Chapter 8 Culture As Process (Bab 8 Budaya Sebagai Proses)

This chapter links concepts and methodology for the study of culture as psychological process. Methods must be based on a conceptual framework. In this chap­ter, the framework is the notion of culture as psychological process (Cole, 1995; Valsiner, 1989). This framework leads to particular kinds of research questions, which, in turn, implicate particular kinds of methods that are capable of address­ing the original research questions. Because of the tight link between theory and method, methods are not theoretically neutral, but contain an implicit theory (e.g., Ochs, 1979). Beginning with a contrast between cultural and cross-cultural psychology, we lay claim to a series of methodological principles and practices that can guide research on cultural action and interaction, while revealing the ontoge-netic, sociohistorical, and phylogenetic origins of human cultural processes. Where possible, sample studies that serve as models for a particular research method will be presented or cited.
Conceptual Framework
As a species, humans are biologically primed to create, acquire, and transmit culture. The creation, acquisition, transmission, and use of culture are psycho­logical and interpsychological processes. It is the study of these processes that is the paradigmatic subject of cultural psychology. This chapter asks what methods and what methodology' are suitable to the study of culture as psychological and interpsychological process.

The social interaction among beings biologically primed for culture creates culture for the group and for the individuals in it. Culture is therefore viewed as a socially interactive process with two main component processes: the creation of shared activity (cultural practices) and the creation of shared meaning (cultural in­terpretation). Empirical methodology must be adequate to study both shared mean­ing and shared activity, the two major embodiments of shared cultural knowledge.
Culture as shared activity is an important focus of methods for the study of everyday life. However, it is the human capacity to create shared meaning that produces the distinctive methodological contribution of cultural psychology. In­deed, the centrality of creating meaning in cultural processes and human life has profound methodological consequences. This is because not only subjects but aiso researchers are engaged in a meaning-making process.
How to do research that minimizes the researcher's projection of meaning onto the subject and sheds maximum light on the subject's creation of cultural meanings is a major methodological issue in cultural psychology. This issue leads to a number of specific topics in this chapter: the importance of researcher per­spective, research as a communication process, and the methodological role of anthropology in cultural psychology.
Both components of cultural process, shared meaning and shared activity, are cumulative in nature. This is because culture is created by processes that oc­cur between, as well as within, generations. Meanings and activities not only accumulate but also transform over historical time. This cumulative and tempo­ral characteristic of culture creates the necessity for developmental methods for studying the transmission and socialization of culture from the older to the newer generation. It also creates the necessity for historical methods of studying culture as psychological process. Developmental as well as historical methodologies for cultural psychology (as well as their various combinations) will therefore be ma­jor topics of this chapter.
Because of the socionistorical specificity of each group's cultural meanings and shared activities, a case is made for the use of culture-specific procedures. An important (and somewhat unusual) daim is that culture-specific procedures are compatible with the discovery of generalizations about the psychology of culture that transcend particular cultural settings. Such discoveries require the field to transcend methodological behaviorism (Hatano, personal communica­tion, 1996). Methodological behaviorism requires researchers to equate proce­dures on the level of the smallest unit of experimenter behavior and subject re­sponse. In so far as cultural psychology engages in the comparative enterprise, its methodological goal must be to equate procedures at the deeper level of cul­tural meaning. (See also discussion by Sinha, this volume, of the compatibility of indigenous and universal psychologies.)
In order to provide a context for the methodological perspective that will be developed, the chapter begins by first providing an extended example of cultural process and then discussing the relationship between cultural and cross-cultural psychology. Major sections on metamethodology and methodology for the study of cultural processes follow. In the final section, the substantive consequences of this methodology are discussed.
An Example of Cultural Process
Both shared activities and shared meanings are intrinsic to the human mode of adaptation for survival. They are two facets of shared cultural knowledge. An example will elucidate this theoretical formulation of cultural processes. At the same time the example indicates the subject matter for which we must specify our methods and methodology.
During the Los Angeles earthquake of 1994, many of the ecocultural sup­ports of everyday survival and life, such as water, electricity, and roads, were destroyed. In small groups and through the media, people developed new shared knowledge concerning survival activities, such as where to get water, how to circumvent damaged roads to get from point A to point B, and methods for de­tecting leaking gas. Expertise was shared with novices, as when a contractor showed his neighbors how to turn off their gas, or a ham radio operator provided news of the location and magnitude of the earthquake in the absence of electric­ity. The nature of culture as a tool for organizing everyday life (Weisner, 1994) was quite apparent.
 Symbolic communication, through both language and visual media, is a criti­cal means by which social sharing takes place; communication processes were quite intense during this period of adapting to the physical conditions created by the earthquake. As a result, new shared activities that enhanced physical sur­vival were created through cuiturai processes of sociai interaction.
Simultaneously, though, shared meanings were also created to rationalize and understand the events that had taken place. Like shared activities, shared meanings arose through communication processes. One shared meaning that developed was the custom of asking people how they fared in the earthquake, the normative reply was, "I was fortunate." The search to create shared meaning for a stunning physical event was particularly apparent when, a few days after the earthquake, a local public affairs radio show host convened clergy from many religions to ask them about the larger meaning of the earthquake. His question was. "Did God send the earthquake to punish Los Angeles?" Clearly, adaptation to the aftermath of the earthquake could not be reduced to a process of adapting to physical conditions; the interpretation of these conditions, that is, processes of creating meaning were part and parcel of the shared culture that developed in response to the earthquake.
This example provides a model of processes that are assumed to occur when­ever a new member of society is born: the creation of shared knowledge, activi­ties, conventions, and meanings through communication and social interaction. This microdevelopmental example of culture recreation by adults occurs in each generation in the macrodevelopmental processes of children. The example is also a modei and metaphor of culture change provoked by new ecological conditions. Finally, this example illustrates the potential for cultural variability as a response to different ecological conditions (Berry, 1976).
If this is the nature of culture as process, the question at hand is what empiri­cal methods will be useful for its empirical study. It is necessary to have an array of techniques that can address the development of shared activities, the develop­ment of shared meanings, and the communicative processes through which they are acquired. Cultural psychology has (and continues to develop) empirical meth­ods that can systematically document processes of cultural creation such as these.

T he Relationship between Cuiturai

ne Keiationsnip between i_uitur; and Cross-Cultural Psychology

The terms cuiturai psychology and cross-cultural psychology are each fuzzy con­cepts with partially overlapping sets of exemplars. There are focal (Rosch, 1973) or paradigmatic instances of the two approaches that are sharply distinctive, but there are also many research exemplars that combine attributes of each approach. This is because the methods are more complementary than competing. However, paradig­matic definitions will be used to illustrate the methodological consequences of each. In cross-cultural psychology, culture is generally operationalized as an ante­cedent variable (Berry, 1976). This is the perspective offered by Van de Vijver and son. Cultural psychologists, in contrast, study cultural processes directly; they rely much less on "packaged" or indexical variables in their research designs.
Just as cross-cultural psychology "packages" culture in independent vari­ables, it also "packages" the individual subject in dependent variables. In the same way that independent variables are stand-ins for complex cuiturai processes, so too dependent variables generally function as indices of individual cultural processes, rather than constituting the processes themselves. A dependent vari­able is something that can be measured; it often functions to summarize a pro­cess. Again, cultural psychology attempts to study the process itself.
An example of this distinction between dependent variable and individual cultural process is the contrast between qualitative process analysis and a continu­ous quantitative scale. This point will be illustrated with a particular study. Greenfield, Raeff, & Quiroz (in press) developed scenarios that are based on real-life value conflicts experienced by Latino families who immigrate to Los Angeles. One scenario is the following:
Anna and Christina both got ten dollars from their uncle. Christina buys a blouse. A week later Anna wants to wear Christina's blouse, and Christina says, "This is my blouse, and I bought it with my own money." Anna says, "But you're not using it now." Christina tells their mother.
The subject is then asked: What do you think the mother should do?
Immigrant mothers (who were given the scenarios in Spanish) were frequently coliectivistic in their responses (Let Anna borrow the blouse). In contrast, their children were most often struggling to resolve and harmonize the coliectivistic orientation of their families with the individualistic orientation of their school in particular and the dominant society in general. For example, one fourth grade girl answered that the mother should tell Anna to pay Christina for the blouse and then Christina can buy another blouse for herself. According to a qualitative process analysis, this is a response that integrates the individualistic concept of personal property with the coliectivistic concept of sharing with extended family members. If, in a contrasting quantitative approach to the data, we were to place this response midway on an interval scale going from collectivism to individual­ism, we would miss the dynamics of the psychological process of cross-cultural value integration that this child has accomplished. Individualism/collectivism as a continuous dependent variable would become a packaged index of a cultural process; the package would hide the process itself.
Cultural Process or Cross-Cultural Comparison
Cross-cultural comparison is the method of choice for cross-cuitural psychology (Triandis, 1980, p. 1). Comparisons are at the heart of quantitative methodology in psychology. Through statistical analysis, psychology uses comparison to pin­point differences. Cross-cultural psychology is based on a comparative method­ology in which statistics are used to identify different frequencies in different cultural groups of a phenomenon of interest (cf. Van de Vijver & Leung, this volume). To do this, it is necessary to assume that what is being measured is the same across groups, that only the frequencies vary. However, this is often not the case. Different adaptations and different systems of meaning, in response to a history of different conditions, imply that phenomena are qualitatively, not merely quantitatively different in different cultural groups. Whereas much discussion in cross-cultural methodology has focused on the best techniques for gaining compa­rable measurement (Berry, et al., 1992), the notion in indigenous psychology (Kim & Berry, 1993; Sinha, this volume), as in cultural psychology (Sn'gler & Shweder, 1990), is that the very phenomena are different. An important methodological conclusion flows from this idea: there can not be comparable measurement of incomparable phenomena. The measuring instruments themselves must change.
