1993, 1994a) contends that discussions of behavioral and ecological processes at Koobi Fora are invalid, given the coarse grain (10,000-year span) of the deposits there. But by using various taphochronometric indicators on time-averaged aggradational and deflational deposits and the process-pattern strategy discussed below, temporal processes occurring at tempos faster than those resolvable geologically or chronometrically may be approached.

stance of technology change, and issues of range compression or expansion or territoriality, in the case of persistent place use, remains to be argued. This discussion especially focuses on the latter, that is, regional place-use histories and their larger implications for understanding a multitemporal process we might gloss as land tenure. Table 5.2 summarizes a variety of processes commonly reported on by anthropologists and others. As Butzer (1982), Bailey (1983), and archaeological Annaliste researchers (Bintliff 1991; Knapp 1992; Smith 1992) have emphasized, different processes unfold at different rates and become manifested over different lengths of time. To diagnose temporal processes archaeologically, two strategies are available. The first depends on recognizing a sequence of conditions or states through time, with “time” monitored using standard chronometric tools and “condition” inferred from time-averaged assemblages. But the resolution of standard chronometric tools is only so fine, and here the second strategy becomes important. This strategy is seen in the study of spatial point processes (Gettis and Boots 1978; Graham 1980), where the task is to infer the generating process (operating at a particular frequency) from the distinctive pattern of points so produced by those processes. Here, I extend this same strategy to identify temporal/spatial processes responsible for distinctive patterning. The taphochronometric indicators discussed here work to inform on the operation of the temporal/spatial point processes that occur more rapidly than can be captured by standard chronometric indicators. The study presented here relies on both of these strategies, with standard chronometric tools used to order assemblages in time and taphochronometric tools utilized to approach “condition,” as inferred from individual place-use histories.

Processes and Place-Use Histories Cultural anthropologists understand and describe hunter-gatherers in terms of how tasks are gendered and organized, how postmarital residence and inheritance are practiced, how identity and institutions are reproduced, the conditions under which sharing occurs, and so forth. Archaeologists, on the other hand, can see that hunter-gatherer technology has changed through time and can recognize that certain places have been extensively utilized or not. How each of these relates to larger issues such as the evolution of the cultural repertoire, in the in-

Wyoming Basin The Wyoming Basin, located in intercontinental northwestern North America (Figure 5.1), consists of a high plateau with many interconnecting smaller basins framed by the Wind River Mountains to the northeast, the Uinta Mountains to the south, and the Wyoming Overthrust Belt to the 64

Time-Averaged Deposits and Multitemporal Processes Table 5.2.  Processes by Length of Term over Which They Are Manifested Term

Process

Example/Reference

Event Response   Very short (subannual)

Mortuary preparations

Olivier 1999

Cyclical   Very short (subannual)   Short (annual–decadal)

  Intermediate (decadal–   subcentury)

  Long (century) Linear   Short (annual–decadal)   Intermediate (decadal–   subcentury)   Long (centuries)   Very Long (centuries–   millennia)

Ritual cycle Delayed reciprocity Logistical planning Monument use, maintenance El Niño climatic oscillation Territory expansion, contraction Social reorganization, alliance ­reconfiguration

Bailey 1983 Bailey 1983 Olivier 1999

e.g., Nunamiut (Amsden 1977) !Kung (Wilmsen 1989)

Market-based economic cycle Ecological community reorganization Colonization, abandonment Paleoclimatic reorganization Frontier evolution Demographic infilling

Kealhofer 1999 Swedlund 1978

Technological change

e.g., Southern California (Broughton 2002) e.g., Wyoming Basin (Eckerle 1997) e.g., Northern Rocky Mountains (Thoms 2007:504)

Soil formation Surface homogenization by biomechanics

west (Fenneman and Johnson 1946). The Rock Springs Uplift is an anticline structure found in the southern-central portion of the basin. Landforms include cobble-mantled terraces, deflation basins, alkali flats, playas, dunes, and badland scarps (Love 1977; Thornbury 1965). Today, the area is best characterized as a cold semiarid desert with sagebrush steppe vegetation dominating in the west, mixed grasses occurring in the east, and saltbrush and greasewood occurring in drier portions of the interior basins (Küchler 1966). On the basin margins and the Rock Springs Uplift, mountain mahogany and other tree species reflecting generally moister conditions are found. Paleoclimatic reconstructions indicate that climatic conditions have not always been as they are today, nor is climate uniform throughout the basin. As summarized by Eckerle (1997; see also Eck-

erle et al. 1999) for the basin proper, the Early and Middle Holocene was characterized by dry and drier conditions, evidenced by dune activation and deflation and accumulations of illuvial calcium carbonate. The development of oxidized B horizons or calcium carbonate horizons indicates localized surface stability. Cooler and moister conditions prevailed between 3300 and 1800 bp, when the Vonalee-Hiland paleosol developed in eastern portions of the basin. Drier conditions, indicated by reactivated dune sands, followed beginning as early as 2500 bp. From work in the Wind River Mountains, Fall, Davis, and Zielinski (1995) suggest that climatic shifts to cooler and moister conditions were first experienced at upper altitudes and later in the lower-altitude basins. In addition, west–east and south–north gradients in temperature and moisture are also seen. 65

LuAnn Wandsnider

Figure 5.1.  Wyoming Basin study sites.

The Wyoming Basin has been an arena for major research on past hunter-gatherer populations over the last 80 years, including pioneering work at the Finely site near Eden (Moss et al. 1951), which established a Paleoindian presence in the area, and on mass kill sites of bison (e.g., the Late Prehistoric Wardell site near Big Piney [Frison 1973]) and antelope (e.g., Austin Wash [Reiss and Walker 1982]). (See Eckerle et al. 1999 for a recent overview.) Most recently, compliance archaeological work prompted by major energy development has been responsible for a great deal of activity here. Interpretations offered through this work reflect the great strides made in hunter-gatherer archaeology during the 1960s–1980s by archaeologists (e.g., Robert Bettinger, David Hurst Thomas) working in the nearby Great Basin, ethnoarchaeological work from around the world, and ­Binford’s theorizing on hunter-gatherer organization. Creas-

man and Thompson (1997) have summarized the salient features of much of this work, offering a model of settlement/subsistence for the area. They report that paleobotanical and faunal indicators in general suggest a winter use of the higher-altitude­ sites and summer use of interior basin sites during both the Archaic and Late Prehistoric time periods. Other work notes the increase in interior ­basin seed processing during Late Prehistoric times (Smith 1988), along with increased evidence for mass kill events starting in the Late Archaic. Here, as well, Ebert (1992) develops the notion of distributional archaeology, which explores multiscalar patterning in surface assemblages and interprets that patterning in terms of supragenerational systemic poses of hunter-gatherer groups. In addition, Harrell, Hoefer, and McKern (1997), Larson (1997a, 1997b), Waitkus and Eckles (1997), and Smith and McNees (1999) have addressed ­issues 66

Time-Averaged Deposits and Multitemporal Processes Table 5.3.  Wyoming Basin Cultural Chronology Uncalibrated Radiocarbon Years (bp)

Period

Phase

Cultural Markers

650–150 1,800–650

Late Prehistoric

Firehole Uinta

2,800–1,800 4,300–2,800

Late Archaic

Deadman Wash Pine Springs

6,500–4,300 8,500–6,500 12,000–8,500

Early Archaic

Opal Great Divide

Poorly known Major increase in radiocarbon dates, mass kill sites, bow and arrow technology, seed processing Trough in frequency of radiocarbon dates Peak in frequency of radiocarbon dates; appearance of stemmed/indented and corner-notched projectile points Pit structures, below-ground storage Poorly represented Poorly represented

Paleoindian

Source: Eckerle et al. 1999; Thompson and Pastor 1995.

of land use stability, especially during the Early Archaic, focusing on facilities that they argue were the focus of deliberate anticipated reuse over hundreds of years. The analysis here relies on excavations carried out at archaeological sites in basin interior and upland marginal areas as well as on the Rock Springs Uplift as part of compliance activities. Because components dating to the Early and Late Archaic and Late Prehistoric are well represented in these data, with Paleoindian and recent components being poorly represented, I only focus on these inter­ mediate time periods. In addition, the majority of the components discussed here come from the Opal (Early Archaic), Deadman Wash (Late Archaic), and Uinta (Late Prehistoric) phases as defined by Thompson and Pastor (1995; see Table 5.3). Thirty Early Archaic to Late Prehistoric components from 11 sites were excavated and reported on by Archaeological Services (Western Wyoming College), Mariah Associates, the Office of the Wyoming State Archaeologist, and the Bureau of Land Management during the 1980s and 1990s (Table 5.4, Figure 5.1). These state-of-the-art endeavors typically present careful descriptions of material culture—pit structure, thermal, and other features; chipped-stone, ground-stone, and bone artifacts; and faunal remains—along with palynological and macrobotanical studies. Importantly for my purposes, almost all include high-quality geomorphological and geoarchaeological ­analyses.

Analyses conducted in the 1980s and 1990s paid very close attention to cultural chronology as well as functional matters; in the 1990s, issues of deposit formation history and settlement stability also were considered, and these are further elaborated on here.

Temporal Structure of Wyoming Basin Components The temporal structure of individual site deposits is owed to both geomorphological factors and human land-use factors operating at many different time scales. Here, I first attempt to assess the nature of the geomorphic packages in which cultural materials were found, highlighting what they can tell us about the availability of that surface to accumulate cultural remains, that is, to record place-use histories. I next consider evidence for actual or realized place-use histories. The final result represents an attempt to recognize variation in place history at the supra-annual, generational, and supragenerational temporal scales.

Geomorphological Factors

Surface stability and sediment accumulation are controlled by both regional and local factors. Regionally, effective moisture and vegetation cover, especially that of grasses, correlate with surface stability (Eckerle 1997:142). These regionwide conditions are differentially expressed in local deposits, documented at archaeological and geological sites, 67

Table 5.4.  Wyoming Basin Excavated Components by Basin Location Site (ID) Interior Basin   48LN1468   Talioferro

  48SW1242

  48UT401   Porter Hollow

Elevation Radiocarbon Comment/ (ft) Component Features Age (bp) Period/Phase Reference 2,600

1

2

5290 ± 190

E.A. Opal

2 3

2 6

E.A. Opal L.A. Deadman Wash

4 5 6 7 8 2

9 6 4 3 0 1

5290 ± 190 1910 ± 110 2590 ± 90 2850 ± 90 1500 ± 70 1310 ± 70 1170 ± 60 960 ± 60

3

5

1 2

1 9

3

0

1

9

2

12

3

6

6,600

4 1

6,860

6,453

6,500

Rock Springs Uplift   48SW2590 7,400   Maxon Ranch

  48SW5175   Sweetwater   Creek

  48SW5215

2170 ± 90 1540 ± 90 1550 ± 80 10,090 ± 120 2200 ± 80 2400 ± 80

L.P. Uinta L.P. Uinta L.P. Uinta L.P. Uinta Recent L.A. Deadman Wash L.P. Uinta

Horizontal and vertical components (Smith and Creasman 1988)

Pit structure

Hoefer 1986

Paleoindian L.A. Deadman Wash

Hoefer 1987

E.A. Opal

Pit structure (Harrell and McKern 1986) Pit structure

11 4

6000 ± 130 6480 ± 90 4760 ± 130 4860 ± 110 2250 ± 100 2180 ± 100 1140 ± 100 5130 ± 90

2 3 1 (A)

3 1 4

4380 ± 200 3170 ± 60 5150 ± 100

2 (B)*

4

1090 ± 60

2 (C)*

4

990 ± 60

E.A. Opal L.A. Pine Spring E.A. Opal Horizontal components (McKern 1987a) L.P. Uinta Reported as part of same component by author but considered separately here L.P. Uinta Designated as component 3 here

E.A. Opal L.A. Deadman Wash L.P. Uinta E.A. Opal

Pit structure; Components 1 and 2 difficult to separate in the field (Newberry and ­Harrison 1986) Vertical components

Table 5.4. (cont’d)  Wyoming Basin Excavated Components by Basin Location Site (ID)

Elevation Radiocarbon Comment/ (ft) Component Features Age (bp) Period/Phase Reference

Basin Margin/Upland   48FR1468 6,890   McIntosh

  48FR1602   Crooks

  48LN2555

  48SU1006   Trappers   Point

  48SW5057   Buffalo Hump

6,920

6,640

7,300

6,770

1

2

2

1

1

4

4850 ± 70

2

9

4300 ± 70 4360 ± 90

E.A. Opal

1

5

5260 ± 90

E.A. Opal

2

13

L.A. Pine Spring

3

5

3070 ± 60 3180 ± 60 3250 ± 90 3420 ± 70 1470 ± 70 2360 ± 90

4 7

0 1

4690 ± 110

E.A. Opal

5

7

E.A. Opal

3

2

5160 ± 210 5390 ± 70 5440 ± 80 5490 ± 60 5510 ± 160 5590 ± 100 5660 ± 100 5720 ± 70 5750 ± 80 5900 ± 160 6010 ± 130 6180 ± 200 7880 ± 60

1 2 3

4 18 18

1480 ± 60 1250 ± 60 1290 ± 60

L.P. Uinta L.P. Uinta

Note: E.A. = Early Archaic, L.A. = Late Archaic, L.P. = Late Prehistoric.

2770 ± 80

L.A. Deadman Wash L.A. Deadman Wash E.A. Opal

Horizontal components (Newberry and Hoefer 1987)

Pit structure (McKern 1987b) Pit structures; one is a reconstruction of Component 1 PS Vertical components (Reust et al. 1994)

L.P. Uinta L.A. Deadman Wash Vertical components (Miller et al. 1999) Antelope bone abundant

E.A. Harrell 1989 Pit structures Pit structure

LuAnn Wandsnider as a consequence of the landform position of that location with respect to potential sources of sediment, local wind directions, and so forth. Geomorphologists reporting on archaeological sediments routinely distinguish among those sediments deposited because of alluvial, eolian, colluvial, and fluvial processes as well as the nature of the contacts between sediment packages. In addition, it is useful to distinguish eluvial (accumulations of silts and clays owed to subsurface weathering) and illuvial (airborne dust that has leached into subsurface horizons) horizons and the paraconformities (deflated surfaces) that commonly cap them. Unfortunately, given the sandy nature of many of these deposits and the fact that archaeological work often precedes geomorphological sampling, the correspondence between archaeologically defined packages of sediments, that is, components, and their boundaries compared with chronostratigraphy defined by geomorphologists in the field and through laboratory analysis is sometimes modest. Here, I rely on these geomorphological interpretations to approach an understanding of surface stability as well as relative rates of sediment accumulation (Table 5.5, page 71.). Abstracted from Table 5.5, Table 5.6 shows that there are the expectable trends in surface stability for components depending on site location. In the interior basin, deflation and aggradation are common, and stable surfaces or surfaces with slow aggradation are rare. In upland areas on the basin margins and the Rock Springs Uplift, sediment accumulation is generally slower and assisted by vegetation entrapment of fine sands, with some evidence of more rapid aggradation for Maxon Ranch (MR) Component 2 (referred to hereafter as Maxon Ranch-2 or MR-2). At this point, actual values cannot be assigned to rates of accumulation or deflation. Also, it is unclear whether we can assume that basin upland “slow accumulation” is equivalent to basin interior “slow accumulation.” Given the different moisture regimes seen in upland and interior basin areas (as reflected by modern vegetation), some differences in sediment accumulation rates under conditions of sparse vegetation likely exist. Table 5.7 reports on those components for which multiple radiocar-

Table 5.6.  Components by Surface Stability and Basin Position (N = 29 Components for Which Information Is Available)

Surface ­Stability Deflation or deflation/   aggradation Stable or slow  ­aggradation Rapid aggradation

Basin Location Margin or Rock Interior Springs Uplift 4

0

3

18

3

1

bon dates, usually from thermal features, are available. Minimum number of radiocarbon occupations (statistically determined using OxCal 3.8 Χ 2 ­goodness-of-fit tests [Bronk Ramsey 2002]) and span of occupation history (determined by simply subtracting the maximum radiocarbon mean from the minimum and by using the OxCal 3.8 span calculation [Bronk Ramsey 2002]) are also presented. Discounting the old wood problem, Table 5.7 indicates that for deposits accumulating on stable or slowly aggrading surfaces, occupation grain values as determined using radiocarbon dates might be either small or large, reflecting actual occupation history. That is, stable surfaces have the potential to accumulate occupational remains over short (Buffalo Hump [BH]-3, Crooks [Crk]-2) or long (Trappers Point [TP]-3, Maxon Ranch-1) time spans, and both extremes as well as values between those extremes are reflected in the Wyoming Basin data. Talioferro (Tal)-3 has a complex deflational and aggradational geologic history. Not surprisingly, widely disparate radiocarbon dates from features are reported here. A similar situation may obtain for LN2555-3, but there is insufficient evidence that deflation has occurred. Rapid aggradation is reported for SW1242-3 and Maxon Ranch-2. For these components, minimal occupational grain is also small. For present purposes, I will assume that for components developed on stable and slowly aggrading surfaces, the potential for a large occupational grain (as distinct from and constrained by geological grain) is high and that there may be some amount of integration in the materials that 70

Talioferro

b

Interfluvial ridge

N

E

No slope

Recent

B4

Medithermal Medithermal

A-N4S2 A-S3

Medithermal

Neoglacial

A-N4

B-2

Neoglacial

A-N2

Medithermal Early Neoglacial

B V

5(B) 7(B) 8(B)

4(B)

3(A) 6(A)

2(A)

2(BD) 1(A)

2 1(A)

Not analyzed by Miller

Deflation; rests in sediments that accumulated during a transition to wetter times; on a paraconformity; stereonet analysis suggests deposits are deflated and severely eroded Associated with an eluvial development in top of a stratum that accumulated during wetter times; also possibly associated with a paraconformity, and cultural remains may represent deflation Dune shadow-type aggradation Dune shadow-type aggradation with significant organic content; likely remains from several occupations present Not analyzed by Miller but likely the same as Component 6(A) above (with differences owed to northerly exposure and near ridge crest position)

Slow aggradation; eolian deposit; fine silts and clays suggest aggradation under vegetated conditions; unconformity attributed to early Medithermal drought at base No information Slow aggradation; wet-phase eolian shadow aggradation in coarse sands derived from conglomerate; no fine sediments available to show illuvial/eluvial modifications

Eolian plain

1

Crooks a

Medithermal

McIntosh

B

Ridge

Site

No slope

Stratum Topographic Thickness Position Aspect Stratum (cm) Emplacement Component Interpretation

Table 5.5.  Deposit Interpretation by Site, Strata, and Components

20 20 20

20

10 60

30

25 0–20

10–20 25

10–20

Miller in Smith and Creasman 1988

Miller in McKern 1987b

Miller in Newberry and Hoefer 1987

Cultural ­Deposit Thickness (cm) Reference

Gently sloping bench

Ridge

Low-relief ridge

Slope

LN2555 c

SW1242 d

Sweetwater Creek

SW5215

Site

SW

SE

No slope

SW

2

1

Neoglacial

Neoglacial

Neoglacial

Medithermal

5 lower

5 upper

C

2(C)

2(B)

1(A)

3 3

2

1

2

Early Medithermal Neoglacial

20

CD (3)

Middle/Late Holocene Middle Holocene

3

IV

0

C/D (3/4a)

Late Holocene

4

Late Neoglacial

25

DE (4a)

Late Holocene

IV

5–10

F (4b)

Slow aggradation; eolian shadow deposit aggrading during mesic intervals; no unconformity Slow aggradation; eolian shadow deposit aggrading during mesic intervals; no unconformity Slow (mesic condition) aggradation; vegetation likely present; “lower” illuvial and eluvial weathering horizons present; no deflation Slow (mesic condition) aggradation; vegetation likely present; “upper” illuvial depositional and eluvial weathering horizons present; no unconformity, no deflation

Rapid aggradation; eolian sand with little pedogenic alteration At interface between lower colluvial deposit and upper eolian deposit Slow aggradation; colluvial deposit with a weak soil; much bioturbation Slow aggradation; eolian deposit aggrading under mesic conditions Rapid aggradation; eolian deposit aggrading under relatively xeric conditions Slow aggradation; eolian shadow deposit aggrading during mesic intervals; no unconformity

Rapid aggradation; eolian sand with little pedogenic alteration; much bioturbation

Stratum Topographic Thickness Position Aspect Stratum (cm) Emplacement Component Interpretation

Table 5.5. (cont’d)  Deposit Interpretation by Site, Strata, and Components

23

24

10–30

10–30

20–30

25–60

20

?

?

?

?

Miller in McKern 1987a

Miller in Newberry and Harrison 1986

Miller in Hoefer 1986

Frederick in Reust et al. 1994

Cultural ­Deposit Thickness (cm) Reference

Finger ridge

Slope

Saddle

Maxon Ranch e

Buffalo Hump

Trappers Point

Site

W

NE

SW

10–18

5–15

10–40

V

VII

Early Neoglacial

Altithermal

Altithermal

Medithermal

VII III

Medithermal

V

VII

V

III

4

3

2

2 1

Medithermal

Altithermal

3

III

20 0

3 2/3

(Late Holocene)

4

1

20–30

4

(Recent)

III?

10

6

Slow eolian accumulation of fine silty sand; anthropogenic A horizon situated on compact eolian sand surface with polygonal cracks Slow eolian accumulation of fine silty sand; anthropogenic A horizon in lightly stained eolian sand

Eolian shadow deposits with illuvial/eluvial development Eolian shadow deposits with illuvial/eluvial development Eolian shadow deposits with illuvial/eluvial development Slow eolian accumulation of fine silty sand; anthropogenic A horizon situated on loamy fine sand; some evidence for deflation

Fluvial deposit with organics suggesting moister conditions Eolian/colluvial deposit with illuvial horizon; much rodent disturbance (the age of which is unspecified) Eolian/alluvial deposit Situated on alluvial fan deposit; overlain by eolian/alluvial deposit No information

Stratum Topographic Thickness Position Aspect Stratum (cm) Emplacement Component Interpretation

Table 5.5. (cont’d)  Deposit Interpretation by Site, Strata, and Components

10–40

5–15

10–18

10

10–15

10–20

10?

20 5

20–30

10

Eckerle and Hobey in Miller et al. 1999

Miller in ­Harrell 1989

Harrell and McKern 1986

Cultural ­Deposit Thickness (cm) Reference

NW

Neoglacial

0

1/2

3–4

Late Medithermal–Recent Medithermal

4–5

1

2

3

Shallow shadow deposits aggrading in a vegetated context with eluvial developments Shallow shadow deposits aggrading in a vegetated context; eluvial developments present Deflation; on disconformity between strata; conflated cultural remains possible 10

8–28

10–30

Miller in Hoefer 1987

Cultural ­Deposit Thickness (cm) Reference

 rooks site assemblages include Component 1, with one pit structure (A), and Component 2, with two pit structures (B and D). Pit structure D was constructed over pit structure A. Miller reports cultural deposits as present in C middle sands from a more arid, active period. But his figures (which show some inconsistencies reflecting the confusion between natural and cultural stratigraphy) suggest that pit structures A, B, and D were constructed when upper sands were deposited, during a more mesic phase. b Talioferro was excavated in two blocks, A and B, each with its own stratigraphy. In addition, different stratigraphies were documented in the northern and southern portions of Block A, distinguished here as A-N2, A-S3, and so forth. Although the stratigraphy reported by Miller for Block A is more finely resolved than that reported by Smith and Creasman, the units developed by the latter are used here. c At 48LN2555, stratigraphy was defined by archaeologists in the field (Strata A–F) and also by Frederick (Strata 1–4). d Component 2 at 48SW1242 was not completely excavated to the base of this unit. e Stratum and component thicknesses interpreted from stratigraphy.

a

Porter Hollow

Site

Stratum Topographic Thickness Position Aspect Stratum (cm) Emplacement Component Interpretation

Table 5.5. (cont’d)  Deposit Interpretation by Site, Strata, and Components

1 2 3 2

3 2 2 3 3 5

3 2 3

Stable/slow aggradation Rapid aggradation Stable/slow aggradation Stable/slow aggradation

Stable/slow aggradation Stable/slow aggradation Stable/slow aggradation Rapid aggradation Slow aggradation Slow aggradation

Rapid aggradation Stable/slow aggradation Deflation/ aggradation

Component Surface Activity

2 2 2 2

2 2 4 2 2 11

2 2 3

2 2 1 1

1 1 >1 2 2 4

1 2** 2

No. Radiocarbon No. Radiocarbon Dates Events

480 100 70 100

40 60 350 890 1,700 850

10 200 940

530–1,430 0–810 0–690 0–120

0–390 0–600 250–840 1,200–1,830 1,860–2,960 380–810

0–500 250–950 1,150–2,050

Years Radiocarbon Span Years Radiocarbon (Maximum Mean– Span (68.2% Confidence Minimum Mean) Interval)*

*OxCal v3.8 (Bronk Ramsey 2002); atmospheric data from Stuiver et al. 1998. **OxCal uses a χ2 test to evaluate the null hypothesis that two or more radiocarbon sample results are actually from the same sample, in this case, occupation event. (If they are from the same event, OxCal recommends combining the results prior to calibration.) For the Porter Hollow results, OxCal returns a value of T = 3.1, which, for df = 1, does not exceed the critical value of Tα = .05 = 3.84 but does exceed critical Tα = .10 = 2.71. On this basis, I argue that it is not impossible that two different radiocarbon events were sampled here.

  SW5215

Rock Springs Uplift   Maxon Ranch

  Trappers Point

Basin Margin   Buffalo Hump   Crooks   LN2555

Basin Interior   SW1242   Porter Hollow   Talioferro

Site

Table 5.7.  Minimum Component Grain

LuAnn Wandsnider accumulate ­during multiple occupation events. That is, the potential for many occupation events and also feature reuse to have occurred on these surfaces is quite high. In deflational settings, the potential for large occupation temporal grain is also present, but the integration of materials between occupation events is variable, depending on when deflation occurred, between or after occupation events. In rapid aggradation settings, small occupational grain is more likely, and, depending on how quickly different occupation events follow each other, components might appear archaeologically integrated. In what follows, I compare and contrast components from basin interior and basin margin/ upland­ settings because previous archaeological work suggests site usage in different seasons and because of the different depositional regimes found there. I also consider the nature of surface activity, for, as discussed above, surface activity constrains the degree to which materials may accumulate on the available surface.

gies were pursued in the Late Archaic and Late Prehistoric, when seed processing and mass harvests of herbivores increased, corresponding to a decrease in pit structure construction. Figure 5.2 reports pit structure dimensions by component and time period. As seen here, Early Archaic and Late Prehistoric pit structures are represented in this data pool. In general, larger pit structures are more shallow than smaller pit structures, but two patterns are visible, shallow-floored and deeper pit structures. Specifically referring to Buffalo Hump pit structures, Creasman and Thompson (1997:277) suggest that small, shallow pit structures with few subfloor features might represent brushy windbreaks around kitchen activity areas rather than actual structures, and they do not distinguish between the slightly more shallow and slightly deeper pit structures seen here. The Crooks site pit structures, both shallow and deep, contain many thermal features and only one feature identified by excavators as a possible storage feature. Larson (1997a) presents a more comprehensive overview of Wyoming pit structures, and those reported on here are generally consistent with her findings, that is, most pit structures are found in margin/upland areas and early pit structures tend to be larger (here, greater than 250 cm in diameter) whereas later ones are smaller. Figure 5.3 graphs fire-cracked rock (FCR) density versus chipped-stone debitage (estimated by summing primary and tertiary flake totals) density, highlighting the differences in composition between basin and marginal or upland assemblages in general. (Density was calculated by dividing totals by amount of area excavated.) In general, higher debitage and FCR densities are found for interior components, and very low debitage and FCR densities are found on the basin margins. The differential availability of toolstone, with sites near the cobble-mantled terraces of the Green River (Love 1977) showing very high debitage and FCR densities, may explain most of this patterning. Also, for interior basin assemblages, as FCR density increases, so does debitage density (Tal-2 being an exception to this trend), suggesting that FCRand debitage-generating activities were being conducted during the same occupation event. There is

Occupation or Place Histories

Geomorphological processes contribute to temporal grain, but so do occupation events. More than 40 years ago, Floyd Sharrock, then working at Pine Springs, nicely laid out the situation: “From the evidence, it was impossible to determine whether the [Component] 1 material represented one group which camped over an extended period of time or returned regularly to the site for a number of years, or if it represented sequent usage by several groups with no significant time elapse between” (1966:22). Using several taphochronometric indicators in concert, I attempt to parse aspects of place-use history so as to partially resolve the conundrum identified by Sharrock. Pit structure presence/absence and form have been used to argue for various interpretations of place histories (Gilman 1987; Larson 1997a). For example, Larson discusses pit structure use by past Wyoming hunter-gatherers in terms of increased sedentism and increased emphasis on plant food storage, especially in the Early Archaic, when possible storage features are reported for pit structures. She argues that more mobile food-storage strate76

Figure 5.2.  Pit structure dimensions by period and location. Symbol designations refer to component and feature identifiers. BH = Buffalo Hump, Crk = Crooks, MR = Maxon Ranch, SWC = Sweetwater Creek, Tal = Talioferro.

Figure 5.3.  Fire-cracked rock versus debitage densities by location. Symbol designation refers to component identifier. Note logged axes. Debitage values for LN2555 components are not compatible with other debitage values and thus are not graphed here. 1242 = SW1242, 2555 = LN2555, 5215 = SW5215. BH = Buffalo Hump, Crk = Crooks, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, SWC = Sweetwater Creek, Tal = Talioferro, TP = Trappers Point.

LuAnn Wandsnider Table 5.8.  Taphochronometric Indicators Derived from Assemblage Information, by Component Bone Site Weathering Component Mode 1242-2 1242-3 2555-1 2555-2 2555-3 5215-1 5215-2 5215-3 BH-2 BH-3 Crk-1 Crk-2 McI-1 McI-2 MR-1 MR-2 MR-3 MR-4 PH-2 SWC-1 SWC-2 SWC-3 Tal-1 Tal-2 Tal-3 Tal-4 Tal-5 Tal-6 Tal-7 TP-3 TP-5 TP-7

1 1 >1 3 3 .00 .00 .00 1 1 3 >1 .00 >1 .00 .00 .00 .00 >1 .00 .00 .00 .00 >1 3 3 3 3 3 >1 1 1

Area (m2) 220 220 102 66 154 48 32 36 219 227 12 35 60 12 115 115 100 100 56 96 96 60 102 58 116 58 92 128 92 27 86 38

FireExternal Standard Debitage Cracked Hearth Mean Deviation Density Rock Density Density Hearth Hearth (per m2) (per m2) (per m2) Area (cm2) Area (cm2) 7.17 29.62 –5.00 –5.00 –5.00 .69 2.09 1.33 .42 .56 1.42 .26 2.45 7.92 11.66 5.40 5.07 8.78 18.50 3.48 2.74 1.75 86.03 97.93 46.77 33.83 55.04 48.77 124.63 359.19 25.28 458.29

.59 2.73 12.26 71.83 11.07 .10 7.97 1.69 2.05 2.45 1.17 .26 .53 .00 .27 .51 .85 2.32 2.70 1.25 1.60 1.60 1.76 .00 7.97 5.34 5.98 3.20 7.74 .00 22.20 .00

.00 .02 .05 .18 .03 .08 .13 .11 .03 .04 .00 .00 .03 .08 .09 .10 .06 .11 .16 .03 .03 .02 .02 .03 .05 .16 .07 .02 .03 .07 .08 .00

1,812.67 2,670.87 1,318.47 4,186.60 3,558.04 2,468.28 2,317.57 1,547.09 1,206.11 1,471.40 3,700.48 2,075.88 2,379.92 594.98 3,498.90 2,647.93 3,263.70 3,173.62 1,096.97 3,426.20 1,258.80 14,338.52 1,888.99 2,047.12 1,712.43 3,240.89 1,666.66 2,302.23 2,799.91 882.93 1,922.99 .00

.00 3,224.87 869.91 2,983.69 1,091.27 1,681.78 2,504.72 673.18 671.66 902.91 2,855.03 1,674.39 .00 .00 2,159.11 1,490.41 1,632.23 1,702.56 836.85 1,672.66 .00 .00 .00 113.49 1,079.89 3,266.45 1,155.52 3,464.50 1,935.68 .00 1,372.11 .00

Note: 1242 = SW1242, 2555 = LN2555, 5215 = SW5215, BH = Buffalo Hump, Crk = Crooks, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, SWC = Sweetwater Creek, Tal = Talioferro, TP = Trappers Point.

some evidence for a negative relationship between FCR and debitage density for the Rock Springs Uplift assemblages (Maxon Ranch, Sweetwater Creek [SWC], SW5215), perhaps reflecting that available toolstone could be used in heat-assisted processing or to make chipped-stone tools but not both. Trappers Point, located at a relatively high el-

evation but near the Green River, shows very high FCR and debitage densities. Table 5.8 summarizes selected taphochronometric indicators suggested to be sensitive to occupation frequency and integration for artifact assemblages and features. Approximately a hundred primary variables, for example, excavated area, 78

Time-Averaged Deposits and Multitemporal Processes

Figure 5.4.  Mean hearth area versus hearth density by location. Symbol designation refers to component identifier. 1242 = SW1242, 2555 = LN2555, 5215 = SW5215. BH = Buffalo Hump, Crk = Crooks, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, SWC = Sweetwater Creek, Tal = Talioferro, TP = Trappers Point.

chert primary debitage, and mean hearth area, were extracted from reports; I have experimented with another 70 indicators, for example, thermal feature density, derived from these primary values. Hearths can serve a variety of functions, for food and material preparation, for heating purposes, for light, and for sweat baths. Ethnoarchaeological work across a range of settings has revealed trends that relate to the history of place use but, again, are likely highly contextual and even idiosyncratic. For example, for the arid Western Desert of Australia, Nicholson and Cane (1991) report a relationship between numbers of occupants and numbers of hearths, in part related to the number of sleeping fires, often large, that are constructed for warmth. Yellen’s (1977) !Kung data indicate that specialized hearths (for head roasting, arrow preparation, etc.) are constructed on the margin of domestic areas and that such specialized hearths, corresponding to rare events, accumulate as occupation length increases. Fisher, Strickland, and Strickland (1989), reporting on their work with Efe camps, find no such pattern but do find several

hearths in each dwelling. As well, hearths may go in and out of use to accommodate social and physical factors that arise as occupation continues (e.g., the Mask site [Binford 1978a]). Finally, hearths may increase in size as they are repeatedly used and cleaned out; Yellen’s (1977) data for mapped !Kung campsites shows hearth “creep” with reuse. A fuller treatment of patterning in hearth size and spatial arrangement by context and occupation history is required but is not attempted here. Suffice it to say that increasing hearth density likely bespeaks greater overall occupation time. Hearth size, perhaps responsive to particular functions (sleeping fire, mass processing of roots) but also reflecting clean out events, may also inform on occupation history. The remaining figures present information on thermal features with respect to other variables. In Figure 5.4, hearth mean area is graphed against overall hearth density. The Crooks components and Porter Hollow (PH)-2 have higher hearth densities, and the former have very high mean areas, compared with other components. For the Crooks 79

LuAnn Wandsnider components, excavation occurred solely within the pit structures found there, so hearth density may appear inflated. The inferred season(s) of use, however, may be of interest; that is, perhaps the high hearth density relates to winter occupation events. In fact, perhaps because information pertains to interior space or because occupation did occur in the winter, no seasonal indicators are reported for these components, in contrast to other components where pollen, fauna, or macrobotanical remains offer some indication. In what follows, because of the likelihood of greater sensitivity to occupation history, I focus on external hearth density, calculated using the numbers of hearths occurring outside of pit structures or in components for which no pit structures are reported. The Crooks components are not represented in these graphs, and the aberrantly high mean hearth size for SWC-3 also precludes its display. Figure 5.5a–d presents mean hearth size versus external hearth density coding components in a variety of ways. I first consider estimated numbers of occupation events determined using numbers of bone weathering modes and radiocarbon events (see Table 5.7). Although Lyman and Fox (1989) challenge the dating potential of bone weathering as initially argued for by Behrensmeyer (1978), southwestern Wyoming archaeologists find it useful to assess relative depositional contemporaneity or the degree to which different preservational regimes existed at a particular site. For example, for Trappers Point-5, researchers argue that the variation seen in bone surface weathering is owed to geomorphic rather than temporal factors (yet radiocarbon assay results would support that multiple occupations occurred here). For this reason, numbers of bone weathering modes were not considered for components with a deflational or ­deflation/​ aggradation­history. (Bone weathering assessments are not reported for the Rock Springs Uplift components, and, furthermore, many hearths there were either lined [Sweetwater Creek components] or built into bedrock [MR-1: 100 percent hearths in bedrock; MR-2: 94 percent; MR-3: 83 percent; MR-4: 9 percent], so hearth area may not monitor actual occupation history in the same fashion as it does for other components.) 80

Figure 5.5a shows that low mean hearth sizes and low external hearth densities are reported for components for which only one occupation event can be determined. As mean hearth size increases, so does the number of estimated occupation events; likewise for external hearth density. The highest numbers of estimated occupation events correspond to those components with very high mean hearth areas and external hearth densities. Indeed, on the basis of qualitative analysis of multiple lines of evidence, Reust and colleagues (1994) explicitly state that LN2555-2 appears to represent several occupation events. Figure 5.5b considers the same relationship, this time coding components according to surface activity. Reports, especially those from the 1990s and later, often discuss whether hearths might be deflated or the degree to which the remains on surfaces may represent a palimpsest (e.g., TP-3 [Francis and Sanders 1999:41]). Thus, we might expect components with a deflational history (Tal-1, Tal-2, Tal-3, Tal-7) to preserve relatively fewer hearths and therefore show an artificially depressed external hearth density. Tal-3 and Tal-7 (Figure 5.3) manifest relatively high FCR densities, consistent with the proposition that hearth features have deflated but FCR has persisted. Miller (in Smith and Creasman 1988) conducted a stereonet analysis of Tal-1 remains and concludes that it was seriously eroded and deflated. Some evidence for the deflation of cultural materials is also reported for Tal-2. Figure 5.5c presents this information again, this time coding components for number of pit structures documented (symbol shape) and for the nature of the site structure (feature tethered, feature coincident, feature tethered/coincident, and feature negative). Gross aspects of site structure can be interpreted by looking at how lobes of high-density artifact, faunal, or FCR distributions (depicted using contour mapping) articulate with features. Typically, high artifact density corresponds to small artifacts accumulating as primary refuse (Metcalfe and Heath 1990; Simms and Heath 1990). If features appear to anchor such lobes (e.g., SW1242-2, Tal-3, Tal-4, Tal-5, Tal-6, Tal-7), then a simple structure, representing a limited span of occupation (hours, days), few occupants, and low debris-generation

Time-Averaged Deposits and Multitemporal Processes activities, is indicated (Wandsnider 1996). On the other hand, high-density (likely, small) artifact areas are sometimes coincident with features (e.g., SW1242-3, TP-3). In this case, specialized hearths may have been constructed in midden areas that developed through extended occupation, as argued by Bamforth and colleagues (2005) for the Allen site. Alternatively, operational hearths may have been shifted to accommodate changing physical or social circumstances, as seems to have occurred at unconstrained open-air sites (in contrast to cave sites [Galanidou 1997b; Goreki 1991]) that were occupied for a day (or longer [Wandsnider 1996]). Both of these alternatives are congruent with an interpretation of a single but extended occupation. Finally, hearths associated with one occupation event may have been constructed, deliberately or not, in the primary refuse of an older occupation event or vice versa. Such an interpretation is consistent with repeated occupation events without integration of activities between those events. Interestingly, for some components, usually in basin margin or upland settings (e.g., LN2555-2, BH-3, MR-4), site structure that is both feature tethered and feature coincident is evident. A feature-negative pattern is reported for TP-5, meaning that features and highdensity artifact concentrations seem to be mutually exclusive. Such a pattern may develop if hearthand chipped-stone- or faunal debris–​generating events occurred contemporaneously or visitors to a littered, nonabsorbent surface sought out unlittered work space. (TP-3 may manifest a similar pattern, but the area excavated here is both small and irregularly shaped, precluding a good assessment of site structure.) In Figure 5.5c, many of the components with a simple, feature-tethered site structure are located in the lower-left quadrant of the graph. Generally more complex, feature-coincident or featuretethered­/coincident site structures are seen for components suggested in Figure 5.5a to have more complex occupation histories. In addition to deflated component Tal-7, components LN2555-3 and McIntosh (McI)-1 are exceptions to this trend. In these instances, perhaps function is trumping occupation history; that is, whatever activities occurred here over the short term, a few very

large hearths were required. The more complex site structure apparent at BH-2 and BH-3 might be attributed to either higher occupant density or extended occupation length of the pit structures there, as opposed to reoccupation. Other components with pit structures (Tal-6, SWC-1; MR-2 is an exception) are also located in the left half of the graph. The location of pit structures may have constrained the ­location of other features and, with site reoccupation, may have invited feature ­reuse if sufficient time had elapsed for vermin to decline (Wandsnider 1992). Indeed, pit structure D (Crooks-2) was located ­directly above pit structure A (Crooks-1), indicating some integration here even between components. If artifact and fauna minimum number of individuals (MNI) amounts might serve as a rough measure of amount of activity (as assumed by Dancey [1973] and argued by Varien and Ortman [2005]), then such densities are generally supportive of the trends reported above, as seen in Figure 5.6a–​d. That is, higher relative densities are seen for components with complex occupation histories. (Relative densities were calculated by producing z‑scores for density values according to geographic location. For example, density values for all Rock Springs Uplift components were standardized independent of the standardization conducted for interior basin or basin margin components, allowing for the gross control of differences in chipped stone sources, animal habitats, and so forth. Trappers Point components were considered with the basin margin classes of components for ground stone and mammal densities but with the basin interior components for FCR and debitage densities. Density z‑scores were then classified.) Note that for chipped stone and FCR, just a few components have extremely high densities, so that the majority of the components appear to have moderate (z‑score: –1 to 0) and high (z-score: 0 to 1) densities. A simple sum of standardized ground stone, FCR, and MNI densities in Figure 5.7a again shows relatively high densities associated with those components that other evidence suggests had complex (longer or extended) occupation histories. The same is seen when chipped-stone debitage density z-scores are added to the preceding score (Figure 81

Figure 5.5.  Mean hearth area versus external hearth density by (a) estimated minimum number of occupation events, (b) surface stability, (c) pit structures and site structure, and (d) seasonal indicators. Symbol designation refers to component identifier. 1242 = SW1242, 2555 = LN2555, 5215 = SW5215. BH = Buffalo Hump, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, SWC = Sweetwater Creek, Tal = Talioferro, TP = Trappers Point.

Figure 5.5 continued.

Figure 5.6.  Mean hearth area versus external hearth density by (a) relative d ­ ebitage density z-score, (b) relative fire-cracked rock density z-score, (c) relative ground stone density z-score, and (d) relative minimum number of individuals density z-score. Symbol designation refers to component ­identifier. 1242 = SW1242, 2555 = LN2555, 5215 = SW5215. BH = Buffalo Hump, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, SWC = Sweetwater Creek, Tal = Talioferro, TP = Trappers Point.

Figure 5.6 continued.

Figure 5.7.  Mean hearth area versus external hearth density by (a) the sum of relative fire-cracked rock (FCR), ground stone, and minimum number of individuals (MNI) density z-scores; and (b) the sum of chipped stone, FCR, ground stone, and MNI density z-scores. 1242 = SW1242, 2555 = LN2555, 5215 = SW5215. BH = Buffalo Hump, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, Tal = Talioferro, TP = Trappers Point. .

Time-Averaged Deposits and Multitemporal Processes

Figure 5.8.  Occupation frequency and integration. 1242 = SW1242, 2555 = LN2555, 5215 = SW5215. BH = Buffalo Hump, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, SWC = Sweetwater Creek, Tal = Talioferro, TP = Trappers Point.

5.7b). (Because of how chipped stone is reported, the LN2555 components are excluded here.) Figure 5.5d considers these assemblages in terms of available seasonal indicators interpreted by Creasman and Thompson (1997) and augmented by individual site reports. Sweetwater Creek (SWC-1, SWC-2) and Maxon Ranch components as well as TP-5 have evidence for spring occupations; many of the interior basin components as well as SW5215 components and McI-1 show evidence (usually in the form of pollen) for summer occupations; and the Buffalo Hump components appear to be owed to fall occupation events. Of course, it is difficult to say if all occupation events responsible for a particular component occurred during only one particular season. The LN2555 components are interesting in this regard, in that both spring (fetal bones and eggshells) and fall (high charred seed counts) occupations are indicated. The Sweetwater Creek and Maxon Ranch components are also interesting because, as mentioned above, hearths here are often lined (Sweetwater Creek) or constructed in bedrock (Maxon Ranch). Perhaps particular spring tasks required a specific hearth environment. In

terms of occupation history, there is no obvious patterning in terms of seasonal indicators. Based on the patterns reported here, I offer the following interpretations in Figure 5.8. The lowerleft quadrant cluster of components (along with McI-1 and LN2555-3) appear to reflect single (or a few) occupation events. The remaining components are the result of several or more occupation events. If features were reused and no new hearths were added to the surface, that is, reoccupation occurred with integration, then the low hearth densities but high mean hearth sizes seen in the upper-left quadrant would result. Reoccupation with little integration resulted, I suggest, in the components seen in the right half of the graph. The degree of reoccupation for these components is reflected by increasing values of external hearth densities, mean hearth areas, and artifact densities.

Discussion Mark Varien (1999, 2002) and colleagues (Varien and Mills 1997; Varien and Ortman 2005) working in the prehistoric Puebloan American Southwest have relied on accumulations of common artifact 87

LuAnn Wandsnider classes, primarily cooking vessel sherds, to measure the occupation spans of various kinds of architectural deposits. With these inferences, they have been able to approach interpretations of land tenure and evolving political economy there. Can similarly rich interpretations be offered for prehistoric hunter-gatherer land use in the Wyoming Basin? The nature of the archaeological record here requires an explicit acknowledgment of geomorphological matters: How active were surfaces, and how might surface availability constrain occupation grain? Table 5.9 summarizes occupational histories for components by site and time period. Abstracted from Table 5.9, Table 5.10 indicates that for some components, geological factors apparently played little role. In the case of deflating and aggrading surfaces, that surfaces may have been “cleaned” by surface processes seems to have made no difference to potential reoccupants: Tal-1 and Tal-2 show little evidence for reoccupation. At Tal-3 and Tal-7, however, though hearth densities are likely artificially low because hearths have eroded away, FCR and chipped stone densities (as well as bone weathering modes, which were not entertained here in the initial analysis) suggest that several occupations occurred. It is difficult to assess the degree to which those occupational events were integrated. In the case of stable surfaces, the picture is likewise mixed. Even though some surfaces were mostly stable or slowly aggrading, they received little in the way of reoccupation for components SW1242-2, SWC-2, LN2555-1, BH-2, BH-3, and TP-7. Other stable surfaces, however, perhaps because of their stability, witnessed multiple occupation events. In some cases, reoccupants apparently reused features (e.g., MR-3 [some integration], 2555-3, Crk-1, Crk-2, McI-1). But for other components (PH-2, SW5215-1, SW5215-2, MR-1, MR-4, LN2555-2, TP-3, TP-5), new features were added with reoccupation. Table 5.11 elaborates further. For the components for which multiple radiocarbon assays were made, it shows that components likely owed to multiple nonintegrated occupations also show large spans in the radiocarbon dates. Where integrated occupations were inferred, typically the number of radiocarbon events determined for the two or more

samples was found to be one. SW5215-2 is an exception to this trend; perhaps the two radiocarbon samples dated here refer to one of the many events that occurred here. The implication of this finding is that the degree of integration may be informing on the nature of reoccupation. Integrated occupation events may be closely spaced in time such that the same persons were involved (à la Yellen’s [1977] Camps 3 and 7) or the facilities there were in good repair (à la Yellen’s [1977] Camps 1 and 4). Such components may reflect a kind of short-term persistent place (sensu Schlanger 1992) use. Components analyzed as owed to nonintegrated reoccupation events may reflect locale reuse. That there exists a bit of a gap (Figure 5.8) between reoccupied but integrated components and components for which repeated occupation but no integration occurred suggests a very simple model of place and land use: single-use places, short-term persistent places, and locales that, for whatever reason, accumulated many single uses. For rapidly aggrading contexts, in some cases ­location reuse occurred within the window of aggradation, resulting in complex, integrated deposits (SW1242-3, Tal-6, SWC-1). For other components (Tal-4, Tal-5, MR-2), it appears that reoccupation occurred at a slower pace than aggradation, resulting in preserved feature-tethered distributions of artifacts and high hearth and FCR densities. And Figure 5.9 shows no simple pattern between water availability and place history. Talioferro, a low-elevation, interior basin site located near Slate Creek (which drains into the nearby Green River), shows evidence for single occupation events (Tal-1, Tal-2) and repeated occupation events both with (Tal-6) and without (Tal-4, Tal-5; possibly also Tal-3 and Tal-7) integration. That seven stacked and horizontal components were documented here indicates that something was bringing people back to this area. Investment in facilities here may have been unnecessary or, because of the inability to project future visits, not warranted. Similarly, the springs near Maxon Ranch site may have contributed to its attractiveness and may in part explain what appear to be repeated occupation events, but, it is interesting to note, there is little evidence for integration. On 88

  Sweetwater Creek   (6,600 ft; ephemeral)

  Maxon Ranch   (7,400 ft; springs)

Rock Springs Uplift   SW5215 (6,860 ft;   ephemeral)

4—Aggrading surface; many occupations without integration 5—Aggrading surface; several occupations without integration 6—Stable surface; several integrated occupations 7—Deflating surface; several occupations

3—Aggrading surface; several integrated occupations (two radiocarbon dates but one radiocarbon event)

Late Prehistoric

1, 2—Stable surface; several occupations 3—Stable surface; several occupations without without integration (2—two radiocarbon integration dates but one radiocarbon event) 3—Stable surface; many occupations with some 4—Stable surface; many occupations without 1—Stable surface; many occupations without integration (two radiocarbon dates but one integration integration (span: 530–1,430 years) radiocarbon event) 2—Rapidly aggrading surface; many occupations without integration (span: 0–810 years) 1—Aggrading surface; several integrated occupations 2—Stable surface; single short occupation

1, 2—Deflating surface; single (short) occupation

2—Stable surface; several occupations without integration (span: 250–950 years) 3—Deflating surface; few (short) occupations (span: 1,150–2,050 years)

Porter Hollow (6,500 ft; ephemeral) Talioferro (2,600 ft; Slate Creek)

       

Period

2—Stable surface; few occupations

Late Archaic

Basin Interior   SW1242 (6,453 ft;   ephemeral)

Site (elevation; nearest water source) Early Archaic

Table 5.9.  Summary of Occupation History by Site and Time Period

3—Stable surface; several occupations without integration 5—Stable surface; many occupations without integration (span 380–810 years) 7—Stable surface; likely single occupation

1, 2—Stable surface; several integrated occupations (2—two radiocarbon dates but one radiocarbon event) 1—Stable surface; several integrated occupations 2—Surface stability unknown; several occupations without integration

2, 3—Stable surface; single extended occupations (3—two radiocarbon dates but one radiocarbon event)

3—Stable surface; several integrated occupations

Late Prehistoric

Note: Format = “component number—surface activity; number of occupations and degree of integration” (see text). Note span indications for those components for which multiple radiocarbon dates were obtained.

