my favourite things object attachment comparative.pdf

8/12/2019 my favourite things object attachment comparative.pdf

1/18

My Favorite Things : A Cross-CulturalInquiry into Object Attachment,Possessiveness, and Social Linkage

MELANIE WALLENDORFERIC J. ARNOULD*

W e explore th e meaning and histories of favorite objects in two cultures usingsurveys and photographs. Favorite object attachment Is differentiated from the pos-sessiveness component of material ism and from at tachment to other people. M ean -ings of favorite objects derive more from personal memories in the U.S. arid fromsocial status in Niger than from object characteristics. Since favorite objects serveas storehouses of persona) meanings, gender, age, and culture reflect differencesin object selected as well as reasons for selection. In the U.S., photographs showgreater proximity to objects that are symbols of others or experiences than toobjects enjoyed for their own attributes.

M aterial objects play many roles in social life. Theyprovide sustenance, shelter, safety, and enter-tainm ent. They serve as tools to accomplish tasks. Theyprovide mobility. They counterbalance the effects ofnature by keeping us dry when nature is wet, warmwhen it is cold, cool when it is hot, shaded when it istoo sunny, and in the light when it is too dark. For 50years paleo-archaeologists have told us that materialobjects have helped us m ak e ourselves as hum anbeings (Childe 1936; Issac et al. 1979).

THINGS AND SELFHOODObjects serve as the set and props on the theatrical

stage of our lives. They situate an ind ividual's characteror personality in a context (Goffman 1959; Holm an1980; Levy 1959; Mick 1986; Turner 1969). We useobjects as markers to denote our characters for others;we also use objects as markers to remind ourselves ofwho we are. In this sense we derive our self-concept

Melanie Wallendorf is Associate Professor of Marketing, Collegeof Business and Public Administration, and Eric J. Amould is AdjunctAssistant Research Scientist, Bureau of Applied Research in Anthro-pology and Office of Arid Lands Studies, both at the University ofArizona, Tucson, AZ 85721. We thank our colleagues at the Uni-versity of Arizona for their useful comments on earlier drafts of thisarticle. We thank the Department of Marketing at the University ofUtah for their hospitality du ring the first author s sabbatical wh ileshe revised this article and the participants at the 1986 Society forEconomic Anthropology meetings for their many stimulating ideason consumer behavior. We would like to extend very special thanksto Russell Belk for his encouragement and extensive comments onthis work. Finally, our thanks to four anonymous reviewers.

531

from objects. That is, we use objects to convey and ex-tend our self-concepts (Belk 1987a) to others as well asto demonstrate the self-concept to ourselves. Objectsconvey our connection to others and help express oursense of self (Levy 1981; McCracken 1986; Rook 1985).

For the most part, modern consumer research pub-lished in marketing has not examined directly the phe-nomenon of attachmen t to objects and the m eaning ofobject ownership (Belk 1985) despite the interest of cer-tain of its forebearers (Veblen 1899). It has, however,examined brand preference and brand loyalty (Jacobyand Chestnut 1978) and involvement (Bloch and Ri-chins 1983), which all tie the individual to the brandor purchase context. Yet these topics focus on the ac-quisition and prepurchase phase of buying, rather thanon ownership and consumption and their meanings toconsumers. Because consumption is an im portant con-cept in understanding dem and and consum er behaviorgenerally, some researchers have begun to addressquestions of ownership and the meaning of consump-tion (Csikszentmihaiyi and Rochberg-Halton 1981;Levy 1981) and more macro issues such as productconstellation meanings (Solomon and Assael 1987) andcultural brandscapes (Sherry 1986).

In anthropolog y, objects have usually been discussedin terms of their role in the production process or ingift exchange (Gregory 1982; Hyde 1983; Levi-Strauss1979; Mauss 1967). Tradition ally, the m ovem ent awayfrom locally made material culture and the adoptionof culturally alien objects was merely viewed as an in-evitable, if regrettable, part ofthe accultura tion process(Douglas and Isherwood 1979; Stout 1947; Wallendorfand Reilly 1983). More recent work has begun to clarify

JCXJRNAL OF CONSUMER RESEARCH Vol. U March 1988


2/18

53 THE JOURNAL O CONSUMER RESEARCH

the historical political-economic uses and meanings ofobjects among cultures of traditional anthropologicalinterest (Appadurai 1986; Douglas and Isherwood 1979;Mintz 1979; Mukerji 1983; Society for Economic An-thropology 1986).

Collectively, the research on ownership in a number

of fields leads us to contend that attachm ent to objectsas symbols of security, as expressions of self-concept,and as signs of one's connection to or differentiationfrom other m embers ofsociety is a usual and culturallyuniversal function of consumption. The primary pur-pose of the work reported here is to conceptually andempirically explore the nature and meaning ofthe at-tachments people form to objects that they designateas special or favorite.

In the United States, the phenomenon extends to theinfant's security blanket (Passman 1976; Passman andAdams 1981; Passman and Halonen 1979; Passman andLongeway 1982; Weisberg and Russell 1971). Such at-tachments develop very early and are common; pref-erence for a favorite object has been found to exist inmore than 70 percent of six-month old infants (Furbyand Wilke 1982). The familiar blanket provides a psy-chological feeling of comfort quite apart from its util-itarian warmth-giving properties. It serves as a transitionobject enabling the chiid to move away from the securityof parents and venture into the physical world.

Since objects carry a self-concept-based meaning,losing or severing our connection to objects nonvolun-tarily can change the meaning of life for individuals.For example, Goffman (1961) has described the strip-ping process that occurs when individuals enter whathe calls total insti tutio ns, such as prisons or mentalhospitals. Upon arrival, an individual's clothing andpersonal possessions are taken away. Institutionalclothing and objects are issued for the person's use butare not under his or her full control. Thus, ownershipof objects disappears as the institution takes on the roleof providing objects for one's use. Connections to no rm al life on the outside are severed, and indi-viduals gradually assum e the dependent role of patientor prisoner. In practice, institutionalized persons findit difficult to claim or reclaim their norm alcy.

Institutionalized mentally retarded patients strippedof objects for maintaining self-definition often attemptto reverse the stripping process by acquiring objects tha tothers ( normal people ) define as useless, such assoiled wrapping paper and expired coupons (Carroll1968). These objects then take on new meaning in dif-ferentiating the self from others. Patients attempt toappear normal to reestablish individuality, and todisplay connection to the outside world by collectingtreasured junk . Their behavior is considered inap-propriate because they confer treasured status on objectsmost people consider rubbish (Thompson 1979).

Social scientists have found that when elderly peoplemove into a nursing home, they feel a loss of status(Sherman and Newman 1977-78). To compensate or

attenuate this feeling, many bring with them a cherishedobject. Their strong attachm ent to this object is usuallnot based on its monetary value. Rather it holds sym-bolic value and provides a sense of security as well ascontinuity in one's link with others. In Mexico Cityeven deeply impoverished families cling to religiou

icons and use a shelf in their homes as an altar to sym-bolize their hopes for a better future in the afterlife(Lewis 1969).

Because objects serve as personal storehouses ofmeaning, losing all of one's material possessions is ex-perienced as a tragedy and a violation of the self inAmerica. The emergence of victim support groups andthe felt inadequacy of safety nets like homeowner's orrenter's insurance indicate how much we dread suchlosses. Loss of objects implies loss of face and statubecause the objects are a representation of self. (SeeBelk 1987a for an extended discussion of loss of possessions leading to a diminished sense of self.)

In summary then, a wide range of phenomena fromthe baby's unselfconscious attachment to objects to thetrauma of loss through theft, catastrophe, or institutionalization indicates how important possessions areto the Am erican sense of self. Data from other cultureprovide comparable illustrations of the fundamentaattachment between people and objects. Although themeaning of self differs cross-culturally and varies in itslink with individualism (Hsu 1985), the fact that thesconcep tions of self are expressed to some degree througobjects seems to be universal.

There are many examples from around the world oftribal peoples' wholesale, ingenuous embracing ofwestern objects, which from a utilitarian viewpoint arecompletely out of place in the tribal context (Arnouland Wilk 1982). In the South Pacific in the wake ofWorld War II, veritable cargo cul ts grew up as bigmen in tribal cultu res sought to obtain western objecby superna tural mean s (Worseley 1968). Acculturatistudies in the forties documented the apparently willinadoption of all m anner of manu factured goods by non-western peoples (e.g.. Stout 1947 on the San Bias Cu naand the pages o ational eographic still contain pic-tures of naked tribespeople enjoying western consuables (e.g.. Devillers 1983: Tw 'eedie 1980; Wentze1978). Such attachment behaviors give expression toself-differentiation by drawing sharp contrast with thecultural context in which they are embedded.