Cross-cultural comparison is not eschewed in cultural psychology (see, for example, Miller, Bersoff, & Harwood, 1990). For example, Shweder and Sullivan (1993) speak of cultural psychology as "a designation for the comparative study of the way culture and psyche make each other up" (p. 498). However, compari­son is not central in the methodology of cultural psychology. Indeed, cross-cul­tural comparison tends to be done both cautiously and differently: distinctive procedures are often used for each culture being compared (e.g., Morelli, Rogoff, Oppenheim, & Goldsmith, 1992). This is also true of the cross-indigenous ap­proach advocated by Kim and Berry (1993).
The methodology of cultural psychology is therefore distinct from the psycho­metric approach to cross-cultural psychology (see chapter by Van de Vijver & Leung, this volume). There, methods used in each culture should, ideally, be formally equivalent. Psychometrics assume that if a questionnaire is used in one culture, the questionnaire format must be used in all others being compared; and that each item must have a corresponding item in all versions/translations. In prac­tice, though, cross-culturalists have frequently accepted variations in question­naire content, in order to take local cultural meanings into account. In contrast, the position of cultural psychology is that one must communicate with subjects in each culture in a way that is comfortable and appropriate to that culture. This will lead to the use of very different methods to study the same issues in different cultures. Comparison and the testing of universals will come at the more abstract conceptual and theoretical levels, not on the level of concrete methods and spe­cific behaviors generated by formally equivalent procedures. The use of parallel procedures across cultures, the hallmark of cross-cultural psychology, works best when cultures are not too different—for example, when all of the samples have had formai education (e.g., Hofstede's 1980 cross-nationai study of individual­ism and collectivism). The use of qualitatively different procedures across cultures works best when the cultures are very different, and when they have different epistemological and communicative presuppositions, topics discussed later in this chapter. As Triandis (1995) points out, cross-cultural psychology provides a methodology for comparing similar cultures, whereas cultural psychology pro­vides a methodology for comparing dissimilar cultures,
An important facet of both cross-cultural comparison and the study of cul­tural process is the selection of cultures to be studied. The ideal in cross-cultural psychology is to select the cultures to be compared, based on theoretical analysis of the independent variables (Berry et al., 1992), In contrast, cultural psychology, like cultural anthropology, puts a premium on deriving procedures and prob­lems from each culture (Wassman, 1995); this requires in-depth knowledge of the culture. Therefore, in cultural psychology, cultures are often chosen because the researcher knows the ianguage or otherwise has a good entree into the culture. Because of this requirement of in-depth cultural knowledge, including language skill, cultural psychologists most often research a single culture. Indeed, they often research their own cultures (e.g., Lave, Murtaugh, & de la Rocha, 1984; Valsiner & Hill, 1989).
When cultural psychologists make cross-cultural comparisons, they often deal with different cultures or social groups within a single country (e.g., Greenfield, 1966; Greenfield, Reich, & Olver, 1966; Wagner, 1984; Wassman & Dasen, 1994; Scribner & Cole, 1981; Serpell, 1993). This type of research design provides better control than a simple comparison between groups in two different countries; it also lies in that overlapping or border area between cross-cultural and cultural psychology.
Because of the labor-intensive methods used within each culture, cultural psychologists are often dependent on other researchers to collect data when they wish to compare cultures across national borders (e.g., Greenfield & Childs, 1978; last chapter of LeVine et al., 1994; Saxe, 1981). However, because comparison is on an abstract theoretical level and because the procedures for each distinctive culture are derived from that culture, the need for exact replication does not come into play.
Removing Unconstructive Stereotypes of Cultural Psychology
To define culture as process and to look for appropriate methods of studying cultural processes removes the present approach from some common miscon­ceptions in the field about cultural psychology and its differences from cross-cultural psychology:
1.         This chapter will not assume the absence of universal psychological pro­
cesses. The view of this chapter is that universal psychological processes clearly
exist and therefore need to be incorporated into methodology and theory. From
the theoretical perspective, this is clearly implicit in the idea, expressed earlier,
that humans are biologically primed for culture, as well as for cultures. Indeed,
there is a psychic unity of humankind.
Shweder's formulation is the notion of "one mind, many mentalities" (LeVine & Shweder, 1995; cf. Shweder, 1995). In this chapter, one mind is viewed as the human capacity for cultural learning (Tornasello, Kruger, & Ratner, 1993), and many mentalities as the result of cultural variability in the conditions presented by different cultures, each with its own distinctive ecology and economy.2
2.         This chapter rejects a dichotomous choice between understanding behav­
ior in its cultural context (stereotypical of cultural psychology) and understand­
ing behavior as indices of universally shared psychological processes (stereotypical
of cross-cultural psychology). Cultural does not mean context-bound, as opposed to universal. All behavior is both relative to a context and representative of uni­versal principles or laws. In other words, to act in a cultural context is in itself a major universal principle of behavior. To place a particular conclusion from re­search in one cultural context into a universal framework is to move to a more general level of conceptualization; it is not to reject the notion of contextualized behavior. Methodology that permits the transfer of contextualized research to general principles of psychology will be one of the foci of this chapter.
Objectivity versus Perspective
Modern psychology was born from the methodological ideology of objectivity, the erasure of perspective, generally known as bias. In sharp contrast to this tra­dition, an important tenet of cultural psychology (also voiced by Miller in her theoretical chapter in this volume) is the logical impossibility of an observer-independent or objective perspective. In cultural anthropology, the notion that results are relative to the position (cultural, class, gender) of the observer has led to self-flagellation and the total rejection of empiricism (Patai, 1994). Cultural psychology, in contrast, is developing methods and concepts suitable to the in­clusion of observer perspective in research.
When studying behavior in one's own culture (as most psychologists do), one has an insider's cultural perspective. Partly because this fact runs counter to the very ideological assumptions of psychological science, the insider's perspec­tive almost always goes unacknowledged (cf. Rogoff & Morelli, 1989). Yet this perspective is crucial. With reference to one's own group, the insider understands the meanings and motives behind in-group behaviors, meanings and motives that may be misinterpreted or devalued by outsiders looking through the lenses of their own cultural values (Berry, 1979).
An example of the unacknowledged insider perspective is the topic of self-esteem in U.S. psychology. Not until the work of Markus and Kitayama (1991) did it become apparent that self-esteem is not a universal quality, but a culture-specific ideal. In sharp cultural contrast, Markus and Kitayama noted the impor­tance of self-effacement, rather than self-esteem, in the development of a Japa­nese person. Note too that this bicultural team of researchers reflects both an insider and an outsider perspective on both cultures they have studied. This is probably an important reason why they were able to remove the cultural blinders informing self-esteem research in the United States.
In essence, the insider 's role is to safeguard the perspective of the subjects, so that it will be represented in the problem definition, methods, and interpreta­tion of results of the research. (See Serpell, 1993, p. 66, for an example of the conscious use of the insider in actual research.) However, in recent years, there have come to be even more direct methods of investigating and therefore safe guarding this perspective have been developed. [The reader is referred to the discussion of Tobin, Wu, and Davidson's (1989; Tobin, 1989) work in the section on video technology and to the section on methods for the study of cultural mean­ing later in this chapter.)
At the same time, however, a knowledgeable outsider has an important per­spective as well. An out-group member can see, and therefore study, aspects of the dominant culture that insiders have taken for granted or even repressed. The outsider can also serve as a cultural intermediary in making one culture more understandable to another.
Perhaps even more important to the methodology of cultural psychology is the role of the culturally marginal person, a particular type of outsider, in 1987 a panel of five leading cross-cultural psychologists assembled at the IACCP meet­ing in Newcastle, Australia to present their intellectual autobiographies. Every one of them had had the experience of being culturally marginal (in the sense of being between two cultures) at some early point in development. Because of their positions between cultures, they had ceased to take culture for granted; it was no longer the air they breathed. The contrast between cultures that they had experi­enced had made them personally aware of culture per se. In crossing over cul­tures in their personal lives, they had arrived at cross-cultural psychology in their professional lives.
There are a number of examples of researchers who have moved permanently into a new culture as adults, who have seen aspects of cultural psychology in their adoptive culture that had been missed by indigenous social scientists. A notable one is John Ogbu, an immigrant from Nigeria to the United States. Upon arriving he became interested in the paradox of explanations for low African-American school achievement. Factors such as poverty and level of parental school­ing were being cited as causal explanations for this psychoeducationai phenom­enon. Yet, having come from Africa where poverty and lack of parental schooling are everyday phenomena, he knew these explanations could not be right, for school achievement was not problematical at all in Africa. At the same time, hav­ing lived as a member of the dominant majority in Nigeria, he saw a social phe­nomenon that Americans had been blind to: caste. This became his explanatory construct for African-American underachievement in Minority Education and Caste (Ogbu, 1978). Ogbu used his outsider's perspective to gain an important insight into the cultural psychology of Americans. At the same time, as a permanent resident in the United States and a product of its university system, he was also very knowledgeable about his adoptive culture. Because of firsthand knowledge of two cultures, he could use the perspective of one to raise to consciousness an important aspect of the other, one that insiders had heretofore been unconscious of.
Cross-cultural psychology has accepted the desirability and possibility of objectivity, while acknowledging ethnocentrism as a barrier to it (Berry et al., 1992). Campbell (1970) has proposed a research design whereby each of two cul­tures, to be compared, is studied by an insider and an outsider. By replicating the same study twice in each culture, one is supposed to be able to separate ethno­centrism from "real" cultural differences. This view provides a good foundation for cultural psychology, but the proposed methodology does not go nearly far enough. Meanings are considered only at the ievei of data interpretation; Campbell assumes that the same procedure will be understood in the same way and will be equally meaningful to subjects in both cultures. However, this is not necessarily the case. Therefore, we still have no idea of what the differences mean. In addi­tion, there is no provision in Campbell's design for both insider and outsider to design their own study. Comparison of insider and outsider research design can go much further in assessing the effect of cross-cultural differences in the con­struction of meaning on comparative research itself. In fact, there are many pro­cedures in cross-cultural psychology (see Berry et ah; 1992, chapter 9) that seek evidence for such variations in meaning (e.g., item bias indicators), and efforts are made to control them.
