Experimental Archaeology: Replicating past objects, behaviors, and processes 9781841714158, 9781407324197

Citation preview

BAR S1035 2002

Experimental Archaeology Replicating past objects, behaviors, and processes MATHIEU (Ed): EXPERIMENTAL ARCHAEOLOGY

Edited by

James R. Mathieu

BAR International Series 1035 2002 B A R

Experimental Archaeology Replicating past objects, behaviors, and processes

Edited by

James R. Mathieu

BAR International Series 1035 2002

Published in 2016 by BAR Publishing, Oxford

BAR International Series 1035 Experimental Archaeology

© The editor and contributors severally and the Publisher 2002 The authors' moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781841714158 paperback ISBN 9781407324197 e-format DOI https://doi.org/10.30861/9781841714158 A catalogue record for this book is available from the British Library

BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 197 4 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd/ Hadrian Books Ltd, the Series principal publisher, in 2002. This present volume is published by BAR Publishing, 2016.

BAR

PUBLISHING BAR titles are available from:

EMAIL

PHONE FAX

BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

Table of Contents List of Contributors ..................................................................................................

ii

Preface ...................................................................................................................

iii

Introduction - Experimental Archaeology: Replicating Past Objects, Behaviors, and Processes - James R. Mathieu ............................................................................................

1

SECTION ONE: OBJECT REPLICATION Monitoring Developments: Replicas and Reproducibility - Genevieve LeMoine ........................................................................................

13

Musical Behaviors and the Archaeological Record: A Preliminary Study - Ian Cross, Ezra B. W Zubrow, and Frank Cowan ........................................... 25 Distinguishing Metal (Steel and Low-Tin Bronze) from Stone (Flint and Obsidian) Tool Cut Marks on Bone: An Experimental Approach - Haskel J. Greenfield ......................................................................................

35

SECTION TWO: BEHAVIORAL REPLICATION Controlled Experiments with Middle Paleolithic Spear Points: Levallois Points - John J. Shea, Kyle S. Brown, and Zachary J. Davis ...................................... 55 Reconceptualizing Experimental Archaeology: Assessing the Process of Experimentation - James R. Mathieu and Daniel A. Meyer ..........................................................

73

Seeing What Is Not There: Reconstructing the Monumental Experience -Alexei Vranich .....................................................................................................

83

Space, Place and lnka Domination in Northwest Argentina - Chad Gifford and Felix Acuto .........................................................................

95

SECTION THREE: PROCESS REPLICATION This Little Piggy Went to Cumbria, This Little Piggy Went to Wales: The Tales of 12 Piglets in Peat - Heather Gill-Robinson ...................................................................................

111

SMELT: Economies of Scale - Carl Blair ......................................................................................................

127

Experimenting with Continental Migration - Ezra B. W. Zubrow .......................................................................................

143

Listof Contributors FelixAcuto

Haskel J. Greenfield

Department of Anthropology

University of Manitoba

SUNY-Binghamton

Department of Anthropology

Binghamton, NY 13902-6000, USA

Fletcher Argue 435

Email: [email protected]

Winnipeg, ManitobaR3T 5V5 Canada

Carl Blair

Email: [email protected]

Dept. of Social Sciences Michigan Technical University

Genevieve LeMoine

1400 Townsend Dr.

The Peary-MacMillan Arctic Museum

Houghton, Ml 49931, USA

Bowdoin College

Email: [email protected]

9500 College Station BrunswickME0401 l-8495, USA

Kyle S. Brown

Email: [email protected]

Anthropology Department SUNY- Stony Brook

James R. Mathieu

Stony Brook, NY 11794-4364, USA

Dept. of Anthropology University of Pennsylvania Museum

Frank Cowan

3260 South Street

Curator of Archaeology

Philadelphia,PA 19103-6398, USA

Cincinnati Museum Center

Email: [email protected]

1301 Western Avenue Cincinnati, OH 45203, USA

Daniel A. Meyer

Email: fcowan9l [email protected]

7 Calandar Road NW Calgary, Alberta T2L OPS

Ian Cross

Canada

Fellow of Wolfson College Cambridge University

Email: [email protected]

West Road

JohnJ.Shea

Cambridge CB3 9DP

Anthropology Department

United Kingdom

SUNY-Stony Brook

Email: [email protected]

Stony Brook, NY 11794-4364, USA Email: [email protected]

Zachary J. Davis

Anthropology Department

Alexei Vranich

SUNY-Stony Brook

Department of Anthropology

Stony Brook, NY 11794-4364, USA

University of Pennsylvania Museum

Chad Gifford

3260 South Street

241 Central Park West, 14B

Philadelphia,PA 19104-6398, USA

NewYork,NY 10024, USA

Email: [email protected]

Email: [email protected]

Ezra B.W. Zubrow

Heather Gill-Robinson

Department of Anthropology

33 Third Avenue

MFAC380

Trenton, Ontario K8V 5N 1

SUNY-Buffalo

Canada

Buffalo,NY 14261-0005, USA

Email: [email protected]

Email: [email protected]

ii

Preface This volume grew out of a symposium session on Experimental Archaeology that was coorganized by James R. Mathieu and Andrew W. Pekin for the Society for American Archaeology annual meetings in Chicago, Illinois in 1999. Of the original eight papers presented at that session, revised versions of six are included here (Blair, Gill-Robinson, Greenfield, LeMoine, Mathieu and Meyer, and Shea et al.). The five other papers consist of an Introduction by the editor and four papers that were solicited after the session (Cross et al., Gifford and Acuto, Vranich, and Zubrow) in order to round out the volume's coverage. Please note that the choice of these additional papers, as well as the overall organization of the volume, were made in part to further elaborate the editor's new typology of experimental archaeological research (see the Introduction), but more importantly, to create a "Reader on Experimental Archaeology" that could be used when teaching courses on the theory and method of anthropological and archaeological interpretation. It is hoped that this volume will provide a clearer understanding of this subfield, show its breadth and vitality as practiced today, and in particular, provide a good starting point for anyone interested in developing and using analogies in archaeological interpretation. During the three years it has taken to put together this volume, I have received the support and encouragement of a number of people. Here I would like to thank especially the authors of the various papers and David Davison of BAR for their patience during this process. I would also like to thank the Kolb Foundation of the University of Pennsylvania Museum for their financial support during my graduate student years and my friends and colleagues at Penn who encouraged my interest in experimental archaeology during various discussions - in particular, my wife Praveena Gullapalli was always there to offer her patience, support, and encouragement when I really needed it. Finally, I would like to express my gratitude to Bernard Wailes for the initial stimulation he provided and his continuing support throughout the years. It is to Bernard that I dedicate this volume - may the road rise with you. James R. Mathieu

iii

Introduction - Experimental Archaeology: Replicating Past Objects, Behaviors, and Processes James R. Mathieu

Abstract

Experimental archaeology is a sub-field of archaeological research which employs a number of different methods, techniques, analyses, and approaches within the context of a controllable imitative experiment to replicate past phenomena (from objects to systems) in order to generate and test hypotheses to provide or enhance analogies for archaeological interpretation. Using a new hierarchical typology of experimental archaeological research, the linkages within and between various types of experiments are delineated and used to suggest new forms of experimental archaeological research.

The Definition and Purpose of Experimental Archaeology

Experimental Archaeology is a sub-field of archaeological research which employs a number of different methods, techniques, analyses, and approaches within the context of a controllable imitative experiment to replicatepast phenomena (from objects to systems) in order to generate and test hypotheses to provide or enhance analogies for archaeological interpretation. This definition is derived from an assessment of the (limited) theoretical literature concerning experimental archaeology (cf. Amick et al. 1989:1; Ascher 196la:793; Coles 1973:13; Ingersoll and Macdonald 1977:xii; Malina 1983; Schiffer et al. 1994: 198; Skibo 1992b). By emphasizing the purpose of experimentation - i.e. the generation and testing of analogies to be used in archaeological interpretation - this definition goes beyond the basic definitions of 'experiment' or 'experimental' . 1 This is intended to clarify some of the misunderstandings currently associated with experimental archaeology. In particular, it is hoped that this definition will dispel the belief that experimental archaeology is solely concerned with issues of technology, subsistence, and the physical properties of materials. 2 Webster's New Collegiate Dictionary (1981 :399) defines 'experiment' and 'experimental' as follows: "experiment n la: TEST, TRIAL b : a tentative procedure or policy c : an operation carried out under controlled conditions in order to discover an unknown effect or law, to test or establish a hypothesis, or to illustrate a known law 2 : EXPERIENCE 3 : the process of testing: EXPERIMENTATION"; and "experimental adj 1 : of, relating to, or based on experience: EMPIRICAL 2 : founded on or derived from experiment 3a : serving the ends of or used as a means of experimentation b : relating to or having the characteristics of experiment: TENTATIVE". 2 This impression has resulted from two related phenomena. One, the bulk of previous research undertaken under the name of experimental archaeology has dealt with subsistence practices and technology, rather than social organization, political organization, and religious beliefs (Ascher 196la:793; Coles 1973:13, 18; 1977:233; 1979:243). And two, many experimenters feel that experimental archaeology's primary aim should concern either "exposing the basic-science underpinnings of dead technologies" (Schiffer and Skibo 1987:609; Schiffer et al. 1994:21O; Skibo l 992b:29) or the quantification of technological matters such as the structural qualities of housing, the functional capability of tools, the performance of implements, and/or the procedures for making objects (Coles 1979:246; cf. Saraydar and Shimada 1973:344-345).

There are four main aspects to this definition. First, the definition emphasizes the context within which research is pursued - i.e. the controllable experiment. In general, experimental research is pursued in an artificial environment where stress is placed more on the control of variables and the understanding of their impact, than on the authenticity of the context of analysis (cf. Hodder 1982:2930; Odell and Cowan 1986: 196; Shimada 1978:212; Spector 1981:7-10). However, the degree to which variables are controlled can vary, as can the degree to which the context of analysis is intended to be authentic (see below). Second, the definition emphasizes that experimental research replicates past phenomena. In other words, replicative experimentation is the hallmark of experimental archaeology as practiced ( cf. the imitative experiments of Ascher 196la:793 and the controlled replication of Ingersoll and Macdonald 1977:xiii). In other words, all experimental research is concerned at some level with the replication of phenomena - from artifacts at the lowest level, to behaviors, processes, and even entire systems at increasingly higher levels. These various phenomena and the implications of conceiving of experimental archaeology in terms of these will be discussed further in the next section. Third, the definition emphasizes the results of experimental research - i.e. the generation and testing of hypotheses. In other words, research is pursued in a scientific vein, trying to develop new research questions while also evaluating current ones (see Beveridge 1957 for a discussion of the art of scientific investigation). For example, experimental research, however loosely or tightly controlled, generates and tests (assesses) hypotheses about materials, behaviors, and/ or beliefs (Ascher 196la:793, 795; Coles 1973:13; Ingersoll and Macdonald 1977:xii; Jewell 1963:12). It does this by allowing researchers to • observe "processes involved in the production, use, discard, deterioration, or recovery of material culture" (Skibo 1992a: 18 cited in Schiffer et al. 1994: 198), • "gain an insight into the importance [meaning] of these objects to their original inventors and owners" (Coles 1979:1-2), and

• generally become aware of, and gain a better appreciation and understanding of, past human enterprises, problems, abilities, and choices (Coles 1979: 1-2).

The typology of experimental archaeological research used here is illustrated in Figure 1.4

Finally, the definition emphasizes the ultimate purpose of experimental research - i.e. the generation of analogies to be used in archaeological interpretation (Charlton 1981: 146; see Stahl 1993 and Wylie 1985 especially for the role of analogy in archaeology, as well as Ascher 1961b; Gould 1980:29-36; Gould and Watson 1982; Hodder 1982:9,11-27; Kleindienst and Watson 1956; Orme 1973, 1974, 1981; Thompson 1956; and Wylie 1982, 1989). In other words, experimental archaeology has a specific purpose and does not encompass all research which simply employs controllable experiments and hypothesis testing ( e.g. methodological experiments - see below).

Object Replication

A Typology of Experimental Archaeological Research

Having stated that the purpose of experimental archaeology is to generate analogies to be used in archaeologicalinterpretation, one might ask: How does experimental archaeology do this? The answer is: via the replication and/or simulation of past objects/materials, behaviors (Ascher 1961a:795; Schiffer et al. 1994:198 citing Skibo 1992a:18), processes, and even entire social systems. At some level, all experimental research is concerned with the replication of past phenomena. This is an important point. By acknowledging that experimental archaeology revolves around replication and simulation, it logically follows that all replicative or simulative activities can be considered (potential) experiments. Therefore, it is suggested here that a better appreciation of experimental archaeology and its potential contributions to our understanding of the past can be gained by considering the types of phenomena that can be replicated or simulated. As stated above, these include in the order o_fincreasing scale: objects, behaviors, processes, and systems. Furthermore, by grouping experimental archaeological research by the scale of the phenomena it replicates, a hierarchical typology can be developed which delineates the linkages within and between certain types of experiments. In other words, one can gain a better appreciation of how higher level experiments incorporate lower level experiments, or in fact, how many higher level experiments merely assume certain lower level information which itself might need experimentation to substantiate. 3 Moreover, using a hierarchical typology based on the type and scale of phenomena replicated, it becomes possible to conceive of new experiments that could be undertaken on similar phenomena, in particular, those beyond experimental archaeology's traditional emphasis on subsistence and technology. 3

Note that in the scheme presented here, most traditional replicative experiments are considered to be lower level ones. This scheme was influenced by Amick et al.'s (1989:2) concept of scales of analysis, though note that their high scale equates to the lower scale here.