  Trappers Point   (7,300 ft; Green   River)

  McIntosh (6,890 ft;   Crooks Creek)

  Crooks (6,920 ft;   ephemeral)

  Buffalo Hump (6,770   ft; ephemeral)

2—Stable surface; many occupations without integration (span: 1,200–1,830 years)

1—Stable surface; single occupation

Basin Margin   LN2555 (6,640 ft;   springs)

Period Late Archaic

Site (elevation; nearest water source) Early Archaic

Table 5.9. (cont’d)  Summary of Occupation History by Site and Time Period

Time-Averaged Deposits and Multitemporal Processes Table 5.10.  Components by Occupation History and Surface Activity Occupation History

Surface Activity Deflation/Aggradation Stable

Single/few

Tal-1, Tal-2

Several occupations Tal-3, Tal-7 (­cannot assess integration) Several integrated ­occupations Several nonintegrated occupations

Rapid Aggradation

1242-2, SWC-2, 2555-1, BH-2, BH-3, TP-7

MR-3 (some integration), 2555-3, Crk-1, Crk-2, McI-1 PH-2, 5215-1, 5215-2, MR-1, MR-4, 2555-2, TP-3, TP-5

1242-3, Tal-6, SWC-1 Tal-4, Tal-5, MR-2

Note: 1242 = SW1242, 2555=LN2555, 5215 = SW5215, BH = Buffalo Hump, Crk = Crooks, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, SWC = Sweetwater Creek, Tal = Talioferro, TP = Trappers Point.

Table 5.11.  Multiple Radiocarbon Dated Components (with 68 Percent Confidence Interval Span Determination) by Number of Radiocarbon Events and Nature of Integration Integration Not Integrated

No. of Radiocarbon Events

Integrated

1

1242-3 MR-3 (some integration) BH-3 Crk-2

>1

5215-2

PH-2: 250–950 years Tal-3: 1,150–2,050 years MR-1: 530–1,430 years MR-2: 0–810 years 2555-2: 1,200–1,830 years TP-5: 380–810 years

Note: 1242 = SW1242, 2555 = LN2555, 5215 = SW5215, BH = Buffalo Hump, Crk = Crooks, MR = Maxon Ranch, PH = Porter Hollow, Tal = Talioferro, TP = Trappers Point.

the other hand, the Buffalo Hump components, SWC-2, and SW1242-2, all located near ephemeral water and with little indication of reoccupation, may represent opportunistic and limited occupation events of the kind described by Yellen (1977) for Camps 5 and 6, where rare climatic events permitted occupation. Finally, Table 5.9 focuses on particular sites through time. Sometimes components with similar histories are found, whereas at other sites, very different histories unfold, yielding components of very different character. For example, the three components described at SW5215 appear very similar during the Early Archaic as well as the Late Prehistoric. The same is true of the Late Prehis-

toric components of Buffalo Hump and the Early Archaic components at Crooks Creek. Others show more variable histories, as for Talioferro and LN2555. There are too few data points to comment on patterns others have noted except cursorily. That is, the greater stability in Early Archaic land use associated with investments in facilities like pit structures and lined cylindrical hearths was nevertheless accompanied by single (Tal-1, Tal-2) or recurring use of places without integration (TP-3, TP-5). Similarly, the higher levels of mobility argued for the Late Prehistoric associated with mass harvests of large mammals and seed harvesting was also ­accompanied by the integrated recurring use of 91

LuAnn Wandsnider

Figure 5.9.  Mean hearth area versus external hearth density by water permanence. 1242 = SW1242, 2555 = LN2555, 5215 = SW5215. BH = Buffalo Hump, McI = McIntosh, MR = Maxon Ranch, PH = Porter Hollow, SWC = Sweetwater Creek, Tal = Talioferro, TP = Trappers Point.

places (e.g., Tal-6, LN2555-3). It is likely that technological organization and place histories speak to different temporal ranges and, thus, can nicely complement each other. With more information of this sort, as well as the rich context provided by a fuller treatment of the excavated materials, we may be in a better position to offer statements translating place histories into land tenure trends through time with the specificity Varien and colleagues have been able to achieve for the American Southwest.

as described by Bailey (1981, 1983, 1987, 2007). As applied to the Wyoming Basin, I have especially highlighted the role of geomorphological processes and how they affect grain (the potential span of time during which surfaces are available to record human activities), occupational grain (the span of time during which occupation actually occurs), and integration (the degree to which features are acknowledged and reutilized between occupational events). Clearly, much work needs to be done in refining and calibrating the taphochronometric indicators we have. Equally important, however, is how to relate various place histories (along with other archaeological materials) to the short-, medium-, and long-term processes of interest to us. Here lies the great interpretative challenge of time perspectivism.

Conclusion Archaeological components are rich mines of temporal information, even in lieu of standard chronometric markers. Here, I have tried to use several taphochronometric indicators, still in need of refinement, to approach an understanding of place history for various Wyoming Basin components. Through the vehicle of place history, in conjunction with other archaeological information, it seems possible to approach reckonings of placeuse histories as well as various ecological and social processes responsible for archaeological variation,

Acknowledgments I thank Steven Creasman, Craig Smith, Lance McNees, and Cynthia Craven for sharing materials on which this chapter is based and hope that 92

Time-Averaged Deposits and Multitemporal Processes I have not abused their ideas too greatly. Simon Holdaway introduced me to the theoretical package called “time perspectivism,” which has helped immensely to move me beyond several impasses. James O’Connell provided the welcome prompt, “So what? What does this buy us?” Earlier versions of this chapter were presented at the University of

Utah, Salt Lake City, at the Midwest Archeological Center, Lincoln, Nebraska, and in a graduate seminar in archaeology at the University of Nebraska– Lincoln; comments received at each venue were most useful. And Ann Ramenofsky, Julie Stein, and an anonymous reviewer provided useful comments on an earlier draft of this chapter—thank you.

93

6

Investigating Persistent Places in the Northern Great Plains, Central North Dakota Mathew A. Dooley University of Wisconsin–River Falls

The concept of time is central to the understanding of change and should play a key role in the interpretation of human pasts. To date, however, time concepts in archaeology remain largely underdeveloped for several reasons, some of which include (1) the tendency for archaeologists to think in terms of synchronic time processes, (2) the poor understanding of multitemporal processes leading to cultural change, and (3) the lack of a theoretical framework that incorporates cultural processes operating at rhythms not visible in the ethnographic present. Bailey (1981, 1983, 1987, see also 2007) confronted these issues very early, suggesting that archaeologists adopt a time perspective, focusing on the temporal nature of both archaeological deposits and the cultural phenomenon under investigation. As he has noted, “Different time scales bring into focus different concepts and different sorts of processes, requiring different sorts of explanatory variables” (1987:​7), and the nature of questions asked of different archaeological records must conform to the temporal structures contained within those deposits. Consequently, cultural processes operating at small time scales, such as those observable ethnographically, are rarely discernable in archaeological deposits. Similarly, Binford has suggested that “the time frame of ethnography is largely inappropriate for archaeological research” because “even under the best of circumstances the archaeological record represents a massive palimpsest of derivatives from many episodes” (1981a:197). He further writes that

the archaeological record is “not a poor or distorted manifestation of ethnographic ‘reality,’ but is most likely a structured consequence of the operation of a level of organization difficult, if not impossible, for an ethnographer to observe directly” (1981a:​198). Later, Smith (1992) expanded on these ­issues, specifically with regard to the study of change within an archaeological context. According to Smith, archaeologists must pay special attention to “the degree of (chronological) refinement possible for a given archaeological situation” and “the degree of (chronological) refinement needed to address specific research objectives” (1992:28; emphasis in original). As does Bailey (1987), Smith emphasizes that the temporal structure of archaeological deposits indeed limits the kinds of past cultural processes that may be investigated (see also Murray 2004, this volume). Though the issues brought forth by Bailey, Binford, and Smith (also see Murray 1999a; ­Olivier 1999) have serious epistemological implications concerning the nature of information that is accessible about the past, time perspectivism has been slow to take hold and has generally made little impact on current archaeological research. Murray criticizes both processual and postprocessual archaeologists because they “have yet to seriously engage with time perspectivism or with the implications of our developing comprehension of the structural properties of archaeological records as phenomena” (1999a:​22). If small-scale processes visible in the ethnographic present are largely 94

Investigating Persistent Places in the Northern Great Plains i­ nvisible in the archaeological record, it is necessary to draw from elsewhere in the formulation of questions regarding human pasts. There is a need for time perspectivism in archaeology; however, this challenge has not easily been met because little is known concerning cultural processes that operate over the medium and long term (Murray 1997) and the formational histories of archaeological deposits at those temporal scales remain poorly understood. Perhaps more importantly, archaeologists lack a set of well-defined methods for analyzing surface archaeological deposits that have accumulated over relatively long periods of time at a landscape scale. Bailey refers to these deposits as “spatial palimpsests” and notes that temporal relationships between spatially disparate deposits have “become blurred and difficult to disentangle” (2007:207). The purpose of this chapter is to explore methods for assessing and evaluating spatial palimpsests in a portion of the Northern Great Plains. Using nonsite methods, the intensity of temporally extensive place use, or persistence, is assessed according to the spatial and compositional structure of surface features, and this distribution is evaluated according to subsurface deposits. As an exploratory exercise, I attempt to assess the qualities of differentially persistent places that may have attracted people, or not, over long periods of time.

tipis once rested. Cairns, or prehistoric rock piles, tend to co-occur with stone rings and are generally thought to have been used to construct drive lines or to mark burials, cache pits, middens, or trails.

The Temporal Nature of Stone Feature Deposits

Here, I discuss the temporal nature of stone feature deposits in terms of chronometric dating and temporal resolution. The dating of stone features in the Northern Great Plains remains a significant challenge for archaeologists. This is largely because stone features have typically been documented in places where little deposition has taken place, resulting in the poor preservation of fragile material such as bone and datable hearths (Deaver 1999). Furthermore, stone features are generally noted for their lack of temporally diagnostic lithic material (Frison 1991). Diagnostic artifacts and radiocarbon dates that have been recovered, however, generally indicate occupations ranging from 5000 to 300 bp (Frison et al. 1996). Research conducted by Deaver (1989) is consistent with this time range. His work, however, suggests that the majority of stone features occurring in Montana and North Dakota were constructed during an 800-year period, ranging from 1,600 to 800 years ago. Of 65 absolute dates associated with stone rings, 63 fall within this time, with only two indicating occupations before 2,000 years ago. Temporally diagnostic artifacts also suggest a similarly tight temporal distribution. Of 121 projectile points, the majority (107) are associated with Besant (2000–1100 bp) and Old Women (1100–​ 200 bp) complex occupations. Very few stone rings indicate occupations earlier than Besant, and even fewer indicate Avonlea (1500–​1000 bp) or Protohistoric (500–200 bp) occupations. Deaver (1989) has suggested that the mere presence of stone rings in Montana and North Dakota likely indicates an occupation from 800 to 2000 bp. The majority of stone features in Montana and North Dakota have been associated with Besant complex groups, which have been characterized as being highly mobile hunter-gatherers depending primarily on bison hunting for subsistence in upland and open prairie environments (Deaver and Deaver 1988; Frison et al. 1996). Besant complex occupations have typically been identified by the

Stone Features on the Northern Great Plains Much of the visible archaeological record throughout the Northern Great Plains is characterized by prehistoric stone features, with stone rings and cairns being most frequently reported. Stone rings, often referred to as tipi rings, are typically 3 to 7 m in diameter and may be composed of a single course or wide bands of stones (Frison 1991). Though the function associated with these features was once a topic of heated debate, extensive archaeological and ethnographic data suggest that many of these features are the structural remains of prehistoric lodges similar to the ethnographically documented tipi. It is assumed that stones were used to secure lodge covers and that upon abandonment, stone rings remained, marking the place where individual 95

M athew A . Dooley presence of large, side-notched, dart-type projectile points, with some corner-notched varieties appearing later (Frison 1978; Frison et al. 1996). The Besant complex was once thought to have occurred in the Northern Plains from 2,000 to 1,100 years ago (Gregg 1985). Deaver and Deaver (1987) have argued, however, that the Besant complex occurred from 2,300 to 800 years ago, thus overlapping the Pelican Lake and Avonlea complexes. More recently, excavation data from the Bees Nest locale suggest that Besant complex occupations may have occurred as early as 3,000 years ago in central North Dakota, extending from the Late Plains Archaic period to the Late Plains Prehistoric period (Peterson and Peterson 1995). The inability to identify individual occupations has been perceived as one of the most limiting factors in stone feature research in the Northern Great Plains. This is in large part because short-term ­ethnographic-scale explanations are being sought. The majority of stone features that have been documented occur in locations where little deposition has occurred, which has resulted in palimpsest, or time-averaged, deposits. Excavation data from Montana and North Dakota suggest that the majority of places demonstrating high feature densities have resulted from sequential occupations rather than individual use episodes. Deaver (1999) reports that a substantial number of ­medium-​density­ sites (5–26 rings) and almost all high-​density­sites (greater than 26 rings) that have been excavated represent deposits from multiple use events (see Davis et al. 1982; Deaver 1983, 1985; Fredlund et al. 1984; Gragson 1983; Quigg 1986). He further suggests that the inability to detect multiple occupations is more likely the product of inadequate sampling rather than owed to single occupation events. Prehistoric settlement systems as commonly understood are inaccessible by means of stone feature analysis because spatially associated features are not necessarily contemporaneous, individual places were likely used by different groups for different reasons, and the chronometric resolution required to distinguish individual occupations is generally lacking. In cases where absolute or relative dates are available, it is difficult to know

96

whether these dates are associated with the construction of a particular stone feature. Thus, rather than investigating settlement systems, I focus here on different magnitudes of temporally extensive place use, or persistence, across the landscape. Although the time-averaged nature of stone feature deposits is typically seen as limiting, these deposits might allow the monitoring of medium- and longterm processes and are only limiting in cases where researchers are interested in understanding shortterm behavioral events (see Holdaway et al., this volume).

Persistent Places on the Northern Great Plains Following Schlanger, persistent places refer to places that have been “used repeatedly during the longterm occupation of a region” (1992:92) and, we assume, have certain qualities that have attracted people over extended periods of time. Persistent places may result from the same groups repeatedly targeting specific locations throughout short ­periods of time or from unrelated groups targeting previously occupied places over the long term. Here, I focus on long-term persistence because the chronometric resolution required to monitor short-term place reuse is typically unavailable for most stone feature deposits. In addition, I have chosen to focus on “generic” place use rather than “site function” because as a place is used repeatedly through time, the functional association of individual occupations becomes increasingly difficult to see (Binford 1980). Research that has focused on prehistoric hunter-gatherer campsite selection has typically focused on environmental variables, some of which include topographic variables (e.g., slope), distance to critical resources (e.g., water), and viewshed (the amount of terrain visible from a single point on the landscape [Allen et al. 1990; Jochim 1976; Kvamme 1985]). In general, these variables are selected because they are relatively stable, relatively easy to measure using maps or geographic information systems (GIS), and help identify less suitable locations for establishing campsites. More recently, however, several archaeologists have explored the role that

Investigating Persistent Places in the Northern Great Plains the material consequences of previous occupations may have played in terms of prehistoric settlement choices (Camilli and Ebert 1992; Schlanger 1992; Wandsnider 1992, 1998). Here, I focus on the cultural landscape, attempting to identify places with longer or deeper histories of use.

This research is relevant here for two reasons. First, the material ­residuals of previous occupations may have attracted humans to certain places over the long term. Second, if a locale containing recyclable material was suitable for occupation, it might have been reoccupied. Prehistoric hunter-gatherers who occupied the Northern Great Plains surely considered many variables when they chose to occupy one place rather than another. Though many of these variables were likely related to the physical environment, people may have been attracted to places that exhibited visible traces of prior settlement for a number of reasons. I am not suggesting that people would have chosen to occupy prior settlements if the resources required for a particular activity were lacking. Rather, I am suggesting that if several portions of the landscape demonstrated similarly attractive environmental settings (i.e., distance to critical resources), prehistoric hunter-gatherers might have preferentially targeted those places where archaeological residuals were found.

The Cultural Landscape and Persistence

Environmental factors were certainly important regarding hunter-gatherer land use in the Northern Plains (see Frison et al. 1996; Quigg and Brumley 1984); however, the cultural landscape might also have been important. I use the term cultural landscape to signify the humanly modified landscape (see Sauer 1963 [1925]) and suggest that as the cultural landscape evolved through time, certain portions of the humanly created environment might have attracted, or deflected, prehistoric occupation. Several researchers have explored how traces of earlier settlement might have affected land use taking place at a later time. For example, Camilli and Ebert (1992) have examined the relationship between the visibility, or discoverability, of reusable archaeological material and the frequency of use events over the long term. They determine archaeological visibility according to the stability of landscape surfaces through time. It follows that ground surfaces that have remained stable will provide conditions suitable for the discovery and scavenging of archaeological material. Their research suggests that where lithic raw material is scarce, stable surfaces demonstrate greater frequencies of reuse because lithic material was available for scavenging. Wandsnider (1992) uses ethnographic data to develop three general rules of reoccupation with respect to previously abandoned facilities (i.e., feature, resource) and sites (i.e., locale). First, if facilities are in good repair, the locale is free of pollution, and resources are not depleted, facilities and locales are typically reused. Second, if facilities are in poor repair or the locale is unsuitable for occupation, scavenging will typically occur, with usable items being transported to a new location. Finally, if a locale demonstrates facilities that no longer hinder reuse, it is available for occupation.

The Study Area The area of focus consists of 18,120 contiguous acres (7,336 ha) in Mercer County, North Dakota, and is situated north of the Knife River and south of Lake Sakakawea (Figure 6.1). Within this area, 1,955 surface features have been recorded throughout approximately 4,700 ha undisturbed by modern agricultural practices. The study area is situated in the glaciated subsection of the Missouri Plateau, with the general ­topography consisting of rolling uplands made up of glacial till and shallow depressions that seasonally hold water. The upland physiographic zone bordering the Missouri River, where the study area is ­located, is generally referred to as the upland plains (Ahler et al. 1991). The physical landscape is typical of that formed by glacial advancement and retreat, with successive layers of glacial till and erratics overlaying tertiary formations (Wyckoff and Kuehn 1983). The two prominent physical features in the area include the Missouri Trench, lying approximately 12 km to the north, and the Beulah Trench, located directly to the east. The study area

97

M athew A . Dooley

Figure 6.1.  Location of study area.

is highly dissected by intermittent and permanent drainages because of its close proximity to the Beulah Trench. The vegetation here is typical of that found in the mixed-grass prairie, with patches of woodlands located in drainages and coulee bottoms. In terms of modern land use, the study area is predominately range and cropland. The Missouri Trench and surrounding area are thought to have been a focus of prehistoric activity within the region because of the variety and abun-

dance of resources available (Toom 1996). Some of the resources that likely attracted people to this area include a variety of plants and animals, Knife River Flint for stone tool manufacture, wood as a source of energy, and water for consumption and processing activities.

Data Collection

Ethnoscience Inc. conducted a pedestrian survey of the study area for compliance purposes during summer 2000. Survey transects were placed at 98

Investigating Persistent Places in the Northern Great Plains a maximum spacing of 30 m, with surface visibility ranging from 5 to 10 percent (Boughton et al. 2000). Within this area, 1,955 surface features were identified, including stone rings, cairns, depressions, and stone alignments, with stone rings (n = 1,347) and cairns (n = 407) predominating. During summer 2001, Ethnoscience Inc. relocated the stone features and recorded their positions using Global Positioning System technology, with positional accuracy being greater than 1 m. In addition, several observations were made that relate to feature siltation, degree of colonization by lichens (i.e., lichening), and stone ring completeness. As discussed below, these three attributes allow the monitoring of different intensities of long-term place reuse across the landscape. Subsurface testing was also conducted during summer 2001. In addition to shovel tests and soil probes, a minimum of one 1-×-1-m excavation unit was placed within at least one stone ring at each of the stone ring sites, with the particular features tested chosen by individual field supervisors while on-site. Over 220 test units were excavated using arbitrary 10-cm levels according to depth below the current ground surface. The sediment was screened through ¼-inch mesh, in which over 6,030 pieces of chipped stone, 138 lithic tools, 44 ceramic sherds, and 105 diagnostic bone fragments were recovered (Boughton et al. 2001:Appendixes E–I). Although absolute dates for the stone features are currently unavailable, projectile points recovered throughout the study area are consistent with the temporal range of 5000 to 300 bp typically associated with stone feature deposits in the Northern Great Plains (see Deaver 1985, 1989; Frison et al. 1996; Peterson and Peterson 1995), with the majority indicating Besant complex occupations ranging from approximately 3000 to 800 bp. It is difficult, however, to be sure that the use of spatially associated projectile points is appropriate for the dating of individual stone features. If a place was used repeatedly through time, a projectile point may have been discarded long before, or long after, the construction of a spatially associated stone feature. While caution must be taken here, it is assumed that the majority of stone features occurring within the study area were constructed from approxi-

mately 3000 to 300 bp, with some constructed earlier and some constructed later.

Quantifying Persistence Although full-scale excavation would likely provide valuable information regarding the spatial distribution of persistence throughout the study area, I focus on the analysis of surface data. Surface data cannot, and should not, replace subsurface investigation; however, information collected at the surface is cost efficient and provides significant information regarding place reuse when archaeological remains are not otherwise endangered. Here, I attempt to quantify both relative age and persistence (degree of use and reuse) using bodies of reference knowledge, or middle-range theory, based in feature siltation, lichening, and the completeness of stone rings. I then evaluate the performance of these variables using subsurface data.

Bodies of Reference Knowledge Siltation

Siltation refers to the degree to which a feature is buried by the natural accumulation of sediment through time and has been used as a means of relatively dating stone features in the Northern Great Plains (see Deaver 1989). It is assumed that in localized settings, stone features exhibiting high degrees of siltation were constructed at an earlier time than those buried only moderately. All stone features within the study area were classified on an ordinal scale from 1 (shallow) to 4 (deep). Contextually speaking, deeply buried features are likely older than those on the surface. Areas demonstrating a greater variety of siltation classes (e.g., features are low, moderately, and highly silted) likely exhibit greater temporally extensive reuse over the long term. In contrast, areas that exhibit a low number of siltation classes (e.g., all features are moderately silted) may demonstrate less temporally extensive reuse because features were likely constructed and abandoned within a relatively short period of time. There are several ambiguities in the application of this reference knowledge to features within the study area. First, siltation is highly contextual, being dependent on the localized geomorphic history of a particular place. In order to appropriately 99

M athew A . Dooley c­ ompare two geographically disparate locations, the locations must have similar depositional histories. For example, an area that has demonstrated very low amounts of deposition through time may contain features that exhibit only low levels of siltation, regardless of the amount of time separating occupations. In contrast, an area that has demonstrated high amounts of deposition might exhibit features with only high levels of siltation, representing only the most recent series of occupations. In this situation, features that were constructed at earlier times would be completely buried and not detectable at the surface. Furthermore, deposition may not have occurred at an even tempo over the long term. For example, if deposition occurred very slowly and later very rapidly, low- and moderately low-silted features might demonstrate a larger temporal gap in their construction than moderately and highly silted features. Though it cannot be assumed that the tempo and intensity of siltation were identical throughout the study area, I assume that different degrees of siltation provide a general indication of different intensities of place reuse over the long term. However, siltation data provide a very coarse relative dating method and will inform about reuse episodes that were separated by a significant amount of time. Reuse over short periods, such as years or decades, is not resolvable using these data. Furthermore, the ability to differentiate between occupation episodes decreases with respect to the age of features. This degree of temporal resolution, however, is highly contextual, depending on different intensities and tempos of deposition through time.

Lichening Here, I use the word lichening in reference to the degree of lichen accumulation on surface stones. Lichenometry, or the dating of a stone’s exposure on the surface according to lichen growth, has proven to be an effective means of establishing relative or absolute dates for surface features in areas where sufficient middle-range research has taken place (see Bettinger and Oglesby 1985; Broadbent 1987, 1990). Lichenometry is based on the slow, and nearly regular, growth rate of certain crustose lichens. An exposed stone that has remained in

conditions favorable to lichen growth for ­longer periods of time will demonstrate greater lichen thalli growth than a stone situated in similar conditions for shorter periods of time. In localized settings, features that exhibit greater overall lichening were likely constructed at an earlier time than those that demonstrate less overall lichening, allowing the relative dating of stone features. All stone features within the study area were classified according to an ordinal scale from 1 (low) to 4 (high). Similar to the case for siltation, areas demonstrating more variation in feature lichening (e.g., low, moderate, moderately high, and high) likely exhibit greater temporally extensive reuse than areas that exhibit less variation in feature lichening (e.g., low and moderate only). Though lichening has been employed as a relative dating method in North Dakota (Deaver 1989), lichenometry in the Northern Plains is rarely employed and remains highly underdeveloped. Consequently, little is known regarding the growth rate and the conditions favorable for lichen colonization throughout the study area. It has been shown elsewhere that the tempo of lichen colonization is contextual, depending on temperature, moisture, and sunlight (Benedict 1967; Curry 1969; Innes 1985). This introduces several ambiguities regarding the use of lichen data as a relative dating method within the study area. First, in order to compare overall feature lichening in geographically disparate locations, the locations must demonstrate at least similarly favorable, or unfavorable, conditions for lichen colonization. Second, the extent to which feature recycling will affect lichen colonization is poorly understood (see Bettinger and Oglesby 1985 for effects of repositioning). Finally, stones will accumulate lichen even before they are used in the construction of features; feature construction will not necessarily provide a “clean slate” for lichen accumulation. So, although several disadvantages exist in the use of lichen data to quantify persistence, it is assumed that different intensities of overall feature lichening provide a coarse indication of relative feature age. However, lichening provides a coarse relative dating method only, with short-term reuse, again, largely irresolvable. Lichening data are likely

100

Investigating Persistent Places in the Northern Great Plains

Figure 6.2.  Frequency diagram showing the difference between siltation and lichening for individual features: –3 indicates features that demonstrate low siltation and high lichening ranks, 0 indicates similar siltation and lichening ranks, and 3 indicates features that demonstrate high siltation and low lichening ranks.

more temporally sensitive in more recent periods of time but are still contextual, depending on local environmental conditions.

Comparing Siltation and Lichening Ranks If both siltation and lichening were equally appropriate for the relative dating of stone features, we would expect to find a positive relationship between the intensity of siltation and lichening among individual features. A Spearman’s rank correlation reveals a weak, but highly significant, positive relationship between siltation and lichening ranks (r = .41, p < .001). Figure 6.2 is a frequency diagram depicting the difference between siltation and lichening among individual features, with –3 indicating features that demonstrate low siltation and high lichening ranks and +3 indicating features that demonstrate high siltation and low lichening ranks. The majority of features demonstrate similar ranks of lichening and siltation (60 percent), with few features (2 percent) deviating substantially from the expected relationship. However, a considerable number of features (38 percent) demonstrate a difference in siltation and lichening by one to two

ordinal ranks. It appears, therefore, that lichening and siltation do not operate similarly or at an even tempo throughout all portions of the study area. Further research is warranted to better understand in which contexts siltation or lichening data will perform better for the relative dating of surface features. Siltation and lichening as chronometric indicators are treated independently within a given area for the quantification of persistence because little is known regarding different siltation and lichening processes throughout the study area. However, more confidence is placed in the siltation data because middle-range research regarding lichen colonization in the Northern Great Plains remains highly underdeveloped.

Rock Robbing Rock robbing, a term used by Deaver (1989), refers to the recycling, or robbing, of stones from previously abandoned stone features for the construction of new features. As locations were used repeatedly through time, many stone rings were likely completely recycled, erasing any evidence of previous occupation. Deaver (1989), however,

101

M athew A . Dooley s­ uggests that sometimes only the closest portion of the ring was “robbed,” leaving incomplete stone rings from prior occupations. If rock robbing can be recorded at a location, it is possible to demonstrate that certain nearby features are not contemporaneous, providing the ability to assign relative dates to stone features. Different degrees of stone ring completeness may provide evidence regarding the frequency of occupation through time. It follows that rings demonstrating a fairly uniform distribution of stones along their perimeter have likely never been robbed and that highly incomplete rings suggest that rock robbing may have taken place, indicating that a location might have been inhabited by tipi-building groups more than once. The incompleteness of stone rings, commonly recorded in terms of the distribution of stone weight around the perimeter of a ring, has typically been attributed to the prevailing wind direction and force, taking place during the time of occupation (Brumley and Dau 1988; Finnigan 1982; Frison 1967; Quigg 1986; Wright et al. 1985). Frison (1967) was one of the first to suggest this explanation when he noted that stone weight concentrations tended to occur on the north and west sides of ring perimeters at the Piney Creek sites in Wyoming. Burley (1990), however, suggests that these studies have failed to produce significant results regarding seasonality, relative dating, or intraring spatial analysis. Moreover, weather patterns in the Northern Plains are dynamic and unpredictable. Surface winds shift dramatically, and often, throughout the course of a single day. If people used a minimal amount of stones, clustering them on the windward side of the tipi, as Finnigan (1983) has suggested, stones would likely have had to be moved daily, and in some cases hourly, throughout the extent of an occupation. If a sufficient amount of stones were placed on all sides of the tipi to compensate for changing wind direction, stones would rarely have to be adjusted. Given the large number of stone rings present throughout the study area, it is unlikely that people were forced to use the minimal number of stones required for tipi construction. Abandoned stone rings likely provided an extensive supply of suitable stones that could be used to construct new tipis. However, data re-

garding the differential availability of surface stone cobbles throughout the study area is necessary but currently lacking. One of the greatest problems with the assessment of ring completeness at the surface is the inability to know what lies beneath the surface. Stone rings that appear incomplete at the surface might reveal themselves as being complete if excavated. Although this is an important factor to be considered, it is likely that a substantial amount of stone rings that appear incomplete at the surface will exhibit the same patterning below surface. While many uncertainties exist regarding the completeness of rings as an indicator of place reuse, it is assumed that areas exhibiting less complete rings result from higher intensities of place use because stone rings were likely recycled through time. Areas that contain more complete rings might have been used to a lesser extent, and rock robbing may never have taken place. Because of the ambiguities discussed above, however, ring completeness is given less weight in the calculation of settlement persistence.

Calculating and Evaluating Settlement Persistence

I consider the variety of feature siltation and lichening rankings, as well as the average completeness of stone rings within a given area, to quantify settlement persistence from approximately 3000 to 300 bp. To do this, a spatial database was created that could be utilized in a GIS. Three independent neighborhood operations were performed for each variable, and these variables were combined in a map overlay operation, as discussed in detail below (Figure 6.3). First, the study area was divided into 30-m grid cells, so that if desired, persistence could more easily be compared with ancillary data sets having the same spatial resolution. The variation of siltation rankings for each 30-m grid cell was computed according to a neighborhood function based on the variety of siltation classes contained within a 60-m search radius from each cell’s centroid. The number of features present within each search radius is not considered. Each grid cell within 60 m of a known feature was assigned a number ranking from 1 to 4. A number 1 indicates that all features

102

Investigating Persistent Places in the Northern Great Plains

Figure 6.3.  Geographic information system operations used to quantify persistence.

within the search radius demonstrate the same class of siltation, and a number 4 indicates that all possible classes of siltation occur within the search radius, regardless of the number of features present. If only one feature is located within the search radius, the cell was assigned a value of 1 regardless of the observed siltation level. Variation in lichen growth was computed in a similar manner. Ordinal ranks of stone ring completeness were calculated according to the variation in number of stones per octant determined by compass direction. Each stone ring was assigned a standard deviation value to describe the variation in stone counts around the perimeter of the ring. Average ring completeness values were computed for each 30-m cell according to a neighborhood function based on the mean standard deviation within a 60-m search radius, and these data were reclassified on an ordinal scale from 1 (high average ring completeness) to 4 (low average ring completeness). Values describing siltation, lichening, and ring completeness were combined in an overlay operation to produce a single data layer describing overall persistence for each 30-m grid cell. The values derived from the siltation data were doubled so that they would play a stronger role in deriving final

persistence values. Final persistence was quantified as follows: persistence index = lichening + (siltation × 2) + ring completeness. The derived values range from 0 to 13. Grid cells assigned a 0 are locations where no surface features were identified; however, it cannot be assumed that these locations exhibit no use by prehistoric groups. These figures were reclassified on an ordinal scale from 1 to 4 according to Jenks’s optimization routine (see Jenks and Caspall 1971) in order to produce a generalized persistence index throughout the study area. Of 8,086 30-m grid cells assigned a value greater than 0, 65 percent are classified as level 1 (low reuse); 23 percent, as level 2 (moderate reuse); 9 percent, as level 3 (moderately high reuse); and 3 percent, as level 4 (high reuse [Figure 6.4]). Approximately 27 locations are described as demonstrating high degrees of temporally extensive place use, with six continuous areas being greater than 600 m2.

Evaluating the Persistence Distribution

The distribution of persistence throughout the study area must be evaluated in order to determine if it contains interpretive potential. The quality associated with the distribution depends on several factors, including (1) the appropriateness of

103

M athew A . Dooley

Figure 6.4.  Persistence index.

the bodies of reference knowledge, (2) the spatial scale of analysis, and (3) the appropriateness of the computational operations. In order to evaluate the quality of this distribution, I use subsurface data collected for compliance purposes. The vertical differentiation of subsurface artifacts is typically interpreted as being evidence of multiple occupations. As sediment accumulates through time, discarded artifacts will be vertically

separated according to the rate of sedimentation and the time at which individual artifacts were discarded. Places that contain artifacts distributed vertically throughout many stratigraphic layers have likely been used to a greater extent than places containing artifacts at only one approximate depth. If the proposed technique measures different intensities of persistence, we might expect to find a positive relationship between the derived ­rankings

104

Investigating Persistent Places in the Northern Great Plains of persistence and the number of excavation levels containing cultural material. It follows that test units located in areas classified as being highly reused will demonstrate a greater number of excavation levels containing cultural material, and test units located in areas classified as being used only minimally will demonstrate a lower number of excavation levels containing cultural material. Though the vertical distribution of artifacts has traditionally been used to detect multiple occupations, several postdepositional (Foley 1981b) or noncultural formation processes (Schiffer 1976) may significantly alter the vertical position of artifacts subsequent to their time of discard. Some of these processes that have likely occurred within the study area include soil formation, plant and animal activity, erosion, and frost thrust and heave, as well as the leaching of small pieces of chipped stone. Furthermore, different rates of deposition throughout the study area will affect the vertical distribution of subsurface materials (Foley 1981b). The excavation techniques employed by field crews constitute a further disadvantage. Arbitrary 10-cm excavation levels may contain only one or possibly many stratigraphic layers, depending on the formational history of sediment matrices. Although these factors pose a significant disadvantage for my purposes, I assumed that general trends regarding the vertical distribution of subsurface artifacts could, at least to some extent, provide information relating to place reuse across the study area. A stratified random sample of 57 1-×-1-m test units, segregated by derived persistence values, was chosen for investigation. Each test unit was evaluated according to the number of excavation levels in which cultural material was recovered. Because these data are significantly skewed, a Kruskal-Wallis­ H test was employed to evaluate if a significant difference exists in the mean number of culturally productive levels and the derived categories of persistence assigned to test unit locations. The results suggest that a moderately significant difference exists among culturally productive levels and the assigned categories of place reuse (χ² = 4.6, df = 3, p = .202). Visual inspection of the associated error bar chart (95 percent confidence interval) and frequency diagrams in Figure 6.5 suggests a simi-

lar trend. The lowest category of place use exhibits the lowest mean number of productive levels, with progressively higher means associated with higher categories of place use. The frequency diagrams demonstrate a similar trend, with a greater number of cultural levels associated with higher categories of place reuse. Although the validity of the persistence distribution cannot be fully verified by these methods, the trend in these data suggests that different intensities of temporally extensive reuse are being detected, at least to some extent.

Feature Attraction and Persistent Places As discussed at the outset, there are several variables that might be considered to better understand why prehistoric hunter-gatherers on the Northern Great Plains targeted some locations more than others over long periods of time. Though highly persistent places tend to occur closer to wooded areas, farther from wetlands, in slightly higher elevations, and where good views are available (Dooley 2004), here, I focus on stone features as a component of the cultural landscape that may have attracted repeated occupation over the long term. If prehistoric hunter-gatherers on the Northern Plains preferentially targeted abandoned stone features, we might expect to find (1) later features tending to concentrate near earlier features, (2) greater intensities of persistence in locations where earlier features occur, and (3) intensities of persistence decreasing as distance increases from early features.

The Cultural Landscape and Feature Attraction

In order to assess the role that previously abandoned stone features might have played regarding persistence throughout the study area, I consider (1) the position of later features with respect to the position of earlier features and (2) the distribution of persistence with respect to the location of earlier features. It follows that if people were mapping onto abandoned places containing archaeological residues for whatever reason, we would expect to find later features and greater persistence clustering near earlier features. The ability to distinguish between earlier and

105

M athew A . Dooley

Figure 6.5.  Comparison of the number of cultural-bearing excavation levels by derived persistence category for a sample of 57 1-×-1-m test units.

later features within the study areas is limited by the lack of high-resolution chronometric data. Here, I collapse all of the stone features into three temporal categories—(1) early, (2) late, and (3) latest—by combining the siltation and lichen rankings of individual features. It should be made clear, however, that these categories serve organizational purposes only and do not reflect contemporaneous or homogeneous settlement periods. Ordinal rankings for siltation and lichening range from 1 to 4, with higher ranks indicating a greater intensity of siltation or lichening. Accordingly, the combined rankings range from 2 to 8. Early features (n = 1,303) represent combined rankings ranging from 6 to 8, late features (n = 373) represent combined rankings ranging from 4 to 5, and latest features (n = 78) represent combined rankings ranging from 2 to 3. The skewed nature of these data likely indicates a bias toward earlier occupations because the chronometric resolution of siltation and lichening data likely decreases as one moves further back in time (see discussion of siltation and lichening data above). Consequently features categorized as “early” likely

represent occupations occurring over a longer time span than those categorized as “later.” In order to assess the spatial congruency among earlier and later features, I compare the position of later features (late and latest combined) with respect to the position of earlier features. Then I compare these results with a random distribution of point locations. Figure 6.6a suggests that the majority of later features occur near earlier features, with their frequency decreasing dramatically as distance from earlier features increases. This trend is significantly different from that produced by the randomly distributed point locations. Figure 6.6b suggests that there is no significant difference between feature type, with later features occurring near stone rings and cairns. Next, I consider the mean distance from earlier features and the presence/absence of early features with respect to the persistence category assigned to 30-m grid cells (n = 1,189). Figure 6.7a suggests that highly persistent places tend to be located near earlier features, with average distance from earlier features decreasing as persistence intensity

106

Investigating Persistent Places in the Northern Great Plains

Figure 6.6.  The position of late features with respect to the position of early features: (a) features compared with random point distribution and (b) comparison of stone rings and cairns.

increases. Similar results are produced when considering the presence or absence of earlier features with respect to persistence category (Figure 6.7b). Here, places classified as 1 indicate locations where only latest features occur. Places classified as 2 indicate locations where late features occur alone or with latest features. Places classified as 3 indicate locations where early features occur alone or with late or later features. It follows that if people targeted places where abandoned stone features are located, greater intensities of persistence should be found where early features occur. Figure 6.7b indicates that average persistence is greater in locations where early features occur, conforming to the expected relationship. These results, however, should be interpreted with caution. Although the persistence categories and the temporal classification of features were calculated independently, the comparison of these data runs the risk of employing circular logic because of the way in which persistence was quantified. Places where early features occur

would have a higher probability of being classified as highly persistent because later features tend to occur near earlier features. Places where only later features occur have a low probability of being classified as highly persistent because features will likely demonstrate less variation in siltation and lichening rankings.

Discussion This study suggests that surface data contain valuable information about different intensities of persistence over the long term and may serve as an alternative information source when archaeological deposits are not otherwise endangered. Here, I use siltation, lichening, and the completeness of stone rings as variables to measure long-term persistence, and I evaluate these results using subsurface data. When compared with an independent test using subsurface assemblages, the combination of siltation, lichening, and ring completeness data

107

M athew A . Dooley

Figure 6.7.  Comparison of persistence and the location of early features: (a) mean distance from early features by persistence category and (b) mean persistence category by place category defined by the presence/absence of early features.

a­ ppears to measure temporally extensive place use, at least to some extent. These results, however, must be interpreted with caution for several reasons. First, though natural formation processes are in some part responsible for the archaeological patterning observed on the surface, the nature and magnitude of these processes are not well documented for this particular landscape. The mapping and interpretation of surface geomorphology would provide significant information regarding natural formation processes, and these data could be used to better distinguish patterns created by natural processes and those resulting from past human behavior (see Wandsnider, this volume). Second, additional research is necessary in order to evaluate the appropriateness of using siltation and lichening data to establish relative feature chronologies within the study area. Although the majority of the features in the study area demonstrate similar rankings of siltation and lichening, a substantial portion of features deviate from this expected rela-

tionship. The development of lichen growth curves and an analysis of features significantly deviating from the expected relationship may offer insight as to the contexts in which lichening, or siltation, will perform best for the relative dating of stone features. Although further research is necessary, this study sheds light on long-term hunter-gatherer land use in this portion of the Northern Great Plains, namely, that people may have mapped onto places where previously abandoned features occur, as opposed to other locations that exhibited similar local environmental variables (see Dooley 2004). One possible explanation of this phenomenon is that prehistoric people preferentially targeted places where previously abandoned stone rings could be recycled for the construction of new features. It remains unclear, however, whether this observation is the result of prehistoric scavenging, optimal environmental conditions, or some other reason not considered here. Minimally, this study

108

Investigating Persistent Places in the Northern Great Plains suggests that the cultural landscape, or humanly created environment, may have influenced prehistoric land-use decision making in this portion of the Northern Great Plains.

Conclusion Time perspectivism, though largely underdeveloped, offers several challenges for contemporary archaeology. One of these challenges is to develop methods for analyzing time-averaged surface de­posits at a landscape scale. The methods explored here offer one approach for analyzing

time-​averaged­deposits when limited temporal information is available and might be used in other locations where large numbers of surface features exist. Though inconclusive, this study suggests that the material remains of previous occupations may have played an important role in land-use decision making in the Northern Plains and that places exhibiting previously abandoned features may have been attractive campsites over long periods of time. Although several challenges remain, this research might serve as a point of departure for time perspectivism in the Northern Great Plains.

109

7

Assemblage Accumulation as a Time-Dependent Process in the Arid Zone of Western New South Wales, Australia Simon Holdaway, Patricia Fanning, and Ed Rhodes University of Auckland, Macquarie University, and Manchester Metropolitan University Interpretative perspectives which ignore the structural properties of archaeological records are ultimately of dubious utility to archaeologists. —Murray 1999b:4

The quote above was penned more than 15 years after Bailey (1981, 1983) first discussed the tendency for archaeologists to undertheorize the relationships between the temporal scales at which the archaeological record can be viewed and the explanations put forward to interpret this record. Murray (1999a) discusses approaches to time and archaeology sourced from the modern social science literature, critiquing the fact that authors pay very little attention to the processes by which the archaeological record is created. This criticism has been leveled in a variety of forums. Stern (e.g., 1993, 1994a; Stern et al. 2002) has commented on the temporal scale at which explanations are formed for the archaeological record in the deep past, and Wandsnider (2003) has reviewed landscape studies in the Mediterranean with a similar objective in mind. Shott (2003) considers the same set of issues in relation to the explanation of Middle Paleolithic assemblage variability, and the contributors to this volume address the problem using archaeological records from a variety of places and time periods. As reviewed elsewhere (Holdaway and Wandsnider 2006), methods of explanation adopted in hunter-gatherer studies frequently rely on syn-

chronic views of hunter-gatherer behavior. Stasis is emphasized over variability, with change coming only in the form of punctuated events. Explanation tends to use a single temporal scale (normally of short duration). Current explanatory approaches to hunter-gatherer archaeology may be divided into two broad explanatory camps, labeled “settlement system” and “evolutionary strategic” (Holdaway and Wandsnider 2006). In the first, either direct ethnographic analogy or a substantive reading of Binford’s (1980) site-type typology for foragers and collectors is used to re-create idealized settlement systems, where site locations are related to such independent variables as environmental zoning and the location of key resources. The archaeological record is interpreted as the outcome of a mobility strategy responsive to these variables. In the second, past behavior is interpreted as a response to risk, the behavior optimized relative to a particular currency measured in terms of access to such items as raw material and food resources. It is the substantive application of these ­models that is of concern because they lead inevitably to a view of the past for hunter-gatherer peoples that is characterized by stability. This, we feel, denies

110

Assemblage Accumulation as a Time-Dependent Process people with a particular economic mode a historical past. The continuity stems largely from what Wandsnider (2003) describes as a synchronic, functional, and essentialist metaphysic. Studies in both camps rely on the stable nature of past environments and past human–environmental interactions, a notion that, as we shall demonstrate below, is anything but applicable in our study area of western New South Wales, Australia. This stability also extends to the characterization of the archaeological record. Although it is recognized that some locations may retain an archaeological record that is disturbed, and therefore of limited analytical value, the distribution of the archaeological record across the landscape is often interpreted as an approximation of the distribution of people in the past. Site locations on maps are taken to represent concentrations of activities, which in turn may be related to environmental clustering of resources (e.g., Veth 1993) or the material representation of short-term social events (e.g., Thorley 1998). Many hunter-gatherer peoples were mobile, so archaeologists expect the archaeological record to represent the results of multiple occupations by peoples who used many locations on a seasonal basis. Nevertheless, there exists a view that too much overprinting reduces the data quality at particular locations. The best-quality sites—Gamble’s “flagship” sites (1999b:​68)—are those that preserve a limited set of activities, the closest equivalent to living floors. Palimpsests mix distinct occupations and therefore must be pulled apart into their constituents if site function is to be determined. According to this view, the greater the degree of behavioral overprinting, the less the analytical utility for functional reconstruction. Therefore, archaeologists make every effort to find assemblages that are as close to pristine living floors as possible, and when the archaeological record departs from this ideal, analysis is directed at accounting for, and thereby removing, the effect of any “disturbance.” This view of palimpsests is valid only if the goal is reconstruction of synchronic lifeways. Time perspectivism, as articulated by Bailey (1983), Stern (1994a), and Murray (1993, 1999a), suggests an alternative view: pattern will emerge only through the passage of time, and therefore analysis should

not be directed at negating its effect. According to this view, palimpsests are not only desirable but also necessary, for without the accumulation of materials from multiple behavioral episodes there will be no patterns to analyze. The task of the archaeologist is not the reconstruction of prehistoric lifeways but the analysis of the variability created through the historical accumulation of material at one or more locations in the landscape (Schlanger 1992). Patterns in the archaeological record occur through the interaction of independent processes occurring at different chronological scales (Dewar and McBride 1992). This chapter aims to illustrate how a time perspectivism view of the past might be constructed using techniques that appeal neither to behavioral reconstruction nor to stable evolutionary systems. We seek to construct a historical view, but one that incorporates variability of place use through time rather than falling back on a set of punctuated transitions from one stable state to the next. We illustrate the approach with data drawn from research conducted in western New South Wales, Australia (Figure 7.1). We begin by investigating the temporal structure of place use and move on to consider duration of occupation.

Holocene Archaeology of Western New South Wales, Australia Abundant silcrete and quartz outcrops and gibber plains (stony deserts) ensured a wealth of raw material for the Indigenous Australians who occupied western New South Wales (NSW) during, and prior to, the mid- to late Holocene. Today, stone artifact scatters are visible on most eroded surfaces in the region, with literally thousands of artifacts lying between the deflated remains of heat-retainer hearths (or earth ovens). These hearths were constructed by digging a shallow pit, lighting a fire, and adding stone or sometimes ant nest fragments that act as heat retainers. When the stones were hot, food was added and slowly cooked under a layer of earth (Holdaway et al. 2002). It is likely that hearths were constructed for a single use or perhaps reused over a number of days; however, once abandoned, they were quickly covered by sediment. Severe erosion since the end of the nineteenth century has 111

Figure 7.1.  Western New South Wales, showing locations mentioned in the text.

Assemblage Accumulation as a Time-Dependent Process

Figure 7.2.  The deflated remains of a heat-retainer hearth. The tape in the photograph measures 50 cm.

e­ xposed the remains of these hearths at the surface, removing the sediments into which the pits were excavated and leaving only small clusters of heatfractured rocks that, in many cases, preserve deposits of charcoal (Figure 7.2). If these hearths were reused in the past, evidence of this is no longer apparent when they are excavated. For archaeological purposes they can be treated as the result of single events. As will be described below, the hearths are important because they provide a ready source of material for radiocarbon determinations. The change in land use from hunter-gathering to European pastoralism in the mid-nineteenth century altered the vegetation cover on the arid slopes and plains of western NSW, promoting surface runoff at the expense of infiltration. The result was widespread loss of topsoil through wind and water erosion; deposition of sediments on the valley floors as a distinctive unit commonly referred to as post-European material (PEM); and subsequent stream incision leading to accelerated surface stripping, channel enlargement, and knick-point retreat (Fanning 1994, 1999). In some contexts, stone artifact scatters and hearths, previously buried in the

topsoil and beneath the PEM, became exposed and lagged onto a common surface. Thus, erosion has exposed a rich archaeological record but one that lacks an obvious means for establishing a chronology. Point forms that seriate, the basis for relative chronologies in many parts of the world, are largely lacking from the NSW stone artifact assemblages, as are other forms of material culture showing clear changes through time. A new suite of artifacts did appear in the mid-Holocene­, including some points, backed pieces, and adzes, but whether these changes occurred rapidly or more gradually is unknown (Hiscock and Attenbrow 1998). Further, once the new forms became established, there are no clear temporal trends from this period until the present. Australian archaeologists have therefore turned to rockshelter sites as a source of stratified deposits from which to assess the degree of change in artifact assemblages through time. Since Harry Allen’s (1972) inaugural study at Burkes Cave more than 30 years ago, excavations in shelters have frequently been conducted in association with survey and analysis projects, the results of the excavations used 113

Simon Holdaway, Patricia Fanning, and Ed Rhodes as a justification for assuming that the technology and typology of stone artifacts show little or no change through time (e.g., Thorley 1998; Veth 1993). The current distribution of artifacts across the surface is thought to indicate both the locations in the landscape used by people and some notion of the tasks that they performed. The temptation is to progress from this description of the archaeological record to a commonsense reconstruction of a prehistoric settlement system. The distribution of archaeological materials becomes an approximation of the distribution of people in the past. Stability in tool forms and stability in the locations where these tools appear suggest stability in the history of land use, punctuated by changes like the appearance of new artifact types in the mid-Holocene. The detail of the reconstructed settlement systems often owes its origin to substantive ethnographic analogy (e.g., M. A. Smith 1989, 1996; Veth 1993). In what Holdaway and Wandsnider (2006) refer to as a tyranny of lifeways interpretation, a conceptual leap is made between the distribution of the archaeological record and the nature of past behavior, with scant regard for the temporal scale at which the evidence can be viewed or the scale at which the interpretation is constructed. As we argue below, this is particularly problematic in western NSW, because the apparent uniformity of the environment, both past and present, is deceptive. The surfaces on which the archaeological record now rests can vary in age by millennia over distances as short as a few hundred meters (Fanning 2002; Fanning et al. in press; Holdaway and Fanning 2003). In addition, age estimates for the heat-retainer hearths indicate a discontinuous history of occupation at specific locations even when surface stability can be demonstrated (Holdaway et al. 2002; Holdaway et al. 2005). Therefore, it appears to us that, far from a uniform settlement system in the past, a discontinuous, complex history of place use better sums up the mid- to late Holocene record. As we discuss below, this has important implications for the way we undertake artifact analysis. Today, the area of western NSW, where this research was undertaken, has an arid climate, with average annual rainfall less than 250 mm and pan

evaporation commonly exceeding 2,000 mm (Hold­away et al. 2000). However, during the midto late Holocene, the climate was both wetter and drier than at present, with most paleoenvironmental evidence indicating a Climatic Optimum around 4000 bp, desiccation and aridity from about 3000 bp to about 1500 bp, and amelioration since that time (Holdaway et al. 2002). The rangelands currently support sheep and cattle grazing on unimproved native vegetation, mostly chenopod shrublands and grasslands, with pastoralists utilizing saline groundwater to water their stock.

