Barter and Economic Disintegration Author(s): Caroline Humphrey Source: Man, New Series, Vol. 20, No. 1 (Mar., 1985), pp. 48-72 Published by: Royal Anthropological Institute of Great Britain and Ireland Stable URL: http://www.jstor.org/stable/2802221 . Accessed: 17/10/2014 05:02

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. 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].

.

Royal Anthropological Institute of Great Britain and Ireland is collaborating with JSTOR to digitize, preserve and extend access to Man.

http://www.jstor.org

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions BARTER AND ECONOMIC DISINTEGRATION

CAROLINE HUMPHREY Universityof Cambridge

The mainstreameconomists' view thatbarter should be seen as a 'natural'phenomenon of human natureand as the originof moneyis rejected.Barter occurs in specificsocio-economic conditionswhich may obtainalso in economieswhich know money.When thereis a verylow supply of currency,money may cease to functionas an index of value forall goods and itself become an item bartered.This is likely to occur when small discretesocial groups wish to maintainautonomy. Unlike moneypayment, which requires a furthertransaction before value is realised,barter satisfies demand immediatelyand is of its naturediscontinuous. As with car trade-insin our economybarter occurs when people cannotafford to keepmoney, and it becomes a systemwhen societyis atomised to the extentthat people do not exploit the variationsin exchangeratios between different communities. Using thecase ofthe Lhomi of north-east Nepal, it is shown thatalthough the exchange of commonproduce, as opposed to rarevaluables, is most likelyto approximateto a notional'equilibrium price', the practice of barterwith no established measuresof weight and volumemeans that there can be no underlyingindex of value /numeraire. Each transactionexists virtuallyon its own. Thus, althoughbarter is an egalitarianmode, it containsno protectionagainst changing exchange ratios which may harmone partner.Barter tendsto takeplace betweenpeople who know one another,because it is onlyby theestablishment of customarytimes and places forexchange that the costs of searchingfor partners, waiting etc. areavoided. Delayed barter,which often occurs with valuables, requires non-economic means of ensuringrepayment, but the ritualisedtrade-partnerships which the Lhomi employ are self- limiting:restricted relations cut tradersoff from the wide, unpredictableworld of the capitalist end sale. Businessoften fails and thetraders suffer.

Barteris at once a cornerstoneof modern economic theoryand an ancient subjectof debate about politicaljustice, fromPlato and Aristotleonwards. In bothdiscourses, which are distinct though related, barter provides the imagined preconditionsfor the emergenceof money. Why should anthropologistsbe interestedin logical deductionsfrom an imaginedstate? No exampleof a barter economy,pure and simple,has ever been described,let alone the emergence fromit of money;all availableethnography suggests that there never has been sucha thing.But thereare economies today which are nevertheless dominated by barter,and hereanthropology can bothlearn from and add to theancient debate. In Plato, non-monetaryexchange provides the firstmeans of satisfying complementaryneeds established by thedivision of labour, which is a condition of the ideal Republic (Plato I908: II, 369-72). Aristotledevelops the idea of 'proportionalexchange', relative to social evaluationsof the worthof persons and theirproducts, in his considerationof whatis just. Here money,which can measureeverything, is introducedby conventionas a 'kind of substitutefor Man (N.S.) 20, 48-72

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 49 need or demand', its value being derived not from nature but from law (AristotleI895: 5, I-I7). Economistsof the contemporaryorthodoxy, on the otherhand, which stems at least fromAdam Smithand was revivedin the marginalisteconomics of Menger andJevons and reachesClower in thepresent, propose an evolutionarydevelopment of economieswhich places barter,as a 'natural'human characteristic, at themost primitive stage, to be supersededby monetaryexchange as soon as people become aware of the latter'sgreater efficiency. Anthropologistshave also tendedto see barteras a transactionof the primitive economy, although in the classic examples, such as Malinowski (I922), Thurnwald(I934-5) and Sahlins(I972) theethnographic diversity and com- plexityof distributionis recognised.Dominated by otheractivities considered to be 'more socially embedded', such as ceremonialexchange, gift-giving, sharingof food, or dues to chiefs,barter is foundin a cornerof the economy -and one that is despised by the people and anthropologistsalike. Barter, accordingto Sahlins summarisinga wide range of ethnography,is 'negative reciprocity,the unsociable extreme'. Characterised as 'haggling',barter is held to takeplace withoutsiders, along with'chicanery' and 'theft',each participant tryingto outwitthe other with an eye to his own benefit(I972: I95). There is somethingunsatisfactory about each of thesemodern arguments. If we take barterto be the more or less simultaneousexchange of one good for anotherwith the possibility of bargaining,the anthropologists have no justifi- cation in regardingit as 'negative reciprocity'.By definition,barter is a complementaryexchange in whicheach participantbargains until he or she is satisfied.It does not necessarilyimply antagonism. As faras the economists' argumentis concerned,we know fromthe accumulated evidence of ethnogra- phy that barter was indeed very rare as a system dominatingprimitive economies.' Money of variouskinds has been aroupd forover two millennia, and in thelast century,in its purest,non 'commodity-money'form, has pen- etratedvirtually every economy on earth,and yet barteris common todayin economieswhich also know money.I shallpropose in thisarticle that barter in thepresent world is, in thevast majorityof cases, a post-monetaryphenomenon (i.e. it coexistswith money),and thatit characteriseseconomies which are, or have become, de-coupled from monetarymarkets. In these circumstances bartercan become a dominantprinciple of transaction,to whichmoney itself is subject.As was well known to Marx and even Aristotle(I895: 5, 14), but has somehow escaped the generalattention of anthropologists,currency may be barteredlike any othercommodity. As an idea 'barter'is partof thehistory of economicsand anthropology,and theassumption that it was theforerunner of monetaryexchange is crucialin the way it is normallyconceptualised. Polanyi's suggestion, in Thegreat transform- ation,that the evolutionarydoctrine of the economistspreoccupied with the emergenceof marketsmight virtually be reversed,has been largelyignored (I957: 58).2 His own view is considerablycloser to thatof the ancientphil- osophers,suggesting as he does theprior existence of long-distancetrade as a resultof regionalenvironmental differences and only laterthe emergenceof local marketsand bargainingor barter.In thisarticle I shouldalso liketo escape

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 50 CAROLINE HUMPHREY fromthe notion of 'truck'or 'barter'as a naturalhuman propensity which still appearsas a mythof our subject.This impliesthe disentangling of two concepts of barter,(a) the notion of a simple exchangeof goods withoutmoney and without specifyingthe mode of the non-monetarytransaction, and (b) the processof barterin timeas a practicewhich involves bargaining. It is thelatter thatwill be my focus. The word barteras I use it impliesan open-ended,potentially innovative, negotiable,transaction, in whichneed not onlyanswers need but can also create a new demand: 'If you don't want thesepotatoes, perhaps you would like this pair of scissors?'Furthermore, barter, which in itselfrefers only to a social relationof transaction and not to economicvalues, can encompassthe idea ofan exchangeof goods whichhave one value to one side and anotherto theother. I use 'barter'in thissense to differentiateit from 'primitive trade' in which, as Malinowskishowed, sociallydetermined rates of exchangeoutweigh bargain- ing in thegreat majority of typesof transaction. For barterto become dominantin an area, such thatit can incorporateeven puremoney, we mustsuppose not only economic but also someparticular social and culturalconditions which allow the 'construction'of barteras a system.I shallattempt to describethis on thebasis of the economy of the Lhomi3 and their neighboursof the Nepal-Tibetborder in I979-80. The Lhomi are farmersbut also traders,and clearly theirhistorical role as intermediariesbetween the highlandpastoral economy of theTibet hillsand theagriculture of the Nepali hills,is a basis forthe cultureof tradingwhich they find so attractive.I would argue, however, thatit is not because the Lhomi were tradersthat they now engagein barter.Similar Himalayan peoples are now almostentirely monetised. Clearlythere are manyconditions ranging from negative consequences of using money4to regulatedsystems of distributionin centralisedeconomies5 which mightimpel people to barteras it were by default.In othercases, themselves differingfrom one another greatlyin ideological context, bartermay be positivelypreferred in itself.We may citethe example of East Coast American farmersof the late eighteenth to mid-nineteenthcenturies, where products were exchangedbetween independent households and even cash was barteredjust outside New York. Merrill(I977: 42-7I) explainsthis rejection of capitalist, functionallymonetised farming as an expressionof radical republicanism, based on essentiallyindividualistic household production. It is certainlynot the case thatsuch examples are simply'survivals' of earlyforms of economy. From historicaland comparativematerials in theHimalayan region it can be seen that alternationsbetween increasingly monetised trade in periodsof accumulation, and barter,which in myview is herea phenomenonof economic disintegration in thevertical economy, have been highlyunstable. To show theoreticallywhy this is thecase we mustlook at theeconomic ideas in moredetail. The prevalenttheory of barter in theeconomic literature looks at barter,not in itself,but with the primary aim ofelucidatingthe role and originof money.Clower (I969), followingJevons(I9I0), statesthat barter is abandoned because of its high transactioncosts. In an imaginedisland economywithout moneyit is assumedthat people have a naturaldesire to acquiregoods theydo not produce themselvesand thereforeengage in exchange. With an elegant