This sort of replication involves the production of a replica that has certain characteristics in common with the original object. This may be done to minimize the cost and effort expended on producing the replica, but it could also be a result of the desire to emphasize only one (or a few) aspects of the original. Visual Replicas

The most common example of this type of replication is a visual replica (Coles 1977:235, 1979:36-43 prefers the term simulation to describe visual replicas). These replicas are most often used to fulfil educational, archival, and/or commemorative functions, providing visual information about real artifacts which can be used to teach, record, and/or symbolically express certain significant aspects of the original. Since the production of other authentic aspects of the original would probably require a lot more time, money and/or effort, visual replicas rarely consist of authentic materials and are often made with modem technology (e.g. plaster casts of artifacts). A particularly good example of the production and varied use of visual replicas is LeMoine's (this volume) tape impressions of the microwear found on bone tools. These visual replicas can be used to study the wear patterns found on the bone, without directly subjecting the bone artifacts to excessive treatment and/or damage. Furthermore, these replicas archive microwear data in a cheap, standardized, and readily accessible format, allowing an easy way for 4

Others prefer to categorize experimental archaeological research differently. For example, Robert Ascher (196la:793) divides experimental research based on the sorts of activities the researcher undertakes: (I) imaginative experiments involve purely cerebral activities, (2) imitative experiments require replicative behavior on the part of the researcher, while (3) comparative experiments involve the comparison of the results of the imitative experiments. In contrast, John Coles (1973, 1979) prefers to group experimental research around topics of interest or subsets of human behavior (e.g. food production or subsistence, settlement, industry, arts and crafts, discovery and exploration, or life and death). This is especially useful for those interested in the relevance of experimental research to specific research questions. Another scheme the current author uses when teaching groups research by the primary materials under study (e.g. experiments with flint, earth, wood, ceramics, stone, etc.). This allows one to somewhat limit the range of discussion while still encouraging consideration of the wide variety of experimental research that could be undertaken with particular materials and/or behaviors. The typology used here would seem to be prefigured by Colcs's ( 1977:235-240, 1979:36-43) "levels of experiment" (e.g. display, production, and function). However, his discussion of these actually describes a linear scheme where experimental research grows from (a) simple simulations which arc purely visual replicas intended for display, to (b) more accurate replicas, "correctly" manufactured with all the appropriate materials and techniques to allow for an understanding of production, to (c) the use of these replicas to test their function, to (d) the repetition of experiments to augment results, to (e) the execution of series tests to build on previous experimental results, to finally, (t) the derivation of social interpretations derived from experimental research. This organization is mainly a recipe for procedure, rather than a typology.

James

microwear analysts to record, learn, and study diagnostic wear patterns from a variety of archaeological and experimental contexts. Functional Replicas

Another type of object replication is the creation of a functional replica. These are replicas that are usually intended to be used in a manner similar to the original object. Therefore, it is necessary that they be accurate with respect to their functional aspects and so they are normally created using functionally appropriate materials and sometimes production techniques. However, functional replicas need not be authentic in every respect, only in those aspects which are deemed relevant to the object's normal functioning. 5 An example of the production and use of a functional replica can be seen in Cross et al.'s (this volume) study of the possible musical capabilities of Upper Paleolithic flint blades. By making replicas of ancient flint blades and then "playing" them to produce recognizable musical notes and tones (a behavioral replication - see below), they use experimental archaeology to show that certain prehistoric artifacts could have been used to make music. Furthermore, by observing the microwear patterns that result on the blades from playing, they identify a form of evidence, which if found on archaeological blades, could be used to indicate past musical behavior.

Using Real Artifacts or "Full Replicas"

This type of object replication requires either the accurate production of a.full replica, using all authentic materials and appropriate techniques for construction (Coles 1977:236; 1979:36-43)7, or a real artifact. Since this sort ofreplication requires a very high degree of accuracy or the use of a somewhat unique artifact, it is often the most expensive and least pursued form of replication. This sort of replication is often employed to generate and test hypotheses concerning the creation and production of phenomena such as artifacts, structures, and/or their associated features and debris (Ingersoll and Macdonald 1977:xiii; Tringham 1978:182). A good example of this sort of replication is Greenfield's (this volume) study of metal and stone tool cut marks on bones. In his study, he makes full replicas of ancient bone cut marks using chipped stone tools and metal knives. Like most microwear analysts, he uses experimental archaeology to produce these real wear patterns in order to understand the types of marks produced by different types of tools. With this comparative information, he is able to identify the different archaeological signatures of stone and metal cutting implements. By applying this information to the analysis of archaeologically recovered bones with cut marks, he is able to track the ubiquity and spread of metallurgy across time in southeastern Europe.

Behavioral Replication

Another interesting example of the use of functional replicas is Andrew Pelcin's ( 1997a, 1997b, 1997c, 1998) experiments concerning flintknapping. 6 In his experiments, standard glass plates and glass lenses are used as functional replicas of flint and obsidian cores. Since each of these materials fracture conchoidally, Pelcin can test hypotheses about the general nature and/or capabilities of all flakeable lithic materials by simulating flintknapping on the glass cores. Furthermore, by using these standardized functional replicas, Pelcin is able to control most of the variables that normally plague traditional experiments with real flint and obsidian cores (e.g. surface morphology and natural defects in lithic material). This suggests that functional replicas in some instances are actually preferable to using real artifacts or specific materials (see below). Overall, though the production of object replicas can lead to the generation of analogies to be used in archaeological interpretation, in most cases, they tend to be used for the purposes noted above (e.g. educational, archival, commemorative) or as necessary prerequisites for the pursuit of the higher level experiments to be discussed below. In other words, functional replicas are often produced as a first step in the higher level of behavioral replication ( cf. Coles 1979:38-39). 5

6

For example, to test the efficiency of a bronze axe head for tree felling, one does not necessarily need to haft the axe head in an authentic prehistoric haft (Mathieu and Meyer 1997). Pelcin's experiments are an example of process replication (see below) - i.e. he replicates the process oftlintknapping by dropping ball bearings onto glass cores to detach flakes.

This type of replication involves the reproduction of past behaviors and activities, including the reproduction of past techniques of use (cf. Tringham's 1978: 182 behavioral experimentation). Since many past activities involved the use of material culture, this type of replication often requires the use of functional replicas, 'full replicas", or real artifacts (Coles 1977:238, 1979:36-43). Functional Replication

The most common type ofbehavioral replication is.functional replication. Typically, this is used to generate and test hypotheses concerning the function and use of particular objects in the past, (Coles 1977:238, 1979:36-43), especially tools and weapons (e.g. Frison 1989; Hansen 1969; Jones and Renn 1982; Odell and Cowan 1986; Tarver 1995), but also instruments (e.g. Cross et al., this volume) and even larger constructions like boats (e.g. Johnstone 1972; Marstander 1976) and architecture (e.g. Jones and Renn 1982; Watson and Keating 1999).8 A classic example of this sort of behavioral replication is Shea, Brown, and Davis's (this volume) experiment which 7

All replication is in some way limited - a truly exact duplicate can never be achieved. Coates ct al. ( 1995:298) refer to accurate boat replicas as Full-Scale Reconstructions.

8

More generally, functional replication can be thought of as replicating any human behavior which involves "doing" something (e.g. cutting, building, hunting, etc.), rather than "perceiving" or "experiencing" something. The section on phenomenological studies below discusses the experimentation with these latter behaviors.

replicates Middle Paleolithic spear hunting behavior. By employing functional replicas of Levallois points and a crossbow that simulates the thrusting behavior of a Middle Paleolithic hunter, they are able to show that Levallois points are very effective as spear points. Furthermore, by analyzing the wear and breakage patterns of their replicated points, they are able to make suggestions concerning the data on archaeologically-recovered points that needs to be collected and recorded in the future in order to determine if these points were indeed used as spear tips.

Typically, such studies lack the control of certain types of variables (e.g. cultural aspects of perception) and rarely repeat the "experiment" by having a number ofwalk-throughs that vary some of the parameters such as architectural elements or different people playing the role of"ego". This is probably related to the fact that these studies are not usually considered experiments. However, by realizing that they are performing a behavioral replication, it would not take much to reconceive them in experimental terms. As such, phenomenological studies could be pursued in a more controlled manner, identifying the different variables at play. In particular, architectural spaces could be reconstructed in a number of alternative ways and walk-throughs could be used to explore the possibilities of each reconstruction, generating and testing different hypotheses about the use and function of space, as well as the power of symbolism. Furthermore, by varying the characteristics of those doing the walk-through (in terms of age, gender, status, cultural background, etc.), one would be better able to identify some of the possible variation in perception. For example, do different types of people perceive space and architecture in different ways? What is the variation based on gender, ethnic, social, and/or other cultural differences? Can we identify a "standard" perception, as well as any significant "alternative" ones?

Comparative Experiments

Another form of behavioral replication is what Ascher (196la:793) refers to as comparative experiments. These basically compare the results of different behavioral replications. For example, comparing the relative efficiency of the use of two or more materials, objects, and/or techniques, particularly to understand the choices people made with regard to adopting one over the other (e.g. Mathieu and Meyer 1997; Saraydar and Shimada 1971, 1973; Schiffer et al. 1994). An example of such an experiment is Mathieu and Meyer's (this volume) comparative assessment of the efficiency of stone, bronze, and steel axe heads. Using real Neolithic stone axe heads and functional replicas of bronze and steel axes, they replicate tree felling behavior and through comparison of their data are able to show that bronze axes were just as efficient as steel axes for felling trees, and that both types of metal axe were more efficient than stone axes. Though this would support arguments for a clear technological choice when deciding between stone axes and metal axes, their experiments raise questions concerning the reasons for opting for an iron-based technology over a bronze one, and more significantly, questions concerning the general importance of technological issues in human decision making. This leads them to reconceptualize the role experimental archaeology can and should play in archaeological interpretation. In particular, they suggest that the subfield's strength lies in its ability to provide a fuller understanding of the context within which past activities occurred.

Two good examples of phenomenological studies are Vranich's (this volume) and Gifford and Acuto's (this volume) threedimensional (3-D) visual reconstruction and analysis of architectural spaces in the South American highlands. In the first paper, Vranich presents a 3-D reconstruction of the Pumapunku Complex at Tiwanaku and takes the reader on a visual tour of the structure, viewing it from different directions in order to determine which side likely functioned as the main entrance. His interpretation that the west side facing away from the center of Tiwanaku was the main entrance turns the traditional understanding of the complex on its head. By visually replicating the appearance of the structure, he is able to suggest that it functioned as a linear complex that channeled pilgrims from outside Tiwanaku to the main plaza within the city. As such, it was a center of relatively public ritual, rather than the more exclusive, elitecentered space others had suggested it to be.

Phenomenological Studies

A third form of behavioral replication includes research typically called phenomenological studies - i.e. the replication of people sensing, perceiving, or feeling certain things. Though these have rarely, if ever, been considered a form of experimental archaeology, their underlying premise is the same -i.e. that researchers can replicate the perceptive behavior (perception) of past people. An example of this can be seen in architectural studies which place the reader in the shoes of a past person and walk them through a building, i.e. recreating the experience of moving through and perceiving that structure as people would have in the past. Other examples of phenomenological studies include replicating the visual perception and experience of past landscapes (e.g. Forte and Siliotti 1997; Tilley 1994) and studies which replicate human perception via the other four senses: smell, touch, sound, and taste (e.g. Sanders 1990; Watson and Keating 1999).

In the second paper, Gifford and Acuto present an even more elaborate 3-D reconstruction of an Inka outpost in northwest Argentina and take the reader on a visual tour and walkthrough of the central building, exploring how the Inka framed the visual perception of, and therefore attempted to affect the sensory experience of, the landscape. By visually replicating a number of structures and behaviorally replicating the movement through one of them, they are able to reconstruct the consciously framed vistas, claustrophobic spaces, and "staggering landscape moments" which the Inka used to manipulate and control subject populations. This illustration of how the Inka framed the perception of the natural and cultural landscape, leads to their conclusion that the Inka did this to establish their own version of order and hierarchy. Though neither of these phenomenological studies fully meet the definition of experimental archaeology laid out at the 4

James

beginning of this introduction 9 , both do indicate the utility and power of using 3-D reconstructions and a phenomenological approach in order to replicate the perceptive behavior of past people. In doing this, they are able to generate and test hypotheses about the use and perception of structures and space.

the same peat bog environment and that the degree of preservation is related to conditions existing at the microenvironmental level, she is able to debunk some of the popular myths associated with bog body preservation and make a strong case for the need for much more experimentation of this sort, particularly ifwe are to identify the likely location of other bog bodies before they are destroyed.

Therefore, by realizing that phenomenological studies are also forms of behavioral replication, the whole arena of human perception can now be approached via experimentation. Eventually, such studies will lead to the generation and testing of more social, political, and even symbolic hypotheses. Though interpretations based on these hypotheses and experiments may always remain more dubious than those based on more materialistic ones (cf. Tringham 1978:184), useful insights will no doubt still be generated. Furthermore, our current social, political, and symbolic interpretations, particularly those generated using phenomenological analysis, will gain a firmer basis as the factors influencing them gradually become better understood through repeated experimentation.

Technological Process Replication

Another common form of process replication are those experiments focused on studying technological processes like flintknapping (e.g. Pekin 1997a, 1997b, 1997c, 1998) and metal production (e.g. Cleere 1971; Sauder 1999; Tylecote and Boydell 1978). An example of this sort ofreplication is Blair's (this volume) SMELT Project which explores early European iron production by replicating traditional shaft furnaces and then using them to smelt iron. By varying the characteristics of the furnaces used, particularly in terms of their shaft height, and by running multiple production cycles over the last decade, he has been able to generate and test hypotheses related to the process of iron smelting. These include observations on the relative productivity and necessary requirements oflow-, mid-, and tall-shaft furnaces. Based on his findings, he is able to suggest that such observations may one day help us understand why the Romans seem to have preferred the more easily managed medium-shaft furnaces, while the peoples beyond the Roman frontier opted for the more productive, yet difficult to operate, tall-shaft furnaces.

Process Replication

This type of replication involves the reproduction of past processes, including natural and cultural processes, from small-scale to large-scale (cf. Ingersoll and Macdonald's 1977:xiii contextual experimentation). Formation Process Replication

A very common type of process replication are those studies concerned with understanding the .formation processes that produce and transform the archaeological record over time (Ingersoll and Macdonald 1977:xiii; Schiffer 1972, 1976, 1983, 1987; e.g. Bell et al. 1996; Crabtree 1990; Evans and Limbrey 1974; Jewell 1963; Jewell and Dimbleby 1966; and see examples cited in Graham et al. 1972:5).

