Alyawara Site Structure and Its Archaeological Implications

Alyawara Site Structure and Its Archaeological ImplicationsJames F. O'ConnellAmerican Antiquity, Vol. 52, No. 1. (Jan., 1987), pp. 74-108.Stable URL:http://links.jstor.org/sici?sici=0002-7316%28198701%2952%3A1%3C74%3AASSAIA%3E2.0.CO%3B2-3American Antiquity is currently published by Society for American Archaeology.Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtainedprior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content inthe JSTOR archive only for your personal, non-commercial use.Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained athttp://www.jstor.org/journals/sam.html.Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academicjournals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community takeadvantage of advances in technology. For more information regarding JSTOR, please contact [email protected]://www.jstor.orgThu Sep 6 15:35:21 2007ALYAWARASITE STRUCTURE AND ITSARCHAEOLOGICAL IMPLICATIONSJamesF.O'Connell Ethnoarchaeologicalresearchhasshownthattheassumptions formerlygu~dingtheinvestigationo fhunter- gatherersitestructureareinvalid,atleastasgeneralrules.FurtherresearchisnoB1 requiredtoidentlfvthe determinants ofsitestructureandassesstheireffect.HereIdescribethoseaspectsof Alyawarabehaviorthat shape site structureatresidential base camps. Ithen compare certain featuresof t he Afyan'ara case withthose of the !Kung andthe Nunamiut.Itentatively inferthatmosto fthevariabilityin sitestructurein these casesisa ,funct~ono fdlfferences in the degree of reliance on foodstorage, seasonalvariationinweather,householdpopu- lation size, and the length oftirne activity areas are in use. Predator pressureand the relative importance of inter- household foodsharing may also be involved.The apparentimportance of-food storage suggeststhatvariationin site structure may parallelBinford'sil980~forager-collectorcontinuum. Furtherresearch is requlred to determine ifthisis so, andwhetherotheraspects ofbehavioralsoaffect sitestructure.These observationshaveimportant ir~~plications forthe conductof archaeological research, especially forthe spatialscalesatwhichanalysesof site structureareundertaken, and,for the sizes c?fr100,000 m2,orabout 1-10ha (Table1). Settlement lccation is determined by access to water. local terrain features. and constraints imposed byEuropean homesteaders. Mostsettlements are situated on exposed,well-drainedsites,200-800 mfrom awater source. Thi s distance is mandatedbythe need to ensure accessto water for cattle. Settlement populationsare divided into households, consisting of groups of kinspeopleor close friends whoregularlyeat, sleep.and interacttogether.Three typesof Householdcan be defined on the basisof the age and sex of their members: ALYAWARA SITESTRUCTURE Post H e a r t hMldden m~xedw~t hashandcharcoal ,"rassw~t hscattered trash RPRoast~ngD I ~,&Charcoal,ashand d~rtp~l e -Sheetmetal w~ndscreen --- Unwalled brush shade Figure 1.Plan of Alyawara household activity area, adjacent roasting pits, and refuse disposal zone. Nuclear familyhouseholds.These include aman,hiswifeor wives(up to three), andtheir pre-adolescent children. They may have as many as12 members, but average about five. They are generally stable in composition,although some members may be absent for short periods for various reasons, including work and inter-settlementvisiting. Women's households.These usuallyconsist of a core group of widows,theirpre-adolescent childrenofbothsexes, andoneormoreunmarriedwomen.Theyaresometimes joinedfor periods of up to a few weeks bymarriedwomen and theirchildren who have temporarilyleft theirresidence elsewhereinthesettlement,eitherbecauseof domesticstrife orbecausethe husband is away from home. Women visitors from other settlements unaccompanied by their husbands also customarily joina women'shousehold for the duration of their stay. Household populations range from 2-28,but average about six. Some are stable in composition, but others vary greatly in membership over short periods of time. Kin relationships among core members and visitors are normally very close and predominantlymatrilateral. 