Processual Archaeology (essays)

Processual Archaeology

Lewis Binford (Essays)

2011

In the memory of Lewis Binford

“The Father of New Archaeology”

(November 21, 1930 – April 11, 2011)

"Lewis Binford led the charge that pushed, pulled and otherwise cajoled archaeology into becoming a more scientific enterprise." David Meltzer, Henderson-Morrison Professor of Prehistory at SMU.

Lewis Roberts Binford, Ph.D. (November 21, 1930 – April 11, 2011) was an American archaeologist most famous for his role in establishing the "New Archaeology" movement of the 1960s. Lewis Binford was born in Norfolk, Virginia, and raised near the Atlantic coast in rural Virginia. His family's usual income came through carpentry, and he put his own skills in this area to use as a graduate student when he started his own contracting business. His interest in anthropology was sparked while working as an interpreter for the U.S. Army in rural villages in postwar Okinawa.[citation needed] He died in Kirksville, Missouri in 2011. Binford's education includes a B.A. degree from the University of North Carolina and an M.A. and a Ph.D. (1964) from the University of Michigan. Binford joined the Southern Methodist University faculty in 1991, after teaching for 23 years as a distinguished professor at the University of New Mexico. Some of Binford's early work was prepared in collaboration with then-wife Sally Binford. Their best-known collaboration is the influential anthology New Perspectives In Archaeology. Binford's last published book, Constructing Frames of Reference (2001), was intensively edited by Nancy Stone. Lewis Binford is mainly known for his contributions to archaeological theory and his promotion of ethnoarchaeological research. As a leading advocate of the "New Archaeology" movement of the 1960s, he proposed a number of ideas that matured into Processualism. He and others argued that there should be emphasis on the application of scientific methodologies as well as the hypothetico-deductive method to archaeology. He places a strong emphasis on generalities and the way in which human beings interact with their ecological niche, defining culture as the extrasomatic means of adaptation. This view reflects the influence of his Ph.D supervisor, Leslie White. Binford's work can largely be seen as a reaction to the culture history approach that

preceded the New/Processual Archaeology. The New Archaeology was long considered a revolution in archaeological theory, but in recent years it has increasingly been seen as the outcome of a more gradual change in archaeological perspectives, thanks to the rediscovery of Walter Taylor's A Study of Archeology, which Binford himself has mentioned as an important influence.

Lewis R. Binford has devoted most of his busy professional life to the archaeology of hunting and gathering. Much of his work on ancient hunting and gathering was done by focusing on the present: He lived with Nunamuit reindeer hunters, learned from Australian elders, studied Bushmen and visited many of the fast-vanishing hunting and gathering groups in today's world. He also compiled thousands of ancient stories and historic reports of the hunting and gathering societies that existed before those groups were displaced by farmers. He compared and contrasted those accounts with a zealous attention to textual detail that could win the admiration of a French literary critic. His vast "data set" on modern and historic hunting and gathering societies constitutes one axis of his research. The other axis is spatial, geographic, natural: a second enormous "data set" of global environmental measures shown (by Binford) to be relevant to hunting and gathering lifestyles, both in the present and in the past. In Constructing Frames of

Reference, Binford arrays the hunter-gatherer data set against the environmental data set in a series of "pattern recognition" experiments. One of Binford's target audiences, of course, is the profession of archaeology, whether of hunting and gathering societies or high civilizations. "Archaeology" is done by professors and curators in art history, religious studies, classics, astronomy and a surprising number of other disciplines, but Binford's archaeology is part of anthropology. He sees anthropological archaeology as a young science, still developing properly rigorous methods for understanding the past. This book is presented as a weighty example of "analytical method for archaeological theory building," in which he (correctly) judges anthropological archaeology as deficient. Binford was involved in several high-profile debates including arguments with James Sackett on the nature and function of style and on symbolism and methodology with Ian Hodder. Binford has spoke out and reacted to a number of schools of thought, particularly the Post-processual school, the Behavioural school, and the symbolic and postmodern anthropologies. Binford was also known for a friendlier rivalry with French archaeologist François Bordes, with whom he argued over the interpretation of Mousterian sites. Binford's disagreement with Bordes over the interpretation of Mousterian stone artifacts provided the impetus for much of Binford's theoretical work. Bordes interpreted variability in Mousterian assemblages as evidence of different tribes, while Binford felt that a functional interpretation of the different assemblages would be more appropriate. His subsequent inability to explain the Mousterian facies using a functional approach led to his ethnoarchaeological work among the Nunamiut and the development of his middle-range theory.

An article in the November–December 1999 Scientific American magazine described Binford as "quite probably the most influential archaeologist of his generation. In 2000, he received an honorary doctorate at Leiden University for his role in the development of a more scientific archaeology. In 2001, Binford was elected to the National Academy of Sciences. The Society for American Archaeology awarded Binford a Lifetime Achievement Award at their annual meeting in Vancouver, British Columbia (May 28, 2008). On May 27, 2010, the International Astronomical Union (IAU) announced that asteroid (213629) Binford was named in honor of Binford.

Works Constructing frames of reference:an analytical method for archaeological theory building using hunter-gatherer and environmental data sets Berkeley: University of California Press, (2001) ISBN 0520223934 Debating Archaeology San Diego: Academic Press, (1989) ISBN 0121000451 Faunal Remains from Klasies River Mouth (1984) ISBN 0-12-100070-2 In Pursuit of the Past: Decoding the Archaeological Record (1983) ISBN 0-520-23339-5

Bones, Ancient Men and Modern Myths (1981) ISBN 0-12-100035-4 Nunamiut Ethnoarchaeology (1978) ISBN 0-12-100040-0 An archaeological perspective New York: Seminar Press, (1972) ISBN 0128077506 New Perspectives in Archaeology (1968) ISBN 0-202-33022-2 Archaeology as Anthropology (1962)

ARCHAEOLOGY AS ANTHROPOLOGY

LEWIS R. BINFORD ABSTRACT

ties and differences characteristic of the entire spatial-temporal span of man's existence (for discussion, see Kroeber 1953). Archaeology has certainly made major contributions as far as explication is concerned. O u r current knowledge of the diversity which characterizes the range of extinct cultural systems is far superior to the limited knowledge available fifty years ago. Although this contribution is "admirable" and necessary, it has been noted that archaeology has made essentially no contribution in the realm of explanation: "So little work has been done in American archaeology on the explanatory level that it is difficult to find a name for it" (Willey and Phillips 1958: 5). Before carrying this criticism further, some statement about what is meant by explanation must be offered. The meaning which explanation has within a scientific frame of reference is simply the demonstration of a constant articulation of variables within a system and the measurement of the concomitant variability among the within the system. ~~~~~~~~~l can then be shown to rechange in One late in a predictable and quantifiable way to changes in other variables, the latter changing in turn relative to changes in the structure of the system as a whole. This approach to explanation PreSuPPosesconcern with process, Or the Operation and structural modification of systems. It is Suggested that archaeologists have not made major explanatory contributions to the field of because they do conceive of archaeological data in a systemic frame of reference. Archaeological data are viewed particulariStically and "explanation" is offered in terms of specific events rather than in terms of process (see Buettner-Janusch 1957 for discusT HAS BEEN aptly stated that "American sion of particularism). archaeology is a n t h r o ~ o l o gor~ it is nothing" Archaeologists tacitly assume that artifacts, ( W i l l e ~and Phillips 1958: 2 ) . T h e Purpose of regardless of their functional context, can be this discussion is to evaluate the role which the treated as equal and "traits." Once archaeological discipline is playing in furthering differences and similarities are "definedn in the aims of anthropology and to offer certain terms of these equal and "traits," suggestions as to how we, as archaeologists, may interpretation proceeds within something of a profitably shoulder more responsibility for fur- theoretical vacuum that conceives of differences thering the aims of our field. and similarities as the result of "blending," Initially, it must be asked, "What are the aims "directional influences," and "stimulation" beof anthropology?" Most will agree that the inte- tween and among "historical traditions" defined grated field is striving to explicate and explain largely on the basis of postulated local or rethe total range of physical and cultural similari- gional continuity in the human populations.

It is argued that archaeology has made few connibutions to the general field of anthropology with regard to explaining cultural similarities and differences. One major factor contributing to this lack is asserted to be the tendency to treat artifacts as equal and comparable traits which can be explained within a single model of culture change and modification. It is suggested that "material culture" can and does represent the structure of the total cultural system, and that explanations of differences and similarities between certain classes of material culture are inappropriate and inadequate as explanations for such observations within other classes of items. Similarly, change in the total cultural system must be viewed in an adaptive context both social and environmental, not whimsically viewed as the result of "influences," "stimuli," or even "migrations" between and among geographically defined units. Three major functional sub-classes of material culture are discussed: technomic, socio-technic, and ideo-technic, as well as stylistic formal properties which cross-cut these categories. In general terms these recognized classes of materials are discussed with regard to the processes of change within each class. Usine the above distinctions in what is termed a svstemic approach, the problem of the appearance and chawing utilization of native copper in eastern North America is discussed. Hypotheses resulting from the application of the systemic approach are: (1) the initial appearance of native copper implements is in the context of the production of socio-technic items; (2) the increased production of socio-technic items in the late Archaic period is related to an increase in population following the shift to the exploitation of aquatic resources roughly coincident with the Nipissing high water stage of the ancestral Great Lakes; (3) this correlation is explicable in the increased selective pressures favoring material means of status communication once populations had increased to the point that personal recognition was no longer a workable basis for differential role behavior; (4) the general shift in later periods from formally "utilitarian" items to the manufacture of formally unonutilitarianfl items of copper is explicable in the postulated shift from purely egalitarian t~ increasingly nonegalitarian means of status attainment.

I

217

218

AMERICAN ANTIQUITY

I suggest that this undifferentiated and unstructured view is inadequate, that artifacts having their primary functional context in different operational sub-systems of the total cultural system will exhibit differences and similarities differentially, in terms of the structure of the cultural system of which they were a part. Further, that the temporal and spatial spans within and between broad functional categories will vary with the structure of the systematic relationships between socio-cultural systems. Study of these differential distributions can potentially yield valuable information concerning the nature of social organization within, and changing relationships between, socio-cultural systems. In short, the explanation of differences and similarities between archaeological complexes must be offered in terms of our current knowledge of the structural and functional characteristics of cultural systems. Specific "historical" explanations, if they can be demonstrated, simply explicate mechanisms of cultural process. They add nothing to the explanation of the pro'cesses of cultural change and evolution. If migrations can be shown to have taken place, then this explication presents an explanatory problem; what adaptive circumstances, evolutionary processes, induced the migration (Thompson 1958: l ) ? W e must seek explanation in systemic terms for classes of historical events such as migrations, establishment of "contact" between areas previously isolated, etc. Only then will we make major contributions in the area of explanation and provide a basis for the further advancement of anthropological theory. As an exercise in explication of the methodological questions raised here, I will present a general discussion of a particular systemic approach in the evaluation of archaeological assemblages and utilize these distinctions in an attempted explanation of a particular set of archaeological observations. Culture is viewed as the extra-somatic means of adaptation for the human organism (White 1959: 8). I am concerned with all those subsvstems within the broader cultural system which are: (a) extra-somatic or not, dependent upon biological process for modification or structural definition (this is not to say that the form and process cannot be viewed as rooted in biological process, only that diversity and processes of diversification are not explicable in terms of biological process), and which (b) function to

[ VOL.28, NO.2, 1962

adapt the human organism, conceived generically, to its total environment both physical and social. Within this framework it is consistent to view technology, those tools and social relationships which articulate the organism with the physical environment, as closely related to the nature of the environment. For examvle. ,we would not expect to find large quantities of fishhooks among the recent archaeological remains from the Kalahari desert! However, this view must not be thought of as "environmental determinism" for we assume a systematic relationship between the human organism and his environment in which culture is the intervening- variable. In short, we are speaking of the ecological system (Steward 1955: 36). W e can observe certain constant adawtive reauirements on the part of the organism and similarly certain adaptive limitations, given specific kinds of environment. However. limitations as well as the wotential of the environment must be viewed always in terms of the intervening variable in the human ecological system, that is, culture. we should not be With such an avvroach -surprised to note similarities in technology among groups of similar levels of social complexity inhabiting the boreal forest (Spaulding 1946) or any other broad environmental zone. The comparative study of cultural systems with variable technoloaies - in a similar environmental range or similar technologies in differing environments is a major methodology of what Steward (1955: 36-42) has called "cultural ecology," and certainly is a valuable means of increasing our understanding of cultural processes. Such a methodolo~v -, is also useful in elucidating the structural relationships between major cultural sub-systems such as the social and ideological sub-systems. Prior to the initiation of such studies by archaeologists we must be able to distinguish those relevant artifactual elements within the total artifact assemblage which have the primary functional context in the social, technological, and ideological sub-systems of the total cultural system. W e should not equate "material culture" with technology. Similarly we should not seek explanations for observed differencesand similarities in "material culture" within a single interpretative frame of reference. It has often been suggested that we cannot dig up a social system or ideology. Granted we cannot excavate a kinshiv terminology or a philosophy, but we can and do exca-

h

~

t

.

I0

~

~

ARCHAEOLOGY AS ANTHROPOLOGY

vate the material items which functioned together with these more behavioral elements within the appropriate cultural sub-systems. T h e formal structure of artifact assemblages together with the between element contextual relationships should and do present a systematic and understandable picture of the total extinct cultural system. It is no more justifiable for archaeologists to attempt explanation of certain formal, temporal, and spatial similarities and differences within a single frame of reference than it would be for an ethnographer to attempt explanation of differences in cousin terminology, levels of socio-cultural integration, styles of dress, and modes of transportation all with the same variables or within the same frame of reference. These classes or items are articulated differently within an integrated cultural system, hence the pertinent variables with which each is articulated, and exhibit concomitant variation are different. This fact obviates the single explanatory frame of reference. The processes of change pertinent to each are different because of the different ways in which they function in contributing to the total adaptive system. Consistent with this line of reasoning is the assertion that we as archaeologists must face the problem of identifying technomic artifacts from other artifactual forms. Technomic signifies those artifacts having their primary functional context in coping directly with the physical environment. Variability in the technomic components of archaeological assemblages is seen as primarily explicable in the ecological frame of reference. Here, we must concern ourselves with such phenomena as extractive efficiency, efficiency in performing bio-compensatory tasks such as heat retention, the nature of available resources, their distribution, density, and loci of availability, etc. In this area of research and explanation, the archaeologist is in a position to make a direct contribution to the field of anthropology. W e can directly correlate technomic items with environmental variables since we can know the distribution of fossil flora and fauna from independent data - giving us the nature of extinct environments. Another major class of artifacts which the archaeologists recover can be termed socio-technic. These artifacts were the material elements having their primary functional context in the social sub-systems of the total cultural system. This sub-system functions as the extra-somatic means of articulating individuals one with an-

219

other into cohesive groups capable of efficiently maintaining themselves and of manipulating the technology. Artifacts such as a king's crown, a warrior's coup stick, a copper from thi. Northwest coast, etc., fall into this category. Changes in the relative complexity of the socio* technic component of an archaeological assem) blage can be related to changes in the structure of the social system which they represent. Certainly the evolutionary processes, while correlated and related, are not the same for explaining structural changes in technological and social phenomena. Factors such as demography, presence or absence of between-group competition, etc., as well as the basic factors which affect technological change, must be considered when attempting to explain social change. Not only are the relevant variables different, there is a further difference when speaking of sociotechnic artifacts. The explanation of the basic form and structure of the socio-technic component of an artifactual assemblage lies in the nature and structure of the social system which it represents. Observable differences and changes in the socio-technic components of archaeological assemblages must be explained with reference to structural changes in the social system and in terms of processes of social change and evolution. Thus, archaeologists can initially only indirectly contribute to the investigation of social evolution. I would consider the study and establishment of correlations between types of social structure classified on the basis of behavioral attributes and structural types of material elements as one of the major areas of anthropological research yet to be developed. Once such correlations are established, archaeologists can attack the problems of evolutionary change in social systems. It is my opinion that only when we have the entire temporal span of cultural evolution as our "laboratory" can we make substantial gains in the critical area of social anthropological research. T h e third major class of items which archaeologists frequently recover can be termed ideotechnic artifacts. Items of this class have their primary functional context in the ideological component of the social system. These are the items which signify and symbolize the ideological rationalizations for the social system and further provide the symbolic milieu in which individuals are enculturated, a necessity if they are to take their place as functional participants

220

AMERlCAN ANTIQUlTY

[ VOL. 28, No. 2,1962

in the social system. Such items as figures of correlations of the structure of artifact assemdeities, clan symbols, symbols of natural agen- blages with rates of style change, directions of cies, etc., fall into this general category. Formal style-spread, and stability of style-continuity. diversity in the structural complexity and in Having recognized three general functional functional classes of this category of items must classes of artifacts: technomic. socio-technic. generally be related to changes in the structure and ideo-technic, as well as a category of formal of the society, hence explanations must be stylistic attributes, each characterized by differsought in the local adaptive situation rather ing - functions within the total cultural svstem than in the area of "historical explanations." As and correspondingly different processes of was the case with socio-technic items, we must change, it is suggested that our current theoretiseek to establish correlations between generic cal orientation is insufficient and inadeauate for classes of the ideological system and the struc- attempting explanation. It is argued that exture of the material symbolism. Only after such planations of differences and similarities becorrelations have been established can archae- tween archaeological assemblages as a whole ologists study in a systematic way this compon- must first consider the nature of differences in each of these major categories and only after ent of the social sub-system. Cross-cutting all of these general classes of such evaluation can adequate explanatory hyartifacts are formal characteristics which can be potheses be offered. Given this brief and oversimplified introductermed stylistic, formal qualities that are not directly explicable in terms of the nature of the tion, I will turn to a specific case, the Old Copraw materials, technology of production, or vari- per complex (Wittry and Ritzenthaler 1956). ability in the structure of the technological and It has long been observed and frequently cited social sub-systems of the total cultural system. as a case of technological "devolution" that durThese formal aualities are believed to have their ing the Archaic period fine and superior copprimary functional context in providing a sym- per utilitarian tools were manufactured, wherebolically diverse yet pervasive artifactual envir- as, during Early and Middle Woodland times onment promoting group solidarity and serving copper was used primarily for the production of as a basis for group awareness and identity. This nonutilitarian items (Griffin 1952: 356). I will pan-systemic set of symbols is the milieu of en- explore this interesting situation in terms of: culturation and a basis for the recognition of (1) the frame of reference presented here, social distinctiveness. "One of the main func- (2) generalizations which have previously been tions of the arts as communication is to reinforce made concerning the nature of culture change, belief, custom, and values" (Beals and Hoijer and (3) a set of hypotheses concerning the rela1955: 548). T h e distribution of style types and tionships between certain forms of socio-technic traditions is believed to be largely correlated artifacts and the structure of the social systems with areas of commonality in level of cultural that thev, represent. The normal assumption when thinking about complexity and in mode of adaptation. Changes in the temporal-spatial distribution of style types the copper artifacts typical of the Old Copper are believed to be related to changes in the struc- complex is that they are primarily technomic ture of socio-cultural systems either brought (manufactured for use in directly coping with about through processes of in situ evolution, or the physical environment). It is generally by changes in the cultural environment to which assumed that these tools were superior to their local socio-cultural systems are adapted, thereby functional equivalents in both stone and bone initiating evolutionary change. It is believed because of their durability and presumed superthat stylistic attributes are most fruitfully iority in accomplishing cutting and piercing that within studied when questions of ethnic origin, migra- tasks. It is a common generalization tion, and interaction between groups is the sub- the realm of technology more efficient forms ject of explication. However, when explana- tend to replace less efficient forms. The Old tions are sought, the total adaptive context of Copper case seems to be an exception. Absolute efficiency in performance is only one the socio-cultural system in question must be investigated. In this field of research archaeolo- side of the coin when viewed in an adaptive gists are in an excellent position to make major context. Adaptive efficiency must also be contributions to the general field of anthro- viewed in terms of economy, that is, energy expology, for we can work directly in terms of penditure versus energy conservation (White

BINFORD]

ARCHAEOLOGY AS ANTHROPOLOGY

1959: 54). For one tool to be adaptively more efficient than another there must be either a lowering of energy expenditure per unit of energy of conservation in task performance, or an increase in energy conservation per unit of performance over a constant enerm -, ex~enditure in tool production. Viewed this way, we may question the position that copper tools were technologically more efficient. The production of copper tools utilizing the techniques employed in the manufacture of Old Copper specimens certainly required tremendous expenditures of both time and labor. The sources of copper are not in the areas of most dense Old Copper implements (Wittry 195I ) , hence travel to the sources, or a t least the establishment of logistics networks based on kin ties extending over large areas, was a prerequisite for the procurement of the raw material. Extraction of the copper, using the primitive mining techniques exemplified by the aboriginal mining pits on Isle Royale and the Keewenaw Peninsula (Holmes 1901), required further expenditure of time and labor. Raw materials for the production of the functional equivalents of the copper tools was normally available locally or at least available at some point within the bounds of the normal exploitative cycle. Extraction was essentially a gathering process requiring no specialized techniques, and could be accomplished incidental to the performance of other tasks. Certainly in terms of expenditures of time and energy, as regards the distribution of sources of raw materials and techniques of extraction, copper required a tremendous expenditure as opposed to raw materials of stone and bone. T h e processing phase of tool production appears to present an equally puzzling ratio with regard to expenditure of energy. The processing of copper into a finished artifact normally requires the separation of crystalline impurities from the copper. Following this processing phase, normal procedure seems to have been to pound and partially flatten small bits of copper which were then pounded together to "build" an artifact (Cushing 1894). Once the essential shape had been achieved, further hammering, grinding, and polishing were required. I suggest that this process is more time consuming than shaping and finishing an artifact by chipping flint, or even the pecking and grinding technique employed in the Groduction of ground stone tools. It follows that there was a much greater expenditure of time and energy in the produc-

22 1

tion of copper tools than in the production of their functional equivalents in either bone or stone. Turning now to the problem of energy conservation in task performance, we may ask what differentials existed. It seems fairly certain that copper was probably more durable and could have been utilized for a longer period of time. As far as what differentials existed between copper and stone, as regards cutting and piercing functions, only experiments can determine. Considering all of the evidence, the quality of durability appears to have been the only possible realm which could compensate for the differentials in expenditure of energy between stone and bone as opposed to copper in the area of procurement and processing of the raw material. What evidence exists that would suggest that durability was in fact the compensatory quality which made copper tools technologically more efficient? All the available evidence suggests the contrary interpretation. First, we do not have evidence that the raw material was re-used to any great extent once an artifact was broken or "worn out." If this had been the case, we would expect to have a general lack of battered and worn out" pieces and some examples of reworked pieces, whereas evidence of use is a common characteristic of recovered specimens, and to my knowledge reworked pieces are uncommon if not unknown. Second, when found in a primary archaeological context, copper tools are almost invariably part of burial goods. If durability was the compensatory factor in the efficiency equation, certainly some social mechanism for retaining the copper tools as functioning parts of the technology would have been established. This does not appear to have been the case. Since durability can be ruled out as the compensatory factor, we must conclude that copper tools were not technologically more efficient than their functional equivalents in both stone and bone. Having reached this "conclusion," it remains to explore the problem of the initial appearance of copper tools and to examine the observation that there was a shift from the use of copper for the production of utilitarian tools to nonutilitarian items. It is proposed that the observed shift and the initial appearance of copper tools can best be explained under the hypothesis that they did not function primarily as technomic items. I I<

222

AMERICAN ANTlQUlTY

suggest that in both the O l d Copper and later cultural systems to the south, copper was utilized primarily for the production of socio-technic items. Fried (1960) discusses certain pertinent distinctions between societies with regard to systems of status grading. Societies on a low general level of cultural complexity, measured in terms of functional specialization and structural differentiation, normally have an "egalitarian" system of status grading. The term "egalitarian" signifies that status positions are open to all persons within the limits of certain sex and age classes, who through their individual physical and mental characteristics are capable of greater achievement in coping with the environment. Among societies of greater complexity, status grading may be less egalitarian. Where ranking is the primary mechanism of status grading, status wositions are closed. There are aualifications for attainment that are not simply a function of one's personal physical and mental capabilities. A classic example of ranking is found among societies with a ramage form of social organization (Sahlins 1958: 139-180). In such societies status is determined bv one's proximity in descent from a common ancestor. High status is accorded those in the direct line of descent, calculated in terms of primogeniture, while cadet lines of descent occupy positions of lower status depending on their proximity to the direct line. Another form of internally ranked system is one in which attainment of a particular status position is closed to all except those members of a particular kin group who may occupy a differentiated status position, but open to all members of that kin group on an egalitarian basis. Other forms of status grading are recognized, but for the purposes of this discussion the major distinction between egalitarian and ranked systems is sufficient. I propose that there is a direct relationship between the nature of the system of status grading within a society and the quantity, form, and structure of socio-technic components of its archaeological assemblage. It is proposed that among egalitarian societies status symbols are symbolic of the technological activities for which outstanding performance is rewarded by increased status. In many cases they will be formally technomic items manufactured of "exotic" material or elaborately decor-

-

[ VOL.28, No. 2,1962

ated and/or painstakingly manufactured. I do not imply that the items could not or were not used technomically, simply that their presence in the assemblage is explicable only in reference to the social system. Within such a system the structure of the socio-technic component as regards "contextual" relationships should be simple. Various status symbols will be possessed by nearly all individuals within the limits of age and sex classes, differentiation within such a class being largely quantitative and qualitative rather than by formal exclusion of particular forms to particular status grades. The degree to which socio-technic symbols of status will be utilized within an egalitarian group should largely be a function of group size and the intensity and constancy of personal acquaintance among all individuals composing the society. Where small group size and general lack of interaction with nearby groups is the normal pattern, then the abundance of status symbols should be low. Where group size is large and/or where between-group interactions are widespread, lowering the intimacy and familiarity between interacting individuals, then there should be a greater and more general use of material means of status communication. Another characteristic of the manipulation of status symbols among societies with essentially egalitarian systems of status grading would be the destruction at death of an individual's symbols of status. Status attainment being egalitarian, status symbols would be personalities and could not be inherited as such. Inclusion as grave accompaniments or outright destruction would be the suggested mode of disposal for status items among such groups. Among societies where status grading tends to be of a nonegalitarian type, the status symbols should be more esoteric in form. Their form would normally be dictated by the ideological symbolism which rationalizes and emphasizes the particular internal ranking system or the means of partitioning the society. The structure of the socio-technic component of the assemblage should be more complex, with the complexity increasing directly as the complexity of the internal ranking system. Possession of certain forms may become exclusively restricted to certain status positions. As the degree of complexity in ranking increases there should be a similar increase in the differentiation of contextual associations in the form of differential

B ~ ~ t o ]n v

ARCtIAEOLOGY AS ANTHROPOLOGY

treatment at death, differential access to goods and services evidenced in the formal and spatial differentiation in habitations and storage areas, etc. W e would also expect to observe differentlatlon among the class of status symbols themselves as regards those which were utilized on a custodial basis as opposed to those that were personalities. Similarly, we would expect to see status symbols more frequently inherited at death as inheritance increases as the mechanism of status ascription. Certainly these are suggestions which must be phrased as hypotheses and tested against ethnographic data. Nevertheless it is hoped that this discussion is sufficient to serve as a background against which an explanatory hypothesis concerning the Old Copper materials can be offered as an example of the potential utility of this type of systemic approach to archaeological data. I suggest that the O l d Copper copper tools had their primary functional context as symbols of achieved status in cultural systems with an egalitarian system of status grading. T h e settlement patterns and general level of cultural development suggested by the archaeological remains is commensurate with a band level of socio-cultural integration (Martin, Quimby, and Collier 1947: 299), that level within which egalitarian systems of status grading are dominant (Fried 1960). The technomic form, apparent lack of technomic efficiency, relative scarcity, and frequent occurrence in burials of copper artifacts all suggest that their primary function was as socio-technic items. Having reached this 'Lconclusion," we are then in a position to ask, in systemic terms, questions concerning their period of appearance, disappearance, and the shift to nonutilitarian forms of copper items among later prehistoric socio-cultural systems of eastern North America. I propose that the initial appearance of formally "utilitarian" copper tools in the Great Lakes region is explicable in terms of a major population expansion in the region following the Nipissing stage of the ancestral Great Lakes. The increase in population density was the result of increases in gross productivity following an exploitative shift to aquatic resources during the Nipissing stage. The increased populations are generally demonstrable in terms of the increased number of archaeological sites ascribable to the post-Nipissing period. T h e shift to

223

aquatic resources is demonstrable in the initial appearance of quantities of fish remains in the sites of this period and in the sites of election for occupation, adjacent to prominent loci of availability for exploiting aquatic resources. It is proposed that with the increasing population density, the selective pressures fostering the symbolic communication of status, as opposed to the dependence on personal recognition as the bases for differential role behavior, were sufficient to result in the initial appearance of a new class of socio-technic items, formally technomic status symbols. T h e failure to perpetuate the practice of the manufacture of copper tools on any extensive basis in the Great Lakes region should be explicable in terms of the changing structure of the social systems in that area during Woodland times. The exact type of social structure characteristic of Early Woodland period is at present poorly understood. I would suggest that there was a major structural change between the Late Archaic and Early Woodland periods, probably in the direction of a simple clan and moiety basis for social integration with a corresponding shift in the systems of status grading and the obsolescence of the older material means of status communication. T h e presence of copper tools of essentially nonutilitarian form within such complexes as Adena, Hopewell, and Mississippian are most certainly explicable in terms of their sociotechnic functions within much more complex social systems. Within the latter societies status grading was not purely on an egalitarian basis, and the nonutilitarian copper forms of status symbols would be formally commensurate with the ideological rationalizations for the various ascriptive status systems. This explanatory "theory" has the advantage of "explaining": (1) the period of appearance of copper and probably other "exotic" materials in the Late Archaic period; (2) the form of the copper items; ( 3 ) their frequently noted contextual relations, for example, placement in burials; (4) their disappearance, which would be an "enigma" if they functioned primarily as technomic items; and (5) the use of copper for the almost exclusive production of "nonutilitarian" items in later and certainly more complex cultures of the eastern United States. This explanatory theory is advanced on the basis of currently available information, and regardless

224

AMERICAN ANTIQUITY

of whether or not it can stand as the correct explanation of the "Old Copper Problem" when more data are available, 1 suggest that only within a systemic frame of reference could such an inclusive explanation be offered. Here lies the advantage of the systemic approach. Archaeology must accept a greater responsibility in the furtherance of the aims of anthropology. Until the tremendous quantities of data which the archaeologist controls are used in the solution of problems dealing with cultural evolution or systemic change, we are not only failing to contribute to the furtherance of the aims of anthropology but retarding the accomplishment of these aims. W e as archaeologists have available a wide range of variability and a large sample of cultural systems. Ethnographers are restricted to the small and formally limited extant cultural systems. Archaeologists should be among the best qualified to study and directly test hypotheses concerning the process of evolutionary change, particularly processes of change that are relatively slow, or hypotheses that postulate temporal-processual priorities as regards total cultural systems. The lack of theoretical concern and rather naive attempts at explanation which archaeologists currently advance must be modified. I have suggested certain ways that could be a beginning in this necessary transition to a systemic view of culture, and have set forth a specific argument which hopefully demonstrates the utility of such an approach. The explanatory potential which even this limited and highly specific interpretative approach holds should be clear when problems such as "the spread of an Early Woodland burial cult in the Northeast" (Ritchie 1955), the appearance of the "Buzzard cult" (Waring and Holder 1945) in the Southeast, or the "Hopewell decline" (Griffin 1960) are recalled. It is my opinion that until we as archaeologists begin thinking of our data in terms of total cultural systems, many such prehistoric "enigmas" will remain unexplained. As archaeologists, with the entire span of culture history as our "laboratory," we cannot afford to keep our theoretical heads buried in the sand. W e must shoulder our full share of responsibility within anthropology. Such a change could go far in advancing the field of archaeology specifically, and would certainly advance the general field of anthropology.

[ VOL.28, No. 2,1962

BEALS, RALPHL. AND HARRY H OIJER 1953 An lntroduction to Anthropology. The Macmillan Company, New York. BUETTNER.JANUSCH, JOHN 1957 Boas and Mason: Particularism versus Generalization. American Anthropologist, Vol. 59, No. 2, pp. 318-24. Menasha. CUSHING, F. H. 1894 Primitive Copper Working: An Experimental Study. American Anthropologist, Vol. 7, No. 1, pp. 93-117. Washington. FRIED,MORTON H. 1960 On the Evolution of Social Stratification and the State. In "Culture in History: Essays in Honor of Paul Radin, edited by Stanley Diamond, pp. 713-31. Columbia University Press, New York. GRIFFIN,JAMES B. 1952 Culture Periods in Eastern United States Archaeology. In Archaeology of Eastern United States, edited by James B. Griffin, pp. 352-64. University of Chicago Press, Chicago. 1960 Climatic Change: A Contributory Cause of the Growth and Decline of Northern Hopewellian Culture. Wisconsin Archeologist, Vol. 41, No. 2, pp. 21-33. Milwaukee. HOLMES, WILLIAM H. 1901 Aboriginal Copper Mines of Isle Royale, Lake Superior. American Anthropologist, Vol. 3, No. 4, pp. 684-96. New York. KROEBER, A. L. 1953 Introduction. In: Anthropology Today, edited by A. L. Krixber, pp. xiii-xv. University of Chicago Press, Chicago. AND DONALD COLLIER MARTIN, PAULS., GEORGE I. QUIMBY 1947 lndians Before Columbus. University of Chicago Press, Chicago.

RITCHIE,WILLIAM A. 1955 Recent Suggestions Suggesting an Early Woodland Burial Cult in the Northeast. New York State Museum and Science Service, Circular No. 40. Rochester.

D. SAHLINS, MARSHALL 1958 Social Stratification in Polynesia. University of Washington Press, Seattle. SPAULDING, ALBERT C. 1946 Northeastern Archaeology and General Trends in the Northern Forest Zone. In "Man in Northeastern North America," edited by Frederick Johnson. Papers of the Robert S. Peabody Foundation for Archaeology, Vol. 3, pp. 143-67. Phillips Academy, Andover.

I ~ I N I O]R D

AKCHAEOLOGY AS ANTHROPOLOGY

225

JULIAN H. STEWARD, 1955 Theory of Culture Ch~mge. University of Illinois Press, Urbana.

WILLEY,GORDON R. AND PHILIPPHILLIPS 1958 Method and Theory in Archaeology. University of Chicago Press, Chicago.

THOMPSON, RAYMOND H. 1958 Preface. In "Migrations in New World Culture History," edited by Raymond H. Thompson, pp. v-vii. University of Arizona, Social Science Bulletin, No. 27. Tucson.

W T ~ R YWARREN , L. 1951 A Preliminary Study of the Old Copper Complex. Wisconsin Archeologist, Vol. 32, No. 1, pp. 1-18. Milwaukee.

WAKING, ANTONIO J. AND PRESTON HOLDER 1945 A Prehistoric Ceremonial Complex in the Southeastern United States. American Anthropologist, Vol. 47, No. 1, pp. 1-34. Menasha. WHITE,LESLIEA. 1959 The Evolution of Culture. McGraw-Hill Book Company, New York.

L. AND ROBERTE. RITZENTHALER WITTRY,WARREN 1956 The Old Copper Complex: A n Archaic Manifestation in Wisconsin. American Antiquity, Vol. 21, No. 3, pp. 244-54. Salt Lake City. OF CHICAGO UNIVERSITY

Chicago, Illinois April, 1962

OF

JOURNAL

ANTHROPOLOGICAL RESEARCH VOLUME 37

*

NUMBER

3

*

FALL

*

1981

BEHAVIORAL ARCHAEOLOGYAND THE "POMPEIIPREMISE" Lewis R. Binford Department of Anthropology, University of New Mexico, Albuquerque, NM 87131 It is argued that "behavioral archaeology," as it has been advocated by M.B. Schiffer, represents a series of partially misguided postures relative to some important points of debate common in the 1960s. Due to failures to understand some points at issue, behavioral archaeologists appear to embrace many of the methods and goals of traditional archaeology, and thereby present a "reactionary "position rather than a "revisionist" position, as it has claimed. The issues upon which behavioral archaeology appears most reactionary are points most urgently in need of change, if archaeology is to progress as a discipline contributing both to our knowledge of the past and to our understanding of historical trends.

WALTERTAYLOR'S (1948) CALL for more attention to "context" and "affinities" stimulated a healthy series of discussions concerning "interpretation" in archaeology. In general, however, archaeologists reacted with skepticism to Taylor's slate of interpretive goals. Most considered the achievement of the "conjunctive" approach to be primarily conditioned by the preservation manifest in the archaeological record. Stated another way, most archaeologists of the time were strict inductivists, and considered that one's ability to construct "cultural" contexts was a function of the relative preservation of associations with clear "indicative" properties. In short, the degree to which the archaeological record preserved what were perceived as clear and directly "meaningful" associations and patterned structure was seen as the factor limiting the attainment of WalterTaylor's goals for archaeology. Skepticism born of an inductivist's approachwas voiced by Hawkes (1954) and supported in the writings of many, including Robert Ascher (1962:368), who stated: The foregoing example is designed as a paradigm to illustrate how observations on ongoing cultures might be adapted to the problems of archaeological reconstruction . . . Unlike endeavors which exclusively emphasize those aspects of culture which archaeologists cannot learn about [e.g. Thompson 1939], it directs attention toward the archaeologically possible (emphasis added).