Research as a Communication Process
Data collection is a process of communication between subjects and researchers. Although human communication is based on universal capacities, the default assumptions about knowledge and communication are culturally variable. The implication of this cultural variability is that data collection must be based on familiar modes of communication in each culture, rather than formally identical modes and means of communication. This is the cultural psychology approach to dealing with a problem noted in cross-cultural psychology—that particular in­strument formats may not be meaningful in all cultures (Berry et al., 1992). How­ever, cross-cultural psychology tries to adjust a single instrument to multiple cul­tures, whereas cultural psychology paradigmatically adopts the notion of using different procedures in different cultures to study comparable issues. But for the reader to understand how this might work, it is first necessary to be more spe­cific about the nature of cultural variability in presuppositions about the commu­nication process itself.
Cognitive Realism or Cognitive Relativism?: Cultural Variability in the Theory of Mind
A radical departure from the assumption that there is a separation of the know­ing subject from the known object can wreak havoc with many psychological procedures. An example follows.
Greenfield (1966; Greenfield & Bruner, 1966 [1969] ) brought tests of conser­vation of quantity to Senegal in order to study the Piagetian stage of concrete operations. After transferring water from a shorter, fatter beaker into a longer, thinner one, unschooled Wolof children were asked (in their native language of Wolof) if the quantity of water was the same, more, or less. After their response, the interview procedure used in Cambridge, Massachusetts (Bruner, Greenfield, Olver, et al., 1966) was continued, and they were asked to justify their quantity judgment: "Why do you think it is the same (or more, or lesser) amount of wa­ter?" This question format met with no response. Even when the wording was changed to "Why do you say it is the same (or more, or lesser) amount of water?", the question elicited only uncomprehending silence.
 Not until the question was changed to "Why is the water the same (or more or less)?" were justifications for the original quantity judgment successfully elic­ited. At that point, the unschooled children gave reasons for their judgments that were as articulate as those given to Piaget and his colleagues in Geneva.
These children had an epistemoiogy of mental realism. According to their implicit theory of mind, they were not making a distinction between the nature of reality and their knowledge of it. Consequently, the idea of explaining a state­ment was meaningless; only the external event could be meaningfully explained (Greenfield & Bruner, 1966 [1969]). Implicit in this theory of mind was an as­sumption that there was only a single way to perceive the event of water transfer and its results.
Had an exact translation of the Cambridge conservation procedure been used, it would have been erroneously concluded that the unschooled Wolof children were not able to explain the reasoning behind their quantity judgments. Their theory of mind would have been confounded with their reasoning about the world. The research publication would have incorrectly concluded that unschooled Wolof children had a major cognitive lack in reasoning skill. Instead, the conclusion from pilot testing was that unschooled Wolof children had a different epistemoi­ogy and therefore required a different interview procedure. When tested with an epistemologically appropriate procedure, the cognitive deficit in reasoning about the world disappeared.
In contrast to the unschooled children, the Wolof children who attended school responded well to a question form that made no sense to the unschooled children: "Why do you say the water is the same (or more, or less)?" They produced articulate reasons indistinguishable from the reasons of children in Switzerland and the United States. Apparently, the process of schooling had changed their epistemological presuppositions to accord with those of the psychological experiment. Our conclusion was that it was the introduction by the school of the written word into an oral culture that had made the differ­ence (Greenfield, 1972): In the medium of writing, thoughts about the world are visibly distinct (on the printed page) from the world itself (Greenfield & Bruner, 1966 [1969]). It seemed likely that the written word had transformed an epistemoiogy of cognitive realism into one of cognitive relativism, an epis­temoiogy in which a given person can have a variety of thoughts about the same thing, or different people can have different thoughts about the same thing.
This difference between schooled and unschooled children has important implications for the kinds of populations to whom one can validly transfer proce­dures: It implicates formal education as a potentially important variable in developing the implicit epistemoiogy required by the communication process common to many psychological procedures and instruments.
Notions of the Nature of Knowledge
An example is the difference between individualistic and collectivistic notions of the nature of knowledge. Many societies think of knowledge as a group, not an individual, process. Co-construction of knowledge, as it normally occurs in thecourse of conversation, is the norm. Interviewing Zinacantecan Maya girls and their mothers about 'learning, experience, and technique in textile production, the author envisioned each girl and each mother as an individual subject with an individual interview protocol. But that is not how Zinacantecans saw it. The no­tion that a girl would have an independent viewpoint, piece of knowledge, or perspective was not within the world view of this Maya group from Highland Chiapas. Instead, they expected more knowledgeable mothers to answer for young girls and for members of the family grouping to answer questions cooperatively. Their perspective seemed to be that the overall information would be as accurate as possible, because it was the product of a group effort. The partitioning of this information individual by individual was at odds with their world view. A simi­lar phenomenon has been described for the A-Chewa people in Zambia (Serpell, 1993, p. 230).
The Zinacantecans and A-Chewa illustrate an assumption that is common to many collectivistic societies. However, this is not just a substantive finding about cultural variability in communication processes. It is also a methodological find­ing that impacts procedures that can be meaningfully used to collect data. The procedure used to collect information in such a society must permit the coopera­tive construction of knowledge. For example, the author allowed anyone who was knowledgeable to provide information about a subject's weaving experi­ence. Usually this knowledgeable person was the girl's mother.
Although the procedure involves the co-construction of knowledge, the indi­vidual subject can still function as the unit of analysis. Thus, in the textile inter­view example, the author emerged with information about each girl and her mother as distinct individuals. What was different from the nature of research communication in the United States was that the information on each subject was co-constructed by more than one person.
Cultural Variability in the Right to Opinions
Closely related are cultural differences regarding who has the right to have opin­ions (Lonner & Berry, 1986). In the United States and other industrialized coun­tries, it is assumed that everyone has the right to have an individual opinion on any subject whatsoever. That is why opinion polls work in these countries. How­ever, in many societies, opinions are formed on the group level, not by individu­als. In such societies, the group's leaders and elders are given the right to an opinion. Others are not encouraged to express an individual perspective. In such a society, polling ordinary individuals on their opinions or attitudes is not an effective way to communicate with subjects (cf. Lerner, 1958).
Independent Questions or Connected Discourse?
Another epistemological surprise that relates to the use of surveys or question­naires is the assumption of independent questions. Surveys are often constructed so that successive questions do not follow the conventions of connected discourse. Instead, they are conceived as a series of independent items. However, this con­vention flies in the face of conversational conventions, in which each turn is a response to the preceding turn as well as a stimulus to the next. A questionnaire or structured interview, with its independent "turns," can therefore seem strange. For example, Zinacantecans seemed very malcontent when a later question seemed to ignore a previous answer. The genre of the structured interview or question­naire was foreign to their culture. As a result, so was the convention of indepen­dent questions specific to this genre.
Culture Specificity of Ignorant or Out-of-Context Questions The absence of such a convention particularly interferes with a common system of internal validity checks in psychological instruments; such a system is based on redundancy, that is, asking for the same information in two different ways. Far from enhancing validity, redundancy could destroy validity in communities unused to the independent "turns" of the psychological instrument genre. For example, Zinacantecan subjects could barely tolerate redundant questions. The attitude transmitted to the author was "Why are you so stupid as to ask the same question twice?"
Closely related was the Zinacantecans lack of tolerance for ignorant ques­tions. It is likely that the Zinacantecos are similar to many other groups in this. The point is that a rigid questionnaire cannot be used to interview on a topic of great unfamiliarity to the researcher. As a prior step, ethnographic field work or focus groups must be used to figure out what the intelligent questions are. (Such prior steps as these will be discussed later in the chapter.)
Another assumption of the interview or questionnaire is that questions can be successfully asked out of context. Not all groups share this assumption. L. Devereaux (personal communication, 1992) developed a methodology based on her perception that Zinacantecans would not provide meaningful responses to questions asked out of the context of an ongoing activity. She developed a tech­nique of interviewing around ongoing activities. For example, she would inter­view about weaving when weaving was taking place, about child development when children were present, and so forth. Greenfield adapted this technique to a structured interview. She was having a hard time eliciting information about the activities of play weaving and play embroidery until she started using a play weaving and a play embroidery as stimuli to ask if subjects had ever made such items when they were little. The recognition and comprehension were instanta­neous. Meaningfulness was evident in the enthusiasm and spontaneity with which subjects responded. This spontaneity and enthusiasm contrasted strongly with the ennui and lack of comprehension that greeted questions about absent objects and events.
Communicating Across Cultures or Age Levels
This same point, that variability in the epistemology of communication between subject and researcher can undermine research communication and invalidate results, has been made with respect to young children (Schubauer-Leoni, Perret-Clermont, & Grossen, 1993; Siegal 1991a, 1991b, in press) Siegal points out that, in a data collection procedure, the violation of communication conventions e.g., failing to minimize redundancy—can cause children to perform at lower levels because of communication failure between researcher and chiid.
Thus, cross-cultural and cross-age communication hold exactly the same methodological dangers of eliciting invalid information because of systematic but unrealized communication failure. The conclusion is that methods of eliciting data from subjects must be adapted to the presuppositions about communication that are held by each cultural group. This is yet another reason why the most valid cross-cultural comparisons will often be based on radically different proce­dures used in each culture.
The final point relates to the methodological significance of a researcher's intersubjectivity. Intersubjectivity, the sharing of a perspective, is the foundation of all communication (Trevarthen, 1980). Applying this idea to the notion of re­search as a communication process with subjects, we must conclude that the va­lidity of data depends on the researcher's achievement of intersubjectivity with his or her subjects. This is because cooperative conversation depends on the shar­ing of conversational goals (Grice, 1975). If such goals are not shared, the subject may be answering a different question than the researcher had in mind. It fol­lows, then, that the more the researcher can share the perspective of a subject, regardless of the culture of that subject, the more valid the data that will be gath­ered and the more valid the interpretation of the data. This conclusion is very different from the received wisdom in psychology. Tne most valid perspective is an objective or detached perspective.