The societal im pact of oss has also been docum entein tribal cultures. Fam ous case studies such as M etrau(1959) on Amazonia, Sharp (1968) on Australia, andTfurnbuU (1972) on Africa do cum ent the breakdown ofsocieties and sociability when key objects in the m ateriaculture inventory were lost or replaced throug h the in-cursion of m anufactures or money. In these cases, ofcourse, loss is culture-wide rather than individual. Re-cently some work has attempted to compare cases, deciphering why some objects were accepted and somrejected by a culture, and why the loss of control ove


3/18

A CROSS CULTURAL INQUIRY 533

only certain kinds of objects results in radical decul-turation (Douglas and Isherwood 1979; Gregory 1982;Leach and Leach 1983; Strathern 1969).

A process similar to Goffman's "stripping" is char-acteristic of rites of passage in nonwestern ritual con-texts. Initiates are often deprived of their possessionsas they assume new social id entities. D uring the ritualtransformation, special objects and foods are designatedfor their use. Upon successful completion of the ritualtransformation, initiates emerge usually with a new so-cial identity, but frequently with new objects as well,such as tribal scars, a spear, a new hairstyle, or a newwrap (Farb and Armelagos 1980; Turner 1969; VanGennep 1960).

From these diverse exam ples, it appears that attach-ment to and derivation of meaning from objects occursamong all peoples, including no madic tribes that placea premium on mobility. For example, for the Samburuand the Nuer of East Africa, cattle take on a multi-layered meaning. For cattle pastoralists, diverse valuesand notions about status ranks are intertwined in onetype of object (Evans-Pritchard 1940; Goldschmidt1969; Lincoln 1981; Verdon 1982). Among the KungSan bushman tribes of Namibia and Botswana, multiplemeanings are conveyed by beaded headbands. Weissner(1984) describes how band affiliation, degree of accul-turation to surrounding Bantu custom, and even beliefin the traditional behavioral norm of "walking softly"are conveyed through headband design elements.

In many cultures in the Third World, the number ofcommodities in circulation and the frequency and mul-tiplicity of occasions for their exchange, consumption,and display have been more limited than in the West(Appadurai 1986). These cultures frequently compressmultiple meanings into a few types o property, ratherthan into the many types of objects such as clothing(McKendrick, Brewer, and Plumb 1982; Solomon 1983;Veblen 1899), automobiles (Evans 1959), homes andhome furnishings (Davis 1955; Felson 1976; Kron 1983;Lynes 1980; W arner, Meeker, and Eels 1960), and foods(Farb and Armelagos 1980) used by Westerners forconveying such m eanings.

In both Western and nonwestern cultures, attach-ment to particular favorite objects as symbols need notbe viewed as something that is evil or bad, as has beenthe perspective toward the more general phenomenonof materialism taken by many religions (Belk 1983) andsocietal critics (Looft 1971; Wachtel 1983). Researchon elderly Americans finds that individuals who lackcherished possessions have lower life satisfaction scoresthan those who have such objects (Sherman and New-man 1977-78). Specific object attachments need nottake over the individual's orientation to life and developinto an all-consuming materialism or attachment toobjects as in the case of fanatical collectors (B audrillard1968). Indeed, fierce competition to obtain kul arm-shells and necklaces, some of which have circulated forgenerations in the New G uinea archipelago, always en-

tails their future exchange for different but equally val-ued markers of status and facilitates the extension ofsocial networks rather than the expression of pure cov-etousness (Leach and Leach 1983; Malinowski 1922).Such objects permit individual differentiation and self-expression for a while, but the meaning of that self-

expression is inextricably intertwined with connectionto a larger group .On the basis ofth e study of beaded headbands among

the IKung San, Weissner (1984) hypothesizes that ob-jects fuel a universal dialectic of style through whichthree fundamental social processes are enacted: differ-entiation, comparison, and integration. Her views havebeen echoed by other scholars working in both Westernand nonwestern contexts (Csikszentmihaiyi and Roch-berg-Halton 1981; Douglas and Isherwood 1979; Ko-pytoff 1986). Although specific m eanings differ in vary-ing cultural contexts, consumption is an activity bywhich co nsum ers create intelligibility in the world andmake visible and stable the categories of culture as theyexperience them (Douglas and Isherwood 1979).

OBJECTIVESThis conceptual foundation leads us to four empirical

research questions in our attempt to understand themeaning of individuals' cherished objects in two quitedissimilar cultures. First, the analysis attem pts to clarifythe nature of attachment to favorite objects for the twogroups of respondents. Second, relationships betweenattachm ent to a specific favorite object and more generalattachment phenomena are addressed. Respondents'

levels of generalized possessiveness a component ofmaterialism as conceptualized and measured by Belk1984) and their social linkages cf. Bott 1971) weremeasured. The extent of overlap of each of these withfavorite object attachment is then examined. The thirdresearch question involves cross-cultural comparisonsof levels of favorite object attachment and the gener-alized possessiveness com ponen t of materialism . Theseresults are presented in an attempt to determine culturaldifferences as reflected in these measures. Finally, therole of three compon ents ofsociety, which are also en-during and distinguishing compon ents of self-concept,namely culture, age (Erikson 1959; Furby 1978; Neu-

garten 1969), and gender (Mead 1949; To urnie r 1981),are examined to see how they structure and explainfavorite object selection cross-culturally. Differences inthese three components would be expected to be ex-pressed through favorite objects. They should then befound to have not only strong and enduring linkagesto self-concept, but also strong linkages structuring fa-vorite object selection. Multiple methods are used toexplore the experiential meaning and history of favoriteobjects as expressions of self (Holb rook and H irschman1982). Analysis focuses on understanding common andcontrasting structures in informants' emic representa-tions of meaning.


4/18

534 THE JOURN L OF CONSUMER RESE RCH

M E T H O D

Samples and SettingsIn-home personal interviews were conducted with

two samples of adults selected to cover variation in cul-

ture, socioeconomic g roup, and domestic group life cy-cle stage. Cultural differences were necessary to assessthe generalizability of favorite object attachments asexpressions of identity. Th us, two highly dissimilar cul-tures with respect to economic development, materi-alistic values, and breadth of opportunities for expres-sion of self through objects were chosen. Within eachculture, socioeconomic and domestic cycle diversitywere desired to adequately capture intracuhural, as wellas intercultural variance.

The first sample, consisting of 300 adults, was drawnfrom a major Southwestern American city. The city ischaracterized by rapid im migration from other parts of

the U.S. Most residences are of recent construction,characterized by open space plans with expansive viewsof the surrounding mountains. The local economy isser\ ice based and economic growth is tied to populationgrowth. Consistent with the hot climate and westernimager>\ lifestyles tend to be casual rather than formal,and because most citizens are recent arrivals, they tendto be open rather than tradition-directed.

The other sample of 45 adults was drawn from theHausa-speaking peasants living in three villages in Zin-der province ofthe Niger Republic (Arnould 1984a).As part of ongoing ethnographic fieldwork, the secondauthor obtained responses from the Nigerien sampleusing a similar semistructured interview. Difficulties oftranslation, sample member identification, the inter-view situation, and establishing rapport resulted in asmaller sample size for the Nigerien group.

Niger is landlocked in the African Sahel and was asite of severe drought from 1969-1973 and again in1983. Located in the center of Niger, Zinder provincelies to the north o fthe Hausa market centers in Nigeria.Victim to the pattern of regional, sectoral, and socialdisarticulation typical of peripheral capitalist devel-opm ent in Africa (Amin 1973, 1976; de Janvry 1981),Zinder's fragile modern economy has never recoveredfrom the collapse of the export-oriented peanut tradein the early 1970s (Franke and Chasin 1980). In 1985-1986, cereal prices were comparable to those in 1977-1978 and prices of many rural handicrafts had hardlychanged in that tim e. Although there is a lively periodicmarket system (A rnould 1985), there is nonetheless littlescope for capital accumulation or discretionary con-sumption. The economy has been characterized as onecondemned to economic involution (Arnould 1984b).

Despite regional, occupation al, eth nic, and class dif-ferences, Islam is cultural ly roo ted in daily life. Un -like Judeo-Christian religions, the Islamic traditionmakes no distinction between religious, civil, and crim-inal law (Schact 1964). In the Islamic conception of

property, limited use rights, as distinct from full owership, are com mo nly recognized (Schact 1964). Prscriptions on usury co-exist with a strong value of ivestment in trade, livestock, and urban real estate. Zinder, fidelity to the practice of almsgiving and beliin the dignity of poverty co-exist with the notion th

wealth brings happiness and the near homology in dato-day life between the status of bourgeois and that pilgrim to Mecca (Hausa, masc. elhadji; fem. hadjiya .

A typical household's compound in rural Zinder icludes a round thatch or adobe thatch-roofed hut feach adult wife. A wealthy m an may also build him sa square adobe house used to entertain male visitoand as a storehouse. Floors are of sand; doors are corrugated metal or matting. Clothes are usually hunon the wall, although some people have cheap valisor metal trunks. Houses are simply furnished withbed and palm fiber or plastic mats. Rich people possea prayer rug or woven hangings. There may be a sma

kerosene lamp. Outside there is often a small area ftethering go ts and sheep. W omen cook in the courtyaon a tripod of stones using clay and gourd vessels anwooden, gourd, and tin utensils. Enamelware foopreparation and serving dishes are now commonplacTools (mo rtars and pestles, axes, hoes, a bucket, a flaslight, and knives are most common) and small woodstools are often scattered around. Thus, both economically and materially, life in Zinder is quite a contrato life in the Southwestern United States.