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY S series of curves, Clower suggests that the costs involved in simple barter (searchingfor exchange partners, establishing a double coincidenceof wants, postponinga desiredtransaction, and wastingtime in bargaining)would lead people firstto barterat fairgrounds, then to establishtrading-posts for particular goods, thento makeuse ofthe one mostcommon item as a meansof exchange at all theposts and, finally,to establishthis commodity also as a standardof value, means of paymentand store of wealth,in otherwords, as money. Here the primaryfunction of moneyis assumedto be as a meansof exchange. Thomas Crump, who has made a surveyof both the economic and the anthropologicalliterature, argues convincingly,however, that Clower is wrong(I98 I: 85). Because thesupply of money must be undersome control,6it is unlikelythat any good emergingas theprimary means of exchangethrough frequencyof use in barter-which is what Clower suggested-would ever 'turninto' money. There is an alternativetheory, again dealingwith barter only as a seed-bed forthe emergenceof money, but one which is, I think,more usefulto us in understandingbarter itself. The economistCharles Goodhart argues against the Clower school on the groundsthat money as themeans of exchangeemerging fromprimitive barter cannot be theinitial key, since thisalready assumes the existenceof a marketeconomy and marketmentality. He is concernedto makea distinctionbetween money as a meansof exchange and itsnarrower function as a meansof payment. He sees thelatter as primary.Barter could distributegoods (and thusrender payment unnecessary) but, becauseof transactioncosts, most exchangeswill not be simultaneousbut will involve the extension of credit.The existenceof timemeans that even in an economyrigidly bound by customthere mustbe uncertainty,e.g. about weather,future technology, or thehonesty of people. Delayed bartercould onlyexist generally in face-to-facecommunities in whichevery exchange partner has knowledgeabout others such that he can trust paymentto be made foritems given. But thisis unlikelyever to be thecase, and lack of informationabout thefuture credit standing of prospectivepurchasers would in some cases forcepeople to use a specialisedmeans ofpayment, money. The proportionof transactionssettled by monetaryexchanges will risewith the growing complexityand dispersionof the economy because of the greater liklihoodof nothaving adequate information about otherpeople. Conversely, the use of money declineswith the developmentof methods,such as credit cards,to increasethe amount of personal information available (I975: 7-8).7 All thisis veryformalistic, but it does helpus to see somethingof the essential natureof barter.Barter makes paymentunnecessary. Payment is the transfer fromone person to anotherof an interestwhich is expressedin termsof a standardof value. It is moneywhich makes paymentpossible, as Crump has forcefullypointed out. The keypoint about money is that,whatever functions it may have, thepayee is in a positionto performthem by virtueof thepayment, but he can only do so by makinga furtherpayment. Thus it is of thenature of moneyto circulateindefinitely (Crump I98I: 3-4). Barter,on theother hand, entailsno furthertransaction. It is, in principle,a one-offaffair, because the objects exchangedare mainlyobjects for use. Barteragreements are made by individualsor groupsseparately and discontinuously,and withoutany implicit

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 52 CAROLINE HUMPHREY standardof value (thisis discussedfurther on p. 60). The implicationis that, unlikethe use of money,barter exchange will have theeffect of dis-integration in theregional economy. This gives some clues as to whicheconomies are likelyto be dominatedby barter.Barter will occur when economiesare atomised(when moneyceases to functionas a standardof value). 'Delayed barter',which in practicemay range from somethingclose to 'primitivetrade' to the activitiesof professional merchantsintent on keeping people in thrall(Leach I959: I45-6 and Gorer I938: I I3-i8) can onlyoccur when there is a largeamount of information about partners(or othersocial pressuresfor repayment).A furtherimportant con- ditionis suggestedby an earlypaper of Helms. She proposed the idea of the 'purchasesociety' which characterisesgroups on the fringesof incompletely centralisedstates (Helms i969). Such groupsexist in a complexrelation between politicalautonomy from and economicdependence on thecentral society. Her pointis thatthe people are 'hooked' on goods fromoutside and characteristically quicklybecome monetised.However, thereare societiesin a similarpolitical situationwhich resist monetisation (Harriss i982), in particularas represented by moneymarkets in land and labour. Other groups may be forced out of the money economy. Crump has suggesteda usefulmodel based on thenotion of upperand lower limitsto the supplyof money (i98i: 83-96). As we havenoted in thecase ofcowries(note 6), an object cannot functionas 'pure money', in Crump's terminology,if the supplyis unlimited(upper limit). Only aftersupply of the money-stuff has been subjectto limitation,either by transportto a differentregion or because it is manufacturedin a situationof control,will it functionas money. Flowing throughthis area, money passes to other regions in progressivelysmaller amounts,ending up in whatCrump callsa 'sink' (thelower limit), 'represented by a populationwhich imports these objects, but not so as to use themfor any recognisablemonetary purpose' (i98i: 86). In thesink, money-stuff cannot be used as money because thereis not enough of it around. A peripheralself- containedsub-economy must maintaina positivebalance of paymentsin its externaltrade with the nationaleconomy if it is to importenough money to maintainits own monetarysystem. In practice,simple poverty makes this very problematicin manyregions of the world, even in placesadjacent to monetised markettowns (i98 I: 2I2). It is what goes on in this'sink' thatwe shallbe concernedwith. I agreewith Crump thatthe supply of money is crucial. Simply to say thatan economy operatesbelow thelower limit of moneysupply tells us little,however, of what actuallyhappens. As we shall see (p. 62) thelevel of thesupply of moneydoes not explainthe value whichmoney has in a bartereconomy. For thiswe must look at barteritself, and hereI turnto theLhomi ethnography.

The Lhomi The Lhomi are a smallcommunity of farmersand tradersliving in thevalley of the Arun River close to the Tibetan border.They are Buddhistsand speak a dialectof Tibetan. Separatedby a precipitous,uninhabited jungle fromtheir

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 53 neighboursto thesouth, they live in a dozen or so compactvillages perched on thesteep mountain slopes. This is a 'verticaleconomy' with many similarities to theAndean case. The landsattached to villagesrange from yak pasturage above thetree-line (i6,ooo ft),through cattle and sheeppasture, swidden, and fieldsfor potato, wheat and barley on the upper levels (I0-I2,000 ft), maize, millet, buckwheat,fruit-trees and vegetablesat village level (6,ooo-io,oooft), and terracesof rice rightdown near the riverat approximately3,000 ft.Although eachvillage produces a rangeof crops, even slight differences in height,sunlight and rainfallmagnify the variations in whatthey can produceas surplus. The Lhomi farmis owned by and productionis organisedon thebasis of the household.Patrilineal clans have rightsto pasturesand residualrights to fields, and landis notfreely available for sale.8 The moresenior of the clans also provide the inheritedpositions of village headman(goba) and tax-collectors(gyembu). Only in thelast io-i 5 yearshas theauthority of these leaders been superseded by thatof electedpanchayat officials of the Nepalese state.About 3 per cent. of householdshave no land,just a fewmore than the number of familiesof'gara', i. e. thoseof 'uncleanmouth'. Otherwise, the overt ideology today is egalitarian, despitethe factthat those clans callingthemselves 'jimi' (originalsettlers) still tendto have morerice land thanlater arrivals. These latter,known as 'Kampa' althoughthey come fromvarious parts of Tibet, are,however, sometimes rich in livestock,since they arrived with herds of yaksand yak-crosseswhich have stillnot yetbeen entirelydepleted. Though mosthouseholds have a fewcattle, pigs and chickens,only around20 per cent. have sheep, and veryfew indeed have yaks. There is virtuallyno wage-labour.Exchange-labour (nga-lak) is by farthe predominant form, though poorer men and women may work forthe odd day or two forpay in cash,grains, salt or garlic.Basically, Lhomi expectto live offthe produce of their farms and to barteror, occasionally,sell the surplus. There is evidencethat the people of theupper Arun valley were richerin the nineteenthcentury than they are today. Referringto the period around i8io, Hamiltondescribes Hatiya on theArun and Alangchang(Wallungchung) on the Tamar Riveras thetwo great'marts' for the important Tibet-India trade, which passed throughnorth-east Nepal. Goods were broughtto the northernmarts from the town of Chayenpur (Chainpur), where therewas a fortand the residenceof the administrativeofficer of the Gorkhagovernment (Hamilton I8I9: I56-7). Hatiya today has no marketof any kind, and Wallungchung, whichwas stillan importanttrade entrepot in theI950's (von Fiirer-Haimendorf I975: I25), was largelyswept away by a landslidearound I970. It is notclear to whatextent the early nineteenth century trade was monetised. Hamiltondistinguishes between periodic markets (hat), where money was used, for which he gives a list for east Nepal, and 'marts', such as Hatiya and Wallungchung,which were not marketsbut entrepots,where goods were exchangedprimarily by barter.Hooker remarksfor Wallungchung in theI840's thatit had no marketor bazaar and thatit was only by threatsthat he could obtain rice there(Hooker I854: 2I8). It is to Sagant thatwe owe the useful distinctionbetween the bazaar, the market and thefair, all ofwhich use money, and the 'mart'which operates primarily by barter.Sagant suggests that Hatiya was probablynot a market(Sagant I968-9: I I I-I2), butagainst this we have the

