Integrating Ethnoarchaeology

8/13/2019 Integrating Ethnoarchaeology

1/30

JOURNAL OF ANTHROPOLOGICAL ARCHAEOLOG Y 11, 33&359 (1992)

Integrating Ethnoarchaeology: A SubtleRealist PerspectiveNICHOLAS DAVID

Archaeology Department, University of Calgary, Calgary, Alberta, Canada TXV IN4Received April 8, 1992

Ethnoarchaeology is characterized by a diversity of products that often seemto have l i tt le in com mo n but which can for the most part be assigned to one or theother of two contrasting, although not incompatible, schoo ls: the scientist and thehermen eutic. A subtle real ist phi losophical perspective and Goodenoughs con-cepts of phenomenal and ideational orders are employed to situate ethnoarchae-ological researches in the conte xt of culture as a whole. It becom es apparent thatthe diversity of studies reflects the expanse of this domain, in the analysis of partsof which i t is appropriate to employ a variety of metho ds and analytical styles .Selected ethnoarchaeological studies are reviewed in demo nstration of the thesisthat the relative simplici ty or com plexity and openness or restriction of the sys-tem s under investigation are the main factors influencing choices o f metho ds andsty les . Ethnoa rchaeolog ical and archaeo logical interpretation should indeed par-take both of scientist explanation and herme neutics. Las t, i t is noted that, whileethnoarchaeology remains closely l inked to archaeology, i ts scope is widening toinclude a broader range of anthropological concerns that can be access ed throughmaterial culture studies. o 1992 cademic FTCSS, Inc.

There is a strong anti-philosophical strand in eth-nographic thinking that places value on the prac-tice and products of research and has little pa-tience with or interest in discussions about re-search. I have some sympathy with this

Hammersley 1992:43This is not a philosophical tract, rather an attempt to locate myself and

others in a fast changing disciplinary landscape. Ethnoarchaeology wasconceived to provide more focused and complete analogies for archaeo-logical application than could be gleaned from ethnographic and othersources (Kleindienst and Watson 1956). Some 35 years later, it is char-acterized by such a disconcerting variety of topics, approaches, andmethodologies that one may well ask whether it is not merely a grab bag:of questions raised, at least initial ly, by the interpretive needs of archae-

330027%4165192 5.00Copyright 0 1 992 by Academic Press, Inc.All rights of reproduction in my form reserved.


2/30

INTEGRATING ETHNOARCHAEOLOGY 331ology; of methods that run the ethnographic gamut from qualitative casestudy to quantitative survey; of philosophical attitudes that, while rarelyexplicit, range from positivist to idealist. It is symptomatic that it is dif-ficult to design a coherent university course on ethnoarchaeology.

Perhaps this is not so surprising; a recent book, Whats wrong withethnography? (Hammersley 1992), signals that one of our parent disci-plines is also in crisis. Hammersleys answer to the question is an at-tempt to integrate both the methodological ideas and methods that theterm ethnography normally refers to into a more appropriate method-ological framework for the social sciences (Hammersley 1992:203). Thebuilding of this framework is founded on a subtle realist philosophy ofscience, to be defined below. My aim in this paper is to extend a similarapproach to ethnoarchaeology in the hope of finding some underlyingunity that can accommodate a variety of problem orientations and re-search techniques. This is attempted not exhaustively but by reference toselected examples, several from research undertaken by members of theMandara Archaeological Project in North Cameroon (David and Sterner1987, 1989; David et al. 1991).

REALIST PHILOSOPHY OF SCIENCEMost (ethno)archaeologists avoid taking an explicit philosophical posi-

tion and are probably wise to do so (see Kelley and Hanen 1988). None-theless, we operate as if we had answers to the following questions. Whatare the things we study? Are there laws of human social life? Whatis explanation? How do we verify our explanations? Realist philos-ophy of science in its subtle form provides answers to these questionsthat serve to orient and direct scientific enquiry without either overlyconstraining its scope or forcing us into bogus intellectual gymnastics.

The realist research program in social science is conveniently set out byGibbon (1989:142-172), from whom the following summary of what arefor our purposes its essential elements is taken.

Realists distinguish between three domains:(a) the real-structures and processes, which are often unobservable

and may be complex stratified composites (e.g., genes, migration); any-thing that can bring about changes in material things is real;

(b) the actual-observable events and phenomena; compounds andconjunctures formed by the real; and

(c) the empirical-experiences and facts generated by our theory ladenperception of the actual.Scientists are concerned to identify, define, and explain things in the

domain of the real, We approach the real through our empirical reading ofthe actual, and according to the scientif ic knowledge of the day. It is this


3/30

332 NICHOLAS DAVIDthat distinguishes subtle from naive realism. Naive realists neglect. . . the way in which theoretical assumptions inform their descriptionsand explanations; they believe that they can have direct contact withreality and achieve knowledge that is certain (Hammersley 1992:32,50).

Society exists as a real object, a complex structure irreducible either toits effects or to people, consisting of the sum of relations, including rela-tions with material culture and the environment, within which individualsand groups stand. Society exists by virtue of the intentional activity ofpeople. It can only be detected by its effects; it generates social life, ismanifest in social behavior and its products, and is conceptualized in theexperience of its bearers. The causal power of social forms is mediated bypeople, and social forms are a necessary condition for social action. How-ever, human behavior cannot be determined by or completely explainedby reference to social forms and rules, because people are purposeful andpossess intentionality and self-consciousness. Psychological and physio-logical as well as social reasons contribute to intentional human behavior.Thus, people act in open systems codetermined by a variety of mecha-nisms of which the social is one. Societies are continuously being trans-formed in practice, are only relatively enduring, and are thus irreduciblyhistorical.Explanation of social phenomena proceeds by the same general processas in the natural sciences:

1. Recognition of a pattern and resolution of events into their compo-nents. Events are viewed as conjunctural, the results of the combinedeffects of a variety of active structures.

2. Redescription of events in the language of social science.3. Creative model-building, the search for generative mechanisms that

might produce the observed pattern. This is an inductive, or more pre-cisely a retroductive attempt to lay out the structural conditions that musthave existed for the events to be present.

4. Theory construction. Candidate mechanisms are reduced to one asthe reality of their postulated structures and powers is checked, in part byevaluation of each mechanism in terms of its plausibil ity and credibility inthe light of other theories, especially those that we currently take to bebeyond reasonable doubt, in part by gathering of independent evidencethat will subject the theory to maximal threat. If they pass these tests,structures and their workings can then be defined in the form of causallaws. However, because of human intentionahty and the openness ofcultural systems, causal laws describing the way real social things operatemust be analyzed as tendencies that may or may not be expressed inparticular historic conjunctures.5. Exploration of the stratum of reality revealed in the previous stepscan now begin.


4/30

INTEGRATING ETHNOARCHAEOLOGY 333As social structuresare only manifest n open systems,decisive ests of

theories are impossible. We may be able to explain precisely and accu-rately, but our capacity to predict remains udimentary. While the validityof claims is subject to stringent criteria of assessment, roposeddefini-tions of the real and theories about the nature of society are ultimatelyaccepted or rejected on the basis of their explanatory fruitfulness orpower. It is this rather than predictive accuracy hat decideswhich of aset of competing models becomes, or the time being, theory.To this partial and abbreviatedaccount of realist philosophy in socialscience, I would add a rider. Although humans live in open systems,certain aspectsof their behavior aremore constrained hanothersand canbe conceptualized n terms of, if not closed, at least restricted systems,some of which are quite simple. Much of the behavior studied by eth-noarchaeologistsand archaeologists, ool-making, for example, or sub-sistence,can be viewed in such terms. The simpler the system and themore it approximates o the closed condition, the greater he predictabil-ity of associatedbehavior. Thus, dependingon the naturesof the realthings and system under investigation, different approaches nd meth-odologiesmay be appropriate,and the ability to predict may be greaterorless.It is becauseethnoarchaeology nd to a much greater degreearchae-ology deal with material things and their relations to both the natural andthe social environment that there is such disagreement etweenauthori-ties as to what we shouldbe doingand how we shoulddo t. Binford (e.g.,1977,1982,1987) elieves that archaeologyshouldbe more like a naturalthan a social scienceand advocates he developmentof middle rangetheory to relate the statics of the archaeological ecord to the dynamicsof the living, systemic context, while Hodder (1986:11 3-146),ho seesarchaeologyas social science,recommendsa contextual approach nwhich context-specific structuring principles replacecross-culturally ap-plicable middle range heories.As Kosso (1991) as convincingly shown,there is little difference n the epistemic structure of their approaches.,*Middle range heories relate the empirical, the perceived actual, to thereal; so do Hodders structuring principles, which may indeed be re-gardedas middle range heories. The content and ustification of theo-ries are [in both cases] strongly influenced by observations,and in turnthe informational content and ustification of observationsare influencedby theories (Kosso 1991:625).Where Binford and Hodder and other(ethno)archaeologists rimarily differ is in (a) the behaviors, relating tomore or less open or restricted, simpler or more complex, systems in

* See No tes section at end of paper for al l footnotes .


