Rhode Island College Digital Commons @ RIC

Faculty Publications

Spring 4-1993 "The ewN Guinea Highlands" Region, Culture Area, or Fuzzy Set? Terence E. Hays Rhode Island College, [email protected]

Follow this and additional works at: https://digitalcommons.ric.edu/facultypublications Part of the Other Anthropology Commons, and the Social and Cultural Anthropology Commons

Citation Hays, T. E., Brown, P., Harrison, S., Hauser-Schäublin, B., Hayano, D. M., Hirsch, E., ... & Westermark, G. D. (1993). " The eN w Guinea Highlands": Region, culture area, or fuzzy set? Current Anthropology, 34(2), 141-164.

This Article is brought to you for free and open access by Digital Commons @ RIC. It has been accepted for inclusion in Faculty Publications by an authorized administrator of Digital Commons @ RIC. For more information, please contact [email protected].

"The New Guinea Highlands": Region, Culture Area, or Fuzzy Set? [and Comments and Reply] Author(s): Terence E. Hays, Paula Brown, Simon Harrison, Brigitta Hauser-Schäublin, David M. Hayano, Eric Hirsch, Dan Jorgensen, Bruce M. Knauft, Rena Lederman, Edward Lipuma, Eugene Ogan, Andrew Strathern, James F. Weiner, George D. Westermark Source: Current Anthropology, Vol. 34, No. 2 (Apr., 1993), pp. 141-164 Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for Anthropological Research Stable URL: http://www.jstor.org/stable/2743972 Accessed: 06/12/2010 09:23

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the 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 at http://www.jstor.org/action/showPublisher?publisherCode=ucpress.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

The University of Chicago Press and Wenner-Gren Foundation for Anthropological Research are collaborating with JSTOR to digitize, preserve and extend access to Current Anthropology.

http://www.jstor.org CURRENT ANTHROPOLOGY Volume 34, Number 2, April I993 ? I993 byThe Wenner-GrenFoundation for Anthropological Research. All rightsreserved OOII-3204/93/3402-0002$2.50

Accordingto Mandeville (I980:549) "studentsagree on two things about New Guinea Highlanders: they ex- changepigs and theydo not conformto Africanmodels. "The New Guinea A good start,but clearlymore is needed." Whatevermay be the limits of agreementamong students,since the Highlands" ig5os anthropologistshave made a cottageindustry out of demarcatinga "region" called "the New Guinea Highlands" and tryingto identifyways in which "the Highlands" can be contrastedwith "the Lowlands" and Region,Culture Area, or what is to be found "there." For example, numerous Fuzzy Set?' differencesin religion and cosmology have been pro- posed as points of contrastbetween "Highlands" and "Seaboard" societies (Lawrence and Meggitt I965). More recently,Lindenbaum (I984:34I) tells us that by Terence E. Hays "from[a] larger,Melanesia-wide perspective,the New Guinea Highlands emerges as a region in which ritualized male homosexual experience is notably ab- sent"-indeed, "the broadest contrasts among Mel- anesian cultures emerge . . . from a comparison be- The criteriafor delineating "the New Guinea Highlands,"a fun- damentalcategory in Melanesiananthropology, are variable, tween the so-called semen groups of the Lowlands vague,and inconsistentlyapplied, with the result that there is lit- and the Highland cultures in which semen is not tle clarityor agreementwith regard to its characteristicsand its the ritualized stuffof life" (p. 342). Whitehead (I986) membership.So faras the literatureis concerned,"the New contends that in "the lowlands" a "manhood empha- GuineaHighlands" is a fuzzyset. The commonresort to notions of "cores,""margins," or "fringes"is an attemptto preservean sis" is to be found in fertilitycults while in "the essentialistapproach but inevitablyleads to the same confusion. highlands""clanhood" is emphasized. The list of char- The continueduse of "the Highlands"as an analyticor theoreti- acterizationsand contrastscould be extended through cal constructcarries the costs ofmisleadingly implied homoge- social and political organization (e.g., Harrison I989) neity,with marginalization of "exceptions,"ahistorical reifica- tionof social and cultural"traits," and deemphasison linkages to warfare(Knauft I990). amongcommunities. A plea is made herefor a shiftfrom studies These few examples are perhaps the kind of "more" ofmorphology to studiesof process-from concerns with what that Mandeville feels is needed, and presumablythey peopleare to concernswith what people do. are the sort of claims that she has in mind in saying that "it makes more than geographicalsense to think TERENCE E. HAYS iS Professorof Anthropology at RhodeIsland about the Highlands as a single area" (ig80:55o). In College(Providence, R.I. o02o8, U.S.A.). Bornin I942, he was ed- any event, they are indicative of how salient "the ucatedat the Universityof Omaha (B.A.,i966), the Universityof Colorado(M.A., i968), and the Universityof Washington (Ph.D., New Guinea Highlands" has become as a fundamental I974). His researchinterests are the ethnographyand ethnology categoryin Melanesian anthropologyas scholars have ofNew Guinea.His publicationsinclude the editedvolumes Eth- tried to develop explanations for social and cultural nographicPresents: Pioneering Anthropologists in thePapua phenomena with referenceto "regions" in which they New Guinea Highlands(Berkeley: University of CaliforniaPress, do or do not occur or in which they take particular i992), Oceania (Boston:G. K. Hall, i99i), and,with others, An- thropologyin theHigh Valleys:Essays on theNew Guinea High- forms. lands in Honorof Kenneth E. Read (Novato:Chandler, i987). My concern here is to examine the categorylabeled The presentpaper was submittedin finalform 2 x 92. "the New Guinea Highlands" as it has been used in several recent studies that offerexplicit comparisons of "the Highlands" with other "regions" (Lindenbaum I984, Whitehead I986, Weiner I988, Knauft I990) or that incorporatemajor surveys of "Highlands" societ- ies (Brown I978, Gelber I986, Feil I987). These works are the result of literaturesurveys from which ethno- graphiccases have been drawn, categorizedas "High- lands" or not, and compared for selected attributes. The criteria employed in these surveys and, conse- quently,their resulting classifications have varied con- i. Thispaper was originallyprepared for a workingseminar entitled "Not in Isolation:Regional Studies in MelanesianAnthropology," siderably,and when their internal inconsistencies are heldin i99I and cosponsoredby theWenner-Gren Foundation for combined with this variation in conceptualizationthe AnthropologicalResearch and the Field Museum of NaturalHis- situation becomes even more muddled. We find our- tory.I am gratefulto thosesponsors and the otherparticipants in selves in a position not only of wonderingwhat we the seminarfor a stimulatingdiscussion of centralissues and to ChrisGosden, Bruce Knauft, Paul Roscoe,Richard Scaglion, Robert know after all about "the Highlands" but of ques- Welsch,and thefour referees for this journal for their very helpful tioningin what sense "the Highlands" is usefullyre- commentson earlierdrafts of the paper. gardedas a "region" at all.

I4I I42 CURRENT ANTHROPOLOGY Volume 34, Number 2, April I993

Delineating "the Highlands" ferredto veryselectively [see StrathernI990]), and in Papua New Guinea it stretchessoutheastward well into It mightbe supposed fromits label that a categorysuch Milne Bay Province, including the country's third- as "the New Guinea Highlands" is basically organized highestmountain, Mount Victoria (at 4,072 m) in the around geographic or physical attributes,but which Owen StanleyRange near PortMoresby (King and Ranck "lands" are "high" is not self-evidenton an island whose n.d. [I982]:88-89). Nevertheless,the StricklandGorge relief extends from tide-washed coastline to snow- on the west and the KratkeRange in the east are often capped peaks at approximately4,5 Io m above sea level the effective,if not explicitlystated, east-westbound- in Papua New Guinea and about 4,740 m in Irian Jaya. aries of consideration.Such truncationscannot be un- Nor has there been agreementon the question among derstoodas motivatedby criteriabased on reliefor con- anthropologists. comitantvegetation or climatic patterns(see King and In one of the earliestattempts to delineatethe region, Ranck n.d. [i9821:92-93, 96-97). Thus, Brown's(I978:2) Read (I954:2) proposed that "the Highlands of New claim that"altitude, climate, temperature, and otheren- Guinea forma regionwhich is ... most simplydescribed vironmentalcharacteristics set the highlandsapart from as a chain of valleys lying at heights of fromfour to the tropicallowlands" may be true for"the highlands" seven thousand feet [I,2I2-2,I2I ml and stretching but it has not in practice been truefor "the Highlands." roughlyfrom east to west across the centerof the coun- When explicit reasons are given forthe exclusion of try."The geographerBrookfield (I96I:436) subsequently some high-elevationpeoples, includingsome withinthe placed "the highlandpeoples" at "between lat. 30 and central cordilleraitself, they tend to focus on subsis- lat. 70 S., at altitudesranging from 4300 to nearlygooo tence types,staple crops,and population density.Most feet[I,303-2,727 m]in valleysof the central cordillera." influentialin this regardhas been Brookfield'sdecision, Focusing on the easternhalf of the island, Bulmer and in his reviewof the "distribution"of "the highlandpeo- Bulmer(I964:39) extended"the Highlands of Australian ples ofNew Guinea" (i96i:437), to dropthe groups of New Guinea" to include "those parts of the Bismarck, "the Vogelkopto the west [in Irian Jaya]and the Kuku- Schraderand Central Ranges above 2,000 feet [606 ml kuku and Goilala areas to the east" because they"have which lie on the northernfringe of the [Eastern,West- not developed the intensiveagricultural forms that are ern,and SouthernHighlands Administrative Districts]." the best distinguishingcharacteristics of the highland For Brown(I 978: I 3), "in highlandvalleys at an altitude peoples."2According to Brookfield(I964:2I), an addi- of about S,000 feet (I, 5 20m), and on the slopes above tional "characteristicof these Highlands people-one them,are the settlementsand gardensof the highland- which distinguishesthem fromclosely-settled popula- ers.Between the mountain ranges surrounding these val- tionsat similaraltitudes in otherparts of the tropics-is leys and the New Guinea lowlands are steep slopes; the theirdependence on root crops,and especially on a sin- inhabitedarea lies between 3,000 feet(goo m) and 7,000 gle root crop, the sweet potato." Thus, using "a sensu feet(2,I00 m). This is the highlandsmargin and fringe." strictoapplication . . . not merelythe peopleson the Accordingto Gelber (i986:5), "[the]societies ofthe New outer slopes of the Cordillera,but also the inner mon- GuineaHighlands . . . lie between4500 and 8ooo feet tane Telefominand Ok Sibil groupsare excluded"; using [I,364 and 2,424 m] in altitude," and Feil, while never "a sensu lato definitionit is possible to include most of explicitlydemarcating the spatial boundariesof the sub- theseother montane people, though it becomes less easy ject of his recentbook, providesa map (I987:38) labeled to distinguishthese [emphasisadded] fromsome of the "peoples of highlandPapua New Guinea" which high- adjacentlowlanders on the bases of agricultureor popu- lights"land over I200 metres." A final example of im- lation density,and hence of ecological adjustment." plicit thresholdsmay be adduced with Weiner's map Brookfieldmust be creditedwith an earlierqualifica- (I988:4), which suggestsSoo-i,Soo m as the "Southern tion (i962:252), proposingthat the notion of a "simple FringeHighlands Area." region"must be "abandoned,and replacedby a series of Clearly, those who have tried to bound "the High- cores showinggradations outward," but the suggestion lands" geographicallyhave taken seriouslythe implied is not a part of his main legacy in the literature.Thus, salience of the adjective, but just as clearly they have Brown(I978:4-5) states that "the most distinctivefea- adopted differentcutoff points, with "high" apparently ture of highland culture is agriculturalspecialization, beginningas little as Soo or as much as I,300 m above which supportslarge concentrationsof people and peri- sea level. odic festivalsat which thousands of visitorsare enter- Elevation per se appears not to be a sufficientcrite- tainedand feasted."For Lindenbaum(i984:343), "High- rion, however,since none of the writerswhose works land societies are . .. based on the intensiveproduction are consideredhere (nor any anthropologistof whom I of sweet potato and domestic pig-herding,. . . whereas am aware)routinely includes in "the Papua New Guinea the smallerLowland groupstend a differentassemblage Highlands" the high-elevationpeoples of the Torricelli, Finisterre,and Owen Stanley Ranges (to cite only the most obvious candidates). Lack of contiguitywith the 2. In particular,Brookfield identified seven features of agricultural omis- methods(dominant fallow cover, method of clearing, ground prepa- central cordilleracannot account for all of the ration,erosion control, water control, mulching and fertilization, sions. Geologically,the cordillerabegins in the farwest and intercropping)as "the principalcriteria for defining a 'central of Irian Jaya(a half of the island usually ignoredor re- highland'region on the basis ofagriculture" (i962:246). HAYS "The New Guinea Highlands" I I43

ofcrops ... accompaniedby huntingand fishing."How- as exclusion ofthat case fromthe "Highlands" category. ever,as Strathern(I990:379) has asserted,when one is In some instances, assignments seem idiosyncratic examining "agricultural intensification,there is cer- whateverthe constraintsimposed by the literature;for tainly no reason why cases fromthe Sepik, fromOk, example,Knauft (I990:277) citesthe Orokaiva and Tau- fromthe highlandfringes, or indeedfrom anywhere else, ade in his surveyof "New Guinea Highlands warfare," should not be chosen fordiscussion." Moreover,the dis- and Whitehead(i 986:87) includes the Awa and Ndumba tributionof sweet potato as the primarystaple crop (southernTairora) of the Eastern Highlands Province (Kingand Ranck n.d. [i9821:50-5 I) matchesnone of the along with such groups as Chambri, Iatmul, Abelam, currentdelineations of "the Highlands," and the same and Arapesh in discussing the "manhood emphasis in can be said forpopulation distribution (King and Ranck the lowlands." In any case, some indicationof the lack n.d.[i9821:20-2i). of consensus among anthropologistsregarding the cate- While environmental,ecological, economic,or demo- gorizationof specificcases can be seen in explicitlabel- graphicattributes are the most common ostensible cri- ing when it occurs. This is most obvious with respect teriafor demarcating "the Highlands," in factit is rare to the "Mountain Ok" groups,those of the "Bosavi re- in the anthropologicalliterature for such featuresto be gion" and of the Karimui area, and the "Anga groups." privilegedin explanationsof, or even consideredas being Craig(I990) has recentlyposed the question"Is the ofmuch causal relevanceto, the social or culturaltraits Mountain Ok culture a Sepik culture?" (forthe prior that are the usual foci of attention.Instead, "big man" question as to "the Sepik as a culture area," see Mead political leadership,ceremonial exchange systems,clan I978, BrownI99I) andanswered it "roughly"in theneg- parishorganization, bride price, and pig festivalsrecur so ative: "the societies most like the Mountain Ok societ- frequentlyin characterizationsof "Highlands" societies ies are to the west, in the centralranges and foothillsof that they almost achieve the status of diagnosticfea- the easternmostinterior of Irian Jaya" (p. i29). Are they, tures.However, not only can one easily point to "Low- then,"Highlanders"? For Feil, the answeris straightfor- land" or island examples of such featuresbut those who ward (I987:7, emphasis added): "Beyond [the "western discuss them are oftenat pains to note and tryto ac- highlandssocieties"], further to the west,are foundsoci- count fortheir variability within "the Highlands." It is eties of a differentsort (forexample those of Telefomin preciselysuch variability(most systematicallysurveyed and other Ok groups),whose adaptation and cultural by Feil [I987] but acknowledgedby nearlyall of the emphases are unrelated to the highlands." For others, scholars discussed here), as well as variation with re- theirambiguous status is made hardlyless so by writers' spect to ecological and subsistence features,that has phrasings. Thus Brown includes them on her map given rise to the increasing tendency to distinguish (I978:6) of "the highlandarea" but explicitlygrants the "core" from"margin," "fringe," and the like. Baktaman,Miyanmin, and Telefolmin qualified mem- bershipas "fringegroups" (p. I3). Similarly,Whitehead (I986:86) considers "some Telefomin area groups" as Who Are "Highlanders"? "fringe"groups, yet later (pp. 89-95) cites Baktamanin herdiscussion of "clanhood emphasisin the highlands." Despite Gelber's (i986:3) claim that "the Highlands So, too, Knauft(i990:280-8i, emphasis added) locates have distinct geographicalboundaries," the attributes Baktaman,Bimin, Miyanmin,and Ngalumin (with the most commonlyused by anthropologistsin definingor latterreferring to Atbalmin,not the Ngalum-speakersof characterizing"the Highlands" as a regional category Irian Jaya)"in the fringeareas . . . west of the Papua are variable,vague, and inconsistentlyapplied. It is not New Guinea highlands." surprising,then, that the ethnographiccases assigned This employmentof the qualifyingadjective "fringe" membershipin it differas well. is also characteristic,although not uniformlyso, of Only rarelydo the writerssurveyed here list the cul- treatmentof the Bosavi area or "Great Papuan Plateau." tural or linguisticgroups included in the "Highlands" Thus, while Brown included a chapter on the Etoro category.When they do, the lists are sometimesincon- (Kelly I976) in her coeditedcollection Man and Woman sistentwith definingstatements. Gelber, for example, in the New Guinea Highlands, in her syntheticover- says (I986:6) that "the Highlands" includes the groups view (I978:6) the Etoroand Kaluli are considered"fringe "fromthe Enga in the west to the Fore in the east (the groups,"as theyalso are by Gelber (I986:6), Whitehead Mendi and Huli beingthe southernmostand the Maring (I986:86), and Knauft(i990:28i) (thelatter adding Be- the furthestnorth)" but then includes in her compara- damini,Gebusi, and Onabasulu fromthe same area). Feil tivetable (pp. i0-i i) the Tairora,who are in factlocated is inconsistent,referring to the "Papuan Plateau" as east of the Fore. Most often,cases are simply adduced "part of the congeries of peoples recently termed forillustrative or analyticalpurposes, and these are too 'SWNG'-southwestern New Guinea coastal fringe" variable to allow systematiccomparisons of inferable (I987:5-6, emphasis added) but includingEtoro and Ka- lists of groups. This is understandable,perhaps, when luli on his map (p. 38) of "peoples of highland Papua surveysfocused on a single topic (e.g., fertilitycultism New Guinea" while excludingthem fromhis table (pp. or ritualizedhomosexuality) must be guidedby available 42-43) showing"language family size in highlandPapua information;failure to cite a givencase in a presentation New Guinea." Weiner (i 988), in contrastto all of the of "Highlands" forms,then, cannot necessarily be taken others,rejects the Bosavi peoples as "fringe"groups, in- I44 1 CURRENT ANTHROPOLOGY Volume 34, Number 2, April I993