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 54 CAROLINE HUMPHREY factof its name which was surelyderived from the Nepali wordfor market (hat). In any case, money was of course present in north-eastNepal in the early nineteenthcentury. The headmen ('guobah')of the Upper Arun and Wallungchungcollected taxes on trade-goodsand transportanimals in rupees (Hooker I854: I6o; von Fiirer-HaimendorfI975: i I6). They were politically powerfulenough to maintainmonopolies on tradethrough these valleys and keep strangersout (HamiltonI8I9: I54 and HookerI854:230-I). Also, the wool trade was always monetised.Traders paid money ir advance to the Tibetanshepherds, the advantage of thisfrom their point of view beingthat in thisway theycould stabiliseprices. In theearly twentieth century, the volume of thewool tradewas so greatthat it determinedthe value of Tibetancurrency. If therewas a largewool crop one yearso manyIndian rupees flowed into Tibet thatthe value ofthe rupee would fallas againstthe Tibetan trang-ka. Conversely, ifthe crop was poor, thevalue ofthe rupee rose (Bell I928: I I7). The wool trade becameimportant in north-eastNepal in themiddle of thenineteenth century. What does seem apparentis thatthe monetisationof particularplaces could rise and fallquite rapidlyin thisregion of the Himalayas. Das mentions,for example,that in I9OI therewas onlya cowshedat Gok, northeastof Darjeeling, whereformerly there had been twelveshops (Das I902: 2I). In theLhomi area furtherevidence of previous prosperity is thelarge number of Buddhisttemples (gompa) attached to each village. Many of themare now virtuallyunused, and all arepoverty-stricken. Their lands have shrunkto small plots, and the buildingsare bare shells. The valuable statuesand paintings, whichsome Lhomi maintainedhad beenstolen, were perhaps sold by thelamas, as I was told by others.In thepast, land was morefreely bought and sold thanit is today. The gompas,which acquired substantialwealth from donations, engagedin tradeand boughtland. Temple land was workedby thevillagers in returnfor the religious services of thelamas. Let us look now at themodern barter economy. Before the virtual closure of theTibetan border by theearly I970's whichfollowed the Chinese invasion, the Lhomi engaged in threekinds of barter.First, there was inter-villagedirect exchangeof surplusfoods, handicrafts and itemsgathered from the forests, e. g. maize forpotatoes, wool fortobacco, wild garlicfor rice, etc. This small-scale barterwas carriedout not only between Lhomi villages but perhaps more importantlybetween Lhomi and theirneighbours to thesouth, Gurungs, Rais, Limbus, Bahuns and Chhetris,all of whom are relativelyprosperous grain farmers. Secondly,regular and large-scalebarter of agriculturalproduce took place each yearin exchangefor the produce of the Tibetan livestock economy, butter, driedfats and meats,woven woollen clothes,ropes, sacks, rugs, and blanketsof yak hair.The Lhomi stillsee pastoralproducts as essentialfor existence in their culture,providing their own kindof food, dress,utensils and objectsof value, radicallydifferent from those of the 'Gorkha' (Nepalese) peoplesto thesouth. In theirown eyesthe Lhomi themselves(shing-sa-wa, 'field-earth-people') provide onlya half-lifefrom their farming produce, and theinferior half at that.But the Tibetannomadic pastoralists (drog-pa) retreated en bloc to thenorthern side when theborder was closed,because that is wherethe good high-altitudepastures are.

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 55 Only thesepastures are suitable for herding yaks, yak-crosses and Tibetantypes of sheepand goats. Thirdly,there was long-distancetrade of salt, acquired by the nomads in Tibet, forgrains (mainly rice) acquiredby the Lhomi in the Nepalese middle hills.This salt-graintrade, supplemented by othervaluable items on eitherside, such as medicinalherbs, musk, vegetabledyes, pashminawool, paper,coral, turquoise,silver and gold, was not primarilyfor use but foronward trading. The goods exchangedwere verysimilar to thoseof theearly nineteeth century (HamiltonI 8 I 9: I 56-7). By themid-twentieth century, however, there were no martsor entrepots.Tibetan salt was acquiredin theborder region for x amount of rice and then transportedto the middle hills where it was barteredwith Gurung,Rai, etc. farmersfor y amountof rice.This ricewas thentaken north and barteredagain forsalt at a ratewhich would give theLhomi an operational surplusfor the next cycle. This barterwas carriedout between'known families' and ofteninvolved delay or debt(bulon). There is evidence from the Dolpo region furtherwest that in the early twentiethcentury the salt-grain trade was conductedwith established exchange ratesand an absenceof competition,i.e. it had featuresof 'primitivetrade' (Jest I975: I64). But thiswas fragile.The disastrouseffect on exchangeratios of the flood of Indian salt in the south fromthe I920'S onwards,and of the shutting down of Tibetansalt supplies by theChinese in the I96o's-70's, have been describedby Rauber (I982) forthe Khyampa, a groupof Himalayantraders in north-westNepal.9 An exactlysimilar process was recountedby the Lhomi when I visitedin I979. The resultis thatthe profitable'business' end of their barterhas been drasticallyreduced. It has ceasedto be profitableto tradeTibetan saltin thesouth, and theLhomis fetch it now onlyfor personal consumption and for theiranimals. The main items of 'business' trade fromTibet today are medicinalherbs and musk, which fetchhigh money pricesin the bazaars of southernNepal and northernIndia. Suppliesof othervaluables and foodsfrom the North have dwindledtoo. The Chinese have officialtrading posts where Lhomi can barteror sell grainand hides forsalt, consumer goods, or Chinese currency,but the ratesare less good thanprivate barter with Tibetansin the communesat Lungde or furtheraway at Karta. The Tibetans,however, can onlyopenly trade the personal shares of wool, butter,salt, etc. allowed themby thecommunes; anything else is illegal.Since very few Tibetans now come over the borderinto Nepalese territory,the Lhomi have to travelnorth on foot, carryingheavy baggage over high passes, sleepingin the open or in caves, in orderto obtainthe pastoral products. Formerly they waited at harvesttime for the nomads, who used sheep as transport,to come to them. Now, the most valued and prestigiousproducts in theirculture are in shortsupply. The Lhomi, like the Humli Khyampasdescribed by Rauber,have had to fallback on their own farmsfor subsistence. To them,this means not only back-breaking toil and a wretchedlyreduced diet but also a senseof injustice. In thissame periodthe Lhomi politicalsystem has fallenapart. In the I930's and 40's thevillages had beenmore or less united under one paramountgobawho managed to subordinatethe others. Goba Sinon was born in the village of Pang-Dok, but later moved to the southernmostLhomi village of Syaksila

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 56 CAROLINE HUMPHREY which standson a craggyoutcrop dominating the trail. He controlledthe area fromHedagna to theborder. Here, apparently,he was able to keep controlof all trafficand takea tax on goods movingthrough, as did hispredecessors of the earlynineteenth century. He maintainedhis positionthrough force of armsand was able to keep officialsof the Nepali stateat bay. His power was probably alreadydiminished by the timeof von Fiirer-Haimendorf'svisit in the I950's (von Fiirer-HaimendorfI975: i i6), and today his descendantsare gobas of separatevillages and have littlebut ancestralprestige left. Regionalbarter in theArun Valley has also beenaffected, since the Lhomi are no longerable to give Tibetansalt in exchangefor rice. Many morepeople than previouslynow traveldown to the middle hills for two to threemonths in winter,doing menial work on Gurung,Rai, etc. farmsfor payment in rice, othergrains, hides and animalcarcasses. In principle,the Lhomis' pay, whichis alwaysin kind,can be takennorth again in springfor trade with Tibet during the summer,but in practicemany of them have suchsmall farms that they need it all forsubsistence. The ratesof pay are highlyvariable. If a Lhomi arriveswith a large number of dependantshe may receivenothing for the winter'swork exceptfood and shelterfor his family.Richer Lhomi, on the otherhand, can eitherremain in theirvillages or use thewinter months in tradingventures in the southernbazaars. From thiswe see thatthe large-scaleeffects of a bartereconomy, like any economy based on 'commodity'transactions, may be unequal, even though each particularexchange is by definitionequal in the circumstances.In barter there are no social mechanismsstrong enough to ensure a stabilisationof exchangeratios in the face of externallyproduced changes in supply. This is unlikethe kind of 'primitivetrade' network alluded to by Polanyiand analysed by Sahlins,where, despite rises in theexternal 'prices' of items,they were still tradedfor the same numberand typeof goods in theinterior over a considerable period of time. We also see fromthe annual cycle of productionand exchange thatthe moderneconomists' suppositions about 'transactioncosts' in a barter economyare mistaken. Who produceswhat, where and when, is customaryand well-knownin theregion. The searchfor coincidence of wants is notnecessary, sincethe time and place of barterfor common items was establishedlong in the past. Activitiesare dovetailedand combined,such thattransport, storage and waitingcosts are minimised. The inter-villagebarter of farming and forestproduce has probablybeen least affectedby thedecline in long-distancetrade. It continuesas before,dominated by the barterof Lhomi potatoes (introducedc. I840's, Hooker I854: 240) for grainsfrom the lower villages.In autumna streamof people fromall over the nearbymiddle hills trudge up to theLhomi villagesbearing baskets of grainfor exchange.They tradeusually with 'known people', thoughotherwise wherever theycan. 10 One common argumentof economistsfor the efficiency of moneyis thatit limitsthe numberof price quotationsnecessary. All items can be quoted in money, whereas in a barter system everythinghas to be quoted against everythingelse. In practice,this is not the case. Althoughthere are no clear 'spheresof exchange', many items are never traded for one another.The reasons

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 57 for this are usually purelypractical, e. g. cattleare too valuable to barterfor transportableamounts of grain.Barter, in practice,follows limited and well- known 'tracks'. The Lhomi todayhave relativelylittle everyday contact with money. Chinese moneyis not acceptedin Nepal, norvice versa.But moreimportant than this is thefact that money is not neededfor subsistence farming: the Lhomi do notuse bought tools, fertilisers,etc., and while theywould buy ploughingoxen or cows ifthey could, theaverage farm inherits its livestock. Even stateland taxes, which are small, are paid to thegyembus if necessary in grains.In everydaylife moneyis onlyabsolutely necessary for buying bazaar items such as cookingoil, keroseneand lamps, iron, cottoncloth, and Westernor Ayurvedicmedicines. Lhomi are not yetin thetransistor belt. But once in a whileevery Lhomi needs large sums of money to pay fora wedding,a funeralor a sacrifice.On these occasionsthey take out loans, a subjectto whichI shallreturn later. The nearestbazaar is at Khandbariin the middlehills, some fivedays walk fromthe Lhomi villages. People make occasional tripsthere to sell Chinese manufacturedgoods, medicinalherbs (Humphrey I980; Burbage i98i), eggs, garlic,pigs, bristles,or beer made fromrice or millet.Most people make no surplusmoney from this. Some Lhomi, both men and women, takeportering jobs in the middlehills to earn money,but thisagain has limitedpossibilities: thereis littletime to sparefrom the farms, and theLhomi arein competitionfor thiswork with the Rais, Gurungs,etc. who live on thespot. Also, Lhomidislike and resenthaving to takepaid workof anykind. All of thismeans that the flow of moneyinto the Lhomi villagesis verylimited. From this brief survey we can see some of the conditionsfor a barter economy. (i) There is a low moneysupply, and absenceof marketsin land and labour. Culturally,the people are not 'hooked on' goods importedfrom the nationaleconomy. (2) The region, because of its ecological diversity,has a certainspecialisation of production. (3) Thereis an absenceof state control of the economywhich would siphonoff surplus into a largesystem of redistribution, as was thecase forexample in theInca verticaleconomy. Lhomi political control has now disintegrated.(4) There is no widelyinstitutionalised system of 'gift exchange',nor is theresocially regulated control of traderates, i.e. bargaining is possible forall goods. (5) Barteris carriedout by individualsbelonging to householdswhich do not requiremoney for essential subsistence. (6) Regional productionand exchange is well-knownand predictable,so that thereare establishedpatterns of dovetailingthe timesand places of barter.Partners are knownto one another,especially in delayedexchange (cf. Goodhart I975).