Simulation Studies

A third type of process replication consists of research traditionally referred to as simulation studies (e.g. Hamond 1978; Hodder 1978; Sabloff 1981 ). 10 Though these computer-aided studies do not experiment with physical phenomena like traditional experimentation does, their underlying assumptions and goals are completely in line with the definition of experimental archaeology outlined above. In other words, they replicate human behaviors and (more importantly) large scale processes in a controlled manner in order to generate and test hypotheses for use as analogies in archaeological interpretation. Therefore, they merit inclusion under the umbrella of experimental archaeology.

A particularly intriguing example is Gill-Robinson's (this volume) attempt to replicate the natural process of bog body preservation in order to understand the environmental variables which promote the preservation of mammalian soft tissue in peat bogs. Using twelve still born piglets as surrogate bodies deposited in different peat bogs, she attempts to replicate the natural process of decomposition and/or preservation to isolate the effects of different variables. By showing that the degree of preservation is highly variable in 9

Note that neither of these papers were originally conceived by their authors as examples of experimental archaeology, and therefore, should not be strictly assessed in those terms. Rather, their inclusion in the volume was at the behest of the editor, who felt that they clearly indicated the potential such phenomenological studies could have if approached in a more experimental manner. In other words, similar studies could be pursued in a more controlled manner, varying some of the parameters of the visual reconstructions, varying some of the characteristics of the people doing the walk-throughs, and running multiple "trials" of each walk-through in order to explore different perceptions of these buildings and landscapes. Instead of producing a single (or primary) interpretation of the past perception of architecture and space, the goal could be to identify the range of possible perception in the past and thereby generate a fuller understanding.

An excellent example of such research is Zubrow's (1990) simulation of European migration into upstate New York during the historic period using a geographic information system (GIS). By replicating a process of migration that used river valleys as the main conduits into the area, and by varying 10

5

Hamond (1978: I citing Schultz and Sullivan 1972:7) defines simulation as "the modelling of a process by a process". Hodder (1978:vi-vii, 134135) refers to simulation studies as "experimental work", though he prefers to consider simulation in archaeology as lying "somewhere between experimental archaeology and mathematical modelling" (Hodder 1978: 134; see also Hodder 1982:30, fig. 4).

the initial starting points for his simulated migrations, he was able to generate a number of different expected settlement patterns given the combination of initial variables. By comparing these simulated patterns with the actual historical information concerning where and when settlements were founded, he was able to challenge the traditional historical description of migration into upstate New York. Specifically, he showed that it was much more likely that Europeans migrated up both the Hudson and the Susquehanna rivers (i.e. via Pennsylvania), than simply up the Hudson as history seemed to emphasize (cf. Anderson and Gillam 2000). This research highlights how simulation studies can be used to test hypotheses.

social entity can be viewed as some sort of trial in a large scale, mostly uncontrolled "social experiment". This is more plausible if the behavior of the social entity can somehow be manipulated (Charlton 1981: 14 7 citing Tringham 1978: 171; Gould 1978: 8-9; Kramer 1979). With such a perspective on ethnographic observations, experimental archaeologists might be able to find "'experimental' data already available" (Coles 1979:39) without having to devise and run an experiment. However, if one considers that the purpose of experimental archaeology is to generate analogies to be used in archaeological interpretation, then it is probably more appropriate for the experimental archaeologists to just acknowledge that ethnographic and ethnoarchaeological observations are also good sources of analogies, rather than to try to cast them as (basically uncontrolled) examples of experimental research.

Another good example of a simulation study is Zubrow's (this volume) replication of population growth and continental migration in order to get a better understanding of the time frame required for the peopling of the world's inhabitable continents. By developing a computer simulation of these processes he is able to vary the initial parameters of the rate of population growth (low vs. high), the type of migration undertaken (wave/area vs. buffer/perimeter), and the role played by chance (determinant vs. robust chance). By running the eight different possible scenarios of the simulation for each of the six inhabited continents, he is able to show the effects of certain variables, particularly how low growth rates and robust chance dramatically increase the time needed for complete continental coverage. However, the most significant conclusion he derives is that regardless of the initial parameters used, the simulation indicates that the peopling of the world should have taken much less time than our archaeological evidence suggests it did. This leads him to suggest that we may be underestimating the role played by chance in this process or that there may be other variables involved that we have yet to identify. This research highlights how simulation studies can be used to generate new hypotheses for further testing.

The Overall Typology

This typology of experimental archaeological research has presented the four types of replicative experimentation as if they were clear cut and independent forms of research. Obviously, this is not really the case. In reality, the four types overlap in their use and are often integrated to a greater or lesser extent in the pursuit of a variety of research objectives. This can be clearly seen when a behavioral replication requires the use of materials produced during object replication. In an ideal experimental design, replication at the lower levels (e.g. object replication) will precede experimentation at the higher levels, creating a firmer foundation for higher level research (see below). Unfortunately, this does not always happen. In fact, most researchers embarking on a program of experimentation begin with fairly high level studies. This "putting the cart before the horse" mainly results from both a desire to understand "more interesting" higher level phenomena and an unawareness of the potentially confounding effects that lower level phenomena (variables) mayhave. 13

System Replication

This type of replication involves the reproduction of numerous processes in what is most easily thought of as a living system (cf. Ingersoll and Macdonald's 1977:xiii living archeology). Since this involves such a high level of replicative activity, it is usually only pursued through research which is normally referred to as ethnography or ethnoarchaeology .11

Finally, the last point to be made here is that this typology conceives experimental archaeology in a more fruitful way than others (cf. Ascher 1961a; Coles 1973, 1979; Reynolds 1999). In other words, by focusing on the sort of phenomena being replicated and by emphasizing the main purpose of experimentation (i.e. the production and evaluation of analogies for archaeological interpretation), new types of experimental archaeology can be developed ( e.g. phenomenological and simulation studies). Besides freeing the subfield from its traditional focus on technological and subsistence matters, this will also bring the strengths of a more controlled methodology to the analysis of economic, social, political, and even symbolic phenomena.

For example, some researchers consider ethnography or ethnographic observations to be a form of experimental archaeology (Amick et al. 1989:1-2; Coles 1977:237, 1979:39). 12 This is plausible, if one considers that a living 11 Ethnoarchaeology

"has most often connoted problem-oriented ethnographic research conducted by anthropologists trained as field archaeologists .... [it] usually involves the explicit integration of ethnographic and ethnohistoric data with archaeological data" (Kramer 1979:4). "The theoretical basis for ethnoarchaeology is the use of analogies derived from present observations to aid interpretation of past events and processes" (Watson 1979:277).

12

13

In contrast, some ethnoarchaeologists consider experimental archaeology to be a sub-discipline of ethnoarchaeology (e.g. Watson in Gould and Watson 1982:356).

6

"Isaac [ 1967:31 cited by Tringham 1978: 173] diplomatically attributes to enthusiasm and eagerness the archaeologist's tendency to leapfrog the tedious basic-level testing of things like depositional processes in favor of investigating more interesting upper-level problems such as population dynamics."

James

Experimentation, but not Experimental Archaeology

Control in Experimentation

One of the major practical goals in the performance of an experiment is the ability to control the variables that affect the phenomena under study (Spector 1981:15-16). The purpose of controlling variables in an experiment is to observe their effects (Amick et al. 1989:1; Odell and Cowan 1986:196) and thereby improve both the hypotheses generated and the testing of hypotheses. In general, increasing the control over the variables in an experiment improves the analogies that can be drawn from it. In other words, a controlled experiment produces the most reliable results (Ingersoll and Macdonald 1977:xii).

Having proposed a new typology, it should be noted that some research previously labeled experimental archaeology does not fit into this scheme and therefore should no longer be considered under the heading of experimental archaeology. The first of these are Ascher's (1961a:793) imaginative experiments. Based on his discussion, these seem to be the thoughts a researcher has about either what happened in the past or what might happen during a replicative experiment. For Ascher (1961a:793), they are used preparatory to field work to imagine how things will work. Though a case can be made that such thoughts can be very useful and may even be considered replicative in nature and intended to generate hypotheses which could be used as analogies to interpret the past, the lack of any sense of an experiment (i.e. a test or trial of some sort) taking place should preclude them from being considered a type of experimental archaeology.

Though most would agree that a highly controlled experiment is an ideal experiment and one that is easily repeatable (Spector 1981:21 ), less controlled and potentially unrepeatable experiments can also have significant value (see Mathieu and Meyer, this volume ). 14 Therefore, it is best to consider all experiments as existing along a continuum of control (Tringham 1978: l 70)(Figure 2). At one end of the continuum are those experiments which control very few variables. These are most often performed during the first generation of experimentation on a topic. 15 They are often the most clearly imitative of past human activity (e.g. researchers cutting down trees naturally using accurate replicas ofNeolithic axes), attempting to replicate past behaviors. Because of their lack of controls, and the problem of confounding variables (Spector 1981: 17), these experiments rarely clarify the cause and effect relationship between the independent variables and the dependent variables. They also tend not to be easily repeatable in terms of results. Because of this, and despite the initial goals set out by the researchers, these experiments tend not to be particularly useful for hypothesis testing. Instead they are best employed for the purpose of hypothesis generation and the identification of new variables, what Amick et al. (1989:6- 7) would consider as exploratory experimentation or what Malina (1983 :75) might call orientational experiments.

The second type which do not fit into this scheme are methodological experiments. These are experiments intended to either (a) test the usefulness of new methods or (b) test "the validity of methodological assumptions by applying them to known data or known results" (Ingersoll and Macdonald 1977 :xiii). Though both of these are performed within the context of an experiment and therefore can generate and test hypotheses (cf. Reynolds 1987:3-4), neither are replicative in any sense, nor are they intended to generate or test hypotheses that can be used as analogies in archaeological interpretation. Instead, they are merely experiments intended to test methods (or hypotheses about methods), not experiments to test interpretations (or hypotheses) about the past. As such, they should not be considered a type of experimental archaeology. Lastly, another type of "experiment" which does not fit this scheme are what one might call social interpretations - i.e. the inferences derived from experimental archaeological studies (Coles 1979:36-43; Ingersoll and Macdonald 1977:xiii). These include assessments of (a) the necessary labor input required to perform certain activities (e.g. treefelling and monumental construction) or (b) the likelihood of particular past behavior having been successful (e.g. ancient sea-faring contact). Though these interpretations (and new hypotheses) are useful and important, they should not be considered a type of experimentation, but rather one of its goals.

At the other end of the continuum are those experiments which employ many controls. These are more often performed during

Control and Confidence in Experimentation

Finally, before concluding this Introduction, a discussion of experimental archaeology is not complete without mentioning the issues of overall control and confidence in experimentation (see Spector 1981 : 11-19 for the basic concepts and terms used when discussing experimentation). 7

14

Though some researchers feel that experimental archaeology produces only particularistic knowledge useful only in time-space specifics (Charlton 1981: 145; Tringham 1978: 177), other researchers choose to emphasize the wider value of experimentation. They note that the process of experimentation makes one "think about all of the characteristics and qualities of his [sic] material, and not just those that are superficially evident" (Coles 1977:243). By concentrating one's "attention upon the material evidence ...the implementation of the experimental process can provide greater insight into practical implications of the archaeological evidence that have previously been unrecognised or ignored" (Reynolds 1977a:32-33, 1978: 144).

15

First generation experimentation often takes the form of what Schiffer ct al. (1994: 198) call archaeological experiments (in contrast to their experimental archaeology - sec next footnote). In other words, they "tend to be one-shot affairs, done in isolation and concluded before there has been an adequate opportunity to conduct the study rigorously. Such experiments often seem like pilot studies, perhaps reported prematurely" (Schiffer et al. 1994: 198). Most experiments performed by or for archaeologists have been of this type (Coles 1979:246)- though notable exceptions exist in the subfield of experimental lithic analysis (e.g. Pelcin 1997a,b,c, 1998).

Acknowledgments

the second or later generation of experimentation on a topic. 16 These experiments often appear to be less clearly imitative of past behaviors (e.g. researchers dropping steel balls on glass plates to replicate flintknapping). Despite this appearance, they are still replicative in nature. Because many, if not all, of the variables in these experiments are controlled, the cause and effect relationships between independent and dependent variables are more easily discernible. These experiments are also advantageous in that they are more likely to be repeatable. For these reasons, such experiments are especially useful in hypothesis testing (and hypothesis generation). Amick et al. (1989:6- 7) refer to such research as confirmatory experimentation, while Malina (1983:75) calls these corroborating experiments.

I would like to thank my colleagues Greg Borgstede, Michael Frachetti, Praveena Gullapalli, Andrew Pelcin, and John Walker.for their comments on previous drafts of this paper. I would also like to express my thanks to the various authors in this volume and the students in my Experimental Archaeology classes at the University of Pennsylvania (Spring 2000) and Washington and Lee University (Fall 2001) for helping me develop my ideas about experimental archaeology.

Biographical Sketch

James R. Mathieu received his B.A. (1992), M.A. (1992), and Ph.D (2001) in Anthropology from the University of Pennsylvania and his M.A. (1995) in Medieval Archaeology from the University of York, England. His field experience includes surveys and excavations in North Carolina, Maine, Virginia, Syria, France, England, Wales, Belize, and Tunisia. His research interests include the development of complex societies, the archaeology of Europe and the Circum Mediterranean area, the use of GIS, spatial analysis, experimental archaeology, and the role played by analytical scale in archaeological interpretations.

Confidence in Experimentation

Lastly, the reliability of, or confidence one places in, an inference generated or tested through experimentation can be increased in three ways. First, increasing the controls used in an experiment can hold more variables constant and result in the isolation of the effects of particular variables. Second, confidence can be increased by running the experiment in as contextually accurate a way as possible - e.g. choosing "experimental materials which were, or could have been, available in an aboriginal setting" (Ascher l 96la:8 l l; Coles 1967:2, 1973:15-16). Note, that these two techniques for improvement often are incompatible with each other! 17 Lastly, the repetition of the experiment in order to confirm that the results are not due to chance and that they can be reliably duplicated can also further increase the confidence in experimental results (Ascher 196la:811; Coles 1973:17, 1977:239, 1979:40, 43; Ingersoll and Macdonald 1977:xiv; Reynolds 1977b:308).