78AMERICANANTIQUITY[Vol. 52, No.1,1987 Figure2.Main shelterat Bendaijerum,September1974. Nuclearfamilyhousehold: twoadults,twochildren. AWen'shouseholds.These are similar to women'shouseholds,except thatage is an important organizing principle. In some men's households, nearly all members are aged 35 or older. Often they are good friends, but not close relatives. Residents of younger men's households are typically close patrilateralkin. The number of households per settlement ranges from about 5-35,and averages about 20. Nuclear familyhouseholdsusuallymake up 65-85%ofthe total, women'shouseholds15-2S0/o,and men's households3-20%(Table1). HouseholdAct i v~t yAreas All activities insettlements take place inone of two kinds of areas, householdactivity areas and special activity areas. Household activity areas are the settings for most activities in the settlement. Each is occupied by one of the householdtypes defined above. Each is marked by a circle of cleared ground1-20mindiameter(3-350m2, = 78.9i 84.5 m2; see Figure1). (Quantitative data are drawn fromasampleof106 householdactivity areas described atBendaijerum.) Within this area, household members prepare and consume food, make and repair tools and equipment, and entertain visitors.Eachhouseholdactivityareacontains one or more shelters and hearths, and is flankedby arefuse disposal zone. AtfainShelters.Theseareplaceswherehouseholdmemberssleep,seekprivacyorprotection from the elements, prepare and eat food during inclement weather, and store goods and equipment. Mainsheltersareusuallylocatedonthewindwardsideoftheactivityarea.Theyvaryinform depending on the season of the year, prevailing weather conditions, anticipated length of occupation, and availability of building materials. The most common types are brush or sheet metal windscreens (often rigged withtarpaulin roofs for shade or protectionfrom rain), sheet metal lean-tos, and fully enclosed, box-likestructureswithstoutwoodenormetalframes and brush,grass,orsheetmetal walls(Figures2-4;see alsoO'Connell1979:Plates1-4).Openunroofedstructuresare favored in the wintermonths;roofedshelters are essentialforshade insummer. Householdersanticipatinga stayofmorethanafewdaysoftenbuildaroofed,enclosedstructureregardlessoftheseason. Structure formis unrelatedto the age and sex composition of the household.Despite differences in 79O'Connell]ALYAWARA SITE STRUCTURE Figure 3.Men's household activity area at Bendaijerum, July 1974. Three men at front shape wooden shields; fourmenatbacklounge onbedding inleeoflowbush windscreen. Sheetmetal main shelterabuts windscreen atright.Carmarkslocationofautorepairstationbehindwindscreen. Edgeofrefusedisposalareavisiblein right foreground markedbyash dump and large cans. overallformor constructionmaterial,nearlyallmainsheltershavesquareor rectangularground plans. Totalfloorareas range from2-40m2 (X = 10.0* 6.9 m2). OtherStructures.Twootherkindsofstructures,sunshadesanddogshelters,arealsooften foundinhouseholdactivityareas.Sunshadesare usuallylocatedbetweenthecenterofthearea and thesideoppositethemainshelter (Figure 4;seealsoO'Connell1979:Plate 4). Theyvaryin form, butmostaresquareor rectangular, flatroofed,brushand polestructures,1-2mwideand 1-8m long (covered area X = 5.1i3.7 m2). Shades are unwalled in winter, but typically have leafy boughscreensontheir northand westsidesinsummer to provideprotectionfromthe afternoon sun. About30% of Alyawarahousehold areas have shades in winter, 75% in summer. Again, there is no correlation between the presenceor absence of a shade and the composition of the household. All women'shouseholds and most nuclearfamilies keep dogs for the personalsecurity of women and for useinhunting.Smaller dogsare allowedtorunfreeinthe settlement, butthe largerones are kept tied up to prevent them from attacking small children or fighting among themselves. Small, domedbrushsheltersarebuiltforthesedogsattheedgeoftheactivityareaoppositethemain shelter, or in the refuse disposal zone (see below). Roughly 10% of all recorded Alyawara households, nearlya!lof them women'shouseholds, have one or more of thesespecial shelters. Wherever dogs are tied, they dig shallow pits or dust wallows up to1 mwide and 50 cm deep, apparentlybecause it helpskeepthemcool.Dogtethersare