Exchangeratios I now takeup thequestion of barter exchange ratios in greaterdetail, as I believe themto be subjectto differentconditions through the range of products. It was statedabove thatthe possibility of bargainingis a conditionfor barter, but does thisimply that exchange rates are dominatedby supplyand demand? The Lhomi materialseems to show thatthe more widely produced an itemis, theless likelyit is to be subjectto 'interference'with the workings of supply and

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 5 8 CAROLINE HUMPHREY demand.This appearsto be onlycommon sense. It is also a pointwell-known in 'economicanthropology', e.g. Bohannan'slowest 'sphereof exchange'among the Tiv (Bohannan& Bohannan I968). However, we can only say thatbarter ratiosfor such productsappear to correspondmore or less with what a 'market equilibrium'rate mightbe. In fact,of course, no such rate is visible. It is somewhatcomplex to demonstratethis, but let us startwith maize-salt barter in autumn I979. Maize becomes cheaperin relationto salt as one travelssouth- wardsin theArun Valley.

Kimathanka(north) i maize = i salt PangDok i maize= i- salt Syaksila(south) i maize= 1/3salt11 This in itselftells us littlebecause these ratios might indicate that maize becomes scarceras one goes north,or thatmaize is equallycommon in all of thevillages, butthat salt, which we know is obtainedin Tibet,becomes progressively more expensivethe furtheraway it is bartered.Or theymight indicate both. Indian saltcan be boughtin Khandbaribazaar for money. In Pang Dok thetwo kindsof salt have more or less equal value againstother things, despite the factthat Tibetansalt is preferredfor its taste.The explanationfor this is thatIndian salt has to be paid forin moneyand can onlybe acquiredsome six dayswalk away, while the borderwith Tibet is only two days away. The supply of maize is equally problematic.No adequate surveyof the countlesstiny fields of the producershas everbeen made. We can besttry to explainthe ratios by looking at maize in relationto anotheritem, e. g. potatoseed.

Kimathanka i maize = 2 potatoseed Pang Dok i maize = I 1/2potato seed Syaksila i maize = i potatoseed Angla (nearKhandbari) i maize = 1/2potato seed We see thatmaize variesagainst potato seed as it goes northin much thesame proportionas it does againstsalt (maize become threetimes more expensive againstboth saltand potatoesfrom Syaksila to Kimathanka).This concurrence would suggestthat maize does actuallybecome scarcerthe furthernorth one goes, and thiswould seemto be confirmedby Lhomiimpressions of the amount of maize planted.The rateschange from year to year:in Pang Dok in a yearof poor maize harvest2 unitsof potatoesobtain only i of maize, but 3 in a year when maize is plentiful.All of thissuggests that there is a grosscorrespondence betweenbarter rates and supplyand demand. Withineach village,barter rates for these basic productsare consistentwith one another.I shall give here rates for I979 in Pang Dok as statedby the villagers.This may be confusingto read but it is an accuraterepresentation of ethnographicreality. The data are re-expressedin termsof i unit of rice as numerairein the table. However, it is importantto realisethat Lhomi do not thinkin termsof a numeraire.

3 potatoseed = 2 maize / millet i salt = 2 maize / millet

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 59

Ifrates were consistentwe would expectto findthat i saltwould exchangefor 3 potatoseed, and thisis confirmedfrom the data: I salt = 3-4 potatoseed. Rice was preferredto maize or milletfor food, but fetchedalmost the same rate in barterbecause it was possibleto exchangeit in Tibet at equal rates. i rice= 1/2salt Consistentwith this is: i rice= i + maize/ millet 2 rice = 3 potatoseed Workalso enteredthe barter economy: i day's labour = 2 unitsmaize / millet hirepair ploughing oxen per day = 4 dayslabour hirepair ploughing oxen per day = 8 unitsmaize /millet

Barterrates in Pang Dok in I979 expressedin one unit(kathi) of rice. * representsabsence of transaction. Rice Maize Millet Pots. Workday Oxen hire Salt Rice * I+ I + I 1/2 1/2 1/8 1/2 Maize * I- I 1/2 1/2 1/8 1/2 Millet * I 1/2 1/2 1/8 1/2 Potatoes * * * 23 Workday * 1/4 Oxen hire * * Salt *

Thereis consistencyin theinternal village barter exchange ratios for common produce (e. g. all grains,salt, chillies, garlic, vegetables, butter and local meat) which would suggestthat 'market forces' are at work here. But, otheritems, such as clothes,Tibetan woven aprons,boots, yak tails,Tibetan driedmeat, rugs, gem stones,etc., which are barteredless frequently,exchange for one anotherand for the agriculturalproducts on an ad hocbasis and withoutany consistency.This can be explainedconventionally by lack ofinformation about comparabletransactions, the cost of obtaininginformation being greaterthan the estimatedsaving of a lower price (the 'thinmarket'; cf. Stigleri96i) but I would account for it ratherby the social practiceof barter:some degree of comparison,and hencecompetition, is unavoidablein frequentlocal exchanges, but not in distantor infrequentbarter. But in eithercase it is not lateral comparison,but the processof bargainingbetween two people, which deter- minesthe 'price'. But let us look more clearlyat what is really going on. There are two importantpoints: (a) the lack of standardunits of measurement,and (b) the hiddeninequalities behind the 'traditional' places and timesof barter. In theUpper Arunthere are threedifferent scales of measurementoperative: (i) Nepali measuresof weightand volume; (2) Tibetanmeasures of weight and volume; (3) Lhomi measuresof volume, themain one beingthe kathi, a round pot with straightsides. While Nepali and Tibetan weightsand measuresare moreor less standardised,this is farfrom true in thecase ofthe Lhomi ones. (Jest

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 60 CAROLINE HUMPHREY

(1975: i68) remarksfor Dolpo, 'A chaque maison, sa mesure'.) The kathiin particularis oftendescribed as 'a largekathi', 'a smallkathi', etc. Some goods are commonlymeasured in one seriesand othersin another.Lhomi use different seriesat theTibetan border, in theirown villages,and in thesouth. This makesit difficultfor them (or us) to compareexchange rates for different goods. This is an importantfactor behind the absence of a commonstandard of value in barter. Barter,though equally 'commodityexchange' is radicallydifferent from the monetarymentality. Most Lhomi could only make approximateguesses as to the conversionof one set of measuresinto another.The same lack of abstract measurementapplies to productionas well as exchange.Lhomi usually estimate fieldsizes by the amountof x or y grainthey sow in them.But thisis largely guesswork,as theyrarely weigh the seed and almostnever weigh the harvest. 12 No doubt, ifthey sold theirharvest in a bazaar theLhomi would weighit. But theydo not, theybarter it. This means thatit is impossiblefor them to assess 'profit'against costs of production. Let us takethe example of a poor farmer,Kun-top, from Pang-Dok. His main crop was potatoes,which he grewfor his own consumptionand forbarter. He obtainedin exchangerice, maize, milletand chillies,mostly from people from outsidethe village. He knew that3 kathisof potatoeswere equal to 2 of maize thatyear (I979) and thatmillet and maize were thesame 'price'. However, he actuallybartered his potatoesin a unitcalled a tobo,a kindof largebasket. He thoughtthat one tobo was equal to about 30 kathis,but some were up to 40 kathisin size. His own tobo was a smallone, around20 kathis,he thought.What he was certainof was thathe aimedto bartertwo tobos ofpotatoes that year and thenhe would have enoughgrains for his consumption. In barter,the participantsonly have to examine what is offered.In the monetisedeconomy, on the otherhand, it is not theamount of goods but the moneyprice which fluctuates. Thus theLhomi reach an impassein thebazaar of Khandbari.No-one in the bazaar truststhe kathi,and the Lhomi therefore,if theyare to sell anything,are requiredto hirea Nepali measureof volume, the kurwa.Prices in thebazaar are well-knownand thereis a certainsocial pressure not to raisethem. 13 Lhomi grumbleat thisand say thatthis relative inflexibility is one of themain reasons why theyprefer barter. The effectof theabsence of exactmeasurement is to makevillage barter rates as quoted by the Lhomi more or less notional, almost ideological. They representmerely the beginning of the bargaining process. As I have suggestedin thecase ofmaize, the state of supply of products can onlyreally be guessedat. In barter,what people knowis demand,and themutual adjustment of demands is a social relation.For thisreason, the most importantfact about barteris thatit takes place betweenindividuals who are socially'understood', whether they belong to the same or different'ethnic groups'. Part of the calculationis the extentto whichpeople can trustone another. Let us look now at relationsof dominationin barter.14 In thecase of common produceperhaps any disadvantage (in transportcosts, timing of the transaction, etc.) takesan 'economic' formwhich can be adjustedvia exchangeratios. But in thecase of goods forwhich something like a monopolycan be maintained,for examplethe vestiges of the wool tradein theUpper Arun,the customary 'track'

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 6i of bartermight seem to establisha persistentlyunequal relation between ethnic groups,such that one sidecan dictateterms of trade. It is significantthat it is here thatrelationships of debtalso flourish.