Ethnic subpopulations represented in the two samplwere too small to allow subanalysis. Census data fthe American sample and fieidworker knowledge population composition for the Nigerien samp le permus to claim that the populations were representative terms of key demographics such as gender, incom e, agand, where appropriate, home ownership and eduction.

Data Collection MethodsThree methods of data collection were employe

surveys, photographs, and focus group interviews. Fmost of the concepts of interest, self-report measurwere deemed appropriate. However, for the primaconcept of interestnature of attachment to the fvorite objectan approach employing more than omethod (Campbell and Fiske 1959) was used. FSouthwest American respondents, two methods of dacollection were employed: (1) individuals were askquestions about their favorite object, and (2) individuawere photographed with their favorite object.

As suggested by Wagner (1979) and Collier (1967the use of photographs in social science should go byond m erely using photos as illustrations (cf. Danfor1982; Lynes 1980; Susman 1973). The photograp hmaterials should be coded to becom e raw data for anaysis (cf. Felmlee, Eder, and Tsui 1985; Rheingold anCook 1975), an approach that is receiving increasi


5/18

CRO SS CULTURAL INOUIRY 535

tention in consumer research (cf. Belk 1987b; Heisleynd Levy 1987; Heisley, McGrath, and Sherry 1987;

Wallendorf and Westbrook 1985). The photos in thisudy became raw data through structured analysis

Collier and Collier 1986) ofthe physical relation be-ween the. respondent and the favorite object. C losehysical proximity was taken to indicate a high level oftachm ent to the object (Mershon 1985). This is similar

o other research in which physical proxim ity has beensed as an unob trusive measure of social connectionnd structure between racial groups (Campbell, Krus-al, and Wallace 1966), within peer groups (Feshbachnd Feshbach 1963; Hall 1969), in families (Milgram977), and in field studies of animals (Imanishi 1960).In this project, the physical proximity between the

espondent and the favorite object in the photographsas coded using a five-point scale. The coding was done

with one coder on two occasions, and once by anotheroder. Intrarater reliability was 0.90; interrater reli-bility with two coders was 0.93. These levels meet re-ability criteria established by Nunnally (1967). Thisnalysis uses the average of the three codings of eachhotograph.

In Niger also, two methods of data collection weremployed: (1) individuals were asked about their fa-orite objects as in the Southwest, but in addition, (2)ocus-group interviews were employed to discuss pat-erns of intro duc tion and diffusion of items identifieds recent popular introductions into the local materialulture inventory. Information necessary to place re-ponses in context has been collected over a num ber ofears using a variety of ethnographic methods (Arnould

1984a).Both samples were administered an interview sched-

ule originally developed for the U.S. but also adaptedo the cultural and linguistic situation of Niger. In Niger,he questions were translated into the Hausa language.

However, this was not sufficient for rend ering themulturally and contextually appropriate. Although itntroduces nonparailel methods in the two cultures,ome scaling and meaning changes were made. For ex-mple, for the Americans, frequency of talking with

others on the phon e was a scale item used in measuringocial linkages. In Niger, other forms of social com-

mun ication were included such as attending village as-ociation meetings, eating with friends, and sharing

Moslem thanksgiving. Thus, cultural appropriatenesswas given priority over linguistic equivalence in scaleconstruction. As part ofth e interview, respondents wereasked a series of questions to identify possessive atti-udes towards possessions in general, to explore the ex-ent and importance of social linkages, and to identify

a favorite object.In the sample drawn from the American Southwest,

he choice of objects was confined to the living room.This limitation enables greater comparability of thecommonalities of expression through favorite objectsbetween the Nigeriens and the more possession-rich

Americans. This area ofthe house is one which is themost public and therefore the most involved in impres-sion managem ent in Am erican culture (Goffman 1959).It is designed to present to others our sense of ourselvesand our personalities (Baudrillard 1968; Kron 1983).Laum ann and House state that the living room reflectsthe individual's conscious and unconscious attempts toexpress a social ident ity (1970, p. 323). In sho n, wewould expect to find favorite objects that are expressionsof some im portant aspects ofthe self n American livingrooms.

Kron (1983) recognizes that there is greater femalethan male influence over the American living room.Restricting choice to the living room could producegender differences in degree of attachment to objectschosen. H owever, since the living room is more gend erneutral than other areas ofthe house (e.g., kitchen andbedroom ), it was selected as the best area for co ntainingboth male and female expressions of social identity.

In Zinder, respondents were simply asked to nameany favorite object with no restriction of location ap-plied to their choice. The justification is that the scopeof consumption for these rural people is simplified incomparison to that of Americans. And since the notionof finely graded responses is culturally unfamiliar, therespondents were asked to rate their liking for the fa-vorite object on a four- rather than seven-point Likertscale as used with the Americans.

RESULTS

Forms of Attachment ossessiveness Although the primary focus in this

study is attachment to a specific object, the relationbetween this form of attachm ent and other more generalexpressions of attachment was also of interest. Thecomponent of materialism that is a general attachmentto possessions has been termed possessiveness. Anine-item summed scale to measure possessiveness(seven items in Niger), which has been demonstratedto have fair reliability, convergent validity, and criterionvalidity in U.S. cultural settings (Belk 1984, 1985) wasemployed to ascertain the respondent's more generalattachment to all material possessions. This scale in-cludes items addressing general attachment toward allof one's possessions as well as control over possessionsand feelings concerning loss of possessions. However,the scale items do not focus on attachment to specificpossessions or what are termed favorite objects inthe current w ork. Some changes were made in the scaleitems to render them contextually ap propriate in Niger,although they remain conceptually comparable to theoriginal scale utilized in the American sample.

Because the scales for generalized possessiveness aredifferent in the two cultures, for com parison, the m eansfor each culture were transformed to standardized scoresby dividing by their standard deviations. This resulted


6/18

536 THE JOURNAL OF CONSUMER RESEAR

EXHIBrr 1

PERSON ATTACHMENT SCALE ITEMS AND FACTOR LOADINGS' FOR THE AMERICAN SAMPLE

EXHIBIT 2

PERSON ATTACHMENT SCALE ITEMS AND FACTOR LOAFOR THE NIGERIEN SA MPLE

Items

Frequency of entertainingothers in the home

Frequency of talking withothers on the phone*

Marital attachmentNumber of people in the

householdNumber of financially

dependent childrenNumber of relatives living in

same townNumber of dub memberships

Factor 1

- . 0 3

- . 0 0 2- . 6 4

81

.08

.19

.04

Factor 2

.63

.51

.10

- . 1 2

.02

- . 0 1.03

Factor 3

- . 0 9

.24

.19

.01

.01

- . 4 5.20

Fn*p oint scale where 1 - never. 5 = frequ ently. Ordinal scale to reflect strength ol m antal attaOvnent using demographic inlonnation:

1 married. 2 - w idowed. 3 separated. 4 - divorced. 5 never manted.

in standardized group scores of 4 01 (raw s.d. = 2.65)for the Nigeriens and 4.62 (raw s.d. = 4.49) for theSouthwest Americans. A /-test of these standardizedgroup scores revealed differences that are statisticallysignificant at the 0.0001 level (/ = 3.79, df = 344). Thesefindings suggest that Americans are substantially morematerialistically possessive than are the Nigeriens. Thisshould not be surprising. The major consumption goalof Zinderois elicited in surveys (Crow and Henderson1979; Republique du Niger 1985) remains nutritionalself-sufficiency. In contrast to the Am erican consum ers,

Zinder's consumers have not yet been taught to con-sume and how much to consum e the good life bymarket mediated consumption and mass media adver-tising (Belk and Pollay 1985).

Social Linkage Also of interest is the individual'sattachment to other people. One might wonder whetherattachment to favorite objects can fill the void of alien-ation from other people. Csikszentmihaiyi and Roch-berg-Halton (1981) found, however, that individualswho claimed not to be materialistic because they didnot have things that had special meaning for them alsolacked special close friendships and relationship s. Thosewho had strong ties to other people represented theseties in special material objects.