cluding them instead in his new "Mountain Papuans" studies. Variabilityin inferablelists of "Highlands" so- category. cieties follows as a matterof course. The Karimuiarea is a bit more confusing,with Brown (I978:I4) consideringthe Daribi a "Highlands fringe" group,Feil (I987:38) placing them ambiguouslyon his "Core" and "Fringe" map but apparentlyconsidering them an "easternhigh- lands" society(p. 3o), Knauft(i990:265-79) including Clearly,much of the apparentdiversity in delineations the Polopa in his discussion of "New Guinea Highlands of the elevationalboundaries of "the Highlands" results warfare,"and Weiner (i988) countingboth Daribi and fromthe variable inclusion of the "fringe"or "margin." Foraba(Polopa) as "Mountain Papuans." As fortheir eth- Thus,Read's proposal (I954:2), citedearlier, placed the nographers,D. J.J. Brown (I979:7i2, emphasis added) valleysof the "centralhighlands" at I,2I2-2,I2I m, comparesthe Polopa to the Melpa, "another Highlands which is perhapsnot significantlydifferent from Brook- people," and Wagner both contrasts the Daribi with field's (i96i:436) range of I,303-2,727 m or Feil's "Highlanders"(i 967: I I) anduses Daribisocial organiza- (i987:38) "overi200 metres."Brown (I978:I3) seemsto tion (I974) to answer the question "Are there social centerthem on i,520 m, and this would be consistent groupsin the New Guinea highlands?" with theirplacement on Weiner's (i988:4) map, where Finally,there are the "Anga groups,"occupying highly the upperlimit of the "southernfringe" is i,5oo m. We diverseenvironments in the EasternHighlands, Morobe, mightbe justified,then, in the inferencethat "the core" and Gulf Provinces.3Brown (I978:6) includes at least of "the Highlands" is generallythought to be foundat some of them on her map of "the highland area" but about i,200 m and above, with the "fringe"extending otherwise does not mention them in her survey of down to 50 m in the south (Weiner)and 6o6 m in the "highlandpeoples of New Guinea." The Baruyaand the north(Bulmer and Bulmer I964:39). (pseudonymous)"Sambia" are consistentlyregarded as However, as with "the Highlands" in general,eleva- "fringegroups" by Lindenbaum (i984), Gelber (i986:6), tion alone seems not to be the criterionfor distinguish- Whitehead (i986:87 [adding Yagwoia as well]), and ing "core" from"fringe"; indeed, some of the writers Knauft(i990:268). Again,Feil seemsunable to makeup surveyeddo not even seek to specifyelevations. Brook- his mind,referring to Baruyaand "Sambia" as "livingat field (i962:253), as we have seen, appears to assess de- the far eastern fringeof the highlands" (i987:I76) and greeof "highlanderness"in termsof intensity of agricul- includingthem on his map but excludingthem from his turaltechniques. Others have focusedon othercriteria. table. Weiner's(I988:3) are ostensiblygeographical and ecolog- We have seen, then,that in an arguablyrepresentative ical: "I referto 'FringeHighlanders' as those people who, sample ofrecent anthropological writings on "the High- like the Mountain Papuans, live in valleys on the edge lands" no single attribute-environmental,ecological, of the central cordillera,valleys that are significantly demographic,social, cultural, or linguistic4-is a reli- lower in altitude and which consequentlyhave a mark- able predictorof which ethnographiccases will be in- edly differentenvironment," with "the special features cluded in the categoryeither by a givenwriter or across of the fringedwellers" including"low population den- sity,broad-based low-intensity subsistence production, and communal longhouse residence" (p. 2). For Knauft 3. It is not uncommonfor authors of books forgeneral audiences (I990:268), "peripheraland fringeareas of the New to use politicalboundaries in demarcatingthe "Highlands"region Guinea highlands"had "much lower population densi- ofPapua New Guinea.Thus, Sinclair ( I 97 Ixix) cites"the four High- landsdistricts-Eastern, Western and SouthernHighlands and the ties, ample land, and placed little if any emphasis on Chimbu,"just as Millerseems, judging from his map (I983:I4-I5), land acquisitionthrough warfare." Gelber (i 986:6) views to employthe now-equivalent Eastern Highlands, Simbu, Western "groupson the fringeareas ofthe Highlands"as differing Highlands,Enga, and SouthernHighlands provinces. None of the "considerablyfrom the Highlandsin populationdensity, anthropologistswhose works are examinedhere uses political boundariesin thisstraightforward manner. Indeed, it maybe worth horticulturalpractices, staple crop,reliance on hunting, notingthat among the peoples sometimesconsidered "highland- comparativeunimportance of pigs, and lack of elaborate ers" thatare foundoutside of the fiveprovinces listed above are exchange,as well as in ritual organizationand in their theKalam, Gende, and some Maring(Madang Province), all ofthe sexual orientationand concerns." Whitehead compli- MountainOk (WestSepik and Western),Waffa and mostYagwoia cates the picture somewhat by combining "fringe" (Morobe),Polopa and some Simbari(Gulf), and some Bogayaand some Duna (Western).Also, thereare groups,usually not consid- groupswith "lowlanders" in her analysis,as well as ap- ered"highlanders," that straddle "highlands" and otherprovinces, parentlyextending her rangebeyond the centralcordil- amongthem the Hewa (Engaand East Sepik)and theBeami, Sonia, lera (I986:84, emphasis added): "On the marginsof the andTomu Riverlanguage-groups (Southern Highlands and Western highlands,and at middle elevations throughoutthe is- provinces). smaller clusters 4. Feil's (i987:42-43) table is apparentlybased on Wurm's(i982) land, distinctly population practice assignmentof languagesto the "East New Guinea Highlands mixed crop cultivation,modest (sometimesvanishingly Stock."A possibleunstated linguistic bias elsewherein thelitera- modest) pig husbandry,and foraging.These groups,of- tureis suggestedby the factthat among the Papua New Guinea ten termed'fringe,' are quite variedin regardto ceremo- groupsthat are not includedin thatstock are the MountainOk, nial exchange."Feil (i987:5-6) proposessimply that the the Anga,the Lake Kutubupeoples, those of the Bosaviarea, and theDaribi and the Polopa-all ofthe groups most frequently con- people of "the highlandsof Papua New Guinea ... have, sideredeither "fringe Highlanders" or "MountainPapuans." forcomparative purposes, been distinguishedfrom the HAYS "The New Guinea Highlands" I I45

so-called 'highland fringe'groups and those of the Pa- elementsin common with one or more otheritems, but puan Plateau . . . and even more so fromthe coastal no, or few,elements are common to all items" (p. 575). 'seaboard' (particularlySepik) peoples on the bases of Common indicatorsin speech behaviorof a fuzzyset geography,subsistence, language, and highlydivergent includethe use ofqualifying adjectives in labelingmem- aspects of society and culture,"but he does not specify bers,as when a color is called "offred" or "blue-green" the natureof the distinctionsbetween any two of these (Kay and McDaniel I978). Upon examination,such us- categories.Finally, for Brown (I978: I 3) "fringe"groups ages point the way to the identificationof exemplarsas "seem mostlyto be betweenlowlanders and highlanders "a prototype(clearest cases, best examples of the cate- in culture; many speak languages of groupsalso found gory) and nonprototypemembers, with nonprototype in the lowlands. They are characterizedby small and memberstending toward an orderfrom better to poorer scattered settlements and partial dependence upon examples"(Rosch and Mervis I975:574). The "mostpro- huntingand gathering." totypicalmembers ... are those which bear the greatest In this arrayof characterizationsof "fringe"peoples familyresemblance to othermembers of theirown cate- are common threads; in particular,"traits" such as goryand have the least overlap with other categories" smallness of populations and mixed subsistence econ- (pp. 598-99), "prototypicality"being "a functionof the omyseem to recurmost frequently.Deferring to another cue validity [or predictiveutility] of the attributesof occasion a systematicreview of the ethnographiclitera- items" (p. 599). ture with respectto these and othervariables, I would With respect to "the New Guinea Highlands," it say here only thatI findthese traitsdifficult to consider seems clearfrom the usages reviewedabove that,despite as aptly describing the Telefolmin (Brown I978:I3; attempts to specify attributes (elevation, population WhiteheadI986:86; Knaufti990:268), the Baruyaand density, agriculturaltechniques, staple crops, settle- "Sambia" (LindenbaumI984; Gelber I986:6; Whitehead ment types, or social institutionssuch as ceremonial I986:86; Feil I987:I76; Knauft i990:268), the Awa and exchange),when anthropologistsassign societies to "the southernTairora (Whitehead I986:87), and the Kewa, Highlands" these attributesare less oftentruly diagnos- Wola,Maring, and Huli (Knaufti990:268), to nameonly tic than loosely employed,with weightingson sliding some of the purportedly"fringe" groups. scales. Thus, "the class of definingattributes that con- stitutesthe intensionof the termis not a class of attri- butes that are severallynecessary and jointlysufficient, "The New Guinea Highlands" as a Fuzzy Set but a 'polythetic group' or 'imperfectcommunity"' (AtranI990:54). Thatthis is evidentto mostis indicated If the worksreviewed here are representativeof the cur- by the groupingof societies into "core" (prototypic)and rentstate of comparativestudies in Melanesian anthro- "fringe"and like extensionsof categorymembership. pology(and I believe theyare in manyrespects), it would Whenanthropologists elaborate "the Highlands"cate- seem that one of the fundamentalcategories used in gory to accommodate prototypic("core") cases and such comparisons-"the New Guinea Highlands"-is ("fringe")extensions through family resemblances, they employed with little consistency or clarity.However do not therebyresolve all of the definitionalcomplexi- fundamentalit may be to anthropologicaldiscourse and ties that this categoryentails. This is only partlybe- however much we may act as if it correspondedto cause differentwriters still sort cases differently,de- a "real region," its use does not resemble that of pending on which attributesthey highlightand how what cognitivepsychologists call "basic categories"- carefullythey apply them. In fact,there is probablysub- "information-richbundles of perceptualand functional stantialagreement, at least regarding"the core." While attributes. . . that formnatural discontinuities"in the groupsthat belong to the "core" have largelyto be iden- world(Rosch et al. I976:385). Indeed,it does not even tifiedthrough a process of elimination (coming down seem to be a categoryin the traditionalsense of that to those that are not designatedas "fringe")or by the termin set theory-a "logical bounded [entity],mem- frequencyof theiruse as main exemplars,one could say bershipin which is definedby an item's possession of a that the language groups called Enga, Melpa, Kuma, simple set of criterialfeatures, in which all instances Chimbu, and those clustered around Goroka and Kai- possessingthe criterialattributes have a full and equal nantu are the ones most oftenregarded unambiguously degree of membership" (Rosch and Mervis I975:573- as "Highlanders."5One importantcomplication is sug- 74). Instead, "the New Guinea Highlands" as used in gested by the popularityof what might be called "the anthropologicaldiscourse exemplifieswell what cogni- continuum game," recentlyplayed most comprehen- tive psychologistswould call a "fuzzyset." In fuzzy-set sively by Feil (i987) in his designationof several main theory(Zadeh i965), "the referentsof a word [orphrase] "societal configurations"distinguished on the basis of need not have common elements in orderfor the word their variable elaboration of ceremonial exchange or to be understoodand used in the normalfunctioning of other"traits." When such continua are developed,even language"(Rosch and MervisI975:574-75). Rather,"a "core"/groups are treatedas constitutingchains of soci- familyresemblance might be what [links] the various eties linked throughfamily resemblances with respect referentsof a word. A familyresemblance relationship consists of a set of items of the formAB, BC, CD, DE. 5. Knauft (I990:275) is probablyidiosyncratic in consideringSiane That is, each item has at least one, and probablyseveral, ''non-core"but apparentlynot "fringe." I46 CURRENT ANTHROPOLOGY Volume 34, Number 2, April I993

to one attributeor another.A second complication is "exemplarytexts" regardingthe "Enga," "Melpa," and illustratedby the diversityto be foundwithin language "Chimbu," with their"big men," ceremonialexchange groups, especially in the north, such as "Enga" and systems,and challenges to or refinementsof "African "Melpa" (hence my use of quotation marks). While models" of descent,have come to represent"the High- in most discussions "the Enga" really means Mae or lands."7 Raiapu Enga, a cautionaryreminder may be found in But clearly those who invest considerableenergy in Dornstreich's(I974:475-87) carefulcharting of a "cul- tryingto establish contrastsbetween "the Highlands" tural typology of Enga-speakingpeoples," in which and other "regions" and then theoreticallyto account he distinguishes "Central" from "Intermediate" and forthem believe that they are doing more than merely "Outer Enga," with the lattertwo kinds of "Enga" al- employingan arbitraryprofessional sorting device. In most certaincandidates for "fringe" status at best. The these "postmodern"times, it is perhaps not surprising same could be said fornorthern "Melpa, " whose ecologi- that none of the writersdiscussed here uses the old- cal situationdiffers strikingly from that of better-known fashionedterm "culture area," yet it would seem that "Melpa" groupsaround Mount Hagen, leading Gorecki when theywrite of "the Highlands"vs. "the Lowlands," and Gillieson (i989) to include them in the northern etc., they are in fact invoking that concept: "Culture "Highland fringe." areas are geographicalterritories in which characteristic culturepatterns are recognizablethrough repeated asso- ciations of specific traits and, usually, throughone or The Utility of "the New Guinea Highlands" moremodes of subsistencethat are relatedto the partic- ular environment"(Ehrich and Henderson I968:563). I Many anthropologistswill perhapsbe neithersurprised will not rehearsehere the problemswhose cumulative nor troubledby the results of my review.6Feil (i987:6) weightplayed a large part in the near disappearanceof seems content with "the rathervague concept of the the term "culture area" fromcontemporary anthropo- highlands as a cultural-ecologicalunit," and Brown logical discourse (at least outside of pedagogical con- (I978:i8) concedes that "the regionwhich we consider texts,where it still thrives),but it is worthnoting briefly here has no precise physical boundary,and any social a few of the costs incurredby the attemptsso far to boundarywould be arbitrary,cutting social linkagesand employ"the New Guinea Highlands" as a theoreticalor traderoutes." For all its "fuzziness," it is virtuallycer- analyticalconstruct. tain that "the New Guinea Highlands" will remain,in First,such usages misleadinglyimply greater environ- some sense, a "region" in which some people do their mental, social, and cultural homogeneitythan can be research and about which much will continue to be demonstratedeven within the prototypicmain exem- written-if forno otherreason because of the social or- plars such as the "Enga," "Melpa," or "Chimbu."8 As ganizationof our discipline or, as Fardon(I990:24) views some(e.g., Strathern I990) havepointed out, the "High- it, because "regionalismis so pronounceda featureof land" groups of Irian Jaya,where ceremonial exchange our professionalpractice." As he elaborates: systems appear to be rare or absent, are routinelyig- nored,a failingthat is doubtless attributableat least in Regional factorsinfluence the entry(in the broadest part to the fact that much of the relevantliterature is sense) of the ethnographerto a fieldthat is neces- published only in Dutch or German. But even within sarilypre-imagined, the circumstancesunder which Papua New Guinea apparentexceptions to depictionsof fieldworkwill be carriedout, the issues which have the most celebratedcases are marginalizedby creating been preconceivedas appropriateand pressing,and, "fringes" or "Mountain Papuans" and then treating in writingup, the canons of adequate reportingand themas ifit were theythat required some special expla- the audience to whom, in part at least, the work nation. will be addressedand whose opinions will be the Second, "Highlands societies" are frequentlypor- most telling. trayed,individually and collectively,with little if any Unquestionably, in Melanesian studies, "the New considerationof the antiquityor historical stabilityof Guinea Highlands" has come to have such influences, the "configurations"or institutionsattributed to them and those influencesextend "outside the narrowcircle (or to theirfoils in "the Lowlands" or wherever).While ofregional specialists. The most pervasiveof these is the projectionfor non-specialists of regionalrepresentations 7. In urbancenters of Papua New Guinea hailan (Tok Pisin for (oftenvia exemplarytexts or, as commonly,secondary "highlands")is in frequentuse as an ethnicmarker. In myexperi- rescensionsof them) which establish an image of place ence, this expressionoften connotes or is used interchangeably with"Hagen" or "Simbu"(Chimbu), just as Bougainvilleansin the in terms of particularproblematics which it typifies" I960s came to form images of "Highlanders-of whom the (Fardon i990:26). Thus, for many anthropologiststhe Chimbuserved as the prototype"(Nash and Ogan I990:7). The growingliterature on ethnicityin the Pacific(see, e.g.,Linnekin and Poyeri990) offersmany fascinatingexamples of "regional" 6. An analysissuch as thisone couldbe carriedout forother "cul- identitiesbuilt in nonacademicdiscourse upon "fuzzysets." tureareas" that are doubtless as "fuzzy"as "theNew GuineaHigh- 8. Suchimplications of homogeneity are often the basis ofcultural lands," suchas "Amazonia," "theNorthwest Coast, " "Melanesia," stereotypes(e.g., "flamboyant"and "bellicose" yet "pragmatic" and "Polynesia"(Thomas i989); see also Knauft's(I993) superb Highlanders),as Herzfeld(i984) has pointedout in relationto an- dissectionof "SouthNew Guinea." otherquestionable "culture area," "the Mediterranean." HAYS "The New Guinea Highlands" I I47