The wooltrade: emergence ofa monopoly By thelate nineteenth century the trade in wool fromTibet was one ofthe most importanteconomic activitiesin theregion. Certainly by thebeginning of the twentiethcentury and probablyearlier this trade was monetised(Bell I928: I I7- i 8). Since theLhomi always obtainedmost of thewool fortheir own use from the exchangewith the Tibetan nomads theyhave never owned many sheep themselves,nor did theyplace muchvalue on theirown inferiorpastures on this side of the Himalaya. At the beginningof this centurytheir leaders sold the rightsof use in thepastures to theGurungs. The positiontoday is thatGurung sheepswamp theonly available pastures, which are Lhomi pastures. Sheep in the Upper Arun todayare of theNepali Baruwal variety,different fromthe Tibetan sheep, and Lhomi have no expertisein thewashing, dyeing, spinning,and weavingoperations for wool of theBaruwal type.The Gurungs, on the otherhand, are the regionalspecialists in the manufactureof Baruwal woolen rugs(radi), jackets, capes, etc. Virtuallyall Gurunghouseholds spin and weave, even if theydo not own sheep. The few Lhomi who own sheep hand over all oftheir wool to theGurungs, either to women who come up to fetchit, or to shepherdspassing through Lhomi villageson theirway to and fromthe pastures.Lhomi take money, or goods measuredin money, fortheir wool. They can expectto gain fairlylarge sums (some 350 Rs forthe wool fromten sheepin a year).Wool pricesin rupeesreflect transport costs to Gurungvillages, and are identicalfor everyone. Although they have risenin the last fiveyears (I975-80) by about 1/3, thisis only keepingstep with the rise in wool pricesin Kathmandu(see note 20). But in factit is relativelyrare for Lhomi actually to sellwool in a simultaneous transaction,since theyhave usually takenout a loan in advance againsteach season's wool crop. The Gurungfarmers know exactlyhow manysheep each ownerhas, and theyare preparedto pay in advanceat slightlylower ratesthan normal,a time-honouredpractice in theHimalayan wool trade.Other families can obtaina loan of 3oo Rs in advancepayment for herding a Gurungflock for a season. But most Lhomi are forcedat one timeor anotherto take out money loans toutsimple. Interest rates are veryhigh: 5 RS a monthor 6o RS a yearfor a loan of I00 RS, witha surcharge(tekki) of I2 RS.15Avoidance of repaymentis frequent.Such loans are veryrarely taken out by one Lhomi fromanother and thenonly witha good deal of persuasion.In such cases loans are oftenin grain ratherthan money and therates are lower (25 RS a yearon I00 RS loan, witha tekkiof 5 RS or one day's labour). The givingof loans is accompaniedby a blessingfrom the donor. Loans are usuallytaken out at lifecrises, when it is necessaryto buy in pigs, chickens,liquor, etc. forfeasts. But theycan also be used, like othervaluables, such as jewelleryor livestock,to startoff a trading venture.Such loans and valuablesare thespring which enable a fewindividual Lhomi tojump clearfrom the maze ofpetty barter operations.

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 62 CAROLINE HUMPHREY It is theGurung farmers with an interestin thenorth who giveout loansand it is apparentthat the operation is notunconnected with the pasture situation. The Lhomi have begun to make claims thatthey are legallyentitled to have their pasturesback. In any case, Gurung shepherdsare dependenton Lhomi for allowing thempassage, forstoring provisions, and givingshelter in thenorth. The ultimateownership of the pasturesis the lever which Lhomi can use to counterbalancethe Gurungwool trademonopoly and to ensurethe continued availabilityof loans. Lhomi avoid takingloans frompeople theywork forin winterand theyspread the number of creditorsas widelyas possible,i.e. they are carefulnot to become too dependent.Lhomi maintaina defiantlyseparate culturefrom the Gurungs. Nevertheless,it is perhapsthe increasingGurung control of the money supply to the Lhomi which is reflectedin the recent electionof a Gurungas a panchayatofficial for a Lhomi ward.16 I have been discussingthe wool tradein thecontext of 'barter'and thismay seem odd in view of the factthat even Lhomis sell wool for money. In fact, moneyitself enters the barter 'system'. Money has a differentvalue in differentvillages, in termsfor example of rice or labour.17 It also has a differentvalue in thesame village in relationto one type ofproduct rather than another, just as othervaluables do. 18 The variationin the value of moneyin differentvillages could be explainedsimply by the money supply. We could rememberhere the situationdescribed by Bell for Tibet in the I920S (see p. 54): in years when therewas plentifulwool for export, rupeesflooded in, and thevalue of Tibetangoods againstrupees went up (i.e. they cost more rupees). On this analogy we could explain the high money price of rice in certainLhomi villages by the greatervolume of sales from these villages resultingin a largermoney supply in them. But in this case, the money pricesof all goods would be highin thesevillages. But thisis not so. The explanationof thefacts lies in thestrategic position occupied by various goods and moneyin the differentvillages in the bartersystem. In Pang Dok, although a relativelyfertile village, people simply do not normallyaccept moneyfor rice. Rice can be used bothas highprestige food and as a bartergood in thetrade with Tibet two dayswalk away. Money, on theother hand, is only usefulin the bazaar, five days arduous walk away. Potatoes and garlic are occasionally sold for money in Pang Dok because they are produced here specificallyfor trade, and are no use in theTibetan barter. The same is trueof wool. But people herewant wages in grains,not money. Contrary to orthodox economicviews on money,even in termsof exchange-value, it is muchbetter in the'sink' to be paid in grainsthan money. The factthat wage labouris normally paid in grainsthus reflectsthe relativelystrong bargaining position of em- ployees. This is what one would expect in a village with so few landless households. Unfortunatelywe do nothave thedata to discussprices and exchangeratios in relationto productivecapacities and regionaldemand. However, it is possibleto make some generalconclusions. The amountof grainor moneypaid in wages does not correlatewith the fertility of the location (Syaksila is a poor villageand yethas highgrain wages and low moneywages17), norwith an equal subsistence

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 63 levelfor workers in all villages(the difference between Pang Dok or Chemtang and Syaksilais considerable).Essentially, money, grain, and work arebartered forone anotherin each location.Lhomi do not exploitthe differences between villages.The extentof lack of consistencyin ratesis an indicationof thelack of integrationof societyand henceof theeconomy.

Barterin society and culture Money is not used to measurethe common productswhich are bartered.A corollaryof thisis thatsmall amountsof moneyare not added up or saved to acquirethings. The 'windfallmentality' of Lhomi with regard to moneyis often commentedupon by theirneighbours. If they have money,they spend it, often on drink.If they need a largeamount of money, they borrow it or sellsomething so as to obtainthe amount they require. I cannotpossibly agree, for such a case, withthe economistswho argue thatbarter gives riseto monetisation.On the contrary,money has a tendencyto sinkinto barter. Of course,money is not always,or necessarily,used in this'barter' fashion. Lhomi can operatein thebazaar as cannilyas anyoneelse, not that the bazaar is a perfectmarket by anymeans (Sagant i968-9;Jones i980). But insidethe Lhomi economymoney 'disappears' because it is notaccumulated as a goal ofwealth. A Lhomi counts himselfrich by virtue of ownershipof land, livestock,and valuablessuch as jewellery.A richman or woman is honouredfor sponsoring specialrituals, requiring beer, grains, butter, meat, etc., at thenumerous gompa temples.But virtuallyno-one can affordto do thisnow. The richerLhomi, of thejimi(original) clans, are membersof associations attached to templesfor the paymentof the costs of regularmonthly rituals. These associationsare egali- tarianin thesense that a groupof wealthy families takes it in turnto providethe expenses,and all membersof the village participatein the feastwhether they have made provisionor not.19 Money thustends to be rapidlyconverted into goods. Recentsevere inflation in lowland Nepal may well play a partin this.20 The expectationthat in thefuture one will be poorer,not richer,and thatother people also will be poorer,which is theLhomi experienceof thepast decades, inclinespeople to preferthe immediacyof barter.Poor people, indeed, who experiencehunger in bad years,have littlechoice. In the past therewas not such a markedtendency to convertout of money. This is becausetwo kindsof money, as metallicvalue and as statecurrency, were stilllinked. Until quite recentlythe Indian silverrupee had an intrinsicmelt down value, as did variousTibetan and Chinesecoins (up to I939 therupee was 92 per cent.silver, thereafter 5o per cent.until the end of Britishrule). But this was less importantthan theircultural role as signifiersof 'treasure'.Among ordinaryLhomi, silvercoins were melteddown to makebracelets or belts,but oftenthey were used injewelleryas coins.All Lhomistill use old Indianrupees as indicationsof wealth, sewn onto women's hats. In Khandbari there is a flourishingtrade in old coins,and Lhomi craftsmenknow how to weld ringsto them and polish themup fordisplay purposes. Goodhart'slogical deduction that the absence of monetarypayment as the normal form of economic transactionwould requirethe demonstrationof credit-worthinessseems to be