Social network linkage was measured in a summedscale of factor scores for seven items in the U.S. sampleand thirteen in the Nigerien sample. The items wereselected to reflect common ways individuals maintainstrong attachmen ts to other people. The items employedin the U.S. sample include frequency of entertainingothers in the home, talking with friends on the phone,marital attachment, number of people in the household,number of financially dependent children, number ofrelatives living in the same town, and number of clubmemberships. In Niger, some culturally irrelevant itemswere deleted while other items were added, including

Items Factor Factor 2 Facto

kana zuwa abinci gidan wane, kunaciyy y l (Frequently eats withfriends) .03 .02 .1

in kana da lokaci kana so kullum kaziy rci abo/ca/'? (Would more oftenvisit friends) .01 .43 .

abokanka suna zuwa ziyartarkakullum (Frequently receivesguests) .04 -. 5 8 .3

in ana kirin taro kana zuwa kullum?(Frequently attend s villageassociation meetings) .09 .44 .2

a ganinka, yara dole ne su bi asbinda iyaye yya ke so? (Childrenshould obey their parents) .70 .07 .0

t>ak an magan ba ta tadama

h nk link (Sensitivity to villagegossip) -. 2 2 .69 -. 3kana ziyarten yan uwa ko abokankai

na garuruwan kewaye da ku?(Frequently travels to othervillages to visit family/friends) .15 -. 0 7

a ganinaka akwai wayyandan sukafinka tsaregan jama a a garin na?(Others are more suspiciousthan I) .03 .74 .3

kana yarda da raiwon mutanen nang rin (Agreement with localcustom) .76 -. 19 .

ka biye da raiwon mutanen nang rin (Follows local custom) .91 .00 -.

kana yarda ka bar ra ayinka ka dauki

na jama a? (Better to shareothers' opinions than to preserveone's own) .31 -. 3 5 .

a wane gida ka samu naman layyab n (Number of persons withwhom I shared Moslemthanksgiving in 1985) .05 .13

mutanen nawa kawa tufan salla?(Numtjer of persons for virtnom Ipurchased Moslem thanksgivinggarments) -. 1 3 -. 03 .

number of persons entertained on a major festivecasion, attendance at village moots, and sensitivigossip (see Exhibits and 2 for comparisons).

These items were factor analyzed and varimaxtation was used on the principal factors. In the sample, three factors were extracted with eigenvagreater tha n accounting for 58 percent ofthe varin these divergent h um an contact item s (see ExhibIn the Niger sample, although five factors with evalues greater than 1 accounted for 67 p ercent ovariance, only the first three were used, given the limsample size { = 45) and the relatively large numbeitems included in the analysis (13). These three fahave eigenvalues larger than 1.5 and account fopercent ofthe variance (see Exhibit 2). Since the pr


7/18

A CROSS-CU LTURAL INQUIRY 537

work is not attempting to empirically decompose socialinkage, but rather is examining its mo re m olar overallelation to favorite object attach ment, factor scores wereummed across the three factors for an overall measuref person atta chm ent. Because different scale items were

used in the two sam ples, cross-cultural statistical com -arisons of soci l network density are not appropriate.

Favorite O bject Attachment. In the portion of thequestionnaire dealing with favorite objects, one ques-ion addressed degree of attachm ent to the favorite ob -ect. In the U.S. sample, this was a seven-point Likertcale item; in the N igerien sample, this w as a four-pointoding of reported attach me nt. W e recognize that such

a measure should be somewhat skewed given the self-election of objects that hold favorite status; the scale

was used only for cross-cultural comparison ofthe ex-remes of expressed attachment. That is, this scale in-

dicates the maximal degree to which an individual in-vests the self in an object. Since one was a seven- andhe other a four-point scale, the group means wereransformed for comparability by dividing by their

standard deviations, resulting in means of 3.72 for theNigeriens (raw s.d. = 0.72) and 4.28 for the SouthwestAmericans (raw s.d. = 1.35 . A /-test oft he differencebetween these standardized group means for attachmento favorite object was statistically significant (/ = 3.52,

df = 344, p 0.0001). Thus, on average, the U.S. samples more strongly attached to their favorite objects than

is the Nigerien sample.

Relationships Among Favorite Objects Possessive-ness. and Social Network Linkage. The relationshipsbetween degree of attachment to favorite object (mea-sured verbally and photographically), possessiveness,and social network linkage are shown in Table 1 asmeasured by Pearson correlation coefficients. There islittle consistent empirical overlap among these threetypes of attachm ent across the two samples. In the U .S.sample, generalized possessiveness bears a weak nega-tive relationship with self-reported attachment to fa-vorite object (r = -0 .15 , p = 0.05). Conversely, in theNigerien sample, generalized possessiveness bears aweak positive relationship to self-reported attachmentto favorite object (r = 0.28, p = 0.028). Overall, thereis evidence for a substitution effect of favorite objectfor possessiveness in the Southwest American sample,but a collaborative effect in the Nigerien sample. How-ever, these relations are sufficiently weak to claim thatgeneralized possessiveness and favorite object attach-ment are conceptually and empirically separable. Al-though they are weakly correlated empirically, it ap-pears that neither is merely an expression ofthe other.Favorite object attachment is not strongly related togeneralized possessiveness or attachm ent to other peo-ple. Since these are distinct p henom ena, favorite objectattachment requires additional contextual analysis tospecify its nature in particular cultural contexts.

TABLE 1

CORRELATIONS BETWEEN FORMS OF ATTACHMENT

Anachment tofavorite object photographic

display)

Posses-

siveness

Person

attachment

Attachment to favoriteobject verbal report)

Attachment to favoriteobject photographicdisplay)

Possessiveness

.03 . 1 5.29

- . 0 0 8

- . 0 1- . 2 0

.0 6

.005- . 3 0

ps 5Soores n me frsi row ara for ttw SouVnvestem sample, thos m the seoond row lor me

Nigerien sample. Photographs were not taken ol the Nigerien sample.

Overall Meanings of Favorite Objects

Respondents in both cultures provided insights intotheir lives when they explained why they liked a par-ticular object. When respondents were asked why theychose a particular object as their favorite, they did notfocus on functionally based performance attributes. Forroughly 60 percent ofthe American sample ( = 171)the reasons given reflected attachments based uponpersonal memories. The object was a favorite becauseit was a reminder ofa friend or family member, a va-cation trip, or an event in the respondent's past. For 6percent ofthe U.S. sample {n = 18), the object was afavorite because it reminded the respondent ofthe per-son who had made it. typically as a gift. The meaningof these objects, then, often derives from symbolic per-son, event, and maker attachments rather than fromtheir physical attri bute s. This is not surprising since 45percent ofthe U.S. respondents received their favoriteobject as a gift, indicating the unique meaning of objectsselected and given as gifts (Caplow 1984; Sherry 1983).Like the infants who do not choose blankets that arephysically similar to their own security blanke ts (Weis-berg and R ussell 1971), adults layer meanings on objectsthat do not derive from physical features, as with sou-venirs and tourist photographs (MacCannell 1976).

Some U.S. respondents chose functional (rather thandisplay) objects such as chairs or clocks (see Table 2).Nevertheless, the reason given for these attachmentstypically derives from a shared history between the per-son and the object, such as between the television char-acter of Archie Bunker and hi s chair. This historj' isnot purchased with the object. After years of use. theweb of semiotic and sym bolic associations spun aroundthe object by which it becomes decommodified and singula rized for the individual (Kopytoff 1986) cometo be the reasons for its selection as a favorite object.


8/18

538 THE JOURNAL OF CONSUMER RESEAR

TABLE 2

TYPE OF FAVORITE OBJECT BY SOUTHWESTAMERiCAN GENDER GROUPS

Type of favorite object

Function ai ciiair, ciocl


9/18


makes people ho t (Hausa, zafi , the loss of possessionsis expected to be borne with patience (Hausa, hakuri .One informant expressed concern that if he amassedtoo many possessions, they might be lost and wastedthrough divine intervention.

Nigerien attitudes towards possessions are clarified

with interpretive contextual data. Periodic droughtsregularly reduce consum ption decisions to the problemof obtaining adequate food, clothing, and shelter.Against the backdrop of Islamic attitudes towardsproperty, the search for psychological well-beingthrough discretionary consumption in Niger is furtherconstrained both by the limited agricultural productiv-ity and state economic policy. Direct and indirect tax-ation (e.g., contributions to festivals and dignitaries)and cheap food policies limit peasant purchasing power(de Janvry 1981; Olivier de Sardan 1984; W atts andBassett 1986). Distortions in regional developmentpatterns and the exchange rate drain human and mon-

etary capital out of Zinder into Nigeria (Evans 1977).As a result, rural Nigeriens are not socialized to chooseamong a plethora of alternate material sources of sat-isfaction as are Southwest Americans.

Material satisfaction in the countryside entails de-pendence upon or power over other people (Baier 1974,1976; Kirk-Green 1974). Zinder is an economy an d aculture (cf. Hyde 1983) in which personal well-being ismeasured not solely in wealth in objects, but in theability to give and to compel persons to reciprocate.Exchanges between kin of clothing, items of adornment,or other possessions, even favored ones, are common-place. Annual tithes are paid to persons in positions ofboth religious and secular authority.

Taken together, findings from the two samples recallthose of Myers (1985) and of Csikszentmihaiyi andRochberg-Halton (1981). The latter note that the as-signment of meaning to objects is flexible since it doesnot derive from the physical characteristics ofthe object.Like dialectical variation in language, the same objectwill have different meanings to different people becauseof its different associations to them . These auth ors statethat things are cherished not because oft he materialcomfort they provide, but for the information they con-vey about the owner and his or her ties to others(1981,p. 239). Our work shows that these individual nuanc esof meaning are overshadowed by cultural differences inthe meaning not only of objects but of possessivenessitself.