5/30

334 NICHOLAS DAVIDwhich they are most interested, (b) in their views of the real things thatstructure these behaviors, (c) in their understandings of what constitutesexplanation and verification, and (d) the corresponding scientist orhermeneutic styles of their arguments.2 I use the terms scientist andhermeneutic to contrast styles of analysis. Each encompasses a range ofphilosophical and methodological positions, and indeed individual studiesoften show blends of the two. Scientist analyses are characterized bypreferential use of quantitative approaches and statistical inference, afocus on behavior rather than its meaning, an emphasis on the context ofjustification over that of discovery, on deduction as against induction, andon cross-cultural laws (whether deterministic or probabilistic) rather thancultural patterns. Hammersleys (1992: 159-173) chapter, Deconstruct-ing the qualitativequantitative divide, is a valuable in-depth treatmentof these apparent polarities.

The distinction between the schools is made clearer by specification ofthe cultural domain that is our object of study.

THE CULTURAL DOMAINThe cultural domain is represented in Fig. 1, at the center of which is

what Goodenough (1964: 11) termed the phenomenal order. This is a stra-tum of real things, activities, and patterning, an abstraction from behav-iors empirically observed in living cultures. As a soil may be representedas constituted by different proportions of sand, silt, and clay, so certainbehaviors are classified as activities that may be categorized in terms oftheir ideological, social, and technical aspects.3 Certain activit ies are re-sponsible for that part of the material output of the phenomenal order(henceforth PO) in which archaeologists are primarily interested, that is tosay artifacts or, more precisely, things affected by human action anddifferentiated by matter, form, and context. A small portion of these isrecovered as the archaeological record.

POs exist as part of an interacting dyad, the other element of which isthe ideational order (Goodenough 1964). This is made up of another orderof real things, unobservable ideas, values, norms, representations, and isitself underlain by a psychocognitive order of reality, the hard wiring ofthe human mind. The ideational order (henceforth IO) is expressed in,although it does not wholly determine, the activit ies and their patterningdefined as the PO. While all 10s are structured by certain real thingsrelating to the biology of modern Homo sapiens sapiens, they vary as aconsequence of differing historical trajectories, varying interactions be-tween the dyadic complex and the environment, and within and betweenthe IO and the PO themselves. Furthermore, since things in the IO do notinteract directly with the environment, they are less constrained by it than


6/30

r

a

h

INTEGRATING ETHNOARCHAEOLOGY

LOCUSIDEATIONAL ORDER Minds

Exchanges ofenergy andinformation,materialoutput

Behavioral archmological

\ARCHAEOLOGICAL RECO RD Ground

335

------Increasing con straint on variation, -----+decreasing openness of system sHer me neu tic -------- Ana lytical style ---------- Scie ntist

FIG. 1. The cultural domain and i ts relationship to interpretive approaches and analyticalsty les.

behaviors in the PO, which, particularly in its technical aspect, comprisesrelatively restricted subsystems. 10s are open systems characterized bysome fuzziness and internal contradiction, competing representations ofreality held by different individuals, genders, age groups, craftspeople,ranks, classes, and so on. The relations between IO and PO influenceactual behavior and thus its material output. It follows that one cannotreconstruct a PO from cultural events without reference to the IO. Thetwo form a hermeneutic circle: system is the source for understandingstructure, and structure is the background for understanding the system,as Kosso (1991:624) puts it.

An archetypal Binfordian scientist approach in ethnoarchaeology in-


7/30

336 NICHOLAS DAVIDvalves attempts to define cross-cultural regularities or lawlike statementsderived from ethnoarchaeological and other, including ethological, evi-dence, in order to apply these to the reconstruction of past POs and theirdevelopment, and to generate new regularities regarding culture change.Before these regularities can be applied to the archaeological record, theymust first be reformulated, in the terminology of behavioral archaeol-ogy as cultural (C-) and natural (N-) transforms, the former theoreticalconstruct relating observed remains to unobservable past behaviors, thelatter variables of culturally deposited materials and variables of thenoncultural environment in which the former materials are found (Schif-fer 1976:14-16). Archetypal Hodderians on the other hand dispute theexistence of most cross-cultural regularities and thus their value in pre-diction. To them, artifacts are not only instruments but also signitiers ofsignifieds that exist at the IO level. Their informational content musttherefore vary as a function of the IO-PO complex that generated them.An axe mealzs something very different to a Canadian lumberjack, aMinoan, and a Stone Age Australian. Hodderians are therefore con-cerned to reconstruct specific IO-PO complexes and their diachronicdevelopment from the inside with, at least in theory, minimal appeal tocross-cultural-as opposed to cross-cognitive-regularities. To discusshumanity in terms of general laws, is ultimately to deny people theirfreedom (Hodder 1986:102) is a fine example of Hodderian oratory,although one which he has denied in practice (Hodder 1982:125-184).Hodderians emphasize the irreducible historicity of societies.

Binfordians and Hodderians are thus primarily interested in definingdifferent sorts of real things, Binfordians the regularities realized in socialand technical behavior in environmental context, Hodderians the struc-tures and principles that underlie ideological and social activities in IOcontext. There are only certain ways in which hunter-gatherers can ex-ploit a tropical rainforest and survive, and the freedom to devise toolkitsand form social groupings is limited. Their religious symbolism, on theother hand, is comparatively free to vary. The greater the constraintsacting on behavior, the less open and the simpler the system, the morepredictable it becomes, and the greater the potential to derive from it across-cultural regularity with predictive power. This being the case it isscarcely inappropriate, at least in the first instance, for Binfordians towork in a style inspired by natural sciences in which variables can oftenbe str ict ly controlled. Where, due to the greater complexity and opennessof systems, variables are much harder to control, much less quantify, it issimilarly appropriate for Hodderians to grapple with their materials inhermeneutic or symbolic interactionist mode.4 Binfordians and Hodderi-ans strive for different things, the former explanations in terms of laws


8/30

INTEGRATING ETHNOARCHAEOLOGY 337that apply cross-culturally, the latter understandings n which laws areapparent, f at all, only as tendencies n sets of historical conjunctions.Neither account of sociocultural ie is complete without the other.

MECHANISMS AND PROCESSES, THE PARTICULAR ANDTHE GENERAL

This contrast between schools, simplified though t may be, providesthe basis for a considerationof ethnoarchaeology,he study of the rela-tions betweenmaterial things and the I&PO complex in living cultures.Ethnoarchaeology nvolves field study of the production, typology, d is-tribution, consumption, and discard of material culture, with particularattention to the mechanisms hat relate variation and variability to socio-cultural context, and to inference rom the mechanisms o processesofculture change.Mechanisms are here defined as configurationsof thefull range of environmental, material, and sociocultural variables thatinteract at one ime horizon to generate atterning n material culture, andprocesses as diachronic changes n mechanisms ncluding their struc-tural change,breakdown,and transformation nto others.