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 64 CAROLINE HUMPHREY borneout by theLhomi practice-paradoxicallyby theuse of 'money-turned- into-treasure'as a signifier. To some extentmodern Nepalese currencystill retains an internalrole as a 'valuable'. This is at odds withthe mentality of the bazaar, but less so eitherwith barteror with ceremonial-statuspayments. A ritualgift of 6i RS is still an essentialtransfer from the bridegroom'sside to the bride's at marriage(a transactionbetween groups of differentstatus). But in barterthe money exchangedfor a valuablecomes itself to have somethingof the valuable about it. This is because the mentalityof barteris one of equal exchange. In a fully monetisedeconomy it is betterto hold money than goods. But in a barter economy,having money rather than goods mayor maynot place someoneat a directadvantage, and theLhomi always behave to one anotheras ifmoney is just another,rather honourable, 'good'. This mustbe determinedby a psychology tendingtowards 'immediatereturns' (cf. Woodburn I980), since Lhomi dis- regardwhat we know fromAristotle: that 'even if we happento wantnothing at the moment,money is a sort of guaranteethat we shall be able to make an exchangeat any futuretime when we happento be in need' (I895: 5' I4). It is perhapsbecause of the impossibility in barterof calling upon an abstractly expressed'just price' (this illusion being createdby the idea of money as a measure of value) that bargainingitself is decorous, markedby silences or absences, as people retreatto 'consider'. We may compare this with the confrontationalmode of higgling described by Geertz for the monetised Moroccan suq, wherethe intensive determining of therealities of theparticular deal coexistswith some sortof extensivesearch for the going pricewhich also seemsto be the'fair price' (1979: 206-7). Inthe suq the 'generally acknowledged' and the'equitable' are the same, though admittedly it is a vexed and problematic matterto establishwhat thismight be. But the Lhomi do not even try.They ensurethat the reference to 'fairprice' becomes an impossibilityby theirattitude to money and to measurement.Barter thus abolishes abstract justice (and the need to appear'honest' in relationto generalor externalcriteria), and replacesit withwhatever is consideredby theparties to bejust in theircircumstances. This is arrivedat by bargaining.This same featureis truein a somewhathidden sense of thesuq, but in barterit becomesovert and dominant.We could see thisagain as a featureof atomisedsociety. Among the Lhomi thereis a continuumfrom petty money transactionsor directbarter, though regular and thendelayed trade by barter,to theoccasional 'big business'.All thiscan be contrastedwith those transfers of value whichare concernedwith reproducing differences in statusbetween clans and maintaining hierarchicalinstitutions (see Clarke I95I). But thesekinds of paymenthave declinedto a minimumamong the Lhomi today,and the questiondoes arise whetherthis is a simpleresult of increased poverty or somethingin thenature of a 'social choice' engenderedby ideological change. I do not at presenthave enoughinformation to disentanglethese two factors.Certainly, egalitarianism is whatthe Lhomi todayadmire. People even grumbleabout others who keep a servant.It is notablethat Lhomi egalitarianismincludes relations between the sexes. Women have never previouslybeen eligible for inheritedpolitical position,but today several of the elected representatives in thepanchayat system

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 65 are women withtheir own households.Two of thesewere ebullientpersonal- ities,not especially rich, but 'merry widows' withseveral lovers. A wealthyand influentialman, on the otherhand, told me thathe would not standfor the panchayatbecause office was heldonly for a shortperiod. In sum,the Lhomi are in the process of reformulationof politicalvalues, to some extentinduced by theirgradual incorporation in thestate of Nepal, butalso fuelledby an anarchic individualismof theirown. It seemedto me in I979 almostas though,alien to Westernpractice, economic transactionswere undertakenwith seriousness borderingon solemnity,while politics, at anyrate of the panchayat, was a realm ofdisrespect and hilarity.Going down thetrail to attendmeetings, the closer the Lhomi deputiesapproached the 'Gorkha' centreof Khandbarithe more out- rageoustheir behaviour became. For theLhomi, thepinnacle of barteris theoccasional big businessdeal. This is normallyconducted between 'known people', oftendistant kin, or 'ritual friends',21 or kinsmenof ritual friends. A deal is oftenpreceded by thegiving of minorsolicitory gifts and a ritualscarf. The arrangementsrequire discussion and time.Much of thistrade is illegaland so thepartners have to trustone another. The watch-muskbarter trade, between India and Tibet,is themost valuable of all. The same channelsare used forlesser valuables, such as Tibetancarpets or religiousobjects. They go, perhapsvia southerntowns such as Dhankuta or Dharan, to Lhomi contactsin Darjeelingor Kathmandu.A tradechain of this kinddoes not consistof people tryingto outwitone another,but of friendsand accomplices. It is a gestureof friendshipto be asked to participate(unlike Westernsociety, where 'friendship'has a differentmeaning and is considered inimicalto business). In all thisthere are, and can be, no set prices.The only guide is the distant chimeraof a notionof what themarket in theend will bear.The ultimatebuyer is part of the capitalistsystem, Indian traders,French perfume houses, West Germancarpet importers, or Westerntourists. This is the part of the Lhomi bartersystem which is leastsubject to the 'traditional'channels of theregional productiveeconomy: prices can varywildly and yetit is themost involved with relationsof friendshipand trust. Sahlins made the perceptiveremark, for 'primitivetrade', that where rates are more or less fixed, the only way to adaptto changesin supplyand demandin thelong runis to revisethe partnerships rather thanthe rates (I974: 3 I2-I 3). Again it appearsthat the 'barter economy' differs radicallyfrom 'primitivetrade' systems.Exchange rates in bartertrade are decided upon separatelyfor each transaction,i.e. thereis no expectationof a 'standardrate' over time.This does not,however, have theeffect of destroying tradepartnerships, which often include prolonged periods of negotiation.The time allowed forrepayment, in the case of a delayedtransaction, can be very elasticand much depends on the energyand social weightof the individual. Some people makes disastrouslosses and theirpartnerships lapse, not because thesocial relationships have ended,but because they have, for the time being, no tradeto conduct. I knew of one man, Lhagpa, who arrivedfrom Tibet with threeyaks, worth several thousandrupees, which he entrustedto a distant Lhomi Khampa relativefor a tradedeal. The businessfailed. Five yearslater, Lhagpa, a mildman, was stillwaiting for the Khampa to makea successfuldeal

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 66 CAROLINE HUMPHREY and repayhim. Lhagpa moved to Kathmanduand tookpaid work. Livinglike a churchmousewith the friendsof kin, he spenthis firstmonth's wages on a second-handsleeping-bag which he sentoff, via a trustedfriend, to the same Khampa fora deal which was to involve medicinalherbs. Some monthslater when I leftKathmandu the outcome of thistrade was stillnot known. There is no radical break between this trade kind of barterand the rest. Essentiallyit is still conductedwithin the community,and the commercial biting edge, the final sale, is done by middlemenof other ethnicgroups. Economically,everything depends here on theinternal gains of each transactor bearingsome relationto theoutside price. Lhomi have fairlygood channelsof information,but in factthey are not very good 'big businessmen'precisely because theythemselves seldom have theopportunity or boldnessto make the finalsale (in this they are much more retiringthan other'Bhote' groups of Nepal). Losses are absorbedinto the complex web of unfinishedtransactions which link exchangepartners. It is highlyprobable that there is not enough wealth around for these ever to be resolved in the foreseeablefuture. The consciousnessof thisis one factoramong many thatmakes Lhomi refuseto give one anotherloans and to prefer,whenever possible, the simultaneous transaction.

Conclusion The conditionsin whichbarter emerges as a dominantform of transactionare discussedon p. 52. Here I mentioncharacteristics of barterwhich appear to be important. i) The relationbetween actual exchange ratios in barterand presumed'equilib- rium rates' (i.e. the prices representingequilibrium between supply and de- mand, cf. Marshall I890: AppendixF on barter)depends on the volume and frequencyof exchangeof the itemsin question.Goods widely producedand barteredare likely to be exchangedat ratios most nearlyapproximating to notional 'equilibriumrates', while those infrequentlytransacted may be bar- teredat quite arbitraryrates. Among the Lhomi, money fallsinto thislatter categoryin thelocal economy (the essentiallynon-monetised 'sink' in Crump's terminology).In effect,the practiceof barterincluding money renders specu- lationabout 'equilibriumr4tes' pointless. The notionof a general'just price'in Aristotelianterms is thereforeabsent. 2) One reason forarbitrary rates lies in the factthat barter is immediate,and being immediateis by its verynature discontinuous. For relativelyinfrequent exchanges people do not concern themselveswith what prices might be elsewhere.Even when barteris delayed and part of a complex of debts, the participantsdo not comparetransactions with one anotheror againsta standard ofvalue (see also Leach I954: I45-6). 4) The requirementof immediate satisfaction of demand is paramountin barter systems.This is one of theexplanations of therejection of monetisation.Very poor people literallycannot afford to keepwealth in money.In thissituation, the absenceof exact measurement is nota featureof low culturaldevelopment but a deliberatestrategy to ensurethe persistenceof one-to-onetransactions. The

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 67 subjectionof moneyto thesame rationalityindicates the strength of thebarter system.Here even money has no temporality.The 'windfallmentality' thus does notindicate irresponsibility, as thestereotypes of other ethnic groups about theLhomi maintain,but a desirefor freedom. 5) Immediatebarter is an expressionof autonomyand thereforeis likelyto be insistedupon in relationswith outsiders. Direct barteralso dispenseswith the need forpayment, i.e. it will be used when thereis littleinformation about the credit-standingof purchasers, or whenthere is a lack oftrust. These advantages of directbarter are so greatthat they outweigh the transaction costs involved. Thus we findthat Gurung farmers, who needseed potatoesand do not trustthe Lhomi to bringthem down in timefor the planting season, will trudgefor days up and down thecraggy Arun slopes to getto theLhomi villages for direct barter of theirgrains for potatoes. 6) Delayed barteroccurs where credit is required,and thereis a correlationwith themore valuable goods becausethese also tendto be less frequentlytransacted (see Douglas and Isherwood, I978). However, delayed barter,as Goodhart pointedout, can onlywork in a relationshipof knowledge and trust.In thecase of the Lhomi, despite the existenceof tradingfriendships with otherethnic groups,this feature limits trade networks and resultsin a curtailingof tradein valuablegoods. The generalimplication of Marx's workis thatwealth accrues throughcontrol of production.Here, however,we see thatthere are ways of profitingthrough strategic domination of barter-trade, the control of the means of exchange. However, if barterrather than monetarysale is the means of exchangewe can see thatit is self-limiting:the trust required for barter to include crediteffectively circumscribes economic operations. 7) The preferencefor barteris thereforerelated to underlyingpolitical con- ditionswhich make autonomya paramountconsideration. The Lhomi have entereda cycle of internalpolitical disintegrationsuch that the economic transfersreproducing hierarchy, whether in the contextof affinalrelations, chiefship,or the Lamaist church,have all but ceased. Many differentsocial groups (villages, clans, old settlers/newsettlers, lamas, households, etc.) coexist and do not wish to be subjectedto one another.This situationhas complex antecedents,both in the earliereconomic collapse of the salt-grain trade, which drasticallyreduced the amount of surplus the Lhomi could produce,and in theadvent of Nepali politicalinstitutions, which are displacing theLhomi ones. In relationto outsidepolitical power, theLhomi stillwish to maintainan autonomyallowing themsome degreeof controlover the Arun routeand theirmonopoly of theTibetan trade. 8) While it is truethat barter is conductedin an egalitarianmode, and most customaryor relativelylong-standing barter arrangements adjust exchange- ratiosto transactioncosts, neverthelessin the long run barter,unlike certain systemsof 'primitivetrade' described in theliterature, cannot save people from the effectsof externalchanges in supply and demand. In an atmosphereof perfectfriendliness, as betweenthe Lhomis and theirrespected Tibetan nomad friends,exchange rates may stillinexorably drop. Indeed,barter, by hindering the emergenceof integrativeeconomic institutions,may make the situation worse. Barteris a responseto increasingpoverty on thepart of people who wish

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 68 CAROLINE HUMPHREY neverthelessto maintaintheir autonomy. Thus, while preservingthe same 'benign' cultureof economic independence,barter may actuallypreside over realeconomic desperation and instability.