Ascher's reference to Thompson (1939) is to the provocative paper in which he discusses seasonal behavioral changes which he believes would mislead many archaeologists, given then current interpretative conventions: . . . the camps and the house types, the weapons and the utensils are of a specialized type and related to the seasonal life, so that viewing these independently at different periods of the year, and seeing the people engaged in occupations so diverse, an onlooker might be pardoned for concluding that they were different peoples (1939:209). 195

196

JOURNAL OF ANTHROPOLOGICAL RESEARCH

This exchange is typical of the time. A claim is made regarding the nature of living systems, with clear implications for the patterning found in the archaeological record. The archaeologist replies that the cited characteristics are aspects of culture "which archaeologists cannot learn about" (Ascher 1962:368). It was in the context of further elaborating this skeptical point about our ability to reconstruct the past that Robert Ascher began developing a point of great interest -I have in mind his very early and in many ways pioneering ethnoarchaeological studies (Ascher 1962, 1968), andhis discussions of what Schiffer (1972) would later term "formation processes": Every living community is in the process of continuous change with respect to the materials which it utilizes. At any point in its existence some proportion of materials are falling into disuse and decomposing, while new materials are being added as replacement. In a certain sense a part of every community is becoming, but is not yet, archaeological data ... (Ascher 1961:324)

Ascher's arguments regarding formation processes were extremely insightful. He argued that the archaeological record is some combination of interruptions in an entropy-linked process, and that it is therefore limited in its information potential. He cited the disorganizing effect of the operation of past systems, as well as the disorganizing processes associated with subsequent modifications of the archaeological record, as the sources of "distortion" in the archaeological record as seen by an archaeologist. In short, as he put it, what the archaeologist digs up is not "the remains of a once living community stopped, as it were, at a point in time"; such an "erroneous notion, often implicit in archaeological literature, might be called the Pompeii premise" (Ascher 1961:324). Ascher argued rather that the archaeological record is the disorganized arrangement of matter regularly generated after the point in time which would be of interest to a "reconstructionist." Ascher was attempting to force archaeologists to realize that the archaeological record is ravaged by "time's arrow," and should be treated as such, rather than as a preserved "past." In this important point I agreed with Ascher. On the other hand, the implications of this empirical point as it reflects upon the aims and goals of archaeology was a point of contention between Ascher and myself. In his view, the archaeological record, viewed from a strict inductivist position, does not generally contain information regarding "cultural context," "social behavior," and other aspects of interest to those who seek to understand the past as it existed. His reasoning was simply that since the Pompeii premise was erroneous, then most reconstructionist goals, as advocated by Taylor (1948) and others, were equally unrealistic. In his view, we should restrict our discussions to subjects which are archaeologically possible.1 From this perspective, it follows that those who advocate the discussion of "cultural context" are really only advocating speculation and "theorizing": One reason for archaeology's continued lag as a contributor to culture theory ... is that we continue to see our main goal as the reconstruction of ancient ethnology. Since we have so great a handicap to begin with here, where speculation and inference mix too confusingly ... I think we are overlooking another order of interpretation . . . one can manipulate the artifacts statistically without much concern whether one understands precisely what they originally were, exactly how they were used, and just what they meant to the ancients" (Wauchope 1966:19).

"POMPEII PREMISE" IN ARCHAEOLOGY

197

Clearly what was advocated was a "realistic" appraisal of the "limitations of the archaeological record," with the understanding that this would guide us to an appreciation of "realistic" goals for archaeological research. My approach to the situation was different, and took exception to both points of view being discussed; I disagreed with Taylor's reconstructionist goals, as well as with inductivist skepticism regardingour ability to learn about the past. I have never viewed the reconstruction of prehistoric lifeways in the form of prehistoric ethnographies to be an appropriate goal for archaeology in general. It has been clear to me that the time frame of ethnography is largely inappropriate for archaeological research. Rates of deposition are much slower than the rapid sequencing of events which characterizesthe daily lives of living peoples; even under the best of circumstances, the archaeological record represents a massive palimpsest of derivatives from many separate episodes. Any structure and repetitive patterns of association and covariation must derive from the operation of "systemic events," or dynamics, with a much longer term, more rigidly determined organization than is true of those observed in the lives of persons and groups which embody the ethnographers' perception of time and human systems. My view was that we should be seeking to understand cultural systems, in terms of organizational properties, and in turn, to explain differences and similarities among these cultural systems, rather than to generate set pieces of descriptive history. This means, however, that those "things" of interest from the past are organizational properties, which cannot be dug up directly. Stated another way: Granted we cannot excavate a kinship terminology or a philosophy, but we can and do excavate the material items which functioned together with these more behavioral elements within the approriate cultural subsystems (Binford 1962:218-19).

In this view, a strict inductivist approach to inference is clearly impossible, since we need to have an understanding as to how living systems functioned in order to make inferences from their static by-products; in short, we need to link our observations of the archaeological record to an understanding of systems dynamics, in order to make reliable inferences about the past. The way in which I diagnosed the situation regardingTaylor and his critics was that while I did not necessarily agree with Taylor's goals, the relevance of his critics' arguments depended on a strict inductivist approach to inference. From the beginning of my professional career, I have considered the archaeological record to be an ordered consequence of levels of adaptive organization which are difficult to appreciate directly through the observation of events and episodes in the "quick time" perspective of the ethnographer or the participant in a cultural system. Nor have I been surprised or alarmed to find that the archaeological record does not generally carry the intuitively obvious information regarding the "quick time" events and "human episodes" which one would expect from an ethnographer or a preservedPompeii. I have considered it necessary for archaeologists to investigate the archaeological record as a different order of reality, the patterned structure of which represents not a simple accumulation of little events, but rather some of the basic organizational constraints and determinants operating on the events or episodes of daily living. The archaeological record is therefore not a poor or distorted manifestation of ethnographic "reality," but most likely a structured consequence of the

198

RESEARCH JOURNALOF ANTHROPOLOGICAL

operation of a level of organization difficult, if not impossible, for an ethnographer to observe directly. This level of organization is in turn likely to be the unit upon which ecological-evolutionary selection operates, rather than the level of the specific event. Consistent with such a viewpoint, I criticized the idealized view of culture (Binford 1962), and suggested that the meanings of artifacts derive from their function in living systems, not from some manufacturer's "ideas" as translated by our classifications for things that archaeologists have found. More directly, my view was that the "Pompeii premise" is important only if one adopts a strict inductivist approach to the archaeological record, expecting to uncover archaeological "facts" with selfevident meaning for the past. I reasoned that if our methods of inference were so faulty that the "realities" of an organized system as reported by Donald Thompson (1939) had to be considered something that "archaeologists cannot learn about" (Ascher 1962:368), then clearly we needed a change in our methods. In the context of arguing that cultural systems are internally differentiated (Binford 1962), rather than internally homogeneous (the idea threatened by Thompson's descriptions), and that such organizational differences would be manifested in the archaeological record, I wrote (Binford 1964:425) that: The loss, breakage, and abandonment of implements and facilities at different locations, where groups of variable structure performed different tasks, leaves a 'fossil' record of the actual operation of an extinct society.

Some years later (Binford 1968a:23) I addressed quite directly the problem of the "limitations of the archaeological record," and amplified my views on its information potential: The practical limitations on our knowledge of the past are not inherent in the nature of the archaeological record; the limitations lie in our methodological naivete, in our lack of principles determining the relevance of archaeological remains to propositions regarding processes and events of the past.2

The degree of preservation of the archaeological record has never been considered limiting in a practical sense, since it has also been maintained that "facts do not speak for themselves," and that all arguments from the archaeological record as to the character of the past are inferences. These in turn are only as good as our understanding of the linkages between dynamics and statics, the causes in the past and the consequences remaining for us to observe archaeologically. The reader cannot imagine my surprise when Michael Schiffer (1976a) years later announced to the world his discovery that the "new archaeology" had been laboring under a false principle-the Pompeii premise! The early years of the new archaeology witnessed the frequent and unquestioning repetition of major methodological principles. One such principle was enunciated by Binford (1964:425) in perhaps its most explicit form: The loss, breakage, and abandonment of implements and facilities at different locations, where groups of variable structure performed different tasks, leaves a 'fossil' record of the actual operation of an extinct society. As often happens . . . few investigators have noticed that the principle is false. It is false because archaeological remains are not in any sense a fossilized cultural system. Between the time artifacts were manufactured and used in the past and the time these same objects are unearthed by the archaeologist, they have been subjected to a series of cultural and noncultural processes which have transformed them spatially, quantitatively, formally and relationally . . . If we desire to reconstruct the past from archaeological remains, these processes

"POMPEII PREMISE" IN ARCHAEOLOGY

199

must be taken into account, and a more generally applicable methodological principle substituted for the one that asserts that there is an equivalence between a past cultural system and its archaeological record. The principle I offer is that archaeological remains are a distorted reflection of a past behavioral system (Schiffer 1976a:11-12).

This is exactly the point which Ascher had made years earlier; namely, that the archaeological record is a distortion of the Pompeii that he claimed we thought we were finding. Schiffer (1976a: 14) describes his Pompeii premise as follows: If the human participants and all other energy sources completely halt their actions, the activities cease, as does the operation of the behavioral system, what remains (assuming no modification by other processes) is the closest conceivable approximation of a 'fossil' of a cultural system-its material elements in a system-relevant spatial matrix.

Ascher had discredited the Pompeii premise as a means of discrediting the reconstructionist goals of Walter Taylor. I had rejected these same goals as unrealistic and wasteful relative to the processes which were likely to be responsible for the archaeological record, i.e., the long-term structure and evolutionary dynamics of a cultural system. In turn, Ascher's arguments against Taylor were beside the point. Archaeological theory consists of propositions and assumptions regarding the archaeological record itself-its origins, its sources of variability, the determinants of differences and similarities in the formal, spatial, and temporal characteristics of artifacts and features and their interrelationships. It is in the context of this theory that archaeological methods and techniques are developed (Binford 1968b:2).

When Schiffer, years later, claimed that my approaches presuppose the Pompeii premise, my initial response was to ignore his misunderstanding.3 However, I have come to realize that instead of minor misunderstandings, there are substantive points of disagreement between Schiffer and myself. WAS DISTORTION GOING ON AT POMPEII?

Schiffer proposed that in order to grasp fully the inferential challenges facing the archaeologist, we have to understand in some detail the sources of, or causes for, the properties remaining in the archaeological record. In this I agree. He further proposed that we must always keep clearly in mind the difference between systemic context, or the dynamics within which matter "participated" in the past, and archaeological context, or the static form in which we find culturally generated matter in the archaeological record; with this I also agree. According to Schiffer it is the "transformation" of matter from the systemic context into the archaeological context that should receive our focused attention. He called these transformational conditions formation processes, and recognized two basic types: C-transforms, or cultural formation processes, and N-transforms, or noncultural formation processes. The question now is why does Schiffer consider the archaeological record distorted? I am quite comfortable with the idea that there may be natural, or noncultural, processes which condition the character of the archaeological record so as to "distort" the organization as it was generated in the "systemic" context. I am somewhat at a loss, however, to understand why events which modified the formal properties of matter during the operation of a cultural system should be considered distorting. The principle I offer is that archaeological remains are a distorted reflection of a past behavioral system. However, because the cultural and non-cultural processes responsible for these distortions are regular, there are systematic (but seldom direct) relationships between archaeological remains and past cultural systems (Schiffer 1976a: 12; emphasis added).

200

JOURNAL OF ANTHROPOLOGICAL RESEARCH

It is clear from Schiffer's statement that he considers C-transforms potentially distorting. From my perspective, archaeologists must understand formation processes-the dynamics of cultural systems which yield derivative residues both in from Schiffer's perspective, properties of form and spatial distribution-while the distortions between must "systemic" and the "archaeoarchaeologists identify contexts. logical" For instance, Schiffer would view the event of a young boy cleaning out a hearth, taking the ash and other unwanted contents out of the house, and tossing it to one side as a C-transform distorting the juxtaposition of hearth and ashes which obtained during the period of active fire in the hearth. From the perspective of the occupants of the site, it was cleaning up. We may reasonably ask how cleaning up should distort the relationship between the archaeological record and the cultural system from which it derives; some might even argue that cleaning up was essential for the continued use of the location. Similarly, one might argue that adding fuel to the fire was essential to maintaining the fire, but alas, it was also "distorting" the prior relationship between the hearth and the adjacent pile of firewood. Thus, I totally disagree with Schiffer's idea that the archaeological record is a distortion of a past cultural system. Such a position could only be true if the archaeological record as produced in the past were destroyed or modified by postdepositional events. The archaeological record is a normal consequence of the operation of living systems, all of which are dynamic, "flow through" systems, in which energy is captured and its potential reduced. Entropy is the inevitable product of a living system, and it is generated continuously. The archaeological record must therefore be viewed as matter transposed and organized during the process of energy use and entropy production. It is the functional linkages between the organization of a system and its energy-capturing tactics, together with its patterned residues (entropy), that yield information about the organization of past systems. Our inferences from the archaeological record to the past may be wrong or unjustified, not because the archaeological record is a distortion of the past, but because we do not accurately understand the relationship between statics and dynamics. The archaeological record can only be considered a distortion relative to some a priori set of expectations; certainly it is not a distortion of its own reality. It is a faithful remnant of the causal conditions operative in the past, and our task is to understand those causal conditions. Put another way, a pattern or arrangement among artifacts at an archaeological site can only be viewed as distorted if one is not interested in the cultural system as manifest, but rather in some property of a cultural system chosen a priori to receive special inferential attention. Cleaning up is distorting only if it destroys some patterned association in which one might have a research interest. PARADIGM CONFLICT: DESCRIBING PAST CULTURAL SYSTEMS VERSUS RECONSTRUCTING THE PAST

I have argued that the idea of the archaeological record as "distorted" by Ctransforms makes sense only if we imagine a set of past conditions which would be "optimal," or ideal, for us to find. The archaeological record is not distorted with respect to its own reality, but only relative to an archaeologist's preexisting expectations. The latter will of course be derived from a specific paradigm, or what we expect the world to be like (see Binford 1981).

"POMPEII PREMISE" IN ARCHAEOLOGY

201

What would Schiffer like to find? In the answer to this question rests the clue to the really major differences between our views of the archaeologist's challenge. Schiffer wants to find Pompeii. What one pictures is a system of Let us begin by visualizing an ongoing cultural system.... If the human participants and all other energy sources completely halt their actions, action.... the activities cease, as does the operation of the behavioral system, what remains (assuming no modifications by other processes) is the closest conceivable approximation to a 'fossil' of a cultural system . . . (Schiffer 1976a: 12-13)

What Schiffer has described is as close as one can come to the ethnographer's notion of the "ethnographic present"-a synchronic cross-section of all the events or actions transpiring at a given time. It is not a picture of a cultural system, but a slice of "history" in the literal sense of the word. A cultural system is not a summation of all the events, behaviors, or other transactional phenomena going on at a given time, but rather the conditioning organizational framework within which all these events transpire. This organizational framework includes all the places, things, and characteristic social relationships and intellectual conventions in terms of which events are played out. It is true that a cultural system has dynamic properties; it is not true that it is completely dynamic and transitory, as is implied by Schiffer's writings.

Given Schiffer's view of a cultural system, placing trash in a dump is a C-transform; as such, it is a distorting action, marking the transition between systemic context and archaeological context for the items discarded. They were in fact "used" at another place, at a prior time, and as such the presence of the item in the dump is a distortion of such a prior state. But is it also true that it is a distortion of the cultural system? Only if the cultural system is equated with specific actions or the "context of use" for the items in question. However, if one considers the maintenance of life space part of the cultural system, the act of cleaning up and disposal itself as a part of the cultural system, even the presence of a dump as a component of the cultural system, how are things distorted? Answer: they are not distorted. What is distorted is Schiffer's view of a cultural system, and his ideas as to how to understand one. The history of anthropology follows research oriented toward offering some answer to the question of why different "peoples" consistently behave differently; in short, why their behaviors are so frequently variable in the face of seemingly similar stimuli. Early answers to this question cited "race" differences as the causes of behavioral differences; to counter this position anthropologists pointed to "culture" as the cause of patterned, long-lasting behavioral differences among "peoples." In this context there was a conflict between those who saw differences in culture as referrable to differences in individual psychology, which then would feed back to reinforce such populational differences, and those who sought the explanation for different cultures not from the properties of the participants, but from the ecological history of the system itself, seen in interaction with its environment. The psychologically oriented view sees cultural systems as abstractions or generalizations of all the specific behaviors considered "cultural" in character. Change reflected through such a normative view is the result of microfactors, which push and pull the frequencies with which specific behaviors are executed.4 In turn, change is seen

202

JOURNAL OF ANTHROPOLOGICAL RESEARCH

in an almost Lamarkian framework, where "most changes result from attempts to cope with the unanticipated consequences of previous decisions" (Schiffer 1979b:366).

I wish to understand in organizational terms the frameworks within which behaviors were carried out, the structural properties of systems within which the dynamic functioning we see as behavior occurred. In turn, I wish to understand in Darwinian terms the pressures on such systems, which result in their structural change, diversification, and evolution.5 Consistent with this contrast between the systems and the interactionist views, Reid and Schiffer state (1973:2) that: It seems appropriate that archaeologists recognize that culture is knowledge-a system of learned information and the rules for processing and transforming that information into action or behavior. Strictly speaking culture is not behavior.

Here we see another fundamental point of disagreement between processualists and Schiffer and his colleagues. For the latter, not only is a cultural system considered to be behavioral in the literal sense of the word, but culture itself is considered to be mental phenomena-a system of learned information. Culture is then manifest in things and behavior, but is not a material phenomenon itself. This is traditional archaeology. Culture provided a mentalist explanation for variability in behavior and manufactured items for the traditionalist; it is again cast in that role in Schiffer's thought. The only difference is that culture is extended to include more than mental templates for producing things, to include plans for "living," manifested behaviorally and transactionally. Ironically, instead of viewing his ideas as conflicting with those advocated by myself and others in the 1960s, Schiffer saw them contributing to harmony in the field: Although the debates of the past decade have resulted in significant advances in archaeological method and theory, they have also led to a fragmentation of the discipline and to overt and covert hostility between rival sub-paradigms and between traditional and processual archaeologists. Thus, before we treat some of the important issues that contribute to the revisionist version of processual archaeology, we offer a scheme for reintegrating the divergent aims and interests of all archaeologists (Reid and Schiffer 1973:2-3).

The intellectual peacemaking which Reid and Schiffer were engaged in would have constituted an intellectual capitulation of processualist views to major components of the traditionalist position. This is neither compromise nor evaluative scientific judgment; it is simple paradigmatic bias. In addition, there are real contrasts between myself and Schiffer regarding the characteristics of evolution-Darwinian versus Lamarkian-and hence the appropriate types of facts needed to understand cultural change. There are paradigmatic differences regarding the concept of culture, and there are also major contrasts in our views on how to do science: Schiffer is primarily an inductivist, seeking to elevate empirical generalizations to the status of "laws," and worrying about the probabilistic "value" of such propositions. Nevertheless, we agree on the basic need for increasing the power of our inferential strategies, and that this depends on giving priority attention to understanding the archaeological record. Since his major work has been in this latter area, we may justifiably ask if his call for a "revisionist" new archaeology should be taken seriously. Has Schiffer illustrated any methodological gains which would back his call for a "revisionist" perspective?

"POMPEIIPREMISE"IN ARCHAEOLOGY

203

BEHAVIORALARCHAEOLOGYAT THEJOINT SITE In Schiffer's model of site formation there are three modes of deposition, resulting in primary refuse, secondary refuse, and defacto refuse. Matter discarded or dropped in the immediate context of use is primary refuse; secondary refuse is matter which has been moved from its primary location, generally through maintenance or clean-up operations; and defacto refuse consists of items lost or abandoned during the abandonment of a site. By definition, defacto refuse is usable items occurring in the context of "late" or terminal use of the site. We may ask how Schiffer would distinguish defacto refuse, deriving from the abandonment by sedentary residents, from site furniture (see Binford 1979), maintained and regularly used in the residentially abandoned rooms by hunting parties or other logistically organized groups who use the "ruin" as a camping spot. The ruin would, of course, offer shelter and raw materials, and water would necessarily have been nearby, since it was once a residential site. The answer seems to be that Schiffer could not recognize such differences, since he never included the successional alternative in his "paradigm" of possibilities for interpreting pueblo "formation processes." In this regard it is interesting that in studying debris associated with what Schiffer considers to be deposits accumulated at the Joint Site during the main occupation, as contrasted with associations developed during the "abandonment," he finds that "the flow rate of chalcedony increases by a factor of two during the abandonment period" (Schiffer 1976a: 170). Schiffer rightfully seeks an understanding of this interesting observation, by adopting a form of argument from elimination (see Binford 1981:83), proposing four "hypotheses" and seeking to "test" each of them. After rejecting three, he is left with hypothesis number four, namely, that "chalcedony was used for the same tasks, but the rate of task performance increased" (Schiffer 1976a: 173). He then suggests that a biased use of chalcedony is most likely to have arisen in hunting. In evaluating this suggestion Schiffer (1976a: 175) argues that: The ratio of food species to total chipped stone tools should increase appreciably between the main occupation and abandonment periods. In Table 12.6 the ratio of the total number of identifiable bones (of the five most common food species) to the total number of stone tools is presented for the two periods. Quite clearly, the ratio increases by a factor of at least 10 for all species.

Schiffer's (1976a: 177) "historical" scenario is as follows: It seems as though the environmental conditions conducive to subsistence agriculture in the Hay Hollow Valley deteriorated considerably during the second half of the thirteenth century. In response to the stresses created by a high rate of crop failure or by reductions in available agricultural products, the inhabitants of the Joint Site diversified their procurement strategies to include more species of wild foods, and a greater dependence on the wild foods ordinarily consumed. The heavier emphasis on hunting brought with it a greater demand for arrow points and other hunting and butchering tools. Because the increased rate of hunting by groups of males resulted in more trips to the vicinity of Point of the Mountain, chalcedony, a material favored for manufacture of arrow points, was retrieved more frequently . . .

We have here a fine example of a set of untested conventions used as a method of inference. Given the use of a limited set of formation conventions, we are lead to a set of conclusions about the past which, at least in my view, are truly bizarre. We may reasonably ask: (1) If game density was sufficient to support the local group in the immediate area, why did they practice agriculture in the first place? (2) Wild resources are normally increasingly depleted in the immediate area of a

204

JOURNAL OF ANTHROPOLOGICAL RESEARCH

sedentary site as a function of increasing occupational duration. If this was the case here, how were the people able to compensate for "crop failure" at the end of their occupation by shifting to the exploitation of wild foods? (3) In general, the speed of game depletion is inversely related to body size; larger body sized animals are overexploited first, so that there is an increasing trend toward animals of smaller body size during the tenure of a sedentary occupation. We note that there is a relative increase in mule deer relative to antelope between the "main occupation" and the "abandonment phase" assemblages. This proportional shift is also true with respect to the smaller animals; there is an absolute increase in mule deer relative to other animals. I find this pattern almost impossible to imagine as characteristic of the closing phases of a permanent settlement. On the other hand, it is perfectly consistent with the post-abandonment utilization of the ruin by hunting parties operating out of another residential base. Presumably, some years after abandonment, the local flora and fauna would recover, and the site of the old sedentary village would make a perfect camping spot for hunting parties exploiting what may well have been a temporary, localized abundance of game. This abundance would be due to the patterns of recovery within the floral community, providing more forage for animals than with climax vegetation. Such an argument might well account for the "defacto garbage" on the floors (this is site furniture), as well as the presence of quantities of fire-cracked rock on the "floors" (see Schiffer's Table 11.2, 1976a:150) of these "abandonment phase" rooms. It is hard to imagine the use of roasting techniques or stone boiling inside the habitation rooms of an ongoing sedentary site, with the fire-cracked rock not cleaned up and disposed of as secondary refuse. On the other hand, fire-cracked rock as primary refuse in hunting camps is quite common, in my experience. DO WE NEED A "REVISIONIST" BEHAVIORAL ARCHAEOLOGY?

First, if we view Schiffer's work through the same critical eyes he has turned on the work of many of the "new archaeologists," then we would have to conclude that Schiffer, too, has failed because of that pesky problem of the Pompeii premise. He has assumed a direct relationship between his unit of observation (the site) and a unit of systemic context (a single occupation, both ethnically and functionally conceived). The possibility that Schiffer has completely distorted the past rests with his failure to appreciate an important formation process (successional use) as a possible contributor to his "recovery context" materials (see Binford 1972:314-26). Using Schiffer's own criteria, then, we are clearly in need of a "revisionist" behavioral archaeology. I think that the reader would agree that this would be an unwarranted conclusion to draw from Schiffer's quite probable failure, yet this is the type of conclusion which Schiffer draws from his critical evaluation of other archaeologists. The "new archaeology" has primarily involved a claim that we did not know all the sources of variability contributing to the archaeological record. Given this condition, new archaeologists have argued that it is premature to offer an a priori evaluation as to the "limitations" of the archaeological record for yielding up information: There has been as yet no attempt to assess the limitations of the archaeological record for yielding different kinds of information, nor does there seem to be the means of accurately

"POMPEII PREMISE" IN ARCHAEOLOGY

205

determining these limits short of total knowledge of all the systematic relationships which characterize past cultural systems. Thus, present discussions of limitations .. are inappropriate and are based on speculation (Binford 1968a:22).

Schiffer distorted this posture into a claim that we know all the sources, and therefore any failure to recognize, say, a particular "formation process," would be clear evidence of a false principle (the "Pompeii premise") at work. Schiffer's position is very close to sophistry. Pompeii is only an ideal for one interested in events, specific behaviors, and event-centered "history." For Schiffer, who obviously has a reconstructionist goal in mind, Pompeii is the most desirable condition of the archaeological record.6 Sadly, after encountering the archaeological record he became disillusioned; it was not a series of little Pompeiis only in need of "dusting off" in order to yield "ethnographies," or complete pictures of the past. Just because Schiffer was disillusioned because of his own false expectations, it is unfair for him to blame that disillusionment on the new archaeology; such behavior can only result in the silly posturing visible in Schiffer (1978). For instance, in striking out against the "Pompeii premise," Schiffer claims that the new archaeologists believe that a complete cultural system is capable of being "seen" at a single site, and that any site contains meaningful information on any subject (Schiffer and House 1977:250):7 While the new archaeologists can be applauded for their desire to promote . . . verifiability in archaeology, they must be faulted for not going far enough and especially for obscuring the fact that ... not all sites are equal in research potential (Sullivan 1976).

I can only respond with Flannery's words "Leaping lizards, Mister Science!" Although presented as a "straw man," this little example of disillusionment points to a major flaw in even the positive aspects of Schiffer's work, namely, his search for "precise moments in the remote past" (Roe 1980:107). Paradigmatic growth is crucial to the growth of science, providing as it does the concepts which may in turn be transformed into an observational language for the science. If we dismiss most of the archaeological record as distorted, mixed, or disturbed, and seek only those provenience units which appear to represent little capsules of human behavior, we will continue to have an impoverished, unrealistic view of the past. We must seek rather to understand the archaeological record in the state in which it is available to us. In most cases, the greater the apparent disorganization, the more intense the use of the place in the past; it is these disturbances we must understand, instead of seeing them as conditions which render the site "insignificant" and the past unknowable. Disturbed deposits, such as mixed "plow zone" aggregates of tools, are the most common remains we encounter; if we hold out for the very few sites where we may "recognize" undistorted "analytical units," then we will have very few remains from the past with which to work.8 The challenge is how to use the "distorted" stuff, not how to discover the rare and unusual Pompeiis. It is clearly impossible to understand the settlement system of a group of mobile huntergathererswhen you demand, as Schiffer has, "undisturbed deposits made by the same group in different locations in the environment" (Schiffer and House 1977:252). Holding up the Pompeii premise as an ideal insures that the imperfect world of "dirt archaeology" will almost always be a frustrating one. Perhaps herein lies a

206

JOURNAL OF ANTHROPOLOGICAL RESEARCH

clue to Schiffer's recent statements that archaeology is not "prehistory," and that many archaeologists are "not concerned with objects at all and some do not even dig!" (Schiffer 1976b:9). It is particularly crucial to have an adequate and realistic observational language for treating properties of the archaeological record as it is; it is the very act of conceptualizing archaeological observations that conveys meaning to them, and in turn implies conditions in the past. If our conceptual inventory is impoverished, relative to the processes which could be responsible for the objects remaining for us to see, then our views of the past will be correspondingly impoverished and unrealistic. It is perhaps ironic that Schiffer, who for many has played the role of the master critic, taking many to task for having "failed" to appreciate the complexities of "formation processes," is at the same time content to view so much processual evidence as distortions.9 Not only is the Pompeii premise rarely justified by the and record, but seeking a reconstructed archaeological Pompeii is an unrealistic we have and the data available to us unprofitable goal in the light of the knowledge in that record. NOTES 1. See Rouse (1964:465) and Wauchope (1966:19) for strong statements supporting Ascher's position on this point. 2. In the minds of many, my claims that we could learn about conditions in the past which were not "directly" recoverable seemed to be a continuation of Taylor's position, and many authors have offered skeptical arguments in the form of "cautionary tales," largely drawn from ethnography, to demonstrate how limited archaeologists' data actually are! Richard Lee's (1966) "lesson" on how limited archaeological remains are when compared to the "richness" of the actual life that left them was amplified in the writings of Carl Heider, who demonstrated how "spurious" associations between things could be generated, and how the dynamics of an actual system would "tend to mislead the archaeologist" (Heider 1967:57). These "cautionary tales" generally assumed a strict inductivist approach to inference, and were usually offered as a skeptical note on some of the goals of archaeological inference, as they were discussed during the 1960s. They were all arguments against a "direct" linkage between the facts of the archaeological record and the past, challenges to the "Pompeii premise." 3. After listening to a major statement of the position presented by M. Collins (1975) in a symposium in San Francisco in 1973, and later championed by Schiffer, I replied as follows (Binford 1975:252):

Raising objections that the archaeological record may be "biased" seems to stem from some strange expectation that the archaeological record is some kind of fossilized picture remaining to us from the past with built-in and self-evident meanings upon which 'time's arrow' has played disturbing tricks. No such meaningful picture exists complete with semantic directions as to how to properly read the record. All meaning comes from us. ... telling me that the archaeological record cannot be translated mechanically into statements about the past according to some set of rules for assigning meaning does not surprise or depress me ... 4. Such a perspective is certainly not unique to Schiffer; (see Jochim 1979:83; Spiess 1979: 1-5). 5. The very practical consequences of these contrastive points of view are well illustrated by comparing Schiffer 1979a with Binford 1979 for vastly different approaches to lithic studies. Similarly, Schiffer's (1976a:56-57) discussion of curation should be compared to my discussion (Binford 1976) for a graphic example of the differences between a search for systems properties and a concern with specific behaviors and their cumulative result. 6. See Dunnell 1980 for his opinion on "reconstructionism. '" 7. This is also true of other recent critics,

"POMPEII PREMISE" IN ARCHAEOLOGY who have suggested that I believe that the archaeological record contains information about all past events (Coe 1978), or that I consider there to be an isomorphism between past behavior and the archaeological record (DeBoer and Lathrap 1979:103, 134). Similarly, see Gould's (1980:6-28) "revelation" that sites are multicomponential and behaviorally stark. Cautionary tales continue to demonstrate, again and again, the fallacy of the Pompeii premise.

207

8. Perhaps the popularity of Schiffer's position with Cultural Resource Management archaeologists arises from its use as a basis for writing off most of the archaeological record as "insignificant," since it does not approach the idea of little prepackaged Pompeiis. This approach renders it impossible even to begin to see something of the organization of past cultural systems. 9. See, for example, Schiffer 1975 and Sullivan and Schiffer 1978.

REFERENCES CITED Ascher, R., 1961, Analogy in Archaeological Interpretation. Southwestern Journal of Anthropology 17:317-25. Ascher, R., 1962, Ethnography for Archaeology: A Case from the Seri Indians. Ethnology 1:360-69. Ascher, R., 1968, Time's Arrow and the Archaeology of a Contemporary Community. Pp. 43-52 in Settlement Archaeology (ed. by K.C. Chang). Palo Alto: National Press Books. Binford, L.R., 1962, Archaeology as Anthropology. American Antiquity 28:217-25. Binford, L.R., 1964, A Consideration of Research Design. American Archaeological Antiquity 29:425-41. Binford, L.R., 1968a, Archaeological Perspectives. Pp. 5-32 in New Perspectives in Archaeology (ed. by S.R. Binford and L.R. Binford). Chicago: Aldine. Binford, L.R., 1968b, Archaeological Theory and Method. Pp. 1-3 in New Perspectives in Archaeology (ed. by S.R. Binford and L.R. Binford). Chicago: Aldine. Binford, L.R., 1972, An Archaeological Perspective. New York: Seminar Press. Binford, L.R., 1975, Sampling, Judgment, and the Archaeological Record. Pp. 251-57 in Sampling in Archaeology (ed. by J.W. Mueller). Tucson: University of Arizona Press. Binford, L.R., 1976, Forty-seven Trips: A Case Study in the Character of Some Formation Processes. Pp. 299-351 in Contributions to Anthropology, the Interior Peoples of Northern Alaska (ed. by E.S. Hall, Jr.). National Museum of Man, Mercury Series, Paper No. 49. Ottawa: National Museum of Canada. Binford, L.R., 1979, Organization and Formation Processes: Looking at Curated Technologies. Journal of Anthropological Research 35:255-73.

Binford, L.R., 1981, Bones: Ancient Men and Modern Myths. New York: Academic Press. Coe, M.D., 1978, The Churches on the Green: A Cautionary Tale. Pp. 75-85 in Archaeological Essays in Honor of Irving B. Rouse (ed. by R.C. Dunnell and E.S. Hall). The Hague: Mouton. Collins, M.B., 1975, Sources of Bias in Processual Data: An Appraisal. Pp. 26-32 in Sampling in Archaeology (ed. by J.W. Mueller). Tucson: University of Arizona Press. DeBoer, W.R. and D.W. Lathrap, 1979, The Making and Breaking of Shipibo-Conibo Ceramics. Pp. 102-38 in Ethnoarchaeology: Implications of Ethnography for Archaeology (ed. by C. Kramer). New York: Columbia University Press. Dunnell, R.C., 1980, Americanist Archaeology: The 1979 Contribution. American Journal of Archaeology 84:463-78. Gould, R.A., 1980, Living Archaeology. Cambridge: Cambridge University Press. Hawkes, C., 1954, Archaeological Theory and Method: Some Suggestions from the Old World. American Anthropologist 56:155-68. Heider, K.G., 1967, Archaeological Assumptions and Ethnographical Facts: A Cautionary Tale from New Guinea. Southwestern Journal of Anthropology 23:55-64. Jochim, M.A., 1979, Breaking Down the System: Recent Ecological Approaches in Archaeology. Pp. 77-117 in Advances in Archaeological Method and Theory, vol. 2 (ed. by M.B. Schiffer). New York: Academic Press. Lee, R.B., 1966, Kalahari-1: A Site Report. The Study of Man, Anthropology Curriculum Study Project, Chicago. Reid, JJ. and M.B. Schiffer, 1973, Prospects for a Behavioral Archaeology. Paper presented at the 1973 meeting of the American

208

JOURNAL OF ANTHROPOLOGICAL RESEARCH

Anthropological Association, New Orleans. Roe, D., 1980, Introduction: Precise Moments in Remote Time. World Archaeology 12:107-8. Rouse, I., 1964, Archaeological Approaches to Cultural Evolution. Pp. 455-68 in Explorations in Cultural Anthropology (ed. by W. Goodenough). New York: McGraw-Hill. Schiffer, M.B., 1972, Archaeological Context and Systemic Context. American Antiquity 37:156-65. Schiffer, M.B., 1975, An Alternative to Morse's Dalton Settlement Pattern Hypothesis. Plains Anthropologist 20:253-66. Schiffer, M.B., 1976a, Behavioral Archaeology. New York: Academic Press. Schiffer, M.B., 1976b, What is Archaeology? South Carolina Antiquities 8(2):9-14. Schiffer, M.B., 1978, Some Issues in the Philosophy of Archaeology. Paper presented at the State University of New York, Binghamton. Schiffer, M.B., 1979a, A Preliminary Consideration of Behavioral Change. Pp. 353-68 in Transformations: Mathematical Approaches to Culture Change (ed. by C. Renfrew and K.L. Cooke). New York: Academic Press. Schiffer, M.B., 1979b, The Place of Lithic Use-wear Studies in Behavioral Archaeology. Pp. 15-25 in Lithic Use-wear Analysis (ed. by B. Hayden). New York: Academic Press.