it may be as difficultto conceive of "Highlands" peoples in criticismof Feil (I987) but with implicationsfor all withoutsweet potatoes as of "Plains Indians" without of us, much recentwork horses,in both cases we are dealing with adoptionsno ignoresthe wider question of relationswith the morethan a fewcenturies old. Feil (I987) has been atten- people of the lowlands and with otherhighlanders tive to pre-Ipomoea precursorsso far as crops and ag- throughthe lowlands which are posed by the re- riculturaltechniques are concerned,but he bases his so- markablesimilarities between highlandpeoples sev- cial reconstructionsprimarily on what we know from eral hundredkilometres apart and with no possible the "ethnographicpresent." With respectto the latter, contactwithin the cordillera.It is simplynot possi- fewhave consideredwith any seriousnessthe degreeto ble to writea credibleaccount of the evolution of which featuresoften regardedas diagnostic of "High- highlandPapua New Guinea societies in isolation. lands societies," such as large-scale ceremonial ex- change and "big man" leadership,were affectedif not If "the Highlands," "fringes,""the Lowlands," and createdby the colonial process (cf.Hughes I978).9 other commonplace abstractionsare to signifysome- Third, at the same time that writersare imprecise thingmore than geographical regions, we need attention and inconsistentwith each otherin drawingthe spatial to all ofthese caveats,but at minimumwe need explicit boundaries of "the Highlands," their descriptionsand statementsof principled criteria for bounding such cate- analysespersist in treatingthe societies as if theyor the gories,and we need to apply them in such a way that "region" as a whole could be understoodin isolation. both inclusions and exclusions are clearly motivated This view can be traced to the firstdecade or so after and consistentlyexecuted. This does not mean thatcon- the "openingup" of "the Highlands" to anthropologists. sensus or uniformitywill necessarilyresult, for the cri- Thus, Read (I954:2) declared: "The Highlands of New teria and resultingsortings will inevitablydepend, at Guinea forma regionwhich is more or less isolated geo- least in part,on factorsextrinsic to "the region" itself. graphicallyfrom the surroundingcountry by highmoun- As "big men" vs. "great men," ritualizedhomosexual- tain ranges," and Watson (i964:2) extended the point ity, and sexual antagonismgo in and out of vogue as beyondmere geography:"The pre-i930 isolation of the theoreticalfoci, "core," "fringe,"and "continua" will Highlandshas provento be more than a question of the doubtlesscontinue to be identifiedin termsof attributes lack of reportsby literateexplorers.... the area is an- or considerationsthat are particularto the agenda of thropologicallya good deal more than a region by de- the researcher.Anthropologists are human beings, and fault.The conditionsthat isolated the Highlandsappear "fuzzy-set"and "prototype/extension"models, after all, to have given it a distinctivecharacter reaching well flow from a general information-processingstrategy back into time." Perhapssuch statementswere intended (Atran1990:55): to stress the fact that fieldworkerswere encoun- only Prototypesfacilitate the patterningof input foruse in I950S I96os, tering, the and early societies that con- in memoryand forone's actual dealingswith the trastedin many ways with those of the better-known day-to-dayworld by describingsimilarities among "Lowlands" and islands. in the of But, any case, image particularlyuseful, salient or familiarclusters of ex- as "isolated" has doubtlesshindered "Highlandpeoples" emplars.Prototypical patterning is thus contingent our of them. understanding on memoryand use. Because memoryand use are Brown(I978:29) has acknowledgedthat "the small influencedby context,prototypical patterning trans- communities and of the fragmentedgroups fringe formsin accordancewith changes in historyand so- area . . . have been intermediariesand always traders, ciety,with the extentand natureof such transforma- new ideas and into the bringing techniques highlands tions varyingas much as individualsand cultures fromthe outside." But few have been made so attempts vary. far to document systematicallyor include in explana- toryefforts the "social linkages and trade routes" in- Definitionalissues will remain, then, and boundary volved(Brown I978:I8). Yet it was preciselysuch link- disputeswill be a continuingfeature of any "essential- ages and routes that connected "the Highlands" with ist" approach to the peoples and societies of New the northcoast and even areas beyond New Guinea in Guinea and Melanesia. But quests forthe "traits" that trade in bird plumes (Healey I980) and in shells and trulydistinguish "Highlanders" from "Lowlanders" and other marine products (Hughes I977), the diffusionof for their correlatesfrom which explanations for such tobacco (Hays i990), and the movement(with localized differencescan be developed are not the only kinds of transformationsand permutations)of cults (Hays I986). explorationsthat can be conducted. It may even be, contraWatson above, that such linkages have been instrumentalin the development of what commonalities can be observed in many "Highlands" From Morphology to Process societies.As Brookfield(I990:69, emphasis added) notes Essentialistapproaches to "the New Guinea Highlands" as a "region"have reliedupon the categorizationof soci- 9. Even morerecently, it was only in the I970S that the Irakia Awa beganto intensifypig productionfor purposes of ceremonial eties in terms of their "culture-bearingaspect." Thus, exchange(Boyd. I 98 5); perhapsthis will be sufficientfor Whitehead "core"/and "fringe"groups, like "ethnic groups,"have (i986:87) to elevatethem to "Highlander"status. been distinguished"by the morphologicalcharacteris- I48 CURRENT ANTHROPOLOGY Volume 34, Number 2, April I993

tics of the culturesof which theyare the bearers"(Barth "the Highlands fringe"to academic equality with "the i969:i2). In this approach,"differences in groupsbe- true Highlands" but as an indication of how little any come differencesin trait inventories;the attentionis such categories help us when we shift our attention drawnto the analysis of cultures,not of ethnic organi- from social morphology to social process-in other zation." words, from nouns to verbs. I am proposing here a An alternativeprogram would take as its primaryob- changeof focusfrom what people are to what people do ject the social linkagesthat are in dangerof beingother- (e.g.,trade, engage in ceremonialexchange, intermarry, wise obscured.In Barth's terms (i969:i5), "the critical fight)or, indeed, do not do. Such a shiftcan, in fact,help focus of investigationfrom this point of view becomes us arriveat betteranswers to both kinds of questions, the ethnicboundary that definesthe group,not the cul- for,as Barth (i969:io) has argued,"ethnic distinctions tural stuffthat it encloses." With attentionto bound- do not depend on an absence of social interactionand aries and linkages,the "fringes"assume criticalimpor- acceptance,but are quite to the contraryoften the very tance. One example of the kinds of linkages that have foundations on which embracing social systems are in fact long characterizedthe actual situation is to be built." These systems,such as those disclosedby studies foundin the traderoutes referred to above by Brown.As of transformationsin cults withinparticular geographi- in Healey's study(I980) ofthe plume tradeand Hughes's cal regions (e.g., Knauft I985, Stratherni99i) or the (I977) regardingsalt, pigments,pottery, stone tools, and "communityof culture" suggestedby linkages among shells, in my own ongoing investigationof the spread diverse language-groupson the north coast of New of tobacco and smokingin New Guinea and the social Guinea (Welsch,Terrell, and Nadolski i992), can incor- dynamicsby which it was effected(see, e.g., Hays I990), poratedeliberately maintained differences or boundaries categories such as "Highlands," "fringe,"and "Low- as well as those thatmay be productsof varying environ- lands" are irrelevant.Thus, speakingonly ofPapua New ments,resources, or local histories.It should be appar- Guinea, I can now demonstratethe existence of trade ent thatsuch systemscan onlybe discoveredand under- networks involving tobacco which linked northern stood by tracingout the connectionsand boundariesof Chimbu with the Ramu Valley and Mae Enga with the particularkinds of interaction-in what are sometimes Sepik foothills.In the south, I am able to document a called "village-outward"studies-rather than by impos- vast system that joined the Huli with the peoples not ing a prioria gridof traitsto demarcatea "region." onlyof the Bosavi area and the StricklandPlain but also Put simply,the "ethnic groups" of "the Highlands," ofareas to the east as faras Lake Kutubu,the Kikoriand "fringe,"and "Lowlands" have long been engagedin nu- Turama Rivers,and the south coast. merous and wide-rangingnetworks of interaction,but How can such linkageshave escaped our attentionfor the result has not been homogeneityor uniformity. so long?To be sure,as Brown(I978:28-29) and others Why not? Whichever "region" interestsus most and (e.g., Weiner I988) have noted, much of our detailed howeverwe choose to defineit, we must wonder why knowledge of such peoples as those on the "southern it is not larger.Of course, constraintsimposed by envi- fringe"has come only fromrecent fieldwork(though ronment,climate, and disease may be a part of the an- Williamslong ago [1I940-4I] demonstratedthe key role swerso faras some traits(e.g., subsistence base and pop- played by those in the Lake Kutubu area in linkingthe ulation density) are concerned (BrookfieldI964), but southernhighlands with the south coast). But this re- what about the rest?If it is people as much as "Nature" cency of attentionto the "fringegroups" may itselfbe that create,maintain, or ignoreboundaries, we need to due less to the fact that "they are sparselydistributed know how and why, and forthat we need new ways of in relativelyinaccessible areas" (BrownI978:29) than to framingour questions. an essentialistview of such societies as "peripheral"or "marginal"in more than a geographicalsense.10 These remarksare intendednot as a plea forelevating Comments

io. RichardScaglion (personal communication) has remindedme PAULA BROWN of anthropologyto the east of New of a parallelin the history W. I2th Guinea,where some partsof Micronesia and the SolomonIslands 59 St.,New York,N.Y. iooii-8527, U.S.A. havelong been categorized as "PolynesianOutliers." Such margin- 5 XI 92 alizationmay be understandablefrom the viewpointof Central Polynesia,with its largechiefdoms and even kingdoms,but from I must respond,for my name has neverbefore been cited a prehistorian'sperspective the situationcan look quite different so manytimes in such a shortspace. Hays can fairlysay (Terrelli986:i2o-21): "Ironicallyenough, we are now beginning to see thatthe trueoutliers of the Polynesianrealm may not be that the central highlands of Papua New Guinea, as a thePolynesian-speaking communities found on thefringes of Mel- geographical-culturalregion, was inventedin the i950S anesia and Micronesiaafter all. If thereare outliersin the Pa- and I96os. Read(I954) was certainlyinfluential. Histor- cific-places offthe beaten path and awayfrom the main arena of ically, it mightalso be said that it is a regionby virtue Pacificprehistory-the real Polynesianoutliers are morelikely to after I930. havebeen those islands, large and small,geographically situated at of discoveryand settlementby Europeans the distantcorners of the greatPolynesian triangle: Hawaii, New Therewas a period,about I946-50 I think,when "Cen- Zealand,and EasterIsland." tralHighlands" was a districtand the name used to iden- HAYS "The New Guinea Highlands" I I49

tifythe area. I can rememberthe firsttime I heardabout sexual antagonism,gender, inequality, leadership, law, the highlands: In I956 I had just arrivedin Canberra. marriage,initiation, sorcery, and history,in additionto The newlyformed Association ofSocial Anthropologists generalessays and books which coverall or partof Mel- (Australianbranch) broughttogether Australians who anesia. had recentlycompleted fieldwork in the highlands;the It may be no more than a convenience in defininga discussion,typical forsocial anthropologyof the time, regionfor teaching purposes and researcharea identifi- centeredon terminologyfor descent, groups, and com- cation-textbooks, the organizationof lectures and as- munities. signments,reviewers of researchproposals and manu- At the AustralianNational Universitythe Nadel pro- scripts-but I thinkthat the highlandsmay have more gram,to be honoredposthumously, was to focus on the in common than some otherpurported regions in Mel- New Guinea highlandsbecause of the concentrationof anesia.3 Highlands intensificationof agriculturehas populationand the excitementof studying people whose combined sweet potato subsistence with pig raisingto areas, social systems,and response to contact could be make massive feastspossible in the area fromChimbu traced while Australia broughtthem into the modem to Enga and again in some sections of West Irian. This world.The ANU was then competing(if I may let it out) pig-feastregion4 is surelynot a continuousarea and not with anthropologicalresearch programs at Sydneyand commensuratewith the highlands,as the Easternhigh- Washington(and Mick Read soon went there);research lands are mostly left out.5 Anthropologistshave made and writinggrew quickly. There were also some geolo- some progressin definingSepik and Massim cultural gists,geographers, mission anthropologists,and others regions,which, like the highlands,seem to be conve- studyingthe region.I got my chance forChimbu field- nient categoriesfor teaching units, symposia,and essay work on social organization and social and political collections.Is that a good enough reason? change a year or two later; at that time Harold Brook- Hays says,"Many anthropologistswill perhapsbe nei- field and I joined to study agriculture,land use, and thersurprised nor troubledby the resultsof my review." economy. He is right.As researchin these interstitialareas (that The I964 papers edited by JimWatson were a major is, between the "seaboard" and the highlands)has been step in establishingthe region,and then definingpapers published,we have learned two importantthings about by Meggitt,Brookfield, and all the restdrove the region- relationsbetween the highlandspeoples and theirneigh- alism ahead. The "fringe"(Mountain Papuans, Ok, and bors.The firstis the importanceof tradeand the move- Anga) was hardlyknown. Irian Jayawas a key area of ment of materialsand goods (Hughes I977); agriculture, studyin the I950S and I960s under the Dutch; afterit pigs, and sweet potatoes were surelycrucial in making became partof Indonesia, research permits were difficult highlandsculture. The second is the extentand impor- or impossible to obtain, and Anglophoneanthropologi- tance of intergrouprelationships-exchange relations cal researchpersevered in Papua New Guinea. The lin- which cross regional and linguisticboundaries, the de- guistic studies of Wurm began at the same time as sirefor and acceptance ofnew practices,including cults, Brookfield'sand mine (we shared a Jeeppurchased by and payment to outsiders for the privilegeof holding the ANU in i958).1 certainceremonies or rituals.Through these studieswe "Core" or "center" was variouslydefined and mostly may downplay the artificialcategories erected in the confinedto Australian New Guinea, which differenti- past generationand gain an understandingof intergroup ated highlandsfrom "fringe." As late as the mid-I970s connections. I could findlittle ethnographic information2 about them, I think,however, that Hays and I may want to draw forthese areas were accessible to researchersonly after differentconclusions. He is dissatisfiedwith fuzzysets, the Australian administrationhad established a patrol while I would now ask: what purpose would be served post. There are still few road connections to much of by clear ones? If we attemptto createexclusive regional the area. categoriesand culturetrait lists, we falsifyall we know If,then, we had inventeda categoryor culturalregion, of culturalinfluences and change,relations with neigh- what could have been the rules forinclusion and exclu- bors,and interculturalinteractions. sion? None of the volumes of collected regionalessays, even those with "New Guinea Highlands" in the title, were restrictedto what is now recognizedas the central 3. Whenwe held the symposium"Man and Womanin the New Highlands.I can thinkof over a dozen such collections, Guinea Highlands"at the AmericanAnthropological Association on topics as wide-rangingas politics, religion,kinship, meetingin I974, therewere severalpapers which were not in- cludedin the laterpublication, and Kelly'spaper was addedto it (Brownand Buchbinder I976). Perhapswe shouldhave changed the title.At thetime of the symposium, Rhoda Metraux protested that i. This workestablished a languagephylum which included some thehighlands seemed overrepresented in symposia, with the Sepik peoplesin Irian Jaya,and my comparativediscussions often in- leftout; I suggested,and she organized,a Sepik symposium(Me- cludedthese. traux I978). 2. A manuscripteditor hired by Cambridge University Press at first 4. Anotherway to look at it is Feil's (I987), whose continuum attemptedto integratemy discussion of "fringe" (I978:28-39) with peaksin theEnga Te pig exchange. thegeneral text. I wonderwhat would have been the meaningof 5. If largefeasts were held in the easternarea beforeintensive "Highlands"if I hadnot objected to thisreorganization and insisted contact,they have not persisted.I wonderif the influenceof the thatmy chapter be reinstated. Seventh-DayAdventist mission can have been decisive. I50 I CURRENT ANTHROPOLOGY Volume 34, Number 2, April 1993