The methodology of cultural psychology impels the field to go beyond tradi­tional psychological concepts of validity. A new type of validity - Maxwell's (1992) notion of interpretive validity—is most relevant to the metamethodological dis­cussion of perspective and communication just discussed. Interpretive validity involves a concern with what "objects, events, and behaviors mean to the people engaged in and with them" (Wells, Hirshberg, Lipton, & Oakes, 1995, p. 288). If we expand this concept to include what questions and other conversational moves mean to the people engaged in them, then interpretive validity will include (1.) understanding the basic communicational and epistemological presuppositions of our subjects, and (2.) making sure that all data collection procedures conform to these presuppositions. Much work remains to be done to establish methods for ensuring interpretive validity.
A second type of validity that is particularly relevant to the methodology of cultural psychology is ecological validity. This, a more well-known type of validity in psychology, involves the extent to which a procedure elicits data that is repre­sentative of behavior outside the research context. Many of the methods for study­ing adaptive behaviors in everyday contexts that we will describe have ecologi­cal validity built into them. In studying naturally-occurring rather than labora­tory behavior, ecological validity is insured. In deriving research problems and methods from the cultural context, rather than from the science of psychology, ecological validity is automatically enhanced. In basing experiments on ethno­graphic observations, a topic to be discussed later, ecological validity is maxi­mized. Thus, ecological validity is implicit in the very foundations of cultural psychology.
When a researcher directly studies a phenomenon of interest, rather than measuring it indirectly through items in an instrument, traditional types of valid­ity checks, such as assessment of content validity, become superfluous. This is because these traditional types of validity concerns are based on using some measurement as an index of the phenomenon of interest. However, when a be­havioral phenomenon is studied directly, rather than indexed indirectly, the ne­cessity to check representativeness through a content validity check becomes unnecessary; it is ensured ipso facto. This point applies to the important cultural psychology method (elaborated later in this chapter) of studying behavior in its natural context.
Another relevant type of validity is theoretical validity (Maxwell, 1992). This involves "the presence of a more abstract explanation of described actions and interpreted meanings" (Miles & Huberman, 1994, p. 279). Theoretical validity is particularly important because it gives culture-specific research its generalizability (Miles & Huberman, 1994).
The Necessity of Methods for Studying Cultural History
A key aspect of human culture is its cumulative quality: culture is both transmit­ted and transformed between and within generations. If cultural psychology is to be grounded in the nature of culture, then a key aspect of empirical methodology will be the development of methods for studying the role of cultural history in the current psychological functioning of individual members of a cultural group. As part of this effort, it will be necessary to have methods that can relate the interactive processes of intergenerational transmission to the cumulative nature of cultural knowledge. At the same time, the other side of cultural accumulation, cultural transformation and change, must also be a focus of empirical methodol­ogy. It is particularly important that the methods of cultural psychology be able to study the ways in which historical roots and cultural change combine to affect the enculturation of individuals in a cultural group at a given point in time (Greenfield & Cocking, 1994a).
Developmental Methods
A theoretically important process in cultural psychology is the expert-novice re­lationship in which someone with greater cultural knowledge in a meaningful domain of activity interacts with someone with lesser knowledge in that domain, enhancing shared knowledge in a way that moves the novice toward expertise (Rogoff, 1990). The repetition of this process across generations leads to the cu­mulative nature of cultural knowledge and the importance of cultural history in cultural psychology (Scribner, 1985). The capacity of the novice to respond to different kinds of input provided by the expert is influenced partly by chrono­logically based maturity. In addition, the movement from cultural novice to cul­tural expert does not occur all at once; it is a step-by-step process. Hence, the importance of developmental methodology to cultural psychology (cf. Eckensberger, 1979).
As Valsiner (1989) notes, an important goal of cultural psychology is to un­derstand how the process of development takes place within a culture. (See also chapter by Valsiner and Lawrence, in Volume 2 of this Handbook). This "entails development of empirical methodologies that document the process of interac­tion between the child and his/her environment... a research paradigm that is primarily directed towards explaining how culture organizes the conditions for children's development, and how children assimilate these conditions, and si­multaneously accommodate to them" (Valsiner, 1989, pp. 4-5). Methods for the study of developmental processes are a methodologica] keystone for the study of culture as psychological process.
The Relevance of Cross-Species and Neuroscience Methodologies
Bruner (1972) discussed the evolutionary foundations of culture and culture learn­ing. More recently Cole (1992 ) has spoken of the necessity to integrate universal biological factors into a general theory of cultural psychology. The biological prim­ing of culture includes what have been termed cognitive constraints (Carey & Gelrnan, 1991; Hirschfeld & Gelman, 1994) or, more accurately, learning biases (Gallistel, Brown, Carey, Gelman, & Keil, 1991). For example, humans are "bi­ased" to learn language (e.g., Pinker, 1994), an important component of cultural processes. The methodological implications of the universal human propensity to culture include investigations of the evolutionary foundation of culture through cross-species comparisons in domains relevant to the learning, use, and trans­mission of culture such as tools (e.g., Goodall, 1986; Greenfield, 1991; Matsuzawa, 1991; McGrew, 1992), symbolic communication (e.g., Plooij, 1978; Greenfield & Savage-Rumbaugh, 1991), cultural variability (Nishida, 1987) and observational learning (e.g., Tomasello, Davis-Dasilva, Camak, & Bard, 1987). (See also chapter by Keller, this volume.)
Tomasello, Kruger, and Ratner's (1993) formulation of three levels of cul­ture learning—emulation, imitation, and collaboration—relevant to the evolu­tion of culture stems from and suggest cross-species comparison of culture learn­ing mechanisms. This methodology can provide insights (and has already done so) into the evolutionary (and therefore biological) foundations for culture learn­ing. An outstanding example is Boesch's (1991) naturalistic study establishing that chimpanzee mothers in the Tai forest of the Ivory Coast use intentional teach­ing techniques to transmit the hammer/ anvil technique of nut cracking to their young. This technique is cultural not only in the sense of involving social trans­mission, but also in the sense of being a technique that is distinctive to chimpan­zees in a particular geographical region.
The human propensity for culture also implies investigations of the biologi­cal, especially the neural foundations of various cultural skills such as ianguage and tools (e.g., Greenfield, 1991; Deacon, in press). Perusse (1993) has done a behavior genetics study establishing the heritability of parental teaching style, an important process in cultural transmission. Segal (1993) has used twin methodol­ogy to explore the importance of genetic relatedness in stimulating important cultural behavior such as helpfulness and cooperation. Her studies indicate that methods addressing not just the biology of individuals but the genetic relation­ships between individuals will be important to understanding cultural behavior.
Methodological Role of Anthropology
It is clear from the inclusion of anthropological approaches in this and other hand­books (Goodenough, 1980; Munroe & Munroe, 1986; see also chapter by Munroe & Munroe, this volume) that anthropology has a role to play in cross-cultural psychology. However, this role is particularly important in cultural psychology where culture is considered a process rather than an independent variable.
First, and probably most important, is the anthropological notion of ethnog­raphy as a methodological concept. Weisner proposes a central role for ethnogra­phy because "it brings the importance of lived experience in a cultural place to the center of attention, transforming it from ground to figure" (Weisner, in press, ms. p. 3). This method is crucially important as the first stage of any cultural psychological research in a new, unfamiliar setting. Ethnography, according to Goodenough (1980), "describes what people must have learned in order to par­ticipate acceptably in most of the activities of that society" (p. 29). This is a broad anthropological notion that assumes that it is possible to master most activities of a society.
For the purposes of studying the psychological processes of culture, a more limited ethnographic goal seems more appropriate. As Packer (1995) puts it, ethnography involves firsthand experience of the settings in which the human activity of research interest occurs. The classical method of experiencing settings is by participant observation. Goodenough (1980) and others have described how to keep systematic records of participant observation. In the course of becoming a participant-observer, the researcher establishes an identity in the setting (Rizzo, Corsaro, & Bates, 1992). An important aspect of ethnography is to be able to com­municate with the people of the study community in their own language.
Sometimes, as Packer (1995) notes, firsthand experience can be approximated by open-ended conversations and interviews. Unstructured focus groups can provide another approximation to firsthand experience.
Even when researching in one's own culture, any new setting can be an un­known cultural niche. Packer (1995) relates his experience with the ethnographic phase of a study of a kindergarten class: "I spent two years hanging out with the children in a pre-school kindergarten class, and that experience was invaluable; it helped me interpret and analyze the video-recordings I made. First it gave me a sense of the tone and climate of the school, and the style and manner of the children.... In addition to the affective level, practical engagement with the arti­facts of a context is quite different from the detached, objective observation of these artifacts.... To understand the common-sense that the participants in a context employ—and, as Clifford Geertz has insisted, this common-sense is a cultural system—we need to encounter it first hand" (pp. 3-4).
This statement stresses ethnography as a way to uncover and discover the subjects' own perspective. But, as Weisner, a psychological anthropologist, points out, "ethnography is not limited to understanding meaning and the construction of experience by culture members. It is also central in understanding social insti­tutions and social structure, demographic trends, economic exchanges, power and influence, and other, presumably more formal, distal or etic influences on development in a cultural place" (Weisner, in press, ms. p. 8).
Weisner (in press) also stresses the complementarity of ethnography with other methods. Combinations of ethnography with other methods often permit the re­searcher to integrate data concerning different levels of the sociocultural system. This is a methodological concept that Rogoff has explicated in speaking of the complementary "lenses" through which the researcher can see the developing child on a variety of planes: as individual, as member of a dyad, as part of a community setting (Rogoff, Baker-Sennett, Lacasa, & Goldsmith, 1995).