Archaeologists are eagerconsumersof cultural laws, regularities, ortendencies n relationshipsbetweenmaterial and total culture that can beapplied retrodictively to the archaeological ecord, especially those pro-vided by middle-level theories, generalizations hat attempt to accountfor the regularities that occur between two or more sets of variables nmultiple instances (Trigger 1989:21). hey vary from the particular tothe most general. As an example of the first take Yellens (1977:130)equation: The number of days a Dobe-/DuDa Kung camp has been oc-cupied = 0.1 (Area in square meters of the Absolute Limit of Scatterminus the Limit of Most Scatter) + 1.87.Such a formula is of course ofrestricted value in archaeologicalprediction. Note that it refers to thesocial and technical. As a regularity at the other end of the scale,relatingPO to IO, consider the following formulation (suggested y a readingofKramer 1985:88),Typology, in the generalsenseof patternedvariationin classes of material culture, reinforces principles of social structure,including gender and power relationships, and reifies other aspectsofworld view. This candidate aw, with which Carol Kramer herselfwouldprobably not agree, s applicable o architecture n Westernsociety, or toceramics n the Mandararegion of North Cameroon David et al. 1988),and, in my opinion, offers a valuable, f only too rarely utilized, approachto interpretation of archaeological nd ethnographic ata of all times andall places.Archaeologistsand ethnoarchaeologists re faced with inverse obser-vational problems. Archaeological data have a temporal dimension and


9/30

338 NICHOLAS DAVIDmanifest the workings over time of cultural processes. The interpretationof a given body of archaeological data must necessarily account for con-tinuity and change through time in terms of both mechanisms and pro-cesses. Neither are directly observable. Ethnoarchaeologists work in theethnographic present. While they can infer from their observations theexistence of mechanisms, they lack opportunities to observe, except overshort time spans, the material manifestations of processes. Ethnoarchae-ological inference to process therefore takes the following form. A mech-anism can be considered as an interplay of variables structured by arelational system that generates a material output. Although we cannotobserve them, we may imagine the effects of changes in the value of oneor more variables on output. As an example I might point to the possibleoutcomes under varying conditions of the socioeconomic mechanism un-derlying the production of Fulani pottery (David and Hennig 1972:26).Within an Islamic context in which male authority over women wasmarked and there was great desire for children, household instability andincome, influencing the desirabili ty and accessibili ty of nonpottery sub-stitutes, were deemed to be critical variables affecting pottery supply andquality. Such exercises, in which the observers imagination interactswith informants accounts of past events and future plans, can extend thedata span to several decades while remaining anchored to reality via theinferred mechanism. More difficult is determination of the boundary con-ditions under which a mechanism will continue to operate without sys-temic change in its nature. In the example cited, changes as different asacceptance of Christian attitudes to monogamy or the eradication of sex-ually transmitted diseases would both result in systemic change (seeDavid and Voas 1981). More difficult still is imagining mechanisms thatexisted in the past but that have no even fairly close analogues today. Thisis of course not str ict ly the task of ethnoarchaeologists, although I wouldargue that our contribution to the world store of analogies is a fertilesource of inspiration in this area.

An ultimate aim of Binfordian archaeology and ethnoarchaeology is toarrive at general laws regarding the nature of cultural variability andchange. While the ultimate aim of Hodderians is different-to renderintelligible the human condition-there is always tension between theparticular and the general, any particular interpretation ultimately relyingupon assumptions regarding the nature of culture and of culture change.Developments in interpretation occur as the result of a dialectic betweenthe particular and the general that, whether explicitly or not, takes theform of retroductive model building and the testing, formal or infor-mal, of hypotheses derived from them against a database. While ethnoar-chaeology is of necessity particularistic, every effort should be made todraw out the broader implications of each study. It is here that, for what-


10/30

INTEGRATING ETHNOARCHAEOLOGY 339ever reason, many ethnoarchaeological tudies eave us unsatisfied (seeSchifTer 1978).

ETHNOARCHAEOLOGY IN SCIENTIST AND LOGICIST MODEIn Archaeological approaches to the present Yellen exemplified andevaluated our types of archaeological se of ethnographic ata: what hetermed the laboratory, buckshot, spoiler, and generalmodels.6The mainaim of his book was the developmentof transforms elating the dynamicsof San behavior to the statics of their abandoned amps.His Ring Model

is a general model of a regularity of relevance o the interpretation notonly of Kung camps, but also, with modifications, of any communityorganizedon a circular plan. Although length of stay was an importantvariable, Yellen was not much concernedwith processas defined above,and the model relates to the social and technicalaspectsof the PO, treat-ing the IO as irrelevant in this context. The symbolic significanceof theSan hut did not enter nto his equation.Neither, one might add, do econ-omists consider the symbolism of sports cars when predicting GeneralMotors third quarterearnings.That economistsare better at explanationthan prediction points to the openness nd complexity of economic sys-tems and to the exclusion rom them of any consideration f the ideationalorder. [Tlhe tastesandhabits of the consumer, indeedmost of culture,are regarded s givens(J. M. Keynes 1939:245, ited in Herskovits 1960:47) and this constitutes a real weakness.The foregoingcomparisonalsohelps us to understandwhy it is appropriate that the scientist style ofYellens analysis s modeled on that of economicsand much natural sci-ence.Comparable although less quantitative in its approach s a study byA.-M. and P. Petrequin(1984)of Tofftnu villages built in and aroundLacNokouC in Benin. This was undertakenwith the aim of improving andcontrolling interpretations of housing n Alpine and other European a-custrine habitats. Variations in architectureand community plans are setin their geographic,cultural, sociopolitical, and economic contexts, andclose attention s paid to differential processes f fossilization of struc-tures and features in relationship to their dry land, liable to flooding,emergingat low water, and fully aquatic situations. The study spans heuncertain boundary between C- and N-transforms and, as might be ex-pectedgiven the marked cultural contrastsbetween he ethnographicandarchaeologicalsituations, its major contribution is taphonomic.Inferences regarding ength of stay at a Kung camp site and as towhether a particular dwelling was built over openwater, in a marsh, or ondry land differ in their applicability to the archaeological ecord. In theformer case he predictive value of the inference ests primarily upon the


11/30

340 NICHOLAS DAVIDextent to which specific sociocultural and economic behaviors are com-mon to the ethnographic instance and the archaeological case under in-vestigation. As such commonality is very hard to demonstrate in theabsence of witnesses, the numerical value of the equation even in aBotswana Later Stone Age context can only be considered as broadlyindicative. Its real value lies in its contribution to research design. In thecase of the Petrequins study, the predictive value is greater preciselybecause, while cultural behaviors are of course implicated in productionof the data, natural processes (wave action, decay under aerobic andanaerobic conditions, fires, the falling of objects in water, etc.) and me-chanical contingencies such as susceptibility of objects to comminutionby trampling, largely determine the patterns of deposition that are thesignatures of different settings of structures.

It is much easier to reconstruct taphonomic, that is to say largely nat-ural, processes such as these than cultural behaviors that are less imme-diately constrained by physical laws, and that are not amenable to exper-imental confirmation. The systems differ in their relative simplicity andopenness. Ease of validation of inferences regarding cultural behaviorvaries inversely with (a) the extent to which that behavior is controlled byproperties of matter and (b) the specificity of the behavior in question.Thus, we may well be able to reconstruct details of pressure flaking in theproduction of arrowheads, but it may be foolish to characterize the re-mains of a compound as the residence of a polygynous family (see David1971).

Stil l in the scientist mode, the marriage of ethnoarchaeology and ex-perimental archaeology, advocated by Tringham (1978) and others, seeksto capitalize upon the enhanced potential of the combination for securinginference. The recent and ongoing work of Longacre (1991a,1991b) andhis Kalinga Ethnoarchaeological Project team in the Phillipines providessome nice examples. Employing an amalgam of both approaches, theseresearchers have achieved a good understanding of the surface alter-ations [of pots] as a direct reflection of a variety of cooking and cleaningbehaviors, and of material and functional characteristics that have onthe one hand influenced the replacement of certain kinds of pots by metalanalogues, and on the other have encouraged a trend toward specializa-tion in pottery production (Longacre 1991b:78). While we must admiresuch work, we must also be conscious of its limitation in scope to thesocial and technical, and to the mechanistic rather than the processual,and in the range of cultural behaviors to which it is applicable.7 We haveonly to remember Goulds (1980:141-160) demonstration that Australianaborigines on occasion used poor quality stone to manufacture tools be-cause they were righteous rocks, to realize that cultural mechanismsstrongly influenced by the IO might very well lead to the monopolization


12/30


13/30

342 NICHOLAS DAVIDmethods both require to be supportedby an analysis of the third type.This involves situation of the regularity in an ethnological context thatdefines he limits of applicability of the regularitiesproposed.These imitsare expressed n the form of hypotheses. nferenceswill be valid only ifthe mechanisms rom which the transculturalrules have been developedwere in fact operating n the segmentof the archaeological ecord underinvestigation.The regularity derivedby Gallay from his ethnoarchaeologicalesearchis that:ifpotters tools are found in a compound, the larger part of the asso-ciated pots is representative of a homogeneous ethnic tradition. Thisrequiresestablishmentof the truth ( instantiation)of a number of hypoth-esesregardingcontext, as follows:

1. There are distinct sympatric ethnic groups.2. Production of ceramics s a specialistactivity.3. There are distinct ethnic traditions of pottery manufacture.4. Ceramicproduction covers he whole rangeof needs hat require theuse of pots.5. The economy is characterized by markets (marches periphe-riques).6. The modesof distribution and purchaseof pots result in mixtures ofpots of different ethnic traditions.Again according o Gallay, application of the regularity proposed o thearchaeological ecord requires hat one s able to demonstrate he follow-ing propositions.