SIMON HARRISON about the frameworkfor comparison,mostly touched Departmentof Sociology, Universityof Ulster, upon only in passing, since the focus of most publi- Coleraine, Co. LondonderryBT52 iSA, Northern cations is on something else, have been widely ne- Ireland. 2 XI 92 glected-except that those who study material expres- sions of culture comparativelyhave never been able to Hays does Melanesian studies a service by examining avoid them. the inconsistentand contradictoryways in which Mel- Tiesler (I990) offersan admirableattempt at system- anesianists have used the term "New Guinea High- atic classificationof New Guinea art and an excellent lands." His paper should promptthem to think more summaryof how differentscholars (since Haddon I984) carefullyabout the terms they use. He argues persua- have approachedthis problem.Some of these anthropol- sivelythat "the Highlands" is a fuzzyor polytheticcate- ogistshave taken "style" as a startingpoint (e.g.,Speiser gory,but I am a littlebothered by his apparentlyassum- I937, GerbrandsI951), and Speiserhas relatedstyles to ing that such categoriesare Bad Things and need to be historical classifications.Another approach links con- expunged from anthropologicaldiscourse. Surely, the siderationsof style primarily to geography(Biihler I 96 1). point that cognitive psychologistssuch as Rosch are Tiesler points out that since Buihlerthere has been no makingis thatvirtually all human thinkingis done (and attemptto develop theoreticaland methodologicalcon- done verysuccessfully) in termsof prototypes and fuzzy cepts forclassifying art in New Guinea. In his introduc- sets. All concepts in anthropologywould probablyturn tion he formulatesthe problemon a generallevel: "Die out upon analysis to be just as fuzzy.In short,while I Erarbeitungvon Gliederungsprinzipienfuir die Vielfalt agree that we need to think more carefullyabout the derErscheinungen" (I 990:23 5). I thinkthis is thecrucial categorieswe use, I do not accept that we should tryto questionfor those who do not want to limit theirefforts stop thinkingin categoriesaltogether as Hays seems to to one specific culture. How can we work out princi- suggest.But of course we must suspend judgementon ples ofclassification for the diversityof cultural phenom- his approachuntil he demonstratesits superiority. ena? Hays's objection to the use of "Highlands" as a con- If we acknowledgethat this question is legitimate,a structis that it has led ethnographersto misrepresent largerange of possibilities arises. Hays's suggestionthat thesesocieties as homogeneous,ahistorical, and isolated anthropologistsshift the focus "fromwhat people are to fromthe outside world. Althoughnot a Highland spe- what people do" is simply one possibilityof many for cialist myself,this seems to me unfair.I doubt that structuringthe continuum of cultural phenomena the manyHighland ethnographerswill agree that tradenet- betterto understandcertain aspects of it. It is obvious workshave up till now "escaped our attention"or that that any classificationinvolves the constructionof a few have consideredwhether big men and large-scale gridto distinguishbetween aspects to be comparedand ceremonial exchange "were affectedif not created by others. Therefore,our attention should shiftfirst and the colonial process." A non-Melanesianistreading this foremostto methodologicalquestions: why we classify, article could gain the impression that New Guinea what criteriaand methods we use to attain this goal, Highland ethnographyhad been stuck theoreticallyin and, finally,whether the systemof classificationchosen the I940S and I950S until the publicationof this paper. is consistentlyapplied. "Commonsense abstractions" Hays seems to some extentto have "essentialised" the such as "the Highlands,"their "core," and their"fringes" Highland ethnographers,imputing to them that very are classifications.If more attentionhad been paid to homogeneity,changelessness, and isolation which he methodology,the comparisonof cultureswould perhaps accuses them of attributingto the Highland societies have produced classificationsin "the Highlands" that themselves. made more sense. In shiftingto "what people do" Hays identifieshis perspectivefor structuring complex reality,but this ap- BRIGITTA HAUSER-SCHAUBLIN proach is not new forNew Guinea. Again, Tiesler pub- Instituteof Ethnology,University of Gottingen, lished (in German and thereforeprobably unintelligi- Theaterplatz15, D-3400 G6ttingen,Germany. 2 XI 92 ble to most anthropologistsworking on New Guinea) a large-scaleregional analysis of tradingand exchange The more I read Hays's paper the less I could refrain networks along the north coast of New Guinea in fromsmiling. His brilliantanalysis of how anthropolo- I969-70. It answersthe question what social processes gists have dealt with "the New Guinea Highlands" lead to the developmentof a culturearea and how one amuses me, but the questions he raises are sobering. can be geographicallydefined on the basis of complex The expression"culture area" is simplyout offashion intertribalrelationships depending on the goods traded forboth modernistand postmodernistanthropologists, and exchanged.Apart fromthis kind of analysis there even thoughmost of them obviouslyhave a similarno- are many otherpossible ways of identifyingsimilarities tionin mind in speakingof a "region."At least, it seems and differencesbetween cultures.One of the most com- that none of them has a betterconcept to offer. monly used is "languages" and their interrelations; Detailed and systematiccomparative studies on New these classifications,often taken as mirroringreality, Guinea have become fewer and fewer in the past 2o are constructed,too, mainly on principlesof lexicosta- years. Thus, methodologicaland theoreticalquestions tistics.A furtherway would be to ask people how they HAYS "Th e New Guinea Highlands " 5Ii I

classifythemselves and others.And even if we go back warfare,ceremonial exchange,and political leadership, to the notion of cultureareas and traits,it is no longer which are moreobviously behaviors? The Awa practiced in order to make cumulative lists of identical mate- warfare(Hayano I974), as did the Hageners,Simbu, and rial objects throughoutan area. Asking questions about others,but there are vast qualitative and quantitative similarities,differences, transformations, and "breaks" differencesamongst them. A simple "absence of/pres- (Briiche),as I have tried to do with ceremonialhouses ence of" codingof behaviorsto discernwhich activities in northem New Guinea (i989), reveals new insights mightbe characteristicof Highlands,fringe areas, and that cannot be achieved otherwise. Lowlands as fuzzysets will producesimilarities that are Therefore,I wonderwho mightbe able to decide what superficialand mask differencesthat are critical. kindof classification is the rightone withouttaking into In a sense Hays's paper is an anachronism.His issues considerationthe context and the purpose for which should have been raised 30 years ago, or earlier,soon such studies are made. after the Highlands first opened to anthropologists. Granted,the data were not complete then,and looking at these newly discoveredpeoples as a continuumfrom DAVID M. HAYANO the coast was apparentlynot as importantas portraying Departmentof Anthropology,California State them as remote and untouched. That was the bias of University,Northridge, Calif. 91330, U.S.A. 27 x 92 anthropology.Part of this bias also shows in Hays's pa- per in the discussion of social processes such as trade. My understandingis that "the New Guinea Highlands" Tryingto recapturean idealized model of Highlandsso- (and, beforethat, "the Central Highlands" or "Central cieties at some generalized,undefined (but apparently North New Guinea" [Nelson I98 2:1I2I]) was originally precolonial)point in time would seem to overlookmore a set of colonial administrativeand political boundaries than a half-centuryof historyand change. and eventuallycame to serve the needs of government I would include on a more contemporarylist of social officials,gold miners,census takers,various exploratory linkages the irreversiblechanges wroughtby the colo- patrols, and anthropologists.Hays argues that since nial and postindependencegovernments. How about those days it has changedfrom words on an administra- groups that grow coffee,drink alcohol, gamble with tive map to a murkyanthropological concept. He ably cards, work on plantations,join the army, go to the demonstrateshow confusingthat concept is, and I am university,travel as tourists, drive Toyota pickups, leftwondering whether the Awa in the Eastern High- watch videos? Here the social linkages (modern"trade lands Province,whom I studied, are Highlands,fringe, routes"?) between village, township, and city can be or Lowlands people. Their gardensand some houses are seen as a complex patternof behaviorsdynamically cre- scatteredover i,ooo m in altitude; some of theirbehav- ated in a much largernational or world system.If we iors resemble those of other Highlandersand some do focus solely on traditionand timelessness and ignore not. Perhapsthey are all three-or none at all. culture history and the contemporary,our academic Hays raises but does not addressthe additionalepiste- view of Papua New Guinea will undoubtedlyremain mological problemof how concepts in human language fuzzy. can everadequately represent "reality." He suggeststhat We seem to have come full circle: the New Guinea "the New Guinea Highlands" should more accurately Highlandsis a convenientlocational termfor designat- be considered,because of its physical and culturalhet- ing known political/administrativeareas but severely erogeneity,as a fuzzyset. I am not thoroughlyconvinced flawedas a usefulanalytical concept. I applaud Hays for that this is not some kind of semantic sleight-of-hand raisingthis issue, but he has not resolvedit. and thatthe olderconcepts of "region"or "culturearea" did not allow forintra-areal heterogeneity and flexible borders.But a furtherproblem arises when one tries to ERIC HIRSCH elucidate what specifictraits, characteristics, or behav- Departmentof Human Sciences, Brunel University, iors are associated with one fuzzy set and not another. Uxbridge,Middlesex UB8 3PH, England. 9 XI 92 Conceptually,"the New Guinea Highlands" and "the fringe,"for example, are not equivalent to "red" and Hays suggeststhat the comparativeethnography of the "off-red."Cultural characteristicsmay change,interact New Guinea Highlands has failed to address the issue with one another,appear and disappear in ways that ofhow the local inhabitantscreate, maintain, and ignore color categoriesdo not. I do not findthe idea of a fuzzy boundariesand thatto do so we need "new ways offram- set more explanatorythan the older terms. ing our questions." Given this conclusion,it is surpris- Hays emphasizes thatanthropologists should concen- ing that he does not draw on Marilyn Strathern'sThe tratemore on processthan on morphology,on what peo- Gender of the Gift(I988). Not only is it a comparative ple do ratherthan on what they are. Assuming,then, and syntheticaccount of Melanesian ethnography(with that we proceed with the notion of a fuzzy set, what its specific emphasis derived froma "highlands"/Mt. specifictraits should be foundin the list?Are we to omit Hagen point of view) but it is directlyrelevant to the factorssuch as altitude,population size and density,and main theme of Hays's paper. In fact,his closing words abstractionssuch as patrilinealitybecause people don't could in many respectsbe read as the startingpoint of do these things?(Or do they?)Should we look instead at Strathern'sbook. I52 CURRENT ANTHROPOLOGY Volume 34, Number 2, April I993

Most ofHays's conclusionswith regard to the compar- We cannot account forthe emergenceand operation ative accounts of the "Highlands" have been rehearsed of the boundaries/differencesadduced both by anthro- in the literaturehe cites, but he has done us a service pologistsand by the inhabitantsof Melanesia until we in bringingthem togethersystematically. To summa- have grappledwith the nature of these shared conven- rise: (i) It is difficultto reach agreementas to what tions. We must also be aware of how our Westerncon- constitutes"the Highlands" when analysis is based on ventions impinge on what we come to see as a "prob- single and/ormultiple factors (e.g., environment,de- lem" in the firstplace (StrathernI988:3I8-23). Harris mography,economics, etc.). (2) Fromone vantagepoint (I986) has arguedthat repeatedattempts to account for societies may appear similar, but on closer inspection the originof writingover the centurieshave failed for this seeminghomogeneity dissolves, leading to the iden- the simple reason that no one has providedan adequate tificationof "core" and "marginal" areas. (3) Most an- answer to a preliminary(but largelytaken-for-granted) thropologistsnevertheless acknowledge the continuing question:what is writing?Similarly, we cannottrace the usefulnessof the concept. (4) Its vagueness or fuzziness historical consequences of the wide-ranging"interac- is part of a more general cognitivecondition shared by tions" highlightedby Hays by assumingthat Melanesian Highlanderand anthropologistalike-in other words, socialityis a versionof Westernsociality; we must first unavoidable. What can be avoided, however,is defini- establishthe conventionsimplicit in these interactions. tion of "the Highlands" in essentialistterms, which has Hays is arguingthat instead of looking at Melanesian focusedattention on issues ofmorphology to the relative societies or regions we should examine the links be- neglectof inter-and intraregionalprocess. tweenthem; Strathernis suggestingthat we call "societ- An altemativestrategy proposed by Hays and recently ies" and "regions"in this culturalcontext have a shared acknowledgedby others(cf. Gell i992) is to give greater feature:the linkages at once produce and are a product attention to linkages between areas formed through of this commonality. trade.One reason forthe relativelack of scholarlyprog- ress in this area is the conditionsgenerated by the ad- ventof a colonial and mission presencethroughout Mel- DAN JORGENSEN anesia: tradebecame less apparentbecause of the influx Departmentof Anthropology,University of Western of both European manufacturedgoods and traditional Ontario,London, Ont., Canada N6A 5C2. 6 XI 92 goodsimported by Europeans. The energiesand interests of local inhabitantswere focused elsewhere.As urban Hays tackles the imprecisionof the "Highlands" cate- centres,cash-cropping, and the circulationof money be- goryin Papua New Guinea ethnographyand criticizes came established, for example, a transformationoc- recent comparative work centred on this. Exposing curredin the objects deployedin ritual and ceremonial weaknesses in our customaryways of lumpingcultures exchange (cf.Hirsch I990). Nevertheless,the existence together,he shows that anthropologistssometimes dis- of wide-rangingtrade networks has been establishedfor agreeabout which culturescount as Highlandscultures particularvaluables by Healey (I980) and Hughes (I977), and that we are oftenunclear about the contentsof the and Hays (I990) has traced a link between the Huli of package with the Highlandslabel. Worse,the Highlands the SouthernHighlands and the populationsof the south designationmay blind us to linkages across zones, ob- coast. This is evidence for long-standinglinkages be- scuringwider processes-trade, for example. Reifying tween areas oftenthought of as separateand relatively the Highlandscategory may also producea sortof "sec- autonomous. ondaryOrientalism" by a reflexlumping of remaining Hays asks rhetoricallywhy the result has not been Papua New Guinea cultures as "the Lowlands." For homogeneityor uniformity,but I contend that these these reasons I think Hays has done us a service,and wide-ranginglinkages have in fact produced a kind of much of what he says fitswith contemporaryefforts to homogeneityand uniformity.This is where I see Mari- dissolve old anthropologicalcategories. But despitefears lyn Strathern'saccount as being of centralimportance. ofmaking essentialist mistakes I thinkwe can continue On the basis of a comparativeanalysis of ethnographic to finduses forthe Highlands categoryand othersof its material fromdiverse "regions" of Melanesia, she has kind. suggestedthat Melanesian societies share an aesthetic Hays talks about the troublewe have decidingwhere (I988:34I). She warns against assuming that this is the the Highlandsbegin and end,but maybe this isn't as bad residue of a common past (p. 342). Rather,it reflects as he thinks.For one thing,it's not clear that we need these societies' being outgrowthsand developmentsof to thinkof the Highlandsas beingsharply circumscribed one another-implicated in one another'shistory. "Spe- afterthe fashion of provincesor states-we've known cific formscome not fromgeneralized ones but from about dialect chains, and so on, fora long time,and the otherspecific forms" (cf. Kulick i992 forlinguistic par- fuzzinessof the Highlandsset should not in itselfthrow allels). Thus the wide-rangingtrade networks described us offbalance. More interesting,however, the fuzziness by Hays at once affirmthese shared conventionsand Hays discusses is not uniform,and I suspect that most sustain them fromdiverse perspectives. To phrase this of us would perceive some of the edges to be relatively in terms closer to Strathern's,they highlightdistinc- sharp(e.g., the boundarybetween eastern Highlands and tions and obscure the common set of conventionsthat the Anga peoples). I think we should be curious about underliesthem (cf.O'Hanlon i992). such things,and differentiatingbetween sharp edges and HAYS "The New Guinea Highlands" I535