Wells, Hirshberg, Lipton, & Oakes (1995) point out that sociocultural planes are not themselves static units that can be defined in advance, instead, the pianes are co-constructions between researchers and subjects. Researchers must learn through interacting with subjects and observing their activities which relation­ships and community institutions are relevant to the focal subjects. This approach is quite different from the notion of random sampling that cross-cultural psy­chology (Lonner & Berry, 1986; Berry et al., 1992) has adopted from traditional psychology.
Another type of complementarily between anthropology and psychology is that of ethnography and experimentation. A wonderful example of a study in which full-blown ethnographic study was the basis for experimental research is Beach's study of the role of external memory cues in learning to become a bar­tender. In the ethnographic phase of his research, Beach (1984, 1992) enrolled in and went through a bartending school's two-week course. Based on his ethnog­raphy, Beach then designed experimental studies to verify and extend the ethno­graphic findings. His experimental procedure was based directly on an existing school practice, the speed drill for mixing drinks. However, in the experimental phase, he couid vary stimuli systematically (e.g., glass shape) in order to pin­point the cognitive issue of interest, microdevelopmental change in the use of external memory cues with increasing expertise.
Another excellent example of this methodology lies in the work of the Bra­zilian team of Nunes, Schliemann, Carraher, and colleagues (e.g., 1993; see also chapter by Schliemann, Carraher & Ceci, Volume 2, this Handbook). For example, in one study, (Schliemann, 1984), naturalistic observation of carpenters at work and of a carpentry school yielded realistic mathematical problems that carpen­ters must solve in their everyday work. These problems were then the basis for using experiments to compare professional carpenters and apprentices at different levels in order to find out the nature of the development of cognitive strate­gies for solving the mathematical problems required in this particular ecological context. In other studies, experimental problems are based on interviews with experts in a particular activity context rather than observation (e.g, Grando, 1988). (See also chapter by Schiiemann, Carraher & Ceci in Volume 2 of this Handbook.)
In cross-cultural psychology, prior ethnographies are recognized as a valu­able foundation for research (Berry et al., 1992). However, whereas cross-cultural psychologists generally rely on anthropologists or iheii local colleagues to learn the language and the culture ethnographically, cultural psychologists more often make learning the language and participating in the culture an important part of their own research, This is because the procedures and methods of cultural psy­chology arise from the culture itself, not from the methodological cupboard of psychology. As a consequence, the ethnographic stage of research is indispens­able. The prior examples make clear the potentially close, and even isomorphic, relationship between ethnography and psychological experimentation.
The integration of anthropological and psychological methods can be facilitated by interdisciplinary teams consisting of a psychologist and an anthropologist. Re­cently Jurg Wassman (anthropologist) and Pierre Dasen (psychologist) have devel­oped a methodology for integrating the two disciplinary perspectives. They "advo­cate the following general research strategy in three steps: (1.) interviews with a few key informants and "jpfs" [just plain folks]; (2.) behaviour observations in everyday settings to get at the application of knowledge in daily life; and (3.) setting tasks, to induce behaviour that is not observable in everyday situations" (Wassmann & Dasen, 1994, p. 23). The first two steps draw primarily on anthropological methodology, whereas the last step draws on psychological methodology.
The research strategy of Wassman and Dasen was developed partly as a way of addressing the sampling problem that stemmed from the anthropological reli­ance on "key informants." Reliance on "key informants" was based on the con­ception of culture as a homogenous entity, ignoring its internal structural diver­sity and the individual differences so important to the field of psychology. Inter­viewing "just plain folks" expands the researchers' sample beyond "key infor­mants" and allows comparison of various viewpoints within a culture. Others in psychology have also argued for sampling as an important way of modifying classical ethnography:
Newman and Saxe both develop the argument that careful sampling can be cru­cial for ethnographic work. Nervman samples particular groups for generational identity differences, and Saxe compares Brazilian street sellers who are adepts and others ivho are novices in folk math tasks (Weisner, in press, ms. pp. 12-13).
In essence, sampling is a way of dealing with the problem of perspective in ethnography. Miles and Huberman (1984) point out that qualitative field research typically involves purposive rather than random sampling, both within and between subjects: "These may be, for example, samples of actors, settings, events, time periods, and processes" (Miles & Huberman, 1984, p. 25). This is something that has been strongly debated in recent years:
Critical and feminist theories alike question the historical bases of gender, power, or control from which ethnographies and ethnographers come (di Leonardo, 1991; Marcus & Fischer, 1986; Weisner, in press, ms. p. 5).
Yet it must be noted that methods for reducing and evaluating bias, as well as methods for enhancing and evaluating validity and reliability in ethnographic data do exist. Validity was discussed earlier, so the examples presented here will concern bias and reliability. Procedures suggested by Miles and Huberman (1994) to avoid the biasing influence of researcher effects include staying as long on-site as possible; spending some time simply hanging around, fitting into the land­scape, taking a lower profile; and using unobtrusive measures where possible (Miles & Huberman, 1994, p. 266). In order to evaluate bias, the authors suggest a number of relevant queries, including:
Can we follow the actual sequence of how data were collected, processed,
condensed /transformed, and displayed for specific conclusion drawing?
Has the researcher been explicit and self-aware as possible about personal
assumptions, values and biases, and affective states—and how they may have
come into play during the study?
Were competing hypotheses or rival conclusions really considered? If so, at
what point in the study? Do other rival hypotheses seem possible (Miles &
Huberman, 1994, p. 278)?
For ethnographic research, reliability involves processes of "quality control" (Miles & Huberman, 1994). Some suggested queries to evaluate reliability include:
Is the researcher's role and status within the site explicitly described?
Were data quality checks made (e.g., for bias, deceit, informant knowl-
edgeability?) (Miles & Huberman, 1994, p. 278).
Another way in which anthropological methods can be integrated into cul­tural psychology is through training in anthropology for psychologists and through training in psychology for anthropologists. Many researchers in cultural psychol­ogy have had this type of cross-disciplinary training. Such training enhances in­terdisciplinary collaboration, as well as the integration of cross-disciplinary meth­odology by a single researcher.
Qualitative Methods are Primary in the Understanding of Process
The conversational analyst, Schegloff (1993) wrote:
In examining large amounts of data, we are studying multiples or aggregates of single instances. Quantitative analysis is, in this sense, not an alternative to single case analysis, but rather is built on its back (p. 102).
Schegloff s point is that it is necessary to understand the phenomenon under study before being able to aggregate multiples of them in a quantitative test of their frequency or typicality. Indeed, it is necessary to discover the phenomenon under study before being able to aggregate multiple instances; in such a case the N of 1 functions as an important existence proof (T. Au, personal communica­tion). Qualitative methods (the study of single instances, that is, data are not aggregated) are particularly crucial when it is important to stay in touch with the structural unity of a process (Fisher, 1994). Here is an example from attachment research of the danger of aggregating data without understanding each individual case, with its structurally unified process.
A baby in the United States participated in Ainsworth's Strange Situation Procedure; the baby was classified as "resistant," not "securely" attached to his mother, because he played alone quietly and contentedly (rather than resisting) when left alone with a stranger. As his mother watched this behavior through a one-way mirror, she "proudly commented to the researchers, "Look how inde­pendent he is! See how he can play by himself? This is what 1 have been working for by having him be with other kids and families while I am working' " (Weisner, in press, ms. p. 11). Weisner (in press) continues:
This mother ivas a single parent by choice. She had told us about her cultural goals for independence for herself and her child, her commitment to feminism, her struggles to sustain tvork and parenting, and many other values. Her con­struction of her child's behavior came from this framework of beliefs and prac­tices. The knowledge gleaned through such informal conversations -with the mother about her ideas about her child, done along with ongoing ethnographic observa­tions ofzuhat she is doing in her even/day world to operationalize those ideas, is surely a powerful tool in understanding trust and attachment in cultural con­text" (ms. p. 11).
Does it make sense to aggregate this baby with other infants showing the same behavior for completely different cultural reasons? This is the sort of question raised by Schegloff's discussion.
Schegloff goes on to say that "We need to know what the phenomena are, how they are organized, and how they are related to each other as a precondition for cogently bringing methods of quantitative analysis to bear on them" (p. 114) (cf., Hatano, 1995). In other words, qualitative analysis must precede and inform quantitative analysis. This principle is illustrated by the following example in­volving eye contact as a communication and socialization process for infants:
Based on an analysis of mother-infant eye contact, LeVine and colleagues (LeVine, Dixon, LeVine, Richman, Leiderman, Keefer, & Brazeiton, 1994) con­cluded that infants receive less maternal eye contact in Africa than in the United States. However, unlike the norm for European-derived cultures, siblings are important caregivers in African families (LeVine et al., 1994; Weisner & Gallimore, 1977; Zukow, 1989). Based on this qualitative analysis of the social organization of child care in a particular ecological context, Sigman et al. (1994) measured the Schegloff s point is that it is necessary to understand the phenomenon under study before being able to aggregate multiples of them in a quantitative test of their frequency or typicality. Indeed, it is necessary to discover the phenomenon under study before being able to aggregate multiple instances; in such a case the N of 1 functions as an important existence proof (T. Au, personal communica­tion). Qualitative methods (the study of single instances, that is, data are not aggregated) are particularly crucial when it is important to stay in touch with the structural unity of a process (Fisher, 1994). Here is an example from attachment research of the danger of aggregating data without understanding each individual case, with its structurally unified process.