1. Pots in potters compounds epresent heir own work.2. Pots of other traditions occur in limited numbers,3. Potters compoundsare identifiable.4. Permanentsettlementsare identifiable.5. Potters tools vary by ethnic group.6. Technological and morphological variation in pots can be linkedwith ethnically varying toolkits.I treat this example at some length for a number of reasons.First, asGaIlay acknowledges, he hypotheses nd propositionsare open to chal-lenge.Not only do certainhypotheses ppearunnecessarye.g., 2 and 5),and others missing (e.g., that potters of different ethnic groups do notcoreside),but also it would be quite impracticable o instantiate a clusterof suchhypotheses n any archaeologicalnstancewhere here s not otherconfirmatory evidence e.g., ethnohistoricsourcesor oral traditions). Asfor the hypothesesso for the propositions. Second, the regularity pro-posed is trivial, and, while this is perhapsnot inherent in the method,


14/30

INTEGRATING ETHNOARCHAEOLOGY 343neither s it, I contend,a particularly productive function of ethnoarchae-ology to provide analogicalsupport for reconstructionof detailed socio-cultural arrangementson specific sites. Gallay (1989:31) s regretfullyprepared o limit the interpretive ambitionsof his school, we shall prob-ably have to definitively renounce rying to attain certain ambitious [ar-chaeological]objectives,particularly thoseconcerning he organizationofsociety or religious beliefs. This would seemunder this methodology obe as true for the discovery of ethnoarchaeologicalegularitiesas t is fortheir application to the archaeological ecord. In which case et us aban-don this alchemy (Gardin 1980:173) nd seekanother way.

Neither logicist nor scientist analysis (and the style of fieldwork withwhich the latter is generally associated) s appropriate to the ethno-graphic (asopposed or example o the experimentalpsychological)studyof mechanisms nvolving the ideational order. Whereas, n a computerprogram, constructs and their relations are precisely defined within aclosed system, human activities are interdependentwith the IO, whichnot only incorporatesa degreeof contradiction but a lso symbols that areprotean and polysemic. To logicist-positivist deconstruction must beaddedhermeneuticsThere are further conclusions o be drawn from this attempt to injectlogicist rigor into our practiceof inference.While it may neverbe possibleto instantiate the hypothesesand demonstrate he propositions requiredto establish he working of nontrivial mechanisms n the past, logicists atleast make their reasoningexplicit and, by this courtesy, open up thepossibilities for constructive debate.Last, rigorous methodologycannotcompensate or inadequate esearchdesign.

TOWARD HERMENEUTICSWe shouldnot be imiting ourselves o the study of mechanisms nvolv-ing the social and technical but should also attempt to comprehenddi-mensionsof variability in IO-PO complexes,both in particular nstancesand in general. To do this the ethnoarchaeological trategy must be toemploy a hermeneuticstyle in the developmentof theoriesregarding heexistenceof IO constructsandprinciples and heir working in the PO. Thetheoriesmust then be validatedby assessingheir internal plausibility andcredibility in the context of (a) the analytic model and other theoriesand(b) empirical constraints mposed by data that are su&iently indepen-dent of superimposednterpretation o challenge, orce revision and evenrejection of the theoretical constructs ntended to explain them (Wylie1982:42&l). A useful criterion is the extent to which proposedconstructsand principles can interrelate and account for variation acrossclassesofmaterial culture. While logicism is not directed to the context of discov-


15/30

344 NICHOLAS DAVIDery-logicist programs of fieldwork are likely to focus on questions oftechnology(e.g., Roux 199O)-our texts will be available or logicist anal-ysis, preferably (but rarely) by ourselves and before publication. Butlogicist explicitation is not prerequisiteeither to practical understandingor to insightful interpretation.Since 1984membersof the MandaraArchaeologicalProject team havebeen studying the culture history and ethnoarchaeology f the Mandarahighlandsand surroundingplains of North Cameroon,a regiondesignedlychosen or its maximal topographical,social, and ethnic variety (Boutrais1984).Within the 7000km2 in which, between 1986and 1990,eight eth-noarchaeologists penta total of 80 monthsaccumulatinga very largedatabase (see David and Sterner 1987,1989), here are societies descendedfrom both precolonialstatesand small-scale nhierarchized ommunities.At least 28 mother tonguesare spoken, mainly of the Central branch ofChadic (Fig. 2). A primary aim of our research s the developmentof atheory of style of general, ndeedbroadly predictive, value in ethnologyand archaeology.We have chosen his focus because he conceptof style

KANU3 IANU3 I

L _ 2km ME LGV A

FULBE

IGISIGA &

FULBE EM -FIG . 2. Parts of northeastern Nigeria and north Cameroon , showing ethnolinguisticgroups.he cross-hatched areas in the northeast represent the Bama ridge. (Map by A. S.MacEachem.)


16/30

INTEGRATING ETHNOARCHAEOLOGY 345is a key to all archaeological nterpretation that is concernedwith thebehavior of humans as members of sociocultural groups. We are inter-ested both in its instrumental aspect, ts uses as a medium of communi-cation, and in accounting or the specific forms style takes in Mandaracultures. We have emphasized tyle and stylistic variancewithin domainsof material culture chosen n part for their intrinsic interestand n part forthe likelihood that they will be represented n the archaeological ecord.These are architecture, ceramics, metallurgy, and mortuary practices.While some researchers ave undertaken n depth studiesof single com-munities (e.g., Gavua 1990), therswere ess narrowly focused.The com-bination of the two approaches rovides a broad comparative base foranalysis of the PO and IO (seeDavid et al. 1991).We take an isochrestic(Sackett 1990) pproach o style, accepting hatit may reside n aspectsof form either that are adjunct to utilitarian func-tion or that involve choice betweenviable functional alternatives.A styleis a polythetic set of attributes presentby virtue of common descent roman identifiable artifact-production system (David 1990: 9). This impliesthe fundamental mportance of studying modalities of production.One basic finding concerns he useof style and is supportedby Wiess-ners (1983)study of Sanprojectile points. In the absence f corporationsresponsible or the production of a material culture set, the vast majorityof stylistic signaling s directedby group members o themselves,servingto reinforce group identity, structure, and values (Sterner 1989;contraWobst 1977).Where corporations hat control the production of materialculture exist, style is not only manipulated or a variety of ideological,commercial, and other ntrasocietalpurposes,but is liable also o be usedfor organized external messaging.This general transcultural regularitymust be taken into considerationwhen making use of typologies or pur-posesof inference o socioculturalmechanisms. t is partly for this reasonthat the reconstruction of manufacturing sequences, chaines opera-toires, has becomesuch a focus of research n France (e.g., Binder andPerlbs1990), nd he conceptof technological style (Lechtman 1977)nthe Americas. In our area, Robertson in press)has shown how the guildof Wandala smiths uses metal-and not only mental-templates to pro-duce varieties of hoe forms designed or different ethnic markets. Suchpackaging is common; style can only be meaningfully nterpreted n thecontext of the production system responsible or its expression.Archae-ologists can scarcely secure nferences egardingexchange, rade, or anyform of social differentiation without reference o modes of production.Production systems mediate the material culture instrumental in theconstitution of ethnicity, defined here following Eriksen (1991)as theaspect of behavior that communicatescultural similarity and difference.Much if not most archaeological nterpretation has as a prerequisite he


17/30

346 NICHOLAS DAVIDdefinition of cultural groupings. While in our area it is possible to delimitethnolinguistic groupings on a map and to assign individuals to them foradministrative and other purposes, this distorts cultural reality. Individualethnicity is in fact situational and experienced at several levels (David etal. 1991). Thus, instead of discrete, formally differentiated, and geograph-ically bounded assemblages of material culture, we find zones with fuzzyboundaries, within and to a lesser extent between which, at the levels ofthe lineage, the multiclan settlement, the political alliance, the dialect,sometimes the language, there takes place a continuous, kaleidoscopicreassembly of forms. Each of these levels may be associated with an IO,but individuals, depending upon the context in which they find them-selves, may relate to more than one, possibly several, the distinctionsbetween which are by no means clear. Borrowing, reinterpretation, in-version, montage, and jumbling of attributes and motifs create a materialworld that is both the context that renders interaction comprehensible,and is itself constantly being invented through that interaction. Whileethnographers are very familiar with such situations (Sanjek 1991:622 andreferences there cited), the very nature of the archaeological record, com-prising artificial and static segments of cultural reality, has led archaeol-ogists first to reify aggregates of materials as archaeological cultures andthen to puzzle over their relationships to ethnographic analogues. It is infact rare for either to exist as a bounded entity.