hazy zones is encouragedby attemptsto describe cul- riods, certain areas become "hot" for studyingcurrent tural areas. topical and theoreticalissues. A given topical emphasis I also thinkwe can learn a lot by asking more ques- can easily be reifiedand magnifiedby ethnographichis- tions about where the notion of the Highlands comes tory,itself coming to define the region. The firstma- fromand what kind of life it leads today. For example, jor monographsbecome "classic" and imprintthe per- a look at Papua New Guinea's historyshows that the ception of later students and comparativeresearchers. extensionof colonial controlinto the Highlands forced Thus,in the I950S to I970S, "the New Guinea High- a massive reorientationof colonial policy (see Downs lands" became a hothouse for studies of big-man I980 fordetails). This was not merelya matterof geogra- politics, competitive gift exchange, pre-statewarfare, phy: the densityand scale of the Highlandspopulations loosely structuredclanship, pre-state agricultural inten- quicklyled to administrativearrangements marking the sification,and sexual antagonism.These featureswere Highlandsoff as a regiondistinct from the restof Papua in various guises present there, but their reification and New Guinea, a historymirrored in today's provin- as anthropologicalcategories tended to configure"the cial boundaries.Not all mountain populations-for ex- Highlands" as a static precolonial ethnographicregion ample,the peoples west ofthe Stricklandor in the Owen definedin significantpart by oppositionto otherareas Stanleys to the east-elicited this kind of treatment. of Melanesia in which its "typical" traitswere assumed This suggeststo me thathistorical contingencies played to be attenuatedor absent.Thus "the Highlands"is con- a role in shapingour notions of the Highlands but also figureddifferently to contrastparticular patterns of lead- that the existence of the Highlands as a region is not ership,gender, exchange, or subsistence with those in simplythe productof anthropologicalpigeonholing. the Sepik,lowland south New Guinea, the Massim, and I can thinkof at least two otheruses forthe Highlands so on (e.g.,Lawrence and MeggittI965; Herdtand Poole label, and followingthese lines up mightbe more diffi- i982; Lindenbaum I984; Whitehead I986; Feil I987: cult if we decided that there was no wheat lurking chap. 7). As the Annales historian Marc Bloch recog- among the chaff.The firstuse is historiographic:any nized (I97I[I9I3:I2-2), the verynotion of a regionde- account of anthropologicalwork in Papua New Guinea pends on the theoretical problems one is concerned that failed to recognize a "Highlands period" would with. Without denyingthe importanceof comparative surely be missing somethingbig. Postwar Melanesian studies, these interregionalcontrasts are increasingly ethnographymoved to centre stage when the New being questioned and put in historicalperspective (e.g., Guinea Highlands were found to be located outside of Godelierand Strathern1990; M. Strathern1990; A. Africa(Barnes i962). The second use turnson the sug- StrathernI990; KnauftI993). gestion that we make a suitably indigenous formour Hays's pointthat fuzzy-set regions call formore rather focus by turningour attentionto the Pidgin termhai- than less specificationof concrete ethnographiccon- lans. Hailans is a Papua New Guinea folkcategory that toursdeserves emphasis. It points to a creativetension has acquired a life of its own quite independentof what in his paper between the use of principledcriteria to anthropologistshave to say; looking to its regionaland identifyethnographic regions and the tracing of net- (novel,fuzzy) ethnic import may serveas a usefuldiver- works and processes that crosscut them. Exactly how sion fromthe ethnographicallyparochial preoccupation these competingviews should be balanced is a key and with My Village (vs. Yours). unansweredquestion. Hays appears to advocate the lat- In the end, I would be reluctantto do without the ter,but too much emphasis on these networkscarries Highlandslabel because I thinkit can tell us something the dangerof obliteratingwhat is distinctiveto particu- real about the cultures we are looking at-cultures lar ethnographicregions. which, afterall, are not all entirelyunique or equally One approachmight be neitherto assertdogmatic re- differentfrom one another. gional boundariesnor to ignore them but ratherto be clear why a given geographicregion is appropriateas a unit of analysis. Not all the interestingranges of varia- BRUCE M. KNAUFT tion can be supposed to line up within a certainregion. Departmentof Anthropology,Emory University, This reflexivemove can at the same time allow us to be Atlanta, Ga. 30322, U.S.A. 27 x 92 morerather than less empiricallyspecific in our compar- isonsand contrasts(Knauft I993). Hays's premise that the New Guinea highlands is a Much currentdisagreement in the assertion of re- fuzzy set is compellinglydocumented and beyond dis- gional characteristicsand definitionscomes fromcon- pute. Below I push fartherthe implicationsof his preg- flict over scales and purposes of analysis: large-scale nant analysis and exploreits inherenttensions. generalitiesand characterizationsseem inadequate to "The New Guinea Highlands" is obviously a proto- characterizea regionwhen its rangeof internal variation type that has been configureddifferently to fitvarious is more closely consideredand the scale of analysis is researchers'analytic agendas. The question Hays raises reduced.Further, it would be foolish to be constrained by implicationis important:what largerhistorical and by regional contours appropriateto previous ethno- theoreticalbiases have informedit? As Appadurai(I986, graphic interestsor time periods, because indigenous I988) has suggested,the pairingof name and place labels networksand regionsthemselves proliferate and change. in anthropologyitself has a history.At particularpe- That regionslike the Highlandsare fuzzymeans neither I54 1 CURRENT ANTHROPOLOGY Volume 34, Number 2, April I993

that all ethnographyis relative nor that inalterable of,shall we say, makinga difference.Narrowing Hays's boundariesmust be imposed. Arguably,it is important prescription,I would suggestthat we centerour atten- to maintaina creativetension between the heuristicde- tion on indigenous comparative discourses,a topic al- lineation of regions and the analysis of networksand ready elaborated in studies, for example, of "gender," processesthat crosscutthem. "exchange," and "myth" (Lederman iggoa, b, I99I, Hays's paperis effectivein raisingissues and possibili- i992). When we do so, we learn thatthe "same"/"differ- ties; it is now up to us to develop and concretelyuse ent" relationis organizedand deployed"differently" in them. local practice and in anthropologicalanalysis. Under- standingthis means reworkingour whole comparativist game: if our "medium" ends up obviatingour (their?) RENA LEDERMAN "message" ("medium"?), at the least we ought to have Departmentof Anthropology,Princeton University, intended that effect!Failing to understandit, we run Princeton,N.j. o8544, U.S.A. I3 XI 92 the risk of "essentializing"social processesjust as Hays suggests we have "essentialized" regions. Social pro- As Hays argues,the use of regional categoriessuch as cesses will also be misrepresentedas homogeneous, "the New Guinea Highlands" in anthropologicalworks ahistorical,and systemicso long as we insiston present- implies (and too oftenasserts) that the peoples referred ing theirmeanings as unitary,adjudicated, and authori- to (i) are culturallyhomogeneous (that is, theircommu- tative. Having downed the disciplinarydragon of Re- nities are "similar"), (2) share (in some sense) histori- gionalism, we may find ourselves set upon by the cally stable,autochthonous traits, and (3) can be under- social-theoreticsnake of a certainkind of Objectivism stood as isolates (that is, studied as if their "linkages" (Bourdieu's"practice" being of no greathelp here). with outsiderswere irrelevant).I applaud Hays forhis "Trade" is a good case in point (Ledermann.d.). As persuasivedemonstration that we have no basis forcon- long as we take it upon ourselves to determinethat an sidering"the New Guinea Highlands" an ethnographic interactionis "trade" and then go on to compareit with "region"in these senses, but I urgehim to take his cri- "similar" acts elsewhere,it will end up beingas incoher- tique farther.Indeed, heterogeneity and shifting,bound- ent an analytical categoryas "the New Guinea High- ary-engendering"linkages" are to be expected given lands." But we already"know" better(even if we don't what we know about Melanesian symbolic and social always recognizethe knowledgeas such). A readingof inventiveness(e.g., Wagner I972, 198I, 199I; Strathern any number of monographs (e.g., Malinowski I1922, I988). What is more,we should expectcultural transfor- GewertzI983, GodelierI986, Healeyi990) revealsthat mations at all orders of regional magnitude (as Hays one person's "trade" may be his partner's "gift ex- hints in his remarksabout the difficultyof identifying change"; that one person's asymmetricalexchange may coherentcultural "cores" such as "Enga"). Melanesia be the other's reciprocity;that one's party's external may be even more, and more interestingly,heteroge- boundarymay be the other's internalrelation. Mutual neous than he thinks. mistranslationsand reinterpretationsof one another's Hays is rightto reinforcethe point (also made by Ap- social formsis mundane. As a step towardsunderstand- padurai I986, Fardon I990, and others)that our disci- ing how Melanesian social process maintains "differ- pline encouragesus to reproduceanalytically loaded re- ence" (rather than producing hegemonic or encom- gional categories.He believes that we would be better passing structuresof relationshipor larger "regions"), servedby a shiftfrom "morphology" to "process." This we need to acknowledge the typicalityof decentered, is an attractivesuggestion on its face, but we need to asystemictransformations in our accounts in the very recognizethat the same problems Hays has identified ways we juxtapose local constructionsof events and in- in our treatmentof "morphology"also exist with re- teractions. spect to "process" (and, conversely,that there are better The question,then, is not why this or that region"is approachesto "morphology"than those Hays criticizes). not larger"but how any "region" is made to appear in Our sharpest(but dangerouslyrecursive) challenge is to the firstplace. The same goes for"processes." In answer- expose and then rework the relationshipbetween our ing these questions, we must take care to distinguish own comparativediscourses (whetherabout "morphol- our own discursiveinterests and those of Melanesians. ogy" or "process") and indigenoussocial practice. Then we mightadvance Hays's criticalcontribution by Hays has highhopes for"process." In studies of social exploringthe potential of differentdisciplinary writing "linkages" regional distinctions like Highlands/Low- styles to "translate" these interestswith the subtlety lands become irrelevant,and our attentionshifts from theydemand. familyresemblances and the like and to "difference" and "boundaries." Hays reaffirmsWelsch, Terrell,and Nadolski's (I992) importantpoint that social interaction EDWARD LI PUMA may deliberatelyreproduce difference in New Guinea, Departmentof Anthropology,University of Miami, and he asserts that in orderto understandwhy regions Coral Gables, Fla. 33124, U.S.A. 2 xi 92 (zones of similarity)are so small we need to understand why and how people create and maintain boundaries. Hays makes the importantand persuasive argument I agree thoroughlywith the importancein Melanesia that no criterialfeature or set of featureswill allow us HAYS "The New Guinea Highlands" I 155

to delineateunambiguously the internalcharacteristics First,focusing on what people do, while certainlycen- or limits of "the New Guinea Highlands." Though ana- tral, can never explain the organizationof trade, the lysts oftensettle on a varietyof featuresranging from formof ceremonial exchange, or other structuralrela- elevation to ceremonial exchange and pig husbandry, tions.When agentsmake an exchangeor negotiatetheir these featuresare neitherdiagnostic nor predictive.The identitythey do so froma determinateposition within crucial implicationis that those who believe that "the an objectivesocial field (e.g., the definitionand history New Guinea Highlands" is anythingmore than an ana- of a clan and the dispositionsinculcated in clan mem- lyticalconstruct are confusingthe logic of science with bers).Second, focusing on action cannotexplain the con- the logic of social formations. structionand politics of identity,how peoples come to Hays's paper calls attentionto a submergedbut long- identifyand representthemselves as Simbu, Kuma, or standingproblem in the anthropologyof New Guinea Enga, and how terms like "Highlander" have entered societies: though the characterof its analytical object into the political discourseand the makingof the Papua oftendeprives it of a sense of limits,theory and method New Guinea nation-state.It also cannot explain why a presumethat real limits exist and that anthropologists certainpolitics of identity(e.g., a vitae which specifies can craftnotions of regionaland culturalclosure. Oppo- "the New Guinea Highlands" as my area of expertise)is sitionssuch as Highlandsversus Lowlands, fringe versus vital to anthropologicaldiscourse, recognition, and posi- core,and termssuch as "externaland intercultural"are tion takingwithin the field. the epistemologicalinstrumentation for attaining such Analystswho focuson what people are and those who closure. However, this viewpoint not only takes too focuson what people do have thisin common: theyboth much forgranted but masks the processof construction graspsocial practice in terms of what is directlygiven of cultural/ethnicidentity. The Maring(usually cited as to ordinaryexperience. But what if "the New Guinea a fringegroup) offer a criticalexample. Collective identi- Highlands" were more akin to a solar systemin which ties such as Maring (an imposed ratherthan indigenous the orbit of any one culture is definedby the gravita- name) as opposed to specificclan and clan-clusteridenti- tional pull and push of all the others?What ifthe reality ties did not exist prior to the incorporationof Papua of the Highlandswere nothingless than a set of cultural New Guinea "within" a nation-state(such as Australia). spaces definedby theirinteraction? On this view, "the Further,the limits or boundariesfor a given cultureand New Guinea Highlands" would not be reducibleeither language are characteristicallygraded. The clan cluster to objective criteriaor to what people say about them- at Kandambiampis comprisedof both Maring and Ka- selves (e.g.,defining themselves as Highlanders),and the lam (culturallysimilar, linguistically distant) peoples. In aim of analysis would be to constructthe ethnographic this regard,Maring culture flows into and overlapswith space that would allow us to grasp the wide range of Kalam culture(LiPuma i988), and the same may be said variationobserved throughout Papua New Guinea. In a forMaring and Manga culture at the other end of the relationalanalysis such as this,there would be no such valley. thingas a peripheral,marginal, or fringesociety; there There are several ways for anthropologiststo deal would only be societies that more or less share degrees withthe issue ofclosure. If closure is the objective,Hays of sameness and difference.Though the notion of fuzzy maintains, we need explicit principled criteria for sets is an improvementon conventionalviews, I would boundingthese categories,and we need to apply them arguethat the tropeof "sets" itselfshould be abandoned. in a systematicway so that both inclusions and exclu- Pushed to its limits, an anthropologyof Papua New sions are adequately motivated and consistentlyexe- Guinea does not need to focus on what people do as cuted. Such a search forlimits stems fromthe a priori opposed to what theyare; it needs to dissolve that very assumptionthat structuraland functionallinkages be- distinction.The "people" in "what people are and do" tween "highland"and "lowland" societies,between one includes anthropologists. cultureand another,and betweenthe centraland periph- eral groups that comprise a given culture are external ratherthan constitutiverelations. This is problematic EUGENE OGAN preciselybecause what we need to determineis how the Departmentof Anthropology,University of relationsbetween groups, cultures, and regionsgenerate Minnesota,Minneapolis, Minn. 55455, U.S.A. 6 xi 92 structuresthat permit socially objective categories to emerge. Hays's thoughtful,scholarly critique of some of the an- Hays concludes that we need new ways of framing thropologicalwriting about Papua New Guinea could the issues. He asks what would happen if we rejected not come at a more appropriatetime. Pacificanthropol- essentialistapproaches to "the New Guinea Highlands" ogyis hardlyimmune to currentattacks on essentialist and stoppedsearching for diagnostic characteristics. He portrayalsof other societies or large sections of the suggeststhat we shiftfrom social morphologyto pro- world.Indeed, concepts once consideredas basic as that cess, changingour analytical focus from"what people of "Melanesia" seem now to obscurerather than illumi- are to what people do (e.g., trade,engage in ceremonial nate our studies (e.g., Green i99i). What makes Hays's exchange . . .)." But this reformulation(derived from contributionvaluable is thathe does not simplycriticize Barth)does not,I would submit,go farenough or answer but also points to ways in which our researchmight be the criticalquestions that surroundthe issue of limits. betterfocussed. I56 CURRENT ANTHROPOLOGY Volume 34, Number 2, April 1993