Schiffer, M.B. and J.H. House, 1977, An Approach to Assessing Scientific Significance. Pp. 249-57 in Conservation Archaeology: A Guide for Cultural Resource Management Studies (ed. by M.B. Schiffer and G.J. Gumerman). New York: Academic Press. Spiess, A.E., 1979, Reindeer and Caribou Hunters: An Archaeological Study. New York: Academic Press. Sullivan, A.P., 1976, The Structure of Archaeological Inference. Manuscript on file, University of Arizona, Tucson. Sullivan, A.P. and M.B. Schiffer, 1978, A Critical Examination of SARG. Pp. 168-75 in Investigations of the Southwestern Anthropological Research Group, Proceedings of the 1976 Conference (ed. by R.C. Euler and GJ. Gumerman). Flagstaff: Museum of Northern Arizona. Taylor, W.W., 1948, A Study of Archaeology. Memoirs of the American Anthropological Association, No. 69. Thompson, D.F., 1939, The Seasonal Factor in Human Culture. Proceedings of the Prehistoric Society 10:209-21. Wauchope, R., 1966, Archaeological Survey of Northern Georgia, with a Test of Some Cultural Hypotheses. American Antiquity 31 (5), Part 2, Memoir 21.

AN

JOURNAL OF RESEARCH THROPOLOGICAL

VOLUME 38

*

*

NUMBER 2

SUMMER

*

1982

PARADIGMS,SYSTEMATICS,AND ARCHAEOLOGY Lewis R. Binford and Jeremy A. Sabloff Department of Anthropology, University of New Mexico, Albuquerque, NM 87131 The ways archaeologists view the past-their paradigms-directly influence their interpretations of the archaeological record. Paradigm change need not be irrational or undirected; such change can best be accomplished by focusing attention on the various ways that dynamic cultural processes can be linked with the static archaeological record.

THE ASSUMPTION THAT SCIENTISTS are capable of clearing their minds and achieving total objectivity is basic to the strict empiricist ideas which dominated early science.1 Boasian anthropology in the United States, and most of the intellectual background of archaeology in general, was founded on this belief in strict empiricism and in the old, Baconian idea of "psychological" objectivity: one simply cleared one's mind of bias and allowed the great truths of nature to be uncovered through the vehicle of one's "bias-free" mind. Such were the views dominating the "great days" when "discoveries" were considered to be the products of a science whose mission was to accumulate and inventory "natural facts." But starting in the late nineteenth/early twentieth century, logical positivists challenged the idea that an "objective" observer is capable of seeing nature directly and accurately. Instead, these philosophers distinguished between the process of discovery (gathering facts and generating theories) and the process of evaluation (of theories used to account for these "facts"). They therefore sought to develop methods largely dependent upon deductive forms of reasoning for evaluating ideas that were already in existence. They also recommended that scientists predict causal occurrences (i.e., if A, then by necessity B) and then test these predictions against the facts. Anthropology, and more particularly archaeology, was relatively slow in absorbing the implications of logical-positivist thinking. In the 1960s, many archaeologists (often labeled "new archaeologists") reacted to the strict empiricism of their colleagues by suggesting that the field at least catch up with the changes in other sciences, i.e., acknowledge the fallacy of being able to achieve "psychological objectivity." New archaeologists argued that the discipline should adopt logicalpositivist methods or develop ones of its own (see Binford 1972; Spaulding 1968). At about the same time that these suggestions for updating archaeological epistomology were being made, some interesting ideas were percolating in the philosophy 137

138

JOURNAL OF ANTHROPOLOGICAL RESEARCH

of science. Perhaps the most influential of these was the thesis of Thomas Kuhn (1962a, 1962b) that one should distinguish between paradigm and theory. A paradigm is one's world view, translating one's experience into meaningful statements. A theory is an argument one makes about why the world is patterned in a particular form. Through arguments involving necessary linkages between classes of data, theory makes prediction possible. Theory is thus derived from one's paradigm: it is a subset of it. Kuhn's argument amounts to a very explicit recognition of the position that meaning comes from humans, that conceptualizations of experience are our inventions, and that nature does not dictate the meanings we assign to it. It recognizes further that when we seek to explain nature through theories, we are seeking to explain our conceptualization of nature, rather than some objective, "true" nature such as Sir Francis Bacon thought himself capable of seeing directly. Kuhn's thesis clearly was significant. He implied that while the logical positivists had recognized that the source of ideas was subjective, they had sought to evaluate these admittedly subjective inventions by referral back to experience, through the use of "objective" observational methods. Kuhn's thesis suggested that this was a fallacy. Objectivity is not attainable either inductively or deductively. Rather, one's observational means for conceptualizing experience are rooted in one's paradigm. The testing of theories was thus an illusion, ultimately bound by paradigmatic subjectivity. Falsification and theory testing as advocated by the logical positivists were thus mere puzzle solving, and were not acknowledged as capable of advancing the paradigm. It is our reading of Kuhn's argument, as well as that of many other critics, that he viewed paradigmatic change as resulting from the irrational interaction of one's domain of thought with other domains. He suggested that such change was responsive to "historical" trends in the wider intellectual domains of society. Science thus did not grow through rational progress but through the chance interaction between investigatory games played out inside a paradigm-bound discipline and "noisy" intellectual conditions outside the discipline, which tended to modify the paradigm under scrutiny (see, for example, Kuhn 1962b:77-78, 1970:208).2 What is interesting in these developments is that the very paradigmatic distinctions which Kuhn so insightfully introduced were ignored by many who have accepted or elaborated his arguments (see, for example, Feyerabend 1975; Toulmin 1972). These distinctions were (1) his conceptual separation of paradigms from theory, (2) his recognition that science may grow by changes in either, and (3) his assertion that logical positivism had not considered paradigmatic change. Having insisted that paradigms are an intellectual domain capable of being unaffected by debate about theory and that many different theories might be offered within the context of a single paradigm, Kuhn stated that paradigm change is not subject to development by methods of rational change. He linked this idea of "irrationality" to the very conceptual recognition of paradigms, and the argument for existential validity of the concept was then cited as rationale for a belief in the associated theory of irrational paradigm change. Such a linkage does not appear to be necessary at all. The challenge to science is to address directly the problem of developing methodological aids to paradigm change and evaluation, as well as the continued perfection of such aids for the evaluation and production of theories. In short, we may

PARADIGMS, SYSTEMATICS, AND ARCHAEOLOGY

139

accept Kuhnian insights regarding the importance of paradigms and their impact on our ideas of objectivity without accepting his particular approach to paradigm change. We may profitably explore alternative theories of change and seek to build into scientific procedure methods for encouraging rational paradigm growth. We believe that archaeologists today are in an excellent position to show how such rational paradigm growth can be achieved. The archaeological literature of the past two decades is replete with arguments which point out that if a major goal of the discipline is to explain culture change, then the traditional ways of looking at the past-the normative paradigms-have not been very productive. A new way of looking at the world, the systems paradigm (see Binford 1962) has been proposed as a potentially more productive means of reaching this explanatory goal. The question is, how should the discipline proceed in developing this (or any other) paradigm? This paper seeks to address the issue both as a first step in the direction of developing rational methods that will foster productive paradigm change and as a means of heightening archaeologists' awareness as to how difficult the challenge of science can be. As the senior author (Binford 1977:3) notes, Science is a method or procedure that directly addresses itself to the evaluation of cultural forms. That is, if we view culture as at least referring to the particularly human ability to give meaning expediently to experience, to symbol, and in turn, view experience through this conceptual idiom, science is then concerned with evaluating the utility of the cultural tools produced. TWO VIEWS OF "CULTURE" Cultural "Cohesion" in Continuity: A New WorldParadigm

Archaeologists are often unaware of how their traditionally held paradigms influence their views of the past. Two such paradigms are presented here to illustrate their effect on archaeological systematics and interpretations. A view which developed largely in North America focuses on the continuity of culture (as manifest in material objects) over wide geographical areas. The perspective stems largely from early empirical studies of North American Indian culture as developed by Wissler (1914) and elaborated by Kroeber (1939). On the basis of comparative study of "culture traits" (particularly material aspects of culture) among certain American Indian groups, Wissler (1914:468) noticed that "while many have called attention to the intergradations of culture, few, for example, have considered the significance of the rarity of abrupt breaks in its continuity." He then pointed out that in known cases where identified "ethnic" groups, such as Cheyenne, Plains Cree, etc., moved from one region to another during historic times, they quickly assimilated to the "culture type" characteristic of the area into which they moved. In turn, they "lost" the culture traits not shared with the groups in their new setting. Wissler (1914:469;emphasis added) concluded that: What evidence we have seems to indicate that by separating a tribe from a center [of a culture area] its material culture is made intermediate [between the form of its parent 'center' and the 'center' toward which it moved] and by joining a tribe to a center its culture is made typical. Hence, unless we find data to support the wholesale movement of a material culture center, we must assume stability of habitat during its historic life. We need not, however, assume stability as to its political, linguistic, and somatic unit constituents ... We have been long familiar with the lack of correlation between culture, language, and somatic type.

140

JOURNAL OF ANTHROPOLOGICAL RESEARCH

Throughout the writings of the founders of American anthropology, we find repeated emphasis that the "essential" feature of culture is that culture traits can be exchanged independently of race, language, or socio-political identity. As Wissler (1914:490) indicated, the one thing culture does not have is "cohesions" among units over time: "Tribal individuality appears only in decoration and a few inessential features, but even so is rarely restricted to a single tribe and tends toward a geographical rather than a random grouping." Basically, Wissler's view (which was shared with many other American anthropologists) was that continuities exist across geographical regions and through time as long as the environment stays roughly the same. Traits and complexes are free to move among the social, linguistic, racial, or other types of groups. Any cohesion, or tendency toward a stable association among elements of material culture, is a phenomenon of large geographical regions and relatively long life spans; not a phenomenon of social or ethnic units. The implication was that cultural dynamics must be understood in terms more comprehensive than simple interaction among persons or social units. Although this view was modified in North America and ultimately reduced to a "psychological" point of view, presented in terms of learning theory and ideas of "historical causation," most archaeologists, who clearly recognized that they were studying "material culture," adopted the view that culture must be described in terms of continuities. It is no accident that the development of observational "tools" by archaeologists, and the growth of archaeological systematics in general, was guided by the above ideas. New World systematics was generated by archaeologists using units (types) which were considered equivalent to the "culture traits" of their ethnographic colleagues. Summing up what was most certainly a view held by a majority of his colleagues, Krieger (1944:272) states that: "The purpose of a type in archaeology must be to provide an organizational tool which will enable the investigator to group specimens into bodies which have demonstrable historical meaning in terms of behavior patterns." He then operationalizes a means of achieving these "bodies": after describing two steps in the sorting of material using judgments of similarity based on morphological criteria, he then discusses the next, crucial step (1944: 280-81): The third step is a process of recombining the groups obtained in the second step, through the study of their comparative distributions. .... Only in this way can it be determined that certain characters are variations of single underlying plans, while others which do not fall together consistently are not variations but culturally distinct ideas. Those details which do consistently combine through site after site, in the same temporal horizon and in the same culture complex, may thus be safely regrouped into tentative types. These differ from all other so-called types in that the cohesiveness of their elements has been proved through the use of archaeological data rather than simply supposed through a variety of assumptions.

For New World analysts, a "culture trait" was something that had "cohesion" among its particular properties (attributes). This cohesion was demonstrable only if the pattern held across a number of spatially distinct cases (sites). In this sense, the criterion of cohesion was used to define a culture trait qualitatively, which could only be demonstrated by a pattern of continuity or repetitive association of properties across a number of cases. Culture itself was seen to represent a "cohesion" of demonstrated culture traits, considered to exist at a level of organization transcending

PARADIGMS, SYSTEMATICS, AND ARCHAEOLOGY

141

the level represented by the social, ethnic, racial, or linguistic "identity" of the participants. Cultural "Cohesion" in Association: An Old WorldParadigm

As far as we can determine, the modem concept of culture commonly used by many Europeans was popularized by V.G. Childe. For example, he (1929:vi) noted that: We find certain types of remains-pots, implements, ornaments, burial sites, house formsconstantly recurring together. Such a complex of regularly associated traits we shall term a 'cultural group' or just a 'culture'. We assume that such a complex is the material expression of what would today be called a 'people'.

Childe's idea of culture appears to have been adopted from the Germanic literature, in particular from the works of Kossinna, who used it to express strong racist sentiments (see Trigger 1980). According to Kossinna, differential cultural achievements should be understood in terms of the degree that societies are kept racially "pure." Childe rejected the linkage of social success (judged in terms of cultural "achievements") with racial purity. He did not, however, reject the linkage between society and culture. For Childe, cultures were the material expressions of particular "peoples." Childe's view, as well as that of many others of the time, represents a poorly analyzed blend of holistic ideas then current in European circles.3 The idea that a "society" is the basic unit of "cultural variability" is perhaps best illustrated by one of the most influential writers who espoused a holistic viewpoint (Durkheim 1938 [1895] :119-20): If we represent historic evolution as impelled by a sort of vital urge which pushes men forward, since a propelling tendency can have but one goal, there can be only one point of reference with relation to which the usefulness or harmfulness of social phenomena is calculated. Consequently, there can, and does, exist only one type of social organization that fits humanity perfectly; and the different historical societies are only approximations of this single model.

He goes on to say that this view is unacceptable for a number of reasons. He then proceeds to develop his argument (1938 [1895] :120-21): If . . . the fitness or unfitness of institutions can only be established in connection with a given milieu, since these milieus are diverse, there is a diversity of points of reference and hence types which, while being qualitatively distinct from one another, are all equally grounded in the nature of social milieus . . . the constitution of the social milieu results from the mode of composition of the social aggregates . . . the considerations just stated lead us back to the idea that the causes of social phenomena are internal to society.

Durkheim's approach to the understanding of "milieus" was based on his view that such social "essences" are formed by coercive forces operating within societies. He states (1938 [1895] :103-4; emphasis added) that: Society is not a mere sum of individuals. Rather, the system formed by their association represents a specific reality which has its own characteristics. Of course, nothing collective can be produced if individual consciousnesses are not assumed; but this necessary condition is by itself insufficient. These consciousnesses must be combined in a certain way; social life results from this combination and is, consequently, explained by it. Individual minds, forming groups by mingling and fusing, give birth to a being, psychological if you will, but constituting a psychic individuality of a new sort. It is, then, in the nature of this collective individuality, not in that of the associated units, that we must seek the immediate and determining causes of the facts appearing therein.

Such views provide archaeologists with a particularly defined reality to monitor along with their classifications. The seat of both causes and perpetuation of cultural

142

JOURNAL OF ANTHROPOLOGICAL RESEARCH

distinctiveness is the internal, "collective" characteristic of each society. It is this inner "milieu" which serves to differentiate and perpetuate societies. In short, any demonstrable cohesion among parts must derive from the operation of internal social factors-the collective. In many places in the Old World, and particularly in France, this view was linked with ideas expressed frequently by the German word Volk and the French esprit. Both have the suggestion of vitalism, where there was something driving the "esprit de corps." One's public spirit, or devotion to one's society, was seen as driven by an inner vitalism, somehow "natural to man." When this vitalism breaks down, "civilization" begins to flounder. In such a view, "cultures" were seen as differentiated packages, isomorphic with ethnically or nationally differentiated peoples. Blendings and mixings represented the breakdowns of one's very "spirit" and hence a kind of degeneracy similar to that which racists considered to result from "race mixing." As Durkheim (1938 [1895] :124) states, "The principle we have just expounded would . . . create a sociology which sees in the spirit of discipline the essential condition of all common life." This emphasis on spirit and the maintenance of distinctiveness through "discipline" is consistent with the frequently expressed view that some measure of a people's worth could be seen in the degree to which external factors could not intrude or impinge upon the cultural expressions of their distinctiveness. Such a point is emphasized by Sonneville-Bordes (1975). Under this view of culture, continuity should be demonstrable among the materials left at different archaeological sites by representatives of the same "people." In a similar manner we should expect conservative patterns of formal change, and we should not expect blended or mixed "traditions," since each people manifests its essential characteristics, its "spirit," in its products. This view is well stated by Bordes (1968:144): "Man is more ready to exchange his genes than his customs, as the whole history of Europe demonstrates." In sum, culture is expressed in terms of formal distinctiveness of artifacts and cohesion, seen in the repetition of a distinctive formal pattern at different sites. When this "ethnic" view of culture was popularized in the United States it was characterized as follows (Rouse 1965:6): All components that have yielded similar assemblages are grouped together. Each group is defined by listing its distinctive traits and is given the name of a typical site .... The name applies not only to the groups of components but also to the traits which characterize it and to the people who lived in the components. The traits constitute a complex which is indicative of the people. Whenever one discovers a new site one can identify the people who lived there simply by determining which complex it contains.

The ethnic view of culture was operationalized by the archaeologist Francois Bordes. His methods of classification consist of a type list, or a set of categories, generated essentially as a "paradigmatic classification" (see Dunnell 1971). For example, he classified Paleolithic tools as a set of combinations and permutations of forms of working edge and the placement of such forms relative to the axis of percussion used in detaching the flake on which the tool was produced. The basic unit of observation is the assemblage, defined in terms of the principle of association; that is, all those tools found together within a recognizable unit of deposition within a site. The units of observation can then be described by tabulating the items in the assemblage relative to the type list. This quantitative pattern was commonly presented in the form of a cumulative graph. Thus, a culture was represented as a

PARADIGMS, SYSTEMATICS, AND ARCHAEOLOGY

143

unit of "cohesion" by the repetitive pattern of similar relative frequencies occurring among assemblages from different places. Cohesion was measured by formal similarity, described in quantative terms. Mixed Paradigms

As one might imagine, there has been some mixing and blending of the Old and New World points of view. There appear to have been two types of mixing or blended views and intellectually blended views. blending-operationally blended views. "La methode Bordes" was developed in its pure Operationally form to treat lithic assemblages from very early time ranges. In general, lithic industries are all that the very early sites yield, except for occasional animal bones. Analysts are fairly comfortable adding, subtracting, and calculating percentages among an array of things scored according to Bordes's type list. However, all know we should not add apples and goats. As long as we have only lithics, the definition of an assemblage in quantitative terms is acceptable. On the other hand, when in later time periods we recover worked bones, art work, ceramics, ground stone, burials in substantial numbers, etc., the definition of an "assemblage" in quantitative terms appears to violate our sensibilities for dimensional mixing. How can we interpret the statement that an assemblage includes 13 percent laterally retouched scrapers, 14 percent extended burials, and 9 percent interior red-slipped pottery sherds? For the more "domains" of things recovered, the more we develop separate, self-contained classificatory schemes to deal with each independently; we then usually define "cultures" in terms of the particular mix of "types" generated independently in each of our separate classificatory schemes. The approach thus begins to appear like the tactics of New World, or Kriegerian, methodology, discussed above. In short, we may begin with an Old World paradigm but be forced to use a New World methodology because a simple quantitative summarization cannot treat all the things found in association within a single paradigmatic classification. Such a development seems to be the type of "mixed" approach which many Old World archaeologists employ when dealing with materials more complicated than the simple lithic industries of the early time ranges. As Childe (1929: 9-10) points out: Societies are represented,not by their members' skeletons, but by . . . pots and houseplans, personal ornaments,and burial sites, the materialsthey fetched from afar.... Such remainsarchaeologistsdivide and classify into types, and when the same types are repeatedly found together at different sites within a limited region they are groupedtogether to represent what we term cultures . . . types are repeatedly found togetherjust because the traditions they embody are approvedand transmittedby a society of persons. ... In this sense archaeologists''cultures'do really stand for societies. It should be clear that Childe's statement is consistent with Kriegerian method, but what a cohesion of types is said to represent is very different from what a typical New World archaeologist, familiar with the extensive distribution studies of culture traits, would conclude. Such studies fail to show any "cellular" distribution of culture traits corresponding to the social boundaries between separate societies (see Childe 1940; for more recent confirmatory studies of this problem, see Klimek 1935; Milke 1949; Clarke 1968; Hodder 1977). All these studies confirm that while some traits may tend to be distributed in terms of social units, configurations of traits representing demonstrable "cohesions" tend to exhibit more regionally extensive distributions, confirming Wissler's and Kroeber's earlier findings.

144

JOURNAL OF ANTHROPOLOGICAL RESEARCH

In this case of operationally blended views, we see researchers starting from a paradigm which assumes the perspective of an internal participant (who sees the distinctiveness of "peoples" arising from the "essential" distinctiveness of their social identities). The researchers are forced by the logic of treating different things not easily included in a single quantitative frame of reference to use methods developed from an almost diametrically opposed set of paradigmatic expectations as to what the world of culture is all about. We can expect that when these conditions hold, the cultures developed by such workers for time periods yielding increasingly diverse archaeological remains will begin to take on the character of "New World" cultures, in direct proportion to the numbers of different data sets recognized by the analysts. Intellectually blended views. Unlike the situation above, where incompatibility derives from the limited relevance of "la methode Bordes" in treating complex archaeological remains, intellectually blended views derive from recombinations of elements in the reasoning used by archaeologists. It is not uncommon to hear archaeologists from some area of North America talking about archaeological materials, described by largely Kriegerian methods, in terms of ethnic or social dynamics. We can only assume that these archaeologists are ignorant of the many empirical studies which repeatedly illustrate the point that there is no equivalence between culture, conceived in terms of cohesions among traits, and specific ethnically or politically defined units. Perhaps the ease with which the "ethnic identity" view of culture is adopted by many archaeologists simply reflects the fact that most archaeologists are products of complex systems. Few have had much direct experience with small-scale systems, and as a growing number of scholars' knowledge of general anthropology becomes smaller, they probably do not know that among small-scale societies, at least, culture as an expression of "one's identity" is a viewpoint which is very hard to defend. PROJECTING VIEWS OF CULTURES INTO THE PAST: OUR OBSERVATIONAL LANGUAGES

The New World and Old World views of the world are paradigms. They summarize expectations as to what "culture" is like. Comparison of these two paradigms should illustrate just how insightful philosophers have been when they argue that our world view, or paradigm, conditions our observation and description of experience. But a paradigm also directly conditions the classificatory procedures which archaeologists have designed to measure culture. The first, or New World, viewpoint is rooted in empirical generalizations regarding the nature of culture which early ethnographers generated from their comparative study of culture traits (largely material objects) across the named social groups of the American Indian. The second, or Old World, viewpoint is rooted in a less systematically studied, but just as empirically based understanding of European history. As we have seen, the perspective of the New World paradigm is one of outside observers looking at variability across socially organized groups of people. In contrast, the Old World paradigm takes the viewpoint of inside observers looking at themselves, relative to the social world of their experiences. Advocates of this Old World paradigm compare other societies from this egocentric point of view.4

PARADIGMS, SYSTEMATICS, AND ARCHAEOLOGY

145

These two paradigms are conscious attempts to describe the world of cultural phenomena accurately. Yet they are generalized from very different data bases and described from different observer perspectives; one from relatively small-scale,lowenergy societies; the other from complex, power-based societies. It is not surprising, then, that the "reality" each view projects is different. Given that persons "seeing" the world through each of these paradigmatic "eyes" have developed different conventions for converting observations made on the archaeological record into descriptive statements about past cultures, and that past cultures have been treated like ethnographically known cultures, it is not surprising that classificatory techniques and methods have been designed to yield cultural information in terms of what the archaeologists thought culture was like. It also should not be surprising that the picture of the past generated by "Krieger's methods" is very different from the view of the past generated by "Bordes's methods." The Kriegerian,or New World, past is a picture of continuity within regions and through limited time spans, a picture of geographical continuities in culture but with temporal punctuations, or lacks of continuity, punctuating regional sequences. It should be kept in mind that "types" are primarily recognized in terms of the principle of continuity as viewed across spatially distinct samples. Since the number of dated samples is almost always far fewer than undated ones, continuity is demonstrable primarilyin a spatial mode. Thus, it is the temporal sequencing which is free to vary somewhat independently of the defining contexts. Approaching the past with the eyes of the New Worldview, we inevitably see the cultural past as a series of growths, followed by declines or collapses. It should be realized there are only two types of change which could possibly have been seen when "cohesion as measured by continuity" is the criterion for defining a culture: (1) growth where new culture traits are being added to the cohesive unit, and (2) disintegration where the cohesion breaks down. Any organizational change which reorganizes cultural phenomena will give the appearance of disintegration, since cohesion of disparate culture traits is the criterion for recognition of cultures themselves (cf. Erasmus 1968). It should further be remembered that a cohesion is an association of culture traits, and thus a qualitative phenomenon. Since the culture traits themselves are each qualitatively defined in terms of the criterion of continuity-that is, the same properties tend to "cohere" in items recovered from many different places-it is not surprising that New World cultures tend to exhibit geographicalcontinuities. In marked contrast is the view of the past generated through the use of "la methode Bordes." Here we see very different cultures living side by side in the same regions, characterized by a lack of geographical continuity sometimes described as "parallel phyla." We see a past where tenaciously unchanging cultures replace one another in confusing historical patterns within a similar region, and a lack of temporal continuity describedas "alternatingindustries" is sometimes claimed. Although this is sometimes a point of controversy, we see a past in which cultures change less, mix less, and are modified gradually through time, with few cases of collapse or decline. The picture one obtains is of replacement, not decline; gradual transitions, not punctuated change. Under the Old World view one expects cultures to exhibit a branching, diversifying pattern, very similar to that of biological evolution.

146

JOURNAL OF ANTHROPOLOGICAL RESEARCH

Table 1 provides a summary in outline form of the points of contrast between "Kriegerian," or "New World," and "la methode Bordes," or "Old World," approaches to classification. It should be apparent that the criterion of similarity has permitted archaeologists using "la methode Bordes" to judge which materials from different places represent a single culture. On the other hand, it has been recognized that "peoples" used different things, i.e., different artifact types. The criterion of association has permitted archaeologists to judge which different things went together within a given "culture." Different things found together (criterion of association) in an archaeological "level" would show what different things a "people" had. Similar assemblages found in different places (criterion of similarity) would show archaeologists where different "peoples" had been. It should be recognized that in the Old World view these conventions for interpretation do not admit the possibility that aspects of a single cultural system could appear as different assemblages at different places. This possibility would, by convention, be designated as an indication of cultural differences per se, used for the definition of different cultures borne by different peoples. In the words of Sonneville-Bordes (1975:3), "les types et leurs proportions sont stables et constants a l'interieur d'une meme culture pour une periode donnee dans une region donn6e, du moins dans certaines limites." If "la methode Bordes" is followed rigorously, it absolutely prevents us from ever seeing any organizational facts about past systems beyond those which may be manifest within a single occupation or a single level at a site. All units of synthesis beyond the assemblage will be internally homogeneous by convention of interpretation. We could thus never gain an appreciation for the organization of internally differentiated components of a system which might be manifest at different places. Turning to the "Krieger method," we face a slightly different problem: the characteristics of sites as such are not studied. In this approach the basic unit of observation is the artifact, in a framework of attributes. Types may be recognizable in many different data classes. Every class of items does not yield types, for some may be judged so generalized in their distributions as to be "nondiagnostic," and as such are most often ignored. Items of this kind traditionally would include most lithic material which is not bifacially worked, many kinds of ceramic utility wares, and classes of tools such as hammer stones, choppers, etc. Another characteristic of cultures defined in terms of "cohesions as measured by continuity" is that features such as pits, house forms, and hearths are frequently not considered to be basic cultural diagnostics because, from a pragmatic perspective, it is recognized that these features are not regularly preserved at most sites, or in some cases, are too expensive to recover. In other words, traits which are frequent, not too generalized, and easily recovered from different places are given priority as the defining characteristics of cultures. The only time that "places" are studied is when it is judged that some extraordinary conditions of preservation are present. Then the aim of archaeology is shifted from studying "culture history" to "reconstructing lifeways," and more attention may be given to the excavation of a site as a location that was used and lived in by past peoples. In general, the alternative aim of reconstructing the past is considered to be possible only under conditions of extraordinary preservation, which frees archaeological interpretation from restrictions thought to be imposed

PARADIGMS, SYSTEMATICS, AND ARCHAEOLOGY

147

TABLE 1 Old World and New World Paradigms "La Methode Bordes"

"The Krieger Method"

A. Framework for Observation

Type List

Selected Attributes

B. Basic Unit of Observation

Assemblage

Artifact "Type"

1. Criterion for recognition

Principle of Association

Principle of Continuity

2. Observational Framework

Recognizable Depositional Strata

Examples from Multiple Locations in a Region

Quantitative Paradigmatic Arrays

Grouping of "types"

Cumulative Graphs

Trait Lists

"A Culture"

"A Culture"

1. Criterion for Recognition

Principle of Similarity

Principle of Association

2. Observational Framework

Assemblages from Different Sites

Patterned Repetition of Types at Different Sites

(Case comparisons)

(Matrix comparisons)

Similar quantitative patterns seen among similar things

Qualitatively defined different things regularly associated at different places

(Similarity)

(Continuity)

Operations

C. Method of Description

1. Method of Presentation D. Unit of Synthesis

E. "Cohesion" Measured by

by the "limitations" of the archaeological record (see Binford 1981b). Thus, one only investigates "places" intensively if they are judged to be little "Pompeiis" which might offer extraordinary "glimpses" of particular past events or conditions. When doing culture-historical research, one normally needs only to recover a sufficient sample of artifacts to permit a "cultural" assessment of the remains. This means that no real understanding of internal differentiations or organizational variability among components of a single system will be revealed by carrying out normal, traditional archaeological work. In sum: (1) The properties unique to sites are generally ignored or, if described, are not discussed in terms of developing arguments about "culture." This is perhaps

148

JOURNAL OF ANTHROPOLOGICAL RESEARCH

most clearly illustrated by citing "features"-hearths, pits, and other structured of in sites. Classes of which are common to many things (2) things arrangements sites over wide areas are often ignored. (3) Frequency variations among classes of things not considered "diagnostic" are frequently ignored. Even more telling is the absence of anything more than impressionistic treatment of frequency variations among different data sets. For example, relationships between bone frequencies and artifact-class frequencies, or feature frequencies and ceramic frequencies, are rarely discussed in describing sites and are almost never discussed when "cultures" and their definition are the subject of discussion. The implications are clear: the organizational properties of cultural systems as manifest in the differential use of places is logically excluded in "la methode Bordes" and is ignored (in favor of measures of cohesion and overriding similarities which tend to exhibit temporal or spatial continuities) in the "Krieger method." Only in rare cases where the archaeologist judges a site to be extraordinarily preserved does the perspective shift toward describing the internal relationships among data sets within a site, and even then the purpose has been to reconstruct small segments of the past, not to seek an understanding of the past in general. CONCLUSION

When new archaeologists, and most particularly the senior author, began arguing for a change in the way archaeologists analyze the archaeological record, they did not argue for a particular theory, or propose a new theory as to how the world worked. Instead, they argued for a change in paradigm. They viewed the need for change as a shift in perspective, and further recognized that it was unlikely that archaeologists would invent on the spot any one "new perspective" which would be the most useful for all archaeological research. The discipline needed to try a variety of new perspectives which would permit it to explore the information potential of the archaeological record. It is no accident that New Perspectives in Archaeology (Binford and Binford 1968) did not call for new theories in particular, nor did it attempt to develop a monolithic approach. If one examines the list of contributors and the directions which their research took after 1968, it would be very difficult to find a single unifying argument or position other than a general dissatisfaction with the conventions of traditional archaeology. Under the more traditional approaches, it had been argued that the archaeological record limits the kinds of information which the archaeologist might refer to characteristics of past cultures. The new archaeologists argued that the discipline had not even begun to explore the archaeological record nor assess its potential for yielding information about the past, since all traditional methods for making inferences had been derived from limited paradigmatic expectations regarding culture. In the mid to late 1960s, some new archaeologists began to discuss the problem of "verification." It was recognized that the "methods," or interpretation, used by traditional archaeologists were simply conventions, not subject to evaluation by reference to the so-called empirical materials with which they worked. Most archaeologists' understanding was that a scientist "took his or her ideas to experience" for evaluation. In other words, science represented a philosophy which

PARADIGMS, SYSTEMATICS, AND ARCHAEOLOGY

149

sought the growth of knowledge through subjecting ideas ("knowledge") to trial by experience! It was clear to many new archaeologists that under the procedures of traditional archaeology, new experiences never affected archaeologists' alleged knowledge. The latter servedto accommodate all new experiences, and the particular tactics of accommodation became the so-called reconstructed history. The discipline was that of "discovering" (Binford 1982). The problem in the above analysis is that it was assumed that experience could be conceptualized independently of the ideas being evaluated by appeal to it. However, in our opinion, no objectivity had been achieved in the practice of traditional archaeology. Thus, the challenge today is how to achieve some independence for the experiences to which archaeologists appeal. Most archaeological reasoning has been a classic example of inductive argument from archaeological observations; no wonder the past never argued back! Archaeological interpretations have been inductively argued, and hence experience (the archaeological record) is simply the vehicle for inference. These inferences are logically tautologies in relation to the ideas which have guided the meanings given to the archaeologicalrecord. Initially, new archaeologists argued that the solution to this problem was to adopt another form of reasoning-deductive argument-where the premises were stated, consequences deduced, and these deductively reasoned expectations taken to . .. what? The past, or the archaeological record? Clearly, archaeologists in the late 1960s and early 1970s who argued for the potential of deductive procedures had not fully thought through the problem of the dependent status of their ideas regardingthe past. Archaeological knowledge of the past is totally dependent upon the meanings which archaeologists give to observations on the archaeologicalrecord. Thus archaeologically justified views of the past are dependent upon paradigmatic views regardingthe significance of archaeological observations. It is this basic point which we have tried to illustrate here. The challenge to archaeologists is the realization that the archaeological record cannot be used to test propositions about the meaning of archaeological observations in any direct sense. Such a realization inescapably leads to the conclusion that testing and verification of received ideas is really only possible at two basic junctures: (1) in an actualistic context, where archaeologists can evaluate propositions regarding the meaning to be attached to archaeological observations, i.e., middle-rangeresearch (see Binford 1981a); and (2) in the context of evaluatinggeneral theories regarding the "causes" of history. The latter of course assumes the existence of a body of unambiguous, independently warranted meanings, which can be attached to archaeological observations for evaluating the accuracy of proposed causal interactions operative in the past. In short, the key to either knowing the past accurately or evaluating theories about past processes is recognizing that both are dependent upon researchin the dynamic mode (actualistic or historical studies which allow archaeologists to assess the necessity of alleged cause-and-effect relationships between static and dynamic states of matter). In other words, middle-range studies make it feasible for archaeologists to attempt to "know" the past. What guides such studies? The answer is archaeologists' paradigmatic understanding of the archaeological record. This conclusion implies

150

JOURNAL OF ANTHROPOLOGICAL RESEARCH

something quite important: namely, that archaeologists must develop ways of increasing the accuracy and utility of their paradigm. Search the literature of science or the philosophy of science on this issue and one finds little aid or comfort. In the very early days of scientific discussion, the strict empiricists addressed the issue of methods for discovery, but the recent literature has tended to relegate this issue to an imponderable comer of "psychological causation" and restrict itself to considerations of evaluating received ideas (ideas already discovered or invented). These approaches are obviously of little aid to archaeologists today, whose task is the production of a useful paradigm, not the testing of theories produced in the context of the extant paradigm. Unlike skeptics such as Kuhn, we are convinced that we can learn to encourage productive paradigm change through rational means. We are also convinced that we can learn to evaluate competing paradigms objectively in spite of the claims for extreme intellectual relativism and noncomparability of theories and arguments generated in the context of different paradigms (see Binford 1982). For example, we can look back to the classic empiricist arguments of Francis Bacon for a guide to paradigm evaluation which is of continuing value to scholars today. Bacon's very arguments appear to be paradigmatic in character and not concerned with theories per se. He states (1947 [1620] :154;emphasis added) that: I am of opinion that if men had ready at hand a just history of nature and experience, and labored diligently thereon, and if they could bind themselves to two rules, the first to lay aside received opinions and notions, and the second, to refrain the mind for a time from the highest generalizations ... they would be able by the native and genuine force of the mind, without any other art, to fall into my form of interpretation. For interpretation is the true and natural work of the mind.