NOTES The fieldworkon which thisarticle is based was carriedout in collaborationwith the Kosi Hill Area RuralDevelopment Programme in Nepal in I979. I am gratefulto theNepali coordinator,Mr Kayastha,and to DickJenkin,Jit Man Gurung,Prem Subba, Samdup Lama, JitBahadur Gurung, AmbarBahadur Rai, PremDutt and Dorje Bhotefor the very great help they gave me. Sean Conlin, Edgar Keller,Jonathan Zeitlin, StephenHugh-Jones, Samdup Lama, Simon Strickland,Lionel Carterand Mrs TseringSangmo broughtmaterials to myattention during the writing of this article. I am indebtedespecially to Quentin Outram and IstvanHont, and also to Edmund Leach, Keith Hart, StephenGudeman, Wynne Godley, ParkerShipton, Charlotte Hardiman, Paul Sant Cassia and Graham Clarke for discussionsrelating to barter.I wish to thank my mother,Margaret Waddington,without whose supportI would neverhave writtenthis article. 1 Crump in his surveyof theliterature could findonly three 'primitive' economies dominated by barter(I98I: 54). 2 'The logic ofthe case is, indeed,almost the opposite ofthat underlying the classical doctrine. The orthodoxteaching started from the individual's propensity to barter;deduced from it thenecessity of local markets,including division of labour;and inferred,finally, the necessity of trade, eventually offoreign trade, including even long-distance trade. In thelight of our present knowledge we should almostreverse the sequence of the argument: the true starting point is long-distancetrade, a resultof thegeographical location of goods, and ofthe "division of labour" givenby location. Long-distance tradeoften engenders markets, an institutionwhich involves acts of barter, and, if money is used,of buyingand selling,thus eventually, but by no meansnecessarily, offering to some individualsan occasion to indulgein theiralleged propensity for bargaining and haggling'.(Polanyi I957: 58). 3 I use the term'Lhomi' forthis community in the Upper Arun because it is establishedin the literature(von Furer-HaimendorfI975; Bista I967). In factethnicity in theregion is complexand shifting.The Lhomi use bothBhote (local Tibetan)and Nepali namesfor people and places,and this is reflectedin thepaper. 4 For example,in cases ofhyper-inflation, or thequite different situation where money is avoided in orderto evade theattention of taxinstitutions of thestate, as in present-dayCalifornia. 5 As in socialisteconomies where scarce goods are allocated,often without the use of money,to social categories,and frequentlypass from these groups to the populationat large via barter transactions. 6 Crumpnotes that, quite apart from the functions of money mentioned by Clower, 'truemoney' can onlyexist under certain real conditions: it mustbe physicallycapable of circulatingindefinitely (i.e. it links the presentto the future);it must have a distinctiveidenity as money and no real importancefor non-monetarypurposes (e.g. it should not be consumable);and it should be in limitedsupply. If theseconditions do not obtain the money-stuffcan appear and disappearin an uncontrolledway whichplays havoc withits role of storingvalue, etc. For example,cowries, the mostwidely used 'truemoney' of the specie (non-manufactured) type, never circulated as moneyon theshores of the Indian ocean wherethey could be simplypicked up on thebeach, but only in distant places wherethe supply was controllable(I98I: 85). 7 It mightappear that Goodhart,in suggestingthat barter takes place insidethe face-to-face community,is reversingMarx's well knownproposition that the origin of commodityexchange is thebarter of products between communities (Karl Marx I887 vol. I: 9I-2). Only in dealingwith the outside world, Marx wrote, are individualsfreed from the social obligations('property rights') which rendertheir goods inalienableinside theirown communities.Commodity exchange, i.e. transactionof alienableproducts, therefore has to takeplace withoutsiders. This idea has takenroot in anthropologyand been elaboratedin many influentialtheories (Sahlins I972; Servet I98I-2; GregoryI 980). Goodhart,however, really agrees. He is sayingthat barter as simultaneousexchange is so difficultas to be virtuallyan impossibilityand thatwhat would occuris delayedexchange, i.e. creditand debt. Comparativelylittle of themovement of goods withina communitywould use the pricemechanism (i.e. barter),as opposed to internaldistribution by direction,and thesefew debts

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 69

would be settledin thecontext of generalmutual knowledge. Transactions outside the community would be forcedto use money.Although we neednot agree with him about the virtual impossibility of barter,it is clearthat in bothcases we aretalking about 'commodityexchange' in Marxistterms. 8 Lhomi own land in 'kipat' (communal) tenureas definedin Nepali law (see Caplan I970: McDougal I979). Kipat tenureis in theprocess of beingtransformed into individual rights, but the legal situationwas uncertainwhen I visitedthe Lhomi in I979 and almostno land was boughtor sold. The main reasonthe Lhomis gave forthis was theunhelpful attitude of Nepali officialsat the landregistration office where Lhomi feelat a disadvantage,as veryfew of them are literate in Nepali. Land is in facttransferred mainly by bandhaki(Nep.): grainor moneyis givento theowner in return foruse, and ifthe payment is not returnedby a specifiedtime the land is acquiredby theuser. 9 Raubersays that, around I900, one unitof rice obtained I0 ofsalt at theTibetan mart of Purang, while it got 1/3 of salt in the districtof Bajura some 50 miles to the south,and only 1/5 or 1/6thin Accham,another 20-30 milessouth. With the infiltration of Indian salt into the hills from the I920'S onwards, exchange ratios of Tibetan salt with rice declinedsteadily in the southerndistrict of Accham: I900 I rice= 1/5thor 1/6th salt I925 I rice = 1/4salt I935 i rice = 1/3or 1/2salt I965 I rice = i salt By theI 960's, theeffect of theChinese invasion of Tibet began to be feltat theother end of the trade route,and theamount of saltgiven for rice at Purangwent rapidly down: I964 I rice= 8 salt I974 I rice= 7 salt I976 I rice = 4-5 salt I 977 I rice= 3-4 salt In I 977, one unitof rice could stillbe barteredfor only one of saltin thesouthern districts of Baljura and Accham (Rauber I982: I 5 I). 10 Baumgartner(I980: I35) notes thatSherpa elders of each villagemeet annually to discussthe exchangepotential of theirpotato crop and thenset a minimumexchange ratio as againstgrains for thatyear. There is a constantdanger that richer households would agreeto lower rates. 11 These are the acceptedannual ratesin kathis.There was no mentionof changesduring the season, whichsuggests that something like theSherpa practice may have beenin force(see above). 12 HenryOsmaston (personalcommunication) notes a similarlack of precisionin measurement among Ladakhis, despiteknowledge of standardTibetan rates.Harris suggests, in a paper with manythemes in common withthis one, that'measurement by volume ratherthan by weightis an indexof thedegree to whichmonetary equivalence in exchangeis avoided' (HarrisI982: 77). 13 Bazaar pricesare fixednationwide by the Nepali governmentfor paddy and wheat, and in Dhankuta thereis also pricefixing over a wide rangeof basic productsby a groupof largetraders. Sellersare preventedfrom trading at lowerbut not higher prices. The reasonfor this is thedesire of tradersin general to keep purchasersfrom travellingdown to the Terai where prices would otherwisebe lower (Joneset al. I982: 33-5). Althoughsuch institutionalisedprice fixing does not occurin Khandbarithe same factorsapply there. 14 The question of 'unequal exchange' in partiallymonetised peripheral economies has been discussedby Platt(I982) and Harris(I982). Plattmaintains that the very features which look least affectedby thecapitalist economy, such as paymentof wages in maize, ensurethe reproduction of households lacking maize production.Low agriculturalprices in marketsresult in an unequal exchangebetween the peasants and therest of theeconomy. The effectis to transfervalue, but the peasantsare not therebypressured to leave the land. The resultis the maintainingof a reserveof labour availableat timesof seasonal demand. This argumentclearly has relevancefor the Lhomi. However, in thisarticle I am discussingthe question of 'unequal exchange' between Lhomi and their immediatetribal neighbours, rather than between the Lhomi and thecapitalist economy as a whole, mainlybecause thereare so manysteps between the Lhomi and anythingresembling capitalism in Nepal. 'Unequal exchange'is not a phenomenonof capitalismalone. It remainstrue, however, that werecapitalism (labour markets) to penetrateinto the middle hills the Lhomi would findthemselves in-'theposition described by Platt. 15 The tekkior thegi(surcharge, deducted from the initial loan or paid in labour, grain,etc.) is