The Study Areas Data from three widely separated locations within western New South Wales form the basis of this chapter. The S2 study area is located within the catchment of Stud Creek, which drains a roughly circular 30 km2 area of Cretaceous sedimentary rocks approximately 30 km east of Tibooburra in far northwestern NSW (Figure 7.1). Tertiary silcrete outcrops form high points in the landscape (mesas and escarpments), and silcrete gibbers (stones) mantle the hillslopes below. The valley floors contain alluvial valley fills of late Pleistocene to late Holocene age into which Stud Creek has incised (Fanning 1999; Fanning and Holdaway 2001; Holdaway et al. 2004), and it is on these alluvial fills that hearths and artifacts have been exposed. Fowlers Gap is a pastoral property approximately 250 km south of Stud Creek, in the northern Barrier Range (Figure 7.1). It comprises low hills, strike ridges, and valleys formed in Precambrian rocks, overlain by a veneer of Quaternary sediments. Scatters of stone artifacts and associated hearths are found on eroded surfaces on valley floors in the ranges and hills. Six locations within these hills were targeted for detailed survey. Burkes Cave is located on an alluvial terrace in the Scopes Ranges 120 km east of Broken Hill (Figure 7.1). Quartz and quartzite nodules occur as part of the conglomerate that makes up much of the Scopes Ranges, whereas silcrete outcrops are restricted to the eastern flank of the ranges, some 5 km from Burkes Cave. Artifacts analyzed here come from a small excavation conducted by Harry Allen (1972:​143–144) as part of his doctoral research.

114

Assemblage Accumulation as a Time-Dependent Process The assemblage he excavated was reanalyzed by Justin Shiner and Simon Holdaway in 2001 and is fully described elsewhere (Shiner et al. 2007).

Creek, 28 hearths provided sufficient charcoal for radiocarbon determinations (Holdaway et al. 2002). These calibrated determinations fall into two clear groups, separated by a period of 200–300 years. Bayesian analysis of the determinations confirms the existence of both phases of occupation and the gap when no hearths were constructed. During phases of hearth construction, hearths were constructed, used, and abandoned every few decades. However, there is no evidence to suggest that the occupants at any one time were aware of those who had occupied the valley previously. For example, hearths with quite different age estimates were built side by side with no evidence for reuse. This spatial proximity also suggests that patterns in the age estimates of hearths are not easily explained away as the result of differential erosion. Thus the period during which no hearths were constructed at Stud Creek reflects human behavior (or lack thereof ) rather than the effects of geomorphologic processes. Following the suggestions made by Bailey (1983) in his outline of time perspectivism, we interpreted the archaeological record at a variety of chronological scales (Fanning 2002; Holdaway et al. 2002). Viewed at a scale related to the geomorphic history of Stud Creek, people have been occupying the valley floor over the last 2,000 years. Within each of the two phases of hearth construction, people occupied the valley continuously at a scale measured in millennia but discontinuously at a scale measured in centuries. The discontinuity apparent when the two phases of hearth construction are compared equates to a worldwide climatic signature known as the Medieval Climatic Anomaly (Holdaway et al. 2002; Stine 1994). The significance of the Stud Creek hearth chronology is that it forces us to think about the record at multiple temporal scales and demonstrates that surface records have just as much potential to inform us of changes through time as do stratified deposits. If people in the past left Stud Creek for significant lengths of time, then explanations for the archaeological record relying on assumptions of synchrony in past human lifeways must be viewed as suspect. How can a uniform view of ­human-​ environmental­interrelationships deal with the fact

Developing a Chronology of Surface Deposits: Stud Creek The Western NSW Archaeology Program has, over the past 12 years, developed a chronological framework for surface artifact deposits that has two main components. First, we establish the age of the surface on which the artifacts and hearths are now resting by dating the stratified sediment deposits beneath. Second, we obtain age estimates for the hearths by dating those that preserve charcoal. This two-stage approach allows us to establish the maximum age of surface artifact deposits and to investigate the pattern of hearth use through time. At Stud Creek, we excavated a 30-×-2-×-5-m (deep) trench adjacent to the modern stream channel, both to observe the subsurface stratigraphy and to obtain samples for optically stimulated luminescence (OSL) and radiocarbon age estimates (Fanning 2002; Fanning and Holdaway 2001). The results are sobering. Though one OSL determination indicates that the surface onto which the surface archaeological materials had lagged is no older than 2,000 calendar years bp, a result that ties in well with radiocarbon age estimates from the heatretainer hearths on the surface, the majority of the stratigraphic record indicates short periods of deposition interspersed with long periods of stability or erosion, stretching back to the late Pleistocene (Holdaway et al. 2004:Figure 7.2). Thus, at Stud Creek, it is likely that the surfaces on which archaeological materials older than 2000 bp accumulated have been buried or removed by erosion. If the record at Stud Creek can be generalized to other locations, the results suggest that the age of the surface archaeological record is likely to be a function of the local geomorphological conditions as much as it is an indication of the antiquity of human occupation of the region, a finding in line with studies conducted elsewhere (e.g., Wells 2001). The second part of our chronological framework requires partial excavation and dating of the large number of heat-retainer hearths with which the surface artifact deposits are associated. At Stud 115

Simon Holdaway, Patricia Fanning, and Ed Rhodes that people abandoned a region for a period measured in centuries? Methods are needed that allow us to look at assemblage accumulation as a diachronic rather than a synchronic process.

Combining the Chronological Framework with Landscape Mapping: Fowlers Gap The Stud Creek results are provocative but come from just one small valley. It is possible that the patterns we have detected could be related to the chance abandonment of a small region in favor of areas only a few kilometers distant. To test this possibility, we moved approximately 250 km to the south, to Fowlers Gap Arid Zone Research Station (Figure 7.1), to examine the surface archaeological record in a variety of different geomorphic settings. Artifact and hearth sampling at Fowlers Gap was divided into three separate levels, reflecting the influence of geomorphic processes operating over various spatial and temporal scales on the preservation, exposure, and visibility of the archaeological record (Fanning 2002; Fanning and Holdaway 2004). At the macro scale, we made use of Mabbutt’s (1973) existing “Land Systems” classification of Fowlers Gap Station. Land systems are defined as “an area or group of areas throughout which there is a recurring pattern of topography, soils and vegetation” (Christian and Stewart 1953:68). This comprises a first-level or macroscale classification of the physical environment, focusing on the preservation of artifacts in geomorphically dynamic landscapes evolving over time scales of hundreds to thousands of years. At the meso scale, land systems are further subdivided into geomorphic units, classified on the basis of both form and process. They reflect the effects of geomorphic processes on the exposure of artifacts on time scales of tens to hundreds of years. Finally, detailed mapping of surface features at the micro scale reflects the visibility of artifacts at the time of survey, which is related to geomorphic and other environmental processes, such as vegetation growth, operating on time scales of as little as a few hours to several years (Fanning and Holdaway 2002, 2004).

The two-stage chronological framework developed at Stud Creek was applied at Fowlers Gap across six separate locations with different geomorphic histories. Three of those locations (Fowlers Creek, Nundooka, and Mulga Dam) are described here to illustrate how geomorphic landscape history and absolute dating of both sediments and hearths can provide the chronology required to give a time perspective to the place-use history represented by the artifacts.

OSL Dating of “Archaeological Surfaces”

The Fowlers Creek (FC) survey area is located on a terrace surface at least 5 m above the bed of the currently active channel of Fowlers Creek. A terrace is, by definition, an abandoned floodplain—a surface no longer modified by flooding originating in the active channel. This is usually because the active channel either has moved away laterally or has downcut below its own floodplain so that regular floods no longer overtop the banks. The latter is the case at FC. Rather than excavating a deep trench across the terrace, as was the case at Stud Creek, sediment samples for OSL dating were instead obtained by driving opaque stainless steel tubes horizontally into the sides of two shallow pits excavated into the surfaces on which artifacts and hearths rested (Fanning et al. in press). The Nundooka (ND) land system comprises north-south-trending strike belts of moderately to steeply dipping Devonian Nundooka Sandstone, forming massive cuestas and uplands with broadly rounded or beveled crests with minor rocky faces. Narrow strike valleys are intersected mostly at right angles with short steep tributaries. A larger drainage line, Sandy Creek, is superimposed on this structure, cutting through the strike ridges from west to east. Jansen (2001) conducted a comprehensive study of the past and present geomorphic dynamics of Sandy Creek. Morphologic and sedimentologic criteria were combined with absolute dating of deposits to develop a flood history within the gorge extending back nearly 4,000 years ( Jansen 2001; Jansen and Brierley 2004). This flood history provides a geomorphic context for the artifact scatters and hearths discussed here.

116

Assemblage Accumulation as a Time-Dependent Process Table 7.1.  Optically Stimulated Luminescence (OSL) Age Determinations on Quartz Extracted from Sedimentary Units Underlying Surface Artifact Scatters at Three Locations at Fowlers Gap in Western New South Wales, Australia Location Fowlers Creek Nundooka

Mulga Dam

OSL Age Estimate Lab Number Depth (cm) (Years before ad 1950 ± 1σ) K0147 K0149 K0122 K0123 K0124 K0128 K0129

28 32 18 28 52

9670 +1250/–2040 5620 +1120/–1280 1390 +1010/–220 2680 +840/–400 2650 +1080/–860 1820 +250/–760 5420 +1050/–800

1σ Age Range (Years before ad 1950) 10,920–7630 6740–4340 2400–1170 3520–2280 3730–1790 2070–1060 6470–4620

Note: The determinations were carried out at the Research School of Earth Sciences at the Australian National University in Canberra, Australia.

The ND artifact and hearth survey location is a terrace veneered with fine-grained alluvium adjacent to an ephemeral water hole. The latter is dry most of the time but holds water for several months after most streamflow events. Hearths were excavated into dark red/brown silty clay loam alluvium probably accumulating as slack water deposition during “erosion episodes” that occurred between 1,530 and 960 years ago ( Jansen and Brierley 2004:​ 918), thereby providing a maximum age for the hearths and associated artifacts. The hearths and artifacts were subsequently buried (and protected) by bright red sandy alluvium, probably dating to the immediate post–European contact period. Sheetwash and rill erosion is now exposing these materials at the surface. Samples for OSL dating were obtained from the sedimentary sequence exposed in the walls of a gully traversing the survey area. The Mulga Dam (MD) land system comprises a narrow (up to 2.5-km-wide) strike belt on moderately to gently dipping Devonian sandstone, with subdued foothill ridges and foot slopes. The west-facing foot slopes are mantled with a veneer of windblown sand, and recent erosion has exposed dense scatters of artifacts and hearths. A gully traversing the foot slope again provided access to the sediments underlying the archaeological record for OSL sampling. Quartz grains in the range of 177–212 μm were extracted from the sediment sample tubes and prepared using standard OSL separation techniques

117

(Rhodes et al. 2003). Single-grain OSL distributions were measured at 125ºC using an automated Risø TL-DA-15 Mini-sys fitted with a laser attachment, employing a single-aliquot regenerative-dose protocol, similar to that employed by Olley, ­Pietsch, and Roberts (2004). Age estimates are based on the minimum age model of Galbraith, Roberts, Laslett, Yoshida, and Olley (1999), and dose rates were calculated using in situ gamma measurements and neutron activation analysis, following the procedure detailed by Rhodes and colleagues (2003). The results of the OSL age determinations on sediments from FC, ND, and MD at Fowlers Gap are shown in Table 7.1. They confirm the hypothesis that each of the locations has experienced a different landscape history, reflected in the ages of the sediments in the valley fills. Sediments now present at the surface of the terrace at the FC location were originally deposited in the early Holocene, whereas sediments at ND within 30 cm of the surface are late Holocene in age. The age range of sample K0122 from ND also falls within the time span of Jansen’s “erosion episode 2” ( Jansen and Brierley 2004:918) and confirms Jansen’s (2001) interpretation of these alluvial drapes as the product of “slack water” deposition at distal floodplain locations during flood events. Near-surface sediments at MD are similar in age to those at ND, possibly reflecting a depositional event that was at least catchmentwide (MD is located upstream of ND within the catchment of Sandy Creek).

Simon Holdaway, Patricia Fanning, and Ed Rhodes

Figure 7.3.  Fowlers Creek (FC) hearth age estimates. Solid black bars represent the likelihood, and black-outlined bars represent the posterior. Five phases of hearth construction are indicated.

These results also confirm that the archaeological deposits currently exposed at the surface represent periods of accumulation of different lengths. The longest period of accumulation is represented by the FC artifact assemblage, and the shortest, by MD. That is not to say that Indigenous Australians did not live at the MD location prior to about 1,100 years ago but, rather, that any record of their existence has been obliterated by the erosion associated with flood events.

2005), but summary information for three of the locations (Fowlers Creek, Mulga Dam, and Nundooka) is provided here, together with an analysis of determinations from all of the locations with sidereal ages dating to the last 1,000 years. The calibrated prior and posterior plots for FC radiocarbon determinations are shown in Figure 7.3. The plots are based on a Bayesian analysis to test the proposition that the age estimates fall into five distinct phases separated by gaps distributed across the period from 6000 cal bp to 500 cal bp. This analysis of radiocarbon age estimates differs from some, in that the determinations are not considered to “date” the artifacts with which they are associated. Instead, the radiocarbon age estimates are treated as a data set and analyzed to determine whether they form a continuous or discrete distribution. Application of Bayesian calibration using the DateLab software ( Jones and Nicholls 2002a, 2002b) permits various groupings of age estimates (i.e., multiple periods) to be tested against the null hypothesis that the determinations form a continuous sequence (i.e., a single period). Following the discussion in Jones and Nicholls (2002a,

Radiocarbon Dating of Hearth Charcoal: Using Bayesian Analysis to Confirm Temporal Patterns

Forty-five radiocarbon determinations were obtained from hearths distributed over six sampling locations at Fowlers Gap. These determinations show patterns similar to those from Stud Creek, not so much in the exact timing of hearth construction but in the presence of gaps in the sequence of hearth determinations, both when determinations from the different sampling locations were considered separately and when they were combined. A complete analysis of the Fowlers Gap hearth ages has been published elsewhere (Holdaway et al. 118

Assemblage Accumulation as a Time-Dependent Process 2002b), Bayesian calibration allows additional information to be included in the chronometric analysis. Information of this kind is made explicit in the analysis via a probability distribution, called the prior, which weights the calibrated age estimates toward values in line with our prior expectations. The data act through a distribution called the likelihood. A calibrated value that makes the observed carbon reservoir assay a likely outcome of the 14C observation process has a high likelihood. The prior and likelihood distributions together determine a new probability distribution known as the posterior. Sets of calibrated dates agreeing with the data, and at the same time plausible in the light of prior information, yield a large posterior probability. In Bayesian calibration this posterior distribution is our analysis result ( Jones and Nicholls 2002b). Nicholls and Jones (1998, 2001) provide details of the Bayesian calibration model used in DateLab. It allows the user to place age estimates into phases that occur in a series. Though there are no constraints placed on the relative order of the age estimates in these phases, the user determines the order of the phases, and there is a further constraint that no overlap occurs between the phases. Phases may be defined in a number of different ways allowing the definition of several different prior models. Comparisons may be made between these models by the use of the Bayes factor. If the Bayes factor for Model 2 is placed over the Bayes factor for Model 1 giving a number B, the result may be read to say that Model 2 is B times more probable than Model 1. DateLab provides the mean likelihood for each model, allowing the user to calculate the Bayes factor. Raftery (1996) provides values for interpreting the Bayes factor reproduced here as Table 7.2. Following this table, Bayes factors with values of 12 and above provide strong evidence for accepting Model 2 rather than Model 1. We interpret the Bayesian analysis of the FC radiocarbon determinations (Figure 7.3) as indicating that, at this location, hearths were constructed during five discrete time periods separated by long gaps during which there is no evidence of hearth construction. Various reasons can be posited for the existence of the gaps, and differential erosion is the one that seems most obvious. The single age es-

Table 7.2.  Interpretation of Bayes Factor Model 2 versus Model 1 Bayes Factor Support for Model 2 vs. Model 1 150

Weak; supports Model 1 Barely worth mentioning Positive Strong Very strong

Source: From Raftery 1996.

timate older than 6000 cal bp may be a remnant of a much more extensive occupation that existed in the past, the evidence for which has largely succumbed to erosion. However, there is compelling evidence that erosion is not the primary determinant of the temporal pattern of radiocarbon determinations at FC. For example, there are eight determinations with ages around 4000 cal bp and a further two with ages around 2000 cal bp. One of the hearths with an age estimate of around 2000 cal bp (FC20) is spatially associated with a group of hearths with age estimates of approximately 4000 cal bp (Figure 7.4). Another hearth with an age estimate of around 900 cal bp is close to a second group of hearths also with age estimates of around 4000 cal bp. The presence of these more recent hearths is significant because they indicate that the gap in hearth construction between circa 4000 cal bp and circa 2000 cal bp is unlikely to be related to the erosion of hearths with intermediate ages. Any such erosion would have to differentially select hearths with ages between circa 4000 cal bp and circa 2000 cal bp. Spatially, it is difficult to see how this could happen. The same argument could apply for the period between circa 6000 cal bp and circa 4000 cal bp, although the presence of only a single ancient hearth (compared with eight at circa 4000 cal bp) makes the case much less clear-cut. The results instead suggest that, at least from 4000 cal bp onward, hearth construction was intermittent, with long periods when no hearths were constructed. The inference we draw is that people were either present in very much reduced numbers, insufficient for the record of their presence to be preserved, or entirely absent. This pattern becomes clearer as locations with 119

Simon Holdaway, Patricia Fanning, and Ed Rhodes

Figure 7.4.  Fowlers Creek (FC) hearth locations with age estimates. The hearth locations indicated by dots are shown over a false-color aerial photo where vegetation is indicated by dark gray. Ages are expressed as posterior 95 percent HPT and indicate five phases of hearth construction.

e­ vidence for a more restricted duration of occupation are considered. The calibrated plot for the prior and posterior distributions for the age estimates from the ND hearths (Figure 7.5) suggests a maximum of three periods of hearth construction. The Bayes factor for the three-period model compared with a ­single-​period model is 292, showing strong support for the multiperiod model. Although the pattern is not as clear as at FC, it suggests periods of construction rather like the pattern demonstrated at Stud Creek. These hearths are closely spaced, and, again, it is difficult to argue that differential erosion would leave the hearths from which age estimates

were obtained but remove those with ages coincident with the gaps illustrated in Figure 7.5. The six age estimates from MD show two clear periods of hearth construction, one dated to the period more recent than 500 cal bp and the other to a period around 850–1000 cal bp (Figure 7.6). The Bayes factor for a two-period model compared with a single-period model is 142. Spatially, the hearths from both groups occupy similar locations. Thus, MD153 and MD154 are approximately 8 m apart but differ in calibrated age by upward of 600 years. MD156 and MD158 are separated by 10 m and are around 500 years apart in age. Only MD183 and MD184 have a wide separation in both

120

Figure 7.5.  Nundooka (ND) hearth age estimates. Solid black bars represent the likelihood, and black-outlined bars represent the posterior. Three phases of hearth construction are indicated.

Figure 7.6.  Mulga Dam (MD) hearth age estimates. Solid black bars represent the likelihood, and black-outlined bars represent the posterior. Two phases of hearth construction are indicated.

Simon Holdaway, Patricia Fanning, and Ed Rhodes

Figure 7.7.  Hearth age estimates for the last 1,000 years from Fowlers Gap. Solid black bars represent the likelihood, and black-outlined bars represent the posterior. ND = Nundooka, FC = Fowlers Creek, MD = Mulga Dam, SC = Sandy Creek, ST = Sandstone Tank, NN = Nuntherungie. Numbers refer to individual hearths. Three phases of hearth construction are indicated.

time and space. As at FC and ND, spatial proximity argues against erosion as a mechanism for differentially selecting hearths of certain ages. Determinations from hearths from all six sampling locations having age estimates more recent than 1000 cal bp (26 determinations) are shown in Figure 7.7. The Bayes factor of 9.90 × 105 indicates very strong support for the existence of three discrete periods of hearth construction when tested against a null hypothesis of a single occupation.

Discussion

The age determinations, on both sediments underlying “archaeological surfaces” and hearth charcoal, from Fowlers Gap illustrate two important patterns. First, the age range of hearths from different geographic locations differ by several millennia, and a strong case can be made, as for Stud Creek, that the differences in maximum ages of hearths from the sampling locations reflect the geomorphological circumstances of hearth preservation. The FC location seems unique in retaining evidence for hearths dating from the mid-Holocene:

in other locations, erosion has removed any trace of earlier occupations. Second, both at single locations, such as MD, and when hearths with age estimates more recent than 1000 bp from several locations are combined, there is good evidence for periods of hearth formation followed by significant gaps when no hearths were constructed. Thus, even when surfaces are stable, the hearth record is one of discontinuous ­human occupation. As the Stud Creek analysis suggested, it is difficult to see how this pattern is compatible with synchronic lifeway models. Repeated evidence for abandonment suggests to us that the nature of human interaction with the environment (either physical or social, or both) changed in the past. Interpreting the present locations of the archaeological record as a direct model for human use of the landscape in the past effectively combines locations that were used over substantially different periods of time and archaeological records that are the product of different occupation histories. These records are then treated as though they are

122

Assemblage Accumulation as a Time-Dependent Process all equivalent, the products of a single settlement system or a single stable evolutionary strategy. Such an assumption, we suggest, is not supported by the temporal nature of the archaeological record that we have demonstrated for our study areas.

tions (e.g., Dibble 1987; Dibble and Rolland 1992; see also comments in Shott 2003). With this growth in the sophistication of stone artifact analyses has come a search for alternative modes of explanation. One such mode that has received some recent attention is time (e.g., Bamforth and Becker 2000; Shott 1997, 2003). The theoretical justification for giving less emphasis to function and considerably more to time was provided a number of years ago by Binford (1978a) in his analysis of the Mask site. According to Binford’s informants, the Mask site functioned as a caribou lookout station, yet the range of activities undertaken by the people who visited the site had little to do with the stated function. Instead, Binford documented a range of maintenance and other activities undertaken by people to pass the time. It was these activities that created the abandoned material record. The longer people spent at the Mask site, the more complex the artifact record became. However, studying this record would not allow an archaeologist to decipher what the function of the site might be. The Mask site should not to be interpreted as a cautionary tale but as an indication that the nature of the question posed by archaeologists may be in error. Even expediently created artifacts were often used for a variety of purposes in the past and transported to different locations before abandonment. Artifacts with longer use-lives were used at many different locations; as Wandsnider (1996) comments, assemblage composition tells us about the location of artifact abandonment, not the location of artifact use. Artifact abandonment reflects a number of processes. Clearly, context must be part of the equation, for people have to want to visit a place and undertake activities for artifacts to be abandoned. But context is also important in determining duration, as a more attractive place (for whatever reason) will hold people longer, therefore increasing occupation span. Duration is in turn important because it will influence the number and nature of artifacts abandoned. For instance, increasing the duration of occupation span (and therefore activity) beyond the use-life of a particular artifact form will lead to an increase in the number of these artifacts deposited

Stone Artifact Assemblages as Indicators of Occupation Duration In the second part of this chapter we approach the issue of occupation duration at different points in the landscape using analyses of the stone artifacts that are found associated with heat-retainer hearths. There are two ways in which occupation duration may be considered. First, the term may be used to refer to the total amount of time people spent at a place, summed across a number of individual occupations. Schlanger (1992) uses the term persistent place to refer to site-use history thought of in this way. Second, occupation duration may refer to the time spent at a place during any singleoccupation site occupation span. In the discussion that follows, we focus on assemblage patterning derived from the accumulation of multiple occupation spans that combine to create a picture of place-use history. The quest to determine the function of stone artifacts in the past has preoccupied the attention of archaeologists at least since the 1960s. However, as archaeologists have come to understand both the determinants of stone artifact variability and the processes that led to the creation of the archaeological record in more detail, ideas of re-creating functioning tool kits from concentrations of flakes, cores, and tools have given way to approaches that emphasize an understanding of the processes that underlie the manufacture, use, and discard of stone artifacts. Summarizing a number of studies, Wandsnider (1996) concludes that discard behavior rather than artifact function is a more productive line of inquiry for understanding the intrasite distribution of artifacts. Studies that emphasize the changing morphology of artifacts as a result of a continual process of reworking have redefined debates that pitted those committed to artifact form and assemblage composition as a reflection of style against those equally committed to assemblage composition as a manifestation of differing func123

Simon Holdaway, Patricia Fanning, and Ed Rhodes and visa versa. Alternatively, increasing the number of visits to a particular location (the total amount of time a persistent place was used) will also increase the number of long-use-life artifacts deposited. Assuming that use-life can be controlled for, assemblage composition may therefore be used as a measure of occupation duration in both senses of the term (Shott 1997, 2003). We may not be able to tell where in a landscape people used artifacts that were subsequently abandoned at one location, but a concentration of locations indicating the presence of persistent places will allow us to differentiate favored from unfavored landscapes.

twice as long at S2 and BC compared with ND and MD. Under a conventional synchronic analysis, we might expect S2 and BC to differ from ND and MD because the former locations were exposed to artifact accumulation for longer (i.e., had been persistent places for longer) and are therefore further along the continuum from living floors to palimpsests. It would be reasonable, therefore, to expect that S2 and BC would have assemblages with more variable composition than either ND or MD. In fact, we are able to show the opposite in the analyses presented below. Longer occupation times mean less mobility for hunter-gatherer groups. Therefore, measures of Artifact Assemblages at Stud Creek, one necessarily involve measures of the other, and Fowlers Gap, and Burkes Cave measures of either must be linked to the nature of Duration of occupation in either sense is a relative the archaeological record. In western NSW, this concept, and so the application of the ideas dis- means stone artifacts. As discussed above, the macussed above works best when assemblages from a terial for artifact manufacture (silcrete and quartz) number of locations are compared. Here we com- is locally available as gibber cobbles and as outcrops pare four assemblages from our three study loca- that were quarried. Precise allocation of artifacts to tions in western New South Wales: Stud Creek a source is difficult in this environment; however, (S2), Mulga Dam and Nundooka at Fowlers Gap, it is generally possible, using hand specimens, to and Burkes Cave (BC). differentiate between materials derived from gibAll three locations share a similar environment ber deposits and those from quarries (Doelman (Figure 7.1). They are located within the southeast- et al. 2001). Stony gibber pavements are the major ern margin of the Australian arid zone. Cheno- source of quartz at Fowlers Gap and Burkes Cave pod shrublands dominate the vegetation, although and of silcrete containing macroscopic quartz clasts mulga (Acacia aneura) are scattered along the at Stud Creek. At all three locations, nonclast and ridges, and gidgee (Acacia cambagei) and River Red amorphous silcrete was quarried from rock outgums (Eucalyptus camaldulensis) grow along the crops. Clast silcrete is also derived from quarries at watercourses. Apart from the Darling River, which Fowlers Gap and Burkes Cave. Thus, hand specicarries water derived from the wetter subtropics, men identification is sufficient to set up a simple tythe stream systems are ephemeral, flowing for short pology of local and imported material for each of periods after intense or prolonged rain. Ground­ the assemblages considered. water bores (or wells) provide water for grazing People who were mobile in the past had a greater stock, largely sheep and cattle. opportunity to access distant raw material sources All of the assemblages share a similar chronol- than people who were more sedentary. Following ogy, although hearth age estimates for the two as- Elston (1990), this behavior can be couched in ecosemblages from Fowlers Gap (ND and MD) have nomic terms. People spending a greater amount of hearths that are no older than 1000 cal bp, whereas time at one location (i.e., greater occupation span) those from S2 and the single determination derived will make proportionally greater use of locally from stratified charcoal from BC have maximum abundant materials because there is less opportuages of 2000 cal bp (Allen 1972; Holdaway et al. nity to resupply from distant sources. These people 2002; Shiner et al. 2007). Thus, in a time-averaged­ will also tend to use up material they have brought sense (Stern 1994a), the envelope of time during with them to a greater extent as occupation durawhich the archaeological record was created is tion increases because this is a less costly option 124

Assemblage Accumulation as a Time-Dependent Process

Figure 7.8.  Raw material proportions calculated in three ways for each of the sampling locations. Number = proportion based on the number of artifacts, minimum number of flakes (MNF) = proportion based on the number of complete and proximal flakes (given here), volume (or maximum dimension) = proportion based on the volume (length × width × thickness) for each artifact. For S2, proportion is based on the sum of the maximum dimension in any orientation.

than obtaining new material from a distant source. Elston suggests that raw material quality has an important impact on the economics of supply and demand. Very good raw material is at once desirable because it is easier to work and relatively rare in the landscape. Much more abundant locally available raw material will be used only when it has to be and then to a lesser extent. Elston’s dichotomy works well for western NSW, where the locally abundant quartz or clast silcrete from the gibber pavements may be used to produce sharp-edged flakes. Such material was clearly used less for tools that either require more effort in manufacture (e.g., backed pieces and unifacial Pirri points) or have extended use-lives (e.g., Tula and Burren adzes [see Holdaway and Stern 2004 for tool type descriptions]). However, sources of higher-quality silcrete are sometimes only a few kilometers away, and, even over these short distances, economizing behavior is evident (Holdaway et al. 2004). Following Elston, it should be possible to differentiate assemblages that represent short occupation spans and high mobility (high proportions

of imported versus local material, relatively few tools), short duration spans and low mobility (high proportions of local material, low levels of core reduction), and long duration spans and low mobility (low proportions of heavily worked imported materials, high proportions of local materials). Elston’s ideas are particularly useful because they translate into a series of established measures of assemblage composition (Holdaway et al. 2004; Shiner et al. 2005, 2007). Thus, measures such as flake-to-core and flake-to-tool ratio provide objective measures of core reduction, as do the cortical flake–to–noncortical flake ratio and the noncortical core–to–cortical core ratio. Equally, measures of the degree of tool reuse are easy to construct and provide an index of the degree of tool curation (following the definition in Shott 1996). Tool proportions indicate the relative contributions made to an assemblage by tools with high versus low utility.

Technological Analyses

Figure 7.8 provides raw material proportions for the four assemblages considered here, based on 125

Figure 7.9.  Flake-to-core ratios calculated as the number of complete flakes divided by the number of cores for each raw material class (clast/clast silcrete, nonclast/nonclast silcrete, quartz). Numbers above the columns give the values of the ratios.

Figure 7.10.  Complete, noncortical flake–to–cortical flake ratio for each raw material class (clast/clast silcrete, nonclast/nonclast silcrete, quartz). Numbers above the columns give the values of the ratios.

Assemblage Accumulation as a Time-Dependent Process three measures: number of artifacts, minimum number of flakes, and volume (or in the case of S2, summed maximum length). The use of three measures in this way helps to compensate for biases introduced by any single measure (Shiner et al. 2005). From left to right in the figure, quartz makes up an increasing proportion of each assemblage at the expense of both types of silcrete, until the S2 assemblage. At this location quartz is quite rare, and clast silcrete forms the locally available stone source in the form of gibber cobbles. Thus, there is considerable variability among the four assemblages. BC and S2 have comparatively low proportions of locally available material, but in the assemblage from MD the proportion is relatively high. This variability is also reflected in the flake-tocore ratio (Figure 7.9). Higher values of this ratio reflect increased reduction intensity. The values for quartz are the lowest, indicating that this predominantly local material was flaked the least intensively at all locations. The relatively high value for quartz at S2 reflects the fact that this material is not local to the Stud Creek system. Three of the assemblages (BC, ND, and S2) show generally similar values for clast silcrete. Of these, S2 has the lowest value, consistent with the easily available clast silcrete gibber stones at this location. The value for MD stands out. At this location clast silcrete was heavily worked. The ratio of complete noncortical flakes to those with cortex provides another measure of core reduction intensity (Figure 7.10). As cortical cobbles are more heavily worked, the proportion of flakes that lack cortex will increase, hence increasing the value of the ratio. The uniformly low values for quartz reflect the availability of cortical gibber cobbles manufactured from this material at BC, ND, and MD. The slightly higher values for S2, like the heightened value for the flake-to-core ratio, reflects the nonlocal origin for quartz at Stud Creek. Clast silcrete cortical flakes are relatively common at ND and S2 in comparison to the values at MD and particularly BC. For the flake-to-core ratio, MD shows a value higher than BC; however, this pattern is reversed for the relative proportion of cortical flakes. At BC, where sources for clast silcrete are at the greatest distance of any of the locations, cores were worked in such a way

that a relatively high proportion of flakes do not retain cortex. This pattern is mirrored for nonclast silcrete, where the ratio values for the ND and S2 assemblages are identical. Both locations have low values compared with BC and MD. As might be expected, the ratio of the number of cores that lack cortex to cores that retain cortex parallels the situation with the flakes (Figure 7.11). Values for quartz are uniformly low compared with those for the two forms of silcrete, with the S2 assemblage showing the highest value of the four. For clast silcrete, the values for ND and S2 are low compared with the values for BC and MD. A similar pattern is evident for nonclast silcrete, except that the difference between ND and S2, on the one hand, and BC and MD, on the other, is more pronounced. The technological ratios paint a clear picture of differential raw material use among the four locations. Some of the difference reflects access to raw material. BC, the location the most distant to raw material, has evidence for the most intensive core reduction, whereas S2 seems to show the least, except for quartz. However, the ND and MD locations, which are only a few kilometers apart, have rather different core reduction intensities. At ND core reduction is light, with ratio values close to those from S2. At MD, on the other hand, ratio values are high, approaching those at BC.

Tools

Tools are defined here as artifacts with macroscopic retouch (a typological breakdown is given below). As for the flakes and cores, a series of indexes in the form of ratios helps clarify the relationships among the assemblages. The flake-to-tool ratio measures the proportion of flakes converted into tools (Figure 7.12). Calculating this ratio for each of the raw materials illustrates the importance of raw material source. Locally available quartz consistently has the highest values, reflecting the ease with which this material was available and hence the relatively low levels of retouch (and therefore low ratio values). The silcretes uniformly have lower values than quartz, indicating proportionally more retouched tools, and of the two silcrete types, nonclast silcrete is consistently lower in value than clast silcrete. In

127

Simon Holdaway, Patricia Fanning, and Ed Rhodes

Figure 7.11.  Noncortical core–to–cortical core ratio for each raw material class (clast/clast silcrete, nonclast/nonclast silcrete, quartz). Numbers above the columns give the values of the ratios.

Figure 7.12.  Ratio of complete flakes to complete tools, by raw material class (clast/clast silcrete, nonclast/nonclast silcrete, quartz). Numbers above the columns give the values of the ratios.

part this reflects the operation of one of Elston’s (1990) criteria for raw material economizing: nonclast silcrete is the easiest material to work and as noted above was selected for the manufacture of points and adzes. Besides the patterns related to raw material access and quality, there is, however,

a consistent pattern in the ratio values among the different locations, one, moreover, that reverses the trends seen for the flakes and cores. Unlike the situation with the flakes and cores, where BC and MD versus ND and S2 formed similar pairs, for the flake-to-tool ratio, MD and ND are similar, with

128

Assemblage Accumulation as a Time-Dependent Process

Figure 7.13.  Tool type proportions for each raw material class. BC = Burkes Cave, ND = Nundooka, MD = Mulga Dam, S2 = S2 (Stud Creek). There were insufficient quartz tools found at S2 to allow the calculation of proportions.

low ratio values, and BC and S2 form a pair with relatively high ratio values (i.e., relatively low proportions of tools). The ratio of flakes to tools indicates the degree to which flakes are converted to tools; however, the measure does not specify the form of the tools. Tool forms consist of notched pieces (single notches and multinotch denticulates), scrapers (side scrapers and small end scrapers termed thumbnails), heavily resharpened adzes (Tula with retouch on the distal end and Burrens with retouch on the lateral margins), and points (geometric microliths termed Backed Blades and unifacial points labeled Pirri [see Holdaway and Stern 2004 for definitions]). In addition, there is a class of utilized flakes with very light retouch on the lateral margins. Proportionally, scrapers, utilized pieces, and adzes dominate the tools from each of the assemblages (Figure 7.13). Only at MD (and BC in quartz) do points attain a value of 10 percent of the assemblage, and only at S2 do notch tools in clast silcrete exceed 20 percent of the assemblage. Thus, it is the proportions of the three most common tool categories that, in large measure, determine assemblage tool

composition. Quartz and clast silcrete assemblages tend to be dominated by scrapers (although this is less so for S2). Quartz scraper proportions are highest for MD, followed by BC and ND. For clast silcrete, BC shows the highest proportion of scrapers, followed by ND, MD, and S2. For clast silcrete and quartz, utilized flakes show the opposite trend. Therefore, to some degree, the assemblages differ in the proportion of tools identified as utilized flakes and those with sufficient retouch to be identified as scrapers. For nonclast silcrete the situation is more complex because of the addition of adzes, both the relatively common Tula and the rarer Burren. In this case ND and MD have the highest adze proportions at the expense of the proportion of scrapers. Based on the proportions of tools, the assemblages can be placed in an order that reflects the increasing proportions of high-utility tools. Following this order, S2 has the highest proportions of low-utility­ utilized flakes, negligible thumbnail scrapers, and the lowest proportion of Tula. BC ranks second, with a high proportion of scrapers but a lower proportion of Tula. ND and MD have higher proportions

129

Simon Holdaway, Patricia Fanning, and Ed Rhodes

Figure 7.14.  The ratio of complete scraper surface area (length × width) to complete flake surface area (length × width). The ratio has not been calculated for the small number of S2 (Stud Creek) quartz artifacts.

of adzes and correspondingly lower proportions of utilized flakes. A final comparative measure for tools is used to assess the proportion of the tool blank removed through retouch. This may be done by creating a ratio of the size of the tool surface area (length × width) to the flake surface area. The ratio for scrapers indicates a largely consistent pattern for the three raw materials (Figure 7.14). Quartz scrapers have the lowest values for this ratio, and the values for clast silcrete are the highest. To some degree, this reflects the size of the flakes. Because quartz flakes tend to be manufactured from cores that are smaller in dimension than clast silcrete cores, the flakes produced are also smaller. If these relatively small flakes are retouched, proportionally more of the tool blank surface area will be removed for the same degree of retouch. However, within each raw material class, flake size is less of an issue. Here, the ratio reflects the relative intensity of tool modification. The values for the tool-to-flake surface area ratio may be used to rank the assemblages in a similar way to the tool proportions above. For clast silcrete, ND shows the highest value (i.e., where scrapers

have been retouched the least). The three other assemblages have ratio values that are extremely close. For nonclast silcrete a clearer order is discernable: ND has the highest value, followed by BC, S2, and MD. As was the case with the flakes and cores, some of the variability indicated by the comparative measures for tools reflects raw material access. At BC, where the two silcrete forms had to be brought to the location from the greatest distance, a greater proportion of the clast silcrete tools have been retouched into scrapers, and utilized flakes form a low proportion of the tools. These scrapers also show relatively high levels of retouch compared with the BC flakes (although some of this result reflects the relatively small size of the BC flakes compared with those at the other locations). The tools with the greatest expended utility (following Shott’s 1996 definition) are the Tula adzes, and these tools occur in the largest proportions in the MD and ND assemblages.

Discussion

Flake and tool proportions can be viewed, in accordance with Elston (1990), in terms of a series

130

Assemblage Accumulation as a Time-Dependent Process of economic choices that reflect a combination of raw material availability, mobility, and occupation duration. Decisions made by people in the past were clearly affected by access to raw material. Where material sources are distant (e.g., BC), core reduction, along with the working of tool blanks beyond the stage represented by utilized flakes, is more intense. At S2, where clast silcrete is relatively abundant and nonclast silcrete is obtainable from nearby quarries, cores are less reduced and tools were less often reused. None of this should be surprising—access to raw material is clearly a major determinant of assemblage composition in many parts of the world (e.g., Montet-White and Holen 1991; Odell 1996). Of more interest is the variability introduced through differential mobility and occupation span. Here the comparison between the levels of reduction of locally available versus imported material is instructive. At BC, more intensive core reduction is suggested for the imported silcretes, but levels of quartz reduction are not correspondingly high. Elston (1990) expects local material to be used up when occupation span increases, particularly when sources of imported materials are relatively distant, as is the case at BC. Longer occupation span will also lead to the discard of expended tools that began life with high utility, yet BC shows lower proportions of adzes than either MD or ND. Thus, despite the levels of core reduction evident at BC, the assemblage does not seem to represent intensive, long-span occupation (Shiner et al. 2005, 2007). At the opposite extreme of raw material access is S2. Here clast silcrete is readily available, and the intensity at which this material was worked is low. Nonclast silcrete was brought to S2 from nearby quarries (Holdaway et al. 2004), but the intensity with which this material was worked is similar to that at ND. As argued elsewhere (Shiner et al. 2005), in this sense BC and S2 are similar. Although the assemblages differ, the responses to raw material availability are similar, at opposite extremes of a continuum stretching from raw material paucity to raw material abundance. The assemblages may in fact represent the outcome of a series of occupation spans that have produced similar site histories.

MD and ND are intermediate in distance to raw material and are in fact separated by just a few kilometers. Thus, to some extent, the effect of distance to raw material can be discounted when comparing these assemblages. Both assemblages show increased proportions of higher-utility tools compared with BC and S2, but only MD has corresponding ratio values suggesting more intense core reduction. The flake-to-core ratio at ND, for instance, is low, as is the ratio for the noncortical to cortical flakes and cores. Whereas MD fits with Elston’s (1990) predictions, the more intense core reduction and the increased proportion of highutility tools arguing for prolonged occupation span, ND represents something of an anomaly, to use Gould’s (1978) terminology, with a longoccupation­suite of tools and a short-span core and flake assemblage.

Accounting for Occupation Histories Whereas Gould searched for an ethnographic explanation for the anomalies he found, we turn to the chronologies outlined earlier. In searching for ways to approach medium- to long-term social ­theory, Murray (1997) proposes looking for anomalies that challenge our understanding of the past. All the locations in our study represent palimpsest deposits resulting from multiple occupation spans. If we start from the position that any of the locations could have seen variable types of occupation in the past (rather than an ethnographically inspired position that each location is likely to have seen only one type of occupation), we can begin to analyze the variability that is present. To begin with, the constraints on variability can be documented. All assemblages reflect a similar technology, in the sense that raw material selection, core technology, and tool types are common to all. Following the discussion above, raw material access is also a constraint, in that much of the variability in ratio values among the assemblages seems to reflect distance to source. As is argued elsewhere (Shiner et al. 2005), this constraint is most apparent at Burkes Cave and S2 (Stud Creek), so much so that it is difficult to see past the raw material–inspired variability for these

131

Simon Holdaway, Patricia Fanning, and Ed Rhodes assemblages. Nundooka and Mulga Dam, however, are different. Their radiocarbon chronologies indicate about half the maximum possible site-use history compared with S2 and BC. Both also have hearth chronologies that indicate multiple periods of occupation or multiple occupation spans. However, at MD the technological and tool ratios more or less provide a picture of repeated, relatively long occupation spans, whereas at ND they do not. It might be possible to seek a scenario within a settlement system or ecological framework that would explain the ND result. But the fact that the assemblage does not fit with Elston’s (1990) predictions seems to us best interpreted as the type of historical anomaly that Murray (1997) seeks. ND does not fit because it does not reflect either one site “type” or one constant adaptation. Instead, Indigenous Australians used the location in a variety of ways in the past, ways that have added to assemblage variability. MD may also have been used in different ways, but the variability is such that these differences are less apparent given the measures that we have used. Perhaps in the same way that raw material access dominates the variability at BC and S2, a more prolonged occupation span dominates the variability at MD. For us, the situation at ND strikes at the heart of the nature of explanation in archaeology. If we are in the business of ethnographic reconstruction, then locations like ND are of little use because, by their nature, they depart so dramatically from the “single-component site” or “living floor.” If, however, we are interested in the history of persistent places (after Schlanger 1992), it is the very departure from the single-component site that makes ND valuable. History takes time to unfold. If we could travel back in time and visit an individual occupation at ND, we would learn virtually nothing of significance in determining the history of place use at this location. Go back in time to an instant, and insufficient time will have passed for anything of substance to have happened. Taken to the opposite extreme, this argument could be read as a justification for the exclusive analysis of long-chronology palimpsests, where as much behavior as possible is conflated together,

thereby providing as much historical variability as possible. However, the results presented above indicate that this is no more desirable than the study of living floors. If too much time passes, depending on the context, variability may become constrained, reflecting only a limited number of processes, as argued above for BC and S2 and to some degree MD. For these assemblages only the average outcome of what may in fact be a large number of different behaviors is detectable. Where further differentiation is possible, individual behaviors take on the status of outliers from the mean and therefore will often be ignored as “noise” (see Kuhn 1995 for a similar argument). Our study of the geoarchaeological chronology at different locations in western New South Wales, Australia, casts serious doubt on the assumption, implicit in many current approaches to huntergatherer archaeology, that the archaeological record represents a relatively undistorted sample of a prehistoric settlement system. It also raises questions about assumptions that hunter-gatherer­–​ environmental­ relationships should be characterized as predominantly stable with only a few punctuated periods of change. Instead, the research summarized here paints a picture of a dynamic past, one that affected the creation of the archaeological record in the time before European arrival in the same way it continues to affect it today. With these results in mind, it seems inadequate to approach analysis of the artifact record from the past with a set of techniques that are derived from theories based on the analysis of the synchronic present. If history is a goal, archaeology must be able to deal with deposits that transcend time. Living floor sites are not the most valuable archaeological occurrence because they provide a record of such a limited set of events. To understand the past, variability is the key. The challenge is to analyze the constraints on this variability and understand how these constraints vary in relation to how much time has passed. Murray (1997) remarks that the outcome from the time perspective approach initiated by Bailey (1983) will be a different prehistory compared with present versions predicated on reconstructing lifeways. Schlanger’s (1992) notion

132

Assemblage Accumulation as a Time-Dependent Process of a persistent place provides a model of what that prehistory might look like, at least for our western NSW research.

the Tibooburra Local Aboriginal Lands Council, the Broken Hill Local Aboriginal Lands Council, and the NSW National Parks and Wildlife Service to conduct this study is gratefully acknowledged. Thanks also go to students from La Trobe and Macquarie universities and the University of Auckland for their valuable input of labor. This chapter has benefited from discussions with Geoff Bailey, Angela Close, Tim Murray, Justin Shiner, and LuAnn Wandsnider.