Although the history of the Mandara has exerted particular influenceson stylistic expression, the pattern described is by no means unique butcharacterizes other societal and cultural groupings. Responding to envi-ronmental, economic, social, and political change, culture sets experiencea continual flux of persons and ideas, moving between cultural subsetsthat in order to function must both be differentiated and mutually intelli-gible. We can account for continuities in the material expression of suchpatterns in terms of the whole IO-PO complex. The environment ofcourse imposes significant constraints on the form of many artifacts, butother formal aspects including decoration, the aspect apparently leastrestricted, often show very limited variation. In all societies a significantportion of stylistic expression refers to and embodies long-lived cosmo-logical and religious themes existing at the IO level and expressed pri-marily through ideological activities. Here style operates as a potent agentin the recreation of society and the instill ing of its values. Thus, we canaccount for continuity and standardization through space and time and,incidentally, justify-within a certain societal range-the commonplaceuse of decoration as the prime index of ethnicity preserved in the archae-ological record (David et al. 1988). In recognition of the connection be-tween style and societys ultimate concern, its religious substance, weborrowed from McIntosh (1989) the concept of symbolic reservoir to


18/30

INTEGRATING ETHNOARCHAEOLOGY 347describe he long-lived Saussurianangue of founding ideas which aremade manifest in behaviors that include artifactual paroles that varyclinally and stepwise n spaceand time while remainingrecognizably hesame(Sterner n press).McIntoshs symbolic reservoir n fact comprisesmuch of the content of our IO. Such symbolic reservoirs, maintainedbythe participation of individuals in a variety of ethnic contexts, are he truereferentsof many cultures, ethnographicand archaeological,and thepopulation that draws on one symbolic reservoir s the signiticantunit forunderstanding ultural processat the macroscale.Hodders (1990) ecentbook, The domestication of&rope, is on precisely this subject, the evo-lution of a Neolithic Europeansymbolic reservoir.It may be nstructive to considerhow we arrived at this conclusion, onethat relates O to PO. First we carriedout archaeological urveysand testexcavation, which gave us some imited insight into regional cultural de-velopment during the Iron Age. We then carried out a combination ofcomparative and in-depth ethnoarchaeological tudies. The scale of ourwork through time and across space helps free us from the constraintdescribedby Wobst (1978) s the tyranny of the ethnographic ecord.We obtained evidence of synchronic and diachronic patterns n the dis-tribution of material forms-and particularly ceramics which are greatlyelaborated n the Mandara-and set hem in context by reference o socialandcultural anthropologicalandethnohistoricsources.Someof these, orexampleJuillerat (1971), aveus early insights nto mechanismsand pro-cess hat we were able to follow up and confirm. It is the mutual supportprovided by all theseapproaches-and not rigoroushypothesis esting orlogicist specification-that led us to our conclusions.The extent to whichwe are justified in generalizing rom the Mandara to other regions andareasdepends n large part upon the plausibility of our case n the largercontext of anthropological heory, andon comparativestudies.Validationof our conclusions must include study of other areas, of which morebelow.The discussion to this point has beenconcernedwith mechanismandprocess,with dimensionsof variability without regard o cultures contentof meaning.To what extent can we, in Hodders (1986:6) hrase,cometo an understanding f eachcultural context in its own right, as a uniqueset of cultural dispositionsand practices. Hodder s concerned ere withan entirely legitimate subject of archaeologicalenquiry, one that mustinvolve some application of regularitiesof IO-PO behavior to the partic-ular case. Having struggled over many months and with only partialsuccess o gain some intellectual appreciationof the Mandara symbolicreservoir, I am pessimistic as to prehistorianshopesof achievingaccessto complex structuresof prehistoric meaning, ar less to any deep empa-thy with 10s. In our work we have approached he study of meaning


19/30

348 NICHOLAS DAVIDthrough contextual analysis, that is to say by tracking and interrelatingforms and motifs as they appearand reappear n a variety of domainsoflife and classes f material culture. In the courseof ethnographic nquiry,perhapsduring discussionwith my assistant,E. Isa Kawalde, as we walkback through the terracedand domesticated andscape fter visiting witha smith, perhaps n the course of argumentwith Judy Sterner when shereturns from her village, ideas are constantly being proposed, aban-doned, developed,refined, and tested on their way to becomingmodelsand theories(cf. Wylie 1982:45-46). his may work well when research-ing the dynamic present, with recursive access o informants and mate-rials, but the cultural and natural transformsaffecting material remainsasthey enter he archaeological ecordresult, except n very specialcircum-stances, n the collapseof the contextual rames hat renderethnographicanalysis possible and fruitful. Paradoxically, it may be that the morecomplex the society, the richer its iconography, the more capital andlabor expended on its material culture, the better are prehistorianschancesof penetrating he ideology, at least of the elite, through herme-neutics. Inversely, it may be that at the other end of the societal scale,restriction of rangeof imagesand metaphor, and omnipresence f thoserelating to the human body and ts processes,will facilitate the approachto meaningcontent.Attempts by archaeologists o comprehendprehistoric meaninghave,and I believe necessarily, emphasized ts structure at the expense ofcontent, and domains of material culture that maximally retain the integ-rity they possessedn the systemiccontext. Rock art is a classic example(e.g., Leroi-Gourhan1982;Lewis-Williams 1981), s arearchitectureandthe intentional time capsules hat we call burials. Thus, Hodder (1984)found structural equivalences etweenNeolithic domestic and mortuaryarchitecture and related these o transformations n the status of womenand n the productive base.But we are ar from experiencing heir originalrichly textured meanings.The domain of mortuary practice, relating to the IO primarily throughideological and social activities, has great potential for ethnoarchaeolog-ical contributions to the IO-although curiously it is one n which there isan extraordinary ack of useful sources. n studyingmortuary practices nthe Mandara (David in press), I found that, despite the realization thatdisposal of the dead nvolves a conceptual ransformation of living soci-ety (Parker Pearson1982), his is not in itself adequate s an approach.We must also ake into considerationwhat Huntington and Metcalf (1979)have called the Hertzian triangle (see Hertz 1907).The triangle (Fig. 3)representsmortuary practices as comprising hree pairs of relationships.That between the Living and Mourners and the Corpse and its Dis-posal determines he scale of the rites and the differentiation of social


20/30

INTEGRATING ETHNOARCHAEOLOGY 349THE SOULTHE &DEAD

Procession of ritesTransformation of persona

THE LIVING &THE MOURNERS Differentiation ofsocial personaeScale of rites

THE CORPSE &ITS DISPOSAL

Mortuary sociologyFIG . 3. The Hertzian triangle.

per.ronae. t is this social aspectwith which archaeologistswere almostexclusively concerned hrough the 1970s.Until the 1980sittle attentionwas paid to the relationship between he Soul and the Dead and theCorpse and ts Disposal. This is apprehendedhroughmetaphors (e.g.,sleep and awakening o resurrection, reversebirth, or germination of anancestor) hat are expressed n the symbolism of the rites. The third pair,linking Living andMourners to the Soul and he Dead, accounts orthe sequence of rites as the social persona is transformed or extin-guished. 2The study of mortuary practices n the Mandarastrongly reaflirmed theexistenceof the symbolic reservoir, emphasizing hat material culture andpractices associatedwith disposalof the deadare intimately linked withideas, values, attitudes, and themes hat repeat n many different aspectsof the lives of these societies. Very striking is the polysemic nature oftraits and the ambiguity and multivalence of metaphors; thus, the tombpartakesof the natureof a hut, a pot, a womb, and an inverted granary.Also apparent s the greatvariability in the dataand that the sevengroupsstudiedchoose o emphasizedifferent aspectsof this common ideologicalheritage n differing ways and n different dimensionsof mortuary prac-tice.


21/30

350 NICHOLAS DAVIDIn this domain, ideology was shown to operate in all three of the ways

identified by Giddens (1979: 193-197), by:1. representation of the interests of dominant sections of society (maleelders) as universal,

2. denial or transmutation of contradictions (between castes), and3. naturalization of the present (structuring of mortuary practices by

natural social statuses: infant/fully human, man/woman, older/younger, etc.).Given the importance of disposal of the dead as a dramatic institutionalframework for the ongoing invention of society, it is inevitable that thiswil l always be the case. l3 Note that this is a candidate cross-cultural lawrelating IO to PO that has been arrived at via a hermeneutic approach, andthat the absence of scientist methodology is irrelevant. I have beenpresent at only one burial and have measured none.