If archaeologists can gain a healthy skepticism regarding received conceptualizations of nature and seek to place themselves in positions relative to nature and experience where the adequacy and/or ambiguity of the received concepts may be evaluated, then they can hope to gain some objectivity relative to the utility of their concepts. Paradigm change is brought about and implemented, we believe, by seeking out new perspectives. The shift from a static to a dynamic perspective offers one example where the utility of concepts and the conventions of an observational language may be evaluated. As we have argued, the "observer perspective" is an important conditioner of the world to be seen. We have illustrated how in the Old World view the observer perspective is that of an internal participant looking from an egocentric position within the system. The New World view is that of an observer high above the cultural geography of a region, looking down at variability across previously identified ethnic units. The internal participant versus the aerial observer perspectives condition very different "realities." Are either of these observer perspectives really appropriate to the observational framework within which archaeologists commonly work? Let us briefly examine this question. In a very real sense, the spatial frames within which we most commonly work are the "site" and the "region." The site provides us with an observational window to the past which may be likened to the perspective of an immovable observer seated at the bottom of a well, looking up. All the observer can "see" from such a perspective is the "fallout" from some part of an organized system which happened to pass over that one stationary well. We would rarely if ever expect a

PARADIGMS, SYSTEMATICS, AND ARCHAEOLOGY

151

whole culture to pass over or be compressed so as to be seen within the confines of our tiny window. This means that what we see is always some "part" of a larger "whole." Given that the whole is an internally differentiated system, the parts seen from this perspective may be quite different in terms of contact, organization, and "role" played in the organized whole-the overall system. The challenge to archaeological methods is how to integrate reliably different assemblages, organizational forms (e.g., site structure), or parts into an accurate picture of the organized whole which existed in the past. It is very clear that the integrating criteria used under older perspectives-association and similarity-are not relevant and yield distortions. The criterion of continuity remains however as a possibility for integrating dissimilar assemblage units when studying noncomplex groups. Linked with the criterion of association at a much larger scale-the region-it at least appears feasible to begin the tentative description of large-scale, organizationally differentiated systems within which mobile hunter-gathererscarried out their lives. Almost certainly such an approach would be only an interim strategy, for once archaeologists learn how to look at systems from the realistic perspective of observer in a well, they will see many new things which can aid in the organizationaldiagnosis of past systems. The recognition of the most fruitful perspective to assume when conducting research is part of paradigm growth. For instance, if archaeologistsplace themselves in a productive research situation where both static and dynamic aspects of a system may be observed (an ethnoarchaeological project, for example) and assume observer perspectives which areunrealisticrelative to the archaeologicalrecord (as for instance those of internal observer or ethnographer), they will always come away having seen the archaeologicalrecord as limited and impoverished (see, for example, Gould 1980: 27-28). Rather, we must learn to see the dynamics from a perspective appropriate to the archaeological record. Such a viewpoint should be (1) nonparticipating, (2) outside, and (3) partitive. No amount of ethnographic observations will help us understand a system when viewed from the perspective of statics. Static remains are not the watered-down, impoverished "residues" of ethnographers' experiences. They are an existential domain which must be understood in terms of their own properties. Attempts to translate archaeological statics into a reconstruction of interpersonal interaction or forms of mental "deep structure" is akin to translating facts of cell biology into scenarios of predator-prey interactions. We need a science of the archaeological record. To achieve this goal, archaeologists need to continue to experiment with methods for both the production and refinement of a new paradigm appropriate to our science. If we have successes along these lines, then archaeology will begin to achieve the status of "archaeology as anthropology." Under the previous paradigms, archaeologists used culture to explain the archaeological record. We need to be able to reverse the situation and use the archaeological record to help further the anthropological goal of explaining cultural differences and similarities. NOTES 1. We wish to thank the students of our Anthropology 507 (Fall 1981) seminar at the University of New Mexico for their intellectual

stimulation; also Paula L.W. Sabloff and Harry Basehart for clarifying our ideas. 2. We are not arguing that Kuhn contends

152

JOURNAL OF ANTHROPOLOGICAL RESEARCH

that the logic which individuals may have used in advocating a scientific revolution was irrational. Rather, we are saying that he appears to us to be arguing that their theories do not necessarily follow in a rational, continuous fashion from the previous theories of the earlier, "normal science" phase. 3. Our argument is based on our interpretation of the relationship of ideas-not on histori-

cal influences among scholars. That is to say, we are not arguing, for example, that either Childe or Bordes were directly influenced by Durkheim. 4. These paradigms should not be confused with "etic" and "emic" approaches to theory building, because either could be used in terms of both paradigms.

REFERENCES CITED Bacon, F., 1947 [1620], Novum Organum (Book I). Pp. 73-154 in The World's Great Thinkers . . . The Philosophers of Science (ed. by S. Commins and R.N. Linscott). New York: Random House. Binford, L.R., 1962, Archaeology as Anthropology. American Antiquity 23:217-25. Binford, L.R., 1972, An Archaeological Perspective. New York: Seminar Press. Binford, L.R., 1976, Forty-seven Trips: A Case Study in the Character of Some Formation Processes. Pp. 299-351 in Contributions to Anthropology: The Interior Peoples of Northern Alaska (ed. by E.S. Hall, Jr.). National Museum of Man, Mercury Series, Paper No. 49. Ottawa: National Museum of Canada. Binford, L.R., 1977, General Introduction. Pp. 1-10 in For Theory Building in Archaeology (ed. by L.R. Binford). New York: Academic Press. Binford, L.R., 1981a, Bones: Ancient Men and Modern Myths. New York: Academic Press. Binford, L.R., 1981b, Behavioral Archaeology and the 'Pompeii Premise.' Journal of Anthropological Research 37:195-208. Binford, L.R., 1982, Objectivity-Explanation-Archaeology 1981. Pp. 125-38 in Theory and Explanation in Archaeology (ed. by C. Renfrew, M.J. Rowlands, and B.A. Segraves). London: Academic Press. Binford, S.R., and L.R. Binford, eds., 1968, New Perspectives in Archaeology. Chicago: Aldine. Bordes, F., 1968, The Old Stone Age. New York: McGraw-Hill. Childe, V.G., 1929, The Danube in Prehistory. Oxford: Oxford University Press. Childe, V.G., 1940, Prehistoric Communities of the British Isles. London: W. and R. Chambers. Clarke, D.L., 1968, Analytical Archaeology. London: Methuen. Dunnell, R., 1971, Systematics in Prehis-

tory. Glencoe: Free Press. Durkheim, E., 1938 [1895], The Rules of the Sociological Method, 8th ed. (tr. by S.A. Solovay and J.H. Mueller; ed. by G.E.G. Catlin). Chicago: University of Chicago Press. Erasmus, CJ., 1968, Thoughts on Upward Collapse: An Essay on Explanation in Archaeology. Southwestern Journal of Anthropology 24:170-94. Feyerabend, P., 1975, Against Method. London: Verso. Gould, R.A., 1980, Living Archaeology. Cambridge: Cambridge University Press. Hodder, I., 1977, Some New Directions in the Spatial Analysis of Archaeological Data at the Regional Scale. Pp. 223-351 in Spatial Archaeology (ed. by D.L. Clarke). London: Academic Press. Klimek, S., 1935, The Structure of California Indian Culture. Pp. 1-70 in University of California Publications in American Archaeology and Ethnology 37. Krieger, A.D., 1944, The Typological Concept. American Antiquity 9:271-88. Kroeber, A.L., 1939, Cultural and Natural Areas of Native North America. Pp. 1-242 in University of California Publications in American Archaeology and Ethnology 38. Kuhn, T.S., 1962a, The Historical Structure of Scientific Discovery. Science 136:76064. Kuhn, T.S., 1962b, The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Kuhn, T.S., 1970, Postscript-1969. Pp. 174-210 in The Structure of Scientific Revolutions, 2nd ed., enlarged. Chicago: University of Chicago Press. Milke, W., 1949, The Quantitative Distribution of Cultural Similarities and Their Cartographic Representation. American Anthropologist 51:237-52. Rouse, I., 1965, The Place of 'Peoples' in

PARADIGMS, SYSTEMATICS, AND ARCHAEOLOGY Prehistoric Research. Journal of the Royal Anthropological Institute 95:1-15. Sonneville-Bordes, D. de, 1975, Discours de Mme. Denise de Sonneville-Bordes, President Entrant. Bulletin de la Societe Prehistorique Francaise 72(2):3-6. Spaulding, A.C., 1968, Explanation in Archaeology. Pp. 33-39 in New Perspectives in Archaeology (ed. by S.R. Binford and L.R.

153

Binford). Chicago: Aldine. Toulmin, S.E., 1972, Human Understanding. Princeton: Princeton University Press. Trigger, B., 1980, Gordon Childe: Revolutions in Archaeology. London: Thames and Hudson. Wissler, C., 1914, Material Cultures of the North American Indians. American Anthropologist 16:447-505.

WILLOW SMOKE AND DOGS' TAILS: HUNTER-GATHERERSETTLEMENT SYSTEMS AND ARCHAEOLOGICALSITE FORMATION Lewis R. Binford Hunter-gatherersubsistence-settlementstrategies are discussed in terms of differingorganizationalcomponents, "mapping-on"and "logistics,"and the consequences of each for archaeologicalintersite variability are discussed. It is further suggested that the differingstrategies are responsiveto differentsecurityproblems presented by the environmentsin which hunter-gathererslive. Therefore,given the beginningsof a theory of adaptation,it is possible to anticipate both differences in settlement-subsistencestrategies and patterningin the archaeologicalrecord througha more detailed knowledgeof the distributionof environmentalvariables. An old Eskimoman was asked how he would summarizehis life; he thoughtfor a momentand said, "Willowsmokeand dogs' tails; when we camp it's all willow smoke,and when we move all you see is dogs' tails wagging in front of you. Eskimolife is half of each." THIS MAN WAS CAPTURING IN A FEW WORDS a way of life now largely vanished from man's experience: mobile man pursuing food, shelter, and satisfaction in different places in his environment. This paper is a discussion of patterns that I have recognized through direct field study as well as long-term research in the historical and ethnographic literature dealing with hunting and gathering adaptations. I am interested in what, if anything, renders differences in man's mobility patterning, and in turn the archaeological "traces" of this behavior in the form of spatial patterning in archaeological sites, both "understandable" and "predictable."

The posture adopted accepts the responsibilityfor a systemic approach.That is, humansystems of adaptationare assumed to be internallydifferentiatedand organizedarrangementsof formally differentiated elements. Such internal differentiationis expected to characterize the actions performedand the locations of different behaviors. This means that sites are not equal and can be expected to vary in relation to their organizationalroles within a system. What kind of variability can we expect to have characterized hunting and gathering adaptations in the past? What types of organizational variability can we expect to be manifest among different archaeological sites? Are there any types of regular or determined variability that can be anticipated among different archaeological sites? Are there any types of regular or determined variability that can be anticipated among the archaeological remains of people whose lives might be characterized as "willows smoke and dogs' tails"? The archaeological record is at best a static pattern of associations and covariations among

things distributedin space. Givingmeaningto these contemporarypatterns is dependentupon an understandingof the processes which operated to bring such patterning into existence. Thus, in order to carry out the task of the archaeologist, we must have a sophisticated knowledge and understanding of the dynamics of cultural adaptations, for it is from such dynamics that the statics which we observe arise. One cannot easily obtain such knowledgeand understandingfrom the study of the archaeological remains themselves. The situation is similar to conditionsduring the early years of the development of medical science. We wish to be able to cure and prevent

disease. Do we obtain such knowledgethroughthe comparativestudy of the symptomsof disease?

Lewis R. Binford, Department of Anthropology, University of New Mexico, Albuquerque, NM 87131 Copyright ? 1980 by the Society for American Archaeology 0002-7316/80/010004-17$2.20/1

Binford]

SETTLEMENT SYSTEMS HUNTER-GATHERER

5

The symptoms are the products of disease. Can they tell us about the causes of disease? In a like

manner the archaeological record is the product or derivative of a cultural system such that it is symptomaticof the past. We cannot hope to understandthe causes of these remainsthrougha formal comparative study of the remains themselves. We must seek a deeper understanding.We must seek to understandthe relationshipsbetween the dynamicsof a living system in the past and the material by-productsthat contributeto the formationof the archaeological record remaining today. In still more important ways we seek to understand how cultural systems differ and what conditions such differences as a first step toward meaningful explanation for patterns that may

be chronologicallypreserved for us in the archaeological record. As in the earlier analogy with medical science, once we know somethingof the disease and its causes, we may codify the symptoms to permit accurate diagnosis. Similarly,in the archaeological world when we understand something of the relationship between the character of cultural systems and the character of

their by-products,we may codify these derivatives to permit the accurate diagnosis from archaeological traces of the kind of cultural system that stood behind them in the past. These are not easy tasks to accomplish. It has been my conviction that only through direct exposure to dynamics-the ethnoarchaeolog-

ical study of living systems-does the archaeologist stand the best chance of gaining sufficient o accurately diagnosing too or othos methods for oing patterned variability. developing tools My most extensive experience with living systems has been among the Nunamiut Eskimo (Inuit) of north-central Alaska. For this reason I will base my discussion of a "diagnostic approach" to

settlement pattern on some of my Eskimoexperiences. I will compare that understandingwith a numberof different settlement systems as documentedethnographicallyby others. I will then proceed to discuss how settlement systems may vary amonghunters and gatherers living in different environments.In the course of these discussions, I will consider the types of archaeological sites generated in different environmentsas well as some of the probable spatial arrangementsamong such sites. Good diagnosis is "theory dependent." I will therefore be concerned with the factors that condition or "cause" different patterns of intersite variability in the archaeological record. COLLECTORSAND FORAGERS In several previous discussions of the Nunamiut I have described them as being "logistically

organized."I have frequentlycontrasted their subsistence-settlementsystem with that of the San or "Bushman" peoples, whom, I have designated foragers. Foragers Figure 1 illustrates some of the characteristics of a foraging system (this figure is largely based on the /Gwi San as reported by Silberbauer [1972] ). Several points should be made here regard-

ing the characteristics of foragers. My model system as shown in Figure 1 illustrates seasonal residential moves among a series of resource "patches." In the example these include the "pans"

or standingwater sources, melon patches, etc. Foragingstrategies may also be applied to largely undifferentiatedareas, as is frequentlythe case in tropical rain forests or in other equatorialsettings. One distinctive characteristic of a foraging strategy is that foragers typically do not store foods but gather foods daily. They range out gathering food on an "encounter" basis and return to

their residential bases each afternoon or evening. In Figure 1, residential bases are represented by solid black dots along tracks indicated by double dashed lines. The circles around each residential base indicate the foraging radius or the distance food procurement parties normally travel out into the bush before turning around and beginning their return trip. Another distinctive

characteristic is that there may be considerable variability among foragers in the size of the mobile groupas well as in the numberof residential moves that are made duringan annual cycle. In relatively large or "homogeneous"resource patches, as indicated by cross-hatchingon the right of Figure 1, the number of residential moves may be increased but the distances between them reduced, resulting in an intensive coverage of the resource "patch." On the other hand, if

AMERICANANTIQUITY

6

RAINY SEASON

.'

-,

HUTS^

FI

r

W...T

f

III

'\

'

.

I

II

F

I UY

.

D

l

WATER

DURING RAINY SEASON

_ / WA EUUI L

G R I Y S

..:./

IA

II .

II / 11 /

i-' "Y . j"

\

'"

.

'S\ ^

^"

' E N.

' I..

OV CATE

II

^

I.,,

A O

"

. "/ \

MELO

4 " WATER

//'-.-.-.--'

*

,

S II

.

/^'\q

rl I,

%, ,'"M- ',&

;

z .55,

I-

I

[Vol..45, No. 1, 1980

RAI DAILNG

:

DYAIL SEO RADIUS \.

FORAGING

CATERPILLARS

\

.

'-,'"

'

It

AGGREGATED RESOURCES

SHELTERS

TRAPS

GENERAL FORAGINGSYSTEM

9

200

C o 25r

Um

Figure 1.

FEET

VALLEY

ABOVE FLOOR

Characterization of a foraging subsistence-settlement

system.

7

HUNTER-GATHERER SETTLEMENT SYSTEMS

Binford]

resources are scarce and dispersed, the size of the mobile group may be reduced and these small units scattered over a large area, each exploiting an extended foraging radius. This situation is indicated by the multiple residential bases on the lower left side of the "seasonal round" shown in Figure 1. I might point out that when minimal forager groups (that is 5-10 persons) are dispersed, there is frequently a collapse of the division of labor, and foraging parties will be made up of both male and female members involved in procuring largely identical resources. Perhaps the use of the desert San as a model for foraging strategies is somewhat misleading, since the most exclusive foragers are best known from equatorial forests. Table 1 summarizes some of the information from equatorial groups on numbers of residential moves, average distances between moves, and total distances covered during an annual cycle. What can be seen from Table 1 is that there is considerable variability among foragers in the duration of stay at different sites. For some extremely mobile foragers such as the Punam, as reported by Harrison (1949), residential sites would be extremely ephemeral; one could expect little accumulation of debris and very low archaeological "visibility." There is another characteristic which may vary among foragers to further condition the "visibility" of the archaeological record: that is the relative redundancy of land use from year to year. One gains the impression, from descriptions of such groups as the Punam (Harrison 1949), the Guayaki (Clastres 1972), and other highly mobile foragers, that camps are not relocated relative to locations of previous use. The resources exploited are scattered but ubiquitous in their distribution and are not clumped or specifically localized as might be the case in deserts where waterholes are limited in number and discretely placed. Under the latter conditions we might expect more year to year redundancy in the occupation of particular places. Extreme examples of limited locations for critical resources may result in what Taylor (1964) has called tethered nomadism, indicating extreme redundancy in the reuse of identical places (water sources) over long periods of time. Such spatial discreteness tends to "tie down" the settlement system to specific geographical areas while other areas would be occupied little and rarely used because of their distance from such limited and crucial resources. We might think of a typical forager pattern of land use as looking like a daisy-the center is the residential base, and foraging parties move out, traversing search loops which resemble the petals of the daisy. Figure 2 illustrates this actual pattern as recorded by John Yellen (1972) for a mobile group of Dobe !Kung. In recognition that there is an alternative strategy which may be executed occasionally by peoples who are basically foragers, I have indicated a different pattern in the lower right-hand corner of Figure 1. We might think of this as a hunting trip where several men leave a residential Table 1. Summaryof GroupSizes and Annual Mobilityfor a Numberof Equatorialand SubequatorialGroupsof Hunter-Gatherers.

Group Name

Modal Group Size

Numberof Annual Residential Moves

Penum

65

45

18 120 75 50-20 45 25 55-18

21 12 26 50 22 31 5 11-12

Semang Mbuti Siriono Guayaki Aeta Hadza Dobe !Kung G/wi

Mean Distance between Sites (miles) 4.2 7.1 8.3 14.2 3.7 8.0 8.2 14.8 16.8

Total Circuit Distance Covered Annually 195 150 100 370 220 178 256 75 193

References Harrison 1949:135 Schebesta 1929:150 Bicchieri 1969:149 Holmberg 1950 Clastres 1972:150 Vanoverbergh 1925:432 Woodburn 1972:194 Lee 1968:35 Silberbauer 1972:297

Note. These values are estimates from either the observers and interviewers or calculations made by me

from indirect informationprovidedby such authors.

AMERICANANTIQUITY

[Vol.45, No. 1,1980

I

"I,'

0

1

2

3

4

MILES COMPOSITETAKEN FROM YELLEN

5 1968

DOBE 0 Figure 2.

Ac.ual map of foraging trips made by !Kung San around base camps.

base, establishing overnightcamps from which they move out in search of game, frequentlyusing what I (Binford 1978b) have termed an encounter of strategy. If they succeed in their hunting endeavors, and if the body size of the animal is large or the distance to camp is great and the temperature is warm, they may elect to dry the meat in the field and transport processed meat back to camp. This possibility is indicated by the little drying rack in the lower right-handcorner. They may then elect to return to the base by the original route or, if more meat is needed, they are more likely to return by a new route where they may even have further success in hunting.This little huntingtrip represents a different type of strategy. It is a specialized work party, in this case made up of men, who establish camps for their own maintenanceaway fromthe residentaial base

Binford]

SETTLEMENT SYSTEMS HUNTER-GATHERER

9

camp where they live. They may conduct special activities which may be only very rarely conducted in the residential base camp. This type of strategy may leave a very different kind of archaeological record and one we will explore in more detail in the next model. Before going on, however, it may be useful to summarize something of our expectations regarding the archaeological remains of foraging strategies. The first point to be made regarding the archaeological remains of foraging strategies is simply that there are apt to be basically two types of spatial context for the discard or abandonment of artifactual remains. One is in the residential base, which is, as we have seen, the hub of subsistence activities, the locus out of which foraging parties originate and where most processing, manufacturing, and maintenance activities take place. I have indicated that among foragers residential mobility may vary considerably in both duration and the spacing between sites; in addition the size of the group may also vary. These factors would condition the character of the archaeological record generated during a single occupation. I have suggested that foragers may be found in environmental settings with very different incidences and distributions of critical resources. In settings with limited loci of availability for critical resources, patterns of residential mobility mabe tethered around a series of very may restricted locations such as water holes, increasing the year to year redundancy in the use of particular locations as residential camps. The greater the redundancy, the greater the potential buildup of archaeological remains, and hence the greater the archaeological visibility. Thus far, I have basically reiterated some of the generalizations Yellen (1977:36-136) formulated from his I (Binford Bushman, as well as some of the experiences with the Kalahari ararguments 1978b:451-497) derived from my observations of Nunamiut Eskimo residential camps. The further characteristics of the residential base will become clearer in contrast with the other type of archaeological occurrence that foragers are apt to produce: the loction. A location is a place where extractive tasks are exclusively carried out. Since foragers generally do not stockpile foods or other raw materials, such locations are generally "low bulk" procurement sites. That is to say, only limited quantities are procured there during any one episode, and therefore the site is occupied for only a very short period of time. In addition, since bulk procurement is rare, the use, exhaustion, and abandonment of tools is at a very low rate. In fact, few if any tools may be expected to remain at such places. A good example of a location generated by foragers, a wood-procurement site, is described by Hayden (1978:190-191). As a rule, they are spatially segregated frombase camps and are occupiedfor short durations(usuallyonly a matter of hours at the most) by task-specific groups; . .. the lithic tools employed are generally very distinctiveand the assemblages highlydifferentiatedin terms of proportionalfrequenciescomparedto base camp assemblages . . . the tools used are often obtained locally near the procurementsite, and are generally left at the site after the activity is accomplished.... if one walked extensively among the mulga grove, one could see an occasional choppingimplement,usually left at the base of a decaying mulgatrunk. Rarely were there more than two choppingimplements,and the overall density must have been about one choppingimplementper 2500 m2 or less. Under low-bulk extraction or low redundancy in localization, the archaeological remains of locations may be scattered over the landscape rather than concentrated in recognizable "sites." Understanding such remains would require data-collecting techniques different from those archaeologists normally employ. So-called "off-site" archaeological strategies are appropriate to such situations. Given that long periods of time are involved and certain resources are redundantly positioned in the environment, we might anticipate considerable palimpsest accumulations that may "look" like sites in that they are aggregates of artifacts; however, such aggregates would commonly lack internal structure and would be characterized by accretional formation histories. Very important research into this type of archaeological distribution was initiated in this country by Thomas (1975). Further provocative investigations of so-called "off-site archaeology" are currently being pursued by Robert Foley (personal communication) of the University of Durham in the Amboseli area of Kenya. What can be summarized is that foragers generally have high residential mobility, low-bulk inputs, and regular daily food-procurement strategies. The result is that variability in the contents of residential sites will generally reflect the different seasonal scheduling of activities (if any) and

10

AMERICANANTIQUITY

[Vol. 45, No. 1, 1980

the different duration of occupation.The so-called "functionallyspecific" sites will be relatively few; given low-bulkinputs and short or limited field processing of raw materials such locations will have low visability thoughthey may well produce considerable "off-site" archaeological remains if long periods of land use are involved.Basically this type of system has received the greatest amount of ethnoarchaeological attention (e.g., Bushmen and central desert Australian Aborigines). Collectors In markedcontrast to the forager strategy where a group "maps onto" resources throughresidential moves and adjustmentsin group size, logistically organized collectors supply themselves with specific resources through specially organized task groups. Figure 3 illustrates some of the distinctive characteristics of a collector strategy. The model is generalized from my experiences with the NunamiutEskimo.In contrast to foragers, collectors are characterized by (1) te storage of food for at least stpart of the year and (2) logistically organized food-procurementparties. The latter situation has direct "site" implicationsin that special task groups may leave a residential location and establish a field camp or a station from which food-procurementoperations may be planned and executed. If such procurementactivities are successful, the obtained food may be field processed to facilitate transportand then movedto the consumers in the residential camp. Logistical strategies are labor accommodations to incongruent distributions of critical resources or conditionswhich otherwise restrict mobility.Put anotherway, they are accommodations to the situation where consumers are near to one critical resource but far from another equally critical resource. Specially constituted labor units-task groups-therefore leave a residential location, generally moving some distance away to specifically selected locations judged most likely to result in the procurementof specific resources. Logisticallyorganizedtask groups are generally small and composed of skilled and knowledgeableindividuals.They are not groups out "searching" for any resource encountered;they are task groups seeking to procure specific resources in specific contexts. Thus we may identify specific procurementgoals for most logistically organizedgroups.They go out to hunt sheep at the salt lick, or pursue big cariboubulls white fish. They or along the upland margins of the glaciers in July.They are fishing for grayling are not just out looking for food on an encounter basis. This specificity and "specialization"in procurementstrategy results in two types of functional specificity for sites produced under logistically organized procurement strategies. Sites are generated relative to the properties of logistical organizationitself, but they are also generated with respect to specific types of target resources. For foragers, I recognized two types of site, the residential base and the location. Collectors three additionaltypes of sites by virtue of the logistical character of their proateast generate curement strategies. These I have designated the field camp, the station, and the cache. A field camp is a temporaryoperationalcenter for a task group.It is where a task groupsleeps, eats, and otherwise maintains itself while away from the residential base. Field camps may be expected to be furtherdifferentiated accordingto the nature of the target resources, so we may expect sheephunting field camps, caribou-huntingfield camps, fishing field camps, etc. Collectors,like foragers, actually procure andlorprocess raw materials at locations. However, since logistically organizedproducer parties are generally seeking products for social groups far larger than themselves, the debris generated at different locations may frequently vary considerably, as in the case of group bison kills on the Plains (see Frison [1970]orWheat [1967]1)or spring intercept caribou kill-butcheringlocations amongthe Nunamiutsuch as the site at Anavik (Binford1978b: 171-178). Sites of majorfish weirs or camas procurementlocations on the Columbia plateau might be examples of locations with high archaeological visibility as opposed to the low-visibilitylocations commonlygenerated by foragers. Such large and highly visible sites are also the result of logistically organizedgroups,who frequentlyseek goods in very large quantities to serve as stores for consumptionover considerable periods of time.

HUNTER-GATHERER SETTLEMENT SYSTEMS

Binford]

/

i -.'.f"

-;

11

LOGISTICAL

SYSTEM

DIDISPERSED SUMSME

X_^ x-, \^;

/(,'',,''''

2

'K

EARLY

^BB

.

^^^^

SUMMER VILLAGE^

_

C^

(-

^

'<<

'S)|

MEAT DRYING

L '?,t@/ ILTd

5

;

1

i

10

MILES

Figure 3.

Characterization of a collector subsistence-settlement

system.

12

AMERICAN ANTIQUITY

[Vol. 45, No. 1,1980

Stations and caches are rarely produced by foragers. Stations are sites where special-purpose task groups are localized when engaged in information gathering, for instance the observation of game movement (see Binford 1978b) or the observation of other humans. Stations may be ambush locations or hunting stands from which hunting strategy may be planned but not necessarily executed. These are particularly characteristic of logistically organized systems, since specific resource targets are generally identified and since for each target there is a specific strategy which must generally be "informed" as to the behavior of game before it can be executed. Caches are common components of a logistical strategy in that successful procurement of resources by relatively small groups for relatively large groups generally means large bulk. This bulk must be transported to consumers, although it may on occasion serve as the stimulus for repositioning the consumers. In either case there is commonly a temporary storage phase. Such "field" storage is frequently done in regular facilities, but special facilities may be constructed to deal specifically with the bulk obtained (see Binford 1978a:223-235). From the perspective of the archaeological record, we can expect residential bases, locations, field camps, stations, and caches as likely types of sites generated by a logistically organized system. Within each class we can expect further variability to relate to season and to the character of the resource targets of logistically organized task groups. There is still an additional source of variability, since all the logistical functions may not necessarily be independently located. In some situations one might be able to use the field camp as an observation point, in others it may equally serve as a hunting stand. On occasion kills (locations) may be made directly from a hunting stand, and the meat may be processed and temporarily cached there. Many other combinations can be imagined. The point is simple, the greater the number of generic types of functions a site may serve, the greater the number of possible combinations, and hence the greater the range of intersite variability which we may expect. Against this background it is perhaps instructive to follow out some of the conditions modeled in Figure 3. Beginning with the winter village (site) at the lower center of the map. several conditions are indicated. The winter village is a cluster of relatively substantial houses located in a stand of willows (winter fuel). To the left of the village a series of expeditions are indicated; these are carried out by special trapping parties for the purpose of obtaining fur for winter clothing. To the right of the village are a series of site types: a field camp, where a hunting party is maintained while away from the residential camp; a station, or observation site, which is occupied and used basically for collecting information on game presence or movement; and several locations, kill sites and cache locations, which might also represent archaeological accumulations. With early summer a residential move is indicated (site B); this move results in a change in housing and a dependence upon dry rather than frozen meat as was the case in the winter village. From such a site, logistically organized parties may range out considerable distances to hunt such game as caribou or mountain sheep. Field camps and stations, like observation points and a variety of kill locations, might be generated. We see additional complexity caused by the differential combination of functions at different locations. For instance, to the far right of the map is a combined field camp and observation point; in other situations these functions might be spatially separated. In the upper part of the map an additional residential move is suggested. This move is accompanied by a reduction in group size as the local group breaks down into family units, each establishing independent residential camps having slightly different logistical patterning. It should be clear by now that we are not talking about two polar types of subsistence-settlement systems, instead we are discussing a graded series from simple to complex. Logistically organized systems have all the properties of a forager system and then some. Being a system, when new organizational properties are added, adjustments are made in the components already present such that residential mobility no longer plays the same roles it did when the system had no logistical component, although important residential moves may still be made. Given basically two strategies, "mapping on" and "logistics," systems that employ both are more complex than those employing only one and accordingly have more implications for variability in the archaeological record. It should be clear that, other things being equal, we can expect greater ranges of intersite variability as a function of increases in the logistical components of the subsistence-settlement system.

Binford]

HUNTER-GATHERER SETTLEMENT SYSTEMS

13

DISCUSSION Thus far I have been talking about the patterningthat I have perceived in the way hunters and gatherers are organized for subsistence purposes. I have been offering certain analytical and descriptive suggestions as to things one might look for in characterizinghunter-gathereradaptations. I have been attemptingto justify a particular way of looking at hunter-gatherersand suggesting that there are some interesting empirical patterns manifested by hunter-gathererswhen they are looked at from the perspectives advocated. Can we now begin the importanttask of buildingan explanation for the variability presented? Canwe begin to understandthe particular adaptive conditionswhich humangroupsdifferentially face by virtue of coping with different environments?Canwe understandwhich conditionswould favor "mappingon" versus "logistically organized" strategies? Beginningwith a more specific question, are there any clues to the factors that favor or select for a foraging or logistical strategy? If we assume that technological and social characteristics contributeto makingup the means and organizationof production,we wish to know if there are not some basic "determinants" conditioningthe distributionof differing "modesof production"(that is, the characteristic mixes of technology and social organization organized for subsistence purposes). Put another way, since systems of adaptation are energy-capturingsystems, the strategies that they employ must bear some relationship to the energy or, more important,the entrophystructure of the environments in which they seek energy. We may expect some redundancy in the technology or means, as well as the organization (labor organization),of production to arise as a result of "natural selection." That is the historical movement toward an "optima" for the setting. Put another way, technology, in both its "tools" sense as well as the "labor" sense, is invented and reorganizedby men to solve certain problemspresented by the energy-entrophystructure of the environmentin which they seek to gain a livelihood. Giventhis viewpointwe would expect that a foragingmode of productionwould serve men well in certain environmentalconditions, but not necessarily in all. What might some of these conditions be? Are there any environmentalsettings where we mightexpect foragingstrategies to offer "optimal"security for groups of hunter-gatherers?I think it is fair to suggest that althoughmost people view seasonal mobilityof residential locations as being responsive to differences in food abundance, most have little appreciation for the environmentalconditionswhich structure food abundance from the perspective of the human consumer.Perhaps Hollywoodcan be blamed for the widespread idea that "jungles"are food rich while deserts and arctic settings are food poor. In turn most laymen and beginningecology students alike expect the greatest residential mobility in arctic and desert settings and most "sedentism" among non-food-producersin equatorial settings. Simply as a means of provocative demonstrationI have adopted as a basis for further discussion data from Murdock(1967) regardingsettlement patterns. Murdockrated 168 cases of hunters and gatherers as to their degree of residential mobility.Each groupwas scaled from one to four as follows (see Murdock1967:159):(1) fully migratoryor nomadicbands, (2) seminomadic communitieswhose memberswander in bands for at least half of the year but occupy a fixed settlement at some season or seasons; (3) semisedentarycommunitieswhose membersshift from one to another fixed settlement at different seasons or who occupy more or less permanentlya single settlement from which a substantial proportionof the population departs seasonally to occupy shifting camps; and (4) compact and relatively permanentsettings. These 168 cases are summarizedin Table 2, which cross tabulates Murdock'sestimates of residential mobility against a measure of environmentalvariability developed by Bailey (1960), called "effective temperature"(ET).This measure simultaneouslydescribes both the total amount and yearly distributionof solar radiation characteristic of a given place. Stated anotherway, ET is a measure of both the length of the growing season and the intensity of solar energy available duringthe growing season. Since biotic productionis primarilya result of solar radiationcoupled with sufficient water to sustain photosynthesis,we can expect a general relationshpto obtainbetween ET value and global patterns of biotic activity and hence production.Other things being equal, the higher the ET value, the greater the productionof new cells within the plant or pro-

AMERICANANTIQUITY

14

[Vol. 45, No. 1,1980

Table 2. Cross Tabulation of Settlement Pattern as Evaluated by Murdock (1967) and ET (Effective Temperature) Values as Calculated from World Weather Records. Effective temperature

Fully Nomadic (1) 2 1 3 2 1

25 24 23 22 21 Sub-total

9(75.0%)

Sub-total

1 3 2 1 2 9 (64.2%)

20 19 18 17 16

SemiNomadic (2)

SemiSedentary (3)

Sedentary (4)

0 0 1 0 1 2 (16.7%)

0 1 0 0 0

0 0 0 0 0

2 2 4 2 2

1(8.3%)

0

12

1.33

1 1 1 0 1

1 o 0 0 0

4(28.5%)

0 0 0 0 0 0

3 4 3 1 3 14

1.42

0 5

15 17 32

2.31

5(9.4%)

28 25 53

2.33

3 4

29 16

7 (15.4%)

45

2.46

10 2 12

1.91

2 1

11 10

1(7.1%) 2 1

Sub-total

3 (9.3%)

21(65.6%)

3(9.3%)

5(15.6%)

13 12 Sub-total

3 1 4 (7.5%)

17 15 32 (60.3%)

4 8

4 1

11 10 Sub-total

2 3

15 6

5(11.1%)

21(46.6%)

9 8

5 0

3 1

1 1

1 0

5(41.6%)

4(33.3%)

2(16.6%)

1(8.3%)

35(20.8%)

84(50.0%)

31(18.4%)

15 14

Sub-total Grand total

12(22.6%) 9 3 12(26.6%)

18(10.7%)

Index Total Value

168

ducer component of the habitat. This means that in a very simplistic sense we might expect "food rich" environments when ET is high and "food poor" environments when ET is low. Table 2 illustrates some provocative facts. We note that "fully nomadic" strategies characterize 75% of the hunter-gatherer cases located in a fully equatorial environment (ET 25-21); high mobility is also found in 64.2% of the cases in semitropical settings. In warm temperate settings we note a drastic reduction of hunter-gatherers who are "fully nomadic" (only 9.3%), and in cool temperate settings the number is still further reduced (7.5%). Then as we move into boreal environments the number of fully nomadic groups increases slightly (11.1%), and in full arctic settings it increases drastically (reaching 41.6%). Thus we see that mobility, as measured by Murdock's categories, is greatest in equatorial settings, where we have the highest production in the world, and in arctic settings, where we have the most consistently low production. Summarizing the data from Table 2 another way, we observe the greatest concentration of sedentary and semisedentary hunters and gatherers in the temperate and boreal environmental zones and the least in equatorial and semiequatorial settings. This empirical pattern shows that mobility among hunter-gatherers is responsive to conditions other than gross patterns of "food abundance." This is indicated by the disproportionate occurrence of reduced mobility in the cooler, less productive environments. I suggest that since mobility is a "positioning" strategy, it may well be most responsive to structural properties of the environment, that is to say the particulars of food distribution that are not directly correlated with the more intuitively appreciated conditions of food abundance.