A baby in the United States participated in Ainsworth's Strange Situation Procedure; the baby was classified as "resistant," not "securely" attached to his mother, because he played alone quietly and contentedly (rather than resisting) when left alone with a stranger. As his mother watched this behavior through a one-way mirror, she "proudly commented to the researchers, "Look how inde­pendent he is! See how he can play by himself? This is what 1 have been working for by having him be with other kids and families while I am working' " (Weisner, in press, ms. p. 11). Weisner (in press) continues:
This mother ivas a single parent by choice. She had told us about her cultural goals for independence for herself and her child, her commitment to feminism, her struggles to sustain tvork and parenting, and many other values. Her con­struction of her child's behavior came from this framework of beliefs and prac­tices. The knowledge gleaned through such informal conversations -with the mother about her ideas about her child, done along with ongoing ethnographic observa­tions ofzuhat she is doing in her even/day world to operationalize those ideas, is surely a powerful tool in understanding trust and attachment in cultural con­text" (ms. p. 11).
Does it make sense to aggregate this baby with other infants showing the same behavior for completely different cultural reasons? This is the sort of question raised by Schegloff's discussion.
Schegloff goes on to say that "We need to know what the phenomena are, how they are organized, and how they are related to each other as a precondition for cogently bringing methods of quantitative analysis to bear on them" (p. 114) (cf., Hatano, 1995). In other words, qualitative analysis must precede and inform quantitative analysis. This principle is illustrated by the following example in­volving eye contact as a communication and socialization process for infants:
Based on an analysis of mother-infant eye contact, LeVine and colleagues (LeVine, Dixon, LeVine, Richman, Leiderman, Keefer, & Brazeiton, 1994) con­cluded that infants receive less maternal eye contact in Africa than in the United States. However, unlike the norm for European-derived cultures, siblings are important caregivers in African families (LeVine et al., 1994; Weisner & Gallimore, 1977; Zukow, 1989). Based on this qualitative analysis of the social organization of child care in a particular ecological context, Sigman et al. (1994) measured the aggregate amount of eye contact Emba babies in Kenya received from all of their caregivers, including siblings. The results were very different: the amount of eye contact, rather than being less than Euro-American norms, was, if anything, greater. The earlier conclusion that East African babies were "deprived" of eye contact was changed to the conclusion that they were "enriched" in this regard. A differ­ent qualitative analysis of what to count, the nature of the phenomenon, led to a different quantitative analysis and, eventually, to different conclusions. The point is that (1.) quantitative results are strongly affected by qualitative analysis of the phenomenon under study; and (2.) the qualitative nature of a given phenomenon (e.g., infant caregiving) varies from culture to culture.
Cultural Adaptation, Practice, and Naturalistic Methodology
If behavioral adaptations to ecological conditions are central psychological pro­cesses in cultural psychology, this point has methodological implications: the study of natural behavior in situ (in contrast to controlled experimentation) becomes critical to answering questions concerning conventionalized (i.e., cultural) be­havioral adaptations under varying ecological conditions. Skills must be stud­ied in practice (Lave, 1988) before they are studied in the laboratory. In situ obser­vation is important because it highlights dynamic processes in contrast to static products of cultural adaptation. The freer the behavior, the less the adaptive pro­cesses are likely to have been distorted by the research procedure. Insofar as cross-cultural psychology has been wedded to the index (item, stimulus) rather than to the process, it can deal only indirectly with behavioral adaptation to eco­logical contexts. However, cross-cultural psychology has recognized the impor­tance of ecological adaptation, and the need to employ ethnographic and obser­vational methods to study it prior to the development of psychological instru­ments (Berry, 1980; Berry et al, 1992).
Everyday activities reflect cultural adaptation to ecological conditions. An excellent example of the study of everyday activity in context is Rogoff's (in press) analysis of developmental transitions in children's participation in sociocultural activities. That analysis is based on in situ observations of preschool children in their interaction with one-year olds in two communities, a Mayan town in the Highlands of Guatemala and a city in the United States. But behavioral observa­tions in situ are not sufficient, as Rogoff argues. In order to understand the mean-ing of the observed in situ behavioral differences in the two cultures, it was neces­sary to also investigate "the social organization of family roles and cultural ex­pectations of childhood in each community" (ms. p. 18, Rogoff, in press).
Methods for the Study of Cultural Meaning
Although interpretive processes of meaning construction and practical activities of materialistic adaptation are always intertwined in any real-world situation of cultural processes (as in Rogoff's example above), cultural psychology highlights the interindividual construction of meaning. It follows that this is also the dis­tinctive methodological contribution of cultural psychology. Therefore, this chapter now turns to a discussion of concepts and methods for the psychological study of cultural meaning. As Feldman, Bruner, Kalmar, and Renderer (1993) point out, it has not been easy to see what form empirical research would take within a cultural cognitive psychology focused on meaning-making processes. None­theless, considerable steps have been taken, and will be outlined in the fol­lowing sections.
Intersubjectivity and Shared Meaning
The sharing of meaning transforms individual meaning into cultural meaning. Methods that explore the conventionalization of meaning between two partners are relevant to the growth of shared meaning as the basis of culture (Bruner, 1990). The foundation for shared meaning is the same intersubjecriviry that makes communication possible. Trevarthen (1980) sees intersubjectivity as the founda­tion of human culture. Appreciation of other individual minds creates inter­subjectivity and therefore culture. According to him, human beings have from infancy an intrinsic motivation to gain knowledge from others, thereby showing themselves to be intrinsically social and cultural. In this view, social sharing, the basis of culture, is as natural as any other human activity. Hence, methods for exploring the development of shared understandings and information exchange in the communication process (e.g., Ochs & Schieffelin, 1983) are basic methods in cultural psychology.
Cross-cultural psychology has recognized the problem of meaning in design­ing comparative studies. It is conceptualized as a problem of comparability (Berry, et al., 1992) or equivalence (Poortinga, 1989; Poortinga & Malpass, 1986).3 How­ever, for cross-cultural psychology, variability in cultural meanings is looked upon as a barrier to cross-cultural equivalence and comparability. Diversity in cultural meaning is seen as something to eliminate in the interest of finding universals; it is not viewed as something to be studied in its own right. Cultural psychology goes deeper in identifying the nature of the problem and develops methods for the direct study of meaning. In cultural psychology, cultural meanings are not a barrier to research; they are the central topic of investigation. A discussion of some methods for studying the construction of cultural meanings follows.
Meaning can be communicated and shared through reciprocal social interac­tion. This process requires a theory of other minds. Hence, the cross-cultural study of theory of mind is central to the study of shared meaning in cultural psychology.
The developmentally earliest paradigm for social reciprocity and sharing is what Marcel Mauss (1954) called le don, the gift. When an infant gives a gift, he/ she is applying social reciprocity to the world of objects. Studies of the develop­ment of forms of reciprocity in different cultures (e.g., Rabain, 1979) are therefore relevant to investigating the ontogeny of cultural meaning.
Many studies document the transmission of and development of cultural values through everyday processes of interaction and communication (e.g., Blake, 1994; Choi, 1992; Greenfield, Brazelton, & Childs, 1989; Greenfield, Raeff, & Quiroz, in press; Heath, 1983; Ochs, 1982; Rabain-Jamin, 1994; Schieffelin, 1983; Schneider, Hieshima, Lee, & Piank, 1994; Shweder & Much, 1987). Greenfield, Raeff, and Quiroz (1995) use discourse analysis to document cooperative and conflicting cultural constructions of the child; their study illustrates the use of discourse analysis to study dynamic cultural processes in everyday processes of communi­cation and interaction.
All of these studies use microanalytic methods for analyzing interactive pro­cesses. These range from discourse to semantics, grammar, and nonverbal com­munication. Methodological sources for discourse analysis are found in Sinclair and Coulthard (1975) and Edwards and Lampert (1993). The latter focuses on transcription and coding.
The study of social ethnotheories is another important method for the study of shared cultural meanings and their acquisition (see also chapter by Super and Harkness in volume 2 of this Handbook). A social ethnotheory expresses human qualities that are valued by a particular group. Ethnotheories of intelligence (Berry & Bennett, 1992; Dasen, 1984; Serpell, 1993; Wober, 1974) and of parenting (e.g., Harkness & Super, 1995; Zukow, 1984) have been of particular interest. Let us look more closely at the latter. A parental ethnotheory is a culture-specific con­ception of the goals of child development. Probably the most seminal paper for cultural psychology was the "three developmental stories" paper by Ochs and Schieffelin (1984). This paper was important for two reasons. First, it depicted the theoretical stance of developmental psycholinguistics as a culture-specific ethnotheory, rather than a universal truth. Second, it linked everyday interaction to the cultural value systems beneath it.
Other studies have used interview methodology to study parental ethnotheories (e.g., Goodnow, 1985). Another technique is the use of scenarios in which different child development goals lead to different resolutions of the sce­nario (e.g., Greenfield, Raeff, & Quiroz, in press). By analyzing open-ended re­sponses and letting the categories emerge from the data themselves, this tech­nique enables researchers to study the social construction of the child both intraculturally and cross-culturally. An important extension of methods for study­ing ethnotheories of child development and cultural socialization is the "multivocal ethnography" pioneered by Tobin, Wu, and Davidson (1989). It is discussed in detail as a use for video technology in a later section. A social psychological ver­sion of multivocal ethnography has been pioneered by Kitayama, Markus, Matsumoto and Norasakkunkit (in press).
Note that methods for studying the construction of meaning all have as their goal to reveal the nature of the research subject's perspective or subjectivity. Jones and Thome (1987) point out that the perspective of the subject can also be a meth­odological tool to evaluate the meanings of assessment procedures, particularly useful in exploring intercultural clinical assessment.
Narrative Methods
Bruner (1986,1990) proposes narrative as ah intrinsically cultural mode of think­ing (Lucariello, 1995). Narrative thought involves structuring the world in terms of characters with intentions who perform actions in settings, using particular means. Narrative therefore highlights the human interpretive understanding of other human beings and their activity as central to culture. At the same time, narrative as a cultural category emphasizes the making of meaning rather than the performing of behavior. Narrative is a dynamic process because it involves interpreting sequences of ongoing events. As a cultural mode, it is also dynamic and process-oriented in that it permits the study of human generative creativity: there are infinite combinations of characters, intentions, actions, settings, and means that may produced in a narrative construction.