Acknowledgments Fieldwork in western New South Wales (NSW) was funded by Australian Research Council grants and a Macquarie University Research Grant. Permission from the Wangkumara Cultural Heritage Management Committee (traditional owners),

133

8

Time Averaging and the Structure of Late Pleistocene Archaeological Deposits in Southwest Tasmania Nicola Stern La Trobe University Time, and its role in structuring configurations of material remains, has been a part of archaeological discourse since the discipline was established in the mid-nineteenth century. During the past 30 years, however, the time scales established for different portions of the sedimentary record, and their implications for the behavioral information that can be generated from archaeological remains, have come under considerable scrutiny (e.g., Bailey 1983, 1987, 2007; Binford 1978a, 1980, 1983a; Dunnell 1982; Fletcher 1992; Isaac 1972; Lucas 2005; Murray 1997, 2001). As the chapters in this volume demonstrate, archaeologists are far from having a complete understanding of the way in which the material record is influenced by the passage of time or from determining the most appropriate analytical and interpretive strategies for generating behavioral information from agglomerations of material remains accumulated over varying time spans. There is a measure of consensus about the idea that the archaeological record is the product of a hierarchy of processes operating at different time scales (e.g., Bailey 1983, 1987, 2007). There is also a measure of consensus about the suggestion that there is often a mismatch between the time span involved in the accumulation of assemblages of material remains and the time span of the observations on which the ecological and anthropological theories employed in their interpretation are based (e.g., Murray 1993, 1997; Stern 1993). There is, however, no consensus about how this hierarchy of pro-

cesses is structured or about the relationships that exist between processes operating at different scales (Bailey 2007). Nor is there any consensus about how to resolve the interpretive conundrum posed by the apparent mismatch between data and interpretive theories. Some researchers have focused on the problem of identifying an explanatory hierarchy that describes processes operating at different scales and on exploring the interactions between processes at different levels in the hierarchy (e.g., Bailey 1983, 1987, 2007; Fletcher 1992; Gamble 1983). Others have been seeking to establish the relationship between debris-generating activities and their manifestation at different scales of observation and analysis (e.g., Binford 1978a, 1980, 1983a; Stern 2004; Stern et al. 2002). Of course, there is also a group of researchers who remain committed to the notion that rare occurrences of debris accumulated over short time spans lend themselves to the application of existing social and ecological theories and that such fine time lines should be the focus of the discipline’s endeavors (e.g., Conard 1994; Parkington et al. 1992; Roberts and Parfitt 1999). Underpinning these different perspectives is an incomplete understanding of the empirical structure of archaeological data. There are questions that have yet to be answered about the categories of behavioral information preserved in archaeological assemblages accumulated over short versus long time spans. There is also a question about whether the time span and time resolution of archaeological

134

Time Averaging and the Structure of Late Pleistocene Archaeological Deposits data are the primary factors contributing to their perceived singularity as a record of human action. This chapter explores just one of the methodological issues raised by these concerns: the way in which site formation processes and measurements of time impinge on the identification of discrete archaeological assemblages. Ultimately, interpretations of archaeological remains are based on comparisons between assemblages, each of which is the product of a unique combination of circumstances. Frequently, however, the procedures used to make these comparisons undermine current understanding of the essential empirical characteristics of those assemblages. Most (if not all) archaeological assemblages can be characterized as time-averaged agglomerations of debris (Walker and Bambach 1971) in the sense that they comprise the material byproducts of activities that took place on different temporal planes (Stern 1991, 1994a).1 Time averaging comes about because material remains and their encasing sediments accumulate through different processes, often at different rates and over different time spans (e.g., Kidwell 1986; Kowalewski 1996). Of course, the judgment about whether time averaging poses a substantive interpretive problem depends both on the questions being asked and on the amount and type of time averaging involved in the formation of the assemblages under investigation. Information about the mode, rate, and time span of accumulation of both archaeological debris and encasing sediments is fundamental to making that judgment. There has been some acceptance of the suggestion that much of the archaeological record comprises time-averaged agglomerations (e.g., de Lange 2001), but consternation has also been expressed about this characterization of archaeological data, in the mistaken belief that it implies that meaningful behavioral information cannot be generated from these types of data (e.g., Bunn and Kroll 1993). There have been some attempts to demonstrate that time-averaged assemblages do not present any particular interpretive problem for archaeologists, but by and large these actually fail to follow through the logical implications of describing an assemblage as time averaged (e.g., McNabb 1998). This is because time-averaged as-

semblages are treated as though they were an average representation of the activities that contributed to their formation. Some activities and events have a greater impact on surviving material remains than do others. Consequently, the mixing of debris from temporally unrelated activities (time averaging) is not the same as creating an assemblage of material remains in which the objects occur in the same proportion as the frequencies with which debrisgenerating activities were undertaken (an average representation). One of the difficulties of following through the logical implications of studying a timeaveraged assemblage is that established analytical and interpretive procedures were not designed to deal with data characterized in this way. This chapter explores a strategy for defining and comparing archaeological assemblages that takes into account their time-averaged status and which offers the prospect for investigating the behavioral information encapsulated in assemblages representing different time spans and resolutions. The springboard for discussion is provided by two late Pleistocene archaeological deposits created through similar processes but accumulated at different rates and over different time intervals. Apart from occasional hearths preserved at depositional lacunae, the processes involved in the accumulation of both deposits preclude the identification of sets of material remains that might arguably relate to individual events or occupational episodes. Thus, the analytical strategy is geared toward the agglomerations of debris that characterize much of the archaeological record, rather than “high-resolution” records created by rare depositional circumstances. The two sites that provide the basis for this discussion, Mackintosh shelter and Nunamira Cave, are located in southwest Tasmania where a fold-structured geology has created a distinctive topography of rugged mountains and north– south-trending valleys, bounded on its northern edge by a series of high, contorted ranges (Figure 8.1). At present, the area is vegetated by wet, temperate rain forest, and the region appears to have been uninhabited when Europeans colonized Tasmania in the early nineteenth century. However, during the late Pleistocene, when average annual temperatures were depressed by approximately 5ºC, 135

Nicola Stern

Figure 8.1.  Mackintosh shelter and Nunamira Cave, located in southwest Tasmania, Australia.

this landscape was covered by alpine and subalpine vegetation communities (Colhoun et al. 1994). With the onset of full glacial conditions, there was further depression of the tree line and expansion of heath and herb communities. One of the distinctive features of these heath and herb communities were small patches of Poa grasslands that grew on the more fertile alluvial soils in the valley bottoms and on areas of limestone bedrock (Colhoun et al. 1994; Cosgrove 1999). These patches of Poa grassland appear to have supported dense populations of wallabies and wombats, which were exploited as a source of meat and marrow for communities of hunter-gatherers. At the end of the last ice age, however, the recolonizing rain forest engulfed these patches of Poa grassland, and both humans and their prey largely disappeared from the landscape (Cosgrove and Allen 2002). The earliest evidence for human presence in this landscape dates to about 35,000 years ago (Cos­grove 1999), and abundant traces of the activities of local hunter-gatherers continued to accumulate until the rain forest was reestablished between 13,000 and 11,000 bp (Cosgrove 1999; Cosgrove and Allen 2002). Evidence for the activities of these late Pleistocene populations is found primarily in limestone caves, shelters, and collapsed dolines, in part because these settings were effective sediment traps, conducive to the preservation of the material

remains, and in part because of their visibility in heavily forested terrain. Mackintosh 90/1 and Nunamira are two of about 20 sites excavated in this region that preserve material traces of late Pleistocene human activities and are two of the eight sites excavated by members of the Southern Forests Archaeological Project (Allen 1996). They provide a useful springboard for the present discussion because despite differences in the textural characteristics of the encasing sediments, the processes affecting the accumulation and preservation of archaeological material at each site were similar, and yet the two deposits accumulated at different rates and over different time spans. Mackintosh is a small shelter in a bluff of dolomitic limestone, located in a narrow, steep-sided valley on the northern margins of the Tasmanian Southwest (Figure 8.2). Although this valley now forms the eastern arm of an artificial impoundment, the Brougham River, which drains the mountains on the edge of Tasmania’s central plateau, it once followed a well-defined boundary between Ordovician quartzarenites and limestones before disgorging into the Mackintosh River, 200 m downstream from the site. The Mackintosh shelter itself lies about 17 m above the original floor of the Brougham Valley, at an altitude of 230 m, currently in the zone between low and high lake ­levels.2 Thirteen 50-×-50-cm squares straddling the livable

136

Time Averaging and the Structure of Late Pleistocene Archaeological Deposits

Figure 8.2.  Mackintosh shelter, located in a narrow, steep-sided valley on the northern margins of the Tasmanian Southwest, showing area excavated.

area of the shelter were excavated, revealing an archaeological deposit with a maximum thickness of 60 cm. Radiocarbon determinations show that most of this organic-rich horizon accumulated between approximately 17,000 and 15,000 years ago (see ­below). Nunamira is a small cave that lies at the base of a large bluff of dolomitic limestone along the middle reaches of the Florentine Valley, on the eastern margin of the Tasmanian Southwest (Cosgrove 1996:​43–47). The present cave floor lies about 8.5 m above the Florentine River, at an altitude of 240 m (Cosgrove 1996:46–47). The site was excavated by Richard Cosgrove (1996), who dug two 1-×-1-m squares, one in each of the cave’s two, unequal-sized chambers (Figure 8.3). Excavation revealed a stratigraphic sequence with a maximum thickness of almost 80 cm, and a series of radiocarbon determinations showed that the archaeology-bearing horizons accumulated from about 30,000 to 11,000 years ago (Cosgrove 1996:62). Investigation of the Mackintosh deposits was designed specifically to generate information about the mode and rate of accumulation of both the archaeological debris and their encasing sediments. This was deemed critical to establishing the scale and resolution of the archaeological materials, identifying the strategies most appropriately

employed in intersite comparisons, and developing interpretive frameworks commensurable with the empirical characteristics of this particular archaeological record. For this reason, the Mackintosh deposit, and the methodology used in its investigation, is described in some detail. This is followed by a comparison with Cosgrove’s (1991, 1996) descriptions of the Nunamira deposit.

The Mackintosh Sequence The macrostratigraphic features identified during the excavation of the Mackintosh shelter not only guided the removal of the deposits but also provided an initial description of the stratigraphic sequence and determined the locations of the first set of charcoal samples submitted for radiocarbon determinations (Stern and Marshall 1993). At the end of the excavation micromorphological samples were collected from the exposed sections, so thinsection analysis provides an additional source of information about the constituents of the sediments, their mode of deposition, and the effects of postdepositional processes on the accumulated sediments and archaeological debris (Porch 1992:59–79). The micromorphological analysis also helped to identify critical gaps in the radiocarbon chronology and to determine the locations of an additional set of samples.

137

Nicola Stern

Figure 8.3.  Nunamira Cave, showing area excavated (after Cosgrove 1996:Figure 6).

Analysis of the coarser components of the sediments generated additional information about the processes that contributed to the accumulation of the sediments and the archaeological debris. All the cultural and noncultural debris retained in the 3-mm sieves were used for this purpose: not only do they contain the full range of coarser materials making up the deposits, but they provide large enough samples to facilitate quantitative analysis. The complete set of radiocarbon dates, together with the analysis of the sieve residues, provided the information needed to interpret the macrostratigraphic features identified in the field. Consequently, the description of the stratigraphic sequence at the Mackintosh shelter presented here supersedes an earlier one based on macrostratigraphic features alone (Stern and Marshall 1993). The archaeological deposit at the Mackintosh shelter is sandwiched between a thick sequence of clay-rich sands of karstic origin and a thin, compacted jumble of archaeological debris and sterile sands derived from a late-nineteenth-century prospector’s trench, which runs along one wall of the

shelter (Stern and Marshall 1993; Figure 8.2). Each of these represents a distinctive sedimentary envelope: not only does each exhibit different characteristics reflecting a distinctive mode of accumulation, but each is separated from the underlying and overlying horizons by an identifiable depositional hiatus (Figure 8.4). The archaeological horizon actually comprises two quite distinct bodies of sediment; however, postdepositional calcium carbonate impregnation has obscured the depositional break that separates them. This stratigraphic hiatus was identified only by integrating information derived from the micromorphological analysis with the results of the residue analysis and the radiocarbon determinations (Porch 1992:71–73). The distinctive features of each sedimentary envelope provide a ­basis for discussing the processes involved in their formation. The lowermost deposit (Horizon I) consists of at least 1.7 m of water-lain clays and sands of unknown age, although its actual thickness has not been determined because it extends below the present floor of the prospector’s trench. This deposit

138

Time Averaging and the Structure of Late Pleistocene Archaeological Deposits

Figure 8.4.  Archaeological deposit at Mackintosh shelter.

contains no cultural material of any kind, and its surface clearly formed the floor of the shelter when archaeological debris first started to accumulate. The sections exposed in the walls of the prospector’s trench show a marked gradation from a lithology unaffected by pedogenic processes to one that exhibits all the hallmarks of a lower soil horizon (Porch 1992:​60–63). Horizon II is a homogeneous dark reddish brown, clayey and silty, fine to medium sand that contains abundant archaeological debris. It varies in thickness across the shelter, filling a 40-cmdeep pit toward the center of the livable floor area but elsewhere rarely more than 15–25 cm thick; it lenses out toward a pile of roof fall that marks the drip line. It is made up primarily of quartz sands, clay nodules, oxidized clay-rich sediments derived from hearths, uncarbonized plant material, burned and fragmented bone, and chipped-stone artifacts. It also contains some soil and gravel particles derived from outside the shelter that were carried in on people’s feet or with the organic materials they carried into the site. The sand and clay components of this deposit were derived from within the shelter itself: the

mineralogy of the sand and silt-grade particles are identical to those of the underlying sterile sands (Horizon I), and the clay nodules themselves exhibit microstructural features consistent with an origin in the underlying deposit (Porch 1992:63–70). In addition, there is a notable decrease in the relative abundance of these clay nodules with increasing distance from the stratigraphic contact between Horizons I and II. Upward reworking of the sands and clays from Horizon I explains the gradational contact between them. The contact surface is rather hummocky and dissected, resulting from lateral variation in the texture of Horizon I and consequent variation in the extent to which its surface was disrupted by human activity. Hard, compacted clays are exposed as hummocks, and the pockets of looser, sandier sediments that separated them have been reworked into Horizon II (Porch 1992:​60). Upward reworking of the clays and sands from Horizon I undoubtedly came about through a combination of trampling, pit digging, and burrowing. The base of this archaeological horizon is characterized by some distinctive, reddened sediment and by an abundance of oxidized clay nodules, which decrease in frequency up-section. These

139

Nicola Stern r­ epresent the remains of hearths placed onto the surface of Horizon I, mostly in the central portion of the shelter. As the archaeological deposit accumulated, upward reworking of sediments continued, resulting in homogenization of its constituents. The micromorphological data and the residue analysis provide some insight into the impact of this process during repeated occupation: up-section the sediments are more crumbly and microstructures are better developed; in addition, the abundance of fine material decreases, the abundance of clay nodules decreases, they are smaller in size, and a greater proportion of them are oxidized (Porch 1992:​63–70). Four relatively undisturbed hearths are preserved at the top of this horizon, indicating a cessation of human occupation for some time before the overlying sediments were laid down. Intensive homogenization of the accumulated archaeological deposit thus ceased. Following the deposition of the overlying carbonates (see below), the upper levels of Horizon II were subject to calcium carbonate impregnation. This is reflected in both the micromorphological data and the residue analysis, both of which show a significant decrease in the amount of carbonate down the Horizon II profile. Horizon III consists primarily of carbonates that accumulated on the surface of Horizon II following a significant depositional hiatus, the duration of which cannot be determined from existing radiocarbon determinations (see below). As these carbonates form under warmer and more moist conditions than those experienced during the accumulation of Horizon II, they must have been deposited during the late Pleistocene or Holocene (at other sites in Tasmania such moonmilk layers are known to have formed within the last 11,000 years). These carbonates impregnated the top of Horizon II, obscuring the depositional hiatus between the two sedimentary layers and creating a gradational contact between them. Horizon III as a whole contains large quantities of small mammal bone, as well as some discrete lenses of reddened sediment whose color comes from scat-derived apatites (Porch 1992:74). The suite of animals contained in these particular carbonate lenses suggests

that either Dasyurus maculatus or Sarchophilus harisii (both small marsupial carnivores) deposited these scats (Wines 1996:50–59). Micromorphological data, the residue analysis, and the faunal analysis all point to mixing of sediments and archaeological debris across the depositional hiatus that marks the boundary between Horizons III and II. By far the greatest proportion of the bones recovered from Horizon II represent large mammals; these have been stained a pale or dark brown color by the organic matter contained in those sediments, and their surfaces are coated in silt or clay. In contrast, most of the bones recovered from Horizon III are from small mammals; these bones are pale in color, and some (but not all) have a coating of calcium carbonate (Wines 1996:75–79). Most of the excavation units straddling the boundary between Horizons III and II contain bones exhibiting both sets of characteristics, suggesting mixing of debris of different age and origin. Other coarse constituents of the sediments also provide evidence for the mixing of debris across this depositional hiatus. For example, Horizon III contains large quantities of dolomite pebbles derived from the roof of the shelter, but the frequency of these decreases markedly down-section. In contrast, the clay nodules (which originated in Horizon I) decrease in abundance across this depositional hiatus, suggesting upward reworking from Horizon II. Micromorphological samples from Horizon III contain quartz grains and clay nodules that are derived from Horizon II (and originally, from Horizon I) and which formed the nuclei around which the micrite was deposited. The micromorphological samples from Horizon III also provide evidence for the formation, disruption, and reimpregnation of micritic nodules, indicating that homogenizing processes continued during the accumulation of Horizon III. Two burrows identified during the excavation of this horizon identify at least one of the mechanisms involved. Occasional habitation of the site during the accumulation of Horizon III is indicated by the presence of two intact hearths that abut, and are overlain by, carbonate deposits and by the ­quantities of

140

Time Averaging and the Structure of Late Pleistocene Archaeological Deposits chipped-stone artifacts recovered from this part of the sequence.3 Although many of the bones from medium and large-sized marsupials in Horizon II were reworked up-section from Horizon II (as just argued), there are two specific area locales where the bones of human prey animals lie well above the Horizon II–III boundary (Wines 1996:​77). These particular bones appear to have been deposited during the accumulation of Horizon III. Horizon IV is a mottled, yellowish brown, clayrich silt containing a heterogeneous mixture of sediments and archaeological debris that represents the deposition and compaction of the deposits dug out from a trench running along one wall of the shelter (Stern and Marshall 1993). Prospectors were particularly active in this area during the 1890s, and there is some evidence to suggest that Europeans lived in the shelter for a brief period of time. This includes two hearths and two pieces of wood cut by a steel axe, all of which lie at the interface between Horizons III and IV, as well as localized disturbance to a distinctive lens of carbonate that marks the top of Horizon III across most of the excavated area. When humans first made use of the Mackintosh shelter they would have encountered a floor made up of ancient sands and clays. Their activities resulted in disruption of the cave floor and the introduction of a variety of organic and inorganic debris. The archaeological horizon built up through a combination of the upward reworking of underlying sands and clays and the accumulation of organic matter, stone artifacts, and gravels derived from outside the shelter. Thus, both sediments and archaeological debris were homogenized as they accumulated, so neither individual hearths nor depositional lacunae have been preserved within this sedimentary envelope, only at its surface. A break in deposition separates the main archaeological horizon from the overlying carbonate horizon. Both depositional (trampling) and postdepositional processes (carbonate impregnation, burrowing by small mammals), however, have obscured this depositional hiatus. There is evidence for at least occasional human activity within the shelter during the accumulation of the carbonate horizon. Another depositional hiatus separates the carbon-

ate horizon from the thin wedge of heterogeneous sediments and debris spread across the floor of the shelter by late-nineteenth-century prospectors.

Mackintosh Chronology

The processes involved in the accumulation of the Mackintosh deposits obviously have considerable implications for the way in which the radiocarbon determinations are interpreted. Homogenization of the archaeological and carbonate horizons as they built up means that all their coarse constituents, including charcoal, have been effectively time averaged. Furthermore, there was undoubtedly considerable lateral variation in the quantities of organic material introduced and the amount of trampling and homogenization that took place. Consequently, radiocarbon determinations could not be expected to show any consistency across the site, although they might be expected to reveal a general trend of decreasing age up-section. On the other hand, a consistent set of age determinations could be expected for a single location within the shelter (e.g., within a single excavation square). Not surprisingly, this is the pattern shown by the radiocarbon determinations obtained (Table 8.1). All 17 determinations are based on charcoal recovered from the archaeological deposits, mostly from Horizon II. The radiocarbon determinations for the base of Horizon II are variable. A single determination of 20,880 ± 390 (Beta-54740) raises the possibility of pre–Last Glacial Maximum occupation of the site. It is also conceivable, however, that this age determination reflects the incorporation of charcoal from an older source into the archaeological deposit, especially as this sample was derived from the zone of mottled Horizon I/II sediments at the very base of the occupation deposit.4 Probably the most reliable determination for the base of Horizon II comes from within the pit, as the fill from the pit, once deposited, would have been protected from further reworking. Thus, a determination of 17,039 ± 390 (Beta-54740), from the base of Horizon II within the pit, is a reasonable estimate for the time at which archaeological debris started to accumulate in this shelter.

141

Nicola Stern Table 8.1.  Radiocarbon Determinations for the Archaeology-Bearing Deposits at the Mackintosh Shelter Horizon Description

Square and Spit Number

Stratigraphic Unit

III III III II II II II II III II II II II II II II II

FN 5, 8, 9 HM 10–13 GL 7, 9, 10 EN 7; FN 14 GL 19, 21, 23 HM 27–29 GP 9, 10 HN 9 FN 11 FN 15 HN 10, 13 HN 18 GP 13, 14, 15 FN 21, 22 HM 22 FN 36, 37 HN 20, 21, 22

Scat deposit Pale ashy Pale ashy Top of 3 Top of 2 2 Top of 4 Top of 4 Top of 4 Base of 3 Base of 4 Top of 2 Base of 2 2 (pit) Base of 2 Base of 2 (pit) Base of 2

Late Pleistocene carbonates Late Pleistocene carbonates Late Pleistocene carbonates Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon Main archaeological horizon

It is difficult to determine precisely when Horizon II ceased to accumulate, the length of the ­depositional hiatus that separated Horizon II and III, or when the Horizon III carbonates started to form. This is a consequence of the mixing of sedimentary particles (including charcoal) across the Horizon II–III boundary. In particular, radiocarbon determinations for the top of Horizon II (14,950 ± 200 [Beta-46304]) and the top of Horizon III (15,570 ± 180 [Beta-45805]), which are based on charcoal samples recovered from the middle of the livable floor area, show some inconsistency. However, the remaining dates for the top of Horizon II overlap at one standard deviation and suggest that it ceased to accumulate about 15,000 years ago. Charcoal samples from the base of Horizon III in other parts of the shelter yield age determinations of 10,440 ± 170 (Beta-54738) and 12,000 ± 200 (Beta-54737). The distinctive lens of apatite-stained carbonate that lies toward the top of the carbonate sequence provides an age determination of 4910 ± 160 (Beta-​46302). Taken together the age determinations for Horizon III suggest lateral variation in the accumulation and homogenization of the deposit, possibly as a consequence

Laboratory Number

C14 Years bp ± 1 S.D.

Beta-46302 Beta-54738 Beta-54737 Beta-46303 Beta-46304 Beta-54739 Beta-46305 Beta-45813 Beta-45805 Beta-45807 Beta-45814 Beta-45815 Beta-46306 Beta-54736 Beta-45612 Beta-45808 Beta-54740

4910 ± 160 10,440 ± 170 12,000 ± 200 14,820 ± 140 14,950 ± 200 15,040 ± 280 15,160 ± 270 15,560 ± 200 15,570 ± 180 15,620 ± 190 15,720 ± 200 15,730 ± 170 16,010 ± 300 16,200 ± 120 16,500 ± 190 17,030 ± 430 20,880 ± 390

of intermittent use of the shelter during the late ­Pleistocene–​Holocene.

Summary

In contrast to many of the deposits found in Middle Eastern and European caves and shelters, the terrigenous component of the Mackintosh deposit not only derived from within the shelter itself but also accumulated primarily because of the human activities that took place there. The only exceptions to this are the carbonates and scats that contributed to Horizon III. Consequently, lenses of sterile sediment do not separate the debris resulting from different episodes of human activity, and each occupational episode resulted in the mixing of the deposits that had already accumulated. The three distinct archaeological stratigraphic units just described could only be identified because there were some depositional hiatuses and some changes in the mode of sediment accumulation over time. The fill from the prospector’s trench was readily identified because it lies unconformably on the Holocene carbonates and represents a distinctly different mode of sediment accumulation. The boundary between the late Pleistocene/Holocene carbonates and the

142

Time Averaging and the Structure of Late Pleistocene Archaeological Deposits

Figure 8.5.  Stratigraphic sequence at Nunamira Cave (after Cosgrove 1996:Figure 4).

late Pleistocene archaeological horizon was more difficult to identify, partly because it represents a subtle shift in depositional mode and partly because the depositional hiatus that separates them has been obscured. This sedimentary boundary was determined in part by a climatic shift that initiated carbonate formation and in part by a change in the frequency or duration of human visitation, which resulted in less disturbance of the shelter floor. So it could be argued that Mackintosh contains two minimum archaeological stratigraphic units, one that accumulated between approximately 17,000 and 15,000 years ago and a second that accumulated intermittently over the next 10,000 years or so. There is no logical basis for separating out any of the archaeological debris contained within either sedimentary envelope from the rest of the assemblage; to do so would belie the time-averaged characteristics of each assemblage.

Nunamira The 50-cm-thick archaeology-bearing horizon at Nunamira Cave is sandwiched between a thin crust of flowstone (or moonmilk) formed during the past

11,000 years and sterile, silty clays lying on top of a bed of river cobbles (Cosgrove 1996:54–58). The archaeological deposit is quite homogeneous, except for its upper levels, which have been affected by calcium carbonate impregnation, and the lower levels, where there is evidence for mixing with the underlying clays. These are features it shares with the Last Glacial Maximum deposits at Mackintosh 90/1. Although the stratigraphic sequence at Nuna­ mira was divided into nine units during excavation (Cosgrove 1996:54–58), these make up two distinct sedimentary envelopes, each separated by a depositional hiatus and each characterized by a different mode of deposition (Figure 8.5). The lower­most of these is a set of fluvial deposits of unknown age that do not contain any archaeological remains (Cosgrove’s units 8–6). The overlying deposit is a bed of clayey, sandy silts containing abundant organic material (including animal bone) and stone artifacts (Cosgrove’s units 5–2). Some intact hearths are preserved at the top of this deposit (Cosgrove’s unit 3), which is sealed by a thin layer of carbonate (Cosgrove’s unit 1). The contact between the fluvial ­deposit and the archaeological deposit is a

143

Nicola Stern g­ radational one, apparently reflecting both the upward reworking of the sterile clays into the accumulating archaeological deposit and the trampling of archaeological debris into the sterile clays making up the cave floor (Cosgrove’s unit 5). These processes have obscured the depositional hiatus that separates these two distinct bodies of sediment. The fluvial deposit consists of a cobble bed, which lies directly on the limestone bedrock, and a homogeneous, yellow, silty clay that lies on top of the cobble bed (Figure 8.5). Similarities in the size, shape, and lithology of these cobbles and those lining the present channel of the Florentine River led Cosgrove (1991:72) to suggest that they were laid down by the Florentine before the river established its present course. As the silty clay lies conformably on the cobble bed and comprises a mixture of quartz particles and clays, and none of its minor constituents originated within the cave (­McConnell 1990:6), its deposition almost certainly reflects lateral channel migration or an altered flow regime. The archaeological deposit itself is a homogeneous layer of sandy, clayey silts, with quartz making up most of the silt and sand-sized particles. It also contains charcoal, finely fragmented bone, traces of carbonized plant material, a suite of minerals derived from the limestone bedrock, and rock fragments derived from outside the shelter (McConnell 1990:7). Abundant stone artifacts, vertebrate faunal remains, and lumps of charcoal were recovered during the excavation of this deposit (Cosgrove 1991:​64–66). The macrostratigraphy at Nunamira led to the identification of four distinct units within this archaeological deposit, and although there are clear differences in the composition and fabric of these units, most of them are the result of postdepositional processes (McConnell 1990:7). The one exception is the “transitional” layer that lies between the sterile yellow, fluvial clays and the homogeneous, organic-rich archaeological deposit. This is a layer of organic-rich silty clay, which contains some finely fragmented bone, charcoal, and other organic material, as well as some small stone artifacts (Cosgrove 1996:56, 61). Micromorphological data show that it contains pockets of the dense, quartz-rich clays from the underlying fluvial

deposits, as well as soil particles (Cosgrove 1996:59, 61). When humans first occupied Nunamira, the quartz-rich, yellow clays of fluvial origin formed the floor of the shelter. As at the Mackintosh shelter, human activities resulted in both upward reworking of the underlying clay and the trampling of soil particles and debris into the disrupted surface of those clays. This process of homogenization took place through at least a 20-cm thickness of deposit (the maximum observed thickness of this unit). It obscured the contact between two sedimentary horizons that were quite distinct in terms of their mode of deposition and whose accumulation was probably separated by tens of thousands of years. The blurring of this depositional hiatus was a consequence of the processes that contributed to the accumulation of the archaeological deposit. The rest of the archaeological deposit is best described as a single envelope of sediment (units 4 and 2) containing a structurally well-preserved hearth (unit 3 [McConnell 1990:4]) and exhibiting calcium carbonate impregnation in the upper few centimeters. Most of this sedimentary envelope (unit 4) is a homogeneous mixture of sand- and silt-sized quartz grains, charcoal, and noncarbonized plant remains embedded in a clay matrix that also contains finely fragmented plant material. The dominant component of the sediments is quartz, and the only obvious source for these grains is the underlying sterile clays. Weathering of the shelter roof and walls contributed only trace amounts of detrital limestone to the accumulating archaeological deposit (McConnell 1990:8). The relative abundance of introduced soil particles and nodules increases up-section through the archaeological deposit, whereas the relative abundance of clay nodules decreases (Cosgrove 1996:​ 59). This undoubtedly reflects the decreased potential for dilution of the introduced debris and organic material by the underlying clays as the deposit accumulated. It is also likely that the thickness of the layer subject to homogenization decreased as the organic-rich silts accumulated. There is evidence for increasing degradation of organic material (both uncarbonized plant material and bone) with increasing depth, which has been interpreted as reflecting a gradual accumulation of the deposit

144

Time Averaging and the Structure of Late Pleistocene Archaeological Deposits over time (McConnell 1990:5, 7). A slow net rate of sediment accumulation (resulting from episodic deposition), however, could have produced the same effect. Numerous hearth features lie toward the top of the archaeological deposit and were identified as discrete patches of charcoal-rich soil during excavation (Cosgrove 1996:57). Micromorphological examination of one of these revealed a structurally well-preserved hearth, containing abundant organic matter, lying on top of a layer of baked sediments. This is an unusual feature in a site in which the accumulation of the deposit involved a process of homogenization. The preservation of this and the other hearths indicates rapid burial, perhaps following a single visit to the site (McConnell 1990:​ 7), and subsequent occupational hiatus. Like the proverbial set of refitting flakes, they represent an identifiable episode of human activity in an other­ wise time-averaged archaeological deposit. The top few centimeters of the archaeological deposit differ in color and composition from the underlying accumulation. Like the rest of the archaeological deposit this layer is a clay-rich sand and silt composed mostly of quartz grains, but it also exhibits extensive calcium carbonate cementation. Micrite occurs as small cementing nodules and as horizontal bands of cement, and microsparitic and micritic calcite coat the voids within areas of cementation (McConnell 1990:3). As the archaeological deposit is sealed by a thin layer of moonmilk (unit 1 [Cosgrove 1996:58]), the calcite cementation is clearly postdepositional in origin and must have followed the formation of the capping flowstone. The flowstone that seals the archaeological deposit is only 2–10 cm thick and contains tens of thousands of bones of small mammals, all of which are well preserved with pale, clean surfaces. Some of these form discrete bundles, and some bones are partially articulated (Cosgrove 1996:58). The moonmilk accumulated after human occupation of the shelter ceased. The presence of some calciumencrusted artifacts on the surface of this layer, together with a small number of artifacts within the deposit, suggests at least rare visits to the site. However, the fact that the moonmilk has not been dis-

rupted (as it was at the Mackintosh shelter) suggests that these visits were brief or widely spaced in time.

Nunamira Chronology

Fourteen radiocarbon determinations were obtained for the Nunamira sequence, and most of these are based on charcoal obtained from hearths or from hearthlike features. Eleven of these determinations come from one of the excavation squares, and together they provide a measure of sediment buildup over time. As is the case for Mackintosh 90/1, an understanding of the site formation processes is fundamental to the interpretation of the radiocarbon determinations. The Nunamira deposit, like the Mackintosh deposit, was subject to a process of homogenization as the sediments were accumulating, a consequence of both the upward reworking of the underlying clays and the downward trampling of organic and inorganic debris. At any given time, the upper level of the accumulated sediments was being turned over, and the charcoal contained in these sediments would have been subject to time averaging. This leads to the prediction that the age determinations should show a tight relationship between age and depth. This indeed is the pattern described by Cosgrove (1996:61–64). Cosgrove (1996:61) identifies one apparently anomalous date, 23,640 ± 310 bp (Beta-25382), based on a bulk sample from near the base of unit 4. He argues cogently that it is not as reliable as a second sample taken from the base of the sequence, which yields an age determination of 30,420 ± 690 bp (Beta-​25881). Once the anomalous date is set aside, a consistent sequence of age determinations is evident. Cosgrove (1996:​62) identifies a tight relationship between age and depth and argues that it reflects the steady buildup of sediments through time. The gradual accumulation of sediments was the rationale for dividing the homogeneous envelope of archaeological sediments into smaller units, each of which represents a comparable time span (Cosgrove 1996:​63–64). However, in view of the fact that the accumulation of the Nunamira deposit was the result of human activities and that it involved both upward reworking and downward trampling of the deposit, this is not the only possible interpretation of these

145

Nicola Stern data. Episodic sediment accumulation combined with homogenization of the upper levels of the deposit as it was accumulating would have the same effect as a steady rate of sediment buildup. The net rate of sediment accumulation ranges from 2 to 2.5 cm/1,000 years, so that even if the zone of homogenization were restricted only to the upper 5–10 cm, subsequent visits to the site could have been turning over 2,500–5,000 years of accumulated debris at any one time. Furthermore, there was almost certainly lateral variation in the effects of trampling and in the way that the organic and inorganic debris of human activity accumulated. This might be reflected in additional radiocarbon determinations from other parts of the cave. Given the textural differences between the substrates at Nunamira and Mackintosh, however, there is no doubt that the potential for homogenization was considerably less at Nunamira than it was at Mackintosh. Using depositional hiatuses and changes in sediment source or mode of deposition to identify the boundaries of distinctive sedimentary envelopes, it is evident that Nunamira contains a single archaeological stratigraphic unit. With the exception of the intact hearths (unit 3) that are preserved at the top of this sedimentary envelope, any subdivision of it is arbitrary. It thus represents an agglomeration of debris that accrued over a 20,000-year time span.

Discussion Analysis of the Mackintosh and Nunamira deposits was based on the strategy of identifying discrete bodies of sediment, each representing the operation of a distinct depositional regime over a definable time span. Sedimentary envelopes were distinguished from each other by the existence of depositional hiatuses and by macroscopic and microscopic evidence for shifts in sediment sources and modes of deposition. Additional information about depositional and postdepositional processes was provided by radiocarbon determinations and by the coarse components of the sediments (including archaeological and noncultural debris). This provided crucial insights into the relationship between the accumulation of the archaeological debris and their encasing sediments. In the absence of discrete clusters of debris that arguably represent

individual events or activities, each sedimentary envelope thus contains an agglomeration of debris accumulated and preserved in similar circumstances, whose time span of accumulation can be determined. In these cave deposits this is the equivalent of the minimum archaeological stratigraphic unit used to identify the distribution of archaeological debris across a paleolandscape (Stern 1991:102–​ 103). In other words, it is the smallest envelope of sediment, and therefore the smallest unit of time, that can be used to identify a homogeneous assemblage of material remains, and it provides a way of analyzing and interpreting those remains that takes into account their time-averaged status. In many archaeological contexts the debris contained within a sedimentary envelope can be shown to have accumulated at a different rate and over a different time span than the encasing sediments. The Nunamira and Mackintosh deposits are unusual because the accumulation of the encasing sediments was dependent on the activities that generated the archaeological debris. This is because the primary source of the clastic sediments containing the organic material and artifacts was derived from underlying, sterile sediments. In both cases the homogenization of sediments and their contents continued throughout the accumulation of the deposits (although the extent of homogenization at Nunamira was considerably less than it was at Mackintosh). So in each of these particular instances site formation processes reveal that the sedimentary envelopes and the archaeological assemblage they contain represent the same time interval. Thus, the extent of time averaging can be determined by establishing the time span of the sedimentary envelopes containing that archaeological debris. With the exception of the intact hearths preserved at depositional hiatuses, the archaeological assemblages recovered from both Nunamira and Mackintosh represent agglomerations of unrelated activities. So the late Pleistocene deposits at Mackintosh and Nunamira present two contrasting archaeological stratigraphic units: one preserves a 15- to 25-cmthick agglomeration of debris accumulated over approximately 2,000 years, and the other, a 50-cmthick agglomeration of debris accumulated over

146

Time Averaging and the Structure of Late Pleistocene Archaeological Deposits a­ lmost 20,000 years. Normal archaeological practice suggests that intersite comparisons should be based on a portion of each sequence that accumulated during approximately the same time interval and, preferably, as short a time interval as possible (in the belief that shorter time intervals preserve less complicated behavioral signals). As argued earlier, however, any arbitrary subdivision of these archaeological stratigraphic units involves treating a time-averaged assemblage as though it were an average representation of past events. Mackintosh and Nunamira thus present an opportunity to explore whether archaeological assemblages representing different amounts of time averaging actually preserve qualitatively different categories of behavioral information. Clearly, there is a relationship between time and the formation of an archaeological assemblage: as more time elapses, the greater the probability that an assemblage will contain debris relating to rare activities or events (e.g., Shott 1998). This, of course, suggests that a focus on “fine time lines” in the belief that they approximate ethnographic time frames and preserve uncomplicated behavioral records is a misguided strategy. However, it is not known how much time must have passed to capture the material residues of infrequent activities such that the composition and characteristics of an assemblage stabilize. Undoubtedly, this will vary from one context to another. It is conceivable, for example, that Mackintosh and Nunamira preserve similar categories of behavioral information despite the order of magnitude difference in the time span involved in the accumulation of each assemblage. Cosgrove (1999) has commented previously on the extraordinary similarity of the artifact and faunal assemblage through the Nunamira sequence, and though this may reflect the impact of time averaging, it may, to some extent, also reflect aspects of the empirical structure of these cave assemblages. Mackintosh and Nunamira seem to have considerable potential for an exploration of the behavioral information encapsulated in material assemblages representing different time spans and temporal resolution. There are remarkable similarities in the mode of accumulation of the archaeological deposits, they are located in the same bio-

geographic zone, and both assemblages exhibit the distinctive features noted previously for late Pleistocene archaeological assemblages in this region (Cosgrove et al. 1990; Wines 1996). It is anticipated that a comparison of the behavioral information encapsulated by these two assemblages will help to clarify at least some aspects of the hierarchy of activities, processes, and trajectories that are archaeology’s unique purview. It has long been recognized that the archaeological record is the outcome of a complex interplay among debris-generating activities, the geomorphic processes that result in net sediment accumulation, and the time span over which a particular depositional regime prevailed. The logical consequences of this observation, however, are rarely followed through, largely because they challenge not only accepted analytical and interpretive procedures but a long-standing and comfortable consensus about the behavioral information that can be generated about the past (Murray 2001). Unconventional ways of grappling with the archaeological data, combined with ongoing discussion about the ontological significance of those data, are needed if archaeology is to achieve the unique insights into human action that time perspectivism suggests are possible.

Acknowledgments The fieldwork on which this chapter is based was undertaken under the auspices of the Southern Forests Archaeological Project, and I thank Jim Allen and Richard Cosgrove for the invitation to participate in this research. The Australian Research Council funded the excavations at the Mackintosh shelter, and the laboratory research was funded through grants from the Faculty of Humanities and the deputy vice chancellor at La Trobe University. The Australian Institute of Aboriginal and Torres Strait Islander Studies funded Richard Cosgrove’s excavations at Nunamira. Enthusiastic groups of students from La Trobe University participated in the excavations at both Nunamira and Mackintosh. Nick Porch undertook the micromorphological analysis of the Mackintosh sediments as part of his B.Sc. thesis at La Trobe University; and Anne McConnell, from the Tasmanian ­Forestry

147

Nicola Stern ­ ommission, undertook the micromorphologC ical analysis of the Nunamira sediments. Rudy Frank provided invaluable assistance in the field and drafted the figures. Richard Cosgrove generously granted permission for the use of his original figures as a basis for Figures 8.3 and 8.5. Many stimulating discussions about the stratigraphy and chronology of the Mackintosh deposit were held with Nick Porch, but he may not wish to be held responsible for the views expressed here.

Notes 1. The “cumulative” and “spatial” palimpsests described by Bailey (2007:204–207) are conceptually similar, in most respects, to “time-averaged agglomerations” recovered from discrete deposits and paleolandscapes, respectively. However, embedded in the concept of time-averaged agglomerations is the additional notion that palimpsests of material remains do not always represent “average tendencies” (cf. Bailey 2007:204). The term time-averaged agglomerations is used here to avoid making assumptions about whether debris from temporally unrelated activities result in an assemblage in which specific types of material remains occur in the same proportion as the activities that generated them. It is conceivable that there are circumstances in which time-averaged agglomerations of debris are equivalent to an average representation of past activities, and it would be useful to identify the activities and circum-

stances in which this is the case. However, some activities and depositional circumstances either generate more debris or have a disproportionate impact on the configuration of surviving material remains than do others. The time-averaged agglomerations that result from these scenarios cannot be interpreted as an average representation of past activities. As Bailey (2007:​ 209) notes, palimpsests are an inherent property of the material world, and an improved understanding of the way they form in different settings is fundamental to the development of the theoretical and methodological tools needed to identify the behavioural information encapsulated within them. 2. The waters of Lake Mackintosh rise and fall gently, and there is no evidence to suggest that fluctuating lake levels have caused any erosion of the archaeological deposit in Mackintosh 90/1. The valley was not logged before the dam was filled, with the consequence that there is no high-speed boating, which might generate wave action. 3. Stratigraphic data thus preclude the suggestion by Allen and Cosgrove (1996:34) that the hearths in Horizon III date to the historic period. 4. In an analysis of radiocarbon determinations across southwest Tasmania, Holdaway and Porch (1996) reject this and some other determinations from Mackintosh 90/1, largely because statistical analyses identified them as outliers. However, if human occupation of a shelter is intermittent and irregular, such outliers might be expected.

148

9

Time Perspectivism and the Structure of Archaeological Records: A Case Study Josara de Lange La Trobe University

The idea that different dynamics can be observed at different temporal scales has become known in archaeology as “time perspectivism” (Bailey 1981, 1983, 2007). Time perspectivism provides strong arguments to counteract a focus among archaeologists on the short-term processes signaled in anthropological studies. Although it involves the dismissal of the notion that archaeological patterning can be fully explained with reference to processes observable “today,” time perspectivism does more than reject so-called substantive uniformitarianist assumptions (see Bailey 1987, 2007). It specifies that not all archaeological patterning can be explained with reference to localized, short-term processes, whether or not these processes can be currently observed. Thus, time perspectivism stimulates an interest in longer-term aspects of human prehistory. Such aspects would include longer-term trends and fluctuations in climate, biosphere, and so on, as well as in human behavior itself, which may well have ­longer-​term manifestations that cannot be reduced to variations in environmental parameters and which cannot be approached adequately by extrapolating what may be observed over a short time span (Bailey 1981, 1987; Fletcher 1992; Murray 1997). It is in the study of longer-term aspects of past human behavior that archaeology has a unique contribution to make to the human sciences: “Archaeology is the only form of academic enquiry that looks at

the nature of human behavior over very long spans of time” (Fletcher 1992:36). Indeed, Bailey (1981:​ 103) has been one of several archaeologists to suggest that archaeological patterning, particularly in prehistoric records, may be more amenable to study from longer-term perspectives than from short-term viewpoints. In this chapter it is argued that one of the greatest potential contributions of time perspectivism to the discipline of archaeology is situated in the fact that it enables a more sophisticated approach to the construction of evidence. As is discussed in more detail below, the quality of evidence is constituted in the interplay between the information gathered as a result of archaeological analysis and the evidential requirements of the research questions asked. Time perspectivism, by stimulating an explicit appraisal of the temporal scale and rhythm of phenomena or processes of interest, allows a more sophisticated assessment of the evidential requirements of hypotheses and thus of the quality of particular evidence in relation to these hypotheses. Clearly, this is an attractive prospect for any archaeologist who feels that a tighter interaction between data and hypotheses is desirable, if not crucial to the ongoing success of the discipline. On the flip side of the coin, a more sophisticated appreciation of the interaction between data and hypotheses will, in all probability, render some long- and dearly held theories about the past inadequate by

149

Josara de L ange our own standards. Indeed, time perspectivism affects the foundations of the archaeological project, and it can be expected to reverberate not merely at the level of particular hypotheses and theories but at deeper levels, where the conceptualization of archaeological phenomena and the construction of meaning take place. As the history of scientific revolutions suggests, it may not be an easy ride. To illustrate how a more explicit consideration of the temporal structure of archaeological data may affect archaeological reasoning, the first part of this chapter focuses on the extensively researched and well-publicized Upper Paleolithic assemblages of Klithi (Epirus, Greece), excavated between 1983 and 1988 under the coordination of the father of time perspectivism, Geoff Bailey (1997c; Bailey and Gamble 1990; Bailey et al. 1983a, 1983b; Bailey et al. 1993). In order to gain an understanding of the nature of the evidence generated in the Klithi project, a twofold discussion is presented. First, the lithics and faunal assemblages are discussed in some detail, as is the temporal structure of the archaeological record. Some of the hypotheses presented by the respective researchers are explored. It is then argued that some of the data generated cannot be construed as evidence in the context of the hypotheses they are purported to support: they do not interact with the hypotheses and thus do not render (aspects of ) the hypotheses in question stronger or weaker than they were at the outset. The second part of this chapter presents a discussion of the manner in which paleobiologists have addressed the construction of evidence in their discipline. The concepts generated in paleobiology are to a degree pertinent to archaeological research because they address the multitemporal nature of the phenomena under study as well as the structural properties of the geological records in which traces of these phenomena are present. Moreover, they are formulated in such a way as to take explicit account of the interaction between data and research questions. However, although this discussion is enlightening in that it provides insight into the pathways by which researchers dealing with similar problems have developed their conceptual arsenal, it is strongly argued that archaeological research presents its own unique challenges that war-

rant further investigation of these concepts rather than wholesale adoption.

The Construction of Evidence: The Case of Klithi The Klithi rockshelter is a large, south-facing limestone rockshelter, situated at an altitude of 430 m in the foothills of the Pindus Mountains, on the eastern bank of the Voïdomatis River (Figure 9.1; Bailey 1997a). Its investigation took place in the wider context of research into patterns of land use and settlement in the Epirus region during the Paleolithic, instigated in the 1960s by Eric Higgs. Higgs excavated several other sites in the region (most notably the rockshelters of Asprochaliko and Kastritsa and the open-air site of Kokkinopilos [see Figure 9.1; Bailey 1997a]) and interpreted the results as indicating that the rockshelters were seasonally complementary home bases forming part of a single system of exploitation, tracking the annual movements of red deer populations (Bailey et al. 1983b). Open-air sites such as Kokkinopilos were thought to represent kill sites or transit sites within the same system (Bailey 1997a). The Klithi project was set up to reexamine and extend the regional database, partly in order to evaluate some of the premises and conclusions of the earlier project. It can be seen as a study of archaeological data on different spatial and temporal scales, a coordinated attempt to provide “a sharper focus on variability and change in archaeological, palaeoeconomic and palaeoenvironmental sequences at intra-site, local inter-site, and regional scales of investigation” (Bailey 1997a:11). Although the site clearly needs to be considered within its various spatiotemporal contexts, the discussion below focuses on aspects of the intrasite patterning in order to show how an explicit consideration of the temporal structure of the data affects the construction of evidence in the context of the study of past subsistence practices.

The Archaeological Record of Klithi: Contents and Structure

The rockshelter of Klithi houses a substantial sedimentary sequence consisting of roughly stratified, poorly sorted, coarse limestone screes within

150

Time Perspectivism and the Structure of Archaeological Records

Figure 9.1.  Klithi and other sites mentioned in the text in their regional setting (after Bailey 1997a:Figure 1.3).

a predominantly silt-grade, calcareous fine matrix (Bailey and Woodward 1997; Woodward 1997). The fine sediments entered the rockshelter in various forms and by several pathways: as a weathering product of overlying limestone and flysch deposits, transported through the interior drainage system of the limestone bedrock, and as eolian dust of uncertain origins (Woodward 1997:374). The top 2 m of the sequence contain the bulk of the archaeological material. Within this part of the sequence no archaeologically sterile deposits were observed, and the distribution of the material throughout the deposit points toward repeated occupation. Excavation took place in several areas (see Figure 9.2): in the V/W area excavation was aimed at establishing an overall stratigraphic sequence, whereas in the P–T area excavation was aimed at exploring lateral variation (Bailey 1997b). Within these areas local layers or contexts were defined on the basis of color, sediment type, and so on and were subsequently

grouped into site-wide strata on the basis of assumptions regarding stratigraphic relationships and consideration of the radiocarbon dates (Bailey and Woodward 1997). The lithic assemblages excavated at Klithi contain manuports, cores, flakes, blades, bladelets, and microbladelets of both local and nonlocal raw materials. The retouched component consists of typically Upper Paleolithic tool forms such as end scrapers, borers, burins, notches, denticulates, side scrapers, truncated pieces, and backed pieces (Roubet 1997a). Around 80 percent of the lithics are of local raw material, which can be collected from the Voïdomatis River gravels in the form of small pebbles. The remaining 20 percent are of raw materials from farther afield: some from adjacent river catchments 15–20 km away, others possibly brought in over much larger distances (Bailey 1997a:20). The vast bulk of the identifiable faunal specimens could be assigned to Capra cf. ibex, Rupicapra 151

Josara de L ange

Figure 9.2.  The layout of excavation trenches within the Klithi rockshelter (after Bailey 1997b:Figure 3.13).

rupicapra, and Capra/Rupicapra (ibex and chamois [Gamble 1997]). Other taxa were identified from only a few bones and include Cervus elaphus, Castor fiber, Felis cf. leo, Felis lynx, Meles meles, Canis lupus, Vulpes sp., and Lepus sp. A canine and a phalanx of Homo sapiens were found, as well as a small number of small rodent, bird, and fish bones. Of the unidentified component, most remains were ascribed to the category of small-sized artiodactyls or Capra-sized animals. Very few bones show signs of carnivore gnawing (Gamble 1997:219). Burned bones and bones showing signs of butchery are more common, and most of these, when identified, belong to caprines or small-sized artiodactyls. Butchery marks consist of traces of cutting, chopping, and splitting. Many bones were processed for marrow, and almost all phalanges with marrow cavities occur as split or broken specimens (Gamble 1997:223). Cut marks were also observed on a tibia of Castor fiber and on a metatarsal of Lepus sp. Dozens of bone and tooth artifacts such as needles, awls, spatulas, points, and perforated teeth were found. Both the predominance of locally available raw materials and the presence of manufacturing waste suggest that at least some of them were made on-site (Adam and Kotjabopoulou 1997).