Binford]

SYSTEMS SETTLEMENT HUNTER-GATHERER

15

A clue to the types of problemsthat different strategies solve is perhaps best soughtin the contrasts between the two basic strategies themselves. Foragers move consumers to goods with frequent residential moves, while collectors move goods to consumers with generallyfewer residential moves. The first strategy, that of "mappingon," would work only if all the critical resources were within foraging range of a residential base. Logistical strategies (by collectors) solve the problemof an incongruousdistributionamongcritical resources (i.e., the lack of a reliable supply of a critical resource within the foraging radius of a residential base camp presumablylocated with regard to an equally critical resource). Underconditionsof spatial incongruityit must be appreciated that a residential move will not solve the problem.A move toward one location reduces the access to the other. It is under this condition that a logistical strategy is favored. Huntergatherers move near one resource (generallythe one with the greatest bulk demand)and procure the other resource(s) by means of special work groups who move the resource to consumers. In the case of temporalincongruity,a storage strategy is the most likely means of solving the problem.One seeks to extend the time utility for one of the resources beyond its period of availability in the habitat. This is accomplishedgenerally by either dryingor freezing. Storagereduces incongruoustemporalphasing of resources, but it may increase the problemof spatialincongruity. Spatial incongruitymay be exacerbated in that storage accumulates considerable bulk in one place, which increases the transport costs of a residential move in favor of other resources that mightbe "comingin" or located some distance away. With increases in storage dependencethere will be an expected increase in the logistical componentof a settlement system. Finally, if the argument is made that incongruityamong critical resources, whether temporal or spatial, is a condition favoring logistical strategies and a reduction and change in the role of residential mobility, it must also be realized that any condition which either (1) increases the numbers of critical resources and/or (2) increases the climatic variance over an annual cycle will also increase the probabilityof greater incongruities among critical resources. Let us consider two logical expectations arising from this postulate. The law of requisite variety states that for maximumstability, the variety of homeostaticresponses requiredin any system is equal to the variety of environmentalchallenges offered to it. We can expect, therefore, that number of operative homeostatic the the more unstable the thermal environment,the greater mechanisms,and hence the greater the numberof critical resources, other things being equal. As the number of critical resources increases, there is a related increase in the probabilitythat a lack of congruencewill occur in their distributions.Therefore,the greater the seasonal variability in temperature, the the expected role of logistical mobilityin the settlement or "positiongreater ing" strategy. Given an equatorial environmentin which species may exhibit patterns of differential production over an annual cycle, but the interdigitationof differing schedules among species ensures that there will be continuouslyavailable foods, a foragingstrategy works very well. In temperate and still colder settings, such continuouslyavailable food is reduced as a functionof decreases in the length of the growing season. Humangroups attemptingto "make a living" must therefore solve the "over-wintering"problem.Basically three methods are available: (1) exploitingspecies who have themselves solved the over-winteringproblem(that is huntingother animals);(2) storing edible products accumulated largely during the growing season; or (3) storing animal resources accumulated during periods of high density and hence availability. Althoughwe must recognize that storage may not always be feasible, the degree to which it will be practiced can be expected to vary with decreases in the length of the growing season. The degree to which storage is practiced will, in turn, increase the likelihoodof distributionalincongruitiesand hence conditionfurther increases in logistically organized settlement systems with attendant reductions in residential mobility,at least seasonally. Both of these conditionsare related to environmentalreductions in the length of the growing season and to the implicationsof this for man, both in terms of foods and of other temperature-regulatedresources. This means that there is an environmentalconvergence of conditionsacting simultaneouslyto increase the number of critical resources and to increase the conditions favoring storage. Given the arguments presented here, we should therefore see a reduction in residential mobilityand an increase in storage dependence as the length of the growing season decreases.

AMERICANANTIQUITY

16

[Vol. 45, No. 1, 1980

It should be pointed out that both of these expectations are supported empirically. As was

previously indicated in Table 2, there is a markedincrease of cases classified as semisedentary and seminomadicin environmentswith ET less than 16?C. Stated another way, we see increases in seasonal sedentism, with attendant increases in logistically organized food procurement inferred, in such environments. Figure 4 illustrates the relationship between ET and storage dependence as estimated by Murdock and Morrow (1970) for a sample of 31 ethnographically documented hunters and gatherers. Storage dependece is indicated by an ordinal scale distributed from one to six, where six in-

dicates the greatest dependence upon storage. What is interesting in this small sample is that there is a clear curvilinear relationship between increased dependence upon storage and decreasing ETvalues, measuringdecreases in lengths of growingseason. It is notable that storage is practiced only amonghunters and gatherers in environmentswith ETvalues less than 15 (i.e., in environmentswith growingseasons less than about 200 days). Exceptionsto the general trend are interesting and perhaps instructive. In warm environmentsthere are only two exceptions, the Andamanese and the Chenchu. It is my impression that the Andamanese are miscoded while the Chenchuare demonstrablyin the process of adoptingagriculture.Exceptionson the "cold" end of the distribution are the Yukaghir, Yahgan, Slave, Copper Eskimo, and Ingalik. I believe the Yukaghirto be miscoded, as well as the Ingalik,while the other cases are probablytruly exceptional in being more mobile and not putting up stores for winter in appreciable amounts. Additional cases of cold-climate groups who do not put up appreciable stores might be the Micmac, Mistassini Cree, Igloolikand Polar Eskimo,and some groups of Copperand NetsilikmiutEskimo, as well as some temperate cases like the Tasmanians.Many of these groups mightbe technically foragers with relatively high residential mobility, nevertheless they are foragers of a different type than most equatorial foragers. As has been pointed out, equatorial foragers move their residences so as to position labor

*

*

*@

6-

*

.

5-J .)

\

w C)

(uI.$

-*

\ *YUAHR**

-

*INGALIK

I U)ll

O

- -

CHENCHU

\ 3

0

19

Z92

-

? COPPERESKIMO

I -

9

10

II

21

12

22

23

24

25

*ANDAMANISLANDERS

SLAVE

**

YoUKAGHIR ?0 YAHGAN

8

20

TEMPERATURE

EFFECTIVE

13

14

15

16

- -'--.

-@---- ---0----

17

18

19

20

21

22

23

0 -

24

25

EFFECTIVE TEMPERATURE Figure 4.

Graph of the relationship between storage dependence and effective temperature.

Binford]

SETTLEMENT SYSTEMS HUNTER-GATHERER

17

forces and consumerswith respect to food-yieldinghabitats considered in spatial terms. The coldenvironmentforagers are what I tend to think of as serial specialists: they execute residential mobility so as to position the group with respect to particular food species that are temporally phased in their availabilitythrougha seasonal cycle. Leavingsuch interestingissues aside for the moment,it should be clear that there are definite geographical patterns to the distributionof environmental conditions that pose particular problems for hunter-gatherers. Some of these specifiable problemsmay be well solved or at least effectively dealt with throughlogisticallyorganized productionstrategies. Such strategies answer the problemof incongruousdistributionsof critical resources. Incongruousdistributionsmay occur spatially and may be furtherexacerbated by storage strategies. Storage always produces a high bulk accumulationin some place, which then has an increased likelihoodof being incongruouslydistributedwith respect to other critical resources such as fuel, water, shelter, etc. High bulk stores necessitate the determinationof the relative cost of transporting consumers and stored goods to the loci of other critical resources versus that of introducingthese other resources to the storage location througha logistically organized productive labor force. I should point out that if there are other factors that restrain mobility, such as increased numbers of social units in the area, competitionamongmultiplesocial units for access to similar resources, etc., then we can expect an accompanyingincrease in logistically organizedproduction. This is not the place to take up such importantissues as the origins of agricultureand other density dependent shifts in both mobility and productive strategy, but I simply wish to point out that with any conditionthat restricts residential mobilityof either foragers or collectors, we can expect (amongother things)a responsive increase in the degree of logistically organizedproduction. SYSTEMSAND INTERASSEMBLAGE VARIABILITY SETTLEMENT CONCLUSIONS: The above discussion obviouslyhas significant implicationsfor our understandingof archaeological assemblages, their variability, and their patterning.I have argued elsewhere that we may think of an assemblage as a derivative of "some organized series of events characteristic of a system" (Binford1978a:483).An assemblage that is the accumulatedproduct of events spanning an entire year is rather gross and may be referred to as coarse-grainedin that the resolutionbetween archaeological remains and specific events is poor. On the other hand an assemblage accumulated over a short period of time, for instance a two-day camp, represents a fine-grained resolution between debris or by-products and events. Having made the above distinctions I previously argued: 1. Insofar as events are serially differentiated, and the compositionof assemblages are responsive to event differences, the more fine-grainedthe assemblage, the greater the probable content variability among assemblages. 2. The factor which regulates the grain of an assemblage is mobility, such that high mobility results in fine-grainedassemblages, whereas low mobilityresults in coarse-grainedassemblages. (For further discussion see Binford 1978b:483-495.) In reference to the initial condition,"the degree to which events are serially differentiated,"it was argued that from a subsistence perspective the major conditionerof event differentiationis seasonal variance in the basic climatic variables: rainfall and solar radiation. It was therefore suggested that interassemblage variability "can be expected to increase with decreases in the length of the growing season" and/or "decreases in the equabilityof rainfall distributionthroughout a seasonal cycle, given moderate to fine-grainedassemblages" (Binford1978b:484). The earlier argumentshad reference primarilyto residential mobility.In this paper I have explored something of the interaction and the determinantsfor differential degrees of residential versus logistical mobility.I have suggested here that there are two basic principles of organization employed by hunters and gatherers in carrying out their subsistence strategies. They may "map on" by moving consumers to resources, or they may move resources to consumers "logistically." I have suggested that the relative roles played by these two organizationalprin-

18

AMERICAN ANTIQUITY

[Vol. 45, No. 1, 1980

ciples in any given subsistence system will also conditionthe nature and character of archaeological intersite variability generated by the system. Foragers who practice primarily a "mapping on" strategy will generate basically two types of sites: the residential base and the location. Variability among forager systems will derive primarily from differences in the magnitude of residential mobility and environmentaldifferences conditioningdifferent subsistence activities through a seasonal cycle. Collectors who tend toward a greater reliance on the logistical strategies can be expected to generate additional types of archaeological sites. That is, in addition to the residential base and the location we can expect field camps, stations, and caches to be generated. It was also argued that the character of residential bases, as well as that of locations, may well be expected to change in accordance with the relative degree of logistically organizedactivity characteristic of a system. I then turned to the interesting question of what conditionsthe relative roles of "mappingon" versus "logistical" strategies in a subsistence-settlementsystem? It was argued that logistically based strategies are a direct response to the degree of locational incongruity among critical resources. It was further argued that the number of critical resources increases as climatic relative increases, and that the dependence upon stored foods increases as the length of severity the growing season decreases. It was pointed out that these characteristics are linked, and both tend to vary with geographical variability in the length of the growing season. Therefore,as the length of the growing season decreases, other things being equal, we can expect increases in the role of logistical strategies within the subsistence-settlementsystem. It was also pointed out that any other conditions that restrict "normal"residential mobilityamong either foragers or collectors also tend to favor increases in logistically organized procurement strategies. We would therefore tend to expect some increase associated with shifts toward agricultural production. I can now integrate my earlier arguments regarding the factors conditioninginterassemblage variabilityat residential bases with the argumentsmade in this paper regardingvariability in the archaeological record stemmingfrom organizationaldifferences in the roles of mappingon and logistical strategies in the subsistence-settlementbehavior of groups living in different environments. It was argued earlier that as seasonal variability in solar radiation or rainfall increased, given assemblage responsiveness to event differentiations,there would be an increase in residential interassemblage variability. This is assuming a roughly constant assemblage grain. In this paper it has been argued that under the same conditionsincreased logistical dependence with an accompanyingreduction in residential mobilitywould be favored. This situation would have the effect of increasing the coarseness of the assemblage grain from such locations. Increased courseness, in turn, shouldhave the effect of reducinginterassemblagevariabilityamongresidential sites of a single or closely related system occupied during comparable seasons. It would be coarseness, in turn, shouldhave the effect of reducinginterassemblagevariabilityamongresidential sites of a single or closely related system occupied during comparable seasons. It would of course also have the effect of increasing the complexityand "scale" of assemblage content referable to any given uninterruptedoccupation, assuming, that is, a responsiveness of assemblage content to event differentiations. The overall effect of what appears to be opposingconsequences is normallysome seasonal differentiationin the relative roles of residential versus logistical mobility.For instance, in some environmentswe might see high residential mobility in the summer or during the growing season and reduced mobilityduring the winter, with accompanyingincreases in logistical mobility.The overall effect from a regional perspective would be extensive interassemblage variability deriving from both conditions. We may also expect minor qualitative difference among assemblages from the winter villages (in the above examples). These are likely to be categorically different from mobile summer residences which would be highly variable and constitute a "noisy" category. Comparisonsamongwinter residences would clearly warrant a categorical distinction of these from summer residences and they would be a "cleaner," less noisy category of greater within-assemblagediversity. Summer sites would be more variable among themselves but also less internally complex.

Binford]

SYSTEMS SETTLEMENT HUNTER-GATHERER

19

The point here is that logistical and residential variability are not to be viewed as opposing principles (although trends may be recognized) but as organizational alternatives which may be employed in varying mixes in different settings. These organizational mixes provide the basis for extensive variability which may yield very confusing archaeological patterning. The next step in the arguments presented in this paper treats the production of special-purpose sites. It was suggested that with logistical strategies new types of sites may be expected: field camps, stations, and caches. It was further argued that the character and visibility at locations also changes in the context of increased use of logistical strategies. We may therefore argue that, other things being equal, we may anticipate regular environmentally correlated patterns of intersite variability deriving from increases in the number and functional character of special-purpose sites with decreases in the length of the growing season. In addition to such quantitative changes, given the more specialized character of resource "targets" sought under logistical strategies, we can expect an increase in the redundancy of the geographic placement of special-purpose sites and a greater buildup of archaeological debris in restricted sections of the habitat as a function of increasing logistical dependence (for a more extended discussion of this point see Binford 1978b:488-495). This last point addresses a subject not discussed in depth in this paper, namely, the long-term land-use strategies of hunter-gatherers in differing environmental contexts. This paper has primarily dealt with short-term organizational and strategy differences. "Short-term" here essentially means the dynamic of yearly cycle. I have argued that there are environmental factors conditioning variability in short-term mobility and land-use strategies among hunters and gatherers. I have not seriously considered the possibility that hunters and gatherers would ever remain sedentary as a security-seeking strategy unless forced to do so. I am aware of many arguments that essentially appeal to what I term the "Garden of Eden" principle, namely, that things were so "wonderful" at certain places in the environment that there was no need to move. I find that a totally untenable opinion, and one which can be countered easily by scholars who understand ecological relationships. This does, however, imply that an understanding of short-term strategies as discussed here is insufficient for treating patterning which derives from variable redundancy in geographical positioning of the total settlement-subsistence systems. A detailed consideration of the factors that differentially condition long-term range occupancy or positioning in macrogeographical terms is needed before we can realistically begin to develop a comprehensive theory of hunter-gatherer subsistence-settlement behavior. The latter is of course necessary to an understanding of archaeological site patterning. Acknowledgments. This paper was originallyprepared at the request of Peter Bleed, who graciouslyinvited me to participatein the 1979 MontgomeryLectureSeries at the Universityof Nebraska-Lincoln.Forboth the opportunityand the encouragementto prepare this paper I am most grateful. My colleague Jeremy Sabloff read and made constructive comments on earlier drafts as did William Morgan and Robert Vierra; for this assistance I am most grateful. Ms. Dana Anderson developed and prepared the illustrations;certainly the quality of her work adds appreciablyto this paper. The field work opportunitieswhich have providedthe stimulusfor muchof the discussionand my appreciation of hunter-gatherermobilitywere supportedby the National Science Foundation,the Wenner-GrenFoundation,and the AustralianInstituteof AboriginalStudies.A grant fromthe FacultyResearchCommitteeof the Universityof New Mexico aided in the preparationof the manuscript,particularlythe draftingof the illustrations. For this I am most grateful. Colleagueswho shared my interest in hunter-gathereradaptationshave provided me with stimulatingintellectual environments;I would particularlylike to mentionHenryHarpending, James O'Connell,Nick Peterson and JohnPfeiffer. REFERENCESCITED Bailey, Harry P. 1960 A method of determiningthe warmth and temperateness of climate. GeografiskaAnnaler 43(1): 1-16. Bicchieri, M. G. 1969 The differential use of identical features of physical habitat in connectionwith exploitative,settlement, and communitypatterns; the BaMbuti case study. In Contributionsto anthropology:ecological essays, edited by David Damas, pp. 65-72. NationalMuseumsof Canada,Bulletin230, Anthropological Series No. 86.

20

AMERICANANTIQUITY

[Vol.45, No. 1,1980]

Binford, Lewis R. 1978a Dimensional analysis of behavior and site structure: learning from an Eskimo hunting stand. American Antiquity 43:330-361. 1978b Nunamiut ethnoarchaeology. Academic Press, New York. Clastres, Pierre 1972 The Guayaki. In Hunters and gatherers today, edited by M. G. Bicchieri, pp. 138-174. Holt, Rinehart, and Winston, New York. Frison, George 1970 The Glenrock buffalo jump, 48C0304. Plains Anthropologist, Memoir 7. Harrison, Tom 1949 Notes on some nomadic Punans. The Sarawak Museum Journal V(1):130-146. Hayden, Brian 1978 Snarks in archaeology: or, inter-assemblage variability in lithics (a view from the Antipodes). In Lithics and subsistence; the analysis of stone tool use in prehistoric economics, edited by Dave L. Davis, pp. 179-198. Vanderbilt University Publications in Anthropology 20. Holmberg, Allan R. 1950 Nomads of the long bow; the Siriono of eastern Bolivia. Institution Institute of Social Anthropology Publication 10. Lee, Richard B. 1968 What hunters do for a living, or how to make out on scarce resources. In Man the Hunter, edited by Richard B. Lee and Irven DeVore, pp. 30-48. Aldine, Chicago. Murdock, G. P. 1967 Ethnographic atlas; a summary. Ethnology 6:109-236. Murdock, G. P., and Diana 0. Morrow 1970 Subsistence economy and supportative practices: cross-cultural codes 1. Ethnology 9:302-330. Schebesta, Paul 1929 Among the forest dwarfs of Malaya, translated by A. Chambers. Hutchinson Press, London. Silberbauer, George B. 1972 The G/wi Bushmen. In Hunters and gatherers today, edited by M. G. Bicchieri, pp. 271-326. Holt, Rinehart, and Winston, New York. Taylor, Walter W. 1964 Tethered nomadism and water territorality: an hypothesis. Acts of the 35th International Congress of Americanists, pp. 197-203. Mexico City. Thomas, David H. 1975 Nonsite sampling in archeology: up the creek without a site? In Sampling in archaeology, edited by James W. Mueller, pp. 61-81. University of Arizona Press, Tucson. Vanoverbergh, Morice 1925 Negritos of northern Luzon. Anthropos 20:148-199, 399-443. Wheat, Joe Ben 1967 A Paleo-Indian bison kill. Scientific American 216(1):44-53. Woodburn, James 1972 Ecology, nomadic movement and the composition of the local group among hunters and gatherers: an East African example and its implications. In Man, settlement and urbanism, edited by P. J. Ucko, R. Tringham, and G. W. Dimbleby, pp. 193-206. Duckworth, London. Yellen, John E. 1972 Trip V. itinerary May 24-June 9, 1968. In pilot edition of Exploring human nature. Educational Development Center, Inc., Cambridge, MA. 1977 Archaeological approaches to the present. Academic Press, New York.

JOURNAL

OF ANTHROPOLOGICAL

Butchering,

ARCHAEOLOGY

3, 235-257 (1984)

Sharing, and the Archaeological

Record

LEWIS R. BINFORD Department of Anthropology, University of New Mexico, Albuquerque, New Mexico 87131 Received May 21, 1984

Richard Gould has recently advocated certain methodological positions he considers to be appropriate to the field of archaeology. These positions have been placed in critical contrast to propositions which this author allegedly advocates with regard to archaeological method and theory. This paper seeks to correct Gould’s representation of this author’s views and to place in a broader perspective the suggestions which he has offered to the field for consideration. Q 1984 Academic Press, Inc.

INTRODUCTION

Richard Gould (1978, 1980; Gould and Watson 1982) has made positive claims for his inferential approach, at the same time implying that approaches used by others are limited in various ways. His arguments touch upon important issues currently under debate. For instance, Gould addresses patterns of butchery and the sharing of meat in his discussions of method and procedure. These are issues currently discussed among archaeologists trying to elucidate the behavioral significance of very early archaeological sites. Glynn Isaac (1978a, 1978b, 1980) has been a strong advocate of the view that early man lived in base camps and was organized into social groups within which the sharing of the meaty parts from animals was basic to their way of life. I (1981, 1984) have challenged this view and have been concerned to learn how we might render from the archaeological record an unambigious interpretation as to whether or not early man was sharing meat (see Binford 1978:272-273). A totally different context of argument stands behind Gould’s interest in sharing and in its potential influence on behavior, and, by implication, on patterning in the archaeological record. Gould has stated his case as follows: Philosophically, I can regard Binford’s analysis of his Nunamiut data as similar to the approach I am advocating here except for the fact that he has not made any explicit effort to incorporate possible ideational factors into his overall argument. 235 0278-4165184 $3.00 Copyright 8 1984 by Academic Press. Inc. All rights of reproduction in any form reserved.

236

LEWIS R. BINFORD Perhaps the Nunamiut really are as “simple” as he suggests, in the sense that he can account simply for their behavior in relation to meat products on a circumstantial basis without recourse to more complex explanations involving ideas and symbols. But the Western Desert Aborigine case contains anomalies that cannot be fully accounted for at this level of reasoning and must be referred to the ideational domain to achieve a satisfactory explanation of all the behavior that was observed . . (Gould and Watson 1982:367-368) The strict adherence to a fixed pattern of initial division of meat was explained more parsimoniously with reference to social relations based upon kin-based sharing of food and access to resources . . . than to the direct influence of the immediate circumstances under which hunting occurs. (Gould and Watson 1982:367) [Emphasis mine.]

In summary, Gould suggests that among the Ngatatjara a rigid and unvarying pattern of butchery is practiced. This rigidity is cited as anomalous, presumably in comparison to the great variety of butchering tactics which I (197847-62) reported from the Nunamiut Eskimo. Given Gould’s (Gould and Watson 1982:368) belief that when one recognizes an “anomaly” it signals the ideational domain as a conditioner or determinent, he naturally seeks such an explanation in the face of this alleged anomaly. Gould focuses on the ideology of sharing and the ethics of kinbased altruism as the most parsimonious “explanation” for the Australians’ rigid butchery pattern. Given Gould’s “archaic empiricist” (see Binford 1984) views, he considers my behavioral descriptions of the Nunamiut as misguided arguments from ethnographic analogy aimed at establishing empirical laws of human behavior. Gould states: For Binford, it is circumstance rather than culture that determines human behavior in relation to meat procurement and fauna1 remains. (Gould and Watson 1982:366)

In fact, in my Nunamiut study, I was examining the archaeological record as manifest in fauna1 remains. I sought to understand patterning and interlocational variability in terms of the behaviors executed by the Nunamiut actors at places studied archaeologically. My concern was how the same actors, bearers of a constant “culture” or ideational domain, generated variable patterns of association and covariation among fauna1 assemblages. Gould appears to misunderstand this research if he sees it as an attempt to establish that the patterned behavior observed among the Nunamiut was generalizable directly to all hunter-gatherers. Given such a view, Gould cites that different formal behavior of the Ngatatjara Australians as anomalous relative to my Nunamiut descriptions. This difference is then considered a justification for his seeking explanations in the “ideational domain”‘~* (Gould and Watson 1982:368). In the following sections I will consider Gould’s use of ethnographic * See Notes section at end of paper for all footnotes.

BUTCHERING,

SHARING,

AND

THE

ARCHAEOLOGICAL

237

RECORD

experiences and his advocacy of “living archaeology” as a way of addressing the important issue of assigning meaning to the archaeological record. Z will consider Gould’s suggestions in the context of seeking to understand how sharing might be manifest archaeologically. My work with the Nunamiut was conducted with the archeological record as the subject of research. Gould’s actualistic studies among the Aborigines do not seem to have been conducted with a similar focus. Instead, he studied the behavior of the aboriginal peoples in the same way an ethnographer might conduct research (see Binford 1983a:391). Having observed the ongoing dynamics of a living system, Gould then seeks to infer the character of the archaeological record. This is an argument from ethnographic inference, and it is this very tenuous practice that Gould advocates to all archaeologists under the label of “living archaeology”: . . . by observing the adaptive behavior of any living human society, we can predictions

about the society’s discards, we are doing

1980: 113) [Emphasis

living

archaeology.

derive

(Gould

mine.]

It is these derived “predictions” which should constitute “archaeological signatures”2 for the “adaptive behaviors” observed. In the case of Gould’s work we are left in the dark regarding the properties which would constitute an “archaeological signature” for the rigid butchering procedure he discusses. Whatever this “signature” might be, Gould would explain it most “parsimoniously” by reference to a strong sharing ethic. It would appear that the “derived prediction” is that the archaeological “signature” would also be rigid and a direct reflection of the butchering pattern, but on these issues Gould is silent. I find this surprising. Why is Gould not addressing the problem of the material correlates for the behavior and mental phenomena which he considers important? I think that the answer hinges on the realization that Gould is not arguing about the archaeological record. A different issue is at stake! EVALUATING

GOULD’S POSITION

In beginning an evaluation, one of the first thoughts to come to my mind is that essentially the same butchering procedure observed by Gould (1967) among the Australians was used by the Nunamiut Eskimo when butchering large to moderate-size birds such as Canada geese, Whistling swans etc. The only difference was (a) wings were generally removed discretely and did not remain attached to the breast, (b) the breast was not cut into halves at the time of initial butchering, and (c) there was no tail. Otherwise, the butchering procedure was the same as described by Gould for the kangaroo. The butchering pattern which fascinated Gould,

238

LEWIS

R.

BINFORD

when seen in the Nunamiut context, can be understood as another practical behavior pattern accommodated nicely to such situational variables as body size and the proportions of the animals being butchered.3 In all fairness, however, Gould has not stressed the formal properties of the butchering sequence as requiring a “cultural” explanation. Instead, he has pointed to the rigidity with which the Australians butcher animals in a “standard” manner as requiring explanation in the ideological domain. Do the Nunamiut

Lack

a Strong Sharing

Ethic?

Approaching Gould’s argument as a scientist, it would seem rational to ask, in the face of a proposed explanation for a given phenomenon, whether the explanation holds up in contexts different from the one which it was invented to accommodate? The strength of a scientific explanation is at least partially measured by the degree to which it facilitates understanding of experiences more comprehensive than those which seemingly rendered the argument rational in the first place. Gould’s argument, translated into a scientific frame of reference, would appear as a proposition which suggests that the greater the adherence to a sharing ethic with regard to meat products, the more rigid will be the butchering procedures employed by the people in question. Given this general proposition,

linked to the observation that the Nunamiut practice extremely variable butchering procedures, we would expect, if the proposition was generally true, that the Nunamiut lacked or had only a very limited commitment to a sharing ethic. Is this true? One of the inevitable sources of gossip among the Eskimo is a situation, in a context considered appropriate for sharing, where a hunter either failed to share or failed to do it in terms of the kin-based conventions for sharing. Whether men immediately share meat on their return to a residential camp or whether they place their game in the stores of the households they represent is primarily a function of a set of distinctions made by the Nunamiut between hunting for stores versus hunting for immediate consumption. This distinction is commonly associated with hunting for caribou versus local game. Most of the time, local game is considered a target in backup hunting strategies and hence such animals are commonly introduced when the group is poorly provisioned. The animals which fall into this category are mountain sheep, moose, and grizzly bear. These are also the animals most likely to be sought in the mountains during summer when hunting for fat to supplement dried meat stores (see Binford and Chasko 1976:101-111; Speth 1983:143-159). The factor which primarily conditions whether a hunter shares or places meat in storage is his judgment as to whether or not everybody

BUTCHERING,

SHARING,

AND THE ARCHAEOLOGICAL

RECORD

239

already has meat. If the verdict is yes, then the hunter turns over his kills to his wife (places them in storage). On the other hand, if kinsmen are judged to have little food and to be in a needy state the hunter shares out anatomical segments much in the same way the Australians share out meat. If the hunter’s judgment is incorrect and there are poorly provisioned families in the camp, yet he did not distribute meat, then the gossip circuit is sure to start, aimed at the “insensitive” hunter. Sharing among the Nunamiut is always a combination of kin-based obligations and sensitivity to need. The importance of the latter factor is illustrated by the fact that in actual sharing situations the quantity distributed is at least partially adjusted to the size of the family receiving the food. If the hunter renders the judgment that there are kinsman in need, he then distributes butchered anatomical segments to others in terms of ego-centered (male) kin relationships. Each person who receives a share then generally gives at least part to his wife, and if she prepares it for a meal she may then share out again in terms of her kindred. Once meat is brought to the household, the woman takes charge of the cooking. It is also her function to present gifts of meat and fat to other persons. During times of stress her first responsibility beyond the members of her household, is to her parents and other lineal relatives. The second is to her siblings, especially if they reside close by or in the same village, the third is to her husband’s parents. Beyond these persons the woman may distribute meat to such close relatives as cousins, aunts, uncles, nieces and nephews of any member of the family, and to friends. No particular cut of meat or part of an animal is associated with or customarily reserved for any particular with or customarily reserved for any particular relative. A gift should be a good piece of meat with fat on it if possible. (Gubser 1965:81-82)

This important distinction between the sex-based responsibilities for sharing is linked among the Nunamiut with the nearly continuous fact of delayed consumption. That is, most of the time consumption is out of stores;4 hence, sharing is commonly a woman’s responsibility carried out at the time of meal preparation and/or serving. If males are hunting for food to go into stores, there is no sharing by the men beyond the division of game among the participants in the hunting party. Only in situations of direct consumption, judged appropriate in terms of general need, does a hunter conduct a meat distribution using male-centered kinship conventions on his return to camp. I must conclude that among the Nunamiut there is a strong sharing ethic. Given Gould’s “law” we should, therefore, expect them to have

a rigid butchering pattern. Alas, this expectation is not met. Long before I studied the Nunamiut, Gubser (1965:Sl) noted: There are many variations in butchering caribou; the determining factors are how the meat is to be transported and how long it is to be stored. If it is to be stored for a long time, the caribou is not skinned but only quartered, beheaded,

240

LEWIS R. BINFORD and eviscerated; then it is stacked up on a hillside. If. say, a woman is planning to carry the meat home by dog pack, she may even sever the meat from the legs and separate the leg bones from each other with her ulu.

My research confirmed

and elaborated the earlier observations

of Gubser:

What has been described is probably viewed as a great deal of variability and certainly a consideration of many different factors-means of transport, amount of meat available, outside temperature, and so on. I find it dificult to provide an “idealized” description of what may be called the “Nunamiut method of butchering” (Binford 1978:87).

These observations support the view that there is no necessary relationship between the strength of the sharing ethic and the lack of variability in butchering pattern. In all fairness to Gould he might argue that he never expected there to be a “necessary relationship” between the strength of the sharing ethic and the rigidity of the butchering pattern. He could point out that such a claim would have to be based on an argument from “ethnographic analogy.” Such an argument would, in this case, be justified by his experience with the Ngatatjara, from which he has told us there was a strong sharing ethic that explained the rigid butchering pattern. Gould has come out strongly against projecting ethnographic descriptions from one group of people as expections to be realized among another group. Why then, does the Nunamiut material bother Gould? If each society can be expected to culturally unique, exhibiting differing ideological justifications for manifest behavior, why is Gould not content to accept my descriptions of Nunamiut culture? Unfortunately, Gould does not tell us. As in previous discussions, we find that Gould is silent on an issue one might anticipate him to speak out about. As we have queried before, why? THE NUNAMIUT DATA AS A PARADIGMATIC

CHALLENGE

The Nunamiut study presented data which would have been distorted and misrepresented if the traditional interpretative conventions used by archaeologists had been applied to Nunamiut sites. In short, the variability manifest within the Nunamiut system challenged the normative expectation of internal homogeneity for behavior executed within a given culture. Gould’s reaction to this challenge is sketched out in the following: Perhaps the Nunamiut really are as “simple” as he suggests, in the sense that one can account simply for their behavior in relation to meat products on a circumstantial basis without recourse to more complex explanations involving ideas and symbols. (Gould and Watson 1982:367-368). On over 70 observed occasions in 1966-1970, the Western Desert Aborigines

of

BUTCHERING,

SHARING,

AND THE ARCHAEOLOGICAL

RECORD

241

Australia, . . . invariably divided macropods (mainly red kangaroo, Magalaeia rufa) into the same initial nine pieces, . . Clearly some kind of normative principle was at work in the Aborigine case. (Gould and Watson 1982:366) [Emphasis mine.]

In the above quotations we see the issue isolated; Gould expects there to be normative ideological guides to behavior. He expects there to be “mental templates” standing behind manifest behavior and he further expects “cultures” to be internally homogeneous in their repertoire of such templates. Gould observed a pattern, a very repetitive pattern, and assumes along with many others that Culture is patterned . . (therefore) the patterning which the archaeologist perceives in his material is a reflection of the patterning of the culture which produced it. (Deetz 1967:6, 7)

Gould is a traditional archaeologist. He assumes certain properties of culture-it is patterned-this patterning is then thought to be manifest in the regular behavior characteristic of the participants in a culturally organized unit. If we do not see regular, internally homogeneous behavior, something is clearly wrong. Gould sees regular behavior in the butchering behaviors of the Ngatatjara. Everything is as it should be. On the other hand, I reported extremely variable behavior as characteristic of participants in a single ethnic group. This is contrary to his paradigmatic expectations. Does he use such seeming contradictions as an opportunity to evaluate the degree to which his preconceived assumptions are productive and appropriate? No. What Gould does is try to cast doubt on the accuracy of my reporting: . . . as in much of Binford’s earlier work, there is a persistent and high level of ego-involvement that affects the presentation of his findings. In a case like this, where we have a book that will be referred to often by archaeologists in their efforts to explain their own fauna1 evidence and by ethnoarcheologists for comparison with findings for other contemporary human societies, this becomes a matter of some concern. (Gould 1979:739) I have commented

previously

on the tactics of strict empiricists:

In the strict empiricists’s view of the world, those who do not see nature in the same way as he does must be suffering from a lack of objectivity; all unbiased, objective observers should see the same things, since “nature does not lie.” Disagreements among observers are generally considered to derive from flaws in the character of at least some of the disputants. (Binford 1983b:372)

In this vein, Gould chooses to question my “objectivity” rather than his assumptions or his paradigmatic views as to what the world of culture should be like. Gould’s bias is further illustrated by the fact that he ignores my discussions of the variability at issue. I had long ago recognized that the behaviors I witnessed among the Nunamiut were different from others

242

LEWIS R. BINFORD

that I had experienced and differed from descriptive accounts of butchering in the literature. My response had been to accept the challenge, to try and explain the differences and similarities noted. For a traditional archaeologist, differences and similarities are not in need of explanation, they are simply accepted as measures of cultural differences and similarities. The latter, in turn, are to be understood in terms of differential histories. Not being a traditional archaeologist, I had offered suggestions as to how we might understand cultural variability itself: . . . when inputs to the system are few but large as in the case of the Nunamiut, each consumer unit is participating in homologous logistical, storage, and consumption sequences. Evaluations as to the utility of parts are made along these sequences. On the other hand, when inputs are small few decisions are made along an extended sequence related either to logistics, preservation potential, or consumption priorities. Instead, anatomical parts are differentially distributed among consumer units. In short, differential distributions occur at the locations of processing and differential consumption in the Nunamiut case, whereas among a group like the !Kung differentiation of parts among consumption units occurs by virtue of a distribution of parts to persons. differentially evaluated in terms of kinship, status, or other social idioms. (Binford 1978: 132)

It should be clear that I was well aware of the differences between the Nunamiut, who are logistically organized (see Binford 1980) and heavily dependent upon storage, and societies where inputs are small but fairly continuous, as in the cases of the !Kung and Gould’s Australians. I noted in my Nunamiut studies the following contrasts: In the Nunamiut case anatomical parts are differentially evaluated and this scale of evaluation is mapped onto different places and times evaluated in terms of transport consideration, anticipated differentials in storability, and so on. On the other hand, the !Kung most certainly have some similar understanding of the differential utility of anatomical parts but this is mapped onto persons differentially evaluated in such terms as kinship associations. (Binford 1978:133)

Instead of seeking an explanation for the differences, Gould would have us believe that either (a) the Nunamiut were misrepresented by me, or (b) they were “simple,” or (c) I failed to investigate the important symbolic determinants standing behind their use of meat. In other words, Gould is attempting to counter a warranting argument. My descriptions, which rendered it rational to expect cultural systems to be internally differentiated, is simply doubted. My argument that the archaeological record is not simply a material manifestation of mental templates, but instead the actual consequences of concrete behavior, is rejected. I have argued that many may attempt to execute wonderfully integrated plans but the contingencies of the natural world in which he lives frequently force modifications of these plans. The real-world contingencies, in terms of which men rationally modify their plans, ensure that the archaeological record will vary with these contingencies and not be a simple

BUTCHERING,

SHARING,

AND

THE

ARCHAEOLOGICAL

RECORD

243

material projection of the plans and mental images held in the minds of cultured men. It is as a counterclaim regarding the nature of the cultural world that Gould’s descriptions of Aborigine behavior take on significance. He claims that in Australia the Aborigines are not “simple” and, to understand their behavior, we must refer to the beliefs which guide and stand behind it. In turn, their behavior is thought to be “properly” normative and internally homogeneous. In short, the Nunamiut challenge to the traditional paradigm may be dismissed as somehow biased and poorly reported. The world which Gould wishes to see, the world of his paradigm, is restored. The Ngatatjara’s repetitive and redundant butchering pattern is to be taken as confirming the old view of culture. Gould’s “objective and unbiased” work should therefore save archaeology from demeaning (Gould 1980:250) itself as it sailed “perilously close to this particular intellectual iceberg” (Gould 1980:250)-the “new archaeology.” HOW IS THE ARCHAEOLOGICAL

RECORD

IMPLICATED?