Gender and Favorite ObjectsDifferences between men and women in their selec-

tion of favorite objects exist in both samples. As shownin Table 2, U.S. women are more likely than men tochoose handicrafts, antiques, and representational itemssuch as photographs of family members. Men, on theother hand, are more likely to choose art pieces, func-tional items, and plants and other living things. The

TABLE 3

TYPE OF FAVORITE OBJECT BY NIGERIEN GENDER GROUPS

Type of favorite object

Marriage/Domestic goodsGewelry, hangings, beds)

Religious/Magical items Koran, liturgical texts, charm s)

LivestockTools

Males

4

4619

8

Females

85

045

NOTE: Numbers are in percent.

overall relation between gender and type of favorite ob-ject am ong Southwest Am ericans as tested by a X^ testwas statistically significant atp = 0.001.

In Zinder, gender is also strongly related to the typeof favorite object selected (see Table 3). Commonlynamed favorite objects were religious books, in cluding

copies of the Koran. They were named by 22 percentof informants, and exclusively by men. Men indicatethat they value these objects for their instrumentalvalue, either as a (spiritual) link w ith the divine or as a(magical) agent of protection against ill-wishers or evilspirits. Men's favorite possessions, including religiousbooks, charms, swords, and horses, are vinually allsymbolic of real or desired authority over persons orthe spiritual world.

Other frequently named items included machine orhandwoven tapestries (32 percent). The former feature H in du scenes or scenes of Mecca. The latter are tra-ditional strip weavings in form, but today typically in-

corporate the Nigerien national colors (orange, white,green) or emblem s. Next in frequency came silver jew-elry (15 percent), including massive bracelets or neck-laces of Zin der crosses. Both types of items werenamed exclusively by women. Cultural ideals of beauty,notions of prestige, and association with senior femalerelatives were all linked to these items. These items areusually given to brides upon marriage. They are com-monly employed in competitive displays betweenwomen on major religious holidays or during householdlife crisis rituals (baptisms and marriages). Thu s, theseitems are symbolic of women's connections to women,both through kinship and informal politics.

In both cultures, women frequently chose items madefor or given to them by others, antiques or heirloomsthat tie them to previous generations, and representa-tional items (e.g., photos) depicting their children,spouses, and grandchildren. Yet, in the Zinder sample,the relation between gender and social linkage (r =-0 .3 l, p = 0.02), indicatesgreaterdensity of men's so-cial networks. This finding may be explained by scaleconstruction for social linkage and the gender rolesspecified in this Moslem culture. In Niger, men's socialnetworks tend to be more extended than wom en's, sincethey have greater freedom of mov emen t. W om en's net-works are comprised of stronger, more private ties.


10/18

540 THE JOURNAL OF CONSUMER RESEA

TABLE 4

TYPE OF FAVORITE OBJECT BY SOUTHWEST AM ERICAN AGE GROUPS

Age groups

Type offavorite object

Totalsample 18-24 25-35 36-44 45-54 55-64 65

FunctionalEntertainmentPersonal ItemsArt pieceRepresentationalPlants and other

living thingsHandicraftAntique

Total

24211410

8

886

100n = 298

3120

979

1194

100n = 45

252811

66

1076

100n = 114

211811198

01111

100n = 62

24624 5

9

99

100= 33

020202020

100

10

100n= 10

133

10n =

NOTE: Numbars are in peroent;eolunins may not sum 10 100 percent due to foundings. X ' - 4.1 wlth 35d (.: p < 0 . 0 1 .

Men's connections, expressed through gift exchangesexternal to the household and village, are best cap turedby the scale. Yet women's favorite objects are nnoreexpressive of social connections. In the two settingsthen, men m ost often chose craftgoods and artworks torepresent ideals, functional objects to depict levels ofcomfort they have obtained, and religious texts, charm s,plants, and pets to demonstrate their mastery over na-ture.

These findings of gender differences are consistentwith previous research. Sherman and Newman (1977-78) found that elderly men and women were equallylikely to have a cherished possession. However, they

differed in what they cherished. Women chose photo-graphs, while men chose what were called consumeritems. Csikszentmihaiyi and Rochberg-Halton (1981)interpret similar findings as indicating that cherishedpossessions of American women serve to maintain anetwork of social ties. Women's role in maintainingsocial ties through gift-giving in America has been notedpreviously (Caplow 1984).

Regardless of type of object chosen, women and mendiffer in their degree of attachment to their favorite ob-jects. Using a self-report seven-point scale, U.S. womenindicate a higher level of liking or attach ment to favoriteobjects than do men (mean for women = 5.98, mean

for men = 5.61, r = 2.38, = 0.018). Although bothwere instructed to select their favorite objects, womenreport a higher level of liking for the object than domen; However, in Zinder, th e four-point scale showedno significant difference between genders in the levelof liking for a favored object. As noted earlier, thesegender differences may be an artifact of locationallyconstraining object choice in the U.S.

Age and Favorite ObjectsAge also mediates the relation between the individual

and favorite object, but the overall relation is weaker

than that for gender (see Table 4). As Southwest Aicans age, they are less likely to choose functionaopposed to display items as favorite objects. In cultures, as adults age, they acquire social historyappears to be represented in objects. Younger peare in a life phase focused on accum ulation ofth e ftional items needed for independent living and expsion ofth e emerging self (Wells and G ubar 1966). Tappear to focus more on hedonic pleasures than omaintenance of intergenerational ties.

In the U.S. sample, the tendency to select an art oas a favorite object increases witli age. This may bterpreted as indicating that as individuals age, they

to establish a sense of purpose in life and a set of ithat are expressed in a favorite piece of art. Simirepresentational objects are often selected by Americans to show intergenerational ties with progeny and spouse. The life review process ofthderly culm inating in ego integration (Erikson 195volves a reflection on one's life. In this stage offamily photographs arranged in secular shrin es mtangible the success and fulfillment found through family of procreation.

Entertainment items are chosen as favorite obby all U.S. age groups, although the type of objects dYounger people chose stereos as their conne ction t

music and beat of their age cohort. A num ber of m iaged women ch ose the pianos that their children pduring childhood. These women, w ho seldom plapiano , apparently use the piano as a symbol of chiand their accomplishments. In fact, in many homeresearch photo graphs show the piano transformed a musical instrument into a secular altar on whichdren's and grandchildren's photographs are dispas a means of memorializing and recalling the meof one's children's music and the (real or imaghappiness and family togetherness at that stage idomestic cycle. Older Southwestern Americanschose entertain me nt objects typically chose a telev


11/18

A CROSS-CULTURAL INOUIRY 541

TABLE 5

REASONS FOR ATTACHMENT TO FAVORITE OBJECTS BYSOUTHWESTERN AMERICAN AGE GROUPS

Ag e

Personal andmaker attachment

n = 156)

Object basedattachment n = 102)

18-24 n = 45)25-35 n = 106)36-4 4 n = 55)45-54 n = 30)55-64 n= 10)65+ n = 12)

Total sample

53

6

6

63

6

83

6

47

4

4

37

4

7

4

NOTE; Numbers are percent

set. They often mentioned that it brought the world

nto their homes. For some with restricted mobility,contact with other humans was one-way, vicariouscontact via the television set.

Among Southwest Am ericans, as age increases, therewas an increase in the mention of personal and makerattachment reasons and a decrease in the mention ofobject-based characteristics as the reason for selectingthe item (see Table 5). This tendency accords with thenterpretation that age increases one s inc lination to

represent social history in a favorite object.However, the data do not indicate that degree of at-

achment to favorite objects increases with age. In fact,Erikson s (1959) interpretation oflife span developmentsees the psychological task ofthe elderly as acceptanceoflife as it was and acceptance ofthe inevitability ofdeath. This would imply that the elderly might exhibita gradual detachment from material objects in generaland favorite objects in panicular. This detachmentprocess is indirectly reflected in the findings of this re-search in two ways. Among Americans, length of fa-vorite object ownership increases with each age groupup to the 55-64 age group. However, it then sharplyurns down (one-way ANOVA F = 6.5, p < 0.0001).

The means for the 55-64 and 65 and older age groupshow a statistically significant difference from each other/ = 2.65, p = 0.015). although other adjoining age

groups do no t show such differences. These findings are

consistent with research on the more generalized phe-nomenon of materialism, which shows that materialismbears a curvilinear relation to age. peaking in middleage (Belk 1986). Is the oldest group gradually partingwith favorite objects, perhaps by passing them on toheir children or grandchildren prior to the time of

death? The question merits further study. second indication ofthe impact of aging on peop le s

elation to their favorite objects is indicated by theeven-point scale probing degree of liking ofthe favorite

object. The mean on this question was high {x = 5.8..d. = 1.4), as should be expected. However, mean re-ponses vary by age group. Favorite object liking in-

creases with age until 65, then declines sharply (one-way ANOVA F = 2.6, p = 0.02). Between the five agegroups from 18-24 through 55- 64, as age increases,liking of the favorite object increases. However, inmoving from 55-64 into the 65 and older age group,

liking ofthe favorite object declines to its lowest level.The difference between these two adjacent age groupsis statistically significant (/ = 2.24, p = 0.04). This isconsistent with Sherman and New man s (1977-78)finding that the old-old (those over 75) are less likelyto have a cherished possession th an the young-old (those65-75).