Again we may ask how such very general transcultural regularities arevalidated. Principles have been presented above. A first crude test is touse oneself as informant and to ask whether the regularity applies both inones own and in other, preferably very different, cultures with which oneis familiar. However, we should go much further than this and designresearch that will analyze the differential output of mechanisms and pro-cesses in cultures that differ in historical and other contexts but that canbe shown to be structurally similar. This is as close as we may aspire tocontrolled experimentation. We have taken init ial steps to extend ourresearch to the Upper East Region of Ghana where the societies, econ-omies, and political histories of the TalIensi and their neighbors havemuch in common with those of Mandara peoples, but whose culturalheritage is markedly different, as is shown by their speaking languages ofa different phylum and family (Congo-Kordofanian Gur as against Afro-Asiatic Chadic). We expect, for example, that similar mechanisms oper-ating on different cultural materials will be expressed in output that isstructurally analogous but different in content. Such propositions are notamenable to simple forms of testing; rather we must again appeal toconsistency of inference and data, plausibility of analytical model andcoherence of theories. There wiIl be dangers of falhng into the trap ofpseudovalidation by cotirmation (Stockowzcki 1991). But if, despitesearching for counter examples and counter models, our theories aresupported by the new evidence, this would suggest that certain at least ofthe mechanisms and, by extension, processes in question express IO andPO patterns characteristic of the psychology, society, and culture ofHomo sapiens sapiens.

ETHNOARCHAEOLOGYANDARCHAEOLOGYSome regularities are specific to one or a limited range of societies,


22/30

INTEGRATING ETHNOARCHAEOLOGY 351others more widely or even generallyapplicable.Once recognized hesemay appear o be truisms, for example hat Buildings are liable to fulfiia seriesof functions over their useful ife, and the precessionof useswillin any one society show regularitiescharacteristicof one or more devo-lutionary cycles (modified from David 1971: 17-l 19).Nevertheless heyhelp direct and constrain archaeologicalobservationand interpretation.Middle-range heoriesexplicating the relationshipbetweendynamics andstatics are of immediate practical use in planning and interpreting re-search. n such cases he approach s generallyscientist and the behaviorinvestigated primarily technical and/or social. The scale can be large;Carol Kramer hasa major monographicstudy, Village Ethnoarchaeofogy(1982),on cultural transformations, relating primarily to technical andsocial aspectsof the PO, in rural Iran, and an analysisat the local andregional evels of scalareffects on the distribution of pottery in two citiesin Rajasthan(Kramer 1991).Whereas he former must influence the re-search design of any excavator of a village site in the Near East, andindeed elsewhere, he second,pioneering n its scope, should be takeninto account in constructing regionalprehistories.However, it is in the context of discovery that ethnoarchaeology, m-ploying a less obviously rigorous hermeneuticapproach,can contributemost to larger issuesof archaeologicalmethodologyand to researchde-sign. More important than the generationby ethnoarchaeologists f reg-ularities that can be applied in more or less formulaic fashion to archae-ological databases, s the deepened nderstanding f the expressions nmaterial culture of linkages between O and PO.Naturally ethnoarchaeologys elationship to archaeology, he otheranthropologicalsubdiscipline hat focuseson things, s very closeand willremain so. Its contribution to archaeology s invaluable at least in theoryand on condition that rigor is used, which too rarely occurs, often forreasons of incomparability of analytical and observational units-towhich we shouldpay more attention (Skibo et al. 1989).However, in thesense hat several of its practitioners are now engagedn more generalanthropological studies of material culture (e.g., Miller 1987),ethnoar-chaeology s becoming ess restricted to topics of specific and mmediatearchaeological oncern. am, for example,presently engagedn showinghow the technologyof Mafa iron smelting dependsupon a parallel tech-nology of plant medicines that draws on metaphorsof human digestionand pregnancy. Like the two sides of a ladder, the siderological andmagicobotanicalconstructsare linked together,reinforcing eachother ina powerful mnemonic that relates IO and PO, ideological and technicalactivities. Similar systems underpin the intergenerationalcontrol andtransmission of technology in many if not all nonliterate societies-another hypothesis o be explored.


23/30

352 NICHOLAS DAVIDEthnoarchaeologys primary service mission is sti ll the raising of the

analogical consciousness of archaeologists, many of whom prefer theirculture dead, sensitizing them to dimensions of variability and the rich-ness of the relationship between humans and their artifacts, includingtheir own bodies. They need, if only vicariously, to experience culturalrealities other than their own in order to combat the ethnocentrism thatcolors arguments and distorts inferences. But only some ethnoarchaeo-logically generated insights offer archaeologists the analogies, tactics, andstrategies needed for interpretation; others relate to broader anthropolog-ical concerns.

The realist philosophy of science is liberating in that it integrates his-tory and science, and encourages archaeologists and ethnoarchaeologiststo get on with what they are best at-doing archaeology and ethnoarchae-ology. Realism does not, for example, require us to engage in mentalgymnastics in order to reformulate our retroductive insights as hypothet-iconomological deductions. A realist approach offers much more thanwhat Binford (1987:391) called a kind of intellectual comfort in a mis-guided paper that would of itself justify the present attempt to reintegratethe field. His attack on archaeological hermeneutics is there founded onmisrepresentation of ethnography-we learn, for example, that ethnogra-phers, unlike archaeologists, do not report data but only information-but I read the conclusion to his paper as falling fair ly and squarely withinthe realist program advocated above. Realist praxis amid systems that aremore or less complex and restricted in fact characterizes most recentresearch, allowing research variety while providing firm although flexibleguidelines to research process and the criteria to be used in evaluatinginferences and theories. Knowledge can progress in a context of transitivefacts, the theory and politics laden science of the day, incorporating theinsights, say, of feminism and postmodernism while retaining crit ical con-trol. Meanwhile the intransitive objects of our knowledge, the real things,may never be known but can be ever more closely approximated.

ACKNOWLEDGMENTSTea mw ork has been vital in the accumu lation of the data and the analyses on which the

later part of this paper is based. M AP team mem bers in 1986, 1989, and 1990 included,besides myse l f , Kodzo Gavua, David Ki l lick, Diane Lyons, Scott MacEach em, MaureenRe eves, Ian Rob ertson, and Judith Sterner. Field research w as carried out under authori-zation from the Ministry of Higher Education, Com puting and Scienti f ic Resea rch ofCam eroon, and within the conte xt of a Protocole dAccord de Cooperation between theUnivers i ty of Calgary and ORS TOM , France. I t was supported by Social Science andHum anities Resea rch Counci l Grants (410-85-1040, 410-88-0361, 410-89-0871) and anSSH RC Leave Fellowship (451-85-1231). Eldridge Moha mm adou, former Chief of theNorthern Station of the Insti tute of Hum an Studies, Cam eroon, and Philippe Mathieu,


24/30

INTEGRATING ETHNOARCHAEOLOGY 353Michele Deh teuf and other memb ers of OR STO M provided logistical support and offered usman y col legial kindness es. Usher Fleising, Jan e Kelley, Carol Kramer, an anonym ous re-viewer, and other col leagues and studen ts have given me valuable advice, and Bob Hro-madiuk a crucial reference. I give them al l m y thank s.

NOTES I fol low Kosso in treating Lewis Binford and Ian Hodder as archetype s. Binfords

phi losophy of science and more obviously his rhetoric have developed over the years fromlogical em piricism/posit ivism toward the real ism which his practice has always tended toexhibit (see Wylie 1989 and below). Hodders view s, approaches, and interests havechanged even more rapidly, and I exclude from this paper any consideration of his andothers po stmode rnist concern with reflexivity and cri t ical theory (on which see Ham mer-sley 199296422).

2 Herm eneutics is a word that turns man y people off. I use i t here in the broad sense ofthe interpretation of layers of meaning of texts, broadly defined to include such things asmaterial culture sets , in the conte xt of other such texts .3 The term ideological is only used in preference to the more correct ideational inorder to avoid confusion with Ideational Order (see below).4 Given the existence o f contradiction within the ideational order, symb olic interactionistanalysis is in som e case s a more appropriate term than herme neutics. Hodders (1991) study

of I lchamu s decorated calabashes is a good case in point. Ha tt (1992) provides both ahistorical example of sym bolic interactionist analysis and a clear explanation of the ditfer-ence between the kindred methods.5 In the article cited we argued an assimilation of pots to people. We can now dem onstratethat the caste division of central Mandara societies is expressed in the manufacture of redand black ceramic wares (David 1990).