I am not normallysympathetic to importingterms ANDREW STRATHERN from other disciplines when anthropologyis already Departmentof Anthropology,University of burstingwith jargon, but "fuzzy set" seems benign Pittsburgh,Pittsburgh, Pa. 15260, U.S.A. 27 x 92 enough, especially since Hays makes clear that this is a general human "information-processingstrategy." Hays's insertionof detailed definitionalchecking and Ratherthan attemptto answer the question in the arti- cognitiveset theoryinto the discussion of the "High- cle's subtitle,however, I would preferto underscoreand lands societies" is welcome and timely,functioning as amplifyslightly in a comparative context the points a device for clearing the groundfor another approach, made about historyand process in Pacific ethnography. the studyof process ratherthan morphology.I add here A failure to take adequate account of historical threecomments. processes in ethnographyis hardly confined to New The creation and acceptance of "the Highlands" as Guinea, althoughstudies of the area providemore exam- a region may owe much more to administrative(and ples than Hays has time to recount.(A particularlytell- subsequentlyinterethnic) usages than we have tended ing one appearsin Godelier's filmabout the Baruya;my to notice. "The Highlands" was originallyassociated undergraduatestudents are always amused by the an- with the patrols of J.L. Taylor and the Leahy brothers thropologist'sembarrassed admission that only belat- and its extension marked in a sense by the limits of edly did he learn that the large pig herds he counted those patrols; tacked on were the patrolsof JackHides were a recentphenomenon, occurring only afterthe in- into what later became known as the SouthernHigh- troductionof steel tools.) The argumentof a recentvol- lands. Later, in postwar years,the area was definedby ume edited by James Carrier (I992; see also Keesing the creation of a Highlands District, which was then i990:i58) is preciselyto highlightthe inadequacy of de- progressivelysubdivided. The ethnographerswho were scriptionsof life in New Guinea and the SouthwestPa- firstpermitted to work in this area themselvesby acci- cific generallythat do not perceive and analyze those dent providedthe "cultural core" for the area or "re- culturalfeatures which certainlyreflect the incursions gion" by workingin a numberof places identifiablewith of a largerworld system. major languages in which recognizable "prototypical" However,as Hays makes clear (e.g.,in his citationof similarities among institutions could be found (viz., Hughes I978), innovation and change did not wait for Gahuku-Gama, Enga, Melpa, Kuma, Chimbu, Mendi, the arrivalof Westernersto shape the culturesof Mel- Huli). The categoryof "region" or pseudo-regionwas anesia (if one may still use that shorthandterm). The built up by cumulative practice,then, ratherthan by isolated Stone Age tribeignorant of any otherhumans any logical criteriathat could produce a "hard" rather may still have a place in supermarkettabloids, but it than a fuzzyset. Given this,it was obviouslyhigh time has no place in ethnography.Inasmuch as moderneth- to deconstructthe "region" in analytical terms while nographybegins with an account of interisland ex- recognizingthat it has today an administrative,politi- change in New Guinea waters,it is embarrassingthat cal, and interethnicsignificance that has a lifeof its own today'spractitioners still need to be reminded"to view apart fromour academic concerns. Hays runs a retro- the tribalworld as comprisingregional systems" (Kees- spectiveanalytical eye over contemporaryacademic us- ing I990:I53). Components and links within such sys- age and finds it wanting; the explanation of how this tems are an obvious place to startinvestigation. fuzzyusage has come about could be given onlyby trac- Perhaps more pressingthan any problemof drawing ing the developmentof the discursivepractices which boundariesbelieved to reflect"the real world" in New have formedusages historicallyand outside of the aca- Guinea or elsewhereis that of drawingthe most useful demic domain. That problemswould eventuallyemerge boundariesaround our descriptions.Even if we follow is shown even at the heart of the originalprototypical Hays's sound advice to approach certainissues froma enterpriseas exemplifiedin JohnBarnes's (i962) query "village-outward"perspective, it is not clear,at least to concerningAfrican models: we can see fromthis already me, what shape(s) our ethnographiesmight then take. that the Huli as describedby Glasse seemed not to fit One can easily see that internationalmarkets link Bou- the accounts given for the Enga, Melpa, Chimbu, and gainville cocoa growersto the Chicago Board of Trade, Mendi. We can discernthe strainingfor a shareddomain but it is hard to visualize the appropriateway to write of comparative discussion also in Mervyn Meggitt's about that kind of systemwithout losing sightof those (i964) earlyattempt to comparesexual practicesand at- islanderswhose storieswe most want to tell. Nor do I titudes regardingmenstruation across the then "core" find much help in the pronouncementsof those who of "the Highlands,"which included the "Eastern High- most loudlyclaim to know the answer(e.g., Clifford and lands" cases. Meggitt'singenuity was severelytaxed in Marcus I 986). the effortto delineate contrastingpractices not only be- Nevertheless,the challenge to move fromthe ethno- tween "the Enga" and "the Kuma" but also between graphicpresent to historicalunderstanding-from mor- the "western"and "eastern"Highlanders (thus "prudes" phologyto process in Hays's terms-cannot be ignored. versus "lechers" as against cases in which elsewhere For those who would meet this challenge in writing separatedsyndromes of attitudesappeared to runin tan- about the Pacific islands and theirinhabitants, salutary dem). Along with most of us at the time, Meggittwas ground-clearingexercises like the present article are engagedin a kind of comparativebricolage, taking data most welcome. as he could findthem within an assumed culturaluni- HAYS "The New Guinea Highlands" I I57

verse.Later work was bound to make difficultiesfor this fieldworkcarried out in i99i by G. Sturzenhofeckerand early enterprise,hence the proliferationof margins, myselfin one tiny but morphologicallystrategic cor- fringes,and (potentially)rival cores thatoccurred subse- ner of the supposed Highlandshas amply demonstrated quently.Perhaps the only overallframework that in aca- the value of an antitypologizingapproach. In this far- demic termscould have influencedus was the linguistic westernpart of Duna-land, at least, it is necessaryto pictureso rapidlydelineated by Stephen Wurm (i964) recognizeimportant links and parallelswith Ok peoples and his colleagues at the Australian National Univer- west of the Stricklandand Papuan Plateau peoples to sity. This definitelysuggested a certain uniformityof the south as much as with the Huli and Paiela peoples to originsand perhapsfunctioned as an image-schemabe- the east (cf.Strathern and Sturzenhofeckern.d., Biersack yond its own limited intentions,for since Boas it had n.d.). As linguists have found with regard to dialect surelybeen known that language and culturewere not chains, we are dealing here with linked chains of cul- isomorphic.Perhaps, though, Hays could have discussed turalprocesses over time thatproduce not homogeneity more explicitly the linguistic picture-which also at but a mosaic of practicesand ideas in flux,and it is this "the margins"can become fuzzy,as with the ambiguous process that we need now to study in detail. In the classificationof the Duna language either in the East course of our doing so, regionsmay disappearand reap- New Guinea Highlands or the South-WestNew Guinea pear in other guises and with overlappingconforma- Stock. tions,but it will become evidentthat formany reasons The ambiguitiesof boundarieschosen as pointsof ref- (demographic,economic, religious, political) boundaries erence can be seen clearly in a latter-daycomparative between them are always permeable and shifting,and enterpriseto which Hays refersonly in passing-the de- typologyis always dissolvinginto history. bate regarding"big-man" versus "great-man"societies/ formsof leadership.The tack chosen in this work has JAMES F. WEINER been to obliteratethe selective focus on the Highlands Departmentof Social Anthropology,University of by bringingin cases fromelsewhere (e.g., Orokaiva, Va- Manchester,Manchester M13 9PL, England. 25 x 92 nuatu,Rossel Island,Mekeo, Arapesh)while concentrat- ing on a chosen set of structuralvariables considered Whatis the differencebetween the way we identify"the significantfor the longue dur6e.It is generallya produc- New Guinea Highlands" or the "Mountain Papuans" tive strategy,but thereis a sense of strainin imposing and the way Hays elsewhere identifies"the Highlands a single evolutionarymodel of a transitionfrom great- sacredflute complex" (i988)? Does he make any less of man to big-mancases; thereis the recursionof compara- an appeal to characteristicsof membershipin his defi- tiveproblems within as well as betweencategories; and, nition of what afterall is as much an identifiedand finally,there is a strainin elevatingthe category"great- bounded "region" of anthropologicaldiscourse as "the man" into a single analyticaltype in contrastwith "big- New Guinea Highlands"? What, in eithercase, are the man," which itself had its prototypicalstart in the criteriafor identifyingthe "sameness" of the traitsof "core" ofthe Highlandsand thus neverachieved analyti- which such regionsare composed,whether subsistence cal rigoruntil it was suddenlyranged against its latter- regime or origin myth? In which of the two cases is day antitype.Indeed, it is arguable that the beginnings thereless of an appeal to boundedness,discreteness, or of this process can be discernedin a kind of structural systematicity?What is the point of criticizingthe fuzzi- contrast that applies between the "Melpa" and the ness of a concept such as "the New Guinea Highlands" "Anga," in spite of the modest variationwithin the for- when one does not also criticize the boundedness and mer categoryand the much broadervariation now dis- referentiallimits of all the termsin the anthropological cernedin the latter.The attemptto createa "hard" cate- repertoirewhich contributeto such a concept,includ- goryof "great-mansociety/form of leadership"tends to ing, for example, "patrilineality,""big man," "initia- founderon the fuzzy shores of ethnographicand his- tion," "flute," "sacred," and "myth"?In any case, how toricaldata. None of this implies that the exercisewas does the ambiguityor disagreementconcerning the dis- not worthwhile;it only underlinesthe hazards of any tinguishingfeatures of a termaffect the use ofthat term? comparativeenterprise, no matter what baselines are How can a case be made that full knowledge of such chosen. featuresis a prerequisitefor "correct" employmentof A shiftfrom morphology to processwill certainlynot the term unless one has arbitrarilyand tautologously enable us to recreatehard sets or units forcomparison specifiedbeforehand which featuresone will accept as making.Instead, as Hays suggests,it can become a focus legitimate? fortheorizing in itself.His deftand well-takensugges- In short, Hays expresses doubt about the terms of tionshere are much in line with ones pursuedin a series currentclassifications without at the same time ques- ofsessions at the Associationfor Social Anthropologyin tioningthe efficacyof the classificatoryprocess. With Oceania yearlyconferences from I988 to I990. In these the authoritieshe cites, Hays confuses the notion of sessions we explicitlytook the whole island of New family resemblance with the observation that many Guinea as a unit of referenceand deliberatelyeschewed terms,academic or otherwise,have "fuzzyboundaries." any concentrationon a single imputedregion, concern- "Fuzzy boundaries"is a characteristicof a termthat has ing ourselves ratherwith concepts of flow and circula- no determinatetruth-value. But, as Norman Malcolm tion such as were pioneered by James Watson. Later remindsus, the determinationof truth-value is not what I58 CURRENT ANTHROPOLOGY Volume 34, Number 2, April 1993

the notion of familyresemblance was intended to ad- tialist" effortto isolate those traitswhich separatethis dress.To say,for example, that the word "chair" applies regionfrom the surroundingareas. He clearlyshows the to a familyof cases is to say that thereis nothingcom- problemsthat exist with this effort,which has been the mon to all chairs which justifies the use of the label focusof considerabletheoretical debate among Melane- "chair." It does not mean that thereis a familyresem- sianists. More important,he brings to our attention blance common to all chairs. "Whenever a word is a what has been lost as Highlands specialists have strug- familyresemblance word thereis no truegeneralization gled to findan agreed-uponset oftraits. By searchingfor that determinesits extension.... If thereis no general- what was common internally,we have ignored(i) the ization determiningthe extensionof a predicatethere is heterogeneityof the region's cultures,(2) the influence no 'truthrule' for that predicate" (Malcolm I978:4i6). of colonial forces,and (3) the exchangesbetween moun- The whole notion of family resemblance was sug- tain and lowland communitiesthat were a criticalpart gestedby Wittgensteinprecisely as a counterto the es- of regionalcultural dynamics. Hays asks how we could sentialist notions that are so carefullyeuphemized in have ignored these things for so long and goes on to the cognitivistwriting to which Hays defers.The intent advocate a shiftfrom the essentialistperspective to one of the contributorsto The Mountain Papuans was pre- that emphasizes culturalprocess. cisely to show that "traits" consideredspecifically char- I believe that it might also be fruitfulto aim a pro- acteristicof the Highlands(e.g., high pig-per-capita ratio, cessual lightat ourselves.By placingthe workof anthro- high population density)were also found in societies pologistsin a historicalcontext we mightbetter under- which otherwisehad verylittle in common with them. stand the disciplinaryfactors that led to the essentialist It was a polemic deployedagainst just such categoriza- effort.Given that it has now been 50 years since Reo tions. Fortune'spioneering ethnographic work with the Kai- Hays calls attemptsto delineate bounded regionses- nantupeoples, this may now be the time forsuch reflec- sentialist.But unlike his own essentialism,the squab- tion. The Highlands gained ethnographicsignificance bles over Papua New Guinea regions are a by-product afterWorld War II, when anthropologyhad reached a ratherthan a goal of academic debate-a debate which stageof some academic maturity.It was onlyjust begin- concerns the theoretical and ethnographicstatus of ning to be interestedin social change and still had the thingslike reciprocity,gender, and language. In other tribalworld as its primaryinterest. What more inviting words,Hays fails to stipulatethe verydifferent ends of opportunitycould therehave been than an entirelynew discussions of differentPapua New Guinea regions.As regionto investigate,and one that had only been "dis- do Papua New Guineans themselves,anthropologists of covered" within the precedingtwo decades? Theoreti- this area articulate oppositional contrastsat different cally, the structural-functionalismof Radcliffe-Brown levels. Contrastsbetween the easternand westernhigh- came to dominateBritish anthropology in the I 95os, and lands are not automatically comparable to those be- what betterethnographic "laboratory" could therehave tween highlands and fringehighlands, and neitherare been forthe developmentof a science of society?In the by definitionimplicated in a broaderhighlands-seaboard United States in the I96os, cultural ecology came to comparison.Hays's suggestionthat the terms of such have a significancesimilar to that of Britishstructural- contrastsought to retain theirsignificance at different ism, and the Highlands created opportunitiesfor the levels and on differentscales is surelya most plangent studyof tribalenvironments. Hays should be in a good essentialism.It implies that the termsof classification positionto commenton how these historicalfactors af- could be strippedof theirrhetorical, didactic, and pre- fectedthe emergenceof the "Highlands" category,since emptiveuses withinan ongoingconversation among Pa- he has recentlyedited a volume of fieldworkrecollec- pua New Guineaists; but these uses, far from being tions.by many of the early Highland ethnographers.It somethingextra which pragmaticallyalters or misrepre- may be that the fuzzy "Highlands" categorywas as sents theircore, semantic value, are the veryconditions much a culturalconstruct emergent from the historical under which such terms and those to which they are circumstancesof anthropologyas it was a set of cultural placed in opposition acquire meaning and force.Hays traitsreflective of the peoples of that region. sees the ambiguityand argumentsurrounding the classi- ficationof Papua New Guinea Highlandssocieties as an undesirableby-product of the "fuzziness" of our terms rather than as the very discursive situation towards Reply which we labour. Thereforehis comments do not ad- dressor repairthe effectiveterms of debate about Papua New Guinea societies. TERENCE E. HAYS Providence,R.I., U.S.A. 23 XI 92 GEORGE D. WESTERMARK Departmentof Anthropologyand Sociology,Santa I am gratefulto my colleagues fortheir thoughtful com- Clara University,Santa Clara, Calif., U.S.A. 9 xi 92 mentson a paperI wrotewith some trepidation.Weiner correctlypoints out that I myselfhave writtenmuch in Hays provides us with an insightfulcritique of the the past about "the New Guinea Highlands" and may "Highlands" category,calling into question the "essen- have been as guiltyof "essentialism" as those I criticize. HAYS "The New Guinea Highlands" I 59

I am, then, understandablyanxious about any sugges- veyed or cited others to which my criticismsdo not tion thatI have been pursuing,or at least writingabout, apply, I inferthat my depiction of the status of "the a chimerafor the past two decades. If I had been a com- New Guinea Highlands" as a descriptiveand analytical mentatorrather than the author of this paper,I would constructis probablya fairand accurateone. Put simply, have read it carefullyindeed. "the New Guinea Highlands" is, in Hayano's words,"a Such a carefulreading would have disclosed,as nearly murky anthropologicalconcept"; in mine (emphasis all of the commentatorshave realized, that the main added),"as used in anthropologicaldiscourse, [it] exem- focus of my paper is not in fact the highlandpeoples of plifieswell what cognitivepsychologists would call a New Guinea but "the New Guinea Highlands."That is, 'fuzzy set."' (Whetherhighlands societies necessarily it is largelyabout a descriptiveand analytic construct, constitutea fuzzyset is a separatequestion, to which I hence my use of quotation marks in the title and shall returnbelow.) throughoutthe paper. My firstobjective was to present With respect to individual works, the constructis a critique of how anthropologists(especially) have for- murkyor fuzzybecause of failureto state what criteria mulated and used the notion of such a "region." Then, are definitivein assigningethnographic cases to the cat- afterdiscussing some ofthe problemswith theseusages, egoryor because of inconsistentor careless application I proposed-too starkly,I now recognize-a shift in of the criteriachosen. Viewing the works collectively, our focus if we are to be successfulin addressingques- the murkinessbecomes perhapseven more consequen- tions that seem refractoryto analyses employingunits, tial, since, given the varyingreferential content of the whether societies or regions, defined on the basis of category(cases included or excluded by various writers), "traits"they possess. the growinglist of "traits" attributedto highlandssoci- My critiquewas promptedin partby a practicalissue eties is illusory. Aftersome 6o years of ethnographic raised by my impendingproduction of a new edition of reportson these societies,we mightexpect a cumulative a bibliographyof "the New Guinea Highlands." In an pictureof "the region"to have emerged,but when differ- earlierversion (Hays I976 :vii) I noted that "determining ent writersrefer to differentranges of empiricalreality the geographicalboundaries of 'the Highlands' is . . . when theytalk about "the New Guinea Highlands" we difficultsince these depend to some extent upon the are left,as I said earlyin the paper,"wondering what we nature of the problem a researcheris tryingto solve." know afterall." Preferringto errby commission ratherthan omission,I Severalcommentators proffer views on how all ofthis opted to include "the Mountain Ok," speakers of lan- came to pass, but the question is most usefully ad- guages of the Angan family,and the ("Bosavi") peoples dressedat two levels. First,human beginsare classifying of the Great Papuan Plateau as well as the occupants of animals, and we continuallyseize upon perceivedsimi- the centralcordillera in Irian Jaya.Inclined to make the laritiesand differencesto createworlds filled with cate- same choices again but mindfulof the needs of others, gories-of "foods," "people," "societies," "regions," I had the issue of scope forcedupon me anew: what etc. LiPuma's suggestion"that the trope of 'sets' itself do anthropologistsusually mean by "the New Guinea should be abandoned" is surelyfutile. The "epistemo- Highlands"? I thereforesurveyed recent, prominent logical problem"Hayano says I raise but do not address comparativeand syntheticstudies that significantlyde- "of how concepts in human language can ever ade- ployedthe phrase in orderto answer my question. quatelyrepresent 'reality"' is, of course,a hoaryone and Thus, the object of my analysis was the "Highlands" one that people strugglewith in everydaylife as much constructas it has been used in these types of works as in scholarlydiscourse. But life cannot be put on hold ratherthan in the ethnographicliterature per se. Har- until it is solved to everyone'ssatisfaction. In doing an- rison'scharge of unfairness in my purported"objection" thropologyas in everythingelse, we create categories, that ethnographershave been led "to misrepresent sets, and constructs,including "communities" as well these societies as homogeneous,ahistorical, and isolated as "regions."While theremay be experientialdomains, fromthe outside world" is, then,misdirected, since rep- such as "natural kinds" in the biological world, that resentationsin the ethnographicliterature were not my constrainus in categoryformation (Atran i990), when directconcern. My argumentwas that highlandsociet- the focus of our interestappears not to come prepack- ies have oftenbeen so misrepresentedin the ethnolog- aged in neatlybounded, clearly demarcated chunks, we ical literature,but I would furtherassert that in many are prone to create "fuzzy sets." Ogan correctlyreads ethnographiesthe "ethnographicpresent" is alive and me to be saying that the constructionof such sets is well and that homogeneityis certainlyimplied in com- "a general human 'information-processingstrategy."' mon labeling practiceswhereby descriptive generaliza- Thus, ratherthan "assuming that such categoriesare tions seem to be offeredabout whole language groups Bad Things and need to be expungedfrom anthropologi- on the basis of studiesof single communities. Moreover, cal discourse," as Harrison puts it, I regardfuzzy sets the increasinguse of pseudonymousdesignations por- as probablyunavoidable componentsof our discourse.I trayscommunities not only as unconnectedto external advancedthe notion of "the New Guinea Highlands" as referencepoints and events but as unconnectable,ex- a fuzzyset not as an explanatorytool, as Hayano seems cept to "insiders" who know the fieldsites. to have understoodme to be doing,but onlyas a descrip- Since none of the commentatorshas eithertaken is- tive one, hopingto drawattention to the pitfallsof reify- sue with my readingsand renderingsof the works sur- ing what is only a constructthat takes its particular i6o I CURRENT ANTHROPOLOGY Volume 34, Number 2, April 1993