References Cited Amick, D.S, R.P. Mauldin and L.R. Binford 1989 'The potential of experiments in lithic technology' in D.S. Amick and R.P. Maudlin (eds) Experiments in Lithic Technology British Archaeological Reports, International Series 528 (Oxford): 1-14. Anderson, D.G. and J.C. Gillam 2000 'Paleoindian colonization of the Americas: implications from an examination of physiography, demography, and artifact distribution' American Antiquity 65 :4366.

However, despite each of these techniques it is important to realize that there may be "several possible solutions to an experimental problem" (Coles 1979:43). Furthermore, it must always be remembered that experimental archaeology does not provide answers. Rather, it merely eliminates possibilities, shows possible answers, and sometimes indicates the degree of probability of certain answers, but it can never "prove anything beyond a shadow of doubt" (Coles 1967:6; Coates et al. 1995:297; Coles 1973:15, 17, 168; 1979:243; Ingersoll and Macdonald 1977:xiv; Reynolds 1977a:32, 1987:5). 16

17

Ascher, R. 1961 a 'Experimental Anthropologist 63:793-816.

archaeology'

American

Ascher, R. 1961b 'Analogy in archaeological interpretation' Southwestern Journal of Anthropology 17:317-325. Bell, M., P.J. Fowlerand S.W. Hill son (eds) 1996 The Experimental Earthwork Project, 1960-1992 Council for British Archaeology (York). Beveridge, W.I.B. 1957 (3' tl ed. [1950]) The Art of Scientific Investigation W.W. Norton & Company Inc. (New York). Charlton, T.H. 1981 'Archaeology, ethnohistory, and ethnology: interpretive interfaces' in M.B. Schiffer (ed) Advances in Archaeological Method and Theory, Vol.4 Academic Press (New York):129-176.

Second generation and later experimentation is more closely associated with what Schiffer et al. (1994: 198) call a program-based experimental archaeology. Their conception of this sees "the findings of one experiment [as] nested within families of related principles (correlates) that, together, furnish a foundation for explaining technological variation and change" (Schiffer ct al. 1994:198). A number of researchers have noted the advantages that accrue by performing later generation , experimental archaeological research (Coles 1979:247; cf. Schiffer ct al. 1994: 197-198). In such long term programs or series tests (Coles 1977:239, 1979:41), "a conscious effort is made to lead from one result, repeated if required, to a second, or to a redirection of effort" (Coles 1979:41), particularly in order to eliminate alternative possibilities (Coles 1979:45).

Cleere, H. 1971 'lronmaking in a Roman furnace' Britannia 2:203217. Coates, J., S. McGrail, D. Brown, E. Gifford, G. Grainge, B. Greenhill, P. Marsden, B. Rankov, C. Tipping and E. Wright 1995 'Experimental boat and ship archaeology: principles and methods' International Journal of Nautical Archaeology 24:293-301. Coles, J. 1967 'Experimental archaeology' Proceedings of the Society of Antiquaries of Scotland 99: 1-20.

The former are often called controlled laboratory experiments while the latter are referred to asfield experiments (e.g. Skibo 1992b:29-30). Each have their strengths and weaknesses.

Coles, J. 1973 Archaeology by Experiment Hutchinson & Co. (London).

8

James

Coles, J. 1977 'Experimental archaeology- theory and principles' in S. McGrail (ed) Sources and Techniques in Boat Archaeology British Archaeological Reports, International Series 29 (Oxford):233-244.

Malina, J. 1983 'Archaeology and experiment' Archaeological Review 16(2):69-78.

Norwegian

Coles, J. 1979 Experimental Archaeology Academic Press (London).

Marstrander, S. 1976 'Building a hide boat: an archaeological experiment' International Journal of Nautical Archaeology and Underwater Exploration 5: 13-22.

Crabtree, K. 1990 'Experimental earthworks in the United Kingdom' in D.E. Robinson (ed) Experimentation and Reconstruction in Environmental Archaeology Oxbow (Oxford):225-235.

Mathieu, J.R. and D.A. Meyer 1997 'Comparing axe heads of stone, bronze, and steel: studies in experimental archaeology' Journal of Field Archaeology 24:333-351.

Evans, J.G. and S. Limbrey 1974 'The experimental earthwork on Morden Bog, Wareham, Dorset, England: 1963 to 1972' Proceedings of the Prehistoric Society 40: 170-202.

Odell, G.H. and F. Cowan 1986 'Experiments with spears and arrows on animal targets' Journal of Field Archaeology 13: 195-212.

Frison, G. 1989 'Experimental use of Clovis weaponry and tools on African elephants' American Antiquity 54:766-784.

Orme, B. 1973 'Archaeology and ethnography' in C. Renfrew (ed) The Explanation of Culture Change: Models in Prehistory Duckworth (London):481-492.

Forte, M. and A. Siliotti (eds) 1997 VirtualArchaeology: Re-creating Ancient Worlds Harry N. Abrams, Inc. (New York).

Orme, B. 1974 'Twentieth-century prehistorians and the idea of ethnographic parallels' Man 9:199-212.

Gould, R.A. 1978 'From Tasmania to Tucson: new directions in ethnoarchaeology' in R.A. Gould ( ed) Explorations in Ethnoarchaeology University of New Mexico Press (Albuquerque): 1-10.

Orme, B. 1981 Anthropology/or Archaeologists: An Introduction Cornell University Press (Ithaca).

Gould, R.A. 1980 Living Archaeology Cambridge University Press (Cambridge). Gould, R.A. and P.J. Watson 1982 'A dialogue on the meaning and use of analogy in ethnoarchaeological reasoning' Journal of Anthropological Archaeology I :355-381. Graham, J., R. Heizer and T. Hester 1972 A Bibliography of Replicative Experiments in Archaeology Archaeological Research Facility (Berkeley, CA). Hamond, F.W. 1978 'The contribution of simulation to the study of archaeological processes' in I. Hodder (ed) Simulation Studies in Archaeology Cambridge University Press (Cambridge):1-9. Hansen, H-O. 1969 'Experimental ploughing with a Dostrup ard replica' in A. Steensberg, A. Fenton and G. Lerche (eds) Tools & Tillage Vol. I (2) G.E.C. GAD Publishers (Copenhagen):67-92. Hodder, I. (ed) 1978 Simulation Studies in Archaeology Cambridge University Press (Cambridge). Hodder, I. 1982 The Present Past: An Introduction to Anthropology for Archaeologists Pica Press (New York). Ingersoll, D. and W.K. Macdonald 1977 'Introduction' in D. Ingersoll, J.E. Yellen and W. Macdonald (eds) Experimental Archeology Columbia University Press (New York):xi-xviii. Isaac, G.L. 1967 'Towards the interpretation of occupation debris: some experiments and observations' Kroeber Anthropological Society Papers 37:31-57. Jewell, P.A. ( ed) 1963 The Experimental Earthwork on Overton Down, Wiltshire 1960: An Account of the Construction of an Earthwork to Investigate by Experiment the Way in which Archaeological Structures are Denuded and Buried British Association for the Advancement of Science (London). Jewell, P.A. and G.W. Dimbleby 1966 'The experimental earthwork on Overtown Down, Wiltshire, England: the first four years' Proceedings of the Prehistoric Society 32:313-342. Johnstone, P. 1972 'Bronze age sea trial' Antiquity 46:269-274. Jones, P.N. and D.F. Renn 1982 'The military effectiveness of arrow loops: some experiments at White Castle' Chateau Gaillard 9/ 10:445-456. Kleindienst, M.R. and P.J. Watson 1956 "'Action archaeology": the archaeological inventory ofa living community' Anthropology Tomorrow 5:75-78. Kramer, C. 1979 'Introduction' in C. Kramer (ed) Ethnoarchaeology: Implications of Ethnography/or Archaeology Columbia University Press (New York): 1-20.

Pelcin, A. 1997a 'The effect of core surface morphology on flake attributes: evidence from a controlled experiment' Journal of Archaeological Science 24:749-756. Pelcin, A. 1997b 'The effect of indentor type on flake attributes: evidence from a controlled experiment' Journal of Archaeological Science 24:613-621. Pelcin, A. 1997c 'The formation of flakes: the role of platform thickness and exterior platform angle in the production of flake irritations and terminations' Journal of Archaeological Science 24: 1107-1113. Pelcin, A. 1998 'The threshold effect of platfonn width: a reply to Davis and Shea' Journal of Archaeological Science 25:615-620. Reynolds, P.J. 1977a 'Experimental archaeology and the Butser Ancient Fann Project' in J. Collis (ed) The Iron Age in Britain A Review University of Sheffield (Sheffield):32-40. Reynolds, P.J. 1977b 'Slash and bum experiment' Archaeological Journal 134:307-318. Reynolds, P.J. 1978 'Archaeology by experiment: a research tool for tomorrow' in T.C. Darvill, M. Parker Pearson, R.W. Smith and R.M. Thomas (eds) New Approaches to our Past: An Archaeological Forum University of Southampton (Southampton): 139-155. Reynolds, P.J. 1987 'Empiricism in archaeology' in P.J. Reynolds (ed) Buts er Ancient Farm Year Book 1986 Butser Ancient Farm (Hampshire):3-11. Reynolds, P. 1999 'Butser Ancient Farm, Hampshire, UK' in P.G. Stone and P.G. Plane! (eds) The Constructed Past: Experimental Archaeology, Education and the Public Routledge (New York):124-135. Sanders, D. 1990 'Behavioral conventions and archaeology: methods for the analysis of ancient architecture' in S. Kent (ed) Domestic Architecture and the Use of Space: An Interdisciplinary CrossCultural Study Cambridge University Press (Cambridge ):43-72. Saraydar, S. and I. Shimada 1971 'A quantitative comparison of efficiency between a stone axe and a steel axe' American Antiquity 36:216-217. Saraydar, S. and I. Shimada 1973 'Experimental archaeology: anew outlook' American Antiquity 38:344-350. Sauder, L. 1999 'The basics ofbloomery' The Anvil's Ring 27(2): 1519. Schiffer, M.B. 1972 'Archaeological context and systemic context' American Antiquity 37:156-165. Schiffer, M.B. 1976 Behavioral Archaeology Academic Press (New York).

Schiffer, M.B. 1983 'Toward the identification of formation processes' American Antiquity 48:675-706.

Tilley, C. 1994 A Phenomenology of Landscape: Places, Paths and Monuments Berg (Oxford).

Schiffer, M.B. 1987 Formation Processes of the Archaeological Record University of New Mexico (Albuquerque).

Tringham, R. 1978 'Experimentation, ethnoarchaeology, and the leapfrogs in archaeological methodology' in R.A. Gould (ed) Explorations in Ethnoarchaeology University of New Mexico Press (Albuquerque): 169-199.

Schiffer, M.B. and J.M. Skibo 1987 'Theory and experiment in the study of technological change' Current Anthropology 28:595622.

Tylecote, R.F. and P.J. Boydell 1978 'Experiments on copper smelting based on early furnaces found at Timna' in B. Rothenberg ( ed) Cha/eolithic Copper Smelting Institute for ArchaeoMetallurgical Studies Monograph No. I (London):27-49.

Schiffer, M.B., J.M. Skibo, T.C. Boelke, M.A. Neupert and M. Aronson 1994 'New perspectives on experimental archaeology: surface treatments and thermal response of the clay cooking pot' American Antiquity 59: 197-217.

Watson, A. and D. Keating 1999 'Architecture and sound: an acoustic analysis of megalithic monuments in prehistoric Britain' Antiquity 73:325-336.

Schultz, R.L. and E.M. Sullivan 1972 'Developments in simulation in social and administrative science' in H. Guetzkow, P. Katler and R.L. Schultz (eds) Simulation in Social and Administrative Science Prentice-Hall (Englewood Cliffs, NJ).

Watson, P.J. 1979 'The idea of ethnoarchaeology: notes and comments' in C. Kramer (ed) Ethnoarchaeology: Implications of Ethnography/or Archaeology Columbia University Press (New York):277-287.

Shimada, I. 1978 'Behavioral variability and organization in ancient constructions: an experimental approach' in R. Sidrys (ed) Papers on the Economy and Architecture of the Ancient Maya Institute of Archaeology, UCLA (Los Angeles):209-235.

Webster's New Collegiate Dictionary 1981 Webster:SNew Collegiate Dictionary G. & C. Merriam Company (Springfield, MA). Wylie, A. 1982 'An analogy by any other names is just as analogical: a commentary on the Gould-Watson dialogue' Journal of Anthropological Archaeology I :382-40 I.

Skibo, J.M. 1992a Pottery Function: A Use-Alteration Perspective Plenum (New York). Skibo, J.M. 1992b 'Ethnoarchaeology, experimental archaeology and inference building in ceramic research' Archaeologia Polona 30:27-38.

Wylie, A. 1985 'The reaction against analogy' in M.B. Schiffer (ed) Advances in Archaeological Method and Theory, Volume 8 Academic Press (New York):63-111.

Spector, P.E. 1981 Research Designs Sage (Beverly Hills).

Wylie, A. 1989 'The interpretive dilemma' in V. Pinsky and A. Wylie (eds) Critical Traditions in Contempora,y Archaeology: Essays in Philosophy, History, and Socio-politics of Archaeology Cambridge University Press (Cambridge): 18-27.

Stahl, A.B. 1993 'Concepts of time and approaches to analogical reasoning in historical perspective' American Antiquity 58:235260. Tarver, W.T.S. 1995 'The traction trebuchet: a reconstruction ofan early medieval siege engine' Technology and Culture 36:136167.