6 The bucksh ot approach refers to the hit-or-miss use of a specif ic ethnographic analogyto sugge st an answer to som e equally specif ic archaeological question. A researcher takinga laboratory approach exam ines the relationship between known ethnographic behavior andobservable archaeological remains. The spoiler approach, related to the cautionary taleof Heider (1967), uses such correlations to evaluate statem ents, mod els, and assum ptionsof a general ly deductive nature appl icable to a certain range of societies, and often leadsto their being discarded or reformulated. General m odels include general analogies anddeductive hypothes es as well as . . . lawlike general izations, which Y ellen (1977:6) arguesshould be stated in the form of hypotheses susceptible to testing. This s tatem ent d oes not of course ap ply to other research, not involving experimenta-tion, carried out by mem bers of the Kalinga project, for example M iriam Starks (1991)thoughtful study of ceram ic production and com mu nity special ization. With regard to theory construction, I agree with Gallay that regularit ies (and thus mech -anisms) canno t be val idated by reference to the archaeological record, but only b y theextension of ethnoarchaeological observations to other ethnographic conte xts, and ult i-ma tely by reformulation of the regularity in the l ight of new evidence (i .e., the standardpractice of model bui lding and testing).9 There is no space here to explore the styles of ethnoarchaeological f ieldwork necessari lyassociated with the scientist and hermene utic approaches. The former is l ikely to involvecensu ses, questionnaires, mapping, and measu remen t of material culture and the asking ofspecif ic questions. The second , while n ot spurning the techniques of the first, requireslonger and deeper exposure to the culture and language, and more argument over themeaning of conce pts underlying behavior and informants answ ers.


25/30

354 NICHOLAS DAVIDlo Thu s, in the central Mandara, there is a transformer caste responsible among other

things for special ist i ron working and ceram ics (Sterner and David 1991). Cas te mem bers aremuc h less constrained in their ma rital, social, and econom ic relations than the ma ss ofordinary folk, and behave in som e wa ys as a supraethnic-l inguistic group, and they aredisproportionately responsible for creation of the a&actual environmen t. A l i tt le furthernorth different production sys tem s give rise to different effec ts (MacE achem 1990).

i i In fact a wea kness of Hodders (1982: 125-W) fascinating Dirt, wom en and men : Astudy in the Nuba Mou ntains, Sudan is that his argument that bel iefs in sex pol lution areimportant in structuring the patterning of refuse and decoration is overly dependent upon across-cultural regularity proposed by Mary Douglas (1966: 142) and not enough on observa-tion and discussion with informan ts. The symb olic principle thus appears somew hat of an expost facto imposition on the data, despite his exem plary attem pt to show , by comp arativeanalysis, contrasting expressions of the principle in two fields of human behavior in twoethnic groups. Hodder and his team were in the field for only a short t ime. Shan ks and Ti l ley (1982) attemp ted to dem onstrate the existence of ideological natu-ral ii tion (the representation of social cons tructs as part of the natural order o f things) anddenial of contradiction (the concealm ent or misrepresentation of social relations) expressedin regularit ies in the spatial arrangements of different classes of bone within [Neoli thic]tom bs. But their argume nts mu st be faulted for ignoring the left-hand side of the triangle,the precession of ri tes during which de camization and probably ancestral ization-becoming an ancestor-took place. Instead they assum ed that only the final jumbled dis-posit ion of the dismembe red and partial ske letons wa s the explanandum.

I3 A major problem facing the archaeologist is then to determine which elements ofmortuary practices are operating in which wa y. To this there can be no general answer,particularly as, since denial of contradiction seem s l ikely in mo st instances to imply a lackof material differentiation, i ts identi f ication wil l require especial ly ful l contextual knowledgeof other aspects of the culture.

REFERENCES CITEDBinder, D ., and C . Perles

1990 Strategies de gestion des outi llages l ithiques au Ntol i thique. Pal&o 2257-283.Binford, L. R.

1977 General introduction. In For theory bui lding in archaeology, edited by L. R.Binford, pp. l-10. Academ ic Press, Ne w York.

Binford, L. R.1982 Objectivity-explanation-archaeology 1981. In Theory and explanation in ar-

chaeology, edi ted by C. Renfrew, M. J. Rowlands, and B. A. Seagraves, pp.125-138. Academic Press, New York.Binford, L. R .

1987 Data , relativism and archaeological science. Man 22~391-404.Boutrais, J. (editor)1984 Le Nord du Cam eroun: des hom me s, une r&ion. Editions de IORSTOM (Col-

lection Mem oires, No . 102), Paris.David, N.1971 The Fulani compound and the archaeologist. World Archaeology 3:111-131.

David, N.1990 Vessels of the spiri ts: P ots and people in North Cam eroon. Video (50 mm ).

Dept. of Comm unicat ions Media, Univers ity of Calgary, Calgary.


26/30

INTEGRATING ETHNOARCHAEOLOGY 355David, N.In press The archaeology of ideology: Mortuary practices in the central Ma&am

highlands, North Came roon. In An qfrican cam mittm ent: Papers in honour ofPeter Lewis Shinnie, edited by J. Sterner and N. David. Univers ity of CalgaryPress, Calgary.David, N., K. Gavua, A . S. MacEach em, and J. Sterner1991 Ethnicity and material culture in North Camero on. Canadian Journal of Ar-chaeotogy 15:171-177.David, N., and H. Hennig

1972 The ethnology of pottery: A Fulani case seen in archaeological perspec tive. Add-ison-Wesley Vol. 21, Modular Pub lications. Ad dison-Wesley, Reading, MA .David, N. , and J. Sterner1987 The Mandara Archaeological Project 1984-87. Nyo me Akumo 292-8.David, N. , and J. Sterner1989 The Mandara A rchaeological Project 1988-89. Nya me Akum a 325-9.

David, N., J . Sterner, and K. Gavua1988 Wh y pots are decorated. Current Anthropology 29:365-389.David, N., and D. Voas1981 The societal caus es of inferti li ty and population decl ine am ong the settled Fulaniof North Cameroon. Man MSM-564.David, W.1990 Style and history in art history. In The uses of style in archaeology, edited by M .

Conkey and C. Hastorf , pp. 18-31. Cambridge Univers i ty Press, Cambridge/London.Douglas, M.1964 Purity and danger. Ro utledge and Kegan Paul, London.

Eriksen, T . H.1991 The cultural con texts of ethnic differences. Man 26~127-144.Gallay, A.1989 Logicism: A French view of archaeological theory founded in compu tationalperspective. Antiquity 63:27-39.Gallay, A.1991 A propos de la c&antique actuel le du delta interieur du Niger (Mali). Approcheethnoarcheologique et regles transculturel les. Pr&irages de com mun ications,

Ethnoarcheologie: Just@ ation, problemes, l imites, pp. 45-53. 12e Renco ntreIntemationale dArchbologie et dHistoire dAntibes, O ctober 1991.Gardin, J.-C.1980 Archaeological cons tructs. An aspec t of theoretical archaeology. Camb ridgeUnive rsity Press , Cam bridge/London.Gavua, K.1990 Style in Mafa m aterial culture. Unpu blished Ph.D . dissertation, Unive rsity of

c&WY.Gibbon, G.1989 Explanation in archaeology. Blackwell , Oxford .Giddens , A.

1979 Central problems in social theory. Al len and Unw in, London.Goodenough, W. H.1964 Introduction. In Explorations in cultural anthropology, edited by W . H. Good-enough, pp. l-24. McG raw-H il l, Ne w York.Gould, R. A.1980 Living archaeology. Cam bridge Unive rsity Press , Cam bridge/London.


27/30

356 NICHOLAS DAVID

Hammers ley , M.1992 Wh ats wrong with ethnogruphy: Methodological explorations. R outledge, Lon-don.Hat t , D.1992 The sym bolic interactionist perspective in historical analysis. Paper presented atthe Quali tative Analysis Confere nce, Ottaw a, M ay 1992.Heider, K. Cl.1967 Archaeological assum ptions and ethnographical fact s: A cautionary tale fromNe w Guinea. Southwestern Journal of Archaeology 23:52&I.Her tz , R.1907 Contribution a une etude sur la representation col lective de la mo rt. An&e

sociologique 10~48-137.Herskov i ts , M. J .1960 Econom ic anthropology: A study in comp arative econ omics. Alfred A. Knopf,New York .Hodder, I .