One would expect that since all this argument is taking place among archaeologists there would be some discussion of the properties of the archaeological record. Gould is totally silent on this issue. This is not surprising since for traditional archaeologists the meaning of archaeological remains was never considered an issue, they are simply material manifestations of “culture.” The problem of what “culture” means is not something that can be solved through the study of material products, since it is a mental phenomenon. As Gould (1978:833) is fond of telling us, “just because the Western Desert aborigines do it this way today does not mean they did it that way in the past.” This is just another way of asserting that there appear to be no necessary linkages between material things and the meanings assigned to them by different peoples. This has always been a troublesome paradox arising from the traditionalists point of view. If one accepts the argument that patterning in material things derives from the ideas and beliefs which guided the observed order being discussed, e.g., the ideological patterns standing behind the material remains, then all one can do is systematize the materially manifest patterns. Then, by an appeal to one’s belief as to the nature of culture, the archaeologist may interpret his systematics in terms of the historical dynamics that are believed to condition the differential distribution of “culture” among different populations. This has always been the posture of traditional archaeology. Today there is an attempt to argue that one can, even must, understand the nature of the ancient ideas and beliefs assumed to guide the observable patterning in material things. This is the posture of Ian Hodder

244

LEWIS

R.

BINFORD

(1982a, 1982b) and many of his colleagues. Hodder’s approach is rationalized by his belief that there are certain characteristics of the human mind which condition ideologies in general. With a knowledge of these principles, used artfully in interpretation, one can reconstruct the “cultural context,” the actual world of symbols, which is said to have determined the reorganization of matter characterizing the different systems of ideas and values-culturesof the past. This posture differs from the mainstream of traditional archaeology. It is true that traditional archaeology shared with Hodder and other “contextualists” and “structuraiists” the belief that it was ideas and values which guided the actors’ behavior and hence “stood behind” observed artifact patterning. Hodder’s position, nevertheless, differs in that he denies that similar forms may refer to similar ideas. It was, of course, the latter proposition that made possible historical interpretations from patterned similarities. Hodder argues that any form or artifactual pattern can only be understood in terms of the ideological frame work which integrated cognitively the actions and lives of the ancient peoples responsible for it. For classic traditionalists, ethnoarchaeology can play only a very limited role. If one can justify the argument that there had been no relevant culture change between the archaeological past and the social present, then one can study a modern society for purposes of “fleshing out” a picture of the past. This is the strategy advocated by Oswalt (1974) and described by Gould as a “continuous model” (Gould 1974:39) of ethnoarchaeological analogy. The reason for studying a living group is then to find out about the “culture” which served to motivate and condition the way the “makers” of the artifacts designed their products, restructured their environments, and conceptually integrated action with philosophy. Under this approach one never really studies the formation processes of the archaeological record, only the rationalizations of the actors for their participation in the processes which might condition an archaeological record. As I suggested earlier, Gould does not report on the archaeological record as such. He only reports on the dynamics as played out by a series of actors, in the same way an ethnographer might report on experiences (see particularly Gould 1980:6-28). It would appear that, for Gould, documenting the relationships between actions and ideology proves the reality of his paradigm. The paradigmatic assumptions are therefore vindicated and may be used again as conventions for assigning meaning to the archaeological record. In his terms, we may “derive predictions about the society’s discards” (Gould 1980:113). If his view of culture is correct, and the archaeological record is simply a material manifestation of the ideas and values of ancient men, then, given an understanding of “culture,” Gould can “derive predictions” about the properties of the archaeological record. Given this perspective, eth-

BUTCHERING,

SHARING,

AND

THE

ARCHAEOLOGICAL

RECORD

245

nographic experience would be cited in justifying assumptions about culture, while archaeological remains would be considered simple patterned reflections of culture.

Ethnoarchaeological approaches in the hands of Hodder and some of his colleagues are more directed toward the archaeological record, but only from an advocacy perspective. That is, Hodder’s recent works (1982a, 1982b) are both organized to illustrate, using ethnographic examples, how patterning in material remains is unintelligible in the absence of a knowledge of the cognitive ideology of the actors that produced it. Once again, the research is not directed toward understanding how patterning in archaeological materials comes into being. It is designed to demonstrate that patterning is dependent upon ideological or cognitive schema. Like Gould, we are generally left to infer the archaeological patterns which are thought to flow from the cognitive schemes “ethnographically” presented. Rarely is any attempt made to document the patterns actually generated. We are expected to “derive predictions” about the archaeological record from verbal descriptions of ideological “guides” to behavior thought to actually control action. Once again, what passes for ethnoarchaeology is not an attempt to explain an archaeological record or develop reliable methods for understanding its formation. Ethnoarchaeology, as practiced in this context, simply becomes studies conducted among living peoples for use as warranting arguments for believing certain things about culture. If one accepts this perspective, or paradigm, interpretive conventions to be used in assigning meaning to the archaeological record simply flow from the internal logic of the paradigm adopted. Under these approaches, the archaeological record is not investigated, and its properties are not explained. Instead, abstracted properties of the record are simply interpreted according to the beliefs of the investigator. Such accommodative arguments serve as middlerange theories for traditional archaeologists. Given the cultural paradigm of your choice, the archaeological record is traditionally linked to this paradigm by the assumption that the patterning in the record is a material manifestation of culture. Dunnell (1971: 122) pointed out some years ago that for traditional archaeologists culture “is a means of explanation.” My approach has been different. I have argued that the challenge to scientific archaeology is to investigate the archaeological record. It is only with an understanding of how this record is formed that we will be able to infer accurately properties of interest about the past, including properties of past cultural systems. Given this perspective, I am interested in how sharing is manifested in the archaeological record. In addition, given the claims by Gould about rigidity in butchering procedures, I am interested in how such procedures might affect patterning in the archaeological record.

246

LEWIS

EVALUATING

The Nunamiut

R.

BINFORD

GOULD’S ARGUMENT: ON SHARING

CONTROL

DATA

Data

While I worked among the Nunamiut Eskimo I had the opportunity to observe many acts of sharing (see Binford 1978:132-133, 142, 456-457, 471-472, 478). However, the archaeological “signatures” derived from these acts were more difficult to document. Unambiguous data referable to sharing were obtained from only two sites, Ingstad (Binford 1978:323) and Palangana (Binford 1978:437). Table 1 summarizes the anatomical part frequencies for Dal1 sheep recovered from different consumer units within each site. On the Ingstad site there were known to have been four sheep killed during the course of the occupation. Two were introduced by unidentified hunters believed to have been of the Tulugak Lake band while the other two sheep were killed by known individuals. The bones from the houses of the two known hunters are summarized in columns 1 and 2 while the fauna1 remains from the other Tulugak band families are tabulated in columns 3 and 4. Sheep remains from all the residential sites of the Killik River band’s families ( a visiting band at this site) is summarized in columns 5 and 6. Inspection of the table illustrates the first important point, namely, that the anatomical parts common at the residence of one family are generally low or absent on the locations of other consumer units within the same site. This condition derives from the fact that different anatomical segments are the units shared out by hunters. This means that the anatomical units which are represented by only one element from a single individual, such as the skull or the neck, if present within one site will only be present at the residence of one of the consumer units within the site. On the other hand, parts which may be broken down into multiple segments, such as ribs, may tend to exhibit overlapping distributions among a set of consumer locations within a single community. In these examples, data were selected because they represented rarely killed animals which are almost always shared among the Nunamiut. Since only a few such animals are represented, the mutally exclusive form of anatomical-part patterning expected among separate consumer units within a single community is clearly visible. We may reason that this pattern may tend to break down if the duration of the occupation is increased andlor there are are different hunters representing most of the consumer units present within the community who regularly distribute meat during the course of the occupation. How the pattern diverges from the one illustrated in Table 1 will depend upon variability in the hunting success of hunters occupying different nodes within the kinship matrix

N 5

a Information b Information

taken taken

from from

Binford Binford

(1978: (1978:

50.0 50.0 loo.0 50.0 50.0 50.0 0 0 0 50.0 50.0 0 50.0 100.0 50.0 50.0 25.0 25.0 25.0 0 0 75.0 75.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0 25.0

1.0 1.0 2.0 1.o 1.0 1.0 0 0 0 1.0 1.0 0 1.0 2.0 1.0 1.0 0.5 0.5 0.5 0 0 1.5 1.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5

0 0 0 0 0 0 1.5 1.5 2.5 2.7 0.9 3.0 1.5 2.0 1.0 0 0 0 0 2.5 2.0 1.0 2.0 1.0 0.5 1.0 0.5 1.0 0 0 0

MAU

%

MAU

Table Table

0 0 0 0 0 0 50.0 50.0 83.3 90.0 30.0 100.0 50.0 66.6 33.3 0 0 0 0 83.3 66.6 33.3 66.6 33.3 16.6 33.3 16.6 33.3 0 0 0

%

Sheep bones at tents of other Tblugakmiut families

Ingstad

site”

0 0 0 0 0 0.5 0.5 0.5 0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 0.5 0.5

MAU

20-25). 17-20).

0 0 0 0 0 100.0 100.0 100.0 100.0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 100.0 100.0 100.0

%

-

DIFFERENT

Sheep bones at tents of Killikmiut families

PARTS AMONG

6.21, p. 323. columns 8.4, p. 437, columns

OF ANATOMICAL

Sheep bones at tents of two lblugakmiut hunters

FREQUENCIES

1

1.0 1.0 2.0 1.0 1.0 1.5 2.0 2.0 2.5 2.7 1.9 3.0 2.5 2.0 2.0 1.0 0.5 0.5 0.5 2.5 2.0 2.5 3.5 1.5 1.0 1.5 1.0 1.5 1.0 1.0 1.0

Man 29.0 29.0 57.0 29.0 29.0 43.0 57.0 57.0 71.0 77.0 54.0 86.0 71.0 57.0 57.0 29.0 14.0 14.0 14.0 71.0 57.0 71.0 100.0 43.0 29.0 43.0 29.0 43.0 29.0 29.0 29.0

%

RESIDENCE

TABLE

0.5 0 0 1.0 1.0 1.5 0.5 0.5 1.0 0.23 0 0 0 0 0 0 0 0 0 0.5 0.5 0 0 0 0 0 0 0 0 0 0

MAU 33.0 0 0 66.0 66.0 100.0 33.0 33.0 66.0 15.3 0 0 0 0 0 0 0 0 0 33.0 33.0 0 0 0 0 0 0 0 0 0 0

%

Sheep remains, House 1

GROUPS DERIVED FROM

0.5 1.o 0.5 0 0 0 0 0 0 0 0 0.5 0.5 0 0 0 0 0 0 0 0 0 0.5 1.o 1.0 1.0 1.0 1.o 0 0 0

MAU 50.0 loo.0 50.0 0 0 0 0 0 0 0 0 50.0 50.0 0 0 0 0 0 0 0 0 0 50.0 loo.0 100.0 100.0 100.0 loo.0 0 0 0

%

sit@

DALL

Sheep remains, House 2

Palangana

“SHARED”

-

Sheep

1.0 1.0 0.5 1.0 1.0 1.5 0.5 0.5 1.0 0.2 0 0.5 0.5 0 0 0 0 0 0 0.5 0.5 0 0.5 1.0 1.0 1.0 1.0 1.0 0 0 0

MAU

SHEEP

67.0 67.0 33.0 67.0 67.0 100.0 33.0 33.0 67.0 13.0 0.0 33.0 33.0 0.0 0.0 0.0 0.0 0.0 0.0 33.0 33.0 0.0 33.0 67.0 67.0 67.0 67.0 67.0 0.0 0.0 0.0

%

remains, total

248

LEWIS

R.

BINFORD

relative to those receiving shares. I stated the situation with regard to sharing and patterning in the archaeological record some years ago: Food distributions are activities conducted in terms of a set of cultural conventions, a mental template if you will, of the proper pattern of distribution in relation to a set of cultural conventions that evaluate people. Since the activities in which such conventions are employed are egocentric, and despite the face that all members of a group may share common cultural conventions, we can expect two things. The first of these is that the patterning manifested in the fauna1 remains will vary with the number of persons making distributions. In addition, patterns may change even if only one person is making distributions relative to any changes that may occur in the composition of the group receiving food. The structure of this situation is such that redundancy in patterning, or a lack thereof among segments of a community, would bear no relationship to the degree that they shared a comon convention for conducting meat distribution! This example illustrates the differences between patterning that results from the execution of a strategy and the conventions used in [guiding] its execution. The archaeological remains derive from the conditions of execution. (Binford 1978:471-472)

The San Data

Gould might object to my use of Nunamiut observations, given his claims for my overly involved state of mind, so it is important to establish the fact of the signature patterning associated with sharing with observations made by others. Yellen collected an important and provocative body of data from San camps during his fieldwork in the Kalahari. I have summarized from his maps, descriptions (Yellen 1977a), and inventories (Yellen 1977b) data referring to Camp 10 (//Gakwe/Dwa). This site (Fig. 1) was occupied for 12 days by 13 adults and 11 children. During that time two gemsbok were killed and parts introduced to the site for distribution by sharing: . . . all men go out after the gemsbok that = toma had wounded the day before. They find it dead and butcher it, eating all the marrow from the cannon bones, some ribs, the liver and the head. The rest of the animal is carried back to camp and consumed there. Later during their stay, = toma wounds another young male gemsbok at !kau !kasi. The men go after it the next day and find it alive but unable to run, and they kill it with a spear. At the kill site, the men eat the marrow from the cannon bones, some ribs and part of the liver. The skin and horns are left at the kill, and the remainder of the animal is carried back to camp. (Yellen 1977b:207) Of course, the above description has only reference to the behavior of the hunters and the particular events taking place before the parts of the introduced animals were actually shared. In another article Yellen (1977a) has an excellent description of the meat sharing within a residential camp. This takes place after animals are introduced to the site. Yellen generalizes that there is commonly a three-stage pattern of sharing associated with meat distributions:

BUTCHERING,

SHARING,

AND THE ARCHAEOLOGICAL

RECORD

249

FIG. 1. Site plan, !Kung Camp 10 (from Yellen 1977a). Meat moves from the hunters and carriers upward through the kinship network to parents and in-laws, and then outward and downward again. When sharing is viewed in a specific instance, such as this one, it can be clearly seen why hard and fast distribution rules cannot apply; each kill must be considered as a specific case; the relationships between families who happen to compose a group at that moment and family size must both be taken into account. (Yellen 1977b:289)

It is clear that Yellen is warning the student that the archaeological record reflects the actual behavior as executed, not the normative “culture” which might be thought of as “counciling” such behavior. Ironically Gould, himself, has supplied descriptive material with similar implications. He notes that with smaller kangaroo and euros there is a two-stage pattern of sharing among Ngatatjara hunters (see Gould 196758). During his fieldwork Gould was able to trace the acts of sharing in six observed episodes and found that the final size of portions consumed after the two stages of dismemberment and sharing was remarkably the same, averaging 1.23 pounds per share (Gould 196759). What this means is the the fauna1 remains, as finally introduced into the archaeological record at a residential camp, will most likely be deposited (a) in units of consumption, not units of initial butchering and sharing, (b) in forms resulting from the modifications made during preparation of parts for consumption

250

LEWIS R. BINFORD

(cooking, drying, pounding, etc.), and (c) distributed and associated spatially in terms of disposal modes relative to the locations of consumption.

This last point has been astutely made by Yellen as follows: patterning is related to a single variable based Thus in [many] studies . usually on either primary butchering or tool use. It is instructive that, in the !Kung case, these factors have . . little effect on the final form of the fauna1 remains, and that far more important are the last stages of butchering for cooking and direct consumption and selective destruction by natural processes after the bone and bone fragments have been discarded. (Yellen 1977b:327)

Yellen’s experiences are in essential agreement with my own when investigating formation processes in ethnographically controlled situations. Keeping this in mind, it is important to return to a consideration of the control data collected by Yellen in the context of known sharing behavior. Table 2 summarizes the frequencies of anatomically identified bone fragments listed by Yellen (1977a:207-210) from each of the seven residential locations used by the occupants of Camp 10. Figure 2 summarizes the data of Table 2 graphically. What is strikingly illustrated is the same structure of variability as was documented among the Nunamiut. The anatomical parts from the two gemsbok are differentially distributed among the separate residential units within the community. The next characteristic to notice is that parts which are capable of easy segmentation are those which are most ubiquitously distributed. Vertebrae appear in six of the seven house sites, ribs occur in all, and pelvic parts were recovered from five of the seven residences. In marked contrast, the bones of the upper legs-femur, humerus, scapula-appear in three residences only for each bone. In the case of Yellen’s data, the head and lower legs were not reported to have been returned to the site; nevertheless, traces of these parts are present. It is not clear from the ethnography how they were introduced. I think it is germane, however, that the lower leg parts occur at the residences of the men who killed the animals or at the location of the storage rack. In my experience with the Nunamiut, parts which are “consumed” by the hunting party, i.e., distributed for consumption among the field party in hunting camps or at kill sites, etc., may occasionally be introduced to the residential site by hunters who did not consume all their shares while in the field. These parts are considered as already allocated to the hunter even though the bulk of the meat he transported to the camp may be distributed to kinsmen. Put another way, parts which are considered “field food” and already allocated by whatever convention to the members of the field party may be returned as part of the hunter’s share not subject to further distribution within the camp. Such parts are most commonly lower legs as well as head and neck elements, with the occasional addition of some ribs and organ meats. Among the Australians,

2

.oo

0

16

Total

7 0 0 2 0 0

1 1

.06 .25 .06 .06 .44 .oo .oo .13 .oo .oo

%

1 4

No.

Cranium Vertebrae Sternum Ribs Pelvis Femur Scapula Humerus Radius-ulna Extermity Unidentified fragment

Fragment

1

FREQUENCIES

12

0

2 3 0 0 0 4 0

3

No.

2

OF IDENTIFIED

.oo

.17 .25 .oo .oo .oo .34 .oo

.25

%

15

0

6 2 2 4 0 0 1

0

No.

GEMSBOK LOCATIONS

TABLE

2

3

.oo

.40 .13 .13 .27 .oo .oo .07

.oo

%

25

0

.oo

0

15

0

0 0

.04

.oo .08

1

7

No.

0 2

.04 .20

%

3 0 0 5

4

area

.40 .04 .04 .16

10 1 1 4

1 5

No.

Hut

5

.oo

.oo .OO

.oo

.20 .oo .oo .38

.47

%

19

0

0 0

1

5 1 0 0

12

No.

6

.oo

.26 .05 .oo .oo .05 .oo .oo

.63

%

16

1

1 0 3 3 2 0 0

6

No.

BONE FRAGMENTS ASSXIATED WITH EACH OF THE RESIDENTIAL USED BY THE OCCUPANTS OF CAMP 10

7

.06

46 .oo .19 .I9 .13 .oo .oo

.38

%

118

1

2 37 1 28 14 6 16 6 4 3

Total

252

LEWIS R. BINFORD HUT NUMBER CRANIUM

LOWER LEG Number of Adults Number of Children Number Bones Hunter's of Huts

t $ 17 .

+ 2

A 15

:9

Scale 0

20

I" Percent 40 60 80

; 16

: 0 12

: 15 .

+ z 25

100

FIG. 2. Distribution of fauna1 fragments among !Kung residences in !Kung Camp 10 (data from Yellen 1977a:206-211).

the parts of the kangaroo that are sometimes consumed in the field are the lower rear legs and the tail. In spite of Gould’s skepticism, the pattern of consumption or disposal of marginal parts in the field is a general characteristic among modern hunters, commonly varying with the size of the prey relative to the transport potential of the field party. I mentioned earlier that meat distributions among the Nunamiut were conducted in terms of kinship conventions but at the same time involved consideratious of need. For instance, families positioned at roughly similar kinship distances from the person making the distribution would receive different amounts of meat depending upon the size of the families represented. Figure 3 illustrates the relationship between the number of bones and the number of consumers on each residental location. It is clear that there is a generally strong positive linear relationship between the size of the shares as measured by numbers of bone and the size of the consumer unit. The two exceptions are the families of the hunters, both of which exhibit anomalously how values for numbers of bones. There is in the literature a kind of “myth” that the sharing ethic is so strong that successful hunters give away all their meat and, hence, tend to eat less well than the dependents of the less skillful hunters. I can understand how some ethnographers might be deceived into thinking this but my experience leads to opposite conclusions. Successful hunters’ families eat better, not only in times of food scarcity, but all the time. The presence of high frequencies of ribs, pelvic parts, and femurs at hut 3 (Fig. 2)-the hunter’s residence -shows that high-quality food was common at the hunters’ hearths. If you add to this the common occurrence of the hunter getting organ meats (no included bones), the lower frequency of bones per number of consumers can be taken to indicate a dominance of higher-quality meat parts introduced to the hunter’s camp. Comparing the San data with that for the Nunamiut, we note the dom-

BUTCHERING,

0

SHARING,

1

AND THE ARCHAEOLOGICAL

RECORD

2

6

3

4

5

253

NUMBERS OF PERSONS IN RESIDENCE

FIG. 3. Number of bone fragments per hut for hunter and nonhunter residences in !Kung Camp 10 (data from Yellen 1977a:206-211).

inance of anatomical parts with a high utility value (see Bindord 1978: 15) on residential sites in both cases. The abandonment or consumption of parts of marginal utility by members of the hunting parties at kill or field butchering locations was also noted among the Nunamiut. Finally, there is a extreme assemblage variability among the residential units as also noted among the Nunamiut. This variability does not arise from the fact that each family in the community has a different culture, or practices a different primary butchering strategy, or differentially adheres to a sharing ethic. It arises from the fact that sharing may be guided by a strong ethic and even by normative rules as to how one should share meat parts; but, nevertheless, the archaeological record is not isomorphic with regard to either the rigidity or the form of the guiding ethics or to the form and sequence of initial butchering activities. The situation is perhaps best explicated by reference to my thoughts after considering this problem as experienced among the Nunamiut Eskimo: This is an interesting situation in terms of the archaeologist’s notions of linkage between culture and the structure of archaeological remains. Food distributions are activities conducted in terms of a set of cultural conventions, a mental template if you will, of the proper pattern of distribution in relation to a set of. . . conventions that evaluate people. Since the activities in which such conventions are employed are egocentric, and despite the fact that all members of a group may share common cultural conventions, we can expect two things. The first of these

254

LEWIS R. BINFORD is that the patterning manifested in the fauna1 remains will vary with the number of persons making distributions. In addition, patterns may change even if only one person is making distributions relative to changes that may occur in the composition of the group receiving food. The structure of this situation is such that redundancy in patterning, or a lack thereof among segments of a community, would bear no relationship to the degree that they shared a common convention for conducting meat distribution! This example illustrates the differences between patterning that results from the execution of a strategy [versus] the conventions [guiding] its execution. The archaeological remains derive from the conditions of execution . These are situational conditions . (Binford 1978:471472)

Put another way, there is clear interlocational variability in fauna1 assemblages generated by the San. As in the Nunamiut case the people producing this within-system variability share a common culture. In addition, the San appear to practice a “rigid” primary butchering procedure, yet marked internal or “functional” variability is produced among their sites. I must conclude that at least two of Gould’s criteria for dismissing by Nunamiut-based arguments are rendered suspect by the San data. (1) I did not collect the San data; hence, my alleged biases cannot be said to explain the facts. Yellen can hardly be accused of trying to prove my point of view while he was in the field. If anything, his bias was in the reverse direction. (2) The San can hardly be described as either lacking a sharing ethic, being “simple,” or having been poorly investigated in terms of their ideological and belief systems. Gould’s objections to my Nunamiut-based arguments are countered by the San data. There is, however, an additional posture which Gould might assume. He could insist that the Australian materials be considered in terms of themselves. His experiences among the Australians could be cited as providing an empirical case to the contrary and therefore a “spoiler” for generalizations projected from other cases. As I have suggested in this paper, Gould has not thus far made such an argument since he has not demonstrated any archaeological patterning said to be a result of the alleged rigidity or strong commitment to a sharing ethic, which he ascribes to the Australians. I am quite confident that such an attempt would fail, since the patterning among fauna1 remains within archaeological sites is much more likely to be referable to units of consumption, not units of initial butchering and sharing. This situation is further complicated by patterns of breakage and dispersion arising from cooking, drying, pounding, etc., during processing and consumption, as well as by the disposal modes common at the site. These factors all ensure that explanations other than references to idealized “rules” for sharing would be required for understanding the archaeological record at any given site. At the level of between-site comparisons, the level which Gould criticizes in my Nunamiut analysis, a regular adherence to sharing ethics results in

BUTCHERING,

SHARING,

AND THE ARCHAEOLOGICAL

RECORD

255

considerable intersite fauna1 variability and not site-to-site similarities or cultural homogeneity imagined by Gould and other traditionalists. Guessing the nature of the archaeological record from abstracted ethnographic observations is as least as dangerous as guessing the ethnographic past from abstracted archaeological observations. Ironically, it is the intersite variability, the lack of homogeneity, and the lack of formal similarities among a set of culturally identical units that provides the information which makes possible the recognition of patterning indicative of the butchering pattern and perhaps also of sharing. This fact is well illustrated by the multivariate analysis of the Nunamiut faunas from a number of functionally different sites (Binford 1981:93). The analysis isolated the basic primary butchering units which were the units generally introduced to, or abandoned on, sites. A similar analysis of numerous residential sites where the major source of intersite variability was the differential distribution of anatomical parts among sharing consumer units, as is documented here for both the Nunamiut and the San, should yield the patterned manner of dismemberment and provide some clues as to how rigidly butchering procedures had been adhered to by the ancients. What would not be clear, however, would be whether interassemblage variability was a consequence of functional differences among sites, or if it arose, as has been illustrated in this paper, from sharing anatomical parts of differing size and abundance among independent consumer units. The challenge to the archaeologist is to develop ways of reducing ambiguity and increasing the accuracy with which we may analytically identify past causes of variability in the archaeological record. Gould’s strategy of doubting my descriptions of variability and suggesting that the Australians were not as “simple” or perhaps were more “cultured” than the Nunamiut since they appeared to have a more ideologically regulated butchering pattern simply misses the point of archaeological investigations and misleads the unsuspecting as to the nature of archaeological research, particularly the role of ethnographic information in such research. NOTES ’ This approach is logical only if Gould accepts the Nunamiut descriptions as lawlike propositions. They were never presented as such. 2 Gould claims that he first heard this phrase from James O’Connell in Canberra (Gould 1980:253). However, his first use of the phrase is in a review of Binford’s (1978: 131) book, in which it is used repeatedly. r Many writers, such as Hodder (1982a, 1982b), condemn cross-cultural comparisons. They claim that the meanings of similar things must be different in different cultural contexts. Gould probably agrees on this point. 4 There has been considerable confusion in the recent literature regarding factors con-

256

LEWIS R. BINFORD

ditioning storage and its role as a conditioner of complex social roles (see Meillassoux 1973; Testart 1982; Woodburn 1982; Weissner 1982; versus Binford 1980; lngold 1983).

REFERENCES Binford, L. R. 1978 Nunamiut ethnoarchaeology. Academic Press, New York. 1980 Willow smoke and dogs’ tails: Hunter-gatherer settlement systems and archaeological site formation. American Antiquity. 45:4-20. 1981 Bones: ancient men and modern myths. Academic Press, New York. 1983a Working at archaeology. Academic Press, New York. 1983b Reply to More on the Mousterian: flaked bone from Cueva Morin, by L. G. Freeman. Current Anthropology. 24~372-377. 1984 Brand “X” versus the recommended product. American Antiquity, in press. Binford, L. R., and W. J. Chasko 1976 Nunamiut demographic history: a provocative case. In Demographic anthropology, edited by B. W. &brow, pp. 63- 144. Univ. of New Mexico Press, Albuquerque. Deetz, J. 1967 Invitation to archaeology. Natural History Press, Garden City. Dunnell, R. C. 1971 Systematics in prehistory. Free Press, New York. Gould, R. A. 1967 Notes on hunting, butchering, and sharing of game among the Ngatatjara and their neighbors in the West Australian Desert. Kroeber Anthropological Society Papers 3641-66. 1974 Some current problems in ethnoarchaeology. In Ethnoarchaeology, edited by C. B. Donnan and C. W. Clewlow, pp. 29-48. Monograph IV, Institute of Archaeology, University of California, Los Angeles. 1978 The anthropology of human residues. American Anthropologist. 80:815-835. 1979 Caribou hunters-Review of Nunamiut ethnoarchaeology, by L. R. Binford. Science (Washington, D.C.) 204(4394):737-739. 1980 Living archaeology. Cambridge Univ. Press, New York. Gould, R. A., and P. J. Watson 1982 A dialogue on the meaning and use of analogy in ethnoarchaeological reasoning. Joi;rnal of Anthropological Archaeology. 1:355-381. Gubser, N. J. 1965 The Nunamiut Eskimos: hunters of caribou. Yale University Press, New Haven. Hodder, I. 1982a Symbols in action: ethnoarchaeological studies of material culture. Cambridge Univ. Press, Cambridge. 1982b The present past: An introduction to anthropology for archaeologists. B. T. Batsford Ltd, London. lngold, T. 1983 The significance of storage in hunting societies. Man n.s. l&553:571. Isaac, G. Ll. 1978a The food sharing behavior of protohuman hominids. Scientific American 238(4):90- 106. 1978b Food sharing and human evolution: Archaeological evidence from the PlioPleistocene of East Africa. Journal of Anthropological Research. 34:31 l-325.

BUTCHERING, 1980

SHARING,

AND THE ARCHAEOLOGICAL

RECORD

257

Casting the net wide; a review of archaeological evidence for early hominid land-use and ecological relations. In Current argument on early man, edited by L. K. Konigsson, pp. 226-253. Pergamon, Oxford. Meillassoux, C. 1973 On the mode of production of the hunting band. French perspectives in African studies, edited by P. Alexandre, pp. 187-204. Oxford Univ. Press, Oxford. Oswalt, W. H. 1974 Ethnoarchaeology. In Ethnoarchaeology, edited by C. B. Donnan and C. W. Clewlow, pp. 3-14. Monograph IV, Institute of Archaeology, University of California, Los Angeles. Speth, J. D. 1983 Bison kills and bone counts: decision-making by ancient hunters. Univ. of Chicago Press, Chicago. Testart, A. 1982 The significance of food storage among hunter-gatherers: residence patterns, population densities, and social inequalities. Current Anthropology 23:523-537. Woodbum, J. 1982 Egalitarian societies. Man n.s. 17:431-451. Wiessner, P. 1982 Beyond willow smoke and dogs’ tails: a comment on Binford’s analysis of the hunter-gatherer settlement systems. American Antiquity. 47:171-178. Yellen, J. E. 1977a Cultural patterning in fauna1 remains: Evidence from the !Kung bushman. In Experimental archeology, edited by D. Ingersoll, J. E. Yellen, and W. MacDonald, pp. 271-331. Columbia Univ. Press, New York. 1977b Archaeological approaches to the present: models for reconstructing the past. Academic Press, New York.

Guts & Brains

28-03-2007

15:12

Pagina 185

The Diet of Early Hominins: Some Things We Need to Know before “Reading” the Menu from the Archaeological Record Lewis R. Binford Dept. of Anthropology Southern Methodist University Dallas, Texas, USA

Introduction What should we know when we research phenomena from the past for clues to identifying and/or mounting robust arguments regarding the type of diet of early hominins? A scientist should demand that we begin by discussing the “initial conditions” that may be assumed to be both germane and common to the phenomena (i.e. hominins) before and after the appearance of stone tools, with their impact on diet. On the topic of diet, one important question we must consider is, how have the diets of our biological ancestors and predecessors changed and/or diversified during the course of human evolution? For instance, today, humans eat a wide array of species and parts thereof, and this huge inventory is geographically partitioned, ethnically differentiated, and found to be variable even at the individual level of comparison. I think that we not only want to address the issue of the “normative” diet through time, but what generally conditioned the patterned geographical and/or seasonal differentiation in dietary habits, as this may be manifested in archaeological site assemblage variability.

Part I. The issue of initial conditions Before addressing these fascinating issues, we must look into the past to investigate the initial conditions at the time stone tools became common adjuncts in food procurement behaviour. I will assume that the hominins at that time were biologically omnivores, that is that they were capable of ingesting and digesting both animal and plant parts. Put another way, I will assume (1) that they were feeding omnivores (2) subsisting upon a diverse array of natural foods obtained directly from both the plant and animal kingdoms. It should be pointed out that omni-

185

Guts & Brains

28-03-2007

15:12

Pagina 186

vores are relatively rare in nature and that a similar condition was likely to have been so in the African ancient past. A recent comparative study of modern tropical rainforest African and neotropical faunas (Fa and Purvis, 1997: 100-101) confirms that feeding specialists outnumber omnivores by a ratio of 2 to 1 among tropical rainforest species, excluding primates. On the other hand, among primates, omnivores outnumber feeding specialists at a ratio of five to one. Omnivores are thus exceptionally common among primates, a point made some time ago (Harding, 1981) for very different reasons. Of the larger body sized non-human primates, baboons are the most omnivorous, with the chimpanzee and gorilla practising a reduced omnivorous feeding pattern. One can say that feeding on both plants and animals is a strong part of our phylogenetic heritage, making it not unlikely that the earliest hominins in our lineage were omnivores at some unknown but probably variable plant to animal ratio, depending upon the environmental setting. Technically, they were hunter-gatherers. If this is accurate, then at least part of what we want to know is, what conditioned a patterned diversification of diets across space and through time among hominins? The latter features could stand as the initial patterned dietary conditions, out of which our species of hunter-gatherers (Homo sapiens sapiens) appeared in the ancient past. Given the historical trends in body size and brain development, I think that most would agree that there should be a parallel, a trend towards the consumption of higher-quality foods, which would primarily be fruits, insect by-products such as honey, and animal products such as marrow, blood and meat.

Part II. Comparing “us” to “them”, a framework for learning It’s a challenge to fill in the information gap between the vague assumptions of the initial conditions and the actual patterns of dietary diversification and changes that are characteristic of modern humans. Just as it would be nice to know the actual details of hominin diet that we begin with as initial conditions, we must also seek to understand the factors conditioning dietary variability among modern human hunter-gatherers. Such knowledge is important for reasoning about the factors that may have inured dietary diversity in the past. This insight is critical to any evaluation of the differences between “us” and “them”. Let’s begin with two kinds of data: information regarding dietary differences related to the environment, and information regarding diverse tactical strategies facilitating these feeding habits among ethnographically documented huntergatherers. With such information, we may then seek to recognize deviations from the documented patterns as indicative of differences between behaviours in the past and those of known modern hunter-gatherers. In addition, such a strategy

186

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 187

may yield clues to the ecological relationships differentially conditioning our diet, as hunter-gatherers, and at least introduce the issue of whether similar factors were operating upon ancient hominin populations. The second important domain of needed information focuses our concerns upon the species available versus the species (and/or biomes) actually exploited on a regular basis by ancient populations. This focus moves our inquiry into the domain of strategies and tactics employed by the ancients for obtaining food, which also implicates the feeding pattern, the organization of labour when feeding, as well as processing products for transport. Finally, we want to know about variability in the techniques employed when preparing food for consumption. As an initial probe into the above issues, I will explore the broad dietary differences documented among near modern hunter-gatherer peoples. I will focus on recognizing any accompanying consequences for time trends as well as organized geographical variability that may be helpful when considering the past. The property space in Figures 1 and 2 is defined by effective temperature (ET) displayed on the y-axis and the log 10 value of net above ground productivity (LNAGP) shown on the x-axis. Effective temperature is an estimate of the warmth at the beginning and the end of the growing season in degrees centigrade (Bin-

World Weather Station Distribution 26 25 24 23 22 21 20

Bailey’s Effective Temperature (ET)

19 18 12 Mo Growing Season

17 16 15

Plant Storage

14 13

Plant Threshold

12 11

Sub-Polar Bottleneck

10 9 8 0

1

2

3

4

Log10 Net Above Ground Productivity (LNAGP)

Fig 1. Global distribution of effective temperature measures relative to net above ground productivity (NAGP) values (see text for explanation).

the diet of early hominins

187

Guts & Brains

28-03-2007

15:12

Pagina 188

ford, 2001a: 58-59). It is a more informative measure of geographical variability in warmth from solar radiation than many other familiar measures, such as latitude, which is not variable with altitude as regards temperature. ET was developed by Bailey (1960) as an analytical measure of warmth particularly germane to differential productivity in the plant community. An ET value of 8.0°C represents the coldest place on earth. A value of roughly 10 suggests the boundary between boreal forest and the polar regions. A value of 14 approximates the earth’s mean biological temperature, while the values that exceed 18 mark the transition to regions with a 12-month growing season, i.e. the tropical and equatorial zones. Net above ground productivity is a measure (in grams per square metre of land surface) of the amount of new cell life produced by the plant community annually (Binford, 2001a: 78-80). It is a measure of biotic production and varies primarily in response to annual differences in the amounts of solar radiation and rainfall delivered to a given location. Figure 1 displays the global distribution of effective temperature measures relative to NAGP values. One can think of this distribution as describing the terrestrial biological variability of the earth as regards plant productivity. It is like a map of terrestrial plant productive differentials over the earth. Keeping this pattern in mind, we may now turn to Figure 2.