Zubrow, E.B.W. 1990 'Modelling and prediction with geographic information systems: a demographic example from prehistoric and historic New York' in K.M.S. Allen, S.W. Green and E.B.W. Zubrow (eds) Interpreting Space: GIS and Archaeology Taylor and Francis (London):307-318.

Thompson, R.H. 1956 'The subjective element in archaeological inference' Southwestern Journal of Anthropology 12:327-332.

1. Object Replication Visual Replicas Functional Replicas Using Real Artifacts or "FullReplicas" 2. Behavioral Replication Functional Replication Comparative Experiments Phenomenological Studies 3. Process Replication Formation Processes Technological Processes Simulation Studies 4. System Replication Ethnoarchaeology Ethnography Figure 1: A Typology of Experimental Archaeological Research iO

James

150 140 130 u, Cl) 120 .c 110 ns 100 'i: ns 90 80 II-70 0 60 Cl) .c 50 E 40 :::s 30 20 10 0

>

...

z

HIGH

LOW Relative Degree of Control Figure 2: The Relationship Between the Number of Variables in a Replicative Experiment and the Relative Degree of Control over the Experiment

ii

Monitoring Developments: Replicas and Reproducibility Genevieve LeMoine

Abstract

Methods of experimental microwear analysis are well understood: making and using tools,followed by microscopic examination to identify traces of use. One difficulty with such experiments is not the methods themselves, but the means of monitoring and recording the results. Using replicas made at regular intervals during experiments, it is possible both to improve our understanding of the development of wear and to create a permanent record of that development. Additional advantages include increased flexibility in both experiment design and analysis and increased standardisation. The development of a highly portable and permanent library of microwear types and the ability to sample collections rapidly for study later are also beneficial.

Introduction

Experimental archaeology is, at its best, characterized by attention to detail, control of variables, and careful recording of results. As a subset of experimental studies, microwear analyses are frequently exemplary in this regard. Although there is some debate in the literature about the degree of control that should be exerted over working conditions (see Keeley 1980; Tomenchuck 1985; and Vaughan 1985 for different approaches), a generally high standard of reporting the pertinent conditions makes it possible to evaluate and compare results. There is still room for improvement in microwear analysis, however. In particular, recording and describing wear patterns remains much more of an art than a science. As McGuire et al. (1982) have noted, describing microwear has always been a challenge. The best way for an analyst to truly understand a particular type of wear is to observe it first hand. Most often this means either creating the wear oneself in a controlled experiment, or, if one has the luxury of working with others studying microwear, examining their samples. In the absence of direct observation, written descriptions alone do not make it possible to identify a specific type of wear, debates and agreements about standardized vocabularies notwithstanding (Hayden 1979). In most situations, analysts must either rely on photomicrographs of wear patterns in order to evaluate the patterns developed by or identified by other researchers, or have direct access to the actual tools. Similarly, microwear analysts usually have to rely on photomicrographs to record wear patterns for their own reference. However, though photomicrographs are indispensable as a means of recording features and communicating about micro wear through publication, they are not necessarily the best way to record microwear or to exchange information about wear among researchers. To begin with, their utility is highly dependent on the quality of the camera and microscope used to generate them. Similarly, as anyone who has worked with more than one type of microscope can

attest, the same pattern can look considerably different under different microscopes. This is most evident when comparing the results from Scanning Electron (SEM) and light microscopes (see Mansur-Franchomme 1986), but it is observable even between different brands of the same type of microscope. In other words, a photomicrograph taken on one system may not be useful for analysts working with other types of microscopes. In this paper I shall discuss another means of recording and communicating patterns of microwear. Specifically, I will discuss the advantages of using highly detailed surface replicas to record and analyze microwear on both archaeological and experimental tools. I will place this discussion in the context of my own research on bone, antler and ivory tools, and my interest in monitoring the development of micro wear, particularly the impact of postdepositional factors on microscopic wear patterns. My goal is show that replicas offer a number of advantages, from flexibility in analysis to ease of communication between analysts.

A Brief History of Replicas and Microwear: 20 Years of Research

Microwear analysis has traditionally been based on direct observation of both tool edges and faces. This is true of both the high (Keeley 1980; Vaughan 1985) and low magnification schools (Odell 1975; Tringham et al. 1972), both of which relied on incident-light microscopes. With the advent of Scanning Electron Microscope (SEM) analysis, replicas came into the picture as a necessity when specimens were either too large for the microscope chamber, or too important (archaeologically speaking) to be physically altered through coating with conductive material. Despite these drawbacks, however, many researchers continue to rely on direct observation (Fedje 1979), which continues to be the standard practice today. Intuitively, this makes a lot of sense - what could be better than the real thing? But in reality, 'the real thing' can be difficult (and sometimes impossible) to work with for a number of reasons. As each of the researchers who

Experimental Archaeology Replicating Past Objects, Behaviors. and

have published on this topic has discovered, using replicas can help overcome these difficulties.

patterns and to investigate the effects of post-depositional factors on microwear (see below). 1 The second was flexibility, both in the design of experiments, and the physical location of analysis. By using replicas I was able both to carry out experiments away from my lab, and to study museum collections more effectively. Finally, a third reason for relying on replicas was the possibility of standardizing observations between experimental and archaeological tools, and between different microscope types.

There is a long history of publication on the use of replicas for microwear analysis. Of these, one of the first (the first that I have been able to identify in English) was presented by J0rgen Ilkjaer in Hayden's seminal (1979) volumeLithic UseWear Analysis. Since that time at least eight other authors have directly addressed the use of replicas (Beyries 1978; Bienenfield 1995; d'Errico et al. 1982, 1984; Hillson 1992; Knutsson and Hope 1984; Plisson 1983; Rose 1983; Runnings et al. 1984), variously recommending the use of either silicone rubber, dental impression compounds, and/or acetate tape or varnish as a primary molding material. Some of these researchers work with these negative molds or impressions, while others go one step further and create positive replicas in epoxy resin or metal.

Research Goals and Problems

My original research goal was to do a high magnification analysis of micro wear on bone tools, building on Semenov's (1964) seminal work, but using SE Microscopy rather than a light microscope (LeMoine 1985, 1987). In particular, I wanted to monitor the development of wear over a period of time on experimental tools modeled on archaeologically recovered deer and bison bone tools of the Great Plains. This would involve evaluating how long it took wear patterns to become established and stabilized and, through exposing used tools to the elements, how well such patterns might be expected to be preserved.

Despite this slow but steady stream of publications over the last 20 years, the idea of using replicas ( except when absolutely necessary for the use of SEMs) has not caught on. In fact, the tone of many of these papers is that of introducing a new idea. For example, as recently as 1992 Hillson seems to have rediscovered the whole process, citing extensively from the microscopy literature (where the use of replicas in a variety of contexts is well established), while entirely neglecting the published archaeological and physical anthropological uses of this technique. Similarly, Bienenfield (1995) recommends the use of silicon rubber molding compounds (and epoxy resin casts) as discussed by Ilkjaer (1979), unaware perhaps of the more recently developed and superior silicone based dental impression materials popularized among physical anthropologists by Rose (1983). Likewise, in France, a variety of publications have recommended the use of latex rubber and acetate varnishes (Beyries 1978; d'Errico et al. 1984; Plisson 1983, 1984), but these publications appear to have been largely, and undeservedly, neglected by the English-speaking world. Overall, only one publication, Jennie Rose's (1983) article in The American Journal of Physical Anthropology, has really caught on and been cited frequently, primarily but not surprisingly by physical anthropologists rather than by archaeologists.

The major problem I was confronted with was how to observe wear patterns on the bone tools. The vast majority of the bone tools I needed to study, experimental or archaeological, were too large to fit into the small chamber of the SEM. Equally problematic, the alternative strategy of cutting the bone tools into pieces small enough to fit in the microscope chamber (i.e. an inch or less) was unacceptable since it would in effect destroy each experimental tool after a single use and irreparably damage archaeological samples. Since direct observation of the bone tools was impossible, the only feasible option remaining was the use of replicas. Settling on a Replica Strategy

Once it was clear that replicas would be necessary, the next step was to decide which sort to use. At that time (1983), silicone rubber seemed to be the obvious choice, and so I began using Dow Coming RTV Silastic, a silicone rubber molding compound developed for industrial mold-making, but also commonly used in the production of molds for preparing casts of archaeological objects. Compounds of this sort have been documented to produce very high resolution molds, faithfully reproducing microscopic details. To use them, one must mix the basic compound with a catalyst. Since the action of stirring in the catalyst produces air bubbles, which will be visible as artifacts on the mold, it is necessary to expose the stirred mixture to a vacuum to remove the bubbles. Objects to be molded must be prepared carefully to ensure that the compound stays where it is intended. In my case, since I was preparing molds only of the working parts of tools, this often involved somewhat clumsy arrangements of wax dams and/or containers (in which the end of a tool could be immersed). These dams

For the most part, each of these researchers turned to using replicas to facilitate examination in SEMs. Although many of them have also pointed out other advantages associated with using replicas (e.g. Ilkjaer 1979; Knutsson and Hope 1984), their use continues to be primarily associated with the special requirements of SEMs.

Repeating History: Trials and Errors

Like most microwear analysts, I initially turned to making replicas for observation in a SEM. However, a number of other concerns, related to the research questions I was pursuing, rapidly led me to rely on replicas for a variety of other purposes. The first and most significant of these was the necessity of maintaining a record of previous wear. This is essential in order to monitor the development of wear

1

i4

Though it is possible to do this with photornicrographs, a photograph can only record a tiny (microscopic) area of wear. In contrast, a replica can preserve a record of the whole working edge, if so desired.

and containers then needed to be maintained as the compound cured over the course of about 24 hours. Despite these awkward contrivances, the molds themselves were successful, producing high resolution replicas of experimental tools that could easily be cut to size, coated with conductive materials (gold) and observed in a SEM. 2 Unfortunately,the results were not so good for the archaeological tools. While the quality of the molds remained high, the impact of the molding compound on dry archaeological bone was disastrous. After being exposed to the silicon rubber molding compound for 24 hours, the archaeological tools became stained, most likely by an inert silicone oil.3 This stain proved difficult or impossible to remove, and I was forced to cut short this stage of my research. As I later learned (Bruno Pulliot 1988,pers. comm.), such stains are more than simply a cosmetic problem; the inert oil is thought to inhibit the uptake of moisture as the object experiences changes in relative humidity. This can cause the unstained portions of the object to expand or contract at different rates, resulting in cracking and exfoliation. This difficulty proved ultimately to be insurmountable. As I will discuss in more detail below, silicone based molding compounds of any sort are incompatible with archaeological bone, antler or ivory. Although this would seem to have been a major set-back, the advantages of working with replicas seemed to out-weigh the disadvantages and so I set out to find a way to make replicas without damaging the archaeological materials. I began experimenting with other replication techniques reported in the literature (see citations above). Like many others, I turned to the dental impression materials recommended by Rose (1983). These too are silicone based materials, but I was hopeful that problems of staining would be minimized because these compounds set in a very short time (3-5 minutes), significantly reducing the time objects must be in contact with them. They are also much easier to work with, requiring neither vacuum de-airing or a system of dams to hold the compound in place while it sets. Unfortunately, my optimism was not warranted. Field testing on un-worked archaeological bone revealed that staining continued to be a problem (see experiment described in more detail below). I next turned to a relatively old replicating material, acetate tape and similar thin films, which were first used to make surface replicas of opaque objects for observation in Transmission Electron Microscopes, before SEMs had been developed (Pameijer 1974). Although that use is now obsolete, researchers in a variety of disciplines continue to use it. Its use, and the use of related varnishes, has been discussed in the archaeological literature by Knutsson and Hope (1984), Plisson (1983, 1984) and Runnings et al. (1984). Acetate tape has a number of advantages over silicone-based products. First, from my perspective, its most obvious benefit

is that it leaves few if any traces behind, making it well suited for use on archaeological bone, antler and ivory. Should any acetate residue be left behind, it is easily removed with acetone. Second, of all the materials discussed here, acetate tape is the quickest to use. This is important not so much for speedy production of replicas (although that can be advantageous in some settings), but because it means the tape is in contact with the object being replicated for only a brief time (i.e. less than a minute), reducing the possibility of any sort of damage to the object. Third, a less evident advantage is that, since it is transparent, acetate tape can be observed with a variety of instruments, including a transmitted light microscope. Finally, acetate tape is much less expensive than other products, making it readily accessible to projects of all budgets, and making it possible to make multiple replicas of a single tool. The technique I used for making replicas with acetate tape are fully described elsewhere (LeMoine 1997). Briefly, it is a matter of using a small, soft paint brush to wet the area to be replicated with acetone, and then applying a piece of the tape. In less than a minute, the acetone will first soften the tape and then evaporate, drawing the tape onto the surface as it does. Once the tape has set it can be gently pulled from the surface. It takes some experience to judge the amount of acetone to use, and the amount will vary depending on the porosity of the material. But once the technique is mastered, it is quick and easy to do and multiple replicas ('peels') can be made of the same area in a matter of minutes. Of course, acetate tape is not without disadvantages. First, because the tape is very thin, it is not suitable for making replicas of areas of high relief, nor of points or holes. For these a three-dimensional material is preferable. Second, because the tape can be attached quite firmly to the surface, it is also unsuitable for making replicas of fragile, and especially exfoliating, material. Of course, these are unlikely candidates for microwear analysis in any case. Through trial and error, then, I found that acetate tape met most of the requirements of my research. It is simple to use, it can be observed with a number of different types of microscopes, and most importantly, it has practically no impact on archaeological materials. Since it cannot be used in all situations, however, there are circumstances when using a silicone based molding compound is necessary. In the experiments described below, each of these different replicating materials was used. Whichever material was used for a particular experiment, using replicas made it possible to maintain a permanent record of experimental and archaeological tools. It also provided some analytical flexibility and increased my ability to standardize observations between experiments and between experimental and archaeological samples.

2

I used negative molds rather than casts throughout this work.

Case Study: Monitoring Wear

3

It is not clear how serious this is for other archaeological materials, in particular stone. Some lithic materials, such as obsidian, are probably impervious, but more porous stones may not be.