1979 Econom ic and social stres s and material culture patterning. American Antiquity44446-454.Hodder, I .

1982 Sym bols in action. Cam bridge U niversity Press, Cam bridge/London.Hodder, I .1984 Burials, house s, wom en and men in the European Neo li thic. In Ideology, power

and prehistory, edited by D . Mil ler and C. Ti l ley, pp. 51-68. Cambridge Uni-versity Press, Cambridge/London.Hodder, I .1986 Reading the past. Camb ridge U niversity Press, Cam bridge/London.

Hodder, I .1990 The dom estication of Europe: Structure and contingency in neoli thic societies.Blackwell , Oxford.

Hodder, I .1991 The decoration of containers: an ethnographic and historical study . In Ceram icethnoarchaeology, edited by W . A. Longacre, pp. 71-94, Univers ity of ArizonaPress, Tucson, AZ.

Hunt ington, R., and P. Metcal f1979 Celebrations of death: The anthropology of mortuary ri tual. Cambridge Uni-versity Press, Cambridge/London.

Juillerat , B .1971 Les bases de l organisation sociale chez les Mo uktC lC (no&Cam eroun). Mem -oire de 1Insti tut dEthnologie 8. Insti tut dEthnologie-Musee de IHom me ,Paris.

Kel ley, J . H., and M. P . Hanen1988 Archaeology and the methodology of science. Univers i ty of New Mexico Press,Albuquerque, NM.Keynes, J. M.1939 The general theory of emp loymen t, interest and mon ey. Macm il lan, London.Kleindienst, M. R ., and P. J. Watson1956 Action archaeology: The archaeological inventory of a l iving com mu nity. An-thropology Tomorrow 975-78.Kosso, P.1991 Method in archaeology: Middle range theory as herm eneutics. Ame rican Antiq-uity S&621627.


28/30

INTEGRATING ETHNOARCHAEOLOGY 357Kramer, C.

1982 Village ethnoarchaeology: Rural Iran in archaeological perspec tive. Academ icPress, New York.Kramer, C.1985 Ceram ic ethnoarchaeology. Annual review ofanthropology, edited by B. Sigel-man , pp. 77-102. Annual Rev iews, Inc., Palo Alto, CA .Kramer, C.1991 Cera mics in two Indian cit ies. In Ceram ic ethnoarchaeology, edited by W . A.Longacre, pp. 205-30. Unive rsity of Arizona Press , Tucs on, AZ .Lechtman, H.1977 Style in technology-Some early though ts. In Material culture: Styles, organi-zation and dynam ics of technology, edited by H. Lechtman and R. S. Merri l l,

pp. 3-20. (1975 Proceedings of the Ame rican Ethnological Society.) W est Pub-l ishing Co ., St. Paul/New York.Leroi-Gourhan, A.

1982 The dawn of European art. Camb ridge Unive rsity Press, Camb ridge/London.Lewis-Wi l liams, J. D.

1981 Believing and seeing: Symbolic m eanings in southern San rock paintings. Ac-ademic Press, London.Longacre, W . A. (editor)

1991a Cera mic ethnoarchaeology. Univers ity of Arizona Press, Tuc son, AZ .Longacre, W. A.1991b The perfect marriage: The essential joining of ethnoarchaeology and experimen-

tal archaeology. Pretirages de com mun ications. Ethnoarcheologie: justi f ication,problemes, l imites, pp. 77-79. 12e Renco ntre Intemationale dArchCologie etdHistoire dAntibes, October 1991.MacEache m, A. S.

1990 Du Kunde: Proces ses of montagnard ethnogenesis in the northern Mandaramou ntains of Cam eroon. U npublished Ph.D . dissertation, Unive rsity of Cal-&w-Y.McIntosh, R. J .1989 Middle Niger terracottas before the Symplegad es gateway. African A rts 22:74-83.

Mil ler, D.1987 Material culture and ma ss consum ption. Blackwell , Oxford.Parker Pearson, M.1982 Mortuary practices, soc iety and ideology: An ethnoarchaeological study . InSym bolic and structural archaeology, edited by I. Hodder, pp. 99-113. Cam -bridge University Press , Cam bridge/London.

Petrequin, A.-M ., and P. Petrequin1984 Hab itat lacustre du Benin: Une approche ethnoarchtologique. Mem oire 39.Editions Reche rche sur les Civi l isations, Paris.Robertson, I . G.In press Hoes and metal templates in North Cameroon. In An Afr ican comm ittment:Papers in honour of Peter Lewis Shinnie, edited by J. Sterner and N. David.

Unive rsity of Calgary Press, Calgary.Rou x, V. (in col laboration with D. Corbetta)1990 Le tour du potier. Special isation artisanale et com petence s techn iques. Mono-graphie du CR A 41. Editions du CN RS , Paris.Sacket t, J . R.1998 Style and ethnicity in archaeology: the case for isochre stism. In The uses of style


29/30

358 NICHOLAS DAVIDin archaeology, edited by M . Co nkey, and C. Ha storf, pp. 3243. Camb ridgeUnive rsity Press, Cam bridge/London.

Sanjek, R.1991 The ethnographic present. Man 26609-628.

Schi f fer, M.1976 Behavioral archaeology. Acade mic Press, Ne w York.

Schi t fer, M.1978 Methodological issues in ethnoarchaeology. In Explorations in efhnoarchaeol-ogy, edited by R . Gould, pp. 229-247. U niversity of Ne w M exico Press, Albu-

querque , NM.Shanks, M. , and C. Ti iey

1982 Ideology, symb olic pow er and ri tual commu nication: a reinterpretation of Neo-l i thic m ortuary practices. In Symb olic and structural archaeology, edited by I.Hodde r, pp. 129-154. Cam bridge Unive rsity Press, Cam bridge/London.

Skibo, J. M., M . B . Schif fer, and N. Kowalski1989 Ceram ic style analysis in archaeology and ethnoarchaeology: Bridging the an-alytical gap. Journal of Anthropological Archaeology 8~388409.

Stark, M.1991 Ceram ic production and com mu nity special ization: A Kalinga ethnoarchaeolog-ical stud y. World Archaeology 2354-78.

Sterner, J.1989 Who is signall ing whom ? C eram ic style, ethnicity and taphonom y among the

Sirak Bulahay. Antiquity 63:451-459.Sterner, J.In press Sacred pots and symbolic reservoirs in the Mandara Highlands of NorthCam eroon. In An African com mittm ent: Papers in honour of Peter Lewis Shin-nie, edited b y J. Sterner and N. David. Unive rsity of Calgary Press, Calgary.

Sterner, J . , and N. David1991 Gender and caste in the Mandara highlands: Northeastern Nigeria and northernCameroon. Ethnology 30:355-369.

Stoczkowsk i , W.1991 Prehistoire, ethuologie et approche predictive: tentation dune epistemologienaive. Pr&irages de com mun ications, Ethnoarcheologie: Justi f ication, prob-ldmes , l imites, pp. 117-119. 12e Renco ntre Intemationale dArchtologie et

dHistoire dAntibes, October 1991.Trigger, B.

1989 A history of archaeological thought. Cambridge U niversity Press , Cam bridge/London.Tringham, R.

1978 Experimentation, ethnoarchaeology, and the leapfrogs in archaeological meth-odology. In Explorations in ethnoarchaeology, edited by R. Gould, pp. 169-199.Univers i ty of New Mexico Press, Albuquerque, NM .Wiessner, P.

1983 Style and social information in Kalahari San projectile points. Ame rican Antiq-uity 48~253-276.Wobs t , H . M .

1977 Styl istic behavior and information exchange . In For the director: Re search es-say s in honour of J. B. Griff in, edited by C. Cleland, Mus eum of AnthropologyMem oir, Univers ity of Michigan, 13, pp. 317-342.


30/30

Wobst , H.1978Wylie, A.1982

Wylie, A.1989

Yellen, J.1977

INTEGRATING ETHNOARCHAEOLOGY 359M .The archaeo-ethnology of hunter-gatherers or the tyranny of the ethnographicrecord in archaeology. American Antiquity 433303-09.Epistemological issues raised by a structural ist archae.ology. In Sym bolic andstructural archaeology, edited by I. Hodder, pp. 3%+6. Cam bridge Univer sityPress, Cambridge/London.The interpretive di lemma . In Crit ical tradit ions in contemp orary archaeology,edited by V. Pinsky , and A. Wylie, pp. 18-27. Camb ridge Unive rsity Press ,Cambridge/London.Archaeological approaches to the present. Academ ic Pre ss, Ne w York.

Documents

Integrating Ethnoarchaeology