formfrom the interestsand prioritiesof those who in- to Jamesand VirginiaWatson, who went therefrom a vent it and to caution against "confusingthe logic of differentinstitutional base) and in the Goroka area (Ken- science with the logic of social formations"(LiPuma). neth Read and RichardSalisbury). Soon after,the "Cen- As for the origins of the particularfuzzy set called tral Highlands" map was filled out with Paula Brown "the New Guinea Highlands," Brown,Hayano, Jorgen- among the Chimbu, Andrew and Marilyn Strathernin sen, LiPuma, and Strathernpoint to colonial admin- the Mount Hagen area, and others.Thus, as Strathern istrators'actions, and Brown, Jorgensen,Knauft, and comments,"the ethnographerswho were firstpermitted Westermarkalso cite the role of anthropologistsand in- to workin this area themselvesby accidentprovided the stitutionalprograms. All of these observationsare well- 'cultural core' for the area or 'region' by workingin a taken,but what is therebyexplained is open to debate. numberof places identifiablewith major languages." I Administrators'demarcations of the "Central High- would add only that they did so not "by accident" but lands," other "districts,"and, later, "provinces" have because of methodological and theoreticalbiases that neverbeen based on ethnographicsurveys (as anthropo- are long-standingin our profession.Among the by- logical "regions" purportedlyare), nor have they ever productsof these choices were that(i) to the extentthat been presentedas reflectingsocial or culturalboundaries "fringesocieties" in facttend to have smallerand more in any simple sense. Probably for this reason, as I scatteredpopulations, "marginalization" and even "be- pointed out (n. 3), the anthropologistswhose works I nignneglect" were simultaneouswith "centralization," considered have not constructed "the New Guinea and (2) "traits" such as "big-manleadership" and "cere- Highlands" followingpolitical boundaries.At the same monial exchange,"which tendedto be foundamong the time,one kindof unwitting "collusion" amongadminis- large,geographically central populations, became reified trators,missionaries, and anthropologistsmay have had as "Highlands traits" (Knauft). a greatdeal to do with the emergentsalience of certain Harrisonperceives me as suggesting"that we should highland language groups and their elevation to the tryto stop thinkingin categoriesaltogether." That of "core"/as "prototypicalHighlanders." coursewould be, literally,humanly impossible. If he has Jorgensensuggests that "the density and scale of inferredthat I propose we abandon the creationof con- Highlands populations" was arguablya key factorin structssuch as "regions"for comparative purposes, then "administrativearrangements marking the Highlands I was not clear in my paper. I share Knauft's concern offas a regiondistinct from the rest of Papua and New thatan "emphasis on . .. networkscarries the dangerof Guinea." I pointedout thatwhile writersdiffer in their obliteratingwhat is distinctive to particular ethno- assignmentsof cases, there is apparentconsensus that graphicregions," and I keenly appreciatethe "creative speakersof Enga (or at least Mae Enga), Melpa (around tension . . . between the use of principledcriteria to Mount Hagen), Wahgi (Kuma), and Chimbu (especially identifyethnographic regions and the tracing of net- the Kuman dialect) are chartermembers of "the New works and processes that crosscutthem." My trueposi- Guinea Highlands," as some of the peoples in the Go- tion is that both goals are worthwhile.Among the com- roka and Kainantu areas also tend to be. The fourlan- mentators,Strathern and Weiner do appear to reject guage groups named are notably similar in that they regionalconstructs as unfoundedor at least unproduc- are very large by New Guinea standards. Moreover, tive, but when Strathern,citing the "big-man"/"great- theirconstituent populations often manifest somewhat man" debate,advocates a focuson "structuralvariables" higherdensities than do many others.Consequently, I regardlessof geographicalprovenience, he still must would suggest,they were perceivedas both logistically confront,as all comparativistsmust do, sampling as easier to work with and promisinggreater returns for well as definitionalproblems. If Weiner's collection the effortsof administratorsand missionaries.The cre- (i988) was "a polemic deployedagainst just such catego- ation ofa "CentralHighlands District" and the targeting rizations" as "the New Guinea Highlands," I find the of Mount Hagen, the Wahgi Valley,the Chimbu Valley, advancing of another, similar category-"Mountain and neighboringBundi forpioneering missionary work Papuans"-an odd strategy.In any case, my goal was logicallyfollowed (see Hays I992). not to "[down] the disciplinarydragon of Regionalism" When systematic programs of anthropologicalre- (Lederman) but to highlight and urge that we face search in "the Highlands" were launched fromSydney squarely its inherentchallenges: in Harrison's words, and the Australian National Universitybeginning in "to thinkmore carefullyabout the terms [Melanesian- I950 (mentionedby Brownand treatedat lengthin Hays ists] use." [i992]), both programswere informedby "structural- Brown, Hirsch, Jorgensen,and Knauft suggest that functional"agendas, as is noted by Westermark.Given there "really is" somethingdistinctive about, and de- a majoremphasis on social organizationand anthropolo- monstrablycommon to, the people who inhabitparticu- gists' long-termpreferences for larger populations, it is lar,boundable areas. If theywant to persuadeus of this, not surprisingthat, from I950 to I955, A. P. Elkin and the burdenis on them as it was on the authors of the S. F. Nadel sent theirstudents and colleagues to conduct works surveyedto address clearly the methodological pioneeringfieldwork among the Enga (MervynMeggitt questions referredto by Hauser-Schaublin: "why we and Ralph Bulmer),the Mendi (D'Arcy Ryan),the Huli classify,what criteriaand methodswe use, . . . and . . . (RobertM. Glasse), the Kuma (Marie Reay),and the Kai- whetherthe system of classificationchosen is consis- nantupeoples (Catherineand Ronald Berndt,in addition tentlyapplied. " I agree with Knauftthat we need to be HAYS "The New Guinea Highlands" I I6

"more ratherthan less empiricallyspecific in our com- act with each otheracross the boundariesimplicit in our parisonsand contrasts"but also endorseHayano's cau- "morphological" constructs,we can gain insight into tionarynote that "a simple 'absence of/presenceof' cod- "how the relations between groups, cultures, and re- ing of behaviors to discern which activities might be gions generatestructures that permitsocially objective characteristicof Highlands, fringe areas, and Low- categoriesto emerge"(LiPuma), or, as Ledermanputs it, lands . . . will [or only may?] produce similaritiesthat "how any 'region' is made to appear in the firstplace." are superficialand mask differencesthat are critical." Accordingto LiPuma, "Maring culture flows into and Thus, my main complaintregarding the workssurveyed overlapswith Kalam culture . . . and Manga cultureat was not that they were intrinsicallyill-conceived but the otherend of the valley," and Stratherncites "impor- that these challenges and strictureswere too oftennot tantlinks and parallels" amongthe Duna, "Ok peoples," met satisfactorily.Rather than denyingthe potential "Papuan Plateau peoples," Huli, and Paiela. These eth- value of "morphological" approaches, my point was, nographers'reports exemplify well what I considerto be with Lederman, that "there are better approaches to one of the correctivesresulting from our attentionto 'morphology"'than those I criticized. "process": while for comparativepurposes, constructs As I said in the paper,the reasons forclassifying soci- such as "Maring," "Kalam," "Duna," etc., are perhaps eties into "regions,"like the criteriachosen, will vary necessaryand useful (ifthey are carefullyand explicitly with the interestsand prioritiesof the classifier.It was defined),their reification carries the risk of "[falsifying] not my intentto suggestthat comparativistshave used all we know of culturalinfluences and change,relations the "wrong" criteriain the past or that I know which with neighbors,and interculturalinteractions" (Brown). are the "right" ones. As Hauser-Schaiublinrecognizes, Carefulattention to linkages not only checks our im- "what kind ofclassification is the rightone" will depend pulses towardsuch reificationbut also can help us un- on "the contextand the purposefor which such studies derstandhow societies and regions,and not just "societ- are made." Contra Weiner's claim, I would not contend ies" and "regions," acquire their distinctivecharacter. "that the terms of classificationcould be strippedof For both ethnographicand wider purposes, I find Li- theirrhetorical, didactic, and pre-emptiveuses within Puma's analogy an excellent example of the potential an ongoing conversation among Papua New Guinea- reconceptualizationsthat come from a shift in focus ists," but I hope and believe that as ethnographersand such as that I proposed: "what if ["the New Guinea comparativistswe are engagedin morethan a "conversa- Highlands"]were more akin to a solar systemin which tion" among specialists. Yet even in that "conversa- the orbit of any one culture is definedby the gravita- tion," we owe it to the peoples we are characterizing tional pull and push of all the others?" and tryingto understand,as well as to each other,to Linkages throughtrade are indeed "an obvious place make the definitionalcriteria we employand our appli- to startinvestigation" (Ogan) of "gravitationalpull and cations of them explicit,principled, systematic, and ac- push," but there are numerous other candidates,some curate. of which have barely, or never, been tapped: warfare In my discussion of a shift from "morphology"to and alliance patterns;intergroup marriage (sometimes "process" in our research,I did not mean to advocate a across language boundaries)that entails not only gene replacementof comparativestudies, nor did I intendto flow but "culture flow"; migrationand resettlement, suggestthat "our researchmight be betterfocused" with whethergovernment-sponsored or as local copingstrate- such a shift(Ogan, emphasis added). Rather,I meant gies (WagnerI97I, Waddell I975); the diffusionand syn- only that in the pursuitof some kinds of questions "re- cretic elaboration of decoration and dance styles (and gions," and perhaps even "communities" and "societ- who knows what else?) that occurs throughsuch cata- ies," are inappropriateor misleadingunits of description lysts as "the Highlands Show"; and the role of planta- and analysis.I cited engagingin tradeas one instanceof tion labor experiencesor even jail sentences in the fos- "what people do" that oftencrosses "regional" bound- teringor blurringof "ethnic boundaries" throughwhat aries, but this was not intended as the only example Ledermancalls "indigenouscomparative discourses." worthyof attention,as Hirsch would have me say. Cer- As Lederman cautions, studies focused on "process" tainly,with respect to trade,Hauser-Schaublin is cor- have theirown problems,and we must not suppose that rect in notingthat "this approach is not new forNew theyare simple to conduct.Ogan expressesconcern that Guinea," and I would not claim authorshipof it, as Har- in a "village-outward"approach "it is not clear ... what risonimplies. Nor would I claim that "focusingon what shape(s) our ethnographiesmight then take." Obviously people do ... can [ever]explain ... structuralrelations" the verydescription, let alone understandingand expla- or the "politics of identity"(LiPuma, emphasis added). nation, of some phenomena is not well served by the Focusingon "linkages" is a strategyfor discovering and model followed in typical community studies. Some depictingwhat needs to be explained,including similar- leads may be suggested by studies such as Finney's ities and differencesamong the units chosen. Thus, as (II973, I987) of entrepreneurshipin the Goroka Valley Strathernsays, a shiftfrom "morphology" to "process" or Sexton's (i986) of the Wok Meri movement in the can "become a focus fortheorizing in itself." Daulo Pass area. But, as in comparativestudies, we will As I triedto suggestby alludingto boundary-creating be forcedto deal explicitlyand satisfactorilywith thorny and boundary-maintainingdevices deployedby "ethnic issues of definition,measures, and delimitationof the groups,"if we examine the ways in which people inter- scope of inquiryin terms of the cases to be included. I62 CURRENT ANTHROPOLOGY Volume 34, Number 2, April 1993

We must be ever mindfulof the fact that,like "societ- DOWNS, IAN. I980. The Australiantrusteeship: Papua New ies" and "regions,""networks" also are constructs,de- Guinea 1945-75. Canberra:Australian Government Publishing mandingcomparable attention to why and how we cre- Service.[DJ] EHRICH, ROBERT, AND GERALD M. HENDERSON. I968. "Cul- ate and use them. turearea," in Internationalencyclopedia of thesocial sci- ences,vol. 3. Editedby David L. Sills,pp. 563-68. New York: Macmillan. FARDON, RICHARD. I990. "Localizingstrategies: The regional- References Cited izationof ethnographicaccounts: General introduction," in Lo- calizingstrategies: Regional traditions of ethnographicwrit- APPADURAI, ARJUN. I986. Theoryin anthropology:Center and ing.Edited by RichardFardon, pp. I-35. Washington,D.C.: periphery.Comparative Studies in Societyand History SmithsonianInstitution Press. 28:356-6i. [BMK, RL] FEIL, DARYL K. I987. The evolutionof highland Papua New .I988. Introduction:Place and voice in anthropological Guinea societies.Cambridge: Cambridge University Press. theory.Cultural Anthropology 3:I6-20. [BMK] FINNEY, BEN R. I973. Big-menand business:Entrepreneurship ATRAN, SCOTT. I990. Cognitivefoundations of natural history: and economicgrowth in theNew Guinea Highlands.Hono- Towardsan anthropologyof science. Cambridge: Cambridge lulu: UniversityPress of Hawaii. UniversityPress. . I987. Businessdevelopment in thehighlands of Papua BARNES, J. A. i962. Africanmodels in the New Guinea High- New Guinea. PacificIslands Development Program Research lands.Man 62:5-9. [DJ, AS] Report6. BARTH, FRED RIK. I969. "Introduction,"in Ethnicgroups and GELBER, MARILYN G. I986. Genderand societyin theNew boundaries:The social organizationof culturedifference. Ed- Guineahighlands: An anthropologicalperspective on antago- itedby FredrikBarth, pp. 9-38. Boston:Little, Brown. nism towardwomen. Boulder: Westview Press. BIERSACK, A. Editor.n.d. New perspectiveson the New Guinea GELL, A. i992. "Inter-tribalcommodity barter and reproductive Highlands.MS. [AS] giftexchange in Old Melanesia,"in Barter,exchange, and BLOCH, MARC. 197I (I913). The Ile-de-France:The country value. Editedby C. Humphreyand S. Hugh-Jones.Cambridge: aroundParis. Ithaca: CornellUniversity Press. [BMK] CambridgeUniversity Press. [EH] BOYD, DAVID J. I985. "We mustfollow the Fore": Pig hus- GERBRANDS, ADRIAAN. I95I. Kunststijlenin WestNieuw- bandryintensification and ritualdiffusion among the Irakia Guinea. TijdschriftIndonesie 4:25 I-83. [BH] Awa, Papua New Guinea.American Ethnologist I : II9-36. GEWERTZ, D. I983. SepikRiver societies. New Haven: Yale Uni BROOKFIELD, HAROLD C. I96I. The highlandpeoples of New versityPress [RL] Guinea: A studyof distributionand localization.Geographical GODELIER, M. I986. The makingof great men. Cambridge: JournalI27:436-48. CambridgeUniversity Press. [RL] . i962. Local studyand comparativemethod: An example GODELIER, MAURICE, AND MARILYN STRATHERN. Editors. fromcentral New Guinea.Annals of theAssociation of Ameri- I990. Bigmen and greatmen: Personificationsof power in can Geographers52:242-S4- Melanesia. Cambridge:Cambridge University Press. [BMKI . I964. "The ecologyof highland settlement," in New GORECKI, PAUL P., AND DAVID S. GILLIESON. I989. "Intro- Guinea: The centralhighlands. Edited by James B. Watson,pp. duction,"in A crackin thespine: Prehistory and ecologyof 20-38. AmericanAnthropologist 66(4, pt. 2). the Jimi-YuatValley, Papua New Guinea. Editedby Paul P. . I990. Reviewof: The evolutionof highland Papua New Goreckiand David S. Gillieson,pp. I-4. Townsville:James Guinea societies,by D. K. Feil (Cambridge:Cambridge Univer- Cook Universityof North Queensland, School of Behavioural sityPress, I987). The AustralianJoumal of Anthropology Sciences,Division of Anthropology and Archaeology. I:68-69. GREEN, R. C. I99I. "Near and remoteOceania: Disestablishing BROWN, D. J. J. 1979. The structuringof Polopa feastingand 'Melanesia'in culturehistory," in Man and a half:Essays in warfare. Man I4:7I 2-33. Pacificanthropology and ethnobiologyin honourof Ralph BROWN, PAULA. I978. Highlandpeoples ofNew Guinea. Cam- Bulmer.Edited by AndrewPawley, pp. 49i-5o2. Auckland: bridge:Cambridge University Press. PolynesianSociety. [EOJ . I99I. Reviewof: Sepikheritage: Tradition and changein HADDON, A. C. I894. Decorativeart of BritishNew Guinea. Papua New Guinea,edited by NancyLutkehaus et al. (Dur- Dublin. [BH] ham: CarolinaAcademic Press, I990). AmericanAnthropolo- HARRIS, R. I986. The originof writing.London: Duckworth. [EHI gist 92:765-66. HARRISON, SIMON. I989. Magical and materialpolities in Mel- BROWN, P., AND G. BUCHBINDER. Editors.I976. Man and anesia.Man 24:I-20. womanin theNew Guinea Highlands.American Anthropologi- HAUSER-SCHAUBLIN, BRIGITTA. I989. Kulthauserin Nord- cal AssociationSpecial Publication8. [PB] neuguinea.Abhandlungen und Berichtedes StaatlichenMuse- BiUHLER, AL FRED. 196I. Kunststileam Sepik.Basel: Museum ums furVolkerkunde Dresden 43. [BH] fiurVolkerkunde. [BH] HAYANO, DAVID M. I974. Marriage,alliance, and warfare:A BULMER, SUSAN, AND RALPH BULMER. I964. "The prehistory view fromthe New Guinea Highlands.American Ethnologist ofthe Australian New Guinea highlands,"in New Guinea: I: 28I-93. [DMH] The centralhighlands. Edited by JamesB. Watson,pp. 39-76. HAYS, TERENCE E. I976. Anthropologyin theNew Guinea AmericanAnthropologist 66(4, pt. 2). Highlands:An annotatedbibliography. New York:Garland. CARRIER, JAMES. i992. Historyand traditionin Melanesianan- . I986. Sacredflutes, fertility, and growthin the Papua thropology.Berkeley: University of CaliforniaPress. [EO] New Guinea highlands.Anthropos 8I:435-53. CLIFFORD, JAMES, AND GEORGE E. MARCUS. I986. Writing . I988. "'Mythsof matriarchy' and the sacredflute com- culture:The poetics and politicsof ethnography. Berkeley: Uni- plex ofthe Papua New Guinea Highlands,"in Mythsof matri- versityof CaliforniaPress. [EO] archyreconsidered. Edited by D. Gewertz.Oceania Monograph CRAIG, BARRY. I990. "Is the MountainOk culturea Sepikcul- 3 3. [JFW] ture?"in Sepikheritage: Tradition and changein Papua New . I990. The centralityof the Mek-Okregion in the diffu- Guinea. Editedby NancyLutkehaus et al., pp. I29-49. Dur- sion oftobacco in New Guinea.Paper presented at the confer- ham: CarolinaAcademic Press. ence "The Mek and TheirNeighbors," Seewiesen, Germany. DORNSTREICH, MARK D. I974. An ecologicalstudy of Gadio . i992. "A historicalbackground to anthropologyin the Pa- Enga(New Guinea) subsistence.Ph.D. diss.,Columbia Univer- pua New Guinea Highlands,"in Ethnographicpresents: Pio- sity,New York,N.Y. neeringanthropologists in thePapua New Guinea Highlands. HAYS "The New Guinea Highlands" I I63