26

Contemporary Hunter-Gatherers Non-packed Cases Only

25 24 23 22 42

Bailey's Effective Temperature (ET)

21 20 19

12 mo. grow

18 17 16 15

Plant Storage

14

SUBSP

Plant Threshold

13 12 11

Sub-Polar Bottleneck

fishing gathering

10 9 8 1.5

hunting 2.0

2.5

3.0

3.5

4.0

Log 10 Net Above Ground Productivity (LNAGP)

Fig. 2 Environmental distribution of near contemporary hunter-gatherers with population densities less than 9.098 persons per one hundred square kilometres.

188

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 189

Figure 2 shows the environmental distribution of near contemporary huntergatherers who have population densities less than 9.098 persons per one hundred square kilometres. The latter value was found to mark a threshold (between hunter-gatherers who strategically rely on residential mobility to insure their wellbeing and those who have intensified their land use through specialization and/or expanding the diversity of their subsistence base). Perhaps the most obvious feature of this distribution is the strong concentration of cases that are dominantly dependent upon aquatic resources in environments colder than the Cool Temperate zone (i.e. concentrated in settings colder than the “plant threshold”). The latter threshold was found empirically to mark the coolest setting below which few, if any, cases are found that were primarily dependent upon terrestrial plant resources. In short, all but one non-packed hunter-gatherer case primarily dependent upon aquatic resources are found in settings colder than the ET value of 12.75°C. On the other hand, in settings warmer than the 12.75°C are found all of the documented non-packed hunter-gatherer cases primarily dependent upon terrestrial plant resources. The exception to the above generalization, case no. 42, is the Guato, who lived in the upper Paraguay River Basin of Brazil. They were swamp-adapted people who have been very poorly documented (Simoens de Silva, 1930; Schmidt, 1942; Métraux, 1946). These people are, nevertheless, interesting particularly in light of the large number of swamp-living peoples from New Guinea, who appear to have been hunter-gatherers in the recent past (Zegwaard, 1959; Van Arsdale, 1978), and the Miskito of Nicaragua (Nietchmann, 1973). Such cases need to be included in the comparative study of near contemporary “hunting and gathering” peoples. However, for now the Guato remains the lone exception to my generalizations. Now let’s turn to the people whose food is predominantly obtained from terrestrial animals, remembering of course that we speak only of people with a very low population density, below 9.098 persons/100 square kilometres. It should be noted that most of the cases exploiting terrestrial animals are found mixed with aquatic resource-dependent people (Fig. 2), which when viewed together describe a “V”-like distribution. The base of the “V” converges at the sub-polar bottleneck with an ET value of 11.75 or roughly 55 degrees latitude both north and south. The terrestrial animal hunters exhibit a biased distribution along the inside of the “V”, which corresponds to the environments with the lowest NAGP. Along the higher NAGP edge of the case distribution, which reverses at the sub-polar bottleneck, are concentrated the aquatic resource-dependent people. Only one case1 is found in settings warmer than the threshold at an ET value of 12.75°C. The latter is coincident with the ET threshold where the shift to dependence upon terrestrial plants may occur, particularly in moderate to low NAGP settings. On either side of this threshold, across a considerable spread of NAGP setting, we see scattered cases

the diet of early hominins

189

Guts & Brains

28-03-2007

15:12

Pagina 190

where terrestrial animals contribute 50% or more to the total diet. These are found primarily in modern grassland settings. It is in high NAGP environments warmer than or essentially at the plant threshold that we note a major change in the setting where terrestrial animal hunting is concentrated; such cases are found exclusively in the highest NAGP locations (greater than a log 10 value of 2.8 for NAGP). These are with one exception located in warm temperate and/or subtropical settings. The one exception is the Yuqui who are reported to be the remnants of a society of hunting specialists (Stearman, 1987), possibly mutualistically articulated with more complex horticulturalists before discovery (Stearman, 1984).2 This case is easily visible on the chart, as it is the only terrestrial animal hunter case in settings warmer than the ET 18 (12month growing season) threshold. It is directly to the right of the Guato (no. 42). If the Guato are misclassified and the Yuqui are the remnants of an economic specialist segment formally part of a larger society, then it is probably important to note that there are no terrestrial animal-dependent or aquatic resource-dependent hunter-gatherers with population densities at or below 9.098 persons per 100 square kilometres known from the modern world in equatorial or tropical climates. All known low population density cases from regions with a 12-month growing season are terrestrial plant-dependent, recognizing of course the two “exceptions” noted above. This, by the way, is the situation also with the New Guinea swamp dweller cases mentioned above. The latter are all predominantly dependent upon terrestrial plants. This general lack of people predominantly dependent upon food sources other than terrestrial plants is also true for the sub-tropical and temperate zones, with a few minor hunting exceptions in the transitional region indicated along the plant threshold in Figure 2 and the narrow line of hunting cases found in high productivity settings which runs almost to the 12-month growing season indicator at ET 18. If we are looking for modern human living conditions as they vary with demography and environments that approximate the probable range of adaptive forms known from the past, prior to demographically driven intensification, the information in Figure 2 is a provocative guide. As an additional means for making very explicit the dietary biases among low population density, near modern hunters and gatherers, I have prepared the following two illustrations. The basis for these illustrations are projections obtained by running multiple regressions between the three variables treated independently, “hunting”, “gathering” and “fishing” (Binford, 2001a Table 5.01, 118-129), with the full suite of environmental variables I reported (Binford, 2001a: Tables 4.01, pp. 60-67, and Table 4.07, pp. 86-93) for only cases with population densities below 9.098 persons per 100 square kilometres. The equations obtained were quite impressive. The projected value of hunting de-

190

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 191

pendence, derived from one equation based upon seven environmental variables, anticipates the percentage dependence upon terrestrial animals. When these projected values are compared to the actual values for the cases to which the equation was fitted, a correlation coefficient R = 0.917 was obtained, with R2 = 0.840. The highest correlation coefficients between observed and expected values were obtained for the projected percentage dependence upon terrestrial plants R = 0.954, the R2 value was 0.911. The latter equation was based on six environmental variables. Finally, the percentage dependence upon aquatic resources was anticipated by a multiple regression equation based on six environmental variables that yielded a fit between observed and expected values as evaluated by an R value of 0.887 and an R2 value of 0.787. These equations were then used to obtain projected values of the three basic sources of foods for the sample of world weather station locations, where values for the equations discussed above could be solved using weather data. The results are, of course, projected percentage dietary estimates at the location of each weather station. These percentage estimates were then mapped with isoplith lines connecting equal values for percentage contributions to the diet by terrestrial animals, terrestrial plants, and aquatic resources, respectively. A global summary of such values is presented in Figure 3. Figure 4 follows with a more detailed view of projected food source values for Western Europe. It should be kept in mind that these maps are based upon equations for low population density hunter-gatherer people only, not the complete set of known hunter-gatherers as was the case for analogous maps previously published (Binford, 2001a: Figs 6.04, p.190, 6.07, p.193 and 6.09B, p.194). These low-density, near modern peoples were fully capable of exploiting aquatic resources effectively. This is something that needs further investigation since the issue of dietary alternatives that includes aquatic resources may not have been a part of the ancient hominin behavioural repertoire. This point should be kept in mind when viewing Figures 3 and 4, where modern human behaviour is projected geographically. Figure 3 graphically illustrates the projected behaviour for low population density, near modern hunter-gatherers onto the contemporary (interglacial) climatic conditions of the earth. On the global scale it is clear that Eurasia exhibits huge areas where terrestrial animals are projected to provide 50% or more of the diet. This is also true of North America as well as much of South America. In dramatic contrast are the continent of Africa, as well as the islands of southeast Asia, and most of Australia where the values below 40 dominate the isoplith maps for percentage dependence upon terrestrial animals. These indicate greater dependence upon terrestrial plants together with aquatic resources. Western European hunter-gatherers are projected at the global scale of comparison for essentially interglacial settings as obtaining not much more than 50% of their diet from terrestrial animals.

the diet of early hominins

191

Guts & Brains

28-03-2007

15:12

Pagina 192

-100

-50

50

0

-50

-150

70

60

50

40

30

20

10

0

50

100

150

0

Fig. 3 Projected behaviour (percentage dependence upon terrestrial mammals) for low population density, near-modern hunter-gatherers onto the contemporary climatic conditions of the earth. See text for explanation.

The latter suggests a more diverse subsistence base, which appears to indicate, other things being equal, a more secure subsistence base for hunter-gatherers than that enjoyed by most specialized subsistence feeders. 70 65 60 55 50 45 40 35 -10

-5 70

192

60

0 50

5 40

10 30

20

15 10

0

20

25

Fig. 4 Projected behaviour (percentage dependence upon terrestrial mammals) for low population density, nearmodern hunter-gatherers onto the contemporary climatic conditions of Western Europe. See text for explanation.

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 193

My interest in this paper is particularly focused upon Western Europe, thus Figure 4 was prepared using a richer sample of weather stations per unit area, providing us with a more detailed view of hunter-gatherer behaviour as projected from the low population density cases of near modern hunter-gatherers. The greater breadth in diet mentioned above is characteristic of modern huntergatherers, who regularly use aquatic resources. This is particularly true of the more northern regions, as shown in Figure 4. On the other hand, inflated values for hunting terrestrial animals may occur along the southern boundary of this 50% dependence zone. Modern hunter-gatherers gain time utility from “large kill” strategies coupled with storage. The boundary between aquatic resource “specialists” occurs at and above the sub-polar bottleneck as shown in Figure 2. For much of the ancient past this aquatic dependence zone would have been uninhabited (see Fig. 5). The southern boundary of this region, with 50% terrestrial animal food dependence, becomes vague since one expects increasing use of plant resources across the land along the northern fringe of the Mediterranean Sea. Yet during glacial events, the line for 50% or greater dependence upon terrestrial animals may well dip down such that it runs east-west along the North African coast, turning slightly northward so that it crosses Israel and the southern tip of Syria. Thereafter, it runs across Iraq and along the northern border of Iran, turning south toward the coast of Pakistan. The area south of this line is where modern hunter-gatherers with a diet of 50% or greater dependence upon terrestrial plants would have lived. The northern fringe of the Mediterranean Sea would have been the “transitional” zone between terrestrial hunters and terrestrial plant-dependent people. This “transitional” region in the south should have exhibited considerable diversity and strategic chronological shifting over time as the environments changed in response to glaciations. Hominin groups behaving like modern hunter-gatherers should have been hunting terrestrial animals largely as Figure 4 suggests for interglacial stages. During glacial stages, however, the hunters may have been distributed southward into the Mediterranean drainage area, as suggested above. There would be a greater extent of uninhabited region extending south of the Baltic Sea area and around mountain areas, considerably south of the distribution shown in Figure 3. It should be noted that a geographic ordination of hunted food use coming close to what is seen across Europe is also found in South America (uninhabited during the era of interest here) as well as South Africa, and Mozambique. The latter region has much in common with the “transitional” zone along the northern fringe of the Mediterranean. The above material may be thought of as a baseline for orienting our expectations regarding hominin behaviour in the ancient past. I would be very surprised if the

the diet of early hominins

193

Guts & Brains

28-03-2007

15:12

Pagina 194

TERRESTRIAL MODEL (SUBSPX) UNINHABITED HUNTING MIXED GATHERING

Fig. 5 Global projections from ethnographically documented hunter-gatherers for primary source of food (terrestrial model subsistence bias). Source: Binford 2001a, Fig. 6.04.

ancient populations of our relatives and ancestors could out-perform the accomplishments of the fully modern human hunter-gatherers whose behaviour is projected here. It should be noted that the distribution of expected subsistence dependence is not homologous with the ancient Palaeolithic data from Europe. For instance, regular exploitation of aquatic resources prior to 60,000 years ago is at best poorly documented, although not unknown, at least in the Mediterranean region, and even up the west coast of the Iberian peninsula. Nevertheless, the Middle Palaeolithic indications of aquatic resource exploitation are consistent with “picking things up” along the beach and at low tide (Stiner and Kuhn, 1992: 322), while an extractive set of strategies common to Homo sapiens sapiens utilization of aquatic resources is not thus far evidenced for the Middle Palaeolithic. Focusing again on the distribution patterns given in Figure 3, one should recognize a strong connection between the pattern of terrestrial plant dependence projected from low population density, near modern hunter-gatherers and the area occupied, or at least penetrated by hominins dated earlier than approximately 500,000 to 600,000 years ago. In fact, current thinking suggests that demographic probes by early hominin populations did not succeed in peopling Europe until roughly 500,000 years ago, and subsequently there appears to have been a continuous presence of hominins in Europe (e.g. Roebroeks, 2001). During the

194

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 195

latter period indicators for diet should not escape a comparison to the area indicated in Figure 3 for a dominant dependence upon terrestrial animals. This area should correspond to the regions unoccupied by populations of Homo prior to the Middle Pleistocene. It should, however, correspond to much of the European evidence for obtaining food from large-bodied animals by the survivors of the successful radiation into Europe at about 500,000 to 600,000 years ago. The honours for initially penetrating the northern boreal and polar regions of Europe and Eurasia goes to Homo sapiens sapiens during the relatively recent, late and immediately post-Pleistocene era. This late event was anticipated nicely for interglacial episodes similar to the present by the “Terrestrial Model” (Binford, 2001a: 192195). Figure 5 reproduces the information from Figure 6.04 in (Binford, 2001a: 190). The contrast between the terrestrial model and the projections from low population density hunter-gatherers of the near modern world (Fig. 3) is simply that the latter is a projection from data, while the former is a predictive model built on the assumption of no major technological aids to food procurement. The diet is modelled as conditioned directly by the availability and relative presence of different consumable resources in the world’s biotic communities. Additional assumptions were that the consumer was a terrestrial creature, omnivorous, and would only access foods that it could reach and/or overpower physically and/or as regards animal products obtain by scavenging. Finally, it was assumed that no aquatic resources were regularly accessible except along beaches and in tidal pools, etc. The implications of the projections and modelled predictions of the dominant sources of foods, not the actual diet and its likely complexity and variability across different regions at different time periods in the past, should seriously be considered by those who tend to sensationally cite evidence for “meat-eating” to be sufficient evidence to infer a predatory hunting adaptation in the warm tropical early home ranges of our hominin ancestors. Modern, technologically aided, fully capable Homo sapiens sapiens did not have such diets and consumptive specializations except in the zones of very short growing seasons. These are also the settings which require many additional “necessary resources” (for clothing, housing, etc.) than are required for people living in warmer, more continuously productive settings. The area uninhabited until recently includes regions that the models and projections do a very provocative job of anticipating. The archaeological data we have at present tend to confirm the predications and projections. Secondly, the gross fit between the models and projections with the geographical patterning indicative of early radiations is encouraging. Differential changes within regions over time, when the early populations of “Homo” in Europe occupied the northernmost edge of the Homo range, are yet to be well understood. However, geography as a means to recognize a differential response to environment is not enough to begin speaking of diet. Let’s expand our horizons somewhat.

the diet of early hominins

195

Guts & Brains

28-03-2007

15:12

Pagina 196

Part III. Looking again to modern humans for clues regarding weapon use and procurement strategies The issue of tactics for obtaining food is probably more closely related to diet than are the documents of discrete animal death sites or the imagined activities of early hominins advanced by many archaeologists. How do we accurately infer tactics and strategies of food procurement from the archaeological record? I will focus specifically on the issue of “hunting” and the use of weapons as well as techniques for disadvantaging animals prior to weapons use. Some years ago I accumulated a data file to aid in addressing the issue of the probability that Middle Palaeolithic “hunters” regularly killed cave bears (Binford, 2002). I have continued to add information to that database and will use it to provide a knowledge baseline for reasoning on the issue of weapons and hunting tactics. I will begin, however, in Table 1 with data accumulated by Churchill (1993) supplemented, as noted, from my data. Table 1. Weapon type and effective ranges. ———————————————————————————————————————————————————————————

(Data taken from Churchill, 1993) ———————————————————————————————————————————————————————————

Weapon type

Range (in metres)

Standard deviation

Number of cases

———————————————————————————————————————————————————————————

Thrusting spear Hand-thrown spear Atlatl spear Bow and arrow

Physical contact 5.7-7.8 39.6 25.8

Not available 2.2-0.9 5.5 2.4

(26) 14 9 25

———————————————————————————————————————————————————————————

Note: Values in parentheses come from my research and refer to the number of species. The number of cases refers only to the number for which range was a measured observation.

Churchill’s data dramatically illustrate that the range effectiveness of the thrusting and hand- thrown spear is so much less than that of the atlatl and bow and arrow. We can reasonably expect that people lacking the atlatl and/or bow and arrow were forced to rely less on “at a distance” shock weapons than the majority of hunter-gatherers known from the relatively recent past. The spear, as a hunting weapon, is severely limited relative to the shock weapons used most commonly by recent hunter-gatherers. In this more modern context, the thrusting spear and the hand-thrown spear have rather specialized uses in most areas. The hand-thrown spear or “javelin” is commonly used for throws directly at specific animals who are commonly already disadvantaged or disabled. If the javelin is used by a group of hunters, most often engaged in a game “drive” or surround, a high arching simul-

196

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 197

taneous throw is made so a “rain” of spears comes down in the area where animals are congregated. This is not uncommon in Tasmania where wallabies are hunted using fire drives from one direction, and men concentrate along the side of the fleeing animals anticipated to be their escape path. The hunters surprise the animals by blocking their escape, and in the confusion may successfully kill a number of animals with javelin and thrusting spear. Churchill (1993) noted the strong association between the use of the spear in conjunction with techniques designed to disadvantage the prey animals prior to the delivery of a, hopefully, killing blow. “Disadvantaging is highly dependent upon physiographic features for success, requiring natural or man-made traps such as narrow arroyos, box canyons, corrals, swamps, snow drifts, or bodies of water, or the assistance of dogs to surround and impede an animal so that the technology can be employed. The relationship of hand-delivered spears to terrain dependent hunting techniques in northern latitudes suggests that, prior to the development of long-range weapons, hunters may have been restricted to places in the environment where the technology was most effective for the hunting of medium to large size terrestrial game” (Churchill, 1993: 18-19).

In short, hunting with thrusting spear and javelin was largely restricted to where the animals were most disadvantaged. As a follow-up to Churchill’s insights, I have summarized data from my larger sample of 131 ethnographic cases regarding hunts and specific kills (Table 2 ) and the way the animals were disadvantaged prior to the use of killing weapons.

Table 2. Context in which game may be disadvantaged relative to the hunter when using the full range of weapons listed in Table 1. ———————————————————————————————————————————————————————————

Context

Number of cases

Percentage

———————————————————————————————————————————————————————————

(1) (2) (3) (4) (5) (6) (7) (8)

Water Snow At dens Use landforms Trapped or poisoned Exhaustion Not disadvantaged Missing data

19 7 22 5 21 11 21 25

14.5% 5.3% 16.8% 3.8% 16.0% 8.4% 16.0% 19.1%

———————————————————————————————————————————————————————————

Total

131

99.9%

———————————————————————————————————————————————————————————

the diet of early hominins

197

Guts & Brains

28-03-2007

15:12

Pagina 198

In 80% of the documented cases of hunting tabulated in Table 2, the basic strategy was to initially disadvantage or disable the game prior to the delivery of “killing” blows with weapons. All such techniques, except possibly no. 5 “trapped or poisoned”, are possible components of the Middle Palaeolithic hunting bag of tricks. Of the cases tabulated as “not disadvantaged”, thirteen were examples where the hunter exploited some known prey weakness or common response to being approached that put the hunter in an advantageous position for using the weapons available. For instance, in Central Australia a common kangaroo hunting technique is to hunt during mid-day when the animals are “shading” under brush and/or small trees. The hunter observes the way the animal is facing and positions himself so that he is directly in front of the shading animal. He begins to approach the animal, standing fully erect, nude so no clothing will move in the wind, with arms rigidly positioned relative to his body. He slips one leg at a time forward so there is no change in his silhouette as seen by the animal. As the animal begins to get nervous by what appears to be something approaching, but with no change in its silhouette, it will jump slightly to one side or the other to get an oblique angle view. The well-practised hunter knows this little trick and is waiting for it so that when the animal jumps, so does the hunter, but forward as well as obliquely, so that he comes down again directly aligned with the animal’s line of sight but much closer to the animal. The animal detects no clear movement but remains nervous. This deadly dance continues until the hunter is close enough to quickly use the atlatl to spear the kangaroo, or the kangaroo detects movement and runs away. I used the term “spear” in conjunction with the word atlatl since the object propelled averages 12 feet in length, is made from tree roots, not fletched in any way, and importantly not tapered as is the thrusting spear, or even modern javelin as known from Olympic distance competition.3 When working with hunter-gatherers, researchers are always impressed with the enormous detail that the hunters have learned about the prey in their region. They know every small detail of the animals’ responses to being hunted, and their reactions to the presence of both hunters and other animals, seasonal shifts in behaviour, differential feeding habits depending on the location, and social state of the prey population, etc. This detailed body of observation-based lore is the huntergatherer’s “knowledge capital”, the means to success or, if used poorly, failure. I might even suggest that the less sophisticated the technology, the greater the necessary dependence upon a complete knowledge of prey behaviour for insuring a steady food supply. With an AK47 you don’t have to know so much! It is quite likely that the first successful occupants of glacial Europe had to have accumulated considerable knowledge about the behaviour of animals and learned to use this knowledge to their advantage. It is also probable that the early populations of European arrivals at about 600,000 years ago had already developed

198

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 199

some learning skills from hunting, even though it is quite likely that their primary sources of foods were commonly obtained from terrestrial plants, as was suggested earlier based on the behaviour of low population density, fully modern human behaviour (Figs. 1, 2, and 3). There is another interesting domain that is rarely talked about when discussing hunting, namely the material aids that may be used by hunters to insure advantage over potential prey. Table 3 summarizes the frequencies with which some common aids are used by the modern hunters to secure an advantage over their prey. Table 3. Hunting aids used to secure an advantage over prey. ———————————————————————————————————————————————————————————

Hunting aids

Number

Percentage

———————————————————————————————————————————————————————————

Fire for driving Use of dogs Use of snowshoes, skis, etc. Organized drives with construction Riding a horse Moving in boats Use of dog or reindeer sled Aids unreported

10 22 7 38 8 17 8 24

7.5% 16.4% 5.2% 28.4% 6.0% 12.7% 6.0% 17.9%

———————————————————————————————————————————————————————————

Total

134

100.1%

———————————————————————————————————————————————————————————

Note: If more than one aid was used, the most difficult to build or organize was tabulated.

Organized drives with construction is the most common hunting aid used by modern hunter-gatherers, followed by the use of dogs in game surrounds. It is a bit sobering to realize that the use of fire with an early “Homo” organized surround and minor constructions on the landscape to direct game towards ambush locations are the only hunting aids on the list that Middle Palaeolithic hunters may have used, or seem likely to have used, given what is known about the other aids ethnographically. When modern hunter-gatherers are examined for patterns of weapon use across major environmental regions, some basic relationships are revealed (Table 4). The bow and arrow is the dominant form used in all environments except among hunter-gatherers living in deserts, where the atlatl and throwing spear are most common. The thrusting spear is very close to being used with equal frequency to the bow and arrow in both the Arctic and the boreal forest. Data reported elsewhere (Binford, 2002) show that the thrusting spear is primarily used as a defensive weapon among modern hunter-gatherers, not only in the Arctic but in all environments. Its “offensive” role is for dispatching wounded or disabled prey. The throwing spear is known among hunters of the Arctic, boreal and temperate

the diet of early hominins

199

Guts & Brains

28-03-2007

15:12

Pagina 200

Table 4. Use frequencies for different weapon use across environmental zones. ———————————————————————————————————————————————————————————

Habitat type

Thrusting spear

Throwing spear

Atlatl

Bow & arrow

Poison b.&arrow

Row total

———————————————————————————————————————————————————————————

Arctic

08 (42.1%)

01 (05.3%)

01 (05.3%)

09 (47.4%)

00 (0.0%)

19 (12.75%)

———————————————————————————————————————————————————————————

Boreal

12 (44.4%)

01 (03.7%)

00 (0%)

13 (48.1%)

01 (06.3%)

27 (18.12%)

———————————————————————————————————————————————————————————

Temperate 15 (25.0%)

02 (03.3%)

00 (0%)

38 (63.3%)

05 (08.3%)

60 (40.27%)

———————————————————————————————————————————————————————————

Desert

03 (20.0%)

04 (26.6%)

04 (26.6%)

02 (12.5%)

02 (12.5%)

15 (10.06%)

———————————————————————————————————————————————————————————

Tropical

08 (28.6%)

07 (25.0%)

01 (03.6%)

10 (35.7%)

02 (07.1%)

28 (18.79%)

———————————————————————————————————————————————————————————

Column total

46 (30.9%)

15 (10.0%)

06 (04.0%)

72 (48.2%)

10 (06.7%)

149

———————————————————————————————————————————————————————————

zones, but it is rarely used in hunting. On the other hand, it is used more commonly in desert and tropical settings than in any other setting. Without question, ethnographically documented hunter-gatherers living in deserts used the most diverse array of hunting weapons. Tropical settings are second on the weapon diversity scale. The temperate zone is the most specialized in the use of the bow and arrow, using the thrusting spear only slightly less than people of the deserts and in tropical settings. The boreal forest and the Arctic regions seem to share a common set of weapons usage patterning. The thrusting spear is almost as commonly used as the bow and arrow. Game in this region are mostly moderate to large creatures, and the size of omnivorous predators is large. The bow and arrow with its greater effective range (see Table 1) may commonly wound an animal, but the thrusting spear is regularly used to dispatch it. This comparative exercise produced few surprises; the weapons known to have been available to the pre-Homo sapiens sapiens populations of Western Europe were generally less effective than those available to near modern hunter-gatherers. This situation promotes the idea that sustaining a population that is more than 50% dependent upon terrestrial animal resources in a glaciated European setting was a challenging and a relatively high-risk (of failure) situation. The one aspect of this situation that makes success feasible could rest upon the tactical

200

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 201

means employed in predation. One possibility could be a dramatic difference in the “athletic” performance by Neandertal hunters (Trinkaus, 1983, 1986)4. Another might be to focus on the regularities in species-specific behaviours that might well be unknown to most of us. This latter focus would be on reliable “weaknesses” characteristic of different species in different contexts. What I have in mind is well illustrated by the distribution of caribou bones on the bottom of every lake I explored in arctic Alaska. From observing how they got there, I can testify that the bones on the bottom of the lake would vary between evidence of wolf kills and consumption to human kills and consumption. Why? One reason is that wolves can outrun caribou on ice, while caribou can generally outrun wolves in moderately deep snow. The result is that wolves will follow caribou, walking in the trails they make in the snow until they approach a river or lake. Then the wolves split up and some rush ahead to the ice on the lake or river, while others continue “driving” the animals along the caribou trails. As the caribou get closer to the icecovered lake or river, the wolves charge the caribou and “herd” them towards the ice. If successful, the “advanced party” of wolves will almost certainly make one or more kills on the ice. They may feed for some time to “fill up” before the carcass freezes. The remains stay on and are absorbed into the ice until the melt in spring, when the bones drop to the bottom, particularly in lakes. Rivers may carry the bones to point bar accumulations, etc. Humans also set up hunting blinds around lakes and on the ice where they may fish and also keep an eye out for wolf-driven caribou, and even have cooperating hunters driving animals toward lakes. What makes this a successful strategy is that caribou slip and slide on the ice. Anything that frightens them tends to also make them try to rapidly change direction, generally insuring that they lose control and fall down. They can be killed when they are so disadvantaged. The bottom of the lake accumulates all the remains from ice fishing, including food remains consumed there, as well as remains from butchering episodes and wolf kills. It would be my guess that any “hard-hoofed” species would have similar problems running on ice, and predators almost certainly know such things. Many species have very regular movement patterns. Reindeer or caribou are a good example. They regularly move out of the winter range – in the forest – to more open landscape in the spring. The cows generally migrate off the winter range the earliest, initially moving slowly but increasing their speed as birth becomes more eminent. When they leave the forest, they may continue to feed, but once the herd size increases and the activity of the foetus increases, they stop feeding and move day and night toward the calving area. When the cows leave the forest, they may be accompanied by some of their calves born the previous spring, but many will leave their mothers as the rate of cow herd movement increases. The calves join male groups who are feeding and moving very slowly toward the

the diet of early hominins

201

Guts & Brains

28-03-2007

15:12

Pagina 202

summer range. The Nunamiut Eskimo report that they would never hunt at the calving area. When asked why, they replied with a laugh that there was always better hunting in “other places”. Wolves, however, do travel to the calving areas where they prey on newly borns, as well as females and calves of the year in poor condition. It is interesting that the migration paths to and away from the calving areas may vary, but the actual location of the calving area tends to be relatively stable. It is generally located in the same place year after year if the size of the breeding population is large. The wolves take advantage of this reliably repetitive behaviour. In this case Eskimos appear to be targeting prey for optimal return when the wolves are targeting prey for reliable returns. In other cases of exploiting “disadvantaged” animals, reliability versus optimality considerations are not necessarily primary. For instance, in the western part of North America, it is well known that horses, either wild or raised on the range, will generally run uphill if frightened. To prevent them from injury or sometimes killing themselves, range managers will frequently fence off box canyons that have loose or large boulder scree talus slopes so that horses that might enter are not frightened by coyotes or other predators, thunder storms, etc., to run up the talus slopes and break their legs. Hunters could both “encourage” or drive horses to such topography and then frighten them up slope. Horses are not the only animals that tend to run uphill when frightened. This tendency was also characteristic of caribou and was the basis for a regular hunting strategy employed by the Nunamiut Eskimo for taking male caribou migrating north in spring. The Eskimo tactics required only a small group of hunters, frequently teenage boys, and with wind-blown snow smoothing the appearance of the slope, the deceived caribou, when surprised by the hunters, regularly ran uphill, slipped between boulders and either broke their legs or could not extract themselves from the deep snow patches among the boulders. These animals made easy prey for the hunters on snowshoes. Good hunters take advantage of their prey’s regular behaviours. In the examples given above, the only weapons needed would be a thrusting spear to dispatch the wounded or “caught” animal. While these tactics sound simple and reliable, they all depend upon the behaviour of the prey, not the ability of the hunter to take prey on his terms, such as when he needs food or when he encounters prey “unexpectedly”. This means that the mobility patterns, the seasons of use for localities, as well as the size and character of groups might vary less with the specific distribution of shelter or landscape positioning tactics common among hunters equipped with reliable shock weapons. Proposition 1. The mobility and activity rhythms of the prey, as well as the prey options available at different seasons or locales, should condition very directly

202

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 203

the land use patterns of hunters unequipped with effective shock weapons. Proposition 2. Poorly equipped people living in less challenging environments where there are more food options relative to any given location should be expected to exhibit greater redundancy in land use patterns. That is, greater redundancy than among peoples where there are few food options and thus less diet breadth during any given season of the year. Among the latter, expedience and the opportunity accommodations to relatively unique situations would make for greater variability in the scale of land use patterns.

One suspects that these two propositions would have been particularly germane to predominantly hunting-dependent populations or even seasonal hunters of Europe during the Middle Palaeolithic. Variable patterns are anticipated across Europe where the hunting of animals contributes differentially to the diet, as is suggested in Figures 2-5. Variability can be expected seasonally in geographic site patterning in response to the tactics of animal exploitation and, in principle, also to stone tool raw material accessibility. The latter might also be expected to co-vary with seasonality. Proposition 3. Assemblage content should be responsive to the above factors as well as the distribution of sheltered locations for sleeping, etc. Shelter access could vary with shifts in species dependence seasonally, at least partially. On the other hand, as potential diet breadth is likely to be reduced with more challenging climatic change, we may observe opposite trends toward greater redundancy in the use of specific places in the landscape to correlate with more extreme climatic conditions.

Let’s explore briefly what I have in mind. I propose that: Proposition 4. Neandertal hunting strategies should, other things being equal, favour tactics that are reliable over those that are “optimal”.

Proposition 4 should apply to the quality and/or the quantity of food. For instance, smaller prey may be favoured over larger prey only if the former can be obtained more reliably and in adequate numbers. Proposition 5. The dietary role of preferred species in “climatically stressed” environments can be expected to shift seasonally, dependent upon the relative reliability or feasibility of “known” hunting tactics accessed among the potential prey species available simultaneously by season.

the diet of early hominins

203

Guts & Brains

28-03-2007

15:12

Pagina 204

Shifting of game targets would condition a pattern of land use very different from more “modern” hunting strategies where the emphasis is upon the taking, for instance, of large numbers of animals in fall, preparing them for storage and consumption over a much longer period than they are actually accessible in a given location (Binford, 1978). Thus far, storage cannot be warranted as part of the tactics used by Neandertals. It should be noted, however, that the storage-linked behaviour cited above was the likely phenomenon that has led many archaeologists to suggest that the hunting by Upper Palaeolithic humans was “specialized”. Insofar as this is correct in a dietary sense, storage is most certainly implicated. There is, however, no evidence that Middle Palaeolithic populations practised tactical storage. Thus, the citation of kill sites where only one species is involved is not evidence for “specialization” in the sense used with respect to recent and Upper Palaeolithic hunters. I should also caution that patterns of land use or “settlement patterns” vary among Homo sapiens sapiens hunters with group sizes as conditioned by tactics and the availability of alternative resources. We need to understand more about Neandertal “settlement patterns” if we are to understand their diet and the social aspects of the adjustments to their territories. I have briefly introduced some ideas derived from my knowledge of modern hunters to be used as a baseline for comparison and even some expectations regarding the procurement strategies of terrestrial animals by Neandertals. We all know that the evidence we might cite to warrant an argument as to the diet of Middle Palaeolithic populations must rest with the sophistication with which we can argue for “indicative” properties in the archaeological record. How do we make provocative, or ideally compelling, arguments from the archaeological data? Let’s explore some data from a Middle Palaeolithic site.

Part IV. A Middle Palaeolithic site: Combe Grenal of south-central France: A case study dealing with the question of Middle Palaeolithic diet I will begin by making available some of the basic data from this important site. It was excavated by François Bordes (1972, for a review). It spans, essentially continuously, all of Marine Isotope Stages (MIS) 4 and 5. The well-preserved deposits are estimated to start between sub-stages 5.4 and 5.51 with a likely beginning date of 113,000 years ago. The rockshelter is essentially filled by sub-stage 3.31 or approximately 53,000 years ago. Thus, the archaeological sequence from this site runs as a continuously accumulating deposit over the span of 60,000 years. All but one recognized level contained artefacts, animal bones and other evidence for what appears to represent regular use by Neandertals from the beginning to the end of the sequence. There are an impressive 144,189 lithic

204

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 205

artefacts from the site, of which 18,904 are retouched tools. There are a minimum of 553 individual animals evidenced in the deposits by 12,121 identifiable bones and teeth recovered intact. In addition to the removable artefacts, there are 127 documented small hearths and 19 large ones. The only additional features were seventeen small pits and one large one documented among the levels at the site. These statistics are certainly impressive, and we could learn much from the basic data summarized above. However, for the purposes of this paper, I shall focus upon the animal remains and any relationships we may uncover about the way in which the Neandertals exploited different species. One way is to relate how Neandertals treated species, places, and situations about which we have good descriptions of how such phenomena were treated by Homo sapiens sapiens. For instance, I developed arguments elsewhere from the patterns of adaptive variability documented among near modern hunter-gatherers and formulated a generalization about diet breadth and systems security. The generalization (Binford, 2001a: 419) states that: Proposition 6. If generic hunter-gatherers successfully radiate into environments with very low terrestrial productivity, the majority succeed through specializations which condition a relatively narrow niche breadth.