As was mentioned above, one of the main reasons I chose to use replicas was in order to monitor the development of wear

patterns, and to evaluate the impact of post-depositional factors such as transport and weathering on microwear. I have discussed the development of wear patterns elsewhere (LeMoine 1993, 1997). Here I will use my studies ofpostuse factors as a case study in the advantages of using replicas.

the experiments prior to and after exposure, using Dow-Corning Silastic. The results were examined in a SEM, revealing the expected microscopic cracking, and in some cases, general roughening of the surface (Figure la, lb, and le). Because the experiment was of relatively short duration, macroscopic effects of weathering did not have time to develop. Nevertheless, as later study of archaeological samples was to show, microcracking is a significant problem, often obscuring wear even though there is not macroscopic evidence of weathering or other forms of damage to the bone.

Because my research focused on bone, rather than the more typically studied stone, I was concerned that non-use factors would have a significant impact on the preservation of microwear. For example, studies of stone tools have shown that both burial and transport can cause the development of characteristic microscopic patterns that can potentially eradicate or be confused with traces of use (e.g. Levi Sala 1986). Given this, it was only reasonable to suppose that bone tools could also be subject to these effects. Similarly, taphonomic studies have documented the severe impact on bone of surface exposure and deposition (e.g. Behrensmeyer 1978; Behrensmeyer et al. 1986; Brain 1981). In particular, bone tools may undergo cracking and exfoliation, depending on how long they have been curated or exposed on the surface before deposition. Finally, other taphonomic factors, such as chewing by carnivores or rodents, root etching, and even transport from the site to the lab also need to be considered as possible sources of wear which could eradicate or be confused with use wear.

A second set of experiments was designed to study the impact of transport during the life of the tool. A small series of worked and used experimental tools was carried around in pockets and bags over a period of several months. Replicas were made before and at intervals during this period (Figure 2a and 2b). As with the weathering experiment, although I anticipated that the effects of transport would be most evident on edges, points, crests or other high points, it was not possible to predict precisely where such patterns might develop. Thus it was necessary to record as much of the original surface as possible. The results of the experiment bore this out. As was expected, polishes, reminiscent of the 'bright spots' described on archaeological stone tools, and randomly oriented striations developed on various parts of the pieces (see LeMoine 1997:39-40 for a detailed discussion). Identifying which marks were experimentally produced and which were present on the original, and documenting the rate at which such marks appeared, would have been impossible without having a replica of the original surface to refer to.

The Experiments Over the past 15 years then, I have designed a number of experiments to monitor some of these effects. Each was designed to investigate a different type of non-use wear; e.g. weathering, pre-depositional transport, and post-excavation transport and processing. Only the first two will be discussed here. The third set of experiments are the subject of the next section.

Both of these experiments provided information useful in subsequent analyses of archaeological bone tools, where both micro-cracking due to weathering and polished characteristics of transport were frequently observed (LeMoine 1997). Neither of them could have been as successfully documented without the use of replicas of most or all of the original surfaces of the experimental tools.

To monitor the development of any sort of wear, it is necessary to observe the same section of a tool at different stages. In most microwear experiments, the location of wear patterns is readily predicted because the actions that produce it are known and controlled. This is not the case for the experiments being considered here, however, because the forces creating the wear (weathering, generalized rubbing, etc.) act not only on the edges and points used in focused tasks, but on the whole surface. Since the location of patterns that may be of interest is unpredictable it is important to be able to compare the original and modified surfaces in some detail. Replicas make it possible to identify areas of interest first on the modified specimen (weathered or otherwise) and then compare these to the unmodified (original) surface.

Case Study: Flexibility Perhaps one of the most often overlooked advantages of using replicas is the analytical flexibility they provide. The current standard for use-wear analysis is direct observation of tools with an incident light microscope (high or low power), with some researchers using SEMs. This means bringing the objects to the microscope, or bringing the microscope to the objects. The latter option is possible in the case of a light microscope, but not in the case of a SEM. For those working in remote locations where electricity is not always readily available, even bringing a light microscope is not always feasible. Making replicas, as I happily discovered, removes this necessity. In my case, it allowed me to monitor the microscopic changes that occur while bone from saturated, permanently frozen contexts dried under a variety of conditions.

The first and most preliminary set of experiments was designed to control for the effects of surface exposure prior to deposition. Both used and unused experimental bone tools were exposed to the elements over the course of a prairie summer and fall in southern Alberta, Canada. A group of seven experimental tools (including one unmodified bone later made into an awl) were enclosed in a hardware cloth box. Since I was primarily interested in the effects of simple weathering on the microscopic traces of wear, the box was intended to prevent animals from gnawing on them. Replicas were made of all the pieces used in

The Experiments This experiment, described in detail in LeMoine (1997), involved monitoring the impact of post-excavation treatment i6

on bone from frozen, saturated sites. Fragments of unused bone (from Gupuk (NiTs-1 ), a permanently frozen site in the Mackenzie Delta region of the Northwest Territories, Canada, which was the subject of my microwear analysis) were replicated immediately after excavation (Figure 3). 4 After the initial replications, the bones were then divided into groups and subjected to different post-excavation treatments. These ranged from careful wrapping in damp moss and plastic wrap (as was done for all bone tools from the site to prevent drying), to rapid air-drying and bagging together with stone and other debris (to represent the worst possible transport conditions worse than any actually practiced during this excavation). After transport to my home lab (and in some cases, further drying, following the procedures for drying saturated bone tools from the site), the bones were observed with a lowpower microscope to identify possible areas of interest. New replicas were then made, again using Reprosil. Both sets of replicas were then subjected to microwear analysis. The results, (described in detail in LeMoine 1997) provided data on both the impact of different transport methods, and on different approaches to drying saturated bone tools. As was expected, the care with which the bones were handled did have an impact on their microscopic appearance. The utility of careful wrapping and slow drying was demonstrated (Figure 4a and 4b ). For my purposes here, however, the significant point is that this experiment could not have been performed using traditional direct observation of the bones. In other words, by using replicas it was possible to record the state of the bones in the field before any post-excavation factors could have an impact. Equally important, the use of replicas made the recording ofrelatively large parts of the bones possible. This was important, since unlike most microwear experiments, it was not possible to predict where significant changes might take place (even had it been possible to observe the tools with a microscope in the field). Without the use of replicas, it would not have been possible to adequately record, either verbally or photographically, all the possible sites of interest. As is clearly demonstrated in the above example, flexibility in designing experiments is a major advantage of using replicas. In particular, the possibility of studying museum collections without having to transport them to a distant research location is very useful. As Knutsson and Hope (1984) pointed out, the use of replicas allows one to examine archaeological collections without the necessity of transporting them to the lab. Collections can be identified, documented and replicated on-site (either in a museum or in the field), leaving the microscopic analysis to be done in a home lab, with ready access to a reference library of wear 4

The bones replicated were selected on the basis of their form, those with prominent edges and points similar to those on actual tools were preferred. In making the replicas, particular attention was paid to points and edges, as these were considered to be the most vulnerable to damage. Since the bones all came from frozen deposits and were saturated when they were excavated, a dental impression material (Reprosil) was used to replicate them. Intended for use in mouths, the accuracy of dental impression replicas is not affected by the presence of water. The fact that the bones would be stained by the Reprosil was not considered to be a problem since none of the bones were artifactual in nature.

patterns. In my own research I was able to examine collections at a number of institutions, select those appropriate to my study, draw or photograph them, and make replicas of areas suitable for analysis. For the most part, I made two replicas of each area to be examined, one for analysis and one for a back-up should the first be faulty or damaged in any way. This also has allowed me to experiment with new and improved mounting techniques. 5 Another form of research flexibility provided by the use of replicas involves the possibility of comparative observation. In other words, there may exist research questions which would benefit from the examination of the same object using both light microscopy and a SEM. It would be possible, for example to take advantage of the distinct advantages of different microscope systems to study polish, seen better with a light microscope, and striations, best observed with in a SEM. With the use of replicas, any sample that can be examined in an SEM can also be examined with a light microscope (whether incident light (for opaque replicas) or transmitted light (for both opaque and transparent replicas)), thereby making replicas ideal for such comparative observations.

Case Study: Standardization Analytical flexibility is a significant advantage, but it is equally important to consider the issue of standardization. Standardization has always played a significant role in microwear studies. Microwear analysts are typically very careful to document the conditions under which experimental wear patterns were created, the methods used to prepare tools for observations, and the microscope systems used for analysis. But not all aspects of micro wear studies have been subject to this amount of control and standardization. This is most obvious when one considers that it is not possible to eliminate variation in the materials of the tools to be analyzed. Variations in color, reflectivity and translucency of some types of stone or bone, whether they be experimental or archaeological, have long plagued microwear studies (see Petraglia et al. 1996; e.g. Grace 1989). For example, analysts working with polishes can have a difficult time identifying them on highly reflective materials. Similarly, the crystalline structures of some stones can interfere with the identification of wear, while slightly translucent experimental (fresh) bone can make edges difficult to decipher (particularly under an incident light microscope). One way researchers try to control for such variations is by performing experiments which use materials as similar as possible to those they intend to analyze. However, this is not always possible, and even when it is, there may be difficulties in making the leap from experimental to archaeological wear patterns. This results either because of the differences between experimental and archaeological materials, or because of the inherent difficulties of observing the material itself. 5

The equipment required to make acetate tape replicas is compact and readily obtainable. Only a small work space is required and no special equipment is needed.

The point to be made here is that the use of replicas can overcome some of these difficulties. Their advantage lies in the fact that a replica reproduces the surface topography only, eliminating variation in color, reflectivity or any other variable, and thereby resulting in standardization of observation. As Knutsson and Hope (1984) demonstrated so long ago, replicas reproduce polishes and other microscopic features, but without all the other distracting features of the originals.

potentially damaging to the objects. In contrast, acetate tape does not appear to have this disadvantage, although it can lift surface deposits, or even fragile surfaces (which are not likely to carry preserved wear traces in any case). These cautions aside, the advantages of using replicas far outweigh the disadvantages. Far from being simply a way around specific problems, they provide a means of increasing the analytical rigor of microwear analysis, and the flexibility of such studies.

A further advantage of using replicas is that they can help standardize other aspects of analysis. In particular, both the intensity of the light, and the angle at which it hits a specimen can dramatically affect the appearance of many features. Observations on actual objects require either that they be hand held (suitable only for observations with low power microscopes) or supported by some means which must be customized for each object (wax and plasticine are two common techniques). Neither of these mounting techniques is particularly conducive to a standardized viewing angle - a necessary precondition to insure that two observers are looking at the same thing. By using replicas, it becomes much easier to standardize observation conditions not only between observers, but from tool to tool. 6 As quantification of microwear becomes more common (particularly through digital analysis), this sort of standardization will become increasingly important.

Discussion The use of replicas to study microscopic microwear on archaeological tools has a long, but checkered history. Numerous researchers have published on the benefits and advantages of using replicas made from various materials (Beyries 1978; Bienenfield 1995; d'Errico et al. 1982; d'Errico et al. 1984; Hillson 1992; Knutsson and Hope 1984; Plisson 1983; Rose 1983; Runnings et al. 1984), but as this repetition suggests, archaeologists have remained unconvinced. Every few years a new voice points out the advantages of using replicas in one context or another, but it seems to fall on deaf ears. ls that to be the fate of this work as well? The use of replicas has long been stigmatized by the idea that with every copy some detail, some information, is lost. While this cannot be denied, the quality of replica materials, combined with the relatively low magnifications that appear to be most useful for wear analysis (up to about 400X) mean that most analysts are working well below the capability of such material to reproduce detail. Some researchers have demonstrated this to their own satisfaction by the simple expedient of making a replica of some object that can be observed directly and then comparing the two (e.g. Runnings et al. 1984: fig. 5). One can just as easily check the comparability of replicas made with different materials, or observed with different microscopes (see LeMoine 1997: figs. 3.la and 3.lb).

Some Words of Caution The combined benefits of analytical flexibility and observational standardization are compelling reasons for using surface replicas for microwear analysis, but a few words of caution are in order, for no technique is perfect for all situations. First, and perhaps most obviously, replicas should not be made of any object that may be subjected to residue analysis. This is a worthwhile caveat for any microwear analysis, since the cleaning required is also contra-indicated for residue studies, but it is particularly important when making replicas. Any of the materials described here will adhere firmly enough to the surface of the object that they may lift residues. In the case of silicon based materials, they may also contaminate archaeological residues with their own residue.

I have tried to demonstrate, with illustrations drawn from my own research, the advantages of working with replicas. These range from increased analytical flexibility to improved standardization, and include the ability to monitor and record wear patterns much more comprehensively than is possible with photomicrographs alone. Since replicas can themselves be replicated, either by making casts from the original mold, or by making multiple copies from the original object, they also dramatically improve our ability to communicate information about wear patterns between analysts. Difficult or interesting examples can be replicated and sent to distant colleagues for consultation. Students can observe the development of wear patterns on experimental tools without having to endlessly recreate the same experiments. Archaeological collections can be studied, and restudied without risk to the objects themselves, and permanent libraries of wear types can be established. All of these benefits can be achieved at a small cost, using well established techniques and materials. It remains only for analysts to take advantage of this simple and useful technique.

Second, replication materials and techniques should be selected with caution. The three most commonly recommended types of material are silicone rubber moldmaking compounds, dental impression materials (also silicone based), and acetate tape or varnish. I have found the first two of these to be impossible to work with on archaeological bone, although they can probably be used safely with stone. In particular, they leave stains that are both unsightly and 6

This is true of any replica, but it is particularly true of transparent, acetate replicas. To be observed, these must be mounted on a flat, rigid surface (such as a standard microscope slide). Once mounted they can be observed in either reflected light (in which case it is still possible to adjust the angle at which the light hits the surface) or in transmitted light (in which the relationship between the slide and the direction of light is fixed).

i8

Acknowledgments

Knutsson, K. and R. Hope 1984 'The application of acetate peels in lithic usewear' Archaeometry 26:49-61.