The majority of radiocarbon determinations on samples from the archaeological end of the Klithi sequence yielded dates between circa 16,500 and 13,500 bp, or between circa 20,000 and 16,500 cal bp (see Gowlett et al. 1997:Tables 2.1–2.2), and it is suggested that this is the period during which most of the archaeological material was deposited. Although some samples produced younger radiocarbon dates, these are taken either to indicate sporadic use of the rockshelter or to testify to a possibly intense occupation of which the remains have been removed by recent leveling of the deposits (Bailey and Woodward 1997). Most of the remains actually present are thus considered to represent a period of roughly 3,000–4,000 years, coinciding with the Last Glacial Maximum (Gowlett et al. 1997). Consideration of the structural properties of the Klithi archaeological record indicates that the archaeological data are relatively coarse grained. Given the limited spatial extent of the rockshelter, postdepositional mixing and disturbance of the deposits by subsequent human activities may well have taken place, and there is clear evidence of recent trampling by goat herds. Trampling may account for the highly fragmented nature of the faunal remains, and the refitting of lithics has confirmed that archaeological debris has moved over 152

Time Perspectivism and the Structure of Archaeological Records c­ onsiderable distances through the sedimentary column (Wenban-Smith 1997). Radiocarbon dates indicate that several strata are not chronologi­cally distinctive, and their contents need to be lumped from a chronological point of view (Bailey and Woodward 1997:​82–83). Moreover, the ­error margins of the radiocarbon dates are such that, altogether, the maximum resolution attainable sitewide is of the order of magnitude of 103 years (Bailey and Woodward 1997:83). The lithics assemblages from the W24–29 area (Figure 9.2) were analyzed by context (Roubet 1997a, 1997b; see below). They represent unknown time spans; however, the maximum mean sedimentation rate calculated for Klithi (.067 cm per year [Bailey and Woodward 1997:Table 4.1]) suggests that the contexts in question may well have taken several hundreds of years to form. Refitting has shown that vertically adjacent contexts may have undergone a significant exchange of material, further enlarging the potential time span represented by the assemblages (Wenban-Smith 1997). The faunal analysis involved a comparison of assemblages from six different strata, and it has been suggested that three of these need to be treated as a single unit from a chronological perspective (Bailey and Woodward 1997). As such, the assemblages each represent a minimum of several hundred years of formation and possibly a thousand years or more. Thus, the available data do not allow for an assessment of the temporal distribution of the lithics and faunal remains (regardless of how the temporal position of an archaeological object is defined) at scales finer than 102 to 103 years (Bailey and Woodward 1997; Galanidou 1997a; Winder 1997).

ent types of cores, blanks, and retouched tools of the various raw materials. These operational chains (called “virtual chains” by Roubet, to distinguish them from the “real” but very short chains established through refitting [see Wenban-Smith 1997]) are interpreted in terms of a knapper’s sequence of operations, incorporating both mentally and physically performed aspects and flexibly geared toward particular and context-specific goals (Roubet 1997a:​130). They were constructed by ascribing the artifacts to “the different stages of the knapper’s sequence of operations”—the selection of raw materials, the production of blanks, the selection of blanks for modification, and the production of tools (Roubet 1997a:​130). The analysis indicates that both local and nonlocal raw materials were worked on-site in similar ways, although the larger blanks (≥ 5 cm) are generally of nonlocal materials (Roubet 1997a:​150). Roubet emphasizes the interdependence between different raw materials, stating that artifacts of nonlocal raw materials “appear to be...intimately mixed with all the unmodified and retouched pieces made on local materials” and that “such a diversity was obviously deliberate and must be taken into account as a ...component of the procurement system” (1997a:147). With regard to the abundant remains of ibex and chamois, the tooth data indicate the dominant presence of young and old animals, whereas the bone fusion data point to a large adult population. Moreover, the skeletal element representation for prime adult animals, when classed according to food utility values, follows a “gourmet curve,” with predominantly medium- and high-food-value elements, whereas that for younger and older animals follows a “bulk curve,” with low- to high-foodvalue elements (Gamble 1997:230). Gamble has proposed that this patterning may be the result of differential carcass processing and transport strategies for animals of different ages. He suggests that the Klithi hunters focused on adult animals but also took young and old animals; they adapted their butchery strategy to the condition of the prey, with heads mainly brought into the site when the food value of the carcasses was low because of the small size or poor condition of the animals. These hunting and carcass-processing strategies are described

Interpretation of the Lithics and Faunal Assemblages

An appreciation of the complex formation histories and coarse resolution of the assemblages is clearly shared by all contributors to the Klithi monograph (Bailey 1997c). Despite the coarsegrained nature of the data sets, however, attempts were made at establishing aspects of stone reduction and hunting and butchery strategies. Roubet (1997a) has constructed chaînes opératoires on the basis of an analysis and comparison of the differ153

Josara de L ange as strategies taking shape in “real time,” involving responses to short-term resource availability and predictability and the scale of provisioning needs (Gamble 1997:​230, 238–239). It has been observed that there appears to be a striking lack of temporal variability in both the lithic and the faunal material from Klithi (Gamble 1997; Roubet 1997a, 1997b). Different excavated areas sampling different parts of the stratigraphic sequence yielded similar lithic assemblages: a large quantity of both large-sized and microlithic material, with elements from all parts of the reduction sequence, spatially associated with faunal remains, shell ornaments, and bone implements (Roubet 1997a:​126). No major changes in the nature, composition, and distribution of the retouched lithics were noted during excavation (Roubet 1997a:126), and subsequent analysis has confirmed that any differences and changes discernible are subtle. Likewise, the Klithi faunal remains are taken to present “a picture of no change through 3000 years” (Gamble 1997:​239). In terms of subsistence strategies it is suggested that the Klithi faunal assemblages provide “an example of redundant behaviour in the face of environmental fluctuations” (Gamble 1997:​ 208). Whereas “climate and resources fluctuated ... on a shorter ecological rather than a longer geological timescale” (Gamble 1997:239), the faunal remains from Klithi apparently do not reflect these fluctuations. Thus, “it may be that one way to cope with unprecedented uncertainty [resulting from such fluctuations] is to increase redundancy, that is to develop specialized and repetitive patterns of behaviour, rather than to multiply innovations to meet each new environmental variation” (Gamble 1997:​239). As outlined above, however, any two or more remains from the assemblages in question are only “associated” to the degree that they were deposited within the same 102- to 103-year interval. No criteria have been proposed for distinguishing units of finer resolution within the context- and stratum-specific assemblages, beyond the level of single artifacts and refitted sets. As is argued in the next section, given the structural properties of the Klithi archaeological record, data relevant to the assessment of some of the hypotheses presented to account for the pat-

terning in both the lithics and faunal assemblages (in particular insofar as these hypotheses refer to strategies and to change—or stasis—in these strategies over time) are simply not available, and alternative hypotheses cannot be discounted.

Assessing the Interaction between Data and Hypotheses

The lack of data relevant to the assessment of the presented hypotheses regarding (the lack of change in) resource procurement and processing strategies is best illustrated using a simplified hypothetical assemblage that is in some ways similar to the Klithi assemblages. Imagine a faunal assemblage with the remains of 130 animals, the majority of which can be ascribed to adult animals (minimum number of individuals [MNI] = 100), with a smaller number of remains of younger and older animals (MNI = 30). Let us assume that this assemblage represents a time span of 500 years, that is, a time span of roughly the same order of magnitude as that covered by each Klithi stratum assemblage. In the absence of temporally more fine-grained data (again analogous to the Klithi situation) it is not possible to say how the deposition of these faunal remains was distributed over the 500 years of assemblage formation. Table 9.1 outlines three scenarios of deposition for this hypothetical assemblage, but many more can be envisaged. Moreover, a sequence of similar assemblages, analogous to the Klithi situation, could have resulted through repetition of one of the scenarios or from any combination of these (or alternative) scenarios. Scenario 1 is of course highly unrealistic in its regularity but is in principle, given the available data in this hypothetical case, a possibility. It is characterized by a constant rate of deposition of the remains of both adult and younger and older animals, and it provides no clues as to any change in the processes responsible for the deposition. Scenario 2, on the other hand, sees a more or less gradual shift in deposition, from a dominance of the remains of adult animals to a dominance of the remains of younger and older animals. Finally, scenario 3 features a punctuated shift, from the exclusive deposition of remains of adult animals to the exclusive deposition of remains of younger and

154

Time Perspectivism and the Structure of Archaeological Records Table 9.1.  Three Possible Scenarios for the Depositional History of a Hypothetical Assemblage Spanning 500 Years and Containing the Remains of Both Adult Animals (Minimum Number of Individuals [MNI] = 100) and Young and Old Animals (MNI = 30) Time (Years)

Scenario 1

Scenario 2

Scenario 3

0–50 50–100 100–150 150–200 200–250 250–300 300–350 350–400 400–450 450–500 Assemblage Content

10 adult, 3 young/old 10 adult, 3 young/old 10 adult, 3 young/old 10 adult, 3 young/old 10 adult, 3 young/old 10 adult, 3 young/old 10 adult, 3 young/old 10 adult, 3 young/old 10 adult, 3 young/old 10 adult, 3 young/old 100 adult, 30 young/old

15 adult 22 adult, 1 young/old 16 adult, 1 young/old 11 adult 21 adult, 3 young/old 6 adult, 4 young/old 4 adult, 6 young/old 3 adult, 5 young/old 2 adult, 6 young/old 4 young/old 100 adult, 30 young/old

10 adult 30 adult 40 adult 20 adult — — — 8 young/old 15 young/old 7 young/old 100 adult, 30 young/old

older animals, separated by a period of 150 years in which no deposition took place. Insofar as these scenarios differ from each other, and insofar as the hypothetical data at hand can be considered informative on procurement practices underlying the formation of the assemblage, they would likely be interpreted differently. However, the conflated result of these different scenarios is identical: 100 adult animals and 30 young and old animals. In the absence of criteria on the basis of which to assign data to finer time slots, all three scenarios, as hypotheses about the formation of the assemblage, have equal probability, as do different interpretations of the assemblage insofar as these are dependent on differences in the scenarios. Given the time spans represented by the Klithi faunal assemblages, any changes in faunal deposition that occurred over time scales significantly smaller than several hundred to a thousand years (such as occur in scenarios 2 and 3) would remain undetected. Thus, it is not necessarily the case that the Klithi occupants displayed highly redundant behavior in the face of climatic instability, as it remains possible that significant shorter-term adaptive changes in procurement strategies (occurring on the scale of decades or centuries) are masked in each of these assemblages. It also remains possible that the sequence of similar stratum assemblages at Klithi was generated in the context of longer-term changes in the stability of these strategies, analo-

gous to a sequence of scenarios of the types outlined in Table 9.1. Some aspects of the interpretation of the lithics assemblages from Klithi suffer from the same problem. Although these assemblages likely represent shorter overall time spans than the faunal assemblages, these time spans are long enough to have produced palimpsests that do not allow for the detection of relatively rapid changes in reduction strategies and raw material selection. For instance, the presence in many of the contexts of pieces from all stages of the reduction sequence could have resulted from a scenario in which a period during which the earlier stages of reduction were preferentially performed on-site was followed by a period that saw a focus on the production of tools from imported blanks. Thus, we may be dealing with the remains of various reduction sequences involving partial chaînes opératoires that, in their conflated state, mimic complete chains. Likewise, the variety of local and nonlocal raw materials retrieved from the contexts, interpreted by Roubet (1997a) as a deliberate component of the procurement system, could mask shifts in raw material selection if these shifts occurred over time spans smaller than those over which the assemblages were formed. The point of the above discussion is not to suggest that hypotheses involving change in procurement and processing strategies are to be preferred over hypotheses of stability and redundancy. 155

Josara de L ange Rather, it is to emphasize that the data that would allow an assessment of the merit of these hypotheses, as well as of their antitheses, have not been generated—in this case because they could not be generated, given the structural properties of the archaeological record in question. In the face of the hypotheses presented, the archaeological data sets in question are significantly time averaged. The concept of time averaging, which has been developed in paleobiology, is discussed in more detail in the following section.

Time Averaging and the Construction of Evidence in Paleobiology Research on paleobiological data has given rise to the notion that fossil records are time-averaged records (Walker and Bambach 1971; see Kowalewski 1996). Time-averaged records are records in which “the remains of organisms that did not live together end up in the same deposit” (Olzewski 1999:226). The prevalence of relatively low sedimentation rates during the time interval of deposition of the phenomena under study can contribute to time averaging, as well as mixing through bioturbation and the reworking of sediments (Fürsich and Aberhan 1990; Kowalewski 1996). The resistance to destruction and the abundance of the phenomena under study also influence the susceptibility of a record to time averaging. Phenomena with a low preservation potential or that are rare are less likely to survive various processes contributing to time averaging than phenomena with a high preservation potential or that occur in large numbers. Also, so-called analytical time averaging may take place as part of research, when samples are pooled (Kowalewski 1996). What the processes mentioned above have in common is that they compromise the applicability of the principle of stratigraphic superposition and its capacity to provide a spatiotemporal framework for the study of fossil assemblages. Time averaging has been considered a special case of stratigraphic disorder or the “departure from perfect chronological order. . .​in a stratigraphic section” (Cutler and Flessa 1990:​227). In relation to Bailey’s definition of time perspectivism, time averaging may be

seen as affecting the ability of researchers to bring into focus certain processes by compromising their ability to arrange data according to the appropriate time scales. Indeed, one of the more problematic aspects of time averaging, and probably one of the reasons for the persistent confusion regarding its impact, is the relative nature of the phenomenon. Time averaging signals the inability of researchers to ascribe their data to time slots that can reasonably be expected to be relevant to the research questions posed. Whether or not a record is time averaged thus depends on the research questions being asked. When a record is time averaged in the face of a particular historical research question, the time series necessary for answering or exploring that question cannot be constructed. In light of the question-dependent nature of time averaging Kowalewski distinguishes between significant and insignificant time averaging: “Significant time-averaging happens when the scale of temporal mixing exceeds the time-scale of a process. Conversely, insignificant time-averaging happens when the time-scale of a process exceeds the scale of temporal mixing. The time-scale of a process is the time-averaging threshold that determines if the events are time-averaged significantly” (1996:​ 320). A record that is significantly time averaged when confronted with ecological questions may be insignificantly time averaged when confronted with evolutionary questions. For instance, a series of superposed shell assemblages, each representing several hundreds of years of deposition, is unsuitable for addressing hypotheses about the makeup of the ecological community at particular times. However, it may well be suitable for addressing hypotheses concerning longer-term dynamics such as gradual shifts in species dominance. It thus requires an idea of the time scale over which a process of interest operates to establish whether a particular record is significantly time averaged or not, or to embark on a quest to select the records most likely to engage with a particular research question. Another interesting phenomenon that has become visible because of the research of Kowalewski and colleagues has been described as “disharmonious time averaging” (Kowalewski 1996). Because fossil and trace records (i.e., records ­consisting

156

Time Perspectivism and the Structure of Archaeological Records of behavioral traces, whether animal or ­human) are compiled from the products of many different processes (inter)acting over different spatial and temporal scales, and because these products are differentially resistant to destruction, different phenomena may be time averaged to different degrees even when preserved in the same geological unit. For instance, imagine a midden assemblage consisting of the shells of two mollusk species, one much more resistant to breakage than the other, contained in a 10-cm-thick deposit representing 100 years of deposition. In this case the more fragile shells, insofar as identifiable, may be time averaged over a much shorter time span than the more durable ones: because of the differences in durability and because the chances of postdepositional breakage increase with time of exposure and subsurface reworking, the fragile shells deposited during the earlier part of the 100 years are more likely to have been broken beyond recognition than the durable ones deposited during that same period. In such a case it is possible that the durable component represents a much longer time span (even close to that of the period of deposition, i.e., in this hypothetical case 100 years), whereas the bulk of the identifiable fragile remains likely stem from the final part of the period of deposition (say, the final 20 years). Indeed, disharmonious time averaging “can be expected to be the norm rather than the exception in the geological record” (Kowalewski 1996:321), because not only do entities vary in their potential for preservation but processes contributing to time averaging can be very selective, even among entities that have a similar preservation potential. Over the last decade, invertebrate paleobiologists have attempted to get an idea of the extent and structure of time averaging in standard fossil samples through the combined application of amino acid racemization and 14C dating (Goodfriend 1987, 1989; Kowalewski et al. 1998). In particular, they have tried to gain insight into the extent of time averaging in geologically uniform and stratigraphically restricted deposits by dating individual shells from such deposits. In a recent study of shell assemblages from the Colorado River Delta (Kowalewski et al. 1998), a large number of shells were dated, and it was found that samples tended to be

time averaged over much longer time spans than anticipated. Even when samples were collected so as to minimize their stratigraphic and lateral extent, substantial time averaging occurred: the age range between the youngest and the oldest shell within small samples exceeded, on average, 600 years (Kowalewski et al. 1998:​296). This particular study shows that the samples habitually taken in paleobiological practice for the study of ecological issues may be unsuitable for this purpose, and conclusions drawn from them, spurious. It may serve as a cautionary tale for archaeologists wanting to use stratigraphic considerations to make a case for the short-term integrity of archaeological assemblages.

Time Averaging and the Construction of Evidence in Archaeology: Lessons from Klithi Despite claims to the contrary, archaeological records are not fossil records. However, research into the structure, extent, and consequences of time averaging of fossil records raises interesting issues and questions for archaeologists. As was argued above, the Klithi archaeological record is significantly time averaged when it comes to an assessment of some of the outlined hypotheses regarding (changes in) resource procurement and processing strategies. Various significantly different hypotheses about the formation of the faunal and lithics assemblages have equal probability, at least insofar as the structure of the available data is concerned. Moreover, there is ample reason to suspect that the lithics and faunal assemblages from single contexts or strata at Klithi are disharmoniously time averaged. Given the maximum mean sedimentation rate for Klithi (see above), a large amount of faunal remains may well have been lost, even if it is granted that bone weathering in the sheltered setting of Klithi proceeded at a slower pace than bone weathering in more exposed environments. The extremely fragmented state of the faunal material (Gamble 1997) suggests that postdepositional processes have had much more destructive effects on the bone component than on the more durable lithic component of the assemblages. As a result, much faunal material may not have (identifiably) survived the processes of mixing for which there is evidence in the lithic

157

Josara de L ange material record (Wenban-Smith 1997). The consequences of this possibility for the development of models of the subsistence practices of the Klithi occupants remain underexplored. More generally, by assessing the structural properties of the Klithi archaeological record and the interpretations generated in its study we are confronted with several layers of interpretative problems posed by time averaging: it challenges our capacity to identify sets of materials related to certain behaviors (an issue discussed in detail in de Lange 2005), as well as our ability to assess hypotheses regarding the coordination of activities (strategies) and the longer-term change or stasis in these. In regard to the first layer of problems, our lasting fascination (to put it mildly) with “high-definition­” occurrences (cf. Gowlett 1997) is testimony to our infamous inability to construct morphological criteria for the recognition of associations with behavioral integrity, especially (but by no means exclusively) in settings lacking built structures. Recognition of the archaeological correlates of particular behaviors is problematic because attempts to retrodict causal relations between different sets of artifacts suffer almost invariably from the fact that the morphological characteristics of assemblages do not allow a reconstruction of the initial conditions in enough detail to warrant a choice between competing hypotheses. Even high-resolution occurrences (e.g., Audouze and Enloe 1997; Roberts and Parfitt 1999; Stapert and Street 1997) may in many cases not constrain the field of hypotheses to the degree necessary for distinguishing behavioral correlates (de Lange 2001, 2005). In regard to the second layer of problems, those involved in the assessment of hypotheses about subsistence strategies and the recognition of change and stasis in these, although it may be true that “in the absence of temporal differentiation we must assume that the mix of activities did not change over time” (Winder 1997:273), the absence in the Klithi assemblages of temporal differentiation on the time scales of human activities has not been convincingly demonstrated. Indeed, given the current state of our methodological and theoretical tools, it is far from clear how it could be. These assemblages are significantly time averaged in the face of hy-

potheses regarding the constitution of and change in the “mix of activities,” and the absence or presence of temporal differentiation at these scales cannot be assessed. As has been noted by Lake (1996), the only situation in which it is not problematic to use significantly time-averaged data is when no change in the parameters investigated has in fact (or according to an independently generated model) taken place. In such a situation, an assemblage can justifiably be regarded as a numerically exaggerated reflection of what was deposited over the shorter time spans of interest. Gamble does briefly refer to independently generated models of rates of change in adaptive strategies. He suggests that, “according to many processual models” (1997:208), significant shifts in adaptive strategies should have occurred during the time of occupation of the Klithi rockshelter, given that climatic fluctuations affecting the availability of resources would have occurred on a “shorter ecological” time scale rather than over the long term. Acknowledgment that these models might have merit warrants a cautious and sophisticated approach to the interpretation of the similarities in the Klithi faunal assemblages. In the absence of a more detailed consideration of the time scales over which these fluctuations and concomitant adaptive responses may have taken place, the significance of the similarities in the assemblages remains unclear. The hypotheses presented to account for the Klithi assemblages effectively require the assumption that the apparent lack of change at the temporal scale of the available data is reproduced at smaller temporal scales. Finally, the study of the wider land-use strategies in which behaviors and localized strategies were embedded typically requires that spatial scales beyond that of the site be taken into account, and an investigation of the local and regional context in which the archaeological deposits of Klithi are situated forms a prominent part of the Klithi project. At such scales, however, there is often no reason to assume that sedimentation was uniform and regular throughout the area under consideration, and problems related to the time transgressiveness of the geological deposits containing archaeological materials gain currency (see Stern 1993, 1994a, 158

Time Perspectivism and the Structure of Archaeological Records 1994b, 1995). These problems, stemming from the fact that the formation of geological deposits is to a degree dependent on local factors that come into play at different times at different places, help explain the difficulty of establishing the (possibly changing) relationship between the rockshelter of Klithi and the smaller rockshelter of Megalakkos situated in the same valley, which contains deposits broadly contemporaneous with those at Klithi.

Conclusions Time perspectivism explicitly embraces the idea that different dynamics can be observed on different temporal scales, and as such it stimulates a more explicit appreciation of the temporal structure of archaeological records as well as of the data requirements of hypotheses about the dynamics that may have been involved in their formation. This presents us with a conundrum. On the one hand, processes that are fairly well understood have presumably left traces in the archaeological record, but the interpretation of these traces often requires a degree of retrodiction that even the most highresolution records cannot sustain. On the other hand, our understanding of longer-term processes is as yet wanting: like the Annales school of historical thought, we still lack “a consistent statement about the magnitude of processes, their effects, the time scales involved, and the relationship between them” (Fletcher 1992:​37). Currently, this may in many cases prevent us from making meaningful statements regarding the significance of time averaging of particular records with regard to specific research questions. In the face of the general lack of research into the structural properties of archaeological records and the undertheorization of the relationships between hypotheses and archaeological data, the bold conviction with which archaeologists in general have on many occasions presented particular versions of developmental trajectories and historical interactions is surprising and counterproductive. One of the most invasive effects of the fact that many archaeological data sets cannot be resolved to the degree necessitated by current interpretative frameworks (and of the fact that current interpretative frameworks are not adequate in the

specification of evidential requirements) is an explosive proliferation of possible formation histories, all equally likely candidates for the status of actual formation history as far as the archaeological data are concerned. It is premature and unprofitable for us as professional archaeologists to pledge allegiance to a particular rendering of past dynamics in the absence of strong archaeological arguments. To do so precludes any independent archaeological assessment of claims made on the basis of nonarchaeological data. Moreover, it detracts from the exploration of the multitemporal character of archaeological information and of the possibility that change at some scales may produce stasis at other scales (and perhaps vice versa). Instead, we might take the opportunity to look closer at the gaps in our understanding of past dynamics and at the ways in which they are created in the interaction between our analytical frameworks and archaeological records. An improved understanding of the structural properties of archaeological records will allow us to present an image of the range of possible historical trajectories against which fellow researchers can assess their own interpretations of past and contemporary dynamics. More importantly, a growing awareness of how the structural properties of archaeological records constrain and fail to constrain different research questions might drive us, in the tradition that spawned time perspectivism, to experiment with alternative modes of analysis and interpretation and to develop a more sophisticated appreciation of the spatiotemporal shapes of archaeological phenomena. It would be illusory to think that as archaeologists, dealing with unique long-term records of human behavior, we can simply bypass the messy arena of the short term and successfully study ­longer-​ term dynamics. On the one hand, as indicated by the Klithi case study, we need to carefully consider the temporalities of the processes to which our hypotheses refer and engage in a critical evaluation of the suitability of available data for the assessment of these hypotheses. On the other hand, we may ask what the significance could be of detected patterns: What might the apparent homogeneity of the various lithics and faunal assemblages indicate about the longer-term dynamics involved in the

159

Josara de L ange generation of the assemblages, given the possibility that this similarity occurred in the presence of significant shorter-term shifts in procurement and processing strategies? What aspects of this homogeneity could be explained solely with reference to change and stability in natural formation processes, and which aspects hint at longer-term behavioral trends? How much behavioral change or variation could conceivably be masked by the structural properties of the Klithi record? Ultimately, coming to terms with the construction of evidence from a time perspectivist viewpoint may well require a thorough reconceptualization of our analytical and interpretative frameworks (cf. Murray 1997) and a rethinking of the concepts of the short term and the long term and the relationships between them. Perhaps, in our exploration of such questions, we will find that “for all those processes having no modern analogues, it

may be impossible to decide whether the geological patterns representing them are time-averaged significantly, time-averaged insignificantly, or simply spurious” (Kowalewski 1996:325). But then again, perhaps a thorough investigation and discussion of the spatial and temporal structure of archaeological records will allow us to create new frames for new visions.

Acknowledgments Although no one but me should be held responsible for the content of this chapter, I would like to thank Tim Murray and Nicola Stern for discussions that helped shape some of my thoughts, as well as the organizers of and the participants in the time perspectivism session at the Society for American Archaeology Annual Meeting in April 2003 in Milwaukee. Greg Deftereos kindly adapted the figures for this chapter.

160

10

No Time like the Present Philip J. Arnold III Loyola University In the long run, all that really matters is the long run.

To many eyes, ethnoarchaeology is the ultimate exercise in futility. Using contemporary behavior to inform archaeological interpretations is likely to peg you squarely on one of the dilemma’s two very sharp horns, either skewered by the tyranny of the ethnographic record (e.g., Wobst 1978) or gored by the reconstructionist implications of the Pompeii premise (e.g., Binford 1981a). Little wonder that a recent overview of ethnoarchaeology laments the lack of young scholars who are eager to fill the ranks of the subdiscipline (David and Kramer 2001:​416–​ 417). Although the above conundrum is understandable, it represents a particular mind-set regarding archaeological practice. This mind-set often includes a view of the past that I call “reel-time archaeology”—a seeming desire to create individual ethnographic vignettes of the past and then string them together like frames in an old filmstrip or a movie reel. Chronological concerns within reel-time archaeology emphasize the temporal resolution of each vignette, apparently in the hope that when shown in sequence the entire reconstruction will appear smooth, natural, and “lifelike.” As a result, these investigations of long-term phenomena tend to emphasize how fast or slow cultural change occurs, as if the rate of change is the only or the most meaningful dimension of temporal variation. As long as research advances this reel-time agenda, ethnoarchaeology will remain a curious intellectual specimen, pinned down firmly by its own temporal 161

limitations and with little to offer our understanding of the past. This chapter explores an alternative to reel-time archaeology. If you will indulge the metaphor for just a moment longer, I propose that audio is conceptually more similar to the archaeological record than is video. Rather than a frame-by-frame sequence of past activities, the archaeological record is one long series of physical manifestations produced by the constant flow of time. In this sense, time is akin to the continuous groove on the archaeological “record,” culture represents the ­needle, and artifact distributions are like vibrations emitted when culture interacts with time. From this vantage, the goal of archaeology should not be the simple creation of discrete ethnographies strung together in a temporal sequence. Instead, research should seek to identify and understand the organizational principles behind the interaction of culture and time; that is, how can the “vibrations” of the archaeological record be fruitfully understood and explored? Seen in this way, the rate of cultural change (i.e., gradual vs. punctuated) is but one potential archaeological question. A question that I find more satisfying involves identifying different types of change and examining the circumstances that would promote one type over another. Given such a research focus, neither the tyranny of the ethnographic present nor the pitfall of the Pompeii premise offers a serious stumbling block. Ethnoarchaeology as used according

Philip J. Arnold III to this approach is not about creating simple analogues for the past; rather, it serves to establish inferential constants or fixed reference points, which in turn permit researchers to identify and measure behavioral variation manifest in the archaeological record (e.g., Arnold 2003; Binford 1981b, 2001a). Below I explore the temporal implications of the reel-time view and actualistic research. In the first section I support the contention that there is a reel-time bias in contemporary archaeology and consider its impact on “chronology theory.” I then address the issue of change and make the case that change is best addressed as a process of “begoing” rather than a state of becoming. The second section discusses actualistic research and the implications of such studies under­ taken over a short time period. I argue that the length of time during which observations are made is less important than establishing secure material consequences of specific activities. I also investigate the value of considering different kinds of change, using examples from ethnoarchaeological studies as well as a recent global synthesis of ethnographic data.

Time and Change Time, for most archaeologists, constitutes a dimension in desperate need of control. This control includes the accurate and precise location of one’s archaeological data along a temporal stream. The increasing reliance on accelerator mass spectrometry (AMS) assays reflects this interest—after all, who would not prefer an AMS date with a onesigma deviation of 10 years over a conventional radiocarbon date with a one-sigma deviation of 150 years? But the issue of temporal resolution often extends beyond the accuracy, precision, and reliability of a given chronometric tool. For some researchers, temporal resolution also involves the span of particular periods or phases. As dating techniques become more refined, the length of these archaeological constructs is increasingly adjusted and usually shortened. Thus, regional chronologies divided into units of 100 years or less can now be found in the literature. The logical extension of this approach would seem to be the viability, and desir-

ability, of temporal spans that conform to our lived, ethnographic experience. Although these finer chronologies may satisfy us on a visceral level, there has been surprisingly ­little theoretical reflection as to why they are either necessary or desirable. One gets the sense that this archaeological truth is thought to be self-evident— when it comes to chronologies, shorter temporal spans are simply better. Nonetheless, the assumptions that underwrite this view merit inspection. Should archaeological chronologies correspond to ethnographic realities? I would answer no, and I argue below that archaeology is becoming overly wed to, and limited by, an ethnographic temporality. Moreover, this conventional emphasis highlights rates of change at the expense of exploring different types of change. This latter objective is potentially more informative and is certainly more consonant with the practice of ethnoarchaeology.

Archaeology in Reel Time

A common way in which the ethnographic record is used for archaeological purposes is akin to envisioning the archaeological record as if it were a film on a movie reel (e.g., Strobach 1998:201–205). Stratigraphic and chronometric units (i.e., archaeological level, phase, period, etc.) are often treated as if they were individual frames within that movie. To create a meaningful sequence one must make the “action” appear as smooth and lifelike (e.g., ethnographic) as possible. Two options would facilitate this goal. Researchers can either increase the overall speed of the film reel while holding the size of each frame (e.g., temporal length) constant, or they could reduce the size of each frame while holding the speed of the film reel constant. Of course, we cannot avail ourselves of the former option; that is, we cannot change the overall “speed” of time. Thus, researchers have focused on the chronological resolution (e.g., time span) of each archaeological frame. This latter option appears to be what Smith (1992) has in mind as part of his discussion of “chronology theory” in archaeology. Using the Pompeii premise exchange between Binford (1981b) and Schiffer (1985) as his point of departure, Smith

162

No Time like the Present explores the possible temporal discontinuities between the ethnographic and archaeological records. Of course, a central issue in the Pompeii premise debate involved how the archaeological record should be made meaningful in anthropological terms (e.g., Murray 1999a). As part of that exchange, Binford (1981b) argued that the archaeological record is not a biased “transformation” of some static ethnographic past that needs to be reconstructed. Instead, archaeological data constitute phenomena realized through continuous, ongoing processes. According to Binford’s (1981b) position, archaeology’s primary goal should be to identify and explore these processes. One cannot simply overlay the ethnographic record onto the archaeological record, as there is a qualitative difference between the two data sets (Binford 1981a:197). Smith (1992) disagrees with Binford (1981a) and suggests that the main difference between the archaeological and ethnographic records is quantitative, not qualitative. Thus, for Smith, investigating change in the archaeological record is a simple function of observational scales: The continuous-versus-step issue [of change] is a question of scale and methods. In order to make comparisons between different points in time, periodization is required, because it is methodologically not possible to study “continuous change.”. . .The real issue is then the degree of refinement of the chronology employed .... The issue of continuous change is a red herring and the crucial question is how finely can we measure past time, or how refined can we make our chronologies? [1992:27–28] Smith’s vision for chronology theory is apparently an archaeology in which the temporal resolution of each prehistoric “frame” is increasingly refined. By stringing these frames together one could conceivably achieve a version of change that appears less staggered and more ethnographically lifelike. This position represents the quintessential reel-time archaeology in which the ultimate goal is to create an account of the archaeological record that is meaningful in ethnographic terms.

Several problems undermine this approach. First, there is no necessary reason why the archaeological and ethnographic records are subject to different temporal scales. Both data sets comprise cumulative phenomena that are addressed by a researcher in the “present.” Ethnographers do not witness some disembodied “slice of time”—embedded within their observations are the consequences of forces that led up to the moment of those observations (see discussion below). In terms of its temporality vis-à-vis the researcher, the ethnographic record may not be so different from the archaeological record. Rather, the main difference is that the ethnographic record exhibits cause-andeffect, dynamic relationships, whereas the archaeological record comprises static patterns. This is not the difference of temporal scales that Smith (1992) emphasizes; rather, it is a fundamental, qualitative difference, and addressing the two databases requires different analytical frameworks (e.g., Binford 1981b). A second implication of Smith’s (1992) discussion is that archaeology is best served by a temporal resolution that matches ethnographic experience. Nonetheless, it is unclear why ethnographic and archaeological accounts should be consonant. As noted above, ethnographic and archaeological data are different phenomena and are best addressed on their own terms. It would seem that archaeology’s traditional reliance on ethnographic concepts is the main culprit here. Finally, Smith’s (1992) discussion promotes the notion that archaeology should devote disproportionate attention to chronological refinement. This position is the real red herring. Even with the most sophisticated chronological controls, many regions of the world could not boast archaeological records readily divisible into units of a decade or a generation. And even if they could, so what? What are the research questions that will be answered with archaeological phases of 10 or 20 years? Surely research questions should drive analytical methods, rather than vice versa? Thus, reel-time archaeology is problematic on several grounds. Not only is the ability to achieve sufficient chronometric refinement suspect, but the implicit goal apparently derives from a ­debatable

163

Philip J. Arnold III logic that seeks similar knowledge from archaeological and ethnographic records. However, do we really want ethnographic answers to archaeological questions? If so, why do archaeology at all? The present is certainly relevant to archaeology, but its value lies primarily in providing the context to establish secure inferential reference points as a means to monitor archaeological variation. Elsewhere (Arnold 1999, 2003), I have argued that ethnoarchaeology is most productive when conducted as part of a middle-range research program (e.g., Binford 1981b). Such a program begins with an archaeological question and looks to the contemporary or historic use of material culture for insight. But such insight does not derive from simplistic analogical reasoning. Rather, the goal of this program is to identify cause-and-effect relationships in the present and use this understanding as a frame of reference for archaeological patterns. Just as using a yardstick does not presume that every­ thing to be measured will be 3 ft long, the body of referential data should not restrict past behavior to the documented present (e.g., Wobst 1978). Rather, ethnoarchaeology allows us to create behavioral indexes and use those points of reference to assess variation in past human activities.

Change

The second concept that merits our attention is “change.” Change weds together several important characteristics; first, change refers to significant variation in the attributes or properties of a given phenomenon. “Significant” is key here, as the nature of the research questions best determines what specific threshold is deemed significant. The phrase “a given phenomenon” is also important, as it is pointless to speak of change between two independent entities. Second, change is the inevitable and perpetual consequence of existence within the dimension we call time. As long as time exists, change exists; because time is ubiquitous from the human perspective (e.g., Munn 1992:94), change is ubiquitous in terms of that same perspective. Change is the concretization of the Second Law of Thermodynamics and thus represents the normal operation of our humanly perceived physical reality.1

Third, change implies the establishment of at least two fixed temporal points. However, given that change is perpetual, the determination of which temporal points to consider is discretionary and should also be a function of the research question. Darwinian archaeology represents one of the few schools of archaeological thought that explicitly treats change as a perpetual phenomenon. For example, Teltser (1995) cogently argues that all too often archaeological discussions emphasize differences rather than change, per se (see also Rice 1984). This substitution privileges the products of change over the process of change. Unfortunately, this valuable insight is often overshadowed by more zealous affirmations of selectionist archaeology. These advocates would subsume most relevant archaeological transformations as conforming to a Darwinian evolutionary model (e.g., O’Brien and Lyman 2002), thereby marginalizing other viable forms of change. Emergence and development are also biologically recognized examples of change, but these operate in a manner different from Darwinian selection (Hall 1998). Despite its emphasis on change as a process, much of archaeology’s selectionist agenda is hamstrung by an assertion that Darwinian evolution is the lens through which meaningful change should be examined. Finally, the process of archaeological change is often presented as one of becoming. To approach the archaeological record in this way should immediately raise a red flag and begs the question, “Becoming what?” Becoming is an awkward term—it implies that the outcome is already known or somehow preordained. As many scholars have noted, modeling change as a form of becoming encourages the “retrospective distortion” (Pinch and Bijker 1987:​28) that promotes such self-fulfilling prophesies as “Whig” history (e.g., Hinsley 1989:80), the “standard view of technology” (Pfaffenberger 1992:​ 493–​495), and other notions of necessary progress (e.g., Pavelka 2002). Rather than a process of becoming, change can be productively recast as a process of “begoing.” Although the final outcome may be clear as we look backward through time, it was certainly not known

164

No Time like the Present to those members of the culture looking forward into their own future. Change builds on thatwhich-came-before—it is not magically drawn toward that-which-has-yet-to-be. This position was articulated over a decade ago, with the observation that archaeologists “will have to stop looking back from their present position in time, trying to recognize the past patterns that are observed in the present. They will have to travel back in time and look forward with those whom they study” (van der Leeuw 1991:​13; emphasis in the original). Although cultural decisions are often predicated on expectations regarding future conditions, there is never a guarantee that such expectations will be realized. Thus, in many ways change represents a perpetual departure, not an ongoing arrival. In sum, a more temporally aware archaeology encourages us to reconsider two hallmarks of conventional anthropology. First, archaeologists should understand how the reel-time view of the past affects their interpretations. Such a view suggests that the outcome of archaeological research should be an ethnographic-friendly body of information. I have suggested that such a desire is misplaced; archaeological research requires archaeological questions and archaeological concepts, not ethnographic questions framed in simplistic archaeological terms. Second, change should be approached as the normal state of our experiential reality; change, therefore, is appropriately viewed as the independent variable in archaeology’s temporal equation. But the constancy of change should not overshadow the fact that change has many different dimensions besides the obvious temporal one. Working toward an understanding of these dimensions is further explored below.

Actualistic Research and Inferential Indexes Actualistic research uses observations made in the here and now as guides for understanding unobserved dynamics in the there and then. This process might at first appear paradoxical, causing some scholars to suggest that “ethnoarchaeological studies are of interest in their own right but they cannot contribute directly to our understanding of the

past” (Hodder 1987:​424). From this perspective, if behavior is contextual and contingent, then we cannot use the present to establish some unknown context in the past. Creating a picture of the past via the present would be at best reductionistic and at worst imperialistic. The linchpin in these arguments, however, is the often-implicit expectations of what the “picture of the past” is supposed to reveal. As noted above, for many researchers this picture is apparently something like an ethnographic vignette, albeit less fleshed out than the ethnographic version (e.g., Bender 2002; Paynter 2002). This view is nicely illustrated by comments regarding a recent collection of essays on anthropology and timekeeping, with one archaeological commentator happily observing that “historical particularism is back, with a narrative twist” (Lucero 2002:S124). Rather than force the ethnographic present onto the time frame of the past, we are better served using the ethnographic present to establish inferential points of reference. These reference points serve as secure knowns for the purpose of identifying and evaluating variation or deviation from expectations.

Temporal Logics

Nonetheless, some researchers may question the degree to which the ethnographic present can be made consonant with the archaeological past. Doubts regarding the archaeological relevance of ethnographic observations are not new and come in two basic forms (e.g., Ascher 1962; Wylie 1985). First, one can argue that a given group does not provide an appropriate direct historical analogue— in other words, there is no clear evidentiary chain linking an ethnographic group to the archaeological record. Without a clear justification to link the source to the subject, an argued connection between the two will have little credibility. Second, one can argue that the ethnographic time frame is too short to be archaeologically relevant. This second point presents the ethnographic record as a “thin event horizon” and is usually framed in one of two ways: either the temporal natures of ethnographic and archaeological observations are so completely incongruent that

165

Philip J. Arnold III c­ omparisons are summarily dismissed or the properties of the ethnographic context are somehow incomplete given the relatively small temporal window of ethnographic observation. The issue of direct historical analogy is intriguing but suggests that the main goal is the simple projection of the ethnographic case into the past. And although there may be instances in which such a projection is useful, the end result is that archaeologists learn scarcely more than they started with (e.g., Bailey 1983; Wobst 1978). Thus, the direct historical approach has minimal utility within the context of identifying and explaining novel (i.e., undocumented) human activities in the past. Arguments regarding the archaeological relevance of ethnographic observations are more pertinent to our discussion and potentially more devastating. Construing the ethnographic record as a thin event horizon is problematic on several grounds (Arnold 1999). As noted above, the temporal logics of archaeological and ethnographic observations need not be so different. Both data sets are engaged in the present day, and both data sets represent the cumulative effects of long-term activities. The difference, of course, is that cause and effect can be monitored in the latter context but not the former. Finally, the proposal that ethnographic studies fail to reflect the full range of behavior begs the issue of what constitutes the “full range.” Such claims are essentially misplaced sampling arguments. Although ethnographic fieldwork is usually conducted over a prescribed time period, it would be erroneous to misconstrue the results of this activity as representing a discrete slice of time. Rather, the intersection of observer and observed reveals the playing out of short- and long-term processes that have led up to the observations. These processes operated before the ethnographic observer arrived and will continue after the researcher leaves. In other words, the ethnographer is somewhat arbitrarily inserted and extracted from a process of ongoing perpetual change. This fortuitous intersection of the ethnographer and the behavioral record is not unlike that of the archaeologist and the material record. First, it

bears repeating that both archaeological and ethnographic observations address contemporary phenomena. Second, both kinds of studies document patterns determined by short- and long-term processes in effect prior to their respective encounters. Of course, in the archaeological case the retrieval of data usually destroys the relational attributes of the material record, whereas the impact of ethnographic study may not be as great (although it certainly can be). Regardless, the important distinction between the two realms of investigation is not some presumed significant temporal difference in their respective observational activities. Were this assumption to hold, it would be almost impossible to justify any use of the present to inform the past. Confusion over the temporal logic of ethnoarchaeological fieldwork is evident in recent discussions that would necessarily favor longitudinal projects over short-term field observations (see comments in Arnold 2000). Here we see the inverse of the reel-time archaeology noted above—if we can only extend the length of actualistic studies, we might create an ethnographic picture that is more consonant with archaeological time spans (e.g., Stark 2003). In either case the goal is the same; it represents a desire to make archaeological and ethnographic records temporally isomorphic. The process by which research questions are developed is of central importance here (e.g., Binford 2001b). Many contemporary ethnoarchaeological studies are essentially undertaken backward; they first identify some ethnographic “cause” and then attempt to envision how the archaeological “effect” might look in the material record (David and Kramer 2001). Given that no archaeological question per se drives their studies, researchers apparently feel compelled to offer observational duration to justify the fit between the present and the past. Thus, Stark (2003:217) has recently argued that longitudinal ethnoarchaeological research best matches the temporal contours of the archaeological record. Of course, a more productive way to ­conduct archaeologically relevant actualistic research is with an archaeological question that serves as the point of departure. Actualistic study is thus ­targeted

166

No Time like the Present t­ oward an appropriate temporal ethnographic context, be it short or long term. Again, this is not a “one-size-fits-all” situation. What you want to know should determine how you go about collecting relevant data, rather than collecting data first and then hoping to fit it to some future archaeological question.

Exploring Types of Change

As noted above, the reel-time archaeology approach tends to focus on rates of change (e.g., Smith’s “continuous-versus-step issue” [1992:27]) rather than types of change. And yet the identification and exploration of different types of change make up precisely the kind of contribution that ethnoarchaeology is positioned to make. Below I present two examples: the first involves the process of what I call supplanting versus supplementing, and the second compares intensification and extensification. Both comparisons represent the value and relevance of contemporary behavior for archaeological studies of change. Supplanting is a familiar form of change; it refers to the replacement of one category of things with another within a given inventory or tool kit. In this case the new category assumes the function of the previous category. For example, Skibo (1994) discusses the replacement of clay cooking pots with metal pots among the Kalinga of the northern Philippines. This process of supplanting occurred under specific circumstances in which wage labor became increasingly common within the region. Metal containers became a symbol of wealth, in addition to their more utilitarian benefits (Skibo 1994:​124; Trostel 1994). However, metal items did not replace all containers; metal was only acceptable for rice-cooking vessels. Thus, the process of supplanting depended on the specific characteristics of cooking technology and dietary values of the group and did not reflect a wholesale change as a result of a new local economy. Supplementing, in contrast, refers to the expansion of categories within an inventory. In this case, differences in the formal or material quality of items heralded additions to the inventory or tool kit. Adding to or augmenting an inventory

increases the richness component of a diversity measure, even if the evenness is not dramatically affected. It is worth noting that supplementing may be confused with mere stylistic elaboration (e.g., Dunnell 1978) if the functional role of the addition is ambiguous or unknown (e.g., Binford 1989). Deal (1998) provides a good example of the process of supplementing. Households in Highland Guatemala are adopting an additional form of the traditional tortilla griddle (comal). This new form of griddle is made of metal, which means that it takes less time to heat up the tortillas. Unlike the case in the situation noted above, however, the metal comal is not taking the place of the conventional clay version (Deal 1998:91). Rather, much like the modern microwave is used to expedite food preparation, the metal comal is usually employed to reheat tortillas. Of course, most U.S. households that use microwaves still retain a larger stove, and so it is that households within the Highland Guatemalan communities retain their clay griddles. Moreover, the presence of metal comales does not directly relate to economic ranking; rather, it would seem that the use of these griddles reflects different labor organizations and activity scheduling at the household level (Deal 1998:21–22). A second example of how archaeology may benefit from ethnographic information is visible in Binford’s (2001a) exhaustive synthesis of huntergatherer data. Binford (2001a:346–347) makes the interesting distinction between systems that are intensifying and those undergoing extensification. The former is a process whereby a growing population is forced to obtain resources from a shrinking region. The result is an increase in the net amount of food extracted from a given area. Intensification may also lead to changes in social hierarchies and increasing sociocultural complexity. Extensification, in contrast, refers to the creation of new exploitable niches through technological developments or adoptions. Included among these developments are transportation innovations, such as sleds and boats, as well as the domestication or incorporation of pack animals, such as reindeer, dogs, and horses. Systems that are changing through extensification may also become

167

Philip J. Arnold III i­ ncreasingly complex, but the mechanisms in play are different than those revealed within intensifying contexts. The archaeological value of the supplanting/ supplementing and the intensification/extensification distinctions is that these forms of change can leave particular material signatures within an archaeological assemblage. Thus, there is an opportunity in the ethnographic present to not only document the types of change but also discern the relevant material patterns associated with each. Importantly, Binford (2001a:347) highlights the problematic manner in which extensification is often dismissed as a “historically” produced phenomenon, rather than treating it as a legitimate process of change in its own right. A similar point may also be raised regarding the adoption of metal containers among pottery-using groups. Ethnoarchaeologists may decry such “modernization” and even use it as justification to hurry out and record “traditional” practices before they are gone for good. Nonetheless, these situations offer a genuine opportunity to document the kind of change that is taking place and explore the context of that change. Blaming “historical” circumstances for cultural transformation simply reaffirms a focus on rates of change over types of change and promotes a reeltime view of the past. One final point bears repeating. It would be a mistake to treat these concepts simply as ethnographic solutions in search of archaeological problems. Rather, these constructs are best used as archaeological reference standards to identify and assess variation in the material record. In other words, the goal is not to discover a “supplanting culture” or an “extensifying culture” archaeologically; instead, these concepts should be used to provide a fixed point of reference, a secure backdrop, against which meaningful variation can be monitored. As noted at the beginning of this chapter, “meaningful variation” is always constituted in the context of a particular research question or problem.

Conclusion This discussion has focused on some of the important temporal distinctions between ethnographic and archaeological studies. On one hand, research-

ers have argued that the present provides an incomplete account of the past; thus, invoking the former invariably limits our ability to understand the latter. On the other hand, scholars note the difficulty in calibrating ethnographic and archaeological time spans. And the more vigorously ethnoarchaeology tries to wriggle free, the more firmly it becomes impaled on the horns of this recalcitrant beast. Nonetheless, there is a potential solution to this predicament. A close inspection of the above dilemma reveals an archaeological practice that owes allegiance to a reel-time version of the past. Such a view seeks to generate prehistoric ethnographies and string them together through time. Refining the length of each vignette supposedly improves the resolution of the film. Such temporal refinement has become a focus for building chronology theory in archaeology. I have suggested that this reel-time archaeology is misplaced. In its stead I would advocate a renewed concern with the processes of change. Change is a normal, continuous phenomenon; it is the rule, not the exception. Accordingly, culture can be recast as a collective response to change. Archaeological attention, therefore, would go toward documenting the manifestations of these responses throughout the material record. But how do we give meaning to these manifestations? Patterns are made meaningful in terms of a particular question and an appropriate index or frame of reference. Ethnoarchaeology best serves archaeology by establishing these reference points. Archaeology, in turn, uses these intellectual datum points to assess variation in the material record of ancient activities. In sum, archaeology should renew its concentration on the process of change. If things did not change, there would be little reason to conduct archaeological research. Everything that has taken place would exist within the present. Our discipline would be moot. But to advance this goal we must reflect on what we mean by change and how our view of culture fits within that definition. We must also pay close attention to our logic and methods of inferences. This kind of realignment is difficult, slow, and at times confusing. Nonetheless, if ­archaeology is to

168

No Time like the Present realize its oft-stated charge of investigating longterm change, then we need to move forward. After all, there is no time like the present.

Acknowledgments Several of these ideas have been knocking around in my head for a while. I would like to thank Jim Calcagno and Chris Fry for their conversation while these ideas were in nascent form. I would also like to thank LuAnn Wandsnider and Simon Holdaway for their invitation to participate in the 2003 Society for American Archaeology (SAA) symposium that gave rise to this chapter. Comments by the participants at that symposium, both invitees and audience members, were extremely valuable. As always, the intellectual and editorial input of Shannon Fie is gratefully acknowledged. Finally, I would like to

dedicate this effort to Susan Kent, who passed away on the final day of those 2003 SAA meetings. Susan was a good friend and an excellent ethnoarchaeologist who put her ethnographic observations into archaeological practice. She set an admirable example that all of us would do well to follow.

Note 1. Critics might argue that the Second Law of Thermodynamics was initially developed to describe activity within a closed equilibrium system and is thus improperly extended to the open system that better represents cultural behavior. Nonetheless, recent advances in the analysis of nonequilibrium thermodynamic systems show that the Second Law of Thermodynamics is particularly applicable to macroscopic processes, such as cultural systems (e.g., Adams 1988; Kneebone 1990; Prigogine and Stengers 1984).