Extremely specialized cases occur in environments in which there are few resource alternatives when viewed from the perspective of environments with abundant terrestrial plants, aquatic resources, as well as terrestrial animals. We need also to keep in mind that proposition 6 refers to modern Homo sapiens where specializations can be due to extreme seasonal variants (such as the Central Eskimo where seals are consumed during winter and fish with caribou during the warmer months). Neandertals living through eras of glacial development across northern Europe would be expected to behave in accordance with the principles of the above generalization. However, the specializations would be expected to be executed in a more fine-grained manner and in the context of species behavioural shifts, not the consumer populations’ food management strategies, e.g. storage, characteristic of our species. For instance, Figure 6 demonstrates some interesting patterns in the degree that specialization on a relatively few species across time is documented at Combe Grenal. High values of the diversity index signal specializations on a single or a very few animal species. Examining the pattern formed by the lowest specialization indices, there is an approximate normal curve beginning at level 45, rising to around level 26, and then more rapidly going down to about the same value, at the lumped levels of 1-3, as where we started the increase at level 45. This pattern is temporally

the diet of early hominins

205

28-03-2007

LEVEL

Guts & Brains

15:12

Pagina 206

Combe Grenal

1.30 8.00 11.00 14.00 17.19 23.00 25.00 27.00 29.00 35.00 37.00 39.00 41.00 45.00 50.00 54.00 0

10

20

30

High values equal greater species specialization

Fig. 6 Diversity index signal for Combe Grenal sequence, levels 54 to 1. See text for explanation.

distributed across oxygen isotope MIS 5.51 at the bottom of the sequence to the beginnings of MIS 5 at about level 28. From level 28 through level 8 is the very cold MIS 4, and the cool interval between cool stage 3.31 to 3.3 is indicated by a return of reindeer dominance in Bordes levels 7-8, before the sequence at Combe Grenal was complete. An interesting area of the graph spans from early MIS 4, level 11, through the most recent three levels of MIS 5 or the coldest period in Europe prior to the onset of the most recent glacial episode between 24 and 16 thousand years ago. We note a regular pattern of increasing specialization on fewer species extensively exploited. The most extreme cold of the MIS 4 glacial episode is represented in Combe Grenal by levels 17-28. The specialization which drove the lower regular pattern, as well as most of the high spikes prior to the peak cold of MIS 4, was upon reindeer. The latter faunal dominance peaked with the coldest phase of MIS 4 (see Fig. 7). I should also point out that coincident with the peak of the MIS 4 cold, there was a massive roof-fall, which changed dramatically the usable space within the cave and accordingly could have conditioned change in the way the rockshelter was used. Leaving this

206

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 207

18

O (normalized)

-1 0

0

dated events

1 1.1

2.0

2.2 2.21 2.23

3.0

3.1 3.13 3.3

50 Combe Grenal Level 23 Stage 4 Maximum

3.31

4.0

4.22 4.23 4.24

5.0 5.1

5.2 100

5.3

5.31 5.33

Time ( kyr BP )

5.4 5.5

6.0

5.51 5.53

6.2 6.3

150

Fig. 7 Global Oxygen Isotope 18 fluctuations in the last 150,000 years, with Marine Isotope Stages and sub-stages referred to in the text.

issue aside for the moment, we need to continue our consideration of the patterns of species use at the site. We see in Figure 8 that horse contributed around 25% to the animal component of the diet during the warmer phases prior to level 38, thereafter horse was a reduced contributor all during the ascendancy of reindeer as the dominant source of animal food. Bovids (primarily bison, Delpech, 1999: 82) tended to fill the gap between reindeer and elk or red deer as the glacial episode of MIS 4 progressed. On the other hand, after the peak cold period and with the onset of a declining role for reindeer in the diet of the Neandertals, horse played a prominent dietary role during the period of glacial retreat and associated amelioration in climate. Horse moved into a significant dietary role beginning at level 20 and shifted to a dominant dietary role in levels 16-14. Bovids (bison) then moved into a dominant role in levels 11-13. Thereafter for a brief period, in levels 4-8, bovids dramatically reduced as reindeer made a brief reappearance at around 50,000 years ago (see Figs. 5 and

the diet of early hominins

207

28-03-2007

LEVEL

Guts & Brains

15:12

Pagina 208

Combe Grenal

1.30 8.00 11.00 14.00 17.19 23.00 25.00

Reindeer

27.00

Roe Deer

29.00 Pig

35.00

Ibex

37.00 39.00

Horse

41.00 Donkey

45.00 50.00

Red Deer

54.00 0.0

Bovid .2

.4

.6

.8

1.0

Proportion of Level MNI Total by Species

Fig. 8 Proportion of level MNI Total by Species for the Combe Grenal sequence, levels 54 to 1.

6). I should comment that all during the cold period, the “bovids” were dominantly Bison priscus (Delpech, 1999: 69). The important role for the bison during the most recent levels of Combe Grenal fits well with their chronologically increased importance at sites such as Mauran, the lower levels at La Ferrassie and Camiac (Delpech, 1999: 69). Bison also played a significant role in the diet of Neandertals during the warming phases after the climax of the MIS 4 cold period. Horses seem to play the role that red deer had occupied during the build-up of glacial conditions. Their prominence probably documents a greater reliability for hunting horses and bovids after the maximum cold of MIS 4 than was realized during the high-stress years leading up to the maximum cold of MIS 4. More likely, however, natural ecological dynamics were at work among competing populations of animals. Namely, populations of both bison and horse built up at the expense of red deer during the era of increasing cold. This pattern may in all likelihood be translated into a slower shift to grassland and steppe relative to the established forest biomes favoured during the pre-MIS 4 era of MIS 5. As the postglacial maximum era started, the more ecologically established species outcompeted the re-establishment of species otherwise more successful during the warm

208

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 209

to temperate settings of the early MIS 5. Open habitats with high biomass turnover rates tend to outcompete forest biome establishment, both with high wind settings (greater desiccation of tree seedlings) and a greater role for fire in the destruction of young tree growth. I think that the basic gross environmental context for a closer look at Neandertal behaviour has been established, and the empirical material summarized in Figures 6-8 support general Proposition 1. The latter was derived from our knowledge of near modern hunter-gatherers, extrapolating to a population facing dramatic shifts in climate-driven environmental changes, particularly cold-stressed environments. Namely, increased specialization upon reduced numbers of species. We have not dealt with propositions 2-4, however, nor have I treated the issue of Neandertal diet relative to Neandertal tactics for food procurement. Let’s begin by exploring some aspects of the relationships between food procurement and biased prey selection. It was suggested (Proposition 4) that, other things being equal, Neandertal hunting strategies should favour tactics that are reliable over those that are “optimal”. What do I mean by this, and is there any empirical support for such a suggestion? Table 5 summarizes the relative frequencies of reindeer of different age categories represented by dental characteristics in the various excavated levels at Combe Grenal. Optimal in this case would conventionally refer to the greatest caloric return relative to the energy expended in procurement. Since we have no realistic way of evaluating the latter, a bias in favour of prime age adults, assuming whole animal carcass introduction, will be considered tactically optimal. This is a position consistent with arguments by Stiner (1994: 371-379) resulting from her comparative work with Italian materials.

Table 5. Age variability of prey: Reindeer. ———————————————————————————————————————————————————————————

Combe Grenal level

Juvenile

Prime adult

Old adult

———————————————————————————————————————————————————————————

7-8 11-20 21 22 23 24 25 26 27 28 29-36

20% 30% 65% 63% 20% 10% 25% 25% 26% 52% 35%

40% 35% 10% 27% 20% 90% 55% 59% 54% 30% 45%

40% 35% 25% 10% 60% 0% 20% 16% 18% 18% 20%

———————————————————————————————————————————————————————————

the diet of early hominins

209

Guts & Brains

28-03-2007

15:12

Pagina 210

Examination of Table 5 shows that reindeer were being exploited in an increasingly optimal manner (increasing percentages of prime age adult prey) from level 36 up through level 24, which was the maximum cold of MIS 4 based on the focal procurement upon reindeer (Fig. 7). Level 23, however, documents a dramatic shift to old individuals immediately after the MIS 4 maximum cold. This could reflect a change in reindeer behaviour associated with the climatic cold conditions peaking during the formation of levels 24 and 25. Alternatively, it could reflect a change in the season of occupation with an associated tactical shift by Neandertals, or the general tactics of prey exploitation, such as hunting vs. scavenging, etc. Immediately after this dramatic shift to the exploitation of old adult reindeer (level 23), there follows another drastic shift in favour of the primary exploitation of juvenile reindeer during the occupations constituting levels 22 and 21. Several features of this table are revealing. Level 23 in Combe Grenal was deposited as the first level after the apex of dependence upon reindeer during the maximum cold of MIS 4 dating to approximately 67,000 years ago. This is also the first level after a major roof collapse, which restructured the spatial character of the rockshelter. We see that the dominant fauna in level 23 was reindeer and approximately 60% of those animals were old individuals, well past their prime of health. Such a “hunting” bias is not an optimal meat nor fat returning strategy. On the other hand, the same level exhibits the first major increase in the numbers of horses introduced to a location since the onset of the colder climates documented at the beginning of MIS 5 – approximately 72,000 years ago (see Fig. 8). These horses were 54% prime adults, and the remainders were dominantly juvenile (see Table 6). The pattern noted among the horses was partially analogous to the pattern characteristic of reindeer during the era of increasing cold that led up to the glacial maximum culminating at Combe Grenal with level 24. The shift in reindeer exploitation tactics seen in Table 5-level 23 contrasts dramatically with the levels above and below. Why did the Neandertals begin taking older reindeer after the glacial maximum? Why were old and prime age adults equally representTable 6. Age variability of prey: Horse. ———————————————————————————————————————————————————————————

Combe Grenal level

Juvenile

Prime adult

Old adult

———————————————————————————————————————————————————————————

1-10 13 14 23 21-28 29-47 49-55

02% 10% 16% 38% 42% 05% 44%

97% 88% 80% 54% 48% 65% 54%

01% 02% 04% 08% 10% 30% 02%

———————————————————————————————————————————————————————————

210

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 211

ed above level 11 or until reindeer were apparently no longer present in the region? Why the dramatic shift to juvenile individuals immediately above level 23? These contrasts marked major changes in the hunting tactics employed with regard to obtaining reindeer. They also reflect differences in diet, assuming changed ways of using animals of different sizes and ages. In turn, the latter could implicate changes in the character of occupations with attendant different food needs by occupants organized differently. I have in mind age and sex differences with attendant different group-related roles. Let’s leave aside the latter possibility for the moment. We need to consider possible changes in the exploitation of reindeer and how their behaviour could have impacted hunting tactics. If reindeer behaviour changed, that could have influenced how the Neandertals hunted them as well as other game. As I have emphasized, the change was immediate in the content between levels 24 and 23. The change of interest followed chronologically after level 24, where almost all the reindeer were prime age adults (7 of 8 animals). According to proposition 3, stress should result in a dramatic change in strategy in the direction of increased reliability rather than an attempt to achieve an increased return rate on the labour invested, which would hardly be possible given the bias toward prime adult animals documented for level 24 as well as levels 25-27. Understanding the tactical shift is further enlightened by the behaviour of the Neandertals as implicated by the contents of levels 22 and 21. Immediately after the climatic conditions that influenced the behaviour behind the facts of level 23, further instability is suggested for levels 22 and 21. There is a dramatic increase in the number of horses exploited coupled with a strategy for exploiting reindeer not previously recognized. The tactics used on reindeer suggest that Combe Grenal is now geographically positioned close to the reindeer calving area. This appears to be the case during the entire decline of the glacial event marked by the MIS 4-3 transition, which also was accompanied by a marked decline in the exploitation of reindeer. This era is informally characterized by the heavy exploitation of juvenile calves of the year as well as adult females accompanied by the presence of foetus remains. This combination of animals could only have been exploited along the spring migration of females to their calving area or even at the calving area itself. It is quite likely that the route was up the Dordogne valley leading towards more “open” grazing areas bordering the hills and ridges of the margins of the Massif Central. Reindeer and caribou females do not feed as they get closer to the calving areas, but they frequently leave the winter range earlier than the bulls. The bulls move slowly toward the summer range feeding all the way. As the females get closer to the calving area, they stop feeding and begin to move both day and night until the calving area is reached. When they stop feeding, some of the calves of the previous year tend to leave their mothers and join with the smaller, slower moving male groups that

the diet of early hominins

211

Guts & Brains

28-03-2007

15:12

Pagina 212

feed over a much wider area in their migration toward the summer range. This means that males and some calves of the year are generally in much better condition, nutritionally speaking, during the spring movement to summer pasture. On the other hand, bulls are much more difficult to find in spring, requiring a considerable search time to locate their small groups as they drift toward the summer range. In marked contrast, the females are generally in very poor nutritional condition with little stored fat left after the foetus has been the privileged nutrient recipient all winter. They are, nevertheless, easy to find. As mentioned earlier, wolves, not uncommonly, hunt at the calving areas, probably for the same “reasons”. It is interesting that the Neandertals chose to exploit the female spring herds, in a poor nutritional state, and further disadvantaged by pregnancy. The recurrent presence of calves of the year in the faunal assemblage with foetus and exclusively female antlers strongly indicates that the end of winter and the beginning of spring was, as is generally true in temperate to arctic settings, one of the worse times of the year for reliable terrestrial foods including animals. The fairly predictable behaviour of the female reindeer herds, coupled with the late April to May period of probable movement by females toward the calving area, supports the earlier propositions (3 and 4) regarding favouring reliability over optimal returns under conditions of stress. I draw the reader’s attention again to Table 5, where the tell-tale presence of juvenile as well as adult cows (identified by antlers) characterizes the reindeer faunal remains until at least level 11 in spite of an ever diminishing contribution by reindeer to the overall diet (Fig. 8). Exploiting early spring female herds en route to or at the calving area lasted until reindeer were no longer taken. When reindeer briefly reappear in the upper levels of the site (levels 7-8), they are rare, and the majority are represented by adult individuals, reminiscent of the pattern seen at the onset of cold as indicated in levels 25-27. Tactically speaking, this represents still another exploitative strategy used by the Neandertals, but the environmental conditions were very changed, and the character of the site at Combe Grenal was, at the time of levels 4-8, an open site. As the reindeer disappeared in response to warming climate changes, first horse, then bison and finally red deer increased relative to one another until the site of Combe Grenal as currently known was no longer used by Neandertals. Bison were unrelentingly represented by prime adults, while red deer and horses exhibited a pattern of exploitation more like reindeer during the cold period. The latter pattern of exploitation differed from the reindeer in the fact that prime age adults were most common in all horse-bearing levels, but they were accompanied by a regular presence of juvenile and newly born youngsters, suggestive of summer kills. The pattern seen in Table 6 shares with reindeer the greater exploitation of

212

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 213

juvenile individuals during the coldest period. However, horses were similarly exploited during the warmest period levels, e.g. 49-55. The latter facts suggest that the vulnerability of pregnant females and females with young was a major “disadvantaged” situation, exploited by Neandertal hunting tactics whenever possible, regardless of species. Modern hunters view pregnant and very young animals as undesirable because of the poor nutritional state of the adults, and the lack of bulk food on a yearling. At the same time, the skin of the female and the yearling reindeer at least are in very poor condition, commonly full of holes from the hatching of insects whose eggs, implanted in the ungulate the previous fall, eat their way through the skin in spring. Hunting spring females and yearlings, with a foetus “bonus” is certainly not an optimal hunting strategy. It does, however, seem to have been reliable. This further supports the general strategy advanced in Propositions 3 and 4, namely, that reliability was favoured over the optimizing of caloric returns. Finally, it should be pointed out that after the glacial climax of MIS 4 was reached, the climate began to shift in favour of warmer environments. As conditions ameliorated, more species, including plants, were likely to have become available to the Neandertals. Their tactics certainly should have changed as suggested earlier in Proposition 5. That is, when species diversity goes up, in general, the potential subsistence security also increases. Such an expected trend is, however, difficult to evaluate from the Combe Grenal data, particularly more recent data than level 11. Levels 1-10 at Combe Grenal were deposited against the cliff face after the rockshelter was filled with depositional materials. In other words, Combe Grenal became an open-air site, not a truly roofed shelter as it had been for so many years previously. It is likely that this change would have conditioned a very different pattern of use at the Combe Grenal location even if the warming trend documented at Combe Grenal had not occurred. In short, one of the features of the site of Combe Grenal changed dramatically. Prior to the formation of level 10, the “rockshelter” characteristic of the site was reasonably considered a constant. Not so for levels 1-10. They are most profitably compared to other open sites. Certainly the Mousterian of Acheulean Tradition (MAT) identity for the contents of most of the latter levels compares favourably with other “MAT” assemblages that are also most commonly known from open sites (Mellars, 2004: 34). I am confident that the issue of tool variability in the archaeological record at Combe Grenal will, upon further study, at least partially reflect variable food procurement tactics. Nevertheless, such suspected functional variability with tactical implications is a long way from placing us in a position to read off a “menu” of the Neandertal diet.

the diet of early hominins

213

Guts & Brains

28-03-2007

15:12

Pagina 214

Looking at the archaeological record on a more detailed scale, what was a meal like? What conditions change consumption patterns?

LEVEL

Much discussion of the Middle Palaeolithic era has focused upon scavenging versus hunting, and bias in the types of anatomical parts actually targeted for consumption by Neandertals. I will use a single indicator of bias in the anatomical parts introduced to the site of Combe Grenal. Bias in the anatomical completeness of animals introduced to the site could implicate diversity in diet among different active segments of the Neandertal population, gender differences for instance, and/or different conditioners for the differentiated use of the Combe Grenal shelter itself. Let’s see what we can learn. The measure employed is a ratio introduced into the literature by Stiner and Kuhn (1992: 318). It is simply the minimum number of bone elements indicative of the presence of heads divided by the minimal number of elements tabulated from the basic bone units from the remainder of the body. Stiner and Kuhn report that 0.30 is the expected value of this ratio for a complete deer skeleton. My data are tabulated in terms of proximal and distal elements for long bones and even phalanges, as well as distinctions between right and left. Because of this different organization

Combe Grenal

01 04 07 10 13 17 20 23 26 29 32 35 38 41 45.46 49 52 0

1

2

3

4

Ratio-Head to body parts (MNE) Reference lines are for the expected values for complete animals

Fig. 9 Combe Grenal: plot by level documenting the ratio head-to body parts (MNE). See text for explanation.

214

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 215

of the data, I recalculated the ratio using my system of tabulating bone elements (originally done for different reasons) for the total ungulate skeleton distinguished as my data was differentiated. I obtained an expected ratio indicative of a complete skeleton of 0.11. The other issue is teeth, and I accepted the suggestion that head parts only be estimated from anatomical elements of bone. This removed the preservation bias in favour of teeth when bone destruction could be a confusing issue regarding variability among deposits. The data used were for both reindeer and red deer, the two species of moderate body size exploited across the full range of environmental conditions documented in the deposits at Combe Grenal. Figure 9 is a bar graph plot by level documenting the proportion of head parts expressed as a ratio to the minimum numbers of elements of the head divided by the minimum numbers sum of all other skeletal parts. The graph has two references lines originating on the x-axis where the ratio values for head to body parts are plotted. The line to the left identifies the 0.11 ratio value indicative of complete animals being introduced. I have also included the 0.30 ratio value used as a standard by Stiner and Kuhn (1992) as mentioned above. It is strikingly clear that there are only two episodes during the entire approximately 60,000 years of Neandertal occupational use at Combe Grenal when complete animals were being regularly introduced to the rockshelter. The earliest is during the moderate but wet period between levels 55 and 49. These moderately thick levels almost certainly represent palimpsests of multiple use episodes. In spite of the likely lack of occupational integrity within each level, there does nevertheless appear to be a regular pattern of occupational re-use during the formation of these levels. At the same time, Figure 9 indicates that red deer were introduced to the shelter as complete animals in very few other levels (none more recent than level 49), in spite of the fact that red deer were present in all levels more recent than level 49. The earliest of the levels being considered here5 are all levels classified by F. Bordes as “Typical Mousterian”. They are also levels in which the sediments suggest extensive rainfall, particularly in summer. Levels 50 and 51 show evidence of increasing rainfall in all seasons. Thereafter, in level 52, this changes to equitably high rainfall all year around. The other levels in which the head/body ratio clearly indicates the introduction of complete animals to the site are levels 21, 22, 23, and 24. Besides representing the maximum cold achieved in Europe prior to the last glacial maximum at around 18,000 years ago, evidence from the sediments strongly supports a moisture regimen of high winter snowfall, followed in more recent levels (19-1) by relatively dry and warming seasonal conditions. We have seen that the level 23-24 comparison documents a major shift in hunting

the diet of early hominins

215

Guts & Brains

28-03-2007

15:12

Pagina 216

tactics as regards the age and maturity of animals taken at the very climax of the MIS 4 glacial event. What is common to the two tactically differentiated levels is the bias for complete animal carcasses introduced to the site regardless of the procurement tactics utilized. They were maximizing return rates when deciding what to transport to the site. It is also likely that the kills were very near the site, given the demonstrable homogeneity in transport decisions. The places in the sequence where the introduction of complete animals is recognized correspond to (1) the highest rainfall indicated at Combe Grenal (levels 5054) and (2) the coldest temperatures experienced during the accumulation of the deposit (levels 24-21), which also saw the greatest precipitation during the winter. I think that we may infer that shelter was the major factor favouring occupation under the climatic conditions inferred. There were severe weather conditions that would have limited mobility, yet in both situations whole carcasses were returned to Combe Grenal regardless of the species taken. Lithic industries were dramatic in their connection with these two uses of animals as well as the associated environmental contexts. Quina Mousterian was the type found during the cold period and Typical Mousterian during the moderate but very wet period. If we look at the other suggested contexts of site use at Combe Grenal as indicated, for example, by those levels where heads of animals are associated primarily with lower leg and foot bones, or other properties of the animals that may permit us to identify the provisional month of dominant occupation, we might obtain a further clue to diet as well as other aspects of Neandertal behaviour. Table 7 utilizes an estimate of the season of occupation developed from the tabulation of observations on both eruption and wear on teeth – properties which inform us about the animal’s maturational state. Eruption data are well documented regarding the age of animals when different teeth erupt and when milk teeth are lost. Wear on the latter as well as on permanent teeth, when a large sample of teeth from a single species is available, do provide the basis for estimating the average month of death for the assemblage of individuals represented. I should comment that without exception the clustering of age indicators informs us that occupations were relatively short and, when multiple in a given level, were redundant as regards the season of occupation when examined from a single species perspective. Table 7 is organized in terms of the percentage of teeth derived from the mandible divided by the total number of teeth from both the mandible and the maxilla. In short, the value is the percentage of the total teeth from the mandible. Low values indicate few mandibles with a corresponding dominance of maxillary teeth, and by implication, the complete head having been commonly introduced to the level. A value of 50% indicates that complete heads were present. Values lower than 50% signal a bias in favour of crania lacking mandibles. As suggested above, this is suggestive of scavenging with the brain being the component most likely to

216

lewis r. binford

Guts & Brains

28-03-2007

15:12

Pagina 217

Table 7. Mandible and cranial transport into Combe Grenal, seasons of deposition, and assemblage associations. ———————————————————————————————————————————————————————————

Animal

% of teeth = Mandible

% Season of death

Highest % of case types

Median % of case types

Low % of case types

———————————————————————————————————————————————————————————

Reindeer

86.4%

100% early spring

36.9% Quina

26.3% Typical

21.0% Denticulate

———————————————————————————————————————————————————————————

Red Deer

77.0%

66.6% winter, 38.2% early spring

46.6% Typical

20.0% Denticulate

16.0% MAT

———————————————————————————————————————————————————————————

Horse

82.0%

57.1% spring

29.0% Quina

29.0% Typical

29.0% Denticulate

———————————————————————————————————————————————————————————

Bovid

57.0%

50% spring 30% winter

46% Typical

23% Denticulate

23% Ferrassie

———————————————————————————————————————————————————————————

Note: In the case of columns 4, 5, and 6, the percentages do not add up to 100 because low frequency assemblage forms were not included in this table.

have been consumed. It is recognized that “other things being equal”, differences in the percentages inform us regarding (a) possible exploitation of scavenged carcasses, (b) biased exploitation of the brain – it is high quality food and good for baby food, (c) bias introduced through transport cost considerations, and finally (d) bias arising from limited food alternatives available to the Neandertals. “Other things” that could also condition variability in the table are factors not recognized in the above list that arise from environmental variability through time either seasonally or with respect to occupations that occurred at different time periods, each being conditioned in turn by variability in the particular species available for reliable exploitation from Combe Grenal. As an example of the phenomena suggested in “c” above, I offer some observations made among the Nunamiut Eskimo. The tongue, frequently butchered out with the mandible, is a choice meaty part, and is the part most commonly immediately processed and consumed in spring Nunamiut residential camps (Binford, 1978: 196). The skull is the anatomical part most commonly abandoned at Nunamiut kill or butchering locations in spring (Binford, 1978: 216). It may be cooked and eaten there as well as simply abandoned unprocessed at kill butchering locations. Another consideration is the weight and bulk of the head relative to the amount of consumable material. If humans-hominins are the transport agents, it seems unlikely when faced with a complete animal that the transport of the head would be a priority. We may speculate that, like the Nunamiut field parties, the head would be consumed very close to the kill or when it was one of a very few usable parts remaining of a carcass which was scavenged when food was

the diet of early hominins

217

Guts & Brains

28-03-2007

15:12

Pagina 218

scarce. I might add that such consumption is likely to have been by the persons obtaining access to the carcass. The transport of body parts for consumption by individuals other than those obtaining the carcass or parts thereof might be expected to be biased in favour of parts with larger quantities of meat and marrow relative to the bulk or overall weight of the part as butchered prior to transport. Similarly, differences in the body size of animals should condition the frequency of transport of complete heads, such that it would be more common with smaller animals, less common with large animals. The information summarized in Table 7 permits us to evaluate the potential relevance for some of the features found in the above suggestions, all of which derive primarily from ethnographic observations. Perhaps the first observation of interest is that season may have more to do with the return of heads as opposed to the relatively light mandibles with tongues probably attached than does body size. If the latter was the primary conditioner of transport, we should see less mandible dominance among reindeer, certainly the smallest animal on the list, and the greatest percentage of the teeth coming from the mandibles of bovids, the largest animals represented. Clearly, the reverse situation is what is documented at Combe Grenal. The highest percentage of mandibles derive from reindeer and the lowest percentage from bovids. Body size hardly seems to be the conditioner at Combe Grenal. The importance of season is further suggested by the comparison between reindeer and horses, where both show a high percentage value for mandibles relative to crania, and both animals were primarily killed in early spring or later in the spring. This is of course the season in the temperate zone when the least food is generally available for humans and other animals as well. Nevertheless, mandibles were returned to Combe Grenal with the highest frequencies among the four animals considered here. In spring when food is least abundant, we might have expected more marginal parts with higher transport costs to have been transported. They were not! It will be recalled that the spring exploitation of reindeer and horses was observed to have changed dramatically as regards the ages of animals taken and the tactics for killing them, namely the exploitation of reindeer at the calving area or at least very close to it, and a parallel exploitation of pregnant horses and newly born foals. The latter patterns were cited as supporting Proposition 4, which stated that Neandertal hunting strategies favoured reliable hunting tactics over optimal ones. In addition, there does not appear to have been any practice of storage, so very shortterm considerations of “abundance” may have ruled the economics of the Neandertal consumer population represented. Another possibility is that there was regular consumption by the entire local group at the procurement location. Instead of transport of parts to consumers not present in the “hunting party”, consumers may have travelled to the procurement location for initial consumption if relative-

218

lewis r. binford

Guts & Brains

28-03-2007

15:13

Pagina 219

ly close to Combe Grenal. If consumption preference was given to parts with high transport costs, as was for instance the case among the Nunamiut Eskimo, then transport to the “camp” would be biased in favour of parts with lower transport costs. The latter certainly seems to be the case at least as regards the head parts of animals. Examination of the red deer and bovid remains tabulated in Table 7 shows that red deer were killed primarily during winter and secondarily in early spring, while bovids were killed primarily in spring and secondarily in winter. The latter documents an overlapping shift from red deer in winter to bovids in spring. This is consistent with greater diet breadth expected in warmer environments, certainly a dominant condition during the warmer phases of occupation at Combe Grenal, when red deer and bovids were exploited most commonly. The information summarized in Table 7 demonstrates bias in the types of Mousterian assemblage associated with mandible dominated assemblages. The exploitation of reindeer was associated with Quina, Typical, and Denticulate in decreasing frequencies. Horse, on the other hand, was associated with identical percentage frequencies among Quina, Typical and Denticulate assemblage types. The occupations during warmer settings at Combe Grenal exhibited a different but consistent pattern of assemblage type associations. Such associations nevertheless differed consistently in the forms of assemblage found with red deer and bovids. Red deer were associated with 46.6% Typical Mousterian assemblages, 20.5% Denticulate and 15.4% Mousterian of Acheulean Tradition. Bovids (dominantly bison) were associated with 46.2% Typical Mousterian, 23.1% Denticulate, and 23.1% Ferrassie assemblages. These facts stand at present as interesting, but there is no clear causal basis given our current understanding of Mousterian assemblage variability. One suggestion, however, might be to explore the bias in tool-making potential (Levallois cores) being transported when game such as bison was being hunted. Bison cover much larger range areas than do red deer. These assemblage-related patterns do provide patterning from the facts of the archaeological record that is in need of explanation (Binford, 2001b). How does raw material variability relate to these patterns? How does raw material usage indicate differential ranges of land use, and of scales of landscape coverage? How do such characteristics relate to seasons of occupation? The research reported here represents the tip of the iceberg as regards research designed to expand our knowledge of the Mousterian. What can be drawn from this short exercise in looking at the archaeological record for clues to the Neandertal diet? For instance, what about consumption patterns? Are meals consumed by “families”, gender-differentiated groups, individuals, all of the above but in different social contexts? The organization of food consumption will certainly condition differential patterning in the archaeological record.

the diet of early hominins

219

Guts & Brains

28-03-2007

15:13

Pagina 220

Among most species, other things being equal, we can see parallel associational patterns where the males and females move together in feeding. Other species exhibit some degree of mutually exclusive feeding strategies, as seen among orangutan where males and females feed independently much of the time. This latter pattern has the advantage of insuring that males and females do not compete for the same food resources and/or patches. One can find examples of yet other species socially organized so as to prevent males from competing with their mates for resources. Some costs, either in mobility and/or mating opportunities, must accompany varying ways of organizing male versus female subsistence. Asking the rather non-specific question of diet, when we have little to guide us in thinking about the character of the consuming unit relative to the procuring unit, makes the challenge a bit stiff. I have every confidence that we will begin to be able to diagnose not only procurement strategies as illustrated here, but also how the producers and consumers were organized and what conditioned variability in the size of associational groupings. Understanding diet is a more challenging problem than arguing that the ancients killed animals, scavenged for food, and seemingly lived in minimal social units. The basis for the latter inferences already seem in place. Solving challenging problems like the issue of diet should make our understanding of the past much more sophisticated than what is currently available.

Notes 1. This case is the Guato who lived in a deep South American swamp and heavily exploited aquatic resources as well as swamp fauna. At the time of actual observation, they practised some horticulture. The subsistence values used were based on a “memory culture” reconstruction. Nevertheless, it is my impression that deep swamps do foster hunter-gatherer adaptation of some stability. It appears that horticulture was a very late arrival in the swamps of southern Louisiana, USA. 2. If so, they would not be an independent society but a specialist group within a larger cultural system. They would be more akin to caste groups than independent huntergatherers. 3. The comments have reference to the recent claims for the sensational wooden spears recovered from Schöningen, in Germany (Thieme, 1997). 4. The arguments by Trinkaus regarding Neandertal athleticism have recently been challenged by Lieberman – “Rethinking the Fossil Evidence for Behavioral Differences between Neandertals and Modern Humans”, paper delivered April 4, 2004, at the 69th Annual Meeting of the Society for American Archaeology, Montreal, Canada. 5. Levels which Bordes considered to be Riss in age and levels essentially eroded away prior to level 54 in Bordes’original numbering system are not included in this analysis and discussion.

220

lewis r. binford

Guts & Brains

28-03-2007

15:13

Pagina 221

References Bailey, H.P., 1960. A Method of Determining the Warmth and Temperateness of Climate. Geografiska Annaler 42(1), 1-16. Binford, L.R., 1978. Nunamiut Ethnoarchaeology. 1st ed. Academic Press, New York. Binford, L.R., 2001a. Constructing Frames of Reference: An Analytical Method for Archaeological Theory Building Using Ethnographic and Environmental Data Sets. University of California Press, Berkeley. Binford, L.R., 2001b. Where do research problems come from? American Antiquity 66(4), 669-678. Binford, L.R., 2002. L’interaction Homme-Ours dans les données ethnograpicques et les gisements européens d’ours des caverns. In: Tillet, T., Binford, L.R. (eds), L’ours et L’Homme, 141-156. Etudes et Recherches Archéologiques de l’Université de Liège. Université de Liège, Liège. Bordes, F., 1972. A Tale of Two Caves. Harper & Row, New York. Churchill, S.E., 1993. Weapon technology, Prey size, Selection, and Hunting Methods in Modern Hunter-Gatherers: Implications for Hunting in the Paleolithic and Mesolithic. In: Peterkin, G.L., Bricker, H.M., Mellars, P. (eds), Hunting and Animal Exploitation in the Later Paleolithic and Mesolithic of Eurasia, 11-24. No. 4 ed. Archaeological Papers of the American Anthropological Association, G.A. Clark, general editor. American Anthropological Association. Delpech, F., 1999. La Chasse au Bison dans le sud-oest de la France au cours du Wurm: choix humain ou contraintes paleoenvironmentales? In: Brugal, J.P., David, F., Enloe, J., Jaubert, J. (eds), Le Bison: giber et moyen de subsistence des hommes du Paléoindiens des Grandes Plaines. Actes du colloque international, Toulouse, 6-10 Juin, 1995, 63-84. Association pour la promotion et la diffusion des connaissances archéologiques, Antibes. Fa, J.E., Purvis, A., 1997. Body size, diet and population density in Afrotropical forest mammals: Comparison with Neotropical species. Journal of Animal Ecology 66(1), 98-112. Harding, R.S.O., 1981. An Order of Omnivores: Nonhuman Primate diets in the Wild. In: Harding, R.S.O., Teleki, G. (eds), Omnivorous Primates: Gathering and Hunting in Human Evolution, 191-214. Columbia University Press, New York.

the diet of early hominins

221

Guts & Brains

28-03-2007

15:13

Pagina 222

Mellars, P., 2004. Climate and the Upper Palaeolithic Revolution in Europe: Evolutionary Perspectives. In: Cherry, J., Scarre, C., Shennan, S. (eds), Explaining social change: studies in honour of Colin Renfrew, 27-44. 1st ed. Studies in Honour of Colin Renfrew. vol. 1, C. Scarre, general editor. McDonald Institute for Archaeological Research, Cambridge. Métraux, A., 1946. The Guato. In: Steward, J.H. (ed.), The Marginal Tribes, 409418. Bureau of American Ethnology, Bulletin No. 143. Smithsonian Institution, Washington, DC. Nietchmann, B., 1973. Between Land and Water: the subsistence ecology of the Miskito Indians, eastern Nicaragua. Seminar Press, New York. Roebroeks, W., 2001. Hominid behaviour and the earliest occupation of Europe: an exploration. Journal of Human Evolution 41, 437-461. Schmidt, M., 1942. Resultados de mi Tercera Expedicion a Los Guatos effectuada en la año de 1928. Revista de la sociedad cientifica del Paraguay 5(6), 41-75. Simoens de Silva, A.C., 1930. Une Moustiquaire des Indiens Guatos (Bresil). Paper presented at the Proceedings of the twenty-third International Congress of Americanists 1928, New York. Stearman, A.M., 1984. The Yuqui connection: Another look at Siriono deculturation. American Anthropologist 86(3), 630-650. Stearman, A.M., 1987. No longer Nomads: The Siriono revisited. Hamilton Press, New York. Stiner, M.C., 1994. Honor among thieves: A Zooarchaeological Study of Neandertal Ecology. Princeton University Press, Princeton. Stiner, M.C., Kuhn, S.L., 1992. Subsistence, technology, and adaptive variation in Middle Paleolithic Italy. American Anthropologist 94, 12-46. Thieme, H., 1997. Lower Palaeolithic hunting spears from Germany. Nature 385, 807-810. Trinkaus, E., 1983. Neanderthal Postcrania and the Adaptive shift to Modern Humans. In: Trinkaus, E. (ed.), The Mousterian Legacy, International Series, No. 154, British Archaeological Reports. Oxford University Press, Oxford. Trinkaus, E., 1986. The Neanderthals and Modern Human Origins. Annual Review of Anthropology 15, 193-218. Van Arsdale, P.W., 1978. Activity Patterns of Asmat Hunter-gatherers. A Time Budget Analysis. Mankind 11, 453-460. Zegwaard, G.A., 1959. Headhunting Practices of the Asmat of Netherlands New Guinea. American Anthropologist 61, 1020-1041.

222

lewis r. binford