Labov and Waletsky (1967) provide a basic source on the methodology of narrative analysis. Narratives from the crib (Nelson, 1989) presents further examples of narrative analysis; at the same time it illustrates how cultural psychology cen­ters on the developmental acquisition of cultural modes of activity, in this case, the production of narrative. Another example of narrative analysis as a technique is provided by Ochs and Taylor (1992) their analysis of dinner table narratives.
Combining the Study of Materialistic and Symbolic Culture: Constraints and Preferences
The fact that cultural processes are both shared activities adapted to ecological conditions and shared meanings for these activities implicates methods that as­sess both constraints (based on material conditions) and preferences (based on values or cultural meanings) (Shweder, Jensen, & Goldstein, 1995). Shweder and colleagues' study of sleeping arrangements in India and the United States was innovative in distinguishing constraints based on material ecological conditions (e.g., size of family, gender composition of family, number of beds available) from preferences that were based on culturally meaningful values (e.g., caring for the young, incest avoidance, chastity anxiety). The authors distinguished constraints from preferences by asking subjects (in India) about sleeping arrangements of a hypothetical seven-person family under varying resource conditions. Such a method allows one to specify both the shared cultural meanings that generate preferences (e.g., care of the young in India, autonomy in the United States) and the shared cultural responses to limited resource conditions (the number of beds). The methodology of this study points out the limitations of simple behavioral observation: Observing who sleeps with whom would not permit the analytical distinction between constraints and preferences to be made (Shweder, Jensen, & Goldstein, 1995). Simple behavioral observation, no matter how thorough, would confound cultural preferences (symbolic or idealist meaning) with cultural re­sponses to ecological constraint (materialist adaptation).
Methods for the Study of Cultural History
All of the empirical methods discussed so far assume that the study of the psy­chology of culture takes place at one point in time. However, an important tenet of cultural psychology is that the present psychology of culture reflects residues of past cultural history (Scribner, 1985). What methods are appropriate for exam­ining this past residue? One method suggested by Vygotsky was to study the influence of cultural tools and artifacts, because each artifact is itself the product of cultural history (Cole, 1995). Cole writes, "human beings live in an environ­ment transformed by the artifacts of prior generations, extending back to the beginning of the species (Geertz, 1973; Ilyenkov, 1977; Sahlins, 1976; Wartofsky, 1979). The basic function of these artifacts is to coordinate human beings with the physical world and each other" (Cole, 1992, p. 9). Thus, when Scribner and Cole (1981) studied the cognitive effects of three different literacies used by the Vai in Liberia, they were in essence studying the effects of a complex cultural history that had produced the three writing systems and the practices with which each is associated. Saxe (1982a, 1982b) has studied number systems and their use as the product of cultural history.
Another important methodological approach to the study of cultural history is the comparison of psychological processes of the same ethnic group in differ­ent societal contexts. Sometimes this is done directly by a single researcher (e.g., Ho, 1989). Other times, it is done indirectly by comparing the findings of differ­ent researchers (e.g., Greenfield & Cocking, 1994a). The commonalities in psy­chological phenomena of the same ethnic group in different societal contexts point to the effect of ancestral cultural influences that predate the divergence of the ethnic group into different societal contexts. On the other hand, differences be­tween ethnic group members residing in different cultural contexts reflect a dif­ferent sort of history, the history of intergroup contact under varying societal conditions. The development of methodology for studying the impact of differ­ent patterns of intergroup contact on the same ethnic group has been pioneered by Ogbu (e.g., 1978, 1994).
Such differences reflect the dynamic, changing quality of cultural history. There have been a number of attempts to capture the psychological nature of cultural change indirectly by synchronic or cross-sectional methods. Perhaps the first was Vygotsky' and Luria's attempt to assess the cognitive impact of the post-revolu­tionary process of farm collectivization in the Soviet Union by comparing cogni­tive processes in peasants living on traditional farms with those who had been collectivized on a number of cognitive tasks (Luria, 1976). Recent examples of this research strategy include Saxe's (1982a, b) studies in New Guinea comparing the mathematical cognition of Oksamin people who have been more touched by the recent introduction of commerce and money with those who have been less touched. Draper and Cashden (1988) have explored the impact of sedentarization on sex role socialization among the Efe by comparing Efe still living as hunter/ gatherers with those living as agriculturalists. (See also chapter by Berry & Sam, in Volume 3 of this Handbook, on the psychological consequences of culture con­tact and change.)
Perhaps the most direct of all of the synchronic approaches to diachronic change is the study of families who have immigrated from the sam.e country in different generations. An example of this strategy is Delgado-Gaitan's (1993,1994) comparison of immigrant and first-generation Mexican-American families' intrafamilial interaction patterns and values. What differs at one point in time between the immigrant and first-generation families is used to modei the cul­tural assimilation process over time in a single family. In all of these research designs, the logic of the method is to take advantage of current variability in the diffusion of a cultural change to reconstruct the impact of that cultural change over historical time.
This indirect methodological strategy has the potential problem of making assumptions about unidirectional cultural evolution (cf. Eckensberger, Krewer, & Kasper, 1984). Consequently, the simulation of cultural change by comparative research must be based on specific knowledge of actual cultural change, not gen­eral theories of cultural evolution. In addition, there is the methodological prob­lem of assuring group comparability in factors other than those connected with social change.
A more direct way of assessing the impact of cultural change is through diachronic methods, termed longitudinal in psychology. Longitudinal studies of historical change are a recent addition to cultural psychology. Three examples of this new method can be cited:
Through the comparison of two generations of children from the same fami­
lies, Greenfield (1993) has studied the impact of economic development on
interactional processes of informal education and individual processes of cog­
nitive representation in a Mayan community in Chiapas, Mexico.4
Through longitudinal study in both Puerto Rico and the United States of Puerto
Rican families before and after immigrating to the New York area, Laosa (in
progress) has been able to trace the impact of cultural change on psychologi­
cal and family interaction processes.
Through the longitudinal observation of dairy workers adapting to comput­
erization of their jobs, Scribner, Sachs, Di Bello, and Kindred (1991), studied
the cognitive adaptation to a major change in the technological side of culture.
Because longitudinal or diachronic methods are intrinsically historical, they are more direct than cross-sectional or synchronic ones. Therefore, they should take on increasing importance in the study of the psychological impact of cul­tural history. However, the two types of method are mutually complementary. Sometimes, for example, massive historical change may make it impossible to delineate exactly what factors caused what psychological changes. Synchronic variability in a range of factors can then be used to model the historical change process with more precision. This strategy is currently being used by Greenfield, Childs, and Maynard to delineate precisely what aspects of the Zinacantecos' historical economic movement from agriculture to commerce are the proximal and distal causes of the associated changes in learning, teaching, and cognition.
The analysis of cultural change, whether the change be historical or evolu­tionary, requires particular attention to an aspect of data that is often ignored in psychology: its variability. The nature of evolutionary change is that, in response to environmental change, natural selection takes adaptive characteristics that are less frequent and makes them more frequent over long periods of time. In cul­tural change as well, minority trends at one time may expand into dominant trends under new conditions. Therefore, infrequent behavioral phenomena may often furnish more clues as to historical or evolutionary cultural change than do normative trends. The methodological implication holds that it is necessary to recognize the theoretical significance of infrequent as well as frequent behavioral phenomena in understanding the psychological dimensions of cultural change. Centra! tendency may often lead one astray in the study of social change.
The Unique Methodological Role of Video Technology
Video is uniquely suited for the study of processes of in situ cultural adaptation and for the study of the construction of cultural meaning. Examples of research that have drawn upon video for the former purpose are Childs and Greenfield's (1980) study of weaving apprenticeship and Stigler's study of U.S., Chinese, and Japanese classrooms (Stigler & Perry, 1990). Video also solves certain problems of reliability and validity, posed but not solved by classical psychology. It provides a permanent record by which other researchers can check the basis for interpreta­tions and conclusions. Whereas the classic criterion of replicability of findings has been called into question (Valsiner, 1989) because of the possibility of change over time (either developmental or cultural), video freezes data in time, thus allowing the analysis to be replicated without repeating the observation.
Whereas video furnishes the data for the study of in situ processes of cultural adaptation, video can also furnish stimuli for studying the construction of cul­tural meaning (Jacobs et al., 1996; Tobin, Wu, & Davidson, 1989; Tobin, 1989). In this latter method, members of different cultures are given the task of evaluating everyday practices on video from the same range of cultures. In the study of Tobin et al. and of Jacobs, they focus on the evaluation of classroom practices in China, Japan, and the United States. The design is cross-cultural at two levels, the level of the stimuli and the level of the subjects. For example, in Jacobs' (in press) study, Japanese and American teachers are given an opportunity to evaluate vid­eos of Japanese and American classroom lessons. The notion of multiple perspec­tives is built right into the design of the study itself.
An important point is the difference between this empirical approach to mul­tiple perspectives and the post-structuralist approach which simply decries the researchers' ethnocentric perspective without trying to do anything about it (cf. Patai, 1994). Tobin (1989) describes the thought behind the Tobin et al. (1989) study in this way: "We have sought to develop a method for doing research and a narrative stance for our writing that would decenter as well as deprivilege the author—anthropologist. Rather than replacing the persona of the omniscient, posi-tivistic, confident, gentleman-scholar with the persona of the apologetic, soul-searching, self-centered, reflexive anthropologist, we strive to shift narrative at­tention and the authority to define meaning away from the author. We strive to give voice—the power to name, interpret, and analyze—to the teachers, students, parents, and children who have traditionally been objects rather than partners in investigation" (Tobin, 1989, p. 174).