In Zinder, age also exerts an effect on the relationbetween individuals and favored objects, although theoverall relation is secondary to that of gender. Youngerpeople focus more on their hedonic pleasures within acultural age-related dialectic, even though th e possibilityof realizing individual hedonic pleasure through mar-

ket-mediated consumption is a recent phenomenon inNiger.For example, young brides-to-be (ages 13 to 16).

lacking the experience to make deliberative consump-tion decisions on the basis of comparison of functionalattributes of products, nonetheless play an innovativerole through their expressed desires regarding the bridaltrousseau. Because they are allowed whimsy and spon-taneity in their prenuptial status, their requests for novelconsumption goods are honored. Thus cheap quartzwatches, which othenvise have no place in Hausa life,have taken their place among the objects of adornment kayan ado) suitable for giving in marriage.

The unprecedented number of products recently in-troduced into the dowr>- is indicative of profoundchanges in Zind er s econo mic culture . Like other ap-parent absurdities in Third World consum ption of de-contextualized western objects (Arnould and Wilk1982). they symbolize both a recognition of the au-thority and power of occidental civilization and a loos-ening of formal strictures on the statuses to which peas-ants m ay aspire (Baudrillard 1968). Dowr>\ extendedfrom the b ride s family to the husba nd s, is unlikebridewealth, which is extended from the husb and s tothe wife s. Though the gifts given may be the sam e,bridewealth here is a form of gift exchange with all itsimplications of reciprocity and sociability (Hyde 1983;Meillassoux 1981), while dowr> isa form of^commoditytransfer (i.e., inheritance) with no such implications(Goody and T am biah 1973). The fact that dowry hasgrown in value and diversity relative to bridewealth in-dicates a change towards a more open-ended acquisi-tiveness on the occidental model.

Traditionally, young peasant men who stood in thedependent g ridu relationship to their fathers (Arnould1984a; Go ddard 1973; Hill 1972). had no well-definedrole in consumption or other realms of sociopoliticallife (Meillassoux 1981) as they had little or no con trolover household income. As in many other nonwesternsettings (Gregory 1982), their migration for wage labor


12/18

54THE JOURNAL OF CONSUMER RESEARCH

in the twentieth century has served as a conduit for theintroduction of novel consumer goods into villagecommunities. They now express a revised age relationthrough their preference for wearing small clo thes(Hausa. k ank anan kaya), which are secondhand, lo-

cally reconditioned western shirts and pants, and so-called functionary suits in opposition to the tradi-tionally styled long, loose shirt, baggy drawstring pants,and embroidered gown and hat (Hausa, manyan kaya)preferred by rich and elderly men. The new dichotomyin rural clothing styles expresses the longstanding ten-sion between fathers and sons (Hausa. biye: cf. Meil-lassoux 1981). Wage labor for the sons provokes tensionwithin the household over the disposition of labor, re-mittances, and o ther resources (Arnould 1984a, 1984b;Meillassoux 1981; Olivier de Sardan 1984). While thematerial terms ofthe opposition in social status betweenmen and their sons has changed, clothing style helps

mediate the tension.Young men prefer their style, which allows them toexpress their social differentiation through Westerngoods. It avoids direct com parison with the elders' styleand downplays any competition for resources betweenthem. Smali clothe s and functionary suits symboli-cally associate young men with the outside world anddisassociate them from the constraints of village life.The style also symbolizes their availability for flirtationwith unmarried women. To elder men. the wearing ofk ank anan kaya connotes an absence of pretense toa voice in public affairs and househo ld decisions. W ear-ing manyan kaya, often first worn when making formal

visits to prospective in-laws or at marriage, provides asymbol ofa younger man's intention to become a se-rious member ofthe community and to shed youthfulways. Thus, favorite objects are also used in Niger todenote age-related differences and statuses.

Relation to Favorite Objects in PhotosOverall. U.S. respondents indicated some physical

closeness to their favorite objects in the photographsby leaning toward the object or touching it, but onlyabout one-third of the respondents chose to hold orembrace the object.

Unlike our original expectations, there was almostno statistical relation between physical proximity to theobject in the photographs and self-reported attachmentto the object (Pearson correlation coefficient = 0.03 . p= 0.68, see Table I). Other interpretations o fthe mean-ing of physical proximity were therefore developed.Since there were no differences in physical closenessbetween either age or gender groups, an explanationbased on object meaning rather than self-concept wasexplored.

Those respondents whose attachment to the objectis based on person- or maker-based reasons tend to bephysically closer to the object when photographed thanare those whose attachment to the object is based on

intrinsic object-related meanings (/ = 2.38. p = 0.02That is, respondents whose attachment to the favoobject is based upon personal memories of other peopast experiences, or the maker of the object tendtouch or embrace their favorite objects in ph otograpHowever, respondents whose attachment to the favoobject is based upon characteristics of the object itare likely to be more physically distant from the obwhen photographed. Rather than being an expressof degree of object attachme nt or liking as was originpostulated, physical closeness to favorite objectsphotographs exemplifies an American expressionpersonal attachment to others vicariously through jects.

DISCUSSION AND CONCLUSION

Attachment to objects is a pervasive phenomen

Responden ts in both cultures identified favo rite jects. Favorite objects express aspects of self-concsuch as gender, age, and distinctive cultural backgro

People describe their favorite objects as reflecpersonal meanings and attachments in both the Uand Niger. However, cross-cultural comparison shthat while the kinds and range of favorite objects vafavorite objects serve as cultural icons that reflect lculture as experienced by the individual. The wmeanings of objects may not be consciously availto the informants, yet they become clear when crcultural comparisons are made of average levels otachment and types of objects selected.

While the emic perception of Southwest Ameriis that favorite objects represent unique, individualtory, in fact, conventional meanings such as male mtery over the environment and female connectionfamily are encoded. Meanings of favorite objectsconventional in Niger, but informa nts do not stresindividuality of such meanings. Instead, conforwith shared meaning is often emphasized.

Nonetheless favorite objects do provide individized cues for self-expression. Among Southwest Aicans, affective mem ories of personal exp eriences operson who made the item for the owner are often bolized. This form of favorite object a ttach me nt isociated with stronger liking for the object than is ocharacteristic-based attachment. Attachments to observing as memory cues co-exist with higher levesocial linkage. Thus, favorite objects most often sas symbols of, rather than replacements for, closeterpersonal ties.

These objects provide individual so lutions to thmogenization of value and emphasis on socially grative meanings inherent in mass-produced objecwell as the need for individual expression. Indivisingularize things through the mutu al transfer of ming and emotion between the objects and the indivi(McCracken 1986). Singularization deactivates obas commodities and turns them into priceless


13/18


seemingly unique icons for individual self-expression(Kopytoff 1986).

For the Nigerien sample, fewer types of objects wereselected as favorites, reflecting not only the smallernumber of consumer objects owned, but also individual

commitment to a restricted set of cultural values. Fewerkinds of objects were favored by Zinderois than bySouthwest Americans and virtually all were handmade.This is a predictable result in a recently mo netized, vir-tually advertising-free culture. The meanings attachedto objects from which and to which people transfermeanings (McCracken 1986) serve to link individualsto reference groups either cooperatively as with men'sKoranic texts and women's silver bracelets, or com-petitively, as with men's horses and women's woventapestries. Nonetheless, within conventional structuresof meaning there is room for innovation and personaldifferentiation. Purchase decisions for such things as

quartz watches or Western clothes are made with ref-erence to culturally available ideas about consu mp tion,gender, and age roles, as well as notions derived fromexotic m odels.

It appears that favorite object attachment is concep-tually and empirically distinct from the more generalpossessiveness component of materialism. The posses-siveness component of materialism has very differentsalience and substantive m eaning cross-culturally. Thisderives from the different world views (Weltanschauung)ofthe two cultures (Judeo-Christian vs. Islamic-animist)and the way in which objects are used, as well as theway object ow nership is used in the self-definition and

self-expression processes. If. in the Sou thwest U.S .. sta-tus is measured by what one has. in rural peasant Niger,wealth in people {arzikin mutanc . expressed throughthe circulation of conventional objects (especiallybridewealth) with shared meanings continues to havecultural significance. From this research, it does notappear that materialism expressed through generalizedpossessiveness is a cultural universal.