Editedby Terence E. Hays,pp. I-36. Berkeley:University of MALINOWSKI, B. I965 (i922). Argonauts of the WesternPacific. CaliforniaPress. New York:Dutton. [RL] HEALEY, CHRISTOPHER J.I980. The tradein birdplumes in MANDEVILLE, ELIZABETH. I980. Review of: Highland peoples the New Guinea region. Universityof Queensland Anthropol- ofNew Guinea,by Paula Brown(Cambridge: Cambridge Uni- ogy Museum Occasional Papers I0: 249-7 5. versity Press, I978). Man I5:549-50. . I990. Maringhunters and traders.Berkeley: University MEAD, MARGARET. I978. The Sepik as a culture area: Com- of CaliforniaPress. [RL] ment. Anthropological Quarterly 5I: 69-75. HERDT, GILBERT H., AND FITZ JOHN P. POOLE. i982. "Sex- MEGGITT, M. j. I964. "Male-female relationships in the High- ual antagonism":The intellectualhistory of a conceptin New lands ofAustralian New Guinea,"in New Guinea: The central Guinea anthropology.Social Analysis I2:3-28. [BMK] highlands.Edited by J.B. Watson,pp. 204-224. AmericanAn- HERZFELD, MICHAEL. I984. The horns of the Mediterraneanist thropologist66 (4,Pt. 2). [AS] dilemma. American Ethnologist II:439-54. MtTRAUX, R. Editor. I978. Sepik politics. Anthropological Quar- HIRSCH, E. I990. Frombones to betelnuts:Processes of ritual terly 5i(i). [PB] transformationand the developmentof "naturalculture" in Pa- MILLER, BRIAN. I983. The highlands of Papua New Guinea. pua New Guinea. Man 25:i8-34. [EH] Bathurst:Robert Brown. HUGHES, IAN. I977. New Guinea StoneAge trade:The geogra- NASH, JILL, AND EUGENE OGAN. The red and the black: Bou- phy and ecology of trafficin the interior. (Terra Australis 3.) gainvilleanperceptions of otherPapua New Guineans.Pacific Canberra:Australian National University, Research School of Studies I3:I-I7. PacificStudies, Department of Prehistory. NELSON, HANK. i982. Taim bilong Masta. Sydney: Australian . I978. Good moneyand bad: Inflationand devaluationin BroadcastingCommission. [DMH] the colonialprocess. Mankind II:308-I8. 0 HANLON, M. i992. Unstable images and second skins: Arte- KAY, PAUL, AND CHAD K. MC DANIEL. I978. The linguistic facts,exegesis, and assessmentsin the New Guinea Highlands. significanceof the meaningsof basic colorterms. Language Man 27:587-608. [EH] 54:6IO-46. READ, KENNETH E. I954. Cultures of the central highlands, KEESING, ROGER. I990. "New lessons fromold shells: Chang- New Guinea. Southwestern Journalof AnthropologyIO: I-43. ing perspectives on the kula," in Culture and historyin the Pa- ROSCH, ELEANOR, AND CAROLYN B. MERVIS. I975. Family re- cific.Edited by Jukka Siikala, pp. I39-63. Helsinki:Finnish An- semblances:Studies in theinternal structure of categories. Cog- thropologicalSociety. [EO] nitive Psychology 7:573-605. KELLY, RAYMOND C. I976. "Witchcraftand sexual relations: ROSCH, ELEANOR, CAROLYN B. MERVIS, WAYNE D. GRAY, An explorationin the social and semanticimplications of the DAVID M. JOHNSON, AND PENNY BOYES-BRAEM. I976. Ba- structureof belief," in Man and woman in the New Guinea sic objectsin naturalcategories. Cognitive Psychology 8:382- highlands.Edited by Paula Brownand GeorgedaBuchbinder, 439. pp. 36-53. AmericanAnthropological Association Special Pub- SEXTON, LORRAINE. I986. Mothers of money, daughters of cof- lication 8. fee: The Wok Merimovement. Ann Arbor:UMI Research KING, DAVID, AND STEPHEN RANCK. Editors. n.d. (i982). Pa- Press. pua New Guinea atlas: A nation in transition. Port Moresby: SINCLAIR, JAMES P. I97I. The highlanders. Milton: Jacaranda RobertBrown and Associatesand Universityof Papua New Press. Guinea. SPEISER, FELIX. I937. Uber Kunststile in Melanesien. Zeit- KNAUFT, BRUCE M. I990. Melanesian warfare:A theoretical his- schriftfur Ethnologie 68:304-69. [BH] tory. Oceania 60:e250-3II. STRATHERN, ANDREW J. I990. Which way to the boundary? . I993. South New Guinea cultures: History, comparison, American Ethnologist I7:376-83. dialectic.Cambridge: Cambridge University Press. In press. . I99I. Circulatingcults in HighlandNew Guinea: KULICK, D. i992. Languageshift and culturalreproduction. Pointersfor research. Australian Journal of Anthropology 2: Cambridge:Cambridge University Press. [EH] 98-IO7. LAWRENCE, PETER, AND MERVYN J. MEGGITT. I965. "Intro- STRATHERN, A. J., AND G. STURZENHOFECKER. Editors. Mi- duction," in Gods, ghosts, and men in Melanesia: Some reli- grationand transformations.Pittsburgh: University of Pitts- gions of Australian New Guinea and the New Hebrides. Ed- burghPress. In press.[AS] itedby PeterLawrence and MervynJ. Meggitt, pp. I-26. STRATHERN, M. I988. The gender of the gift.Berkeley: Univer- Melbourne:Oxford University Press. sityof CaliforniaPress. LEDERMAN, R. i99oa. Big men, large and small? Towards a com- . I990. "Negativestrategies in Melanesia,"in Localizing parative perspective. Ethnology 29:3- I 5. [RL] strategies:Regional traditions of ethnographicwriting. Edited . iggob. Criticalperspectives on the comparativemethod. by Richard Fardon, pp. 204-i6. Washington, D.C.: Smithso- Paperpresented at the ColumbiaUniversity Seminar on Ecolog- nian InstitutionPress. [BMK] ical Systemsand CulturalEvolution, New York.[RL] TERRELL, JOHN. I986. Prehistoryin the Pacific Islands: A . I99I. "'Interests'in exchange:Mendi big-men in con- studyof variationin language,customs, and human biology. text," in Big men and great men: The development of a com- Cambridge:Cambridge University Press. parisonin Melanesia. Editedby M. Godelierand M. Strathern. THOMAS, NICHOLAS. The force of ethnology: Origin and sig- Cambridge:Cambridge University Press. [RL] nificanceof the Melanesia/Polynesiadivision. CURRENT AN- I992. Reply.Pacific Studies. In press.[RL] THROPOLOGY 30:27-42. n.d. Review of: Maring hunters and traders,by C. Healey TIESLER, FRANK. I969-70. Die intertribalenBeziehungen an (Berkeley:University of CaliforniaPress, i990). AmericanEth- derNordkiiste Neuguineas im Gebietder kleinen Schouten- nologist.In press.[RL] Inseln.Abhandlungen und Berichtedes StaatlichenMuseums LINDENBAUM, SHIRLEY. I984. "Variations on a sociosexual furVolkerkunde Dresden 30-3I. [BH] theme in Melanesia," in Ritualized homosexuality in Mela- . I990. Problemeeiner Darstellung der Kunst Neuguineas. nesia. Editedby GilbertH. Herdt,pp. 337-6I. Berkeley:Uni- Abhandlungenund Berichtedes StaatlichenMuseums fur versityof CaliforniaPress. V6lkerkunde Dresden 44:235-55. [BH] L I P U MA, E D WA R D I1 98 8. The giftof kinship. Cambridge: Cam- WADDELL, ERIC. I975. How the Enga cope with frost:Re- bridgeUniversity Press. [EL] sponsesto climaticperturbations in the CentralHighlands of MALCOLM, N. I978. "Thinking,"in Wittgensteinand his im- New Guinea. Human Ecology 3:249-73. pact on contemporarythought (Proceedings of the Second In- WAGNER, ROY. I967. The curse of Souw: Principles of Daribi ternational WittgensteinSymposium). Edited by E. Leinfellner, clan definitionand alliance in New Guinea. Chicago:Univer- W. Leinfellner,H. Bershel,and A. Hiubner.Vienna: Holder-Pich- sityof ChicagoPress. ler-Tempslay.[JEW] I64 | CURRENT ANTHROPOLOGY Volume 34, Number 2, April 1993

. I97I. A problemof ethnocide:When a Chimbumeets a Historicaland comparativeperspectives from New Guinea Karimui.New Guinea 6(2):27-3I. fringehighlands societies. Edited by JamesF. Weiner,pp. . 1972. Habu: The innovationof meaning in Daribi reli- I-38. Ann Arbor:University of Michigan Press. gion.Chicago: University of ChicagoPress. [RL] WELSCH, ROBERT L., JOHN A. TERRELL, AND JOHN A. NA- . I974. "Are theresocial groupsin the New Guinea high- DOLSKI. i992. Languageand cultureon the NorthCoast of lands?"in Frontiersof anthropology:An introductionto an- New Guinea. AmericanAnthropologist 94:568-6oo. thropologicalthinking. Edited by Murray J. Leaf, pp. 99-I22. WHITE HEAD, HARRIET. I986. The varietiesof fertility cultism New York:D. Van Nostrand. in New Guinea. AmericanEthnologist 13:80-99, 27i-89. . I98I. The inventionof culture.Chicago: University of WILLIAMS, F. E. I940-4I. Natives of Lake Kutubu,Papua. Oce- ChicagoPress. [RL] ania II:I2I-57, 259-94, 374-40I; I2:49-74, I34-54. . I99I. in Bigman and greatman: The developmentof a WURM, S. A. I964. "AustralianNew Guinea Highlandslan- comparisonin Melanesia. Editedby M. Godelierand M. Strath- guagesand the distributionof theirtypological features," in ern.Cambridge: Cambridge University Press. [RL] New Guinea: The centralhighlands. Edited by JamesA. Wat- WATSON, JAMES B. I964. "Introduction:Anthropology in the son,pp. 77-97. AmericanAnthropologist 66(4, pt. 2). [AS] New Guinea highlands,"in New Guinea: The centralhigh- . I982. Papuan languagesof Oceania. (ArsLinguistica 7.) lands. Editedby JamesB. Watson,pp. I-I9. AmericanAnthro- Tiibingen: Gunter Narr. pologist66(4, pt. 2). ZADEH, L. A. I965. Fuzzy sets. Informationand Control WEINER, JAMES F. I988. "Introduction,"in MountainPapuans: 8:338-5 3.

Serials

AnthropologicalJournal on European Culturesis a new and Thomas Gerholm,"The CulturalStudy of Scandina- periodicalfocussing on currentEuropean dynamicsre- via: Where Are the Frontiers?"; Ina-Maria Greverus, sultingfrom fundamental structural changes, increasing "AnthropologicalHorizons, the Humanities, and Hu- complexity,and individualization,on the one hand, and man Practice"; and Christian Giordano, "Is There a forcedhomogenization, on the other. It is edited by MediterraneanAnthropology? The Point of View of an ChristianGiordano and Ina-Maria Greverusand has an Outsider."Future issues will be devotedto urbanEurope internationaleditorial board. Volume i (nos. i and 2) is and to worldview, political behaviour,and economyin entitledAnthropologizing Europe and includes, among the post-Communisttradition. Subscriptions are avail- others,the followingarticles: AnthonyP. Cohen, "Self able forSfr 48/DM 58 forinstitutions, Sfr 32/DM 39 for and Other in the Tradition of BritishAnthropology"; individuals,and Sfr i8/DM 22 forstudents from AJEC, Mincho Draganov,"Scientific Heritage of Social Anthro- S6minaired'Ethnologie, Universit6 de Fribourg,Miseri- pologyin Bulgaria and Prospectsfor the Future"; Lena corde,CH-I7oo Fribourg,Switzerland.

Wanted

Cooperationfrom colleagues and institutionsconcerned ticipatein fieldworkin the Mediterraneanarea and else- with researchon the Mediterraneanin a newly estab- where,and drawingattention to Slovenia as a relatively lished Centre forMediterranean Studies at the Institu- unexplored and promising ethnographicfield. Please tum StudiorumHumanitatis, Ljubljana, Slovenia. The write: Iztok Saksida, Institutum Studiorum Humani- work of the centrewill be focused on the anthropology tatis,Riharjeva i, SI-6iooo Ljubljana,Slovenia. and historyof the Adriaticand otherregions of the Med- iterranean.A Master's and Ph.D. programwill receive Contributionsto an internationalguide to anthropologi- its firststudents in October I993. Anthropology'schar- cal resourcesin the process of compilation by the Li- acteristictopics of kinship, friendship,gender, migra- braryAnthropology Resource Group under the general tions,ethnicities, social conflict,etc., will be coveredby editorshipof Lee Dutton. The guide, to be published visitinglecturers with the appropriateregional interests. by Garland in I994, will present currentand detailed The language of teaching will be English. The centre informationon nonartifactanthropological resources in will give priorityto networkingactivities aimed at es- major libraries,museums, and repositoriesthroughout tablishingeffective collaboration with other research theworld. Please write:Lee Dutton,Founders Memorial institutions, offeringyoung Slovene anthropologists Library,Northern Illinois University,DeKalb, Ill. 6oi I 5, opportunitiesto studyanthropology abroad and to par- U.S.A.