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 70 CAROLINE HUMPHREY

known throughoutNepal and is standardisedin given regions. Simon Stricklandnoted an interestingcase among Gurungsof Siklisin CentralWestern Nepal ofa 'strike'by debtorscaused by the suddenraising of the tekkirate by creditors.The debtorsbelonged to one set of clans and the creditorsto anotherhigher group. The 'strike'consisted in refusingritual services to the upper groupsof clans(Strickland, personal communication). 16 Electionsto thepanchayat are expensive.One of themain lamas of Hatiyamanaged to gethis son elected,but onlyby sellingalmost all of his cattlein orderto give theelectors presents, etc. 17 Chemtang i day's workis paid i.6 kathisgrains or I0 RS Pang Dok I day's workis paid 2 kathisgrains or 8-io RS Hatiya I day's workis paid 2.4 kathisgrains or I 5 RS Syaksila I day's workis paid 2.6 kathisgrains or 4-5 RS Of course,work is notusually paid at all-exchange labourprevails. Employers differ in whatthey offer(the figures above areaverages). There is no freelabour moving from village to village.We see thatas one proceedsnorthwards from Syaksila, a day's workis paid progressivelyless in grains,and thisis whatwe mightexpect given the increasing value of grains vis-a-vis other products from south to north.But the money pay is much less clear. For one thing,grains often cost more to buy separatelyfor money than theirmoney value as wages. These prices do not reflecttransport costs-the moneyprice of rice in Pang Dok, forexample, is farhigher than the cost of buying rice in the bazaar and payingporters to carryit north.When we comparethe moneyprice of grainsin differentvillages and thepayment for work in thesame placesthere is a markedinconsistency. One day's workis paid: Pang Dok 2 kathisgrains or 8-io RS pay, but moneyprice of grainsin i6 RS Hatiya 2.4 kathisgrains or I 5 RS pay, butmoney price of grainsis I 5 RS Syaksila 2.6 kathisgrains or 4-5 RS, but moneyvalue of grainsin I3 RS The pointis thatit is virtuallyimpossible to buy grainsfor money in thenorthern villages. 18 Because moneyis virtuallynever exchanged for common produce it can maintainvalues which areinconsistent. For example,in Pang Dok I kathiof potatoes had a priceof 2 RS, and as we saw, I kathiof ricebarters for I . 5 of potatoes.We would expectthe money price of riceto be 3 RS, butin factit is 8 RS. Thereis thesame effectwith maize. Witha moneyprice for potatoes of 2 RS perkathi and a barterrate of I maize to I. 5 potatoes,we shouldexpect the money price to be 3 RS perkathi. However, in factit is 7 RS perkathi, which is consistentwith the cultural preference for rice (8RS per kathi)and withthe fact that, because of difficult conditions in theTibetan communes, it is possibleto barterall grainsin Tibet at equal rates,e.g. againstsalt. 19 These groupsare known as go ('door') and seemto be similarto theTibetan 'kidu' discussed by Miller(Miller I956) (my Tibetanassistant assumed that they were identical). 20 I am indebtedto Edgar Kellerfor the following information from West German government sources.Prices in Kathmandu(I972-3 = IOO): I977 I978 I979 I980 198I Meat and fish i66 i8o i87 235 290 Ghee and cookingoil I4I 20I i85 202 275 Herbs I94 266 262 23I 227 Vegetables i6i Ig9 205 243 262 Cloth I44 I52 I58 I70 I77 Wool I53 i6i I78 203 239 Between I972/3 and I978/9the price of raw coarserice in Dhankutarose from I.90 to 2.7I Rs/Kg, ghee fromi5.54 to 20.75 Rs/Kg, muttonfrom 7.50 to I2.54 Rs/Kg and sugarfrom 3.48 to 8.28 Rs/Kg. The priceof themain item produced by theLhomi, potatoes,rose by farless: fromI .32 ro I. 85 Rs/Kg (Joneset al. I 982: 37). Lhomi wereaware of inflationand constantlymentioned the rise in wool prices,money (as opposed to grain)wages, and livestockprices. 21 The Lhomishave tradingfriends (trogpo male, and trogmofemale) among theTibetan nomads. These relationshipsare inheritedpatrilineally, but theyare not exclusive.An activetrader might have 5-6 such friends.A few Lhomi also have more ritualisedfriendships (mit) with people of 'Gorkha' ethnic groups. The mit ritual friendshipis commonly establishedbetween people belongingto hierarchicallydistinct groups, or betweenendogamous groups, allowing an extension of close tiesin a societywhich is dividedand ranked.The relationinvolves an initiatingritual in the presenceof a priest,promises of help in timesof trouble,the use of respectfulterms of address,

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions CAROLINE HUMPHREY 7I avoidanceof themit's spouse, no marriagebetween descendants, and funeraryobligations. There is some debatein theliterature as to whethermiteri are commonlyinvolved in tradewith one another, and the evidenceseems to suggestthat regions of Nepal differin thisrespect (Gorer I938; Miller I956; Okada I957; CaplanI970; McDougalI979). In thecase of Lhomimit do tradewith one another,and thekinsmen of mitare also involved.In myexperience, Lhomi setup mitrelationships in orderto trade.

REFERENCES

Aristotle(translated by F. H. Peters)I895. TheNicomachaean ethics London: Kegan Paul. Baumgartner,R. I980. Trekkingund Entwicklung ins Himalaya: die Rotwaling Sherpa in Ost-Nepal. Diessenhofen:Riiegger. Bell, Charles I928. Thepeopleof Tibet. Oxford: Clarendon. Bohannan,P. & Laura Bohannan I968. The Tiv economy.Evanston: Northwestern Univ Press. Burbage,M. B. I98I. Reporton a visitto Nepal: themedicinalplant trade in the KHARDEP area:a study ofthe developmentpotential. London: TropicalProducts Institute. Caplan, Lionel I970. Land andsocial change in east Nepal. London: Routledge. Clarke, Graham E. I95I. The lama people of Nepal: the templeand kinshipamong a Buddhist people of theHimalaya. Thesis, Universityof Oxford. Clower, R. W. I969. Introduction.In Monetarytheory (ed.) R. W. Clower. Harmondsworth: Penguin. Crump,Thomas I98I. Thephenomenon ofmoney. London: Routledge& Kegan Paul. Das, SaratChandra I902.Journey toLhasa andcentral Tibet. London: Murray. Douglas, M. D. & B. IsherwoodI978. The worldofgoods. London: AllenLane. Geertz,C. I979. Suq: thebazaar economyin Sefrou.In Meaningand order in Moroccansociety (eds) C. Geertzet al. Cambridge:Univ. Press. Goodhart,C. A. E. I975. Money,information and uncertainty. London: Macmillan. Gorer,G. I938. Himalayanvillage. London: Nelson. Gregory,C. I980. Gifts to men and giftsto god: giftexchange and capital accumulationin contemporaryPapua. Man (N. S.) I5. Haimendorf,C. von FuirerI975. Himalayantraders. London: Murray. Hamilton,F. I9I9. An accountof the kingdom of Nepal. Edinburgh:Constable. Harris,Olivia I982. Labour and producein an ethniceconomy, northern Potosi, Bolivia. In Ecology andexchange in the Andes (ed.) D. Lehmann.Cambridge: Univ. Press. Helms, Mary W. I969. The purchasesociety: adaption to economic frontiers.Anthrop. Q. 42, 325-42. Hooker,J. D. I854. Himalayanjournals,vols I & 2. London: Murray. Humphrey,Caroline I980. Reporton a fieldstudy in northSankhuwasabha (Nepal) (KHARDEP Rep. 7). Kathmandu:KHARDEP. Jest,Corneille I975. Dolpo: communautesdelangue tibetaine du Nepal. Paris:Editions CNRS. Jevons,W. S. I910. Moneyand the mechanism ofexchange. London: Kegan Paul. Jones,S. F. I980. Kosi Hill Area RuralDevelopment Programme: agricultural marketing survey, vols I & 2. London: TropicalProducts Institute. , J. B. Innes,M. Pant & S. Subba I982. Agriculturalmarketing survey vol. 3: Surveyresults. Kathmandu:KHARDEP. Leach, E. R. I954. Politicalsystems of highland Burma. London: Bell. I960/I. The frontiersof 'Burma'. Comp.Stud. Soc. Hist.3, 49-68. McDougal, Charles I979. The KulungeRai. Kathmandu:Bibliotheca Himalayica. Malinowski,B. I967. Kula: the circulatingexchange of valuablesin the archipelagoesof A New Guinea. In Tribalandpeasant economies (ed.) G. Dalton. Austin,London: Univ. of Texas Press. Marx, Karl I954. Capital,vol. I(I887 edn). London: Lawrence& Wishart. Marshall,A. I920. Principlesof economics. London: Macmillan. Merrill,M. I977. Cash is good to eat: self-sufficiencyand exchangein the ruraleconomy of the UnitedStates. Radical Hist. Rev. I:I. New York:JohnJay College. Miller,Beatrice D. I956. Ganyeand Kidu: two formalizedsystems of mutualaid amongTibetans. SWest.J.Anthrop. 12.

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions 72 CAROLINE HUMPHREY

Okada, FerdinandE. I957. Ritual brotherhood:a cohesivefactor in Nepalese society.SWest. J. Anthrop.13. Plato (translatedby B. Jowett)I908, The Republic.Oxford: Clarendon Press. Platt,Tristan I982. The roleof the Andean ayllu in thereproduction of the petty commodity regime in northernPotosi, (Bolivia). In Ecologyand exchangein the Andes (ed.) D. Lehmann. Cambridge:Univ. Press. Polanyi,K. I957. Thegreattransformation. Boston: Beacon Press. Rauber, Hanna I982. Humli-Khyampasand the Indiansalt trade:changing economy of namadic tribesin farwest Nepal. In Contemporarynomadic andpastoralpeoples: Asia andthe north (ed.) P. S. Salzman. Williamsburg:College ofWilliam & Mary. Sagant,Philippe I968-9. Les marchesen paysLimbu (notessur les troishat bajar de Taplelunget de Terhathum).Ethnographie 62, 90- III. Sahlins,Marshall I972. Stoneage economics. London: Tavistock. Servet,J.-M.I98 I-2. Primitiveorder and archaictrade. Econ. Soc. I0, 45 I-66, II, 22-59. Stigler,George G. I96I. The economicsof information.J.polit. Econ. 69, 2I2-25. Thurnwald,Richard I932. Economicsinprimitive communities. London: OxfordUniv. Press. Woodburn, James I980. Hunter-gatherereconomies. In Soviet and westernanthropology (ed.) E. Gellner.London: Duckworth.

This content downloaded from 110.22.32.138 on Fri, 17 Oct 2014 05:02:16 AM All use subject to JSTOR Terms and Conditions