The research described here was supported by graduate fellowships from the Social Sciences and Humanities Research Council of Canada, the Alberta Heritage Trust and the Killam Foundation. I would like to thank Nicholas David, Scott MacEachern, and James Mathieu.for providing much needed support and comments on various aspects of this work. All errors and omissions are, of course, mine.

LeMoine, G.M. 1985 'Experimental Use Wear Analysis of Bone Tools' Unpublished M.A. Thesis, Department of Archaeology, University of Calgary (Calgary).

Biographical Sketch

Genevieve LeMoine received her Ph.D. in 1991 from the Department of Archaeology, University of Calgary. She is currently curator/registrar at The Peary-MacMillan Arctic Museum, Bowdoin College, Brunswick, Maine. Her research interests include skeletal technology, Arctic prehistory and ethnohistory.

LeMoine, G. M. 1987 'Experimental use wear analysis ofbone tools' Archaeozoologia IIl/1,2:211-224. LeMoine, G.M. 1993 'Applications oftribology to a study of use wear on bone tools' American Antiquity 59:316-334. LeMoine, G.M. 1997 Use WearAnalysis on Bone and Antler Tools of the Mackenzie Inuit. British Archaeological Reports, International Series S679 (Oxford). Levi Sala, Irene 1986 'Use wear and post-depositional surface modification: a word of caution' Journal of Archaeological Science 13:229-244. Mansur-Franchomme, Marie Estella 1986 'Microscopie du materiel lithique prehistorique: traces d'utilisation, alterations naturelles, accidentelles et technologiques, examples de Patagonie' Cahiers de Quaternaire 9, Editions du CNRS (Paris). McGuire, Randal H., John Whittaker, Michael McCarthy and Rebecca Mc Swain 1982 'A consideration of observational error in lithic use wear analysis' Lithic Technology 11(3):59-63.

References Cited

Odell, G. 1975 'Micro-wear in perspective: a sympathetic response to Lawrence H. Keeley' World Archaeology 7:226-240.

Behrensmeyer, A.K. 1978 'Taphonomic and ecologic infonnation from bone weathering' Paleobiology 4:150-162.

Pameijer, Comelis H. 1974 'Replica techniques' in M.A. Hyatt (ed) Principles and Techniques of Scanning Electron Microscopy, Biological Applications Van Nostrand Reinhold (New York):4593.

Behrensmeyer, AK., K.D. Gordon and G.T. Yanagi 1986 'Trampling as a cause of bone surface damage and pseudo-cutmarks' Nature 319:768-771. Beyries, S. 1978 'Etude de traces d'utilisation sur des empreintes en latex' Bulletin de la Societe Prehistorique Francaise 78: 198199.

Petraglia, M., Dennis Knepper, Petar Glumac, Margaret Newman, Carole Sussman 1996 'Immunological and microwear analysis of chipped-stone artifacts from Piedmont contexts' American Antiquity 61:127-135.

Bienenfield, P. 1995 'Duplicating archaeological microwear polishes with epoxy casts' Lithic Technology 20(1):29-39.

Plisson, H. 1983 'An application of casting techniques for observing and recording microwear' Lithic Technology 12:17-21.

Brain, C.K. 1981 The Hunters or the Hunted? An Introduction to African Cave Taphonomy University of Chicago Press (Chicago).

Plisson, H. 1984 'Prise d'empreinte des surfaces osseuses: note complementaire' Bulletin de la Societe Prehistorique Francaise 81: 267-268.

d'Errico, F., G. Giacobini and P.-F. Puech 1982 'Varnish replicas: a new method for the study of worked bone surfaces' In OSSA, 911:29-51. d'Errico, F., G. Giacobini and P.-F. Puech 1984 'Les repliques en vernis des surfaces osseuses faconees: etudes experimentales' Bulletin de la Societe Prehistorique Francaise 81(6): 169-170. Fedje, D. 1979 'Scanning electron microscopy analysis ofuse-striae' in B. Hayden (ed) Lithic Use-WearAnalysis Academic Press (New York): 179-187. Grace, Roger 1989 Interpreting the Function of Stone Tools: The Quantification and Computerisation of Microwear Analysis British Archaeological Reports, International Series 474 (Oxford). Hayden, Brian ( ed) 1979 Lithic Use-Wear Analysis Academic Press (New York). Hillson, S. W. 1992 'Impression and replica methods for studying hypoplasia and perikymata on human tooth crown surfaces from archaeological sites' International Journal of Osteoarchaeology 2:65-78.

Rose, J. 1983 'A replication technique for Scanning Electron Microscopy: applications for anthropologists' American Journal of Physical Anthropology 62:255-261. Runnings, A., C. E. Gustafson and D. Bentley 1984 'Use-wear on bone tools: a technique for study under the Scanning Electron Microscope' in R. Bonnichsen and M.H. Sorg (eds) Bone Modification Center for the Study of the First Americans, Institute of Quaternary Studies, University of Maine (Orono, ME):259266. Semenov S.A. 1964 Prehistoric TechnologyBarnes and Noble (New York). Tomenchuck, John 1985 'The Development ofa Wholly Parametric Use-wear Methodology and its Application to Two Selected Samples of Epipaleolithic Chipped Stone Tools from Hayonim Cave, Israel' Unpublished Ph.D. Dissertation, Department of Anthropology, University of Toronto (Toronto).

Ilkjaer, J. 1979 'A new method for observation and recording of use-wear' in B. Hayden ( ed) Lithic Use-Wear Analysis Academic Press (New York):345-349.

Tringham, R., G. Cooper, G. Odell, B. Voytek and A. Whitman 1972 'Experimentation in the formation of edge damage: a new approach to lithic analysis' Journal of Field Archaeology l: 171196.

Keeley, L. 1980 Experimental Determination of Stone Tool Use University of Chicago Press (Chicago).

Vaughan, P.C. 1985 Use Wear Analysis of Flaked Stone Tools University of Arizona Press (Tucson).

Figure 1a: The surface of a broken deer bone before weathering.

Figure 1b: The same surface after weathering - the topography is both rounded and roughened.

20

Figure 1c: The surface of a deer scapula after weathering showing microcracking.

Figure 2a: Pocket wear - before.

Figure 2b: Pocket wear - after.

Figure 3: Making replicas in the field using dental impression material.

Figure 4a: Unmodified, freshly excavated, saturated bone fragment. A fine crack can be seen in the middle of the image.

Figure 4b: The same bone as in Figure 4a, after drying. The crack has expanded considerably.

Musical Behaviors and the Archaeological Record: A Preliminary Study Ian Cross, Ezra B.W. Zubrow and Frank Cowan

Abstract The re-emergence of an evolutionary perspective on the human mind over the last decade has resulted in a range of different theories about the origin of cognition. Although the origin of language has a long history in cognitive studies, the origin of music is essentially a "tabula rasa ". Does music have its evolutionary roots in animal 'song' or is it a uniquely human behavior? Is music an evolutionary 'by-product', an accident of nature that is inessential to human survival, or is it central to the evolution of the modern human mind? Is the origin of music causal, resultant, or incidental in the separation of modern humans from Neanderthals? There is both historic and prehistoric evidence for the use of lithophones in making music. This paper discusses the results of using simulation and experimental archaeology to reconstruct lithophones from the Upper Paleolithic. Modern knapped blades are analyzed for their sound characteristics during percussion. Some Upper Paleolithic-type flint blades are similar to chime bars producing extraordinarily clearly pitched and resonant notes when played by experienced musicians who have taught themselves to play "the blades". The tonal quality and the actual notes are not only consistent but predictable. Acoustic engineering allows one to actually write equations to predict the notes and tones. A microscopic analysis of these simulated stone-age musical instruments shows that the production of "musical sound" resulted in unique "use-wear" patterns. Examining the archaeological record, several blades were found with this musical "use-wear" pattern at Cro-Magnon. The usefulness of having a means of identifying musical use of stone - over bone, wood, etc. - is that stone goes much further back in the archaeological record; if 'flint tools' were indeed being used to produce musical sound it might be possible to track such artefacts back into the depths of the Paleolithic in an attempt to make out just when the production of musical sound might have begun.

Introduction

Experimental archaeology has several distinct functions. First, it is a method to test a hypothesis about past behavior or past technology. In this sense, it is one of a series of scientific endeavors similar to statistics, test bed experimentation, simulation, or stratigraphic analysis that allows one to disprove a hypothesis. There are many examples. Archaeologists have been known to bury artifacts in the ground and plow the soil in order to determine the degree of pattern deformation. Similarly, they have been known to put artifacts on a beach to examine the changes that take place with the changing tides or record artifactual movements using a wave machine. Second, it is a method to learn how to use "prehistoric technology or behavior". In this case, it is an educational technique for the archaeologist. There are lithic specialists who learn how to make fluted points or Mousterian cores by trial and error as well as by lessons from more experienced lrnappers. Similarly, metallurgical archaeologists are concerned with the investigation of all aspects of the metallurgical process - from smelting to metal finishing. They frequently build prehistoric forges to test out their ideas. Zubrow's brother who is a blacksmith frequently pronounces the only way to learn to smith a tool is to do it. Not all experimental archaeology is about technology. Consider the Maya ball game. Although considerable information and literature exists, it is remarkably inadequate. Somehow it does not answer those detailed questions that only occur when one is actually a player. For example, what should one do when the ball leaves the ball court in order to begin play again? Does one throw the ball in from the sidelines as in soccer/

football or is it centered as in basketball? Third, the function may be to educate others. There are numerous reconstructed "iron age camps", medieval farmsteads, and colonial villages where ethnoarchaeology and experimental archaeology are used to educate the "populace". Finally, experimental archaeology is a method to create domains of information that would otherwise be totally lost. The domain of prehistoric music is rather ephemeral. Experimental archaeology allows us to reconstruct the musical instruments, play them, and hear the music for the first time since the prehistoric population last played it. This paper is an experimental archaeological study attempting to discover the earliest origin of music. The research discussed here has three different points of origin: the relation between music and human evolution, specifically, human cognitive evolution; the nature of the evidence for musical behaviors in the archaeological record; and the issue of making musical sounds with stones. One might suppose that music, as a cultural phenomenon, has little to do with evolution. But, from a cognitive-scientific perspective, music is inescapably material, being evidenced in musical behaviors; behind human behaviors lie human minds, and behind human minds lie embodied human brains. Accepting a materialist basis for human behaviors, consideration of evolution's role in those behaviors seems inescapable. Taking an evolutionary approach to human behaviors does not necessitate adoption of a gene-centered ontological reductionism; indeed, it may be that evolutionary perspectives afford excellent frameworks within which an understanding of music as individual minded behavior -

material practice - can be reconciled with an understanding of music as embedded in a nexus of shared ways of understanding - music as culture. The existence of an evolutionary basis for music is unlikely to be explanatory of most of the attributes, significances, purposes and interpretations that can be borne by the music of any particular culture. But it can provide some hypotheses about the dynamics of cognition and interaction that may underlie those attributes, significances, etc. And some recently developed hypotheses about the relation between music and evolution (see Cross 1999; Brown 2000; Dissanayake 2000) constitute the broad context for the present research specifically, that the emergence of 'musicality' played a significant role in the evolution of modem humans, Homo sapiens sapiens. Turning to the nature of the evidence for musical behaviors in the archaeological record, we run into several problems. What traces would musical behaviors leave? Given that the earliest such behaviors were likely to have been vocal, we are left with trying to make inferences about whether or not any of our predecessors or sibling species had the vocal capacity to articulate the complex timbral and pitch patterns that music requires on the basis of fragmentary human and pre-human remains, and several equally plausible theories appear to lead to different conclusions (see Lieberman 1991; Frayer and Nicolay 2000). In any case, all that such research can tell us is whether or not our ancestors had the capacity to produce 'musical' sounds - it can't tell us whether they produced music. Artefacts provide clearer evidence - one might suppose. But there is controversy over just what the earliest musical artefact might be. On one reading, the earliest artefact is an unambiguously musical bone pipe from Geissenklosterle in Germany, dated to about 36,000 BP and associated with modern humans (see Hahn and Munzel 1995); on another reading, the earliest evidence is a fragment ofa bone pipe from Divje babe in Slovenia, dated to around 45,000 BP and associated with Homo neanderthalensis (Kunej and Turk 2000) - though, alternatively, this 'bone pipe' might have been a hyena's lunch (D'Errico and Villa 1997). One of the aims of the current experimental archaeological research is to attempt to work out ways of identifying whether or not an artefact has been purposively produced by human activity and has been unambiguously employed to make sounds. And finally, we tum to making musical sounds with stones. It seems that most human cultures either do this or have done this; evidence for the use oflithophones - lithic idiophones stretches from Sweden to southern Africa, from the Canaries through Kenya through Vietnam through China to Potosi in the Bolivian Andes (see the entry for 'Lithophones' in The New Grove Dictionary 1980). It even crops up in Victorian England, where the brothers Richardson performed on their specially constructed 'geological piano' before Queen Victoria (her response is not recorded). The possibility that our ancestors might have exploited the materials and technologies that they knew best- flint, and the processes of working flint to produce artefacts - for sound-production constitutes the narrow context for the experimental research now sketched out, the Lithoacoustics Project.