Based on this research, we question whether it is. infact, possible to abstract the meaning of materialismfrom particular cultural contexts. Clearly, the Belk(1984. 1985) materialism scales are well designed tomeasure Western informants' ethnocentric conception

of materialism. Belk's seminal conceptual and scale de-velopment w ork is an an chor for later research, but thescales themselves are not a universal empirical solutionto measuring materialism cross-culturally. Fortunately,our interpretation of the Nigerien data was not solelydependent upon scaled surveys for developing an un-derstanding of relations to objects. At this point, werecommend that the original Belk scales be treated asappropriate only to the culture in which they were de-veloped. For cultures other than the United States, scaledevelopment should be based on thorough ethnographicstudies of the meaning and expression of materialismin that culture. We recognize that our suggestion maypreclude the development ofa scale to measure mate-

rialism that is generalizable across cultures becausethe concept may vary so widely in its cross-culturalmeaning.

From this research, it appears tha t favorite object at-tachment is also conceptually and empirically distinct

from social linkage. Rather than serving as substitutesfor a social network, favorite objects serve to solidifyand represent both on e's connection s to and differencesfrom others. Thus, favorite object attachment does notappear to be an expression of loneliness, but rather anexpression of connections to others. Our research sug-gests this relationship is valid cross-culturally. The eth-nology of exotic, gift-based economies shows sociallinkage, object attachment, and possessiveness developparticular logical relationships all of their own (e.g.,Goodaie 1985; Gregory 1982; Leach and Leach 1983;Malinowski 1922).

The data indicate that women emphasize social tiesthrough favorite objects. Men represent their accom-plishments and mastery in favorite objects. Given thepatriarchal structure of both cultures studied, this resultis not surprising, but it would be necessary to comparethese results with data collected in matrilineal or ma-triarchal societies before generalizing to a constant gen-der effect rather than culture effect.

Age differences in favorite object attachment seemto represent changing m eanings during different lifephases and in cultural and economic history. In theSouthwestern U.S.. the break between the groups aged55-64 and 65 and older showed a sharp disjuncture incontrast to the more continuous evolution through theearlier life phases. The oldest age groups showed amarked decline in length of ownership as well as likingofthe favorite object. In Zinder. distinctive consump-tion behavior was found among young marriage-agepeople, who are most likely to be exposed to novel ob-jects.

An effort to cross-validate degree of attachment tofavorite objects using sur ey and photographic meth odsinstead provided two different, but complementary,perspectives on object attach men t. Photographs capturea different aspect of a person's relation to a favoriteobject than do survey self-reports. Physical closeness tothe favorite object in the photograph more clearly ex-

pressed closeness to the individual for whom the objectstood rather than degree of attachment to the favoriteobject.

Our perspective has been primarily social structural(e.g., culture, age, and gender) and economic in speci-fying object m eanings. We go beyond Douglas and Ish-erwood (1979), who see objects primarily as points thatmark patterns of social relationship. Our research sup-ports the idea that object preference is built up afterpurchase through a dialectical process in which meaningand affect are transferred between individuals and ob-jects over time, as suggested by Baudrillard (1968),Csikszentmihaiyi and Rochberg-Halton (1981), Levy(1981), and McCracken (1986). Building on this work.


14/18

54 4THE JOURN L O CONSUMER RESE RCH

we have tried to decipher some of what is transferredin this process. In addition, we have tried to capturesome ofthe dynamism and conditionality inherent inthese processes that allow for both stability in meaningsand changes in types of preferred object through time.

To more fully understand the meaning of possessionsand the dynamics of such systems, further research isneeded to systematically explore the transmission ofobjects between individuals within families or house-holds. The research finding that many favorite objectswere gifts points to the importance of gifts to recipients.Gift-giving, particularly the giving away of one's ownpossessions, needs to be systematically explored lon-gitudinally by studying gift-givers and the system ofmeanings they attempt to convey with the gift. In thiscontext, studies of matched pairs of heirloom gift-giversand receivers would be particularly enlightening, aswould studies of systems in rapid transition such as that

in Zinder.In addition, our cross-cultural perspective has shown

that more research is needed to explore how preferencesfor favorite objects change both within the lifespan ofindividuals and through time as changes in consump-tion patterns occur, particularly in developing econ-omies.

This research has not addressed the reasons why par-ticular objects become cultural icons and not others.Why pianos and silver bracelets rather than guitars andcalabash covers? Later research should build on the un-derstanding that objects veil an underlying flow of socialrelationships (Douglas and Isherwood 1979) to deter-mine why particular objects are chosen for this task. Byfocusing on particular favorite objects, this research hasattempted to explore a portion of the consumption andownership processes.

[Received January 1987 Revised September 1987 ]

REFERENCESAmin. Samir (1973). Neo-Cokmialism in West .ifrica. New

York: Monthly Review Press.

(1976). Unequal Development: . in Essay on SocialFormations of Peripheral Capitalism. New York:Monthly Review Press.

Appadu rai. Arjun (1986), Introd uction : Comm odities andthe Politics of Value, in The Social Life of Things:Commodities in Cultural Perspective, ed. Arjun Appa-durai. Cambridge University Press, 3-64.

Arnould. Eric J. (1984a), Ma rketing and Social Reproduc-tion in Zinder, Niger Republic, in Households: Com-parative and tlistorical Studies ofthe Dom estic Group .eds. Roben M. Netting. Richard R. Wilk. and Eric J.Arnould. Berkeley. CA: University of California Press.1.10-162.

(1984b). Process and Social Fo rmation: Petty Com -modity Producers in Zinder. Niger, Canadian Journalof African Studies. 18 (3). 501-523.

(1985). Eva luating Regional Econom ic Develop-

ment: Results ofa Marketing Systems Analysis in ZindProvince, Niger Republic, The Journal of DevelopinAreas, 19 (2), 209-24 4.

and Richard R. Wilk (1982), Why Do the NativWear Adidas? Anthropological Approaches to ConsumResearch, in Advances in Consumer R esearch, Vol. ed. Thomas C. Kinnear, Provo, UT: Association fConsumer Research, 748-75 2.

Baier, Stephen A. (1974), African Merchants in the ColonPeriod: A History of Commerce in Damagaram CenNiger). 1880-1960. Ann Arb or: Univers ity Microfil

(1976), Econ omic History and the Sahelian Ecomies of Niger, Journal of African H istory, (1), 1-16

Baudrillard. Jean (1968), Le Systeme des Objets, Paris: Glimard.

Becker, Howard S. (1975), Photog raphy and SociologyAterimage. 3 (May-June), 1-2.

Belk. Russell W. (1983), W orldly Possessions: Issues anCriticisms, in Advances in Consumer Research. Vo10, eds. Richard P. Bagozzi and Alice M. Tybout, AnArbor. MI: Association for Consumer Research, 514519.

(1984), Th ree Scales to Measure Constructs Relato Materialism: Reliability, Validity and Relationshipin .Advances in Consumer Research. Vol. 11. ed. ThoC. Kinnear. Provo. UT: A ssociation for Consumer Rsearch. 2 91-297 .

(1985). Materialism: Trait Aspects of Living in thMaterial World. Journal of Consumer Research, (December). 265-280 .

(1986). Generational Differences in the MeaningThings. Products and Activities. in .Advertising Consumer Psychology, eds. Keith Sentis and Jerry Cson. New York: Praeger. 199-213.

(1987a). Possessions and the Extended Self. workpaper. Depanitient of Marketing. University of UtSalt Lake City. UT 84112.

(1987b), The Role ofthe Odyssey in Consumer Bhavior and Consumer Research. in .Advances in sumer R esearch, Vol. 14. eds. Melanie WallendorfPaul F. Anderson. Provo. UT: Association for ConsuResearch. 357-361.

and Richard Pollay (1985), Imag es of Ourselve Good Life in Twentieth C entury Adven ising. JourofConsumer Research, 11 (March), 887-89 6.

Bloch. Peter H. and Marsha L. Richins (1983), A TheoModel for the Study of Product Im ponan ce PerceptJournal of Marketing. 47 (Summer), 69-81.

Bott. Elizabeth (1971). Family and Social Network: RNorms, and External Relationships in O rdinary UrFamilies. New York: Free Press.

Campbell. Donald T. and D.W. Fiske (1959). Converand Discriminant Validation by the Multitrait-Mmethod Matrix, Psvchological Bu lletin, 56 (2), 81-

, W.H. Kruskal. and W.P. Wallace (1966). SeaAggregation as an Index of Attitude, Sociometry(1). 1-15.

Caplow. Theodore (1984). Rule Enforcement W ithouible Means: Christmas Gift-Giving in MiddletowAmerican Journal of Sociology. 89 (May). 1 306-13

Carroll. Marilyn (1968). Junk Collections Among MeRetarded Patients. .American Journal of Mental ciency. 73(2 ). 308-314.

Childe. V. Gordon (1936), .Man Makes Himself LonWatts & Com pany.


15/18

A C ROSS CULTURAL INQUIRY 54 5

Collier, John, Jr. (1967), Visual Anthropology: P hotographyas a Research Method. New York: Holt, Rinehart &Winston.

and Maleom Collier (1986), Visual Anthropology-Photography as a Research Tool, Albuquerque: Univer-sity of New Mexico Press.

Crow, John E. and Helen K. Henderson (1979), Append ix9: A Report of Sur%ey Findings from Three Villages inZinder, Nige r, in F inal Report ofthe Natural ResourcePlanning Project for the Province of Zinder. Niger, eds.Arid Land s Natural R esources Comm ittee, Tucson , AZ:University of Arizona.