Tobin et al. accomplished exactly that in their innovative study. They first videotaped a preschool in each of three cultures, China, Japan, and the United States. They then showed the tapes to preschool staff, parents, and child develop­ment experts in each country, asking them to evaluate their own and each other's schools. This method led to "a multicultural discussion of such issues as free­dom, conformity, creativity, and discipline" (Tobin et al., 1989, dust jacket). This method leads the authors beyond cultural differences in educational practices to the values and child development goals that lie beneath the practices.
Some of the most interesting discussion took place when the video sequences from one country violated norms of another. Nothing is as effective in revealing cultural norms and values as the reactions people have when the norms are bro­ken. In many cases, the norms of one culture, shown on video, violated the norms of the "outsider" observing the video. The cross-cultural showing of videos pro­vides a systematic opportunity to comment on norm violation, thus revealing the very existence of normative values.
Video technology involves a host of other methodological fine points, includ­ing transcription, coding, and interobserver reliability. The focus here will be on a new method developed by J. Stigler. In the course of a cross-cultural study of classroom practices, he and his colleagues have developed software that allows coding to be done on or next to the video frames, for instant retrieval and statis­tical access. After a tape has been digitized onto CD-ROM, the software allows instant access to any of the codes or video frames. This system has the potential to speed up video coding tremendously. For one reason, it will no longer be nec­essary to transcribe simply to keep a record of what was going on at the time a code was made; the original video clip is stored with the code. At the same time, the advances in video and computer technology enable frames to be captured on computer and, ultimately, paper, to illustrate results and to constitute a visual part of discourse transcription (e.g., Goodwin, in press).
Quantitative Methods for the Study of Behavior as Part of a Cultural System
Once qualitative understanding has been achieved, quantitative methods involv­ing issues of frequency are useful and appropriate (Gaskins, 1994). However, not all statistical techniques are equally suitable for answering cultural questions. Insofar as cultural psychologists are interested in the varying levels and layers of culture from the most macro to the most micro, particular kinds of statistical techniques are often more appropriate. One major problem with traditional tech­niques such as analysis of variance and regression, is the division into indepen­dent and dependent variables. Although cross-cuiturai psychology generally interprets culture as an independent variable or even a set of independent vari­ables (Jahoda, 1990), this approach has major problems.
First, culture is a sy3tern with interrelated, not independent parts (Berry, 1983). An excellent statistical technique to capture these interrelations is structural equa­tion modeling (e.g., Bentler, 1989). This technique permits a model with multiple interacting variables. Any variable can have links with any or all of the other variables. Analyses of covariance structures are another example of a statistical technique that has these characteristics. These methods are discussed by Van de Vijver and Leung, this volume.
Second, variables cannot be neatly divided into cause and effect. Most vari­ables function both as causes of some things and effects of others. For example, Greenfield and Childs (Greenfield, 1993) have proposed a causal model in which it is hypothesized that historical factors influence economic activity , which in turn, influences how weaving is transmitted mtergenerationaily. In this model, economic activity is treated as the effect of historical era and the cause of style of informal education. Given data on each of these levels, such a model can then be tested by structural equation modeling.
Third, in cultural systems, variables often function as both cause and effect. An example is the influence of symbolic cultural tools such as video games on individual development, and the effect of individual differences on the use of these tools. Such reciprocal two-way relationships can also be tested by struc­tural equation modeling (e.g., Greenfield, Brannon, & Lohr, 1994). In general, structural equation modeling is consistent with the concept of culture as a system that is not apart from the individual. It allows the testing of models involving a whole web of social, cultural, and personal factors over time in a single model. However, the very same flexibility that is an asset of these models can also be a disadvantage. If a researcher is using them in an exploratory fashion, as a rule, some structure will emerge. However, the post hoc interpretation of such an out­come is open to question. In such a case, the validity of a given model depends on testing and ruiing out theoretically plausible alternatives.
Cultural Adaptation of Procedures and Measures
A cogent example of the conceptual and theoretical problems that arise when measuring instruments are transported across cultures without adaptation comes from the study of attachment (LeVine & Miller, 1990; Takahashi, 1990), Attach­ment is a topic particularly suitable for cultural and cross-cultural investigation because adult-infant attachment is a key foundation for cultural transmission and for the social relations upon which human culture is based. (See also chapter by Keller, this volume.)
LeVine and Miller (1990) recount how a procedure, the Strange Situation, that had been devised by Ainsworth as a culture-specific adaptation to measure attachment in the United States came to be used universally in the interest of cross-cultural comparison. In moving her study of attachment from Uganda (Ainsworth, 1967) to the United States, Ainsworth had changed her attachment measuring instrument to accord with the greater autonomy and independence encouraged in U. S. infants. Nonetheless, the deceptive standardization of the Strange Situation subsequently sent it all over the world, with no thought as to its cross-cultural validity and dubious comparative results (see Takahashi's critique, 1990). LeVine and Miller (1990) point out that in 1978 Ainsworth and colleagues wrote, "It seems entirely likely that Ainsworth's (1967) Ganda infants and Konner's (1972) Bushmen babies could not have tolerated the strange situation" (Ainsworth, Blehar, Waters, & Wall, 1978, p. iv). Recently, Takahashi (1986, 1990) announced that the Japanese mothers of her sample would not consent to leaving their ba­bies alone in an unfamiliar situation.
Clearly, the same situation had totally different meanings for babies and moth­ers in these different cultures. Mindful of this critique, Harwood (1992) did a cross-cultural study of the interpretive meaning of the Strange Situation and the behaviors that occur in it for Puerto Rican and Euro-American mothers. Indeed, she found that the meaning of responses to this situation was different in the two cultural groups: In keeping with their more individualistic orientation, Euro-American mothers "describe an active yet related infant as most desirable and a clingy, distressed infant as most undesirable" (Harwood, 1992, p. 831). Puerto Rican mothers, in keeping with their more sociocentric or collectivistic orienta­tion, "describe as most desirable a quiet, responsive infant whose behavior is tipped more toward proximity maintenance than toward active exploration" (Harwood, 1992, p. 831). If the culturally desired response to the Strange Situa­tion is culturally variable, then it is not possible to standardize behavioral catego­ries and the interpretations of these categories across different cultures. In line with the earlier discussion of single case analysis as applied to the Strange Situa­tion, cross-cultural quantitative comparisons also become problematic.
Going one step further, note that the Strange Situation measures attachment through response to separation. Given that the ability to cope with separation is normative in only a minority of cultures (often termed individualistic), one might also want to measure attachment as proximity maintenance behavior, in line with the Puerto Rican (and other coiiectivistic societies') ideal of attachment behavior. In other words, this analysis of attachment reveals how culture-specific a mea­suring instrument can be. To use the Strange Situation in both an individualistic and collectivistic culture is to have a biased comparison. The infants in one soci­ety will be tested by the same measure that their upbringing has utilized; the infants in the other society will be tested on how well they can do in a culturally-foreign situation. If only one single testing situation is to be used in cross-cultural comparison, it should be one that provides room for a range of ideals to be actu­alized. A procedure for measuring attachment that allowed both proximity main­tenance and tolerance for separation to be manifest as normative attachment would provide this range. Such a situation would be culturally fairer than the classical Strange Situation that forces mother-child separation on all babies, whether or not they have ever experienced it in their everyday lives (Takahashi, 1990).
Substantive Consequences of Empirical Methodology for Cultural Psychology
Cultural Adaptation of Procedures Leads to Generalizable Processes of Cognition and Culture Acquisition
Generalizations concerning universal processes must be a part of cultural psy­chology, insofar as the culture-making capacities of the human species are being considered. However, cross-cultural generalizations must arise from a more ab­stract level than the level of similar or identical measuring tools. An example illustrating this point is taken from a study of the development of kinship termi­nology carried out by Greenfield and Childs (1978) in Zinacantan, a Mayan com­munity in Chiapas, Mexico. Children of different ages were asked about various sibling relations in their own households, using the complex sibling terminology of their Tzotzil language in which there are separate terms for older and younger siblings, as well as for the sibling of a boy and the sibling of a girl. Samples of the type of question asked are, "What is the name of your older sister?" and "as for older sister, Shunka, what is the name of her younger brother?" From anthropol­ogy, linguistics, and developmental psychology, Greenfield and Childs extracted theories that might have explained the results. The theories from anthropology and linguistics were relativistic, emphasizing the role of culture-specific values (the importance of the older-younger relationship) and language-specific struc­ture (the complexity and organization of the sib terminology system). The theo­ries from psychology, Piagetian and information processing, were universalistic, emphasizing common responses as a function of chronological age (Piaget) and memory load (information processing).

Piaget (1928) had done a study in the same domain with Swiss children. Greenfield and Childs' procedure, however, was quite different from his. Whereas he asked quantitative questions ("How many brothers do you have?"), Greenfield and Childs asked qualitative ones ("What is the name of your older brother?). (To have been required to answer quantitative questions would have demanded an unfamiliar skill for Zinacanteco children.) Whereas Piaget asked simply about the two terms, brother and sister, Greenfield and Childs asked about the six different Tzotzil sibling terms. (Note that comparability would have dictated aggregating sets of sibling terms, such as older sister, girl's younger sister, and boy's younger sister. Yet, to do this would have violated the organization of Tzotzil sibling organization.)
A maxim of the use of "comparable" instruments for cross-cultural comparison is that comparable instruments are necessary in order to elucidate universal processes. A corollary would be that the lack of comparable instruments will lead only to culture-specific conclusions. This maxim, with its corollary, was belied by Greenfield and Childs' results. Whereas their culture-specific procedure should, according to this way of thinking, have led to the verification of culture specific processes (triggered, in this case, by cultural values or linguistic terminology), the results showed no evidence of any such culture-specific processes. Instead, all of the results led to the validation of putatively universal developmental and cognitive processes. For example, Piaget had posited the development of a stage of concrete operations in middle childhood, a stage where children can symbolically represent the transformations of concrete objects. In Greenfield and Chiids' data, there was evidence of concrete operational development, from egocentrism to