The Lithoacoustics Project

The practical origins of the project arose from posing the question "what traces would musical behaviors leave"? It was eventually agreed that it would be worth exploring the materials and the percussive processes involved in flintknapping to find out (I) whether sounds that could be interpreted as musical could be produced, and (ii) whether producing musical sounds would leave any unambiguous traces. To leap to the interim findings (the project is not yet completed) the answer to the two questions appears to be "yes" and "yes". It was decided to focus on the first instance on the tools and

the technologies of the Aurignacian period (about 40,000 to 20,000 BP). This is because peoples of around that time used stalagmitic rock formations in caves as lithophones (Dams 1985), so it seems reasonable to assume that they might have used other types of stones as well. As soon as we began the process of flint-knapping we realized that we had some potential musical instruments in the blades produced (Figure 1).1 The easiest way to 'play' a blade is to suspend it between thumb and forefinger (or middle finger) about a quarter of the way along its length and strike it in the middle or at the bottom end, as shown in Figure 2. It transpires that flint blades can be used as idiophones -

musical instruments or vibrating objects in which energy input and sound output systems are one and the same - which behave like chime bars; when struck, their first mode of vibration (lowest pitch) has nodal points (points of null displacement) at about 0.224 along their length, and they can produce very clear and quite long-lasting pitched sounds (Figure 3). Formal protocols were developed for: (I) categorizing the blades - specimens - on the basis of their physical dimensions and attributes; (ii) formalizing and quantifying the procedures to be used in sound production; (iii) analyzing and categorizing the resulting sounds; and (iv) analyzing and typologizing the damage that accrued to the surface of the blades when they had been used to make sounds. The third author produced and categorized the specimens; then one of the two assistants on the project (the 'performers') used each specimen to make sounds, assessed its 'playability' according to a number of parameters, recorded the sound at the outset of trialling, percussed the blade for five minutes, recorded the sound again, percussed for a further five minutes, and recorded the sound for a last time. The recorded sounds were then analyzed (using CERL's Lemur software); each specimen was then examined under an optical microscope and digital images taken. Finally, the surface damage or use-wear on each specimen was assessed and coded. Some 116 specimens were used, of which 'before' and 'after' microscope photographs were taken of fifteen; two of the specimens were used as percussors. All measurements were entered into a database and the process of analysis was started. 1

Typical blanks were produced using a prepared core technology.

Ian Cross,

BJ/11Zubrow and Frank Cowan Musical Behaviors and the Archaeological Record: A Preliminary Study

The Results

of surface coning is shown in Figure 4 and an instance of surface polish is shown in Figure 5.

Sound and Performance Taking the sound data first, it was found that, overall, the frequencies, durations and intensities of all specimens conformed to normal distributions; taking the rating of each specimen by the 'performers' into account, clear differences in frequency were found between those specimens rated 'good' and those rated 'acceptable' or 'bad' (see Table 1). The mean principal frequency of the 'good' specimens is at the upper end of the usable 'musical' frequency range (after Attneave and Olson 1971), while those of the 'acceptable' and 'bad' specimens was well outside this range; the duration of the 'good' specimens was considerably greater that both the 'acceptable' and 'bad'; while the intensity of the 'bad' specimens was much lower than both 'good' and 'acceptable' intensities. The consistency of these physical values suggests that the categories in which the specimens were placed by the raters in respect of"playability" are (I) directly relatable to the sound-producing characteristics of the specimens and (ii) real. And substantial inter-rater reliability was evident in a series of t tests which showed no significant differences between additional ratings given by each of the two raters to each specimen on dimensions of"pitchedness", "resonance", "power" and "piercingness". Further t tests on the recorded sound values for all specimens at the outset and at the end of trialling yielded no evidence that repeated percussion changed the sounding qualities of any specimen.

In many instances, edge damage in the form of small, abrupt, step-terminated or hinge-terminated flake scars were found where playing percussion was near an edge. An instance of this is shown in Figures 6a and 6b. The cone-cracking results from direct, head-on percussion, while the polishes and scratches may result from a softer and more "stroking" impact against the flake surface. In many instances, the cone-fracturing consisted of multiple, overlapping cone-cracks that often occurred in great density, as can be seen in Figure 7 which shows the same surface area before and after percussion. One of the most salient features of the cone-cracking wear is its placement. Because of the nodal regions on the chimebar-like blades, musical wear tended to occur most frequently either at the midpoint of the specimen or at the end of the specimen (beyond the nodal region furthest from the suspension point). This non-random distribution is probably unique to musical play, and is localized to the faces on the antinodal areas. The effect can clearly be seen in Figures Sa and Sb showing the same surface area (bounded by a circle drawn on the specimen at the distal, far, end). Of the three different kinds of damage, the cone-cracking was most consistent and is undoubtedly the most diagnostic usewear criterion. No other behavioral or geological forces that we can think of are likely to produce the kind of very patterned clustering of cone-cracks as were experimentally produced in musical use. Microscopic images clearly show the patterns of use-wear resulting from this musical use:

A series of multiple regression analyses (with principal frequency, principal intensity and principal duration as the respective dependent variables and length, width and thickness as the independent variables) showed that for all specimens both length and thickness had highly significant predictive value for the intensity and duration of the sounds produced. However, a more complex variable obtained by dividing the thickness of each specimen by the square of its length (T /L 2) provided a very highly significant predictor for frequency in simple regressions for all rated categories. This complex variable was derived from an equation describing the physics of "chime bars", where principal frequency is a complex function of, among other things, length and thickness (though not width). Its functionality as a predictor of the frequencies of the sounds produced confirms that the chime-bar model is operational in respect of these lithic resonators. This set ofresults can be read as indicating that to a "player" a heuristic indication of the sound-producing capacity of the specimen is immediately available from estimation of its length and (secondarily) its thickness.

It is also noteworthy that use-wear intensity varied with the player. One 'performer' produced a wide range of use-wear patterns, including soft, stroking polishes on the surfaces and very seldom produced intensive edge attrition. The other 'performer', on the other hand, tended to strike the resonators more directly and with greater force. Hence, this performer's instruments tended to accrue, very rapidly, much more densely clustered cone-cracks. This latter performer's instruments also were extensively and intensively "retouched" along the marginal edges. Several specimens that were initially unretouched blades or flakes became typologically identifiable "tool" types with extensive alteration of specimen outline. These "retouched" edges were formed by "play" near the edge of the piece, and the force was sufficient to strike off multiple, overlapping retouch flakes. Nonetheless, the patterns of edge retouch are not very similar to intentional technological retouch.

Use-wear

Initial Survey of Museum Collections

While formal use-wear analysis is not yet complete, it was immediately clear that repeated percussion resulted in the consistent appearance of small densely clustered surface cones or of multiple small, densely clustered small areas of surface polish. Occasionally, small scratches occurred. An instance

A preliminary examination was then conducted of some of the flint-tool holdings of the Cambridge University Museum of Archaeology and Anthropology. Approximately 425 (10 kg) archaeological specimens were examined from Aurignacian levels of Laugerie Haute, Cro-Magnon, Le

Experimental Archaeology Replicating Past Objects, Behaviors. and

Moustier, Masnaigre, and other French Upper Paleolithic sites (from the Museum's holdings of an estimated 3000 flint specimens of the period). All were scanned for traces of surface use-wear, especially cone-cracking, with a l0x hand lens. Three important observations can be made from this pilot study. First, cone-fracturing on the ventral surfaces of flakes, blades or tools is extremely rare in the archaeological record. Four specimens out of 425 were observed to have a few potential surface cone-cracks on the ventral surface. This means that this kind of damage is not a common result of either a) prehistoric behavior, b) fortuitous geological processes after deposition in the archaeological deposits, c) excavation damage, or d) post-excavation curation damage (bag-damage). Second, cones are potentially recognizable on ancient archaeological specimens, despite raw material variability, surface patination, breakage, or other altering forces. Third, none of the identified specimens approximated the patterns of wear routinely observed on the experimental specimens. It is therefore clear that musical use of flint blades will result in a very different overall pattern and distribution of cone-cracks than other behavioral or fortuitous causes. So far as our limited exploration of the archaeological record is concerned, there appear to be very few instances of blades or flakes with small surface coning, so if it occurs as a result of "natural" circumstances it would seem to be rare and likely to be differentiable from the type of wear that arises from lithic chime percussion.

Conclusions At present, the use-wear coding remains to be completed, hence our present conclusions must be qualified somewhat; however, the results that emerge seem to indicate that there are patterns of use-wear on the flint blades that we made and experimented with that are diagnostic of use for sound production. What might be the implications of this? To return to the issues considered at the outset, it appears that we are now in a position to say whether or not Aurignacian-type flint blades have been used as lithophones. We know that they can be 2 , and it appears that doing so leaves diagnostic traces, so we may now be in a position to identify unambiguously traces of sound production, and, perhaps, 'musical' performance, in the archaeological record, which will involve examining whether or not any artefacts that have been interpreted as flint tools were in fact used for sound production and perhaps for music. Differentiation between simple sound production - for example, using a flint blade as a sort of Paleolithic doorbell - and 'musical' use will always be a matter of interpretation of both the artefact and the find context, but finding, say, a grouping of lithic blades all of which exhibit appropriate and localized cone-cracking would be likely to point towards something like music. In this context it would also be of interest to explore whether or not other and earlier 2

Examples of sounds produced can be found at http://www.mus.cam.ac.uk/~cross/lithoacoustics/.

w 00 , 0

A 0 0e 00

lithic tool technologies can be exploited in a similar way for sound production, and if so, what traces of use-wear might result. This project has also shed light on some considerations in exploring the nature of the evidence for sound production in the archaeological record. While it is evident that there will be a relation between patterns of use-wear and the acoustical properties of the objects used to produce sounds, here, a very close fit has been found between acoustical properties and use-wear. This close fit derives from the nature and from the configuration of the materials used and from the constraints that these impose on sound producing action. Indeed, the patterns of use-wear found here should have been predictable in advance from an understanding of the chime-bar like acoustical properties of flint blade idiophones. Although the fit is unlikely to be so close in respect of other materials and configurations (the case of pipes made from bone is one such), it would be worthwhile exploring other materials - bone, wood, and perhaps bamboo - and configurations, particularly where these afford the capacity to be used as idiophones as here the relation between acoustical attributes and use-wear can be expected to be very close. The outcome of the project might have some significance for our understanding of the relation between music and evolution. Music has been posited as sharing its origins with language (Brown, 2000), and as having been adaptive in precipitating the emergence of the cognitive ands social flexibility characteristic of modem humans. But whether or not music has been adaptive, exaptive or even neutral in respect of human evolution, it is still of interest to discover just when a capacity or propensity for music appeared. Music certainly appears early in the behavioral repertoire of Homo sapiens sapiens; the Geissenklosterle pipe at 36,000 BP is a complex artefact that must post-date - and most likely by some considerable period- the emergence of a capacity for music, which pushes the emergence of that capacity back towards the very emergence of Homo sapiens sapiens. The longevity of music as a human behavior is evident (if seldom recognized). The results of the present project and the directions that it suggests for future research might help answer some questions about the extent of that longevity and whether or not music is a capacity that we shared with our sibling and predecessor species. And finally, it is clear that experimental archaeology opens new domains of archaeological information. Music may be one of the civilizing forces of human behavior. Clearly, it is as important a domain as technology and ideology.

References Cited Attneave, F. and Olson, R. K. 1971 'Pitch as a medium: a new approach to psychophysical scaling' American Journal of Psychology 84:147-166. Brown, S. 2000 'The "musilanguage" model of music evolution' in N.L. Wallin, B. Merker and S. Brown (eds) The Origins of Music MIT Press (Cambridge, MA):271-300.

Ian Cross,

BJ/11Zubrow and Frank Cowan Musical Behaviors and the Archaeological Record: A Preliminary Study

Cross, I. 1999 'Is music the most important thing we ever did? Music, development and evolution' in Suk Won Yi (ed) Music, Mind and Science Seoul National University Press (Seoul):1039. Dams, L. 1985 'Paleolithic lithophones: descriptions comparisons' Oxford Journal of Archaeology 4(1):31-46.

and

D'Errico, F. and Villa, P. 1997 'Holes and grooves: the contribution of microscopy and taphonomy to the problem of art origins' Journal of Human Evolution 33(1):1-31. Dissanayake, E. 2000 'Antecedents of the temporal arts in early mother-infant interaction' in N.L. Wallin, B. Merker and S. Brown (eds) The Origins of Music MIT Press (Cambridge, MA):389410

Frayer, D. W. and Nicolay, C. 2000 'Fossil evidence for the origins of speech sounds' in N.L. Wallin, B. Merker and S. Brown (eds) The Origins of Music MIT Press (Cambridge, MA):217-234. Hahn, J. and Munzel, S. 1995 'Knochenfloten aus dem Aurignacien des Geissenklosterle bei Blaubeuren, Alb-Donau-Kreis' Fundberichte aus Baden-Wiirttemberg 20:1-12. Kunej, D. and Turk, I. 2000 'New perspectives on the beginnings of music: archaeological and musicological analysis of a Middle Paleolithic bone "flute"' in N.L. Wallin, B. Merker and S. Brown (eds) The Origins ofMusic MIT Press (Cambridge, MA):234-268. Lieberman, P. 1991 Uniquely Human Harvard University Press (Cambridge, MA). 'Lithophones' Entry 1980 The New Grove Dictionary of Music and Musicians Macmillan (London).

Figure 1: Examples of the blades produced and used in the project.

Figure 2: Playing a blade, here suspended between thumb and middle finger of left hand.

nodal is

uency of first mode of vibration v- -voung•s modulus of elasticity 6=densily t=thickness L=length 1

Figure 3: The acoustical functioning of a chime bar in usual horizontal position (top) and in the vertical position employed in playing the flint blades (middle). The equation (bottom) shows the terms involved in determining the frequency of the first mode of vibration.

30

Ian Cross.

BJ/11Zubrow and Frank Cowan Musical Behaviors and the Archaeological Record: A Preliminary Study

Figure 4: Surface coning on specimen 60 resulting from sound production.

Figure 5: Surface polish on specimen 18 resulting from sound production.

Figure 6a: Surface and edge of specimen 104 before percussion.

Figure 6b: Surface and edge of specimen 104 after percussion. Note the extensive surface coning and the edge damage.

34

Figure 7: Surface of distal end of specimen 101 before and after percussion.

Figure 8a: Bounded area on surface of distal end of specimen 108 before percussion.

Figure 8b: Bounded area on surface of distal end of specimen 108 after percussion.

32

Ian Cross,

BJ/11Zubrow and Frank Cowan Musical Behaviors and the Archaeological Record: A Preliminary Study

Table 1: Mean principal frequencies, durations, and intensities of good, acceptable, and bad specimens, and results of a series oft tests between values.

goodsignificantly different goodsignificantly different nosignificant difference (p20 cm deep). Three made partial wounds (5-10 cm deep), and none failed (