Csikszentm ihaiyi, Mihaly and Eugene R ochberg -Halton(1981). The Meaning of Things: Domestic Symbols andthe Self. Cambridge: Cambridge University Press.

Danforth. Loring M . (1982). Death Rituals of Rural Greece.Princeton, NJ: Princeton University Press.

Davis. James A. (1955). Status Symbols and the M easure-ment of Status Perception, Sociometrv, 19 (3), 154-165.

de Jan vr>', Alain (198 1). The Agrarian Question and Reform-ism in Latin .America, Baltimore: Johns Hopkins Uni-versity Press.

Devillers, Carole (1983). W hat F uture for the Wayana In-dians? National Geographic. 163 (January). 66 -82 .

Douglas, Mary and Baron Isherwood (1979). The World ofGoods: Towards an Anthropologv of Consumption, New ^'ork: W.W. Norton.

Erikson. Erik (1959). ldentitv and the Life Cvcle, New York-W.W. Norton.

Evans. Franklin (1959). Psychological and Objective Factorsin the Prediction of Brand Choice: Ford Versus Chev-rolet. Journal of Business. 32 (October). 340-369.

Evans. Peter (1977), Dependent Development. Princeton. NJ:Princeton University Press.

Evans-Pritchard. E.E. (1940). The Nuer. London: OxfordUniversity Press.

Farb. Peter and George .^rmelagos (1980). Consuming Pas-sions: The Anthropology of Fating. New '^'ork: Wash-ington Square Press.

Felmlee. Diane. Donna Eder. and Wai-Ying Tsui (1985). Peer Influence on Classroom Attention. Social Psy-chology Quarterly. 48 (September). 215-226.

Felson. Marcus (1976). Th e Differentiation of MaterialLifestyles: \925-l966. Social Indicators Research. 3 0-4). 397-421.

Feshbach, Seymo ur and Norma Feshbach (1963), Influence

ofthe Stimulus Object Upon the Complementary andSupplementar>' Projection of Fear, Journal ofAbnormaland Social Psychology 66 (5). 498-502.

Franke. Richard W. and Barbara H. Chasin (1980), Seeds ofFamine: Ecological Destruction and the DevelopmentDilemma in the West African Sahel, Totowa, NJ: Row-man & A llanheid.

Furby. Lita (1978), Possessions: Toward a Theory of TheirMeaning and Function Th roughout the Life Cycle, inLife Span Development a nd Behavior ed. Paul B. Baltes.New York: Academic Press, 297-336.

and Mary Wilke (1982), Som e Characteristics of In-fants' Favorite Toys. The Journal o f Genetic Psvchologv140 (June), 207-219.

Godd ard, A.D. (1973), Changing Family Structures Amongthe Rural Hausa, .Africa, 43 (3), 207-219.

Goffman, Erving (1959), Th e Presentation of Self in Everyda yLife, Garden City, NY: Doubleday.

(1961), Asylums, Garden City, NY: Anchor.Goldschmidt, Walter (1969), Kambuya's Cattle: The Legacy

of an African Herdsman, Berkeley, CA: University ofCalifornia Press.

Goo daie. Jane C. (1985), Pig's Teeth and Skull Cycles: BothSides ofthe Face of Humanity, American Ethnologist,12(2), 228-244.

Goody, Jack R. and S.J. Tambiah (1973), Bridewealth andDowry, Cambridge: Cambridge University Press.

Gregor>\ Christopher A. (1982), Gifts and Commodities,London: Academic Press.

Hall, Edward T. (1969), The Hidden Dimension, Garden City,NY: Anchor.

Heisley, Deborah D. and Sidney J. Levy (1987), Fam iliarInterlude: Autodriving in Consum er Analysis, workingpaper. Department of Marketing, Northwestern Univer-sity, Evanston, IL 60201.

, Mar>' Ann McGrath, and John Sherry (1987), Farmers' Market: An Analysis of An Alternative Mar-keting System. paper presented at the American Mar-keting Association Winter Educators' Conference, SanAntonio. TX. February.

Hill. Polly (1972). Rural H ausa: A Village and a Setting.Cambridge: Cambridge University Press.

Holbrook. Morris B. and Elizabeth C. Hirschman (1982). The Experiential Aspects ofConsumer Behavior: Con-sumer Fantasies. Feelings, and Fun. Journal of Con-sumer Research, 9 (September). 132-140.

Holm an. Rebecca H. (1980). Clothing as Com mu nication:An Empirical Inv estigatio n. in Advances in ConsumerResearch. Vol. 7. ed. Jerry C. Olson. Ann Arbor. MI:Association for Consumer Research. 372-377.

Hsu. Francis L.K. (1985). Th e Self in Cross-Cu ltural Per-spective. in Culture and Self: Asian and Western Per-speciivcs. eds. Anthony J. M arsella. George D eVos. andFrancis L.K. Hsu. New '^'ork: Tavistock. 24-55.

Hyde. Lewis (1983). The Gift: Imagination and the EroticLife of Property. New York: Vintage.

Imanishi. K. (I96 0). Social Organization of Subhum an Pri-mates in Their Natural Habitat. Current Anthropologv.1 (5-6). 393-407.

Issac, Glynn. et al. (1979), Human .Ancestors, readings fromScientific Am erican, San Francisco: W. H. Freeman.

Jacoby, Jacob and Robert W. Chestnut (1978), Brand Loyalty: Measurement and Management New' York: John Wiley.

Kirk-Green. Anthony H.M . (1974), M utum in Kirkii: The

Concept of the Good Man in Hausa. unpublishedmanuscript. African Studies Program, University of In-diana. Bloomington, IN.

opytoff Igor (1986), Th e Cultural Biography of Things:Comm odization as Process, in The Social Life ofThings: Comm odities in Cultural Perspective, ed. AjunAppadurai, Cambridge: Cambridge University Press. 64-95 .

Kron. Joan (1983), Home Psych. New York: Crown.Laum ann. Edward O. and James S. House (1970). Living

Room Styles and Social Attributes: The Patterning ofMaterial Artifacts in a Modern Urban C om mu nity.So -ciology and Social Research, 54 (April). 321-342.

Leach, Jerry W. and Edmund Leach (1983). The Kuta: New

Perspectives on Massim Exchange, Cambridge: Cam-bridge University Press.


16/18


17/18

A CRO SS CULTURAL INQUIRY 547

Tournier, Paul (1981), The Gift of Feeling London: SCMPress.

Turnbull, Colin (1972), The Mountain People New York:Simon & Schuster.

Turner, Victor (1969), The Ritual Process: Structure and Anti-

Structure Chicago: Aldine.Tweedie, Penny (1980), Arnh em Land Aboriginals Cling toDreamtime, National Geographic 158 (November),644-663.

Van Gennep, Arnold (1960), The Rites of Passage Chicago:, University of Chicago Press.

Veblen, Thorstein (1899), T he Theory ofthe Leisure ClassNew York: Random House.

Verdo n, Michael (1982), W here Have All Their LineagesGone? Cattle and Descent Among the Nuer,AmericanAnthropologist 84 (3), 566-579.

Wachtel, Paul L. (1983), Th e Poverty of Affluence: Psycho-logical Portrait of the American Way of LifeNew York:Free Press.

Wagner, John. ed. (1979). Ima ges of Information: Still Pho-tography in the Social Sciences Beverly Hills: Sage.allendorf Melanie and Michael Reilly (1983), Eth nic Mi-

gration, Assimilation, and Consumption, Journal ofConsumer Research. 10 (December). 292-302.

and Rob en Westbrook (1985), Em otion s and Cloth-ing Dispo sition, paper presented at the Annual M eetingof the Association for Consumer Research, Las Vegas,NV, October.

Warner, W. Lloyd, Marchia Meeker, and Kenneth Eells(1960), Social Class in America: Manual of Procedurefor the Management of Social Status New York: Harper.

W atts, Michael J. and Th om as J. Bassett (1986), Crisis andChange in African Agriculture: A Com parative Study ofthe Ivory Coast and Nigeria, African Studies Review28 . 3-27.

Weisberg, Paul and James E. Russell (1971), Prox imity andInteractional Behavior of Young C hildren to Their 'Se-curity' Blankets, Child Development 42 (5), 15 75-1579.

Weissner, Polly (1984 ), Reco nside ring the Behavioral Basisfor Style: Case Study among the Kalahari San, Journalof rchaeological Research 3, 190-234.

Wells, William D . and Geo rge Gu bar (1966 ). Th e Life CycleConcept in Marketing Research, Journal of Marketing

Research. 3 (November), 355-36 3.Wentzel, Volkmar (1978). Zulu King Weds a Swazi Prin-cess. National Geographic 153 (January ), 47 -61 .

Worseley, Peter (1968), Th e Trumpet Shall Sound New '^'ork:Schocken.


18/18

my favourite things object attachment comparative.pdf

Documents