oftensurrounded bypiles of excrement. Hearths.All household activity areas have hearths. Their location within the activity area varies seasonally.Inwinter,theyaremostlikelytobefoundinthemainshelter; insummer,theyare equally distributed betweenthe main shelter and outside areas. Those at entrances to enclosed main sheltersor inotherpartsoftheactivityareaare unlined,circularinplan,and rangefrom30-50 AMERICAN ANTIQUITY[Vol. 52, No. 1, 1987 Figure 4.Women'shousehold activity area at Gurlanda C, September 1973. Brush backed, tarp roofed main shelter atleft,sunshade at right. cm in diameter.Those located inside enclosed shelters or associatedwith windscreensare typically oblonginplan,and measure30-50cm wideand 90-150cm long.Where thereismore thanone hearth,hearthsare generallyarranged in line parallelto the longaxisof the shelter, withthe long axis of each hearthperpendicular to thatof the shelter (Figure1). Sleepers lie betweenthe hearths, parallel to the axis of the hearth. Hearths are typicallysetonflatsurfaces thathave beenpreparedonlyinthat grass,stones, and otherloosedebrishavebeenscrapedaside.Thegroundbeneathhearthsmaybescoopedout occasionally to formsmall roastingpits, about30-40cm in diameter and 20-30cm deep, usually to cookIpotnoeatubers(O'Connelletal.1983:84). Becausehearthsare kept smouldering for long periodsandseldompurposelysmothered, their contentsconsistprimarilyoffinewhiteashwith relativelylittlecharcoal. Thoseinusefor more thana weekmay"wander"somewhat,producing large patches of fire-reddened, ash-stained soil. The number of hearths in use at any one time varies directly, thoughloosely, with the number of adults in the household (r = .61). The linear regression that describes this relationship bestis: whereyisthenumberofhearthsinuse,andxthenumberofadultresidents.Thismeansthat householdswithtwoadults are likelyto haveonlyone hearthin use,whileinlargerhouseholds, the number of hearthsincreases withthe number of adult residents at arate of about1: 1. Size ofActivity Areas.Activityareas varyinsizewithhouseholdpopulationand the lengthof timethe area hasbeenoccupied(multiple r= .73). Amultipleregressionanalysisshows thatthis relationship is bestexpressed: wherey is thesize of the activity area, x, the number of days the camp has beenoccupied, and x, the totalhouseholdpopulation,includingadults and children.This means thatwherethespanof 81O'Connell]ALYAWARA SITESTRUCTURE Figure 5.Abandoned household activity area at Bendaijerum, September 1974. Large cleared space is activity area proper. Mainshelterlocationinbackground,sunshade support posts atright center. Refuse disposalzone in foreground. occupation is heldconstant, thesize of the activityareaincreases withhouseholdpopulation, but atadeclining rate.Forexample, activityareas occupiedbytwomember households for35 days are likelytomeasure about28m2, orabout14 m2 peroccupant. Thoseoccupiedby12 member householdsforthesamelengthoftimeshouldmeasure110 m2, orabout9m2 perperson.The declining rate should not be surprising because larger households are more likely to include children. Similarly,wherehouseholdpopulationisaconstant,theactivityareashouldgrowataslow, steadilydecreasingratethroughtime.Activityareasoccupiedbygroupsofsixforfivedaysare likely to cover18-20m2, and to increase in size by less than 2.3 m2each day they are subsequently occupied. Areas occupied bysix people for 60 days should cover about 92 m2, and should grow by less than1m2 each day after that. The longer a camp is occupied, the broader the range of activities likely to takeplacethere, the greater the number of people likely to be involved, and the larger the spacerequired to accommodatethem.Householdmembers reactto this bygradually clearing and expanding the activityareaas necessary. Populationand span of occupation togetheraccount for only about53% of the total variance in size of household activity areas. Some of the residual variation is almost certainly a product of error in measuring the variables. especially household population, which fluctuated substantially in many ofthehouseholds inthesample. Correlationandregressionfigures reflecthouseholdpopulations at the time activity areas weremapped.Preciseestimates of