169

11

Paradigms and Metaphysics, or “Is This the End of Archaeology as We Know It?” Tim Murray La Trobe University Chaos, Cosmos! Cosmos, Chaos! Who can tell how all will end? Read the wide world’s annals, you, and take their wisdom for your friend. Hope the best, but hold the Present fatal daughter of the Past, Shape your heart to front the hour, but dream not that the hour will last. —Alfred Lord Tennyson, Locksley Hall Sixty Years After (1886)

Paradigms and Metaphysics The human fascination with time is unremarkable because it is so commonplace. Notwithstanding some spectacular poetry by such as Tennyson and, more recently, T. S. Eliot (e.g., Burnt Norton [1941]), and some even more spectacular cosmology, the contemplation of time—of process, creation, transcendence, and, of course, memory—is central to human identity. Archaeologists are often reminded that time—its quantification and ordering—​is fundamental to the proper practice of our discipline. How else can trajectory, process, contemporaneity, or succession, indeed history, be established? Even now, when we are right to question whether an archaeology expressing concepts, categories, and methodologies that are generally accepted actually exists, stories about how archaeologists have come to understand and control time (be it the Three Age System, the discovery of high human antiquity, the Midwest Taxonomic ­System, or the effects of radiometric dating) play a significant role in identity formation among budding practitioners. Without time there can be no

­ istory; equally there can be no prehistory or arh chaeology. It is significant that there is more to these stories than a celebration of our capacity to measure time (chronometry). At a superficial level this might simply be a recognition of a developing capacity to determine duration and succession (a calculation of the direction of Time’s Arrow and one aspect of the “speed” of its flight). For example, the link between the law of superposition and the use of patterned distributions of material culture as “fossil markers” was fundamental to the development of culture history in the nineteenth century. Very much more difficult issues arise when we seek to understand the ways in which time (and the playing out of complex formation processes over time) can structure archaeological phenomena—literally the things archaeologists observe and seek to understand and explain. This I take to be a fundamentally ontological issue, raising questions about the nature of archaeological data and the possibility that achieving an understanding of their meanings may, in some contexts at least, require a ­rethinking

170

Paradigms and Metaphysics of conventional relationships with cognate disciplines. For this reason it is significant that these aspects of time stories (of archaeologists’ relationship with time) are not widely discussed or understood. Whereas philosophers of history regularly debate matters that go to the core of their discipline, those few who pursue philosophical discussion in archaeology have tended to focus their attention on epistemological matters—generally either to support or to restrict the application of different research agendas to the discipline. The larger context of my comments in this final chapter flows from the argument that one of the most effective means of promoting discussion of archaeological epistemology and, more fundamental still, of archaeological ontology is to consider the relationship between time and the constitution of archaeological phenomena (thereby defining the categories and predicates of “archaeological being,” which are fundamentally ontological concerns). In this sense my object is to very briefly explore the role time has played (and might yet play) in creating what has been called “the archaeological perspective.” This object can, in part, be expressed as a question: Do the kinds of phenomena routinely observed and analyzed by archaeologists constitute a domain (or even domains) of categories and predicates that are distinctive (in the sense that they are different) from those at the heart of cognate disciplines such as anthropology and history? Expounding the virtues of exploring disciplinary ontology might seem hopelessly naive at a time when archaeology has become so diverse in its interests, perspectives, and data sets that no agreement about what constitutes the ontology of archaeology is likely to be possible. In this sense distinctiveness can be internal to archaeology, as well as external. For example, a consideration of time as a structuring principle might be more significant when dealing with periods before recorded history or where it is not feasible to use oral histories or to deploy direct historical analogy in reasoning. Indeed, notwithstanding the fact that archaeologists who have an interest in the philosophy of science might accept the notion that there is no timeless, universal “scientific method” and that each disci171

pline or research area essentially sorts these things out through “doing” science at any point in time, there is much less agreement about whether or not the data or perspectives of our discipline are more than methodologically distinct from those of disciplines such as anthropology or history. On this basis it might be argued that there is (and can be) no distinctive archaeological ontology and that devoting time and resources to search for it is a wasteful exercise doomed to failure and entirely at odds with the ways in which human beings are to be “properly” studied. In connection with this point, it may be of some significance that, notwithstanding the near continual turbulence in theoretical archaeology since the 1960s, there have been very few detailed discussions of disciplinary metaphysics, although Clarke (1968) and Meltzer (1979) provide notable exceptions. If one takes a different view, however, about the distinctiveness of archaeological epistemology and (perhaps more significantly) the distinctiveness of disciplinary data sets from those routinely explored (and made sense of by social/cultural anthropologists or historians using social theory), then a consideration of a distinctive archaeological ontology is more than justified—if only to emphasize that such theories may well require substantial transformation in order to capture the different essences of archaeological phenomena. Furthermore, although it is highly unlikely (and probably highly undesirable) that a clearer understanding of the fundamentals of our discipline (or some other abstract notion such as “the archaeological perspective”) should lead to a single reading of archaeology or of archaeological ontology, such an understanding would at least have the benefit of helping us to “reconnect” archaeological phenomena to the substance of general inquiries into human beings (given that changes in the way we perceive one should lead to changes in the other). Of course, we are a very long way off from gaining general agreement that either claim has been sustained. Nonetheless, as the editorial introduction and the contribution by Bailey to this collection demonstrate, such claims rest on a perceived need to take time (and its role in creating the phenomena archaeologists study) much more seriously than has previously been the case.

Tim Murray

Taking Empirical and Theoretical Time Seriously Bailey’s contribution to this volume (part personal history, part recognition of the significant challenges palimpsests pose to the conventions of archaeological interpretation and explanation) provides an excellent summary of the forces—theoretical and empirical—that are fundamental to his reading of time perspectivism. Bailey very usefully describes the milieu from which his initial formulation of time perspectivism sprang. Discussing the importance of uniformitarian assumptions, and stressing that temporal scale was crucially important to exploring what David Clarke thought of as the “possibility space” of archaeology, threw into sharp relief the widely differing objectives of prehistoric archaeologists (even in Cambridge) during the 1980s. Indeed, the connections and contrasts between Bailey’s interests and those of Foley and Binford, two other major contributors to the development of interpretations that sought to integrate the temporality of archaeological records in a more systematic fashion, also demonstrate that even at its birth time perspectivism could be a very broad church (see, e.g., Fletcher 1992). Other contributors to this book make such a breadth of view and focus even more apparent, ranging from the microscale analysis of places and sites (Wandsnider, Dooley, Holdaway et al., and Stern) to more general considerations of assemblage composition (Shott) and of theories or approaches that might be borrowed from disciplines ostensibly confronting similar issues (see especially the contribution from de Lange). For me the problems encountered in this first phase of time perspectivism flowed from two sources that were at once distinct and related. The first source was what I have referred to as the disciplinary traditions of archaeology—those elements of disciplinary “culture” that were so deeply ingrained that they held the status of “givens” rather than potentially contentious assumptions about nature and purpose. Central to this “culture” was (and is) the primacy of sociocultural anthropology both as source of theory and as arbiter of what it was meaningful and valuable to know about the archaeological past—an argument only recently

recognized by Lyman (2006). The second source was a clear incapacity for time perspectivism, as originally formulated, to operationalize Schiffer’s (1987) insights into site formation processes and Binford’s (1981a) highly significant observations concerning the “Pompeii premise” (Murray 1999a; Schiffer 1985). I have written extensively about both matters, so there is no need to rehash core arguments here, save to observe that initial conceptualizations of time perspectivism as being mostly about the implications of different temporalities for the ways we posed and answered archaeological questions were somewhat naive. The need to articulate time perspectivism in archaeology has required the development of both theory and method. The introduction of new concepts such as the structural properties of archaeological records and the “normalization” of interpretation as well as a new approach to archaeological metaphysics are cases in point (Murray 2001; see also Cosgrove 1995; Stern 1993, 1994a; Williamson 2004). In this chapter I will extend this background somewhat further by identifying some of the serious questions that confront us and then relate these to the themes and issues that have surfaced in the contributions to Time in Archaeology. In doing so I will make a few general points about my own understanding of the relationship between time and the empirical and theoretical in archaeology, and I will very briefly outline the impact that a developing understanding of the phenomenology of archaeological records might have on the discipline of archaeology. The issue of theory building, especially of ­middle-​range theory building, remains a powerful tool for diagnosing the current state of theoretical archaeology. It allows us to explore the sources of theory, the nature of their application to archaeological phenomena, the consequences of such applications for the original source disciplines (or perspectives), and of course the security of the knowledge we claim from such applications. The ways in which archaeologists deploy theories to make sense of empirical phenomena, and the ways in which they make and defend the knowledge that is claimed as a result, also tell us much about the

172

Paradigms and Metaphysics c­ ohesiveness of archaeology as a discipline and the relationships between archaeology as a discipline and other disciplines commonly engaged in the search for an understanding of human beings. Working at this level we can establish that the core issue here is not about time per se but about the consequences of changing characterizations of the phenomenology of archaeological records, both for archaeological theory and for the structures that currently regulate the terms of an archaeological contribution to knowledge about humans. One of the major causes of my interest in the role of time in archaeology stems largely from a more abstract fascination with the potential power of archaeological records to disorder our imaginations, thereby posing real challenges to human ingenuity and giving us a vantage point from which to begin a deconstruction of the terms of human self-perception. Here I refer to ontological matters—time scale and spatial scale, issues of entropy, problems of causality, and (most significant) the vexing task of developing a conceptual framework for archaeology that is both meaningful and scientifically defensible. Although I have previously discussed the considerable consequences of the discovery of “deep time” in the nineteenth century (e.g., Murray 1993, 2001), it is the quantification of archaeological time (especially with the introduction of a variety of radiometric chronologies), and its link to a more complete understanding of archaeological formation processes, that challenges the conventions and “taken for granteds” of archaeological practice. Many of these conventions (or traditions) were framed over a century ago. Although they have transformed to suit changing fashions in social theory, and to accommodate new sources of information deriving from the rise of world archaeology and the application of scientific analysis to new sources of archaeological data, there has been no change in the metaphysics of archaeology. As Meltzer trenchantly observed 25 years ago in the course of demolishing the paradigm status that had been claimed for the “New Archaeology”: Ultimately, there is very little of the New Archaeology that cannot fit the same linear continuum with the Old Archaeology. The limits

of the continuum are defined by its metaphysic; the contents of the continuum are characterized by theoretical, methodological and technological variability. It is only when a new metaphysic is introduced, one which by definition cannot fit on the same continuum, that a revolution results .... There has been no revolution in archaeology. [1979:654] Of course, it is beyond the scope of this brief essay to satisfactorily sustain a claim that other forms of archaeology (such as the postprocessual archaeology of the 1980s and 1990s or the evolutionary archaeology of more recent times) have also not ushered in changes to disciplinary metaphysics. Nonetheless, there is ample evidence that both postprocessual archaeology and evolutionary archaeology strongly subscribe to what Meltzer called the “archaeology as anthropology metaphysic” (1979:​ 654), although their various forms were all significant departures from the functionalism and behaviorism of the New Archaeology. The persistence of disciplinary metaphysics, even when there have been such great changes in our comprehension of the phenomenology of constituent data, thus becomes a significant issue in the philosophy of archaeology (e.g., Murray 1987). Contemporary archaeological theory, or indeed accounts of the nature and purpose of archaeological knowledge, has not engaged with our changing understanding of the phenomenology of archaeological records. In fact, the opposite has been the case, with much “archaeological theory” being borrowed from sources that (notwithstanding their possible effectiveness in helping us understand our lives in the present) comprise quite different types of information than archaeologists (especially prehistoric archaeologists) now more generally understand that they work with. Such traditional sources of borrowings more closely approximate traditional understandings of archaeological records. We should, therefore, expect that the traditional plausibility of interpretations and explanations that result from the application of such borrowings to archaeology should now also be reassessed. Change in these key areas of practice has also not occurred. I have discussed the ­consequences of this

173

Tim Murray “­dependant” relationship elsewhere (Murray 1987, 1999a, 2001), but Meltzer’s linkage of the dominant disciplinary metaphysic with the lack of archaeological theory applies equally well to processual, postprocessual, and evolutionary archaeologies (see also Murray 2002, 2006). Several reasons, all of which stem from the disciplinary “culture” of archaeology, might go some way toward explaining this situation. Though it can be argued that these apply to all aspects of the archaeological enterprise, they strike particularly hard in prehistoric archaeology: • that archaeological records are widely thought of as being especially impoverished; • that change and variability have traditionally had meaning only through the teleology of social evolution; • that ethnography has always provided the exemplar of what constitutes the goal of any reconstruction or account of prehistoric human action; • that archaeologists, to convincingly explain change and variability, have generally attempted to reconstruct precise causal links between beginning and end states, with any failure generally put down to poor data or, indeed, the improbability of there ever being clear links between behavior and material consequences; and • that the archaeologist places the weight of conviction or plausibility on statements that are primarily generated from contemporary social theory​—​problems with application thus become the fault of the data rather than the ­theory. These reasons also reflect a constant theme in the history of our discipline—that of the “normalization” or “humanization” of archaeological information​—​the process whereby the ontology of archaeological records (as distinctive records of ­human action) is lost in favor of intelligibility in terms of the conventions of social theory (Murray 1997, 2001). The link between this theme and the notion of an impoverished (and theoretically inconsequential) record can be seen to sustain a disciplinary culture that creates archaeological knowledge

of kinds that might be radically at odds with the structural properties of the phenomena we are fundamentally seeking to comprehend. Given this kind of historical interaction, I have for some time now argued that the inertia of tradition is very strong, both within and without the discipline. But the proposition of time perspectivism by Bailey (see especially 1983) provided a standpoint from which to argue that there is no necessary reason why such traditions need to be continued unchanged or unchallenged, and indeed it offered the possibility of archaeologies that could make novel and intellectually respectable contributions to anthropology. The only potential downside in this process was (and remains so) that the “anthropology” in prospect was likely to be different to the one that had underwritten over 150 years of social archaeology. There was also the sense that the problems that could be explored through time perspectivism, especially the links between scale and question, or the investigation of the structural properties of archaeological records and their links to traditional questions, supported an argument that theoretical debate had been prematurely consigned into fixed blocks of processual and postprocessual archaeology. Observing the career of time perspectivism after 20 years is a salutary experience. Most obviously it remains largely unknown in archaeology. For example, most published discussions about the nature of time focus on social and cultural anthropological discourse about the importance of recovering the lived time of the actors—or at the very least couching interpretation in that frame (e.g., Gell 1992; Gosden 1994; Ingold 1993; Thomas 1996). Although there has been continuing discussion about chronology and the pluses and minuses of various radiometric dating systems, with few exceptions (e.g., the chapter by Holdaway et al. in this book) there has been a focus on the technology of dating rather than its use by archaeologists to describe and define phenomena (two core archaeological objectives). Over the same period there have been quite frequent discussions about what Schiffer has termed the formation processes of the archaeological record, but again few discussions of the implications of these insights for

174

Paradigms and Metaphysics our capacity to meet core objectives (obvious exceptions being the contributions by Sullivan, this volume; and Stern 1993, 1994a, this volume) have seriously entered the realm of theoretical discourse in archaeology. All evidence indicates that time perspectivism is ­peripheral to contemporary archaeological ­theory. This peripheral status continues despite the potential of time perspectivism to promote new visions of the nature of archaeological data, new ­relationships between archaeological knowledge and other sources of knowledge about human beings, a reduction in the level of logical circularity (which is so much a feature of archaeological reasoning), and the development of archaeological theories that connect more directly and more convincingly to the phenomena they are supposed to explain. But is the peripheral place of time perspectivism more a matter of because of rather than despite? The idea that archaeologists are unlikely to abandon traditional disciplinary culture for something that is still somewhat abstract and undefined should be entirely uncontroversial. Similarly, the notion exists that time perspectivism might promote a damaging hyper-skepticism about the security of archaeological knowledge claims—thereby using “spoiler” arguments to undermine valuable (and time-​honored­) interpretive strategies. But surely these are unnecessarily reactionary responses to the argument that to raise our expectations about the success of such strategies, to reduce logical circularity in archaeological reasoning, to demonstrate the lack of connection between theory and phenomena, and to propose pathways whereby these connections may be strengthened are all good and positive things for archaeology? Some years ago, in very different circumstances, Michael Schiffer discussed a tendency to close down debate before too much of the “core” of archaeology became the subject of close ­scrutiny:

tation may be deceiving. Complacency, too, is evident in many quarters, and it is symptomatic of a profound neglect of many unresolved epistemological, theoretical, and methodological issues of the past decade. Complacency represents premature closure of discussion and debate on a number of important topics .... A period of normal science...has emerged in archaeology. As in many normal science periods, basic concepts and principles are beginning to submerge into the murky inaccessible depths of the discipline, where they no longer can be easily questioned or challenged—until of course the next major paradigm clash .... It is time to step back from the calm of the current archaeological scene and ask if the paradigm under which we now labor, ushered in by the new archaeology of the 1960s, is fully adequate for genuine scientific studies of the past. [1976:​2–3] The adoption of a time perspectivist agenda inevitably leads to a scrutiny of the intersection of contemporary theory with new understandings of the nature of archaeological records (and through this of archaeological ontology). Writing in response to the challenge to archaeologists made by Schiffer, Lewis Binford observed:

At the present time, when the paradigm of the new archaeology is becoming consolidated and even entrenched, I sense in the discipline both ferment and complacency. . . . The observable ferment may be taken to mean that archaeology is in a healthy state. However, such an interpre175

The assumptions made regarding the conditions under which the archaeological record was formed directly condition the character of inferences about the contents of the archaeological record .... [W]e may be frequently incorrect or at least highly uncertain about our reconstruction of the past. The challenge to archaeologists is simply this: How do we proceed? How do we unify the world of archaeological things with our ideas as to the character of the past? How may we use the empirical world of archaeological phenomena to stimulate ideas about the past and at the same time use these empirical experiences to evaluate the resulting ideas? How can we proceed so as to develop confidence that our ideas of the past are informative about the actual past? We face the challenge of science itself

Tim Murray how to keep our feet on the “empirical” ground and our heads in the “theoretical” sky. Basic to the development of a science is a recognition of the domain to which scientific procedures might be profitably addressed—empirical with respect to what? Theoretical with respect to what? [1983c:​411] Here Binford’s notion that archaeologists need to develop strategies for controlling the a priori in contemporary archaeological theory, and that connecting phenomena and theory through the development of transparent methodologies, can be directly contrasted to the contextual or postprocessual archaeologies that were to follow. For Binford the empirical was consequential, but for Hodder (e.g., 1985) and others (see especially Miller 1985; Tilley 1981a, 1981b, 1982) the idea that the constitution of archaeological phenomena should constrain the application of contemporary social theory to archaeology was precisely the problem. For them an archaeology that does not have people as active agents, that accepts a split between function and meaning, process and norm, system and culture, subject and object, fact and theory, and process and history, dehumanizes the past. In Miller’s opinion the inherent shortcomings sourced to the traditions of the disciplinary culture of archaeology were to be overcome by the development of material culture studies that are concerned with all aspects of the relationship between the material and the social, and are not determined by the logistical constraints of any particular discipline. The aim is to achieve a model capable of representing the complex nature of the interaction between social strategy and artefactual variability and change. It is inevitable that some of this sophistication should be lost with the formation of the archaeological record, but this loss should be regarded as such rather than minimised by starting with a social theory more compatible with the paucity of the evidence. [1985:4] The very purpose for doing archaeology was to become a crucial point of contention. Miller’s view

of the primacy of theoretical over empirical context was a natural outgrowth of a conflict between his belief that any other type of account of human action lacked “meaning” and what he clearly recognized as being the empirical limitations of archaeological records. On other occasions Miller also remarked that for accounts of past human action to be “meaningful” social theory must come first—​to be later modified at a “secondary” level to “accommodate specific problems of archaeological methodology” (1985:4). The rejection of the idea that the constitution of archaeological phenomena could have anything more than methodological consequences for the discipline was further strengthened by a more general argument made by Tilley, that the broader context of human studies should set the agenda for archaeology: Failure to tackle problems within sociology and philosophy can only lead to a blind, unsystematic, groping towards an understanding of the past. It is sheer dogmatism to suggest otherwise, to suggest that problems within philosophy and social theory can be neatly circumvented in the practical business of carrying out research. In tackling these problems archaeologists can, themselves, contribute towards a wider understanding of social form and social dynamics. [1982:36] Over two decades later, it is worth reflecting on whether archaeology has made such a contribution (as distinct from the development of material culture theory, which has tended to draw much more on the analysis of the historical past and the present). Again, it is well beyond the scope of this essay to sustain an argument that the flow of influence and understanding has been much more to archaeology than from it, but it is worth reflecting about how much the theories and perspectives that have been derived from social theory have been modified in the ways stipulated by Miller and others. For me, the absence of evidence for even a methodological transformation of such theories (let alone the application of such transformed instruments to their source disciplines) indicates that a serious engagement with archaeological ­phenomena, through

176

Paradigms and Metaphysics a program of recursive reasoning, remains an unfulfilled promise. Naturally this raises the question of whether archaeologists have been making a contribution “towards a wider understanding of social form and social dynamics” that is recognized outside our discipline (again with the possible exception of material culture theory). It also raises the question of whether transformations are difficult to achieve without rethinking the fundamentals of such theories, a process that is highly unlikely to make practitioners of source disciplines receptive to the work of archaeologists. Of course, there is a more basic question still, and this relates to whether the primary purpose of such activity is to circumvent the “threat of the past” (Murray 1993) by extending the hegemony of the present, rather than seeking new understandings that incorporate the evidence of archaeological phenomena. The peripheral status of time perspectivism in contemporary theoretical archaeology is thus clearly at odds with even the minimalist (i.e., “methodological”) recognition that archaeologists have to work at building connections between the empirical and the theoretical. Whether this is Miller’s “methodology” or Binford’s middle-range theory is somewhat beside the point, as either divergent reading of anthropology recognizes that past and present are different (notwithstanding the “presentness” of pasts or the “pastness” of presents). It is significant that neither processual nor postprocessual archaeology fully appreciated that connecting the empirical and the theoretical through “methodology” or middle-range theory would pose tremendous challenges to practitioners. Again, some two decades later, consider Binford’s optimistic analysis of what lay in store: Archaeology is perhaps in a fortunate position. Although there is much contemporary “culture” or paradigmatic bias regarding the nature of man or the causes of history, there is very little folk knowledge regarding the formation of the archaeological record. This means that there is little explicit prior development of cognitive devices and frames of reference for accommodating archaeological phenomena in the literate static sense of the word. For fur-

ther developments in archaeology, the growth of a paradigm, developing cognitive means for identifying properties of the past or diagnosing the archaeological record and thereby giving meaning to the archaeological record, is crucial. [1983c:415] Binford had not counted on the strength of disciplinary cultural norms (some of which he overtly accepted) that could effectively deflect or delay consideration of these vital issues while at the same time fostering the continuation of perspectives that were framed in quite different terms.

Moving Forward The contributions to Time in Archaeology foster a more detailed understanding of several core elements of time perspectivism. The first of these has to do with time and the creation of archaeological phenomena; the second is a focus on history and anthropology; and the third relates to what might be interpreted as “grammatical” issues in archae­ ology. The chapters directly concerned with exploring time scales and temporalities (specifically those by Bailey, Wandsnider, Stern, and de Lange) seek to explicate how the measurement of archaeological time is a critical determinant of the structural properties of archaeological phenomena (the minimum chronological unit being one such outcome). These properties might range from the apprehension of palimpsests (especially the contribution by Bailey), via a consideration of the constitution of archaeological assemblages (see especially Sullivan and Shott), through to the detailed consideration of site histories—​incorporating concepts such as “persistent places” (see especially the contributions from Dooley and Holdaway et al.). The chapters concentrating on specific sites and places provide excellent examples of the complexity of formation processes and the growing perception (first really discussed in Stern 1993) that the hominid behaviors thought to be found in archaeological records of that era were anything but the subject of commonsense propositions. Perceiving complexity and potentially different kinds of patterning is one thing; developing plausible accounts of past human action is another.

177

Tim Murray Here the contributors are right to explore anomalies and to identify interpretive puzzles and problems that are not readily solved by contemporary theory. Behind much of this discussion is a very serious question: What behaviors are archaeological phenomena records of ? This is the approach taken by Holdaway, Fanning, and Rhodes and by Wandsnider, Dooley, and Stern as they work the theoretical interstices between geomorphology and prehistoric archaeology to discuss the consequences of time-averaged deposits (which it would now seem are very much more common than previously credited). Here the link between a changed understanding of the structural properties of such records and a perception that previous strategies for assessing the plausibility of knowledge claims made about them can be most effectively demonstrated. These examples also allow us to understand that no magic body of theory to make sense of all this complexity currently exists (especially in contemporary social theory). These are firmly archaeological problems, and archaeologists are those with the greatest need to build the theory to solve them through a process of recursive reasoning. Here archaeologists must take existing perceptions, preconceptions, theories, and cultural norms; identify the circumstances where these cease to function as convincing guides; and then, by practical reasoning, build and try out alternatives. There is no reason to shy away from the use of inductive strategies in this process of practical reasoning. The second major element concerns the relationships between archaeology and history and between archaeology and anthropology. These are profoundly difficult issues that concern all archaeologists, not just those with an interest in time perspectivism. However, a focus on the archaeology of palimpsests (an issue that went right to the heart of the “Pompeii premise” debate between Binford and Schiffer) is regarded by Bailey (2007, this volume) and others as a major focus for future research. Similarly, Arnold confronts the many problems raised by the adoption of ethnographic scales of analysis in prehistoric archaeology and sensibly concludes that ethnographic perspectives have their uses but that these should not become the

benchmark for all archaeological analyses. Further research by those seeking to advance time perspectivism in archaeology will need to concentrate on working through, and then integrating, currently divergent archaeological, historical, and anthropological perspectives about the terms under which we seek an understanding of human beings past, present, and future. The final major element, that of the development of archaeological grammar (in the sense of a set of rules and principles), reflects the search for appropriate terminology to describe the new archaeological phenomena that are both the cause and the effect of time perspectivism. In Behavioral Archeology (1976), Schiffer gave substance to this new archaeological perspective by proposing and defining archaeological formation processes that relate archaeological context and systemic context. Sullivan’s discussion of traces (trace loads, trace parsing, etc.) represents one way in which archaeologists can develop new concepts and units of analysis that come closer to capturing the essence of the phenomena they seek to understand. Although a great deal more work remains to be done in this aspect of time perspectivism, there need be no doubting its necessity.

Concluding Remarks I have stressed the importance of the interplay between the empirical and the theoretical in this brief consideration of some of the issues that have arisen, and will arise, as archaeologists explore the significance of changes to our understanding of the phenomenology of archaeological records and then seek to develop frameworks with which they might make new meanings from them. Every contribution to this book adds something to our store of cases and exemplars in this field of research, but my specific concern in this chapter has been to focus on more abstract matters. Here I have argued that a critical link exists among debates about archaeological ontology, differences of opinion about the form of appropriate archaeological epistemologies, and the variable readings of the significance of the phenomenology of archaeological records for the constitution of archaeological ontology and epistemology.

178

Paradigms and Metaphysics Since the advent of the New Archaeology now over 40 years ago, the critical tension in our discipline has been between what archaeologists seek to know about past human action and what they believe they can plausibly demonstrate in terms of the structural properties of archaeological records. Throughout this period the bulk of practitioners have (in positive moments) seen theoretical archaeology as a vehicle for the expansion of archaeological horizons, and there is little doubt that archaeology now intersects with a wider range of interests and discourses than it has at any time in its history. The existence of diversity, however, has also increased the importance of traditional epistemological concerns—how do practitioners justify knowledge claims, and how do we choose between different (and frequently opposing) accounts? It is also true that choice can be (and is being) made on theoretical rather than empirical grounds, where the phenomenology of archaeological records is regarded as being simply a methodological problem of small moment alongside the great benefits that are thought to flow from the “connection” of archaeology with the concerns of contemporary social theory. On this basis we might then re-express the critical tension of the archaeology of this period as being a conflict between the determinants of plausible archaeological knowledge claims. Will new understandings of the phenomenology of archaeological records (of archaeological ontology itself ) become as significant a determinant as anthropological or “social” science models of human action or the possibilities of postempiricist epistemology? In this sense the constitution of archaeology as a discipline also becomes a fundamental point of contention. Should archaeologists redefine their discipline in terms of an increasing perception of the singularity of archaeological records as records of human action—a perception largely stemming from developments in techniques of dating and analysis, as well as from the rise of a more concerted interest in site formation processes? Or should they seek to increase the relevance of their discipline by virtue of a closer association with anthropology, cultural studies, literature and philosophy, and the debates about the proper understanding of

human action that continue to occur there? What is significant in the second position is that archaeology can participate in such debates only if the singular constitution of its data becomes a methodological problem of no particular moment. The balance of plausibility is tipped in favor of the source disciplines from which such theories have been ­borrowed. I have noted that there are powerful cultural reasons for the maintenance of traditional relationships between archaeology and other disciplines concerned with understanding the human past and present (e.g., Meltzer’s “archaeology as anthropology metaphysic”), which have tended to reduce the significance of differences between archaeological records and other records of human action. However, the arguments originally advanced by Bailey and developed by others (including the contributions to this book) have established the logic of a counterargument and provide exemplification of some of its elements. In this view (which I share) archaeologists are now presented with a sufficiently strong rationale for expecting more from borrowed theory (and indeed from the archaeologists who borrow it), in the sense that we now have a clear warrant to more closely examine how such theories should be transformed through an encounter with different (but not necessarily impoverished) information about past human action. There is a clear potential for this to become much more than a limited methodological discussion, as recursive reasoning (the free interplay of data and theory) begins to transform source theories, as well as our understanding of the possibilities of archaeological records. On offer is the archaeological redefinition of anthropological concepts and categories and the development of theories that seriously consider the ontological singularity of archaeological records as records of human action, thereby releasing the potential of those records to meaningfully contribute to the great debates about human beings. Part and parcel of developing the logic of this counterargument is an exploration of the cultural norms of our discipline as an aid to understanding why the traditions of archaeological practice have consistently placed the discipline in the position of a consumer rather than a producer

179

Tim Murray of theory (with the notable exception of material culture theory). If one accepts that the application of absolute dating systems to archaeology and the development of studies of archaeological site formation processes give good grounds for arguing that archaeological records are ontologically singular records of human action, then, logically, archaeologists need to establish two things: (1) whether serious attempts to understand the meanings of such singularity will profoundly disturb preexisting relationships between archaeology and its cognate disciplines, and (2) the nature of ontological and epistemological antinomies that have long provided the context of our search for meaning in the human past. Not only will archaeologists require great intestinal fortitude as they contemplate the building of higher-level archaeological theory, they will also need a strong sense of community to work in this very much more complex theoretical environment. Understanding the history, philosophy, and sociology of archaeological knowledge is an essential element here, if only because it is apparent that archaeologists have a great deal of work ahead of them “deconstructing” their own discipline, developing alternative approaches to the archaeological past, and “hammering out” theories and perspectives. Tilley puts the issues well: How are archaeologists to account for, understand and explain the changes that may be perceived in material culture patterning? A number of crucial questions arise for consideration: is an adequate conception of archaeological thinking inevitably radically pluralistic or should our explanatory and conceptual frameworks approximate to a single fundamental form? Do different conceptual frameworks involve any common suppositions? Are all kinds of conceptual frameworks equally ultimate or do some, more than others, depend upon the context of questioning? [1982:36]

or that archaeologists have to abandon the goal of making substantive contributions to debates about the nature of humanity, past, present, or future. It is vital that we recognize that we do not have to continue old disputes between a “dehumanized” scientific archaeology and the perspectives of contemporary social theory, which have been so much a feature of the landscape of theoretical archaeology over the past 20 years. Many other options that do not recommit us to these tired antinomies are possible. Nonetheless, a recognition that archaeological records pose more than methodological problems for the application of contemporary social theory to archaeology (and that neither the borrowed theories nor the phenomena under review are much advanced beyond logical circularity) poses some considerable challenges to us all. In this sense our task really is to confront disciplinary tradition, in the expectation that contemporary social theory will be made all the stronger and more interesting through a meaningful encounter with archaeological records. Of course, there is every possibility that the inertia of tradition, the fear of the unknown, or possibly just the desire to maintain existing power relations within and outside archaeology will act to stymie these possibilities. Notwithstanding this potential problem, we should proceed by making the logic of our arguments clearer, by demonstrating the shortcomings of existing interpretations and explanations, by developing alternative accounts that do not dismiss the significance of archaeological records, and by restructuring social theory to account for the new information we have been able to create as a result of this profound revolution in archaeology. Time perspectivism in its past, present, and future guises should play a significant role in ushering in this new archaeology. The possibilities are boundless:

Finding answers to these questions does not in any way imply that archaeologists must “disconnect” archaeological information from the human sciences 180

When individuals sense they are living through a period of crisis, when foundations seem to be cracking and orthodoxies breaking up, then a public space is created in which basic questions about the human condition can be raised anew. (Bernstein 1976:xiii)

References Cited

Adam, E., and E. Kotjabopoulou 1997 The Organic Artefacts from Klithi. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 245–​259. McDonald Institute for Archaeological Research, Cambridge. Adams, R. N. 1988 The Eighth Day: Social Evolution as the SelfOrganization­ of Energy. University of Texas Press, Austin. Ahler, S. A., T. D. Thiessen, and M. K. Trimble 1991 People of the Willows: The Prehistory and Early History of the Hidatsa Indians. University of North Dakota Press, Grand Forks. Allen, H. 1972 Where the Crow Flies Backwards: Man and Land in the Darling Basin. Unpublished Ph.D. dissertation, Department of Prehistory, Australian National University, Canberra. Allen, J. F. (editor) 1996 Report of the Southern Forests Archaeological Project, Vol. 1: Site Descriptions, Stratigraphies and Chronologies. Archaeology Publications, La Trobe University, Melbourne. Allen, J. F., and R. Cosgrove 1996 Fieldwork Methods: Excavations, Surveys and Laboratory Procedures. In Report of the Southern Forests Archaeological Project, Vol. 1: Site Descriptions, Stratigraphies and Chronologies, edited by J. F. Allen, pp. 31–39. Archaeology Publications, La Trobe University, Melbourne. Allen, K. M. S., S. W. Green, and E. B. Zubrow (editors) 1990 Interpreting Space: GIS and Archaeology. Taylor and Francis, New York. Ammerman, A. J., and M. W. Feldman 1974 On the “Making” of an Assemblage of Stone Tools. American Antiquity 39:610–616.

Amsden, C. 1977 A Quantitative Analysis of Nunamiut Eskimo Settlement Dynamics: 1890 to 1969. Unpublished Ph.D. dissertation, Department of Anthropology, University of New Mexico, Albuquerque. Andrefsky, W. 2006 Experimental and Archaeological Verification of an Index of Retouch for Hafted Bifaces. American Antiquity 71:743–757. Arnold, P. J., III 1999 On Reconciling Ethnographic and Archaeological Time. Paper presented at the 99th Annual Meeting of the American Anthropological Association, Chicago. 2000 Working without a Net: Recent Trends in Ceramic Ethnoarchaeology. Journal of Archaeological Research 8:105–133. 2003 Back to Basics: The Middle Range Program as Pragmatic Archaeology. In Essential Tensions in Archaeological Theory, edited by T. L. Van Pool and C. S. Van Pool, pp. 55–66. University of Utah Press, Salt Lake City. Artelius, T. 2001 The Framework of Pagan Identity. “Time” in the Southern Scandinavian Late Iron Age. In It’s about Time: The Concept of Time in Archaeology, edited by H. Karlsson, pp. 1–28. Bricoleur Press, Göteborg. Ascher, R. 1962 Analogy in Archaeological Interpretation. Southwestern Journal of Anthropology 17:317–325. Ascher, R., and M. Ascher 1965 Recognizing the Emergence of Man. Science 147:​ 243–250. Audouze, F., and J. G. Enloe 1997 High Resolution Archaeology at Verberie: Limits and Interpretations. World Archaeology 29:​ 195–​207.

181

References Cited Bailey, G. N. 1981 Concepts, Time-Scales and Explanations in Economic Prehistory. In Economic Archaeology, edited by A. Sheridan and G. N. Bailey, pp. 97–117. British Archaeological Reports International ­Series 96. Oxford. 1983 Concepts of Time in Quaternary Prehistory. Annual Review of Anthropology 12:165–192. 1987 Breaking the Time Barrier. Archaeological Review from Cambridge 6:5–20. 1997a The Klithi Project: History, Aims and Structure of Investigations. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 3–26. McDonald Institute for Archaeological Research, Cambridge. 1997b Klithi Excavations: Aims and Methods. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 43–60. McDonald Institute for Archaeological Research, Cambridge. 1997c Klithi: A Synthesis. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 2: Klithi in Its Local and Regional Setting, edited by G. N. Bailey, pp. 655–680. McDonald Institute for Archaeological Research, Cambridge. 1999 Eric Higgs (1908–1976). In Encyclopedia of Archaeology: The Great Archaeologists, edited by T. Murray, pp. 531–565. ABC-CLIO, Santa Barbara. 2006 Time’s Arrow: The Measurement and Theory of Archaeological Time. Antiquity 80(309):​717–​ 720. 2007 Time Perspectives, Palimpsests and the Archaeology of Time. Journal of Anthropological Archaeology 26:198–223. Bailey, G. N. (editor) 1997 Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece. 2 vols. McDonald Institute for Archaeological Research, Cambridge. Bailey, G. N., T. Cadbury, N. Galanidou, and E. Kotjabopoulou 1997 Open-Air Sites and Rockshelters: Survey Strategies and Regional Site Distributions. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 2: Klithi in Its ­Local and Regional Setting, edited by G. N. Bailey, pp. 521–537. McDonald Institute for Archaeological Research, Cambridge.

Bailey, G. N., P. L. Carter, C. S. Gamble, and H. P. Higgs 1983a Asprochaliko and Kastritsa: Further Investigations of Palaeolithic Settlement and Economy in Epirus (North-West Greece). Proceedings of the Prehistoric Society 49:15–42. 1983b Epirus Revisited: Seasonality and Inter-Site Variation in the Upper Palaeolithic of North-West Greece. In Hunter-Gatherer Economy in Prehistory: A European Perspective, edited by G. N. Bailey, pp. 64–78. Cambridge University Press, Cambridge. Bailey, G. N., and C. S. Gamble 1990 The Balkans at 18000 bp: The View from Epirus. In The World at 18000 bp, edited by C. S. Gamble and O. Soffer, Vol. 1:148–167. Unwin-Hyman, London. Bailey, G. N., G. C. P. King, and I. Manighetti 2000 Tectonics, Volcanism, Landscape Structure and Human Evolution in the African Rift. In Human Ecodynamics, edited by N. Bailey, R. Charles, and N. Winder, pp. 31–46. Oxbow, Oxford. Bailey, G. N., G. C. P. King, and D. A. Sturdy 1993 Active Tectonics and Land-Use Strategies: A Palaeolithic Example from Northwest Greece. Antiquity 67:292–312. 1994 Active Tectonics, Complex Topography and Human Survival Strategies. Journal of Geophysical Research 99(B10):20063–20078. Bailey, G. N., and J. Woodward 1997 The Klithi Deposits: Geology, Stratigraphy and Chronology. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 61–94. McDonald ­Institute for Archaeological Research, Cambridge. Balée, W. 1998 Historical Ecology: Premises and Postulates. In Advances in Historical Ecology, edited by W. Balée, pp. 13–29. Columbia University Press, New York. Bamforth, D. B., and M. S. Becker 2000 Core/Biface Ratios, Mobility, Refitting, and Artifact Use-Lives: A Paleoindian Example. Plains Anthropologist 45(173):273–290. Bamforth, D. B., M. S. Becker, and J. Hudson 2005 Intrasite Spatial Analysis, Ethnoarchaeology, and Paleoindian Land Use on the Great Plains: The Allen Site. American Antiquity 70(3):561–​ 580. Barrett, J. 1994 Fragments of Antiquity. Blackwell, Oxford.

182

References Cited Behrensmeyer, A. K. 1978 Taphonomic and Ecologic Information from Bone Weathering. Paleobiology 4(2):150–162. 1982 Time Resolution in Fluvial Vertebrate Assemblages. Paleobiology 8(3):211–227. Behrensmeyer, A. K., and D. Schindel 1983 Resolving Time in Paleobiology. Paleobiology 9(1):​1–8. Bender, B. 2002 Time and Landscape. Current Anthropology 43(4):​S103–S112. Benedict, J. B. 1967 Recent Glacial History of an Alpine Area in the Colorado Front Range, U.S.A.: Establishing a Lichen Growth Curve. Journal of Glaciology 6(48):​ 817–832. Benjamin, G. 1985 In the Long Term: Three Themes in Malayan Cultural Ecology. In Cultural Values and Human Ecology in Southeast Asia, edited by K. L. ­Hutterer, A. T. Rambo, and G. Lovelace, pp. 219–278. Michigan Papers on South and Southeast Asia 27. Center for South and Southeast Asian Studies, University of Michigan, Ann ­Arbor. Bernstein, R. J. 1976 The Restructuring of Social and Political Theory. Methuen, London. Bettinger, R. L., and R. Oglesby 1985 Lichen Dating of Alpine Villages in the White Mountains, California. Journal of California and Great Basin Anthropology 7(2):202–224. Bettis, E. A., III, and R. E. Mandel 2002 The Effects of Temporal and Spatial Patterns of Holocene Erosion and Alluviation on the Archaeological Record of the Central and Eastern Great Plains, U.S.A. Geoarchaeology 17(2):​141–​ 154. Binford, L. R. 1962 Archaeology as Anthropology. American Antiquity 28(2):217–225. 1977a Forty-Seven Trips: A Case Study in the Character of Archaeological Formation Process. In Stone Tools as Cultural Markers, edited by R. V. S. Wright, pp. 24–36. Australian Institute of Aboriginal Studies, Canberra. 1977b General Introduction. In For Theory Building in Archaeology, edited by L. R. Binford, pp. 1–10. Academic Press, New York. 1978a Dimensional Analysis of Behavior and Site Structure: Learning from an Eskimo Hunting Stand. American Antiquity 43:330–361.

183

1978b Introduction. In Nunamiut Ethnoarchaeology, L.  R. Binford, pp. 1–14. Academic Press, New York. 1980 Willow Smoke and Dogs’ Tails: Hunter-Gatherer­ Settlement Systems and Archaeological Site Formation. American Antiquity 45:4–20. 1981a Behavioral Archaeology and the “Pompeii Premise.” Journal of Anthropological Research 37:​195–​ 208. 1981b Bones: Ancient Men and Modern Myths. Academic Press, New York. 1982 The Archaeology of Place. Journal of Anthropological Archaeology 1:5–31. 1983a Long Term Land Use Patterns: Some Implications for Archaeology. In Lulu Linear Punctuated: Essays in Honor of George Irving Quimby, edited by R. C. Dunnell and D. K. Grayson, pp. 27–​53. Anthropological Papers, Museum of Anthropology No. 72. University of Michigan, Ann Arbor. 1983b In Pursuit of the Past. Thames and Hudson, London. 1983c Working at Archaeology. Academic Press, New York. 1989 Styles of Style. Journal of Anthropological Archaeology 8:51–67. 1998 Time as a Clue to Cause? Proceedings of the British Academy 101:1–35. 2001a Constructing Frames of Reference: An Analytical Method for Archaeological Theory Building Using Ethnographic and Environmental Data Sets. University of California Press, Berkeley. 2001b Where Do Research Problems Come From? American Antiquity 66:669–678. Binford, S. R., and L. R. Binford (editors) 1968 New Perspectives in Archeology. Aldine Publishing Co., Chicago. Bintliff, J. L. 1991 The Contribution of an Annaliste/Structural History Approach to Archaeology. In The Annales School and Archaeology, edited by J. L. Bint­liff, pp. 1–33. Leicester University Press, Lei­ cester. 2004 Time, Structure, and Agency: The Annales, Emergent Complexity, and Archaeology. In A Companion to Archaeology, edited by J. L. Bintliff, pp. 174–194. Blackwell, London. Bintliff, J. (editor) 1991 The Annales School and Archaeology. Leicester University Press, Leicester. Blake, E. 2003 The Familiar Honeycomb: Byzantine Era Reuse

References Cited of Sicily’s Prehistoric Rock-Cut Tombs. In Archaeologies of Memory, edited by R. M. van Dyke and S. E. Alcock, pp. 203–220. Blackwell, London. Blumenschine, R. J., and F. T. Masao 1991 Living Sites at Olduvai Gorge, Tanzania? Preliminary Landscape Archaeology Results in the Basal Bed II Lake Margin Zone. Journal of Human Evolution 21:451–462. Blumenschine, R. J., and C. R. Peters 1998 Archaeological Predictions for Hominid Land Use in the Paleo-Olduvai Basin, Tanzania, during Lowermost Bed II Times. Journal of Human Evolution 34:565–607. Boughton, J., B. Fandrich, L. Litwinionek, L. A. Peterson, and L. M. Peterson 2001 Coteau Properties Company: Testing and Criterion D Evaluation of Prehistoric Sites Located in Permit Extension Areas D and H and the West Permit Area, Mercer County, North Dakota. Ethnoscience Inc., Billings. Boughton, J., L. Litwinionek, and S. Walker-Kuntz 2000 Cultural Resource Inventory of Permit Area Extensions D and H and the West Permit Area, the Coteau Mine, Beulah, North Dakota. Ethnoscience Inc., Billings. Bradley, R. 1993 Altering the Earth: The Origins of Monuments in Britain and Continental Europe. Society of Antiquaries of Scotland, Edinburgh. 1998 The Significance of Monuments. Routledge, London. 2002 The Past in Prehistoric Societies. Routledge, London. 2003 The Translation of Time. In Archaeologies of Memory, edited by R. M. van Dyke and S. E. Alcock, pp. 221–227. Blackwell Publishers, Malden, Massachusetts. 2005 Ritual and Domestic Life in Prehistoric Europe. Routledge, London. Broadbent, N. D. 1987 Lichenometry and Archaeology. Research Report No. 2. Center for Arctic Cultural Research, University of Umeå, Umeå. 1990 Use of Lichenometric and Weathering Rates for Dating. Norwegian Archaeological Review 23(1–2):​3–8. Bronk Ramsey, C. 2002 OxCal Program. 3.8 ed. University of Oxford, Radiocarbon Accelerator Unit, Oxford. Broughton, J. M. 2002 Prey Spatial Structure and Behavior Affect Archaeological Tests of Optimal Foraging Models:

Examples from the Emeryville Shellmound Vertebrate Fauna. World Archaeology 34(1):60–83. Brumley, J. H., and B. J. Dau 1988 Historical Resource Investigations within the Forty Mile Coulee Reservoir. Archaeological Survey of Alberta, Manuscript Series No. 13. Edmonton. Bunn, H. T., and E. M. Kroll 1993 Comments on “The Structure of the Lower Pleistocene Archaeological Record.” Current Anthropology 34:216–217. Burley, D. 1990 Tipi Rings and Alberta History: Toward a Historical and Critical Review of a Legislated Archaeology. Plains Anthropologist 35(132):343–357. Butzer, K. W. 1982 Archaeology as Human Ecology: Method and ­Theory for a Contextual Approach. Cambridge University Press, Cambridge. Camilli, E. L. 1983 Site Occupation History and Lithic Assemblage Structure: An Example from Southwestern Utah. Unpublished Ph.D. dissertation, Department of Anthropology, University of New Mexico, Albuquerque. 1988 Interpreting Long-Term Land-Use Patterns from Archaeological Landscapes. American Archaeology 7(1):57–66. Camilli, E. L., and J. I. Ebert 1992 Artifact Reuse and Recycling in Continuous Surface Distributions and Implications for Interpreting Land Use Patterns. In Space, Time and Archaeological Landscapes, edited by J. Rossignol and L. Wandsnider, pp. 113–136. Plenum Press, New York. Camilli, E. L., L. Wandsnider, and J. I. Ebert 1988 Distributional Survey and Excavation of Archaeological Landscapes in the Vicinity of El Paso, Texas. U.S. Bureau of Land Management, Las Cruces. Carnap, R. 1966 An Introduction to the Philosophy of Science. Basic Books, New York. Chapman, J., and B. Gaydarska 2007 Parts and Wholes: Fragmentation in Prehistoric Context. Oxbow Books, Oxford. Chatters, J. C. 1987 Hunter-Gatherer Adaptations and Assemblage Structure. Journal of Anthropological Archaeology 6:​336–375. Christian, C. S., and G. A. Stewart 1953 General Report on Survey of Katherine-Darwin Region, 1946. Commonwealth Scientific and Industrial Research Organisation, Australian Land Research Service 1. Canberra.

184

References Cited In Report of the Southern Forests Archaeological Project, Vol. 1: Site Descriptions, Stratigraphies and Chronologies, edited by J. Allen, pp. 21–30. Archaeology Publications, La Trobe University, Melbourne. 2002 Prey Choice and Hunting Strategies in the Late Pleistocene: Evidence from Southwest Tasmania. In Histories of Old Ages: Essays in Honour of Rhys Jones, edited by A. Anderson, S. O’Connor, and I. Lilley, pp. 397–427. Coombs Academic Publishing, Australian National University, Canberra. Cosgrove, R., J. F. Allen, and B. Marshall 1990 Palaeoecology and Pleistocene Human Occupation in South Central Tasmania. Antiquity 64:​ 59–78. Creasman, S. D., and K. W. Thompson 1997 Archaic Settlement and Subsistence in the Green River Basin of Wyoming. In Changing Perspectives of the Archaic on the Northwest Plains and Rocky Mountains, edited by M. L. Larson and J.  E. Francis, pp. 242–305. University of South Dakota Press, Vermillion. Curry, R. L. 1969 Holocene Climatic and Glacial History of the Central Sierra Nevada, California. In United States Contributions to Quaternary Research, edited by S. A. Schumm and W. C. Bradley, pp. 1–​ 47. Geological Society of America Special Paper 123. Boulder. Cutler, A. H., and K. W. Flessa 1990 Fossils out of Sequence: Computer Simulations and Strategies for Dealing with Stratigraphic Disorder. Palaios 5:227–235. Damm, C. 2001 Another Time, Another History? or How to Write the History of the Bugakhwe in Botswana. In It’s about Time: The Concept of Time in Archaeology, edited by H. Karlsson, pp. 29–44. Bricoleur Press, Göteborg. Dancey, W. S. 1973 Prehistoric Land Use and Settlement Patterns in the Priest Rapids Area, Washington. Unpublished Ph.D. dissertation, Department of Anthropology, University of Washington, Seattle. David, N., and C. Kramer 2001 Ethnoarchaeology in Action. Cambridge University Press, Cambridge. Davis, L. B., S. A. Aaberg, R. Ottersberg, and M. Wilson 1982 Stone Circles in the Montana Rockies: Systematic Recovery and Cultural-Ecological Inference. Montana State University, Bozeman, for Montana Army National Guard, Department of Military Affairs, Helena.

Ciolek-Torrello, R. 1984 An Alternative Model of Room Function from Grasshopper Pueblo, Arizona. In Intrasite Spatial Analysis in Archaeology, edited by H. Hiet­ala, pp. 127–153. Cambridge University Press, Cambridge. Clark, G. 1954 The Study of Prehistory. Cambridge University Press, Cambridge. 1961 World Prehistory. Cambridge University Press, Cambridge. 1989 Prehistory at Cambridge and Beyond. Cambridge University Press, Cambridge. 1992 Space, Time and Man. Cambridge University Press, Cambridge. Clarke, D. L. 1968 Analytical Archaeology. Methuen, London. 1973 Archaeology: The Loss of Innocence. Antiquity 46(185):​6–18. 1978 Analytical Archaeology. 2nd ed. Columbia University Press, New York. Colhoun, E., G. van de Geer, S. J. Fitzsimons, and L. E. Heusser 1994 Terrestrial and Marine Records of the Last Glaciation from Western Tasmania: Do They Agree? Quaternary Science Review 13:293–300. Conard, N. 1994 Discussion and Criticism: On the Prospects for an Ethnography of Extinct Hominids. Current Anthropology 35:281–282. Cormack, J. L. 1996 Big, the Bad and the Ugly: Heavy Duty Tools Revealed. Kaupia 6:37–46. Cosgrove, R. 1991 The Illusion of Riches: Issues of Scale, Resolution and Explanation of Pleistocene Human Behaviour. Unpublished Ph.D. dissertation, Department of Archaeology, La Trobe University, Melbourne. 1995 The Illusion of Riches: Scale, Resolution and Explanation in Tasmanian Pleistocene Human Behaviour. British Archaeological Reports International Series 608. Oxford. 1996 Nunamira Cave. In Report of the Southern Forests Archaeological Project, Vol. 1: Site Descriptions, Stratigraphies and Chronologies, edited by J. F. Allen, pp. 43–68. Archaeology Publications, La Trobe University, Melbourne. 1999 Forty-Two Degrees South. The Archaeology of Late Pleistocene Tasmania. Journal of World Prehistory 13:357–402. Cosgrove, R., and J. F. Allen 1996 Research Strategies and Theoretical Perspectives.

185

References Cited Deal, M. 1998 Pottery Ethnoarchaeology in the Central Maya Highlands. University of Utah Press, Salt Lake City. 2008 Abandonment Patterning at Archaeological Settlements. In Archaeological Concepts for the Study of the Cultural Past, edited by A. P. Sullivan III, pp. 141–​157. University of Utah Press, Salt Lake City. Deaver, K. 1983 Archaeological Investigations on the Northern Border Pipeline: Montana Segment, Vols. 1–3. Professional Analysts, Omaha. 1985 Mitigation of the Anderson Tipi Ring Site (32ML111), McLean County, North Dakota. Ethnoscience Inc., Billings. 1989 Identifying Ring Site Occupations. In Households and Communities, edited by S. MacEachern, D. Archer, and R. Garvin, pp. 256–265. University of Calgary Press, Calgary. 1999 Cultural Investigations along the Montana Segment of the Express Pipeline, Vol. 8: Class I Investigations of Shallow Stone Feature Sites in Central Montana. Ethnoscience Inc., Billings. Deaver, K., and S. Deaver 1987 Dancing Grouse, a Tipi Ring Site in Central North Dakota. Ethnoscience Inc., Billings. 1988 Prehistoric Cultural Resource Overview of Southeast Montana. Ethnoscience Inc., Billings. DeBoer, W. R. 1974 Ceramic Longevity and Archaeological Interpretation: An Example from the Upper Ucayali, Peru. American Antiquity 39:335–342. 1983 The Archaeological Record as Preserved Death Assemblage. In Archaeological Hammers and Theories, edited by A. S. Keene and J. Moore, pp. 19–​36. Academic Press, New York. DeBoer, W. R., and D. W. Lathrap 1979 The Making and Breaking of Shipibo-Conibo Ceramics. In Ethnoarchaeology: Implications of Ethnography for Archaeology, edited by C. Kramer, pp. 102–138. Columbia University Press, New York. de Lange, J. 2001 Matter as Process. The Spatiotemporal Constitution of Archaeological Data. Archaeological Dialogues 8:152–163. 2005 The Structural Properties of Pleistocene Archaeological Records. Unpublished Ph.D. dissertation, Archaeology Program, School of Historical and European Studies, La Trobe University, Melbourne.

de la Torre, I., and R. Mora 2005 Unmodified Lithic Material at Olduvai Bed I: Manuports or Ecofacts? Journal of Archaeological Science 32:273–285. Dewar, R. E., and K. A. McBride 1992 Remnant Settlement Patterns. In Space, Time, and Archaeological Landscapes, edited by J. Rossignol and L. Wandsnider, pp. 227–255. Plenum, New York. Dibble, H. L. 1984 Interpreting Typological Variation of Middle Paleolithic Scrapers: Function, Style, or Sequence of Reduction? Journal of Field Archaeology 11:​ 431–​436. 1987 The Interpretation of Middle Palaeolithic Scraper Morphology. American Antiquity 52:109–117. 1995 Middle Paleolithic Scraper Reduction: Background, Clarification, and Review of the Evidence to Date. Journal of Archaeological Method and Theory 2:299–368. Dibble, H. L., and N. Rolland 1992 On Assemblage Variability in the Middle Paleolithic of Western Europe. In The Middle Paleolithic: Adaptation, Behavior, and Variability, edited by H. L. Dibble and P. Mellars, pp. 1–28. University Museum, University of Pennsylvania, Philadelphia. Diehl, M. W. 1998 The Interpretation of Archaeological Floor Assemblages: A Case Study from the American Southwest. American Antiquity 63:617–634. Doelman, T., J. Webb, and M. Domanski 2001 Source to Discard: Patterns of Lithic Raw Material Procurement and Use in Sturt National Park, Northwestern New South Wales. Archaeology in Oceania 36:15–33. Dooley, M. A. 2004 Long-Term Hunter-Gatherer Land Use in Central North Dakota: An Environmental Analysis. Plains Anthropologist 49:105–127. Downum, C. E., and G. B. Brown 1998 The Reliability of Surface Artifact Assemblages as Predictors of Subsurface Remains. In Surface Archaeology, edited by A. P. Sullivan III, pp. 111–​ 123. University of New Mexico Press, Albuquerque. Drennan, R. D. 1996 Statistics for Archaeologists: A Commonsense Approach. Plenum, New York. Dunnell, R. C. 1978 Style and Function: A Fundamental Dichotomy. American Antiquity 43:192–202.

186

References Cited Evolutionary Theory and Archaeology. In Advances in Archaeological Method and Theory, edited by M. B. Schiffer, Vol. 3:35–99. Academic Press, New York. 1982 Science, Social Science, and Common Sense: The  Agonizing Dilemma of Modern Archaeology. Journal of Anthropological Research 38(1):​ 1–25. 1992 The Notion Site. In Space, Time, and Archaeological Landscapes, edited by J. Rossignol and L. Wandsnider, pp. 21–41. Plenum, New York. Earman, J. 1967 Irreversibility and Temporal Asymmetry. Journal of Philosophy 64:543–549. Ebert, J. 1992 Distributional Archaeology. University of New Mexico Press, Albuquerque. Eckerle, W. P. 1997 Eolian Geoarchaeology of the Wyoming Basin: Changing Environments and Archaic Subsistence Strategies in the Holocene. In Changing Perspectives of the Archaic on the Northwest Plains and Rocky Mountains, edited by M. L. Larson and J. E. Francis, pp. 139–167. University of South Dakota Press, Vermillion. Eckerle, W. P., and J. Hobey 1999 Geoarchaeological Assessment. In The Trappers Point Site (48SU1006): Early Archaic Adaptations in the Upper Green River Basin, Wyoming, edited by M. E. Miller, P. H. Sanders, and J. E. Francis, pp. 58–90. Cultural Resources Series 1. Office of the Wyoming State Archaeologist, University of Wyoming, Laramie. Eckerle, W. P., J. Hobey, and J. E. Francis 1999 The Environmental and Cultural Setting. In The Trappers Point Site (48SU1006): Early Archaic Adaptations in the Upper Green River Basin, Wyoming, edited by M. E. Miller, P. H. Sanders, and J. E. Francis, pp. 5–22. Cultural Resources Series 1. Office of the Wyoming State Archaeologist, University of Wyoming, Laramie. Elston, R. G. 1990 A Cost-Benefit Model of Lithic Assemblage Variability. In The Archaeology of James Creek Shelter, edited by R. G. Elston and E. E. Budy, pp. 153–​ 164. University of Utah Anthropological Papers 115. Salt Lake City. Evans, C. 1999 Christopher Hawkes. In Encyclopedia of Archaeology: The Great Archaeologists, edited by T. ­Murray, pp. 461–479. ABC-CLIO, Santa Barbara. 1980

Fagan, B. 1995 People of the Earth. 8th ed. Harper Collins College Publishers, New York. 2001 People of the Earth: An Introduction to World Prehistory. 10th ed. Prentice Hall, Upper Saddle River, N.J. Fall, P. L., P. T. Davis, and G. A. Zielinski 1995 Late Quaternary Vegetation and Climate of the Wind River Range, Wyoming. Quaternary Research 43:​393–404. Fanning, P. C. 1994 Long-Term Contemporary Erosion Rates in an Arid Rangelands Environment in Western New South Wales, Australia. Journal of Arid Environments 28:173–187. 1999 Recent Landscape History in Arid Western New South Wales, Australia: A Model for Regional Change. Geomorphology 29:191–209. 2002 Beyond the Divide: A New Geoarchaeology of Aboriginal Stone Artefact Scatters in Western NSW, Australia. Unpublished Ph.D. dissertation, Graduate School of the Environment, Macquarie University, Sydney. Fanning, P. C., and S. J. Holdaway 2001 Temporal Limits to the Archaeological Record in Arid Western NSW, Australia: Lessons from OSL and Radiocarbon Dating of Hearths and Sediments. In Australasian Connections and New Directions Proceedings of the 7th Archaeometry Conference, edited by M. Jones and P. Sheppard, pp. 85–​104. Department of Anthropology, University of Auckland, Auckland. 2002 Using Geospatial Technologies to Understand Prehistoric Human/Landscape Interaction in Arid Australia. Arid Lands Newsletter 51. Electronic document, http://cals.arizona.edu/OALS​ /​ALN/aln51/fanning.html, accessed December 2007. 2004 Stone Artifact Exposure and Visibility at Open Sites in Western New South Wales, Australia: A Geomorphic Framework for Survey and Analysis. Journal of Field Archaeology 29(3–4):255–271. Fanning, P. C., S. J. Holdaway, and E. Rhodes In press A New Geoarchaeology of Aboriginal Artefact Deposits in Western NSW, Australia: Establishing Spatial and Temporal Geomorphic Controls on the Surface Archaeological Record. Geomorphology. Fenneman, N. M., and D. W. Johnson 1946 Physical Divisions of the United States. U.S. Geological Survey Physiographic Committee, Reston, Virginia.

187

References Cited Ferring, C. R. 1986 Rates of Fluvial Sedimentation: Implications for Archaeological Variability. Geoarchaeology 1(3):​ 259–274. Finnigan, J. T. 1982 Tipi Rings and Plains Prehistory: A Reassessment of Their Archaeological Potential. National Museum of Man, Mercury Series, Paper No. 108. Archaeological Survey of Canada, Ottawa. 1983 Tipi to Tipi Ring: A Transformational Model. In From Microcosm to Macrocosm: Advances in Tipi Ring Investigation and Interpretation, edited by L. B. Davis, pp. 17–28. Plains Anthropologist Memoir 19. Lincoln. Fisher, J., J. W. Strickland, and H. C. Strickland 1989 Ethnoarchaeology among the Efe Pygmies, Zaire: Spatial Organization of Campsites. American Journal of Physical Anthropology 78:473–484. Fletcher, R. J. 1977 Alternatives and Differences. In Archaeology and  Anthropology: Areas of Mutual Interest, edited by M. Spriggs, pp. 49–68. British Archaeological Reports Supplementary Series 19. Oxford. 1992 Time Perspectivism, Annales, and the Potential of Archaeology. In Archaeology, Annales, and Ethnohistory, edited by A. B. Knapp, pp. 35–49. Cambridge University Press, Cambridge. 1995 The Limits of Settlement Growth. Cambridge University Press, Cambridge. 1999 David Clarke. In Encyclopedia of Archaeology: The Great Archaeologists, edited by T. Murray, pp. 855–​868. ABC-CLIO, Santa Barbara. Flew, A. 1983 A Dictionary of Philosophy. 2nd ed. Macmillan, London. Foley, R. A. 1981a A Model of Regional Archaeological Structure. Proceedings of the Prehistoric Society 47:1–17. 1981b Off-Site Archaeology: An Alternative Approach for the Short-Sited. In Pattern of the Past: Essays in Honour of David Clarke, edited by I. Hodder, G. Isaac, and N. Hammond, pp. 157–183. Cambridge University Press, Cambridge. 1981c Off-Site Archaeology and Human Adaptation in Eastern Africa. British Archaeological Reports International Series 97. Oxford. 1995 Humans before Humanity: An Evolutionary Perspective. Blackwell, Oxford. Francis, J. E., and P. H. Sanders 1999 Excavation Summary. In The Trappers Point Site (48SU1006): Early Archaic Adaptations in the

Upper Green River Basin, Wyoming, edited by M. E. Miller, P. H. Sanders, and J. E. Francis, pp. 38–​57. Cultural Resources Series 1. Office of the Wyoming State Archaeologist, University of Wyoming, Laramie. Fredlund, L. B., D. Herbort, and G. Munson 1984 The Mitigation of Archaeological Site 32OL270, Glenharold Mine, North Dakota. CM Services, Inc., Bismarck. Frison, G. C. 1967 The Piney Creek Sites, Wyoming. University of Wyoming Publications in Science 33:1–92. 1973 The Wardell Buffalo Trap, 48SU301: Communal Procurement in the Green River Basin, Wyoming. Anthropological Papers, No. 48. Museum of Anthropology, University of Michigan, Ann Arbor. 1978 Prehistoric Hunters of the High Plains. Academic Press, New York. 1991 Prehistoric Hunters of the High Plains. 2nd ed., rev. Academic Press, New York. Frison, G. C., D. Schwab, L. Hannus, P. Winham, D. Walter, and R. C. Mainfort 1996 Archaeology of the Northern Plains. In Archaeological and Bioarchaeological Resources of the Northern Plains, edited by G. C. Frison and R. C. Mainfort, pp. 8–40. Arkansas Archaeological Survey, Fayetteville. Fritz, J. 1972 Archaeological Systems for Indirect Observation of the Past. In Contemporary Archaeology: A Guide to Theory and Contributions, edited by M. Leone, pp. 135–157. Southern Illinois University Press, Carbondale. Fürsich, F. T., and M. Aberhan 1990 Significance of Time-Averaging for Palaeocommunity Analysis. Lethaia 23:143–152. Galanidou, N. 1997a Lithic Refitting and Site Structure at Kastritsa. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 497–520. McDonald Institute for Archaeological Research, Cambridge. 1997b The Spatial Organization of Klithi. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 275–​304. McDonald Institute for Archaeological Research, Cambridge. Galbraith, R. F., R. G. Roberts, G. M. Laslett, H. Yoshida, and J. M. Olley 1999 Optical Dating of Single and Multiple Grains of

188

References Cited Quartz from Jinmium Rock Shelter, Northern Australia: Pt. I. Experimental Design and Statistical Models. Archaeometry 41:339–364. Gamble, C. S. 1983 Culture and Society in the Upper Palaeolithic of Europe. In Hunter-Gatherer Economy in Prehistory, edited by G. N. Bailey, pp. 201–211. Cambridge University Press, Cambridge. 1997 The Animal Bones from Klithi. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 207–​244. McDonald Institute for Archaeological Research, Cambridge. 1999a Lewis Binford. In Encyclopedia of Archaeology: The Great Archaeologists, edited by T. Murray, pp. 811–834. ABC-CLIO, Santa Barbara. 1999b The Palaeolithic Societies of Europe. Cambridge University Press, Cambridge. 2007 Origins and Revolutions. Cambridge University Press, Cambridge. Gardner, M. 1976 The Relativity Explosion. Vintage Books, New York. Gell, A. 1992 The Anthropology of Time: Cultural Constructions of Temporal Maps and Images. Berg, Provi­ dence. Gellner, E. 1985 Positivism against Hegelianism. In Relativism and the Social Sciences, edited by E. Gellner, pp. 4–67. Cambridge University Press, Cambridge. Gettis, A., and B. Boots 1978 Models of Spatial Processes: An Approach to the Study of Point, Line and Area Patterns. Cambridge University Press, Cambridge. Giddens, A. 1979 Central Problems in Social Theory: Action, Structure and Contradiction in Social Analysis. Macmillan, London. Gilman, P. A. 1987 Architecture as an Artifact: Pitstructures and Pueblos in the American Southwest. American Antiquity 52:538–564. Gingrich, A., E. Ochs, and A. Swedlund 2002 Repertoires of Timekeeping in Anthropology. Current Anthropology 43(4):S3–S4. Goodfriend, G. A. 1987 Evaluation of Amino-Acid Racemization/ Epimerization Dating Using RadiocarbonDated Fossil Land Snails. Geology 15:698–700. 1989 Complementary Use of Amino-Acid Epimeriza-

tion and Radiocarbon Analysis for Dating of Mixed-​Age Fossil Assemblages. Radiocarbon 31:​ 1041–1047. Goreki, P. 1991 Horticulturalists as Hunter-Gatherers: Rock Shelter Usage in Papua New Guinea. In Ethnoarchaeological Approaches to Mobile Campsites: Hunter-Gatherer and Pastoralist Case Studies, edited by C. S. Gamble and W. A. Boismier, pp. 237–​262. International Monographs in Prehistory, Ethnoarchaeological Series 1. Ann Arbor. Gosden, C. 1994 Social Being and Time. Blackwell, Oxford. Gosden, C., and Y. Marshall 1999 The Cultural Biography of Objects. World Archaeology 31(2):169–178. Gould, R. A. 1978 Beyond Analogy in Ethnoarchaeology. In Explorations in Ethnoarchaeology, edited by R. A. Gould, pp. 249–293. School of American Research, University of New Mexico Press, Albuquerque. Gould, S. J. 1965 Is Uniformitarianism Necessary? American Journal of Science 263:223–228. Gowlett, J. A. J. 1997 High Definition Archaeology: Ideas and Evaluation. World Archaeology 29:152–171. 1999 Charles McBurney. In Encyclopedia of Archaeology: The Great Archaeologists, edited by T. ­Murray, pp. 713–726. ABC-CLIO, Santa Barbara. Gowlett, J. A. J., R. Hedges, and R. Housley 1997 Klithi: The AMS Radiocarbon Dating Programme for the Site and Its Environs. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 27–​40. McDonald Institute for Archaeological Research, Cambridge. Gragson, T. L. 1983 A Preliminary Look at Stone Circle Site Distribution in Montana. In From Microcosm to Macrocosm: Advances in Tipi Ring Investigation and Interpretation, edited by L. B. Davis, pp. 139–145. Plains Anthropologist Memoir 19, Lincoln. Graham, I. 1980 Spectral Analysis and Distance Methods in the Study of Archaeological Distributions. Journal of Archaeological Science 7:105–129. Green, S. F. 1997 Interweaving Landscapes: The Relevance of

189

References Cited ­ thnographic Data on Rural Groups in Epirus E for Palaeolithic Research. In Klithi: Palaeolithic ­Settlement and Quaternary Landscapes in Northwest Greece, Vol. 2: Klithi in Its Local and Regional Setting, edited by G. N. Bailey, pp. 637–654. McDonald Institute for Archaeological Research, Cambridge. 2005 Notes from the Balkans: Locating Marginality and Ambiguity on the Greek–Albanian Border. Princeton University Press, Princeton. Gregg, M. L. 1985 An Overview of the Prehistory of Western Central North Dakota. Bureau of Land Management Cultural Resources Series No. 1. U.S. Department of the Interior, Bureau of Land Management, Montana State Office, Billings. Grunbaum, A. 1969 The Meaning of Time. In Essays in Honor of Carl G. Hempel, edited by N. Rescher, pp. 147–177. D. Reidel Publishing, Dordrecht. Hall, B. K. 1998 Evolutionary Developmental Biology. 2nd ed. Kluwer Academic Publishers, Dordrecht. Harding, J. 2005 Rethinking the Great Divide: Long-Term Structural History and the Temporality of Event. Norwegian Archaeological Review 38(2):88–101. Harrell, L. L. 1989 The Buffalo Hump Site: Late Prehistoric Occupation in the Great Divide Basin, Wyoming. Cultural Resource Series No. 7. Bureau of Land Management, Vermillion, S.D. Harrell, L. L., T. Hoefer III, and S. T. McKern 1997 Archaic Housepits in the Wyoming Basin. In Changing Perspectives of the Archaic on the Northwest Plains and Rocky Mountains, edited by M. L. Larson and J. E. Francis, pp. 334–367. University of South Dakota Press, Vermillion. Harrell, L. L., and S. T. McKern 1986 The Maxon Ranch Site: Archaic and Late Prehistoric Habitation in Southwest Wyoming. Cultural Resource Management Report No. 18. Archaeological Services, Western Wyoming Community College, Rock Springs. Harris, E. C. 1979 Principles of Archaeological Stratigraphy. Academic Press, New York. Hayden, B. 1977 Stone Tool Functions in the Western Desert. In Stone Tools as Cultural Markers: Change, Evolution and Complexity, edited by R. Wright, pp. 178–188. Australian Institute of Aboriginal Studies, Canberra.

Hayek, L. C., and M. A. Buzas 1997 Surveying Natural Populations. Columbia University Press, New York. Hegmon, M. 2003 Setting Theoretical Egos Aside: Issues and Theory in North American Archaeology. American Antiquity 68:213–243. Heirich, M. 1964 The Use of Time in the Study of Social Change. American Sociological Review 29:386–397. Higgs, E. S. 1968 Archaeology—Where Now? Mankind 6:617–620. Hinsley, C. M., Jr. 1989 Revising and Revisioning the History of Archaeology: Reflections on Region and Context. In Tracing Archaeology’s Past, edited by A. L. Christenson, pp. 79–96. Southern Illinois University Press, Carbondale. Hiscock, P., and V. Attenbrow 1998 Early Holocene Backed Artefacts from Australia. Archaeology in Oceania 33:49–62. 2003 Early Australian Implement Variation: A Reduction Model. Journal of Archaeological Science 30:​ 239–249. Hodder, I. 1985 Postprocessual Archaeology. Advances in Archaeological Method and Theory 8:1–26. 1987 The Meaning of Discard: Ash and Domestic Space in Baringo. In Method and Theory for Activity Area Research, edited by S. Kent, pp. 424–​ 448. Columbia University Press, New York. 1999 The Archaeological Process: An Introduction. Black­well, Oxford. 2001 Introduction: A Review of Contemporary Theoretical Debate in Archaeology. In Archaeological Theory Today, edited by I. Hodder, pp. 1–13. Polity Press, Cambridge. Hodder, I., and C. Orton 1976 Spatial Analysis in Archaeology. Cambridge University Press, Cambridge. Hoefer, T., III 1986 Archaeological Data Recovery at Site 48SW1242: LaBarge Natural Gas Project. Cultural Resource Management Report No. 21. Archaeological Services, Western Wyoming Community College, Rock Springs. 1987 The Porter Hollow Site: A Late Prehistoric and Archaic Hunting Camp, Unita County, Wyoming. Cultural Resource Management Report No. 30. Archaeological Services, Western Wyoming Community College, Rock Springs. Holdaway, S. J., and P. C. Fanning 2003 Time, Period, Place and Preservation: An

190

References Cited ­ nalysis of Aboriginal Heat Retainer Hearths A from Fowlers Gap, Western NSW. Unpublished report to the New South Wales National Parks and Wildlife Service, Sydney. Holdaway, S. J., P. C. Fanning, M. Jones, J. Shiner, D. C. Witter, and G. Nicholls 2002 Variability in the Chronology of Late Holocene Aboriginal Occupation on the Arid Margin of Southeastern Australia. Journal of Archaeological Science 29:351–363. Holdaway, S. J., P. C. Fanning, and J. Shiner 2005 Absence of Evidence or Evidence of Absence? Understanding the Chronology of Indigenous Occupation of Western New South Wales, Australia. Archaeology in Oceania 40:33–49. Holdaway, S. J., P. C. Fanning, and D. C. Witter 2000 Prehistoric Aboriginal Occupation of the Rangelands: Interpreting the Surface Archaeological Record of Far Western New South Wales Australia. The Rangelands Journal 22:44–57. Holdaway, S. J., and N. Porch 1996 Dates as Data. In Report of the Southern Forests Archaeological Project, Vol. 1: Site Descriptions, Stratigraphy and Chronologies, edited by J. Allen, pp. 251–277. Archaeology Publications, La Trobe University, Melbourne. Holdaway, S. J., J. Shiner, and P. C. Fanning 2004 Hunter-Gatherers and the Archaeology of Discard Behavior: An Analysis of Surface Stone Artifacts from Sturt National Park, Western New South Wales, Australia. Asian Perspectives 43:34–​ 72. Holdaway, S. J., and N. Stern 2004 A Record in Stone: Decoding the Australian Flaked Stone Record. Aboriginal Studies Press, Canberra. Holdaway, S. J., and L. Wandsnider 2006 Temporal Scales and Archaeological Landscapes from the Eastern Desert of Australia and Intermontane North America. In Confronting Scale in Archaeology: Issues of Theory and Practice, edited by G. Lock and B. Molyneaux, pp. 183–202. Kluwer Plenum Press, New York. Holdaway, S. J., D. Witter, P. C. Fanning, R. Musgrave, G. Cochrane, T. Doelman, S. Greenwood, D. Pigdon, and J. Reeves 1998 New Approaches to Open Site Spatial Archaeology in Sturt National Park, New South Wales, Australia. Archaeology in Oceania 33:1–19. Holliday, V. T., E. Johnson, and T. W. Stafford Jr. 1999 AMS Radiocarbon Dating of the Type Plainview and Firstview (Paleoindian) Assemblages: The Agony and the Ecstasy. American Antiquity 64:​ 444–​454.

191

Hopkinson, T., and M. J. White 2005 The Acheulean and the Handaxe: Structure and Agency in the Palaeolithic. In The Hominid Individual in Context, edited by C. Gamble and M. Porr, pp. 13–28. Taylor and Francis, London. Hosfield, R. T. 2001 The Lower Palaeolithic of the Solent: “Site” Formation and Interpretive Frameworks. In Palaeo­ lithic Archaeology of the Solent River, edited by F. Wenban-Smith and R. T. Hosfield, pp. 85–97. Lithic Studies Society Occasional Paper No. 7. London. Hull, K. L. 2005 Process, Perception, and Practice: Time Per­ spectivism in Yosemite Native Demography. Journal of Anthropological Archaeology 24(4):​ 354–​377. Ingold, T. 1993 The Temporality of the Landscape. World Archaeology 25:152–174. Innes, J. L. 1985 Lichenometry: A Review. Progress in Physical Geography 9:187–254. Isaac, G. Ll. 1972 Chronology and the Tempo of Cultural Change during the Pleistocene. In Calibration of Hominoid Evolution, edited by W. W. Bishop and J. A. Miller, pp. 381–430. Academic Press, Edinburgh. 1981 Stone Age Visiting Cards: Approaches to the Study of Early Land Use Patterns. In Pattern of the Past: Studies in Honour of David Clarke, edited by I. Hodder, G. Isaac, and N. Hammond, pp. 131–​ 155. Cambridge University Press, Cambridge. Isaac, G. Ll., J. W. Harris, and E. M. Kroll 1997 The Stone Artefact Assemblages: A Comparative Study. In Koobi Fora Research Project, Vol. 5: Plio-Pleistocene Archaeology, edited by G. Ll. Isaac and B. Isaac, pp. 262–299. Cambridge University Press, Cambridge. James, W., and D. Mills (editors) 2005 The Qualities of Time: Anthropological Approaches. Berg, Oxford. Jansen, J. D. 2001 Bedrock Channel Morphodynamics and Landscape Evolution in an Arid Zone Gorge: Sandy Creek Gorge, Northern Barrier Range, SouthEastern Central Australia. Unpublished Ph.D. dissertation, Macquarie University, Sydney. Jansen, J. D., and G. J. Brierley 2004 Pool-Fills: A Window to Palaeoflood History and Response in Bedrock-Confined Rivers. Sedimentology 51:901–925.

References Cited Jenks, G. F., and F. C. Caspall 1971 Error on Chroplethic Maps: Definition, Measurement, Reduction. Annals of the Association of American Geographers 61(2):217–244. Jochim, M. A. 1976 Hunter-Gatherer Subsistence and Settlement: A Predictive Model. Academic Press, New York. Johnson, M. 1989 Conceptions of Agency in Archaeological Interpretation. Journal of Anthropological Archaeology 8:​189–211. Jones, M., and G. Nicholls 2002a DateLab 1.2. Electronic document, http://www​ .datelab.org, accessed March 2007. 2002b New Radiocarbon Calibration Software. Radiocarbon 44(3):663–674. Jones, P. R. 1994 Results of Experimental Work in Relation to the Stone Industries of Olduvai Gorge. In Olduvai Gorge, Vol. 5: Excavations in Beds III, IV and the Masek Beds, 1968–1971, edited by M. Leakey and D. Roe, pp. 254–298. Cambridge University Press, Cambridge. Juell, K. E., and D. A. Edwards 1994 On the Structure of the Lower Pleistocene Archaeological Record. Current Anthropology 35:​ 167–​170, 172–173. Karlsson, H. (editor) 2001 It’s about Time: The Concept of Time in Archaeology. Bricoleur Press, Göteborg. Kealhofer, L. 1999 Creating Social Identity in the Landscape: Tidewater, Virginia, 1600–1750. In Archaeologies of Landscape, edited by W. Ashmore and A. B. Knapp, pp. 58–82. Blackwell, Walden, Massachusetts. Kelly, R. L. 1988 Hunter-Gatherer Land Use and Regional Geomorphology: Implications for Archeological Survey. In Issues in Archeological Surface Survey: Meshing Method and Theory, edited by L. Wandsnider and J. I. Ebert, pp. 49–57. American Archeology 7(1). Aticheston Press, Ridgefield, Connecticut. Kerrich, J. E. 1957 Statistical Note. South African Archaeological Bulletin 12:137–143. Kidwell, S. 1986 Modes of Fossil Concentrations: Palaeobiologic Implications. Paleobiology 12:6–24. Kimura, Y. 1999 Tool-Using Strategies by Early Hominids at Bed

II, Olduvai Gorge, Tanzania. Journal of Human Evolution 37:807–831. King, G., and G. Bailey 2006 Tectonics and Human Evolution. Antiquity 80:​ 265–​286. King, G., D. Sturdy, and G. N. Bailey 1997 The Tectonic Background to the Epirus Landscape. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 2: Klithi in Its Local and Regional Setting, edited by G. N. Bailey, pp. 541–559. McDonald Institute for Archaeological Research, Cambridge. Kirch, P. V. 2005 Archaeology and Global Change: The Holocene Record. Annual Review of Environmental Resources 30:409–440. Kleindienst, M. R. 1961 Variability within the Late Acheulian Assemblage in Eastern Africa. South African Archaeological Bulletin 16:35–52. Knapp, A. B. 1992 Archaeology and Annales: Time, Space, and Change. In Archaeology, Annales, and Ethnohistory, edited by A. B. Knapp, pp. 1–21. Cambridge University Press, Cambridge. Knapp, A. B. (editor) 1992 Archaeology, Annales, and Ethnohistory. Cambridge University Press, Cambridge. Kneebone, R. R. 1990 Energy Flow, Spatial Organization, and Community Structure at Matacapan, Veracruz, Mexico. Unpublished Ph.D. dissertation, Department of Anthropology, University of New Mexico, Albuquerque. Kosso, P. 2001 Knowing the Past: Philosophical Issues of History and Archaeology. Humanity Books, Amherst, Massachusetts. Kowalewski, M. 1996 Time-Averaging, Overcompleteness, and the Geological Record. Journal of Geology 104:​317–​ 326. Kowalewski, M., G. A. Goodfriend, and K. W. Flessa 1998 High-Resolution Estimates of Temporal Mixing within Shell Beds: The Evils and Virtues of Time-Averaging. Paleobiology 24:287–304. Krause, R., and R. Thorne 1971 Toward a Theory of Archaeological Things. Plains Anthropologist 16:245–257. Küchler, B. 1966 Potential Natural Vegetation. U.S. Geological Survey, Washington, D.C.

192

References Cited Kuhn, S. 1995 Mousterian Lithic Technology: An Ecological ­Perspective. Princeton University Press, Princeton. Kvamme, K. L. 1985 Determining Empirical Relationships between the Natural Environment and Prehistoric Site Locations: A Hunter-Gatherer Example. In For Concordance in Archaeological Analysis: Bridging Data Structure, Quantitative Technique, and Theory, edited by C. Carr, pp. 208–238. Westport, Kansas City, Missouri. 2006 Magnetometry: Nature’s Gift to Archaeology. In Remote Sensing in Archaeology: An Explicitly North American Perspective, edited by J. K. Johnson, pp. 205–233. University of Alabama Press, Tuscaloosa. Lake, M. 1996 Archaeological Inference and the Explanation of Hominid Evolution. In The Archaeology of Human Ancestry: Power, Sex, and Tradition, edited by J. Steele and S. Shennan, pp. 184–206. Routledge, London. Larson, M. L. 1997a Housepits and Mobile Hunter-Gatherers: A Consideration of the Wyoming Evidence. Plains Anthropologist 42(161):353–369. 1997b Rethinking the Early Plains Archaic. In Changing Perspectives of the Archaic on the Northwest Plains and Rocky Mountains, edited by M. L. Larson and J. E. Francis, pp. 106–136. University of South Dakota Press, Vermillion. Leakey, M. D. 1971 Olduvai Gorge, Vol. 3. Cambridge University Press, Cambridge. Lightfoot, R. R. 1993 Abandonment Processes in Prehistoric Pueblos. In Abandonment of Settlements and Regions: Ethnoarchaeological and Archaeological Approaches, edited by C. M. Cameron and S. A. Tomka, pp. 3–​7. Cambridge University Press, Cambridge. 1994 The Duckfoot Site: Archeology of the House and Household, Vol. 2. Occasional Paper No. 4. Crow Canyon Archaeological Center, Cortez, Colorado. Love, C. 1977 Geological Influences on Prehistoric Populations of Western Wyoming. In Twenty-Ninth Annual Field Conference, pp. 15–29. Wyoming Geological Association Guidebook, Laramie. Lucas, G. 2005 The Archaeology of Time. Routledge, New York.

Lucero, L. J. 2002 Comments. Current Anthropology 43(4):S123–​S125. Ludwig, B., and J. Harris 1998 Towards a Technological Reassessment of East Africa Plio-Pleistocene Lithic Assemblages. In Early Human Behaviour in Global Context: The Rise and Diversity of the Lower Palaeolithic Record, edited by M. Petraglia and R. Korisettar, pp. 84–​107. Routledge, London. Lyman, R. L. 1994 Vertebrate Taphonomy. Cambridge University Press, Cambridge. 2006 Archaeology’s Quest for a Seat at the High Table of Anthropology. Journal of Anthropological Archaeology 26(2):133–149. Lyman, R. L., and G. L. Fox 1989 A Critical Evaluation of Bone Weathering as an Indication of Bone Assemblage Formation. Journal of Archaeological Science 16:293–317. Mabbutt, J. A. 1973 Geomorphology of Fowlers Gap Station. In Lands of Fowlers Gap Station, New South Wales, edited by J. A. Mabbutt and M. E. Sullivan, pp. 85–​121. Fowlers Gap Arid Zone Research Station Research Series 3. University of New South Wales, Sydney. Maxwell, N. 1969 Can There Be Necessary Connections between Successive Events? British Journal of Philosophical Science 19:1–25. May, R. M. 1975 Patterns of Species Abundance and Diversity. In Ecology and Evolution of Communities, edited by M. Cody and J. Diamond, pp. 81–120. Belknap, Cambridge, Massachusetts. McBurney, C. B. M. 1967 The Haua Fteah (Cyrenaica) and the Stone Age of the South-East Mediterranean. Cambridge University Press, London. McConnell, A. 1990 Micro-Morphological Analysis of Sediments from the Nunamira Cave Archaeological Site (TASI 4046), Florentine Valley, Tasmania. Unpublished MS, Silviculture/Forest Practises Unit, Forestry Commission, Tasmania, Hobart. McKern, S. T. 1987a Archaeological Investigations at 28SW5215: A Short-​Term Habitation Site at the Eastern Edge of the Green River Basin. Cultural Resource Management Report No. 33. Archaeological Services, Western Wyoming Community College, Rock Springs.

193

References Cited 1987b The Crooks Site: Salvage Excavation of an Archaic Housepit Site. Cultural Resource Management Report No. 36. Archaeological Services, Western Wyoming Community College, Rock Springs. McNabb, J. 1998 On the Move. Theory, Time Averaging and ­Resource Transport at Olduvai Gorge. In Stone Age Archaeology. Essays in Honour of John Wymer, edited by N. Ashton, F. Healy, and P. ­Petit, pp. 15–​22. Oxbow Monograph 102. Ann Arbor. Meltzer, D. 1979 Paradigms and the Nature of Change in American Archaeology. American Antiquity 44(4):​ 644–​657. Meskell, L. 2001 Archaeologies of Identity. In Archaeological Theory Today, edited by I. Hodder, pp. 187–213. Polity Press, Cambridge. Metcalfe, D., and K. M. Heath 1990 Microrefuse and Site Structure: The Hearths and Floors of the Heartbreak Hotel. American Antiquity 55(4):​781–796. Miller, D. 1985 Artefacts as Categories. Cambridge University Press, Cambridge. Miller, M. E., P. H. Sanders, and J. E. Francis (editors) 1999 The Trappers Point Site (48SU1006): Early Archaic Adaptations in the Upper Green River Basin, Wyoming. Cultural Resources Series 1. Office of the Wyoming State Archaeologist, University of Wyoming, Laramie. Mithen, S. 2000 After the Ice. Methuen, London. 2001 Archaeological Theory and Theories of Cognitive Evolution. In Archaeological Theory Today, edited by I. Hodder, pp. 98–121. Polity Press, Cambridge. Moberg, C. A. 1971 Archaeological Context and Mathematical Methods. In Mathematics in the Archaeological and Historical Sciences, edited by F. Hodson, D. Kendall, and P. Tautu, pp. 551–562. Edinburgh University Press, Edinburgh. Montet-White, A., and S. Holen (editors) 1991 Raw Material Economics among Prehistoric Hunter-Gatherers. University of Kansas Press, Lawrence. Moore, H. 1981 Bone Refuse—Possibilities for the Future. In Economic Archaeology, edited by A. Sheridan and G. N. Bailey, pp. 87–94. British Archaeological Reports International Series 96. Oxford.

Moss, J. H., K. Bryan, G. W. Holmes, L. Satterthwaite, H. P. Hansen, C. B. Schultz, and W. D. Frankforter 1951 Early Man in the Eden Valley. Museum Monographs 6. University Museum, University of Pennsylvania, Philadelphia. Munn, N. D. 1992 The Cultural Anthropology of Time: A Critical Essay. Annual Review of Anthropology 21:93– 123. Murray, T. 1987 Remembrance of Things Present: Appeals to Authority in the History and Philosophy of Archaeology. Unpublished Ph.D. dissertation, Department of Anthropology, University of Sydney, Sydney. 1992 Tasmania and the Constitution of “the Dawn of Humanity.” Antiquity 66(252):730–743. 1993 Archaeology and the Threat of the Past: Sir Henry Rider Haggard and the Acquisition of Time. World Archaeology 25:175–186. 1997 Dynamic Modeling and New Social Theory of the Mid- to Long-Term. In Time, Process and Structured Transformation in Archaeology, edited by S. E. van der Leeuw and J. McGlade, pp. 449–​ 463. Routledge, London. 1999a A Return to the “Pompeii Premise.” In Time and Archaeology, edited by T. Murray, pp. 8–27. Routledge, London. 1999b Introduction. In Time and Archaeology, edited by T. Murray, pp. 1–7. Routledge, London. 2001 On “Normalizing” the Palaeolithic: An Orthodoxy Questioned. In Studying Human Origins: Disciplinary History and Epistemology, edited by R. Corbey and W. Roebroeks, pp. 29–44. University of Amsterdam Press, Amsterdam. 2002 Evaluating Evolutionary Archaeology. World Archaeology 34:47–59. 2004 Archbishop Ussher and Archaeological Time. In The Archaeologist: Detective and Thinker, edited by L. B. Vishniackiy, A. A. Kovalev, and O. A. Scheglova, pp. 204–215. St. Petersburg University Press, St. Petersburg, Russia. 2006 Structure and Event—Again. A Response to Harding. Norwegian Archaeological Review 39(1):​ 82–89. Murray, T. (editor) 1999 Time and Archaeology. Routledge, London. Nash, S., and J. Dean 2000 The Surprisingly Deficient History of Archaeochronology. In It’s about Time: A History of Archaeological Dating in North America, edited by S. Nash, pp. 2–11. University of Utah Press, Salt Lake City.

194

References Cited Newberry, J. C., and C. Harrison 1986 The Sweetwater Creek Site. Cultural Resource Management Report No. 19. Archaeological Services, Western Wyoming Community College, Rock Springs. Newberry, J. C., and T. Hoefer III 1987 Archaeological Investigations at the McIntosh Site (48FR1468) on the Exxon Bairoil/Dakota CO2 Pipeline. Cultural Resource Management Report No. 31. Archaeological Services, Western Wyoming Community College, Rock Springs. Nicholls, G. K., and M. D. Jones 1998 Radiocarbon Dating with Temporal Order Constraints. Technical Report No. 407. Mathematics Department, University of Auckland. Electronic document, http://www.math.auckland.ac.nz/​~​ nicholls, accessed March 2007. 2001 Radiocarbon Dating with Temporal Order Constraints. Journal of the Royal Statistical Society (Series C) 50(4):503–521. Nicholson, A., and S. Cane 1991 Desert Camps: Analysis of Australian Aboriginal Proto-Historic Campsites. In Ethnoarchaeological Approaches to Mobile Campsites: HunterGatherer and Pastoralist Case Studies, edited by C. S. Gamble and W. A. Boismier, pp. 263–354. International Monographs in Prehistory, Ethnoarchaeological Series 1. Ann Arbor. O’Brien, M. J., and R. L. Lyman 2002 Evolutionary Archaeology: Current Status and Future Prospects. Evolutionary Anthropology 11:​ 26–​36. O’Connell, J. F. 1987 Alyawara Site Structure and Its Archaeological Implications. American Antiquity 52(1):74–108. Odell, G. H. 1996 Economizing Behavior and the Concept of “­Curation.” In Stone Tools: Theoretical Insights into Human Prehistory, edited by G. H. Odell, pp. 51–​80. Plenum Press, New York. Olivier, L. 1999 The Hochdorf “Princely” Grave and the Question of the Nature of Archaeological Funerary Assemblages. In Time and Archaeology, edited by T. Murray, pp. 109–138. Routledge, London. 2001 Duration, Memory, and the Nature of the Archaeological Record. In It’s about Time: The Concept of Time in Archaeology, edited by H. Karlsson, pp. 61–70. Bricoleur Press, Göteborg. Olley, J. M., T. Pietsch, and R. G. Roberts 2004 Optical Dating of Holocene Sediments from a Variety of Geomorphic Settings Using Single Grains of Quartz. Geomorphology 60:337–358.

Olzewski, T. 1999 Taking Advantage of Time-Averaging. Paleo­ biology 25:226–238. O’Neill, R. V., and A. W. King 1998 Homage to St. Michael; or, Why Are There So Many Books on Scale? In Ecological Scale: Theory and Applications, edited by D. L. Peterson and V. T. Parker, pp. 3–16. Columbia University Press, New York. Orton, C. 1993 How Many Pots Make Five? An Historical Review of Pottery Quantification. Archaeometry 35:​ 169–​184. Parkington, J., P. Nilssen, C. Reeler, and C. S. Henshil­ wood 1992 Making Sense of Space at Dunefield Midden Campsite. South African Field Archaeology 1:​63–​ 70. Pavelka, M. S. M. 2002 Change versus Improvement over Time and Our Place in Nature. Current Anthropology 43(4):S7–​ S44. Paynter, R. 2002 Time in the Valley: Narratives about Rural New England. Current Anthropology 43(4):S85– S101. Peterson, L. M., and L. A. Peterson 1995 The Bees Nest Site, Mitigation of a Multi-Component Stone Ring Site in Central North Dakota. Ethnoscience Inc., Billings. Pfaffenberger, B. 1992 Social Anthropology of Technology. Annual Review of Anthropology 21:491–516. Pinch, T. J., and W. E. Bijker 1987 The Social Construction of Facts and Artifacts: Or How the Sociology of Science and the Sociology of Technology Might Benefit Each Other. In The Social Construction of Technological Systems, edited by W. E. Bijker, T. P. Hughes, and T. Pinch, pp. 17–50. MIT Press, Cambridge, Massachusetts. Plog, F. T. 1973 Diachronic Anthropology. In Research and Theory in Current Anthropology, edited by C. Redman, pp. 181–​198. John Wiley and Sons, New York. 1974 The Study of Prehistoric Change. Academic Press, New York. Porch, N. A. 1992 Cultural and Non-Cultural Processes and the Structuring of Archaeological Deposits at Three Rockshelters in North-Central Tasmania. Unpublished B.A. (Honours) thesis, Departments

195

References Cited of Geology and Archaeology, La Trobe University, Melbourne.

Potts, R. 1991 Why the Oldowan? Plio-Pleistocene Toolmaking and the Transport of Resources. Journal of Anthropological Research 47:153–176. 2001 Behavioral Response to Variable Pleistocene Landscapes. In In the Mind’s Eye: Multidisciplinary Approaches to the Evolution of Human Cognition, edited by A. Nowell, pp. 137–153. International Monographs in Prehistory, Archeological Series 13. Ann Arbor. Potts, R., A. K. Behrensmeyer, and P. Ditchfield 1999 Paleolandscape Variation and Early Pleistocene Hominid Activities: Members 1 and 7, Olorgesailie Formation, Kenya. Journal of Human Evolution 37:​747–788. Prigogine, I., and I. Stengers 1984 Order out of Chaos: Man’s New Dialogue with Nature. Bantam Books, New York. Quigg, J. M. 1986 Ross Glen: A Besant Stone Circle Site in Southeastern Alberta. Manuscript Series No. 10. Archaeological Survey of Alberta, Edmonton. Quigg, J. M., and J. H. Brumley 1984 Stone Circles: A Review, Appraisal, and Future Directions. State Historical Society of North Dakota, Division of Archaeology and Historic Preservation, Bismarck. Raftery, A. E. 1996 Hypothesis Testing and Model Selection. In Markov Chain Monte Carlo in Practice, edited by W. R. Gilks, S. Richardson, and D. J. Spiegelhalter, pp. 163–187. Chapman and Hall/CRC, London. Ramenofsky, A. F. 1998 The Illusion of Time. In Unit Issues in Archaeology, edited by A. F. Ramenofsky and A. Steffen, pp. 74–​84. University of Utah Press, Salt Lake City. Ramenofsky, A. F., and A. Steffen 1998 Units as Tools of Measurement. In Unit Issues in Archaeology, edited by A. F. Ramenofsky and A. Steffen, pp. 3–17. University of Utah Press, Salt Lake City. Reeves, B. O. K. 1990 Communal Bison Hunters of the Northern Plains. In Hunters of the Recent Past, edited by L. B. Davis and B. O. K. Reeves, pp. 168–194. Unwin Hyman, London. Reher, C. A., and G. C. Frison 1980 Vore Site, 48CK302, a Stratified Buffalo Jump in the Wyoming Black Hills. Plains Anthropologist, Memoir 16. Lincoln.

Reid, J. Jefferson, M. B. Schiffer, and W. L. Rathje 1975 Behavioral Archaeology: Four Strategies. American Anthropologist 77:836–848. Reiss, D., and D. N. Walker 1982 Pronghorn Utilization at 48UT390 in Southwestern Wyoming. Wyoming Contributions to Anthropology 3:1–25. Renfrew, C. 1981 Space, Time and Man. Transactions of the Institute of British Geographers NS 6:257–278. 2001 Symbol before Concept: Material Engagement and the Early Development of Society. In Archaeological Theory Today, edited by I. Hodder, pp. 122–140. Polity Press, Cambridge. Reust, T. P., W. M. Harding, E. H. Schneider, B. R. McClelland, L. M. McNees, R. Weathermon, B. A. Barrows, D. W. Newton, and S. Powley 1994 Data Recovery Excavations at Site 48LN2555, Lincoln County, Wyoming. Mariah Associates, Laramie. Rhodes, E. J., C. Bronk-Ramsey, Z. Outram, C. Batt, L. Willis, S. Dockrill, and J. Bond 2003 Bayesian Methods Applied to the Interpretation of Multiple OSL Dates: High Precision Sediment Age Estimates from Old Scatness Broch Excavations, Shetland Isles. Quaternary Science Reviews 22:​1231–1244. Rice, P. M. 1984 Some Reflections on Change in Pottery Producing Systems. In The Many Dimensions of Pottery: Ceramics in Archaeology and Anthropology, edited by S. E. van der Leeuw and A. C. Pritchard, pp. 231–​291. CINGULA 7. Institute for Pre- and Proto-History, University of Amsterdam, Amsterdam. Roberts, M. B., and S. A. Parfitt 1999 Boxgrove. A Middle Pleistocene Hominid Site at Eartham Quarry, Boxgrove, West Sussex, Vol. 17. English Heritage Archaeological Report, London. Roebroeks, W. 1988 From Find Scatters to Early Hominid Behaviour: A Study of Middle Palaeolithic Riverside Settlements at Maastricht-Belvédère (the Netherlands). Analecta Praehistorica Leidensia 21. Roebroeks, W., D. De Loecker, P. Hennekens, and M. van Ieperen 1992 A “Veil of Stones”: On the Interpretation of an Early Middle Palaeolithic Low Density Scatter at Maastricht-Belvédère (the Netherlands). Analecta Praehistorica Leidensia 25:1–16. Rogers, M. J. 1996 Lithic Analysis in Landscape Archaeology:

196

References Cited ­ vi­dence from East Turkana. In Aspects of AfriE can Archaeology: Papers from the 10th Congress of the PanAfrican Association for Prehistory and Related Studies, edited by G. Twiti and R. Soper, pp. 103–​109. University of Zimbabwe Publications, ­Harare. Rogers, M. J., and J. W. Harris 1992 Recent Investigations in Landscape Archaeology at East Turkana. Nyame Akuma 38:41–47. Roos, C. I., A. P. Sullivan III, and C. McNamee In press Anthropogenic Fire for Long-Term Landscape Management: Paleoecological Evidence for Systematic Burning in Southwest Coniferous Ecosystems. In The Archaeology of Anthropogenic Environments, edited by R. Dean. Southern Illinois University Press, Carbondale. Rosenberg, A. 1975 Propter Hoc, Ergo Post Hoc. American Philosophical Quarterly 12:245–254. Roubet, C. 1997a The Lithic Domain at Klithi: Technology of Production and the Chaîne Opératoire. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 125–​153. McDonald Institute for Archaeological Research, Cambridge. 1997b The Backed Pieces at Klithi—Analysis of the 4000 Series and the Techniques of Segmentation. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. N. Bailey, pp. 155–176. McDonald Institute for Archaeological Research, Cambridge. Rowlands, M. 1993 The Role of Memory in the Transmission of Culture. World Archaeology 25:141–151. Rowley-Conwy, P. 1999 Sir Grahame Clark. In Encyclopedia of Archaeology: The Great Archaeologists, edited by T. Murray, pp. 507–529. ABC-CLIO, Santa Barbara. Sahnouni, M., K. Schick, and N. Toth 1997 An Experimental Investigation into the Nature of Faceted Limestone Spheroids in the Early Palaeolithic. Journal of Archaeological Science 24:701–713. Salmon, W. C. 1992 Explanation in Archaeology: An Update. In Metaarchaeology, edited by L. Embree, pp. 243–​ 253. Kluwer Academic Publishers, Dordrecht. Sauer, C. O. 1963 The Morphology of Landscape. Reprinted in   [1925] Land and Life: A Selection from the Writings of

Carl Ortwin Sauer, edited by J. Leighly, pp. 315–​ 350. University of California Press, Berkeley. Schiefflelin, B. B. 2002 The Dichotomizing Discourse of Multiple Temporalities. Current Anthropology 43(4):S5–S18. Schiffer, M. B. 1972 Archaeological Context and Systemic Context. American Antiquity 37:156–165. 1976 Behavioral Archeology. Academic Press, New York. 1985 Is There a “Pompeii Premise” in Archaeology? Journal of Anthropological Research 41:18–41. 1987 Formation Processes of the Archaeological Record. University of New Mexico Press, Albuquerque. 1995 Behavioral Archaeology: First Principles. University of Utah Press, Salt Lake City. Schindel, D. E. 1982 Resolution Analysis: A New Approach to the Gaps in the Fossil Record. Paleobiology 8(4):​ 340–​353. Schlanger, S. H. 1992 Recognizing Persistent Places in Anasazi Settlement Systems. In Space, Time, and Archaeological Landscapes, edited by J. Rossignol and L. Wandsnider, pp. 91–112. Plenum, New York. Shanks, M., and C. Tilley 1987 Social Theory and Archaeology. Polity Press, Cambridge. Sharrock, F. W. 1966 Prehistoric Occupation Patterns in Southwest Wyoming and Cultural Relationships with the Great Basin and Plains Culture Areas. University of Utah Anthropological Papers 77. Salt Lake City. Shennan, S. 1989 New Perspectives in Archaeology. Antiquity 63:​ 831–​836. 2002 Genes, Memes and Human History: Darwinian Archaeology and Cultural Evolution. Thames and Hudson, London. Shiner, J., S. J. Holdaway, H. Allen, and P. C. Fanning 2005 Understanding Stone Artefact Assemblage Variability in Late Holocene Contexts in Western New South Wales, Australia: Burkes Cave, Stud Creek and Fowlers Gap. In Rocking the Boat: New Approaches to Stone Artefact Reduction, Use and Classification in Australia, edited by C. Clarkson and L. Lamb, pp. 67–80. British Archaeological Reports International Series, Oxford. 2007 Burkes Cave and Flaked Stone Assemblage Variability in Western New South Wales, Australia. Australian Archaeology 64:35–45. Shott, M. J. 1989 On Tool Class Use Lives and the Formation of

197

References Cited Archaeological Assemblages. American Antiquity 54:​9–30. 1996 An Exegesis of the Curation Concept. Journal of Anthropological Research 52:259–280. 1997 Activity and Formation as Sources of Variation in Great Lakes Paleoindian Assemblages. Midcontinental Journal of Archaeology 22:197–236. 1998 Status and Role of Formation Theory in Contemporary Archaeological Practice. Journal of Archaeological Research 6:299–329. 2000 The Quantification Problem in Stone Tool Analysis. American Antiquity 65:725–738. 2003 Size as a Factor in Old World Middle Palaeolithic Assemblage Variation: A North American Perspective. In Lithic Analysis at the Millennium, edited by N. Moloney and M. Shott, pp. 137–149. Archetype, London. 2004 Hunter-Gatherer Aggregation in Theory and Evidence: The North American Paleoindian Case. In Hunter-Gatherers in Theory and Archaeology, edited by G. Crothers, pp. 68–102. Southern Illinois University Press, Carbondale. 2005 The Reduction Thesis and Its Discontents: Review of Australian Approaches. In Australian Perspectives on Lithic Reduction, Use and Classification, edited by C. Clarkson and L. Lamb, pp. 91–​112. British Archaeological Reports International Series 1408. Archaeopress, Oxford. Shott, M. J., and J. A. Ballenger 2007 Biface Reduction and the Measurement of Dalton Curation: A Southeastern United States Case Study. American Antiquity 72:153–175. Simms, S. R., and K. M. Heath 1990 Site Structure of the Orbit Inn: An Application of Ethnoarchaeology. American Antiquity 55(4):​ 797–813. Skibo, J. M. 1994 The Kalinga Cooking Pot: An Ethnoarchaeological and Experimental Study of Technological Change. In Kalinga Ethnoarchaeology: Expanding Archaeological Method and Theory, edited by W. A. Longacre and J. M. Skibo, pp. 113–126. Smithsonian Institution Press, Washington, D.C. Smith, C. S. 1988 Seeds, Weeds, and Prehistoric Hunters and Gatherers: The Plant Macrofossil Evidence from Southwest Wyoming. Plains Anthropologist 33(120):​141–158. Smith, C. S., and S. D. Creasman 1988 The Talioferro Site: 5000 Years of Prehistory in Southwest Wyoming. Cultural Resources Series No. 6. Bureau of Land Management, Vermillion, S.D.

Smith, C. S., and L. M. McNees 1999 Facilities and Hunter-Gatherer Long-Term Land Use Patterns: An Example from Southwest Wyoming. American Antiquity 64:117–136. Smith, M. A. 1989 The Case for a Resident Human Population in the Central Australian Ranges during Full Glacial Aridity. Archaeology in Oceania 24:93–105. 1996 Prehistory and Human Ecology in Central Australia: An Archaeological Perspective. In Exploring Central Australia: Society, the Environment and the 1894 Horn Expedition, edited by S. R. Morton and D. J. Mulvaney, pp. 61–73. Surrey Beatty and Sons, Chipping Norton, Oxfordshire. Smith, M. E. 1989 The Spatial Organization of Settlement at Late Post-Classic Sites. In Households and Communities: Proceedings of the 21st Annual ChacMool Conference, edited by S. MacEachern, D. J. W. Archer, and R. D. Garvin, pp. 450–459. Archaeological Association of the University of Calgary, Calgary. 1992 Braudel’s Temporal Rhythms and Chronology Theory in Archaeology. In Archaeology, Annales, and Ethnohistory, edited by A. B. Knapp, pp. 23–​ 34. Cambridge University Press, Cambridge. Stapert, D., and M. Street 1997 High Resolution or Optimum Resolution? Spatial Analysis of the Federmesser Site at Andernach, Germany. World Archaeology 29:172–194. Stark, M. T. 2003 Current Issues in Ceramic Ethnoarchaeology. Journal of Archaeological Research 11:193–242. Stein, J. K. 1987 Deposits for Archaeologists. Advances in Archaeological Method and Theory 11:337–395. 1993 Scale in Archaeology, Geosciences, and Geoarchaeology. In Effects of Scale on Archaeological and Geoscientific Perspectives, edited by J. K. Stein and A. R. Linse, pp. 1–10. Geological Society of America Special Paper 283. Boulder. 2000 Stratigraphy and Archaeological Dating. In It’s about Time: A History of Archaeological Dating in North America, edited by S. E. Nash, pp. 14–40. University of Utah Press, Salt Lake City. Stern, N. 1991 The Scatters-between-the-Patches: A Study of Early Hominid Land-Use Patterns in the Turkana Basin, Kenya. Unpublished Ph.D. dissertation, Department of Anthropology, Harvard University, Cambridge, Massachusetts. 1993 The Structure of the Lower Pleistocene Archaeological Record: A Case Study from the Koobi

198

References Cited Fora Formation in Northwest Kenya. Current Anthropology 34:201–225. 1994a The Implications of Time-Averaging for Reconstructing the Land-Use Patterns of Early ToolUsing Hominids. Journal of Human Evolution 27:​89–105. 1994b Reply to “On the Structure of the Lower Pleistocene Archaeological Record.” Current Anthropology 35:170–172. 1995 The “Blue” Tuff Locality at Koobi Fora in Northern Kenya: Archaeological Time and the Record of Past Human Behaviour. The Artefact 18:​49–​ 59. 2004 Early Hominin Activity Traces at FxJj43, a One and a Half Million Year Old Locality in the Koobi Fora Formation, in Northern Kenya: A Field Report. Proceedings of the Prehistoric Society 70:​233–258. Stern, N., and B. Marshall 1993 Excavations at Mackintosh 90/1 in Western Tasmania: A Discussion of Stratigraphy, Chronology and Site Formation. Archaeology in Oceania 28:​8–17. Stern, N., N. Porch, and I. McDougall 2002 FxJj43: A Window into a 1.5-Million-Year-Old Palaeolandscape in the Okote Member of the Koobi Fora Formation, Northern Kenya. Geoarchaeology 17:349–392. Stine, S. 1994 Extreme and Persistent Drought in California and Patagonia during Mediaeval Time. Nature 369:​546–549. Strobach, N. 1998 The Moment of Change: A Systematic History in the Philosophy of Space and Time. Kluwer Academic Publishers, Dordrecht. Stuiver, M., P. J. Reimer, and T. F. Braziunas 1998 High-Precision Radiocarbon Age Calibration for Terrestrial and Marine Samples. Radiocarbon 40(3):​1127–1151. Sturdy, D., D. Webley, and G. N. Bailey 1997 The Palaeolithic Geography of Epirus. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 2: Klithi in Its ­Local and Regional Setting, edited by G. N. Bailey, pp. 587–614. McDonald Institute for Archaeological Research, Cambridge. Sullivan, A. P., III 1978 Inference and Evidence in Archaeology: A Discussion of the Conceptual Problems. In Advances in Archaeological Method and Theory, edited by M. B. Schiffer, Vol. 1:183–222. Academic Press, New York.

1986

Occupational Dynamics of a Small Kayenta Anasazi Settlement. In Prehistory of the Upper Basin, Coconino County, Arizona, edited by A. P. Sullivan III, pp. 47–190. Arizona State Museum Archaeological Series No. 167. University of Arizona, Tucson. 1988 Prehistoric Southwestern Ceramic Manufacture: The Limitations of Current Evidence. American Antiquity 53:23–35. 1992a Investigating the Archaeological Consequences of Short-Duration Occupations. American Antiquity 57:99–115. 1992b Pinyon Nuts and Other Wild Resources in Western Anasazi Subsistence Economies. In LongTerm Subsistence Change in Prehistoric North America, edited by D. R. Croes, R. A. Hawkins, and B. L. Isaac, pp. 195–240. Research in Economic Anthropology, Supplement 6. JAI Press, Greenwich, Connecticut. 1995a Behavioral Archaeology and the Interpretation of Archaeological Variability. In Expanding Archaeology, edited by J. M. Skibo, W. H. Walker, and A. E. Nielsen, pp. 178–186. University of Utah Press, Salt Lake City. 1995b Artifact Scatters and Subsistence Organization. Journal of Field Archaeology 22:49–64. 2007 Archaeological Anthropology and Strategies of Knowledge Formation in American Archaeology. In Archaeological Anthropology: Perspectives on Method and Theory, edited by J. M. Skibo, M. W. Graves, and M. Stark, pp. 40–56. University of Arizona Press, Tucson. 2008 An Archaeological View of the Archaeological Record. In Archaeological Concepts for the Study of the Cultural Past, edited by A. P. Sullivan III, pp. 7–23. University of Utah Press, Salt Lake City. Sullivan, A. P., III, R. A. Cook, M. P. Purtill, and P. M. Uphus 2003 Economic and Land-Use Implications of Prehistoric Fire-Cracked Rock Piles, Northern Arizona. Journal of Field Archaeology 28:367–382. Surovell, T. 2003 The Behavioral Ecology of Folsom Lithic Technology. Unpublished Ph.D. dissertation, Department of Anthropology, University of Arizona, Tucson. Swedlund, A. 1978 Historical Demography as Population Ecology. Annual Review of Anthropology 7:137–173. Tani, M., and W. A. Longacre 1999 On Methods of Measuring Ceramic Uselife: A Revision of the Uselife Estimates of Cooking

199

References Cited Vessels among the Kalinga, Philippines. American Antiquity 64:299–308. Teltser, P. A. 1995 Culture History, Evolutionary Theory, and Frequency Seriation. In Evolutionary Archaeology: Methodological Issues, edited by P. A. Teltser, pp. 51–​68. University of Arizona Press, Tucson. Thomas, J. 1996 Time Culture and Identity. An Interpretive Archaeology. Routledge, London. Thompson, K. W., and J. V. Pastor 1995 People of the Sage: 10,000 Years of Occupation in Southwest Wyoming. Archaeological Services, Western Wyoming Community College, Rock Springs. Thompson, S. 2004 Side-by-Side and Back-to-Front: Exploring ­Intra-​Regional Latitudinal and Longitudinal Comparability in Survey Data. Three Cases from Metaponto, Southern Italy. In Side-by-Side Survey: Comparative Regional Studies in the Mediterranean World, edited by S. Alcock and J. Cherry, pp. 65–​85. Oxbow, Oxford. Thoms, A. V. 2007 Fire-Cracked Rock Features on Sandy Landforms in the Northern Rocky Mountains: Toward Establishing Reliable Frames of Reference for Assessing Site Integrity. Geoarchaeology 22(5):​ 477–510. Thorley, P. 1998 Shifting Location, Shifting Scale: A Regional Landscape Approach to the Prehistoric Archaeology of the Palmer River Catchment, Central Australia. Unpublished Ph.D. dissertation, Department of Anthropology, Northern Territory University, Darwin. Thornbury, W. D. 1965 Regional Geomorphology of the United States. John Wiley and Sons, New York. Tilley, C. 1981a Economy and Society: What Relationship? In Economic Archaeology, edited by A. Sheridan and G. Bailey, pp. 131–148. British Archaeological Reports International Series 96. Oxford. 1981b Conceptual Frameworks for the Explanation of Sociocultural Change. In Pattern of the Past: Studies in Honour of David Clarke, edited by I. Hodder, G. Isaac, and N. Hammond, pp. 363–​ 386. Cambridge University Press, Cambridge. 1982 Social Formation, Social Structures and Social Change. In Symbolic and Structural Archaeology, edited by I. Hodder, pp. 26–38. Cambridge University Press, Cambridge.

Tomka, S. A. 1993 Site Abandonment Behavior among Transhumant Agro-Pastoralists: The Effects of Delayed Curation on Assemblage Composition. In Abandonment of Settlements and Regions: Ethnoarchaeological and Archaeological Approaches, edited by C. M. Cameron and S. A. Tomka, pp. 11–​24. Cambridge University Press, Cambridge. Toom, D. 1996 Archeology of the Middle Missouri. In Archaeological and Bioarchaeological Resources of the Northern Plains, edited by G. Frison and R. C. Mainfort, pp. 56–76. Arkansas Archeological Survey, Fayetteville. Toth, N. 1985 The Oldowan Reassessed: A Close Look at Early Stone Artifacts. Journal of Archaeological Science 12:​101–120. Trigger, B. G. 1970 Aims in Prehistoric Archaeology. Antiquity 44(173):​26–37. 2006 A History of Archaeological Thought. 2nd ed. Cambridge University Press, Cambridge. Trostel, B. 1994 Household Pots and Possessions: An Ethnoarchaeological Study of Material Goods and Wealth. In Kalinga Ethnoarchaeology: Expanding Archaeological Method and Theory, edited by W. A. Longacre and J. M. Skibo, pp. 209–224. Smithsonian Institution Press, Washington, D.C. van der Leeuw, S. E. 1991 Variation, Variability, and Explanation in Pottery Studies. In Ceramic Ethnoarchaeology, edited by W. A. Longacre, pp. 11–39. University of Arizona Press, Tucson. van der Leeuw, S. E., and C. L. Redman 2002 Placing Archaeology at the Center of Socionatural Studies. American Antiquity 67:597–605. van Dyke, R. M., and S. E. Alcock (editors) 2003 Archaeologies of Memory. Blackwell, London. Varien, M. D. 1999 Sedentism and Mobility in a Social Landscape: Mesa Verde and Beyond. University of Arizona Press, Tucson. 2002 Persistent Communities and Mobile Households: Population Movement in the Central Mesa Verde Region, ad 950 to 1290. In Seeking the Center Place: Archaeology and Ancient Communities in the Mesa Verde Region, edited by M. D. Varien and R. H. Wilshusen, pp. 163–184. University of Utah Press, Salt Lake City.

200

References Cited Varien, M. D., and B. J. Mills 1997 Accumulations Research: Problems and Prospects for Estimating Occupation Span. Journal of Archaeological Method and Theory 4(2):141–191. Varien, M. D., and S. G. Ortman 2005 Accumulations Research in the Southwest United States: Middle-Range Theory for BigPicture­ Problems. World Archaeology 17(1):​132–​ 155. Veth, P. 1993 Islands in the Interior: The Dynamics of Prehistoric Adaptations within the Arid Zone of Australia. International Monographs in Prehistory, Archaeological Series 3. Ann Arbor. Waitkus, B. R., and D. G. Eckles 1997 Semi-Subterranean Pithouse Structures in Wyoming. In Changing Perspectives of the Archaic on the Northwest Plains and Rocky Mountains, edited by M. L. Larson and J. E. Francis, pp. 306–​333. University of South Dakota Press, Vermillion. Walker, K. R., and R. K. Bambach 1971 The Significance of Fossil Assemblages from Fine-Grained Sediments: Time-Averaged Communities. Geological Society of America Abstracts with Programs 3:783–784. Wandsnider, L. 1992 The Spatial Dimension of Time. In Space, Time, and Archaeological Landscapes, edited by J. Rossignol and L. Wandsnider, pp. 257–284. Plenum Press, New York. 1996 Describing and Comparing Archaeological Spatial Structures. Journal of Archaeological Method and Theory 3:319–384. 1998 Landscape Element Configuration, Lifespace, and Occupation History: Ethnoarchaeological Observations and Archaeological Applications. In Surface Archaeology, edited by A. P. Sullivan III, pp. 21–39. University of New Mexico Press, Albuquerque. 2003 Solving the Puzzle of the Archaeological Labyrinth: Time Perspectivism in Mediterranean Surface Archaeology. In Side-by-Side Survey: Comparative Regional Studies in the Mediterranean World, edited by S. Alcock and J. Cherry, pp. 49–​62. Oxbow Press, Oxford. 2004 Solving Meno’s Puzzle, Defeating Merlin’s Subterfuge: Bodies of Reference Knowledge and Archaeological Inference. In Processual Archaeology: Exploring Analytical Strategies, Frames of Reference, and Culture Process, edited by A. L. Johnson, pp. 315–338. Praeger, Westport, Connecticut.

Wells, L. E. 2001 A Geomorphological Approach to Reconstructing Archaeological Settlement Patterns Based on Surficial Artefact Distribution. In Earth Sciences and Archaeology, edited by P. Goldberg, V.  T. Holliday, and C. R. Ferring, pp. 107–141. Kluwer, New York. Wenban-Smith, F. F. 1997 Refitting of Lithic Artefacts. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. Bailey, pp. 95–​ 104. McDonald Institute for Archaeological Research, Cambridge. Wilcox, D. R. 1975 A Strategy for Perceiving Social Groups in Puebloan Sites. Fieldiana Anthropology 65:120–​ 174. 1982 A Set-Theory Approach to Sampling Pueblos: The Implications of Room-Set Additions at Grasshopper. In Multidisciplinary Research at Grasshopper Pueblo, Arizona, edited by W. A. Longacre, S. J. Holbrook, and M. W. Graves, pp. 19–​27. Anthropological Papers No. 40. University of Arizona Press, Tucson. 1996 Pueblo III People and Polity in Relational Context. In The Prehistoric Pueblo World, ad 1150–​ 1350, edited by M. Adler, pp. 241–254. University of Arizona Press, Tucson. Wilcox, D. R., T. R. McGuire, and C. Sternberg 1981 Snaketown Revisited. Arizona State Museum Archaeological Series No. 155. University of Arizona, Tucson. Williamson, C. 2004 Contact Archaeology and the Writing of Aboriginal History. In The Archaeology of Contact in Settler Societies, edited by T. Murray, pp. 176–199. Cambridge University Press, Cambridge. Willoughby, P. R. 1985 Spheroids and Battered Stones in the African Early Stone Age. World Archaeology 17:44–60. Wilmsen, E. 1989 Land Filled with Flies: A Political Economy of the Kalahari. University of Chicago Press, Chicago. Winder, N. 1997 Spatio-Temporal Pattern in the Klithi Assemblages. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 1: Klithi Excavations and Intra-Site Analysis, edited by G. Bailey, pp. 261–274. McDonald Institute for Archaeological Research, Cambridge. Wines, D. 1996 The Origin and Behavioural Significance of

201

References Cited ­ aunal Assemblages from Late Pleistocene F Southwest Tasmania. Unpublished B.A. (Honours) thesis, Archaeology Program, La Trobe University, Melbourne. Wobst, M. 1978 The Archaeo-Ethnography of Hunter-Gatherers or the Tyranny of the Ethnographic Record in Archaeology. American Antiquity 43:303–309. Woodward, J. 1997 Late Pleistocene Rockshelter Sedimentation at Klithi. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, Vol. 2: Klithi in Its Local and Regional Setting, edited by G. Bailey, pp. 361–376. McDonald Institute for Archaeological Research, Cambridge. Wright, B., C. Poole, and R. Balcom 1985 The Tipi Rings of Crawling Valley: Final Report. Archaeological Survey of Alberta, Edmonton. Wyckoff, J., and D. D. Kuehn 1983 Physiographic Background. In Archaeology of the Northern Border Pipeline, North Dakota, Pt. 1: Survey and Background Information, edited by

M. J. Root and M. L. Gregg, pp. 135–176. University of North Dakota, Department of Anthropology, Contribution No. 194. Northern Border Pipeline Co., Omaha. Wylie, A. 1985 The Reaction against Analogy. In Advances in Archaeological Method and Theory, edited by M. B. Schiffer, Vol. 8:63–111. Academic Press, New York. 1989 The Interpretive Dilemma. In Critical Traditions in Contemporary Archaeology, edited by V. Pinsky and A. Wylie, pp. 18–27. Cambridge University Press, Cambridge. Yellen, J. E. 1977 Archaeological Approaches to the Present. Academic Press, New York. Yoffee, N., and A. Sherrat 1993 Introduction: The Sources of Archaeological Theory. In Archaeological Theory: Who Sets the Agenda? edited by N. Yoffee and A. Sherrat, pp. 1–​9. Cambridge University Press, Cambridge.

202

Index

Page numbers in italics refer to figures and tables. Aberhan, M., 156 abandonment, 116, 122–23; of artifacts, 123; of rooms, 43 accumulation, 5, 9-11, 14, 43, 59, 62, 63, 87, 123–24, 144, 146–47; and artifacts, 7; and features, 3; and palimpsests, 14, 111; of traces, 32, 62 accumulation span, 46, 48, 59, 118, 135; and assemblage size, 49, 56 actualistic studies, 27, 162, 165–167 Adam, E., 152 Adams, R. N., 169 agglomerations of debris, 135, 148 Ahler, S. A., 97 Allen, Harry, 113, 114, 124 Allen, J. F., 26–27, 28, 136, 147, 148 Allen, K. M. S., 96 Allen Site, 62, 63, 81 Ammerman, A. J., 47 Analytical Archaeology (Clarke), 2, 18 Andrefsky, W., 47 Annales School, 1, 6–7, 16, 64, 159 anomaly (Gould’s argument of anomaly), 131 anthropological archaeology, 24 archaeological deposits, 2–5, 7, 8–9; chronology of surface artifact deposits, 115–16; as cumulative palimpsests, 40–41; formation of, 32–34, 46–47, 61; and geological deposits, 158–60; geomorphological factors in, 67, 70–76, 88, 113, 115–17; as spatial palimpsests, 39, 41, 43, 45; as temporal palimpsests, 40, 43, 131–32; temporal structure of, 10–11, 26–27, 61–64, 67, 94–96; and time averaging, 135–47, 178; as true ­palimpsests, 34, 43 archaeological epistemology, 170–71, 175, 178–80 archaeological metaphysics, 111, 171–74, 179 archaeological ontology, 171, 174, 175, 178–80 archaeological record, 1–12; and essentialist metaphysic, 5–6, 8; and formation processes, 172–80; and materialist metaphysic, 5–6, 7; as palimpsest, 1–2; and syn-

chronic view, 4, 8–11, 47, 94, 110–11, 122–24, 132; and taphonomic or formational metaphysic, 4–5, 7–8, 10, 26 archaeology and “disciplinary culture,” 172 Arnold, Philip J., III, 11, 161–69, 178 Artelius, T., 6 artifact assemblages, 10–11 artifacts and use lives, 123 Ascher, M., 32 Ascher, R., 32, 165 assemblage composition, 9, 11, 39, 40–41, 47–49, 50, 55–60, 123–25, 135, 172 assemblage formation, 9, 10–11, 31, 33–34, 45, 46, 47, 49, 56, 57 assemblage size, 47–49, 52, 53–60 assemblage variation, 46–49, 57–60 Attenbrow, V., 47, 113 Audouze, F., 158 Avonlea Complex, 95–96 Bailey, Geoff N., 13–31, 171, 179; and archaeological vari­ ation, 91; and ethnoarchaeology, 166; Klithi rock­ shelter research, 11, 150–53; and palimpsests, 33, 34, 95, 148, 177–78; and temporal structure, 62, 64, 94, 110, 134; and time perspectivism, 1–3, 46, 61, 115, 149–50, 172, 174; Time’s Arrow, 9 Ballenger, J. A., 47 Bambach, R. K., 135, 156 Bamforth, D. B., 62, 63, 81, 123 Barrett, J., 16 Bayesian analysis, 115, 118–19 Becker, M. S., 62, 123 behavioral archaeology, 3–5, 8–9, 10 Behavioral Archaeology (Schiffer), 178 Behrensmeyer, A. K., 62, 63–64, 80 Bender, B., 6, 31, 165 Benjamin, G., 21

203

index Bernstein, R. J., 179, 180 Besant Complex, 95–96 Bettinger, Robert, 66 Bettis, E. A., III, 61 Bijker, W. E., 164 Binford, Lewis, 1–2, 19; and ethnoarchaeology, 94, 161– 64, 166–68; and hunter-gatherers, 66, 96, 110; and Mask Site, 79, 123; and middle range theory, 27–28, 164, 177; New Perspectives in Archaeology, 2; and ­palimpsests, 3–5, 14, 31–32, 94; and Pompeii premise, 2, 161–63, 172, 178; and temporal grain, 62; and theory building, 61–62; and time perspectivism, 134, 172, 175–77; tool kits, 46 Bintliff, J. L., 1, 6, 7, 16, 64 biological frame of reference, 8, 16, 21, 28 Blake, E., 6, 7 Blumenschine, R. J., 51, 55, 59 bone weathering, 63, 78, 80 Boots, B., 64 Bordes, François, 46 Boughton, J., 99 Bradley, Richard, 7, 15, 16, 21, 31 Braudel, Fernand, 6 Brierley, G. J., 116–17 Brumley, J. H., 97, 102 Buffalo Hump (Wyoming), 66, 69–70, 73, 75–79, 81–88, 90–92; surface stability, 70, 75 Bunn, H. T., 135 Burkes Cave (New South Wales, Australia), 113–14, 124, 127–32 Burley, D., 102 Butzer, K. W., 64 Buzas, M. A., 48, 57–58 cairns. See stone features Calcagno, Jim, 169 Cambridge University, 13, 17–20 Camilli, E. L., 5, 96–97 Cane, S., 79 change, 164–65, 167–69 Carnap, R., 31 ceramic artifacts, 36–41, 43 chaînes opératoires, 153, 155 Chapman, J, 33 Christian, C. S., 116 chronologies, 25–26 Clark, Grahame, 15, 17–18, 25, 30; World Prehistory, 24 Clarke, David, 17–19, 27, 30, 45; Analytical Archaeology, 2, 18; and archaeological record, 33, 46; and New ­Archaeology, 4, 18; and “possibility space,” 172; and theory building, 61–62, 171 Close, Angela, 133 closed find, 10

cognitive archaeology, 24 Coles, John, 17 Colhoun, E., 136 Conard, N., 134 Cormack, J. L., 49 Cosgrove, Richard, 26–27, 28, 136–37, 143–45, 147, 148, 172 Craven, Cynthia, 92 Creasman, S. D., 66, 68, 71, 74, 76, 80, 87, 92 Crooks (Wyoming), 66, 69–71, 74–81, 90–91 cultural landscape, 97, 105, 108–9 cumulative palimpsests, 33, 34, 36–37, 39–41 curated, 52 Cutler, A. H., 156 Damm, C., 6 Daniel, Glyn, 17 Darwinian archaeology, 164 dating methods, 25; lichening, 99–103, 106–8; optically simulated luminescence (OSL), 115–17; radiocarbon dating, 113, 115, 118–19, 132; siltation, 99–103, 106–8 Dau, B. J., 102 David, N., 161, 166 Davis, P. T., 65 de facto refuse, 32 de la Torre, I., 51 de Lange, Josara, 5, 8, 149–60, 172, 177; Klithi rockshelter (Epirus, Greece), 26, 150–55, 157–60; and time averaging, 11, 135, 156–59 Deadman Wash Phase (Late Archaic), 67–69 Deal, M., 167 Dean, J., 32 Deaver, K., 95–96, 99, 100–102 DeBoer, W. R., 4, 5, 9; and the archaeological record, 31–32 deep time, 173 Deftereos, Greg, 160 depositional history, 5, 100, 105, 135–47, 152, 154–57 Developed Oldowan culture, 55 Dewar, R. E., 5, 48, 111 Dibble, H. L., 9, 46, 47, 48, 56, 123 Diehl, M. W., 36 Dooley, Mathew A., 94–109, 172; and GIS, 10, 96, 102–3; Northern Great Plains research, 95–109; and persistent places, 96, 105–7, 177; and stone features, 95–96, 99–102, 105–8; and time averaging, 178 Drennan, R. D., 52 Dunnell, R. C., 3, 8, 24, 134, 167; and evolutionary archaeology, 5; and site concept, 45, 47, 51 durational present, 16 Early Archaic, 67–69, 76, 89–91 Earman, J., 31 East Africa, 9, 46, 48, 49, 50, 51, 55–56

204

index Ebert, J., 10, 47, 51, 59, 66, 96–97 Eckerle, W. P., 63, 65, 66, 67, 73 Eckles, D. G., 66–67 Edwards, D. A., 51, 52–55, 59 Elston, R. G., 124–25, 128, 130–31, 132 Enloe, J. G., 158 environmental archaeology, 5–6, 15, 24 essentialist metaphysic, 5–6, 111 ethnoarchaeology, 3, 9, 11; actualistic research, 165–66; benefit for archaeological studies, 167–68; “reel time,” 161–68 ethnographic analogy, 8 ethnographic reconstruction, 132. See also lifeway reconstruction evenness, 47–48, 55, 57–58 evidence, construction of, 148, 156–59 evolutionary archaeology, 5–6; and disciplinary culture, 173–74 evolutionary strategy, 110, 123 extensification, 11, 167, 168 Fagan, Brian, 15, 24 Fall, P. L., 65 Fanning, Patricia C., 10, 110–33, 178; Australian research, 111–33 faunal analysis, 151–55, 157–58. See also bone weathering faunal materials, 3, 11, 63, 80–81, 139–41, 143–45; assemblage formation, 151–55, 157–59 Feldman, M. W., 47 Fenneman, N. M., 65 Ferring, C. R., 62 Fie, Shannon, 169 Finnigan, J. T., 102 fire-cracked rock, 63, 76–78, 84, 86 Fisher, J. J., 79 flagship sites, 111 Flessa, K. W., 156 Fletcher, Roland, 17, 19, 61, 159; and time perspectivism, 134, 149, 172 Foley, R. A., 27; and archaeological deposits, 3, 4, 14, 105, 172 formation theory, 46, 59–60 formational metaphysic, 4, 5, 7–8, 10, 27 Fortes, Meyer, 17 48FR1468 (Wyoming), 69 48FR1602 (Wyoming), 69 48LN1468 (Wyoming), 68 48LN2555 (Wyoming), 66, 69–70, 72, 74–75, 78–88, 90–92 48SU1006 (Wyoming), 69 48SW1242 (Wyoming), 66, 68, 70, 72, 74–75, 77–89, 91–92 48SW2590 (Wyoming), 68

48SW5057 (Wyoming), 69 48SW5175 (Wyoming), 68 48SW5215 (Wyoming), 66, 68, 72, 75, 77–79, 82–89, 91–92 48UT401 (Wyoming), 68 Fowlers Creek (New South Wales, Australia), 116–20, 122 Fowlers Gap (New South Wales, Australia), 114, 116–18, 122, 124 Fox, G. L., 80 Francis, J. E., 80 Frank, Rudy, 147 Fredlund, L. B., 96 Frison, G. C., 64, 95–97, 99, 102 Fritz, J., 31 Fry, Chris, 169 functionalist paradigm, 19 Fürsich, F. T., 156 Galanidou, N., 26, 81, 153 Galbraith, R. F., 117 Gamble, C. S., 17, 25, 134, 150, 152–54, 157–58; flagship sites, 111 Gardner, M., 31 Gaydarska, B., 33 Gell, A., 6, 174 Gellner, E., 20 geoarchaeology, 5, 8, 9, 10, 11 geographic information systems (GIS), 10, 96, 102–3 geological and geomorphological frame of reference, 4–5, 10, 16, 28, 32 geomorphological factors, 28, 61, 67, 70, 76, 88, 92; aggrading surfaces, 63–64, 70–72, 74–76, 80, 88–89, 91; deflated surfaces, 63–65, 70–71, 73–76, 80–81, 88–89, 91 Gettis, A., 64 Giddens, Anthony, 20, 30 Gilman, P. A., 76 Goodfriend, G. A., 157 Goody, Jack, 17, 20 Goreki, P., 81 Gosden, C., 22, 174 Gould, S. J., 17, 27, 131 Gowlett, J. A. J., 17, 152 Gragson, T. L., 96 Graham, I., 64 Green, S. F., 28 Gregg, M. L., 96 Grunbaum, A., 31 Hall, B. K., 164 Hanson, John A., 45 Harding, J., 16

205

index Harrell, L. L., 66–67, 68–69, 73 Harris, E. C., 32, 46 Harris, J. W., 49, 50, 51 Harrison, C., 68 Hayden, B., 47 Hayek, L. C., 48, 57–58 hearths, 38, 43, 62–63, 79–88, 91, 139–41, 143–46; artifacts associated with, 123; earth ovens, 111, 113, 117; periods of formation, 114–15, 119–20; use in dating, 115–16, 118, 120–22, 124, 132 Heath, K. M., 80 Hegmon, M., 3, 8 Heirich, M., 31 heterogeneity, 47, 55, 57–58 hierarchical causation, 15–16 Higgs, Eric, 17–19, 150 Hinsley, C. M., Jr., 164 Hiscock, P., 47, 113 historical integrity, 4–5 Hobey, J., 73 Hodder, Ian, 3, 16, 62, 165, 176; and the importance of the individual, 22; and postprocessual archaeology, 17, 19, 21 Hoefer, T., III, 66–67, 68–69 Holdaway, Simon J., 1–12, 110–33, 172; and aggradational and deflational deposits, 63; and assemblage variation, 47, 58–59; Australian research, 28, 111–33; and dating technology, 174; and the importance of the individual, 21; and integrated occupations, 61, 62; and per­sistent places, 177; Tasmanian research, 148; temporality of the landscape, 10; and time averaging, 56, 96, 178 Holen, S., 131 Holliday, V. T., 63 Holocene, 111, 113–14, 117, 122 Hopkinson, T., 22 Hosfield, R. T., 56, 59 Hudson, J., 62 Hull, K. L., 15 hunter-gatherers, 7, 48, 64, 66, 76, 136; and archaeological record, 51, 88, 110–11, 132, 153; environmental relationships, 96–97, 105, 132; occupation duration/mobility, 95, 108, 124 hypothesis testing, 149–50, 154–60 index fossils, 46, 59–60 individual, the importance of the, 21–23 Ingold, T., 7, 20, 66, 174 integrated components, 62, 76, 88–90 intensification, 11, 167, 168 interactive causation, 16 internalist schools of thought, 15 Isaac, G. Ll., 4, 46, 49, 51; and time perspectivism, 134

James, W., 15 Jansen, J. D., 116–17 Jochim, M. A., 96 Johnson, D. W., 65 Johnson, E., 63 Johnson, M., 22 Jones, M., 118–19 Jones, P. R., 50 Juell, K. E., 51, 52–55, 59 Kelly, R. L., 5, 63 Kent, Susan, 169 Kerrich, J. E., 52 Kidwell, S., 135 Kimura, Y., 50 King, A. W., 62 King, G., 28 Kirch, P. V., 61 Kleindienst, M. R., 46, 49 Klithi rockshelter (Epirus, Greece), 11, 16, 26, 28, 150–60, 151, 152; dating, 152–53; faunal analysis, 151–55, 157–58; lithic analysis, 151–55, 157–58; project history, 150; site structure, stratigraphy, 152–54 Knapp, A. B., 16, 20, 64 Kneebone, R. R., 169 Koobi Fora (Kenya), 5, 26, 49, 63, 64 Kosso, P., 32–33 Kotjabopoulou, E., 152 Kowalewski, M., 135, 156–57, 160 Kramer, C., 161, 166 Krause, Richard, 32 Kroll, E. M., 46, 135 Küchler, B., 65 Kuehn, D. D., 97 Kuhn, S., 132 Kvamme, K. L., 32, 96 Lake, M., 158 landscape perspective, 5, 6, 7, 9, 10 land systems, 116 Larson, M. L., 66–67, 76 Laslett, G. M., 117 Late Archaic, 66–69, 76, 89–90 Late Holocene, 14, 72–73, 114, 117 Late Prehistoric, 66–69, 76, 89–91 Lathrap, D. W., 9 Leach, Edmund, 17, 20 Leakey, M. D., 49, 55 lichening, 99–103, 106–8 lifeway reconstruction, 1, 4, 5, 8, 12. See also ethnographic reconstruction lithic analysis. See Stone artifacts living floors/surfaces, 8–9, 10, 32, 55, 111, 124, 132

206

index Love, C., 65, 76 Lower Okote Member (Kenya), 49, 51–52, 55–56, 58–59 Lower Paleolithic, 25, 49–51, 56 Lucas, Gavin, 6–7, 15, 21, 32, 134; and the importance of the individual, 22–23 Lucero, L. J., 165 Ludwig, B., 49, 50 Lyman, R. L., 27, 80, 164, 172

Murray, Tim, 4, 5, 11, 131–32, 170–80; and archaeological ontology, 171, 174–75, 178–79; and Australian projects, 26; and ethnoarchaeology, 8; and palimpsests, 7, 111; and Pompeii premise, 163; and postprocessual ­archaeology, 20; and structure of the archaeological record, 32, 110, 147; and temporal structure, 61, 62, 134, 170–71; and time perspectivism, 2, 19, 20, 23, 94– 95, 149, 160, 172–80

Maastricht-Belvèdére, 55 Mabbutt, J. A., 116 Mackintosh shelter (Tasmania, Australia), 135–43, 136, 137, 139, 142, 145–47; chronology, 141–43; deposits, 137–41 Mandel, R. E., 61 Marshall, B., 137–38, 141 Marshall, Y., 22 Marx, Karl, 31 Masao, F. T., 51, 55 Mask Site, 4, 79, 123 materialist metaphysic, 5–6, 7 Maxon Ranch (Wyoming), 66, 68, 70, 73, 75, 77–89, 91–92 Maxwell, N., 31 McBride, K. A., 5, 48, 111 McBurney, Charles, 17–18 McConnell, A., 144–45, 147 McIntosh site (Wyoming), 66, 69, 71, 77–79, 81–87, 90–92 McKern, S. T., 66–67, 68–69, 73 McNabb, J., 135 McNees, L. M., 63, 66–67, 92 Medieval Climatic Anomaly, 115 Meltzer, D., 171, 173–74, 179 Metcalfe, D., 80 Middle Holocene, 65, 72 Middle Paleolithic, 47, 55–56, 110 middle-range theory, 27, 99, 164, 172, 177 Midwest Taxonomic System, 170 Miller, D., 176–77 Miller, M. E., 69, 71–74, 80 Mills, B. J., 63, 87–88 Mills, D., 15 Mithen, S., 3, 22 Moberg, C. A., 32 Mode 1 technology, 55 Montet-White, A., 131 Moore, H., 20, 21 Mora, R., 51 Moss, J. H., 66 Mousterian assemblages, 46 Mulga Dam site (New South Wales, Australia), 116–18, 120–22, 124–32 Munn, N. D., 6, 31, 164

narrative time, 23–26; and archaeological approaches to time, 110 Nash, S., 32 New Archaeology, 2, 17, 18, 24, 173, 175, 178–79; and synchronic functionalist interpretation, 4, 8 New Perspectives in Archaeology (Binford and Binford), 2 New South Wales, Australia, 112 Newberry, J. C., 68–69, 71, 72 Nicholls, G., 118–19 Nicholson, A., 79 North Dakota, 95–109 Northern Great Plains, 95–109 Nunamira Cave site (Tasmania, Australia), 135–37, 136, 138, 143, 143–47; chronology, 146–47; deposits, 143–45 Nunamiut seasonal camps, 4 Nundooka site (New South Wales, Australia), 116–18, 120–22, 124–32 O’Brien, M. J., 164 O’Connell, James, 93 occupation duration, 123–24 occupation span, 123; and raw material use, 124 Odell, G.H., 131 Old Women Complex, 95 Oldowan culture, 55 Olivier, L., 9–10, 16, 61, 65, 94 Olley, J. M., 117 Olorgesailie site (Kenya), 55 Olzewski, T., 156 O’Neill, R. V., 62 Opal Phase (Early Archaic), 67–69 operational chains, 153, 155 optically simulated luminescence (OSL) dating, 115–17 Ortman, S. G., 63, 81, 87–88 Orton, C., 19, 56 paleobiology, 11, 150, 156–57 paleoecological frame of reference, 28 Paleoindian, 48, 56, 62, 66–68 palimpsests, 1–4, 7–10, 12, 13–14, 29, 31, 89, 172, 177–78; definition, 16; and global scale, 25; and the ­importance of the individual, 21; at Klithi rockshelter (Epirus, Greece), 155; in the Northern Great Plains, 96; at Site 17, Arizona, 33–45; spatial palimpsests, 95; and

207

index synchronic view, 111; and time-averaged agglomerations, 148; and time perspectivism, 23–24; and varying resolutions, 25–27; in Western New South Wales, ­Australia, 124, 131–32; in the Wyoming Basin, 61. See also cumulative palimpsests; spatial palimpsests; temporal palimpsests; true palimpsests paradigm, 173, 175, 177 Parfitt, S. A., 134, 158 Parkington, J, 134 Pastor, J. V., 67 Pavelka, M. S. M., 164 Paynter, R., 4, 6, 8, 165 PEM (post-European material), 113 People of the Earth, 24 persistent places, 58, 64, 88, 95–99, 177; in the Lower Okote Member, 58; in the Northern Great Plains, 96, 102–8; in Western New South Wales, Australia, 123– 24, 132–33; in the Wyoming Basin, 88 Peters, C. R., 59 Peterson, L. A., 96, 99 Peterson, L. M., 96, 99 Pfaffenberger, B., 164 Pietsch, T., 117 Pinch, T. J., 164 Pine Springs Phase (Late Archaic), 67–69, 76 pit structures, 63, 67–69, 76–77, 80–82, 91 place-use histories, 5, 63–64, 67, 76, 79; in the ­Northern Great Plains, 102–3, 105, 107–8; in Western New South Wales, Australia, 111, 114, 116, 123, 132; in the Wyoming Basin, 67, 76, 88, 91–92. See also persistent places Pleistocene, 11, 26, 27, 114–15, 135–36, 146–47 Plog, F. T., 8 Pompeii premise, 2, 161–63, 172, 178 Porch, N. A., 137–40, 147–48 Porter Hollow (Wyoming), 66, 68, 74–75, 77–79, 82–89, 91–92 postmodernist critique, 1 postprocessual archaeology, 1, 2, 4, 17, 19–21, 24, 94; and disciplinary culture, 173–74; and middle range theory, 177 Potts, R., 49, 51, 55 Prigogine, I., 169 primary refuse, 32, 36, 80, 81 processual archaeology, 8, 19, 21, 24; and adaptive strategies, 158; and disciplinary culture, 174; and middle range theory, 94, 177 Quigg, J. M., 96–97, 102 radiocarbon dating, 25, 63, 67, 68, 69, 70, 75, 80, 89–90, 91, 95, 113, 115, 118–19, 132, 142, 190, 191 Ramenofsky, A. F., 6, 7, 31, 62, 93

Ramsey, Bronk, 70 Redman, C. L., 2 Reeves, B. O. K., 63 Reher, C. A., 64 Reiss, D., 66 Renfrew, Colin, 7, 15, 19, 20 Reust, T. P., 69, 72, 80 Rhodes, Ed J., 10, 110–33, 178; Australian research, 111–33 richness, 47, 48, 55, 57–58, 167 risk (behavioral response), 110 Roberts, M. B., 134, 158 Roberts, R. G., 117 rock robbing. See stone features Roebroeks, W., 55–56 Rogers, M. J., 51, 55 Rolland, N., 46, 48, 56, 123 Rosenberg, A., 31 Roubet, C., 151, 153–55 Rowlands, M., 7 Rowley-Conway, P., 17 Sahnouni, M., 49 Salmon, W. C., 31 sampling paradox, 29 Sanders, P. H., 80 scatters and patches, 51–56, 59 Schiefflelin, B. B., 6 Schiffer, Michael, 9, 28, 43; behavioral archaeology, 4; Behavioral Archaeology, 178; and formation process, 27, 172, 174–75, 178; and Pompeii premise, 2, 162–63, 178; and structure of the archaeological record, 32, 105 Schindel, D., 62, 64 Schlanger, S. H., 58, 88, 96–97; analysis of variability, 111; and persistent place, 123, 132–33 Schroeder, Sissel, 45 secondary refuse, 32, 96 settlement system, 110, 114, 123, 132 Shanks, M., 20 Sharrock, Floyd, 63, 76 Shennan, S., 2, 6 Sherrat, A., 2 Shiner, Justin, 115, 124, 125, 127, 131, 133 Shott, Michael J., 46–60, 110; artifact use-life histories, 9–10; and assemblages, 46–60, 123–24, 125, 130, 147; Koobi Fora (Kenya), 49–52, 55–56, 58–59; and scatters and patches, 51–56, 59; and site formation process, 172, 177 siltation, 99–103, 106–8 Simms, S. R., 80 single component site. See living floors/surfaces Site 17 (Arizona), 33–44, 35, 36, 38, 39–40, 41, 42, 44 site concept, 46–47, 51, 55–56, 58–59 site formation processes, 135, 145, 172–80

208

index Skibo, J. M., 167 Smith, C. S., 66–67, 68, 71, 74, 80, 92 Smith, M. A., 114 Smith, M. E., 3, 6, 7; and Pompeii premise, 162–63; and temporal processes, 64, 94, 167 source-side knowledge, 27 spatial palimpsests, 29, 32, 33, 39–41, 43–45, 61, 95, 148 Stafford, T. W., Jr., 63 Stapert, D., 158 Stark, M. T., 166 Steffen, A., 62 Stein, J. K., 2, 3, 5, 8, 93 Stengers, I., 169 Stern, Nicola, 124, 125, 129, 134–48, 158–59, 160; and Lower Okote Member, 49, 51, 59, 63, 135–47; and site formation process, 172, 175, 177; and temporal structure, 62, 64, 110; and time averaging, 6, 11, 26–28, 124, 135, 141–43, 145–48; and time perspectivism, 5, 111, 172 Stewart, G. A., 116 Stine, S., 115 stone artifacts, 3, 10, 26; assemblage formation, 124–25, 141, 143–44, 151–54, 155, 157–58, 159; assemblage vari­ ation, 46, 47–55; as cumulative palimpsest, 34–37, 39–​41, 43; debitage density, 76–78, 81, 87, 88; and ­occupation duration, 123–27; raw materials and tool formation, 127–32; resharpening, 3, 9; and “scatters and patches”, 51–56; use in dating, 99, 113–14 stone features, 95–96, 99–102, 105–8; cairns, 95, 99, 106– 7; rock robbing, 101–2; stone rings, 95, 97, 99, 101–2, 106–8 Street, M., 158 Strickland, H. C., 79 Strickland, W., 79 Strobach, N., 162 Stud Creek (New South Wales, Australia), 114–15, 116, 118, 120, 122, 124–31 Sturdy, D., 28 subject-side phenomena, 27 Sullivan, Alan P., III, 4, 29, 31–45, 62; and palimpsests, 31, 33–45; and place-use histories, 5; and Site 17, Arizona, 33–45; and site formation process, 175, 177; and trace loads, 9, 178; and use-life histories, 3 supplanting, 167 supplementing, 167 Surovell, T., 54 Sweetwater Creek (Wyoming), 66, 68, 72, 77–85, 87–89, 91–92 synchronic lifeways, 111, 115

taskscape, 7 Tasmania, 11, 25, 28, 135–47, 136 Teltser, P. A., 164 temporal logics, 165–67 temporal palimpsests, 33, 40, 45 temporal process, 11, 64, 94 temporal structure of deposits, 5, 61–64, 67, 94, 116, 123; geomorphological factors, 67, 70–76; place-use ­histories, 63–64, 67, 76, 78–79 Thomas, David Hurst, 66 Thomas, J., 174 Thompson, K. W., 66, 67, 76, 87 Thompson, S., 51 Thorley, P., 111, 114 Thornbury, W. D., 65 Thorne, Robert M., 32 Three Age System, 170 Tilley, C., 20, 21, 176, 179, 180 time, social conceptualization, 6–7 time-averaged deposits, 5, 6, 7, 56, 62, 64, 135; at Klithi rockshelter (Epirus, Greece), 11, 157–60; and paleo­ biology, 150, 156–57; paleontological deposits, 8 time averaging, 5, 8, 11, 45, 55, 56, 59, 62, 96, 109, 124, 134– 35, 146, 147, 148; of archaeological record, 146, 147, 157–59; and construction of evidence, 156–60; disharmonious, 156–57; of fossil records, 156–57; significant and insignificant, 156–58 time perspectivism, 1–4, 5, 7, 9, 13–33; definitions, 15, 30; first phase of, 172, 174–75, 177–80; implementation, 26–29, 45, 46, 111, 115, 149–50, 156, 159; structure of the archaeological record (palimpsest), 13–14 time resolution (palimpsest), 13–14, 16 time scale, 13–14 Time’s Arrow, 9, 31, 170 tool assemblage variation, 127–30 Toth, N., 49 trace loads, 32, 33–34, 37, 40, 43, 45, 178 Trappers Point (Wyoming), 66, 69–70, 73, 75, 77–88, 90–92 Trigger, B. G., 17, 31 Trostel, B., 167 true palimpsests, 33, 34, 43 typology, 47, 49–51 tyranny of lifeways interpretation, 114

Talioferro (Wyoming), 66, 68, 70–71, 74–89, 91–92 taphochronometric indicators, 62–64, 76, 78, 92 taphonomic processes, 8, 27

van der Leeuw, S. E., 2, 165 Varien, M. D., 63, 81, 87–88, 92 Veth, P., 111, 114

Uinta Phase (Late Prehistoric), 67–69 uniformitarianism, 3, 16–17, 27 Upper Paleolithic, 23, 150–51 use-life histories, 3, 9–10, 48, 55–57, 59–60, 123–24

209

index Waitkus, B. R., 66–67 Walker, D. N., 66 Walker, K. R., 135, 156 Wandsnider, LuAnn, 1–12, 29, 58, 60, 61–93, 97, 111, 114, 133, 169; depositional history, 5, 10, 178; and discard behavior, 123; and essentialist metaphysic, 111; and temporal structure, 108, 110, 177; and time averaging, 178; and time perspectivism, 172 Wells, L. E., 115 Wenban-Smith, F. F., 153, 158 White, M. J., 22 Wilcox, David R., 32, 34 Williamson, C., 172 Willoughby, P. R., 50 Winder, N., 153, 158

Wines, D., 140–41, 147 Wobst, M., 161, 164; and ethnoarchaeology, 166 Woodward, J., 151, 152–53 World Prehistory (Clark), 24 Wright, B., 102 Wyckoff, J., 97 Wylie, A., 62, 165 Wyoming Basin, 10, 61–62, 64–88, 92, 102 Yellen, J. E., 63–64, 79, 88, 91 Yoffee, N., 2 Yoshida, H., 117 Zielinski, G. A., 65

210