thesize of the activityarea werealso difficult to obtain. In addition, other factors, such as the characteristics of individual locations (e.g., slope, distributionof naturallyoccurringshade) mayalso beinvolved. Refuse Disposal Zone.Household activity areas are flanked by refuse disposal zones or middens where much of the trash produced by the household is deposited (Figure1).This includes tin cans, bones, old clothing. ashes from hearths, and other material swept. raked, or tossed from the activity area.Athouseholdareas occupied bysmall groupsforveryshort periodsoftime,thesefeatures may consist of no more than a few items scattered justbeyondthe edge of the activityarea. Where 82AMERICANANTIQUITY[Vol. 52, No.1,1987 occupationhasbeenmoreintensive (largerhouseholds.longerperiodsofresidence),themidden mayberepresentedbyabroadbandofgarbagethatcompletelyencirclestheactivityareaand extendsawayfromitfordistancesofupto30m.especiallyonthedownwindside.Withinthe midden,debris densityishighestat theedgeoftheactivityarea,and diminishessharplyasone moves awayfrom it(Figures1and 5). Notallrefuseproducedintheactivityareaisremovedtotherefusedisposalzone.Infact.a strong patternof size sorting is apparent. Large items (maximum length2 5 cm) are generally tossed onto the nearestpartof the middenat time of discard.Small items are more oftendropped in the activityarea at or near their locationof last use. Some of thesesmall items are subsequentlyswept to theedge ofthemidden.usuallywithinafewdays.Inspiteofthesweeping. manysmallrefuse itemsremainintheactivityarea.thoughnotnecessarilyatthespotwheretheywereoriginally dropped. Variationinthesizeofmiddenareais notwellcontrolled, butshouldberelatedtothesizeof thehouseholdpopulation,thelengthoftimetheactivityareaisinuse,thesizeofrefuseitems produced bythehousehold, and the rateat whichtheyare produced.The size of theactivityand middenareas combinedshouldalsovary withtheseparameters. Distributionof ActivitiesWithin Household Areas The distributionofactivitieswithinahouseholdactivityareais determinedbyseveral factors, includingthenatureoftheactivity,theprevailingweatherconditions,thecompositionofthe household, and the size and composition of the group physicallypresentat any giventime. *VatureofActivity.Broadly speaking, there are six kinds of activities that take place in Alyawara householdareas:cooking.eating.talking,cardplaying.sleeping,and themanufactureandmain- tenanceof tools and equipment. Sleeping is almost always confined to shadedor sheltered spots in thecamp.i.e..to structures or totheleeor shade ofatreeorbush.Cooking activities, including allstagesinthepreparationofmeals.are carriedoutintheimmediatevicinityofhearths, either in the main shelter or elsewhere in the activityarea. Because all foods.with the exception of larger game animals (see below), are cooked directly in the ashes or in small pots or pans, there is nothing inherent inthe cooking operations themselves that conditions the use of different types of hearths, or hearths locatedin differentsections of the householdarea. Eating,talking.card playing.and the manufacture andmaintenance oftoolsand equipment cananddo occuranywhereintheactivity area. The only exceptions are those manufacturing tasks that involve the use of fire. e.g., straightening wooden tools, making or using spinifexresinadhesive(Binford1984). or shaping the bits on metal digging sticks.These operations mustbeperformedat or neara hearth. T.t7eatherConditions.These exert a strong influence on the distribution ofactivities in a household area.Duringwinter,daytimehighsaverageabout20-22C(ca.70F) butare oftenaslowas10- 12C (50-52F)forseveralconsecutivedays.Prevailingwindsaretypicallysteadyat10-20kph. Undertheseconditions,peopleusuallyprefertoconfinetheiractivitiestothemainshelter.In summer, daytime highsaverage about36-38C(96-100F)for6-8hours on anygivenday, which means that shaded areas become the favored locations for the performance of most activities. People generallypreferto beundershades ratherthan inmain shelters during this season becausethe air circulationis better. Itis important to emphasize thatsignificantchanges inweatherconditions over shorter periods oftime.ofteninthecourseofasingleday(e.g.. strongwinds,rainstorms). mayhaveanequally strongeffecton thelocationofactivities withinhouseholdcamps.Nearlyalldomestic activities, even those requiring access to a hearth. may be repositionedin response to very short-term changes inwinddirectionandvelocity.and air temperature. Cornposition of t heHousehold.Where household populationis fairly constant, as in most nuclear families, the distribution of hearths and structures, and the activities associated with them are likely to be correspondingly stable. However, where householdpopulation is subject to frequentchanges in size and composition (as inmost single sex households.especially the larger ones), the structural organizationoftheactivityareaand, byextension, thedistributionofactivities withinthatarea. 83O'Connell]ALYAWARA SITE STRUCTURE mayvarygreatlywithinashortperiodoftime.If,forexample, agroupofvisitors joinsasingle sex household for aperiodof several days, the mainshelter will beincreasedinsize. The number of hearths in the shelter will also be increased. When these or other household members subsequently leave the camp.partof the shelter will beabandoned, and the number and distributionof hearths inusewillchangeaccordingly.Furtherfluctuationsinhouseholdsizewillleadtofurthermodifi- cations ofcamp structure. Changes inhouseholdsize and compositionmayalsocausechanges in the structure of other parts of the activity area, especially the size, number.and distribution of sun shades, and the number and distribution of hearths in use. All of these changes affect the distribution ofactivities withinthe household activity area. Size, Cornposition,and BehaviorofGroup PhysicallyPresent.The distributionofactivities in acampisalsoaffectedbythesizeofthegroupphysicallypresentatanygiventime,andthe compatibility of the activities inwhich its members are engaged. Asingle manmakinga throwing stickonacooldaymayprefertositonor nearhisownbedrollinthemainshelterarea; butif several people are working on various manufacturing or maintenance tasks near a hearth in another partoftheactivityarea, hemaydecideto jointhem(Figure3). He mayalso electto workaway from the mainshelter if one of hisfellow householdersis sleeping there. Similarly, a groupof men playingcards maydo so in an areaaway from another group who are eating and talking, primarily becausetheactivities aretosome degreemutuallydisruptive.Becausevariationinthesizeand composition of the group physicallypresentis likely to be greater insingle sex than nuclearfamily households,variabilityinthe locationofactivitiesis alsolikelyto becorrespondingly greater. Special Activity Areas Certainkinds ofactivities regularlytakeplaceoutside the ordinaryhouseholdliving space. and thus create special activity areas. Four types of special activity area are defined: shady spots, roasting pits,auto repairstations, and defecationareas. Shady Spots.On warmdays.especiallyinsummer. individuals from one or morehouseholds oftenspend allor partofthe dayin naturallyshaded areas inother partsof thesettlement. These areas maybecoolerand more comfortable than householdareas themselves.or else offer alarger shaded areainwhichpeoplecancongregate.Activitiescarried out in theseareas are generallythe sameasthosethattakeplaceinhouseholdareas.asare thefactorsthatdeterminetheirlocation withineach area. RoastingPits.Thelargerlocalgameanimals,includingredkangaroo,euro.emu,andplains bustardare cooked wholeinroasting pits.These features vary in shape and size, depending on the number, size and type of animals to be cooked, butmost are oval in plan, basinshaped insection, and measure80-120cmlong,30-50cmwide, and 20-30cm deep. Theyare dug withshovels or digging sticks, and the backdirt piledon thedownwind side. Alarge fire is kindledonthe upwind sideandallowedtobumdowntocoals.The carcassissingedonthefire,thenplacedinthepit coveredwithcoals, ashes, and hotsand or soil from thefire area. Once cooked, itis removedand butcherednearbyon abedof treeboughsorasectionofsheetmetal. (See Gould[1967], Spencer andGillen [1899:23-241, and Binford [1983: 165-1 72, illustrations 10 1-1071 for detailed descriptions of this process.)Apartfrom the fact that the butcheringplatformis seldomlocated more than5 m from the roasting pit, there is no consistentspatial relationship between these features. Metatarsals are often. though not always left near the roasting pit when the butchering process is complete; other elements seldom are. Pit fill typicallyincludes a thick layer of chunky charcoal mixed with ash and burned earth.Roastingpits maybelocatedon the marginsof household activityareas, or inopen spotsbetweenhouseholdareas.Theyaresegregatedinthiswaybecausetheirusecreatesafire hazard. Auto RepairStations.Many Alyawaramenown motor vehicles, mainly second- or third-hand passengercarsorlighttruckspurchasedprivatelyfromotherAboriginesorfromEuropeans,or boughtfromretaildealersinAliceSprings.Nearlyallthesevehiclesareinpoorconditionand require constant maintenance and repair. This work is usually done at a convenient spot immediately adjacent to the owner's household activity area. Most auto repair stations consist oftwo components: AMERICAN ANTIQUITY[Vol.52,No.1,1987 Figure 6.Distribution ofoccupied household activity areas at Bendaijerum, 2 September 1974. anopenactivityarea,10-20mindiameter,whererepairsandmaintenancetakeplace;anda surrounding dump, where larger discarded parts are put to keep the working area clear. Apart from this, there is littleobvious regularity in the internal spatial organization of theseareas. Dejecation Areas.Most Alyawara defecate in sheltered areas at the edge of the settlement where they are screened from view by trees and bushes. Separate areas are used by men and women. Waste is left on the ground surface where it quickly dries and crumbles to dust. This activityis restricted to specialareas for reasonsof personalmodesty. Only the very old and very youngdefecate in the settlement. Distributionof Activity AreasWithin Settlements The distribution of household and special activity areas varies partly with settlement size. Where the number of households is less than12-1 5, individual household areas are likely to be uniformly distributed within the settlement at distances of 25-45m apart, center to center. Auto repair stations and roasting pits are generally immediately adjacent to individual household areas; defecation areas are locatedup to 200-300moff site. Inlargersettlements,householdandspecialactivityareasareoftengroupedinwell-defined O'Connell] ALYAWARASITE STRUCTURE Table 2.HouseholdCluster Populationand ClusterArea atThree AlyawaraSettlements WhereClustersAre Well Defined."Num-berNumber of Households ofCluster Worn- Resi- Area SiteDate MappedClusterFamilyen'sMen'sTotaldents( xl o3m2) Gurlanda AAugust4,197 1 I41- 52 18.4 I131- 4184.0 1112- - 2122.0 IV42- 63 112.0 Gurlanda BNovember 29,1971 I21- 381.8 I161- 72 89. 2 I1172- 94312.0 IV2- - 2101.8 V2- - 2122.4 BendaijerumMay2,1974 I5- 16235.5 Ridge I1511 73 18.5 I11511 7445.7 IV2- - 271.4 September 2,1974 I6- - 6304.1 V211 4203.8 VI821117211.5 March7,1975 VI5117346. 2 XI41- 53 16. 0 I4--4122.4 V32- 5263.5 aBendaijerumRidge is representedmore than once because on two occasions household distribution changed therefollowingthedeath of asiteresident. Cluster areais thatof thesmallestpolygondefinedbycontempo- raneouslyoccupiedhousehold activity areas. All1971 datafrom Denham(1971-1972). clusters, each containing 2-12households (R= 5.2-t2.4), or about10-75people ( a = 25.9-t15.7; Figure6,Table 2). Clustersare composedprimarilyofnuclearfamilies, butmost containat least one women'shouseholdas well.Youngmen'shouseholdsareoftenassociatedwithclusters,but senior men's households seldom are. The average distance between nearest neighboring households is 25-45m, the same as in smaller settlements. The distance between neighboring clusters (edge to edge) is usuallyabout100 m, butmaybeas great as 300m. Total cluster area and total settlement area are closely correlated with cluster population and site population, respectively (Table 3). For example, the relationship between total settlement area and number of households is bestexpressed: Table 3.Correlation and RegressionRelationships BetweenHouseholdClusterArea and Cluster Population, andSite AreaandSitePopulation,Respe~t i vel y. ~ Number Correlation andRegressionof Relationships Between:Casesrr2Sig.ab Cluster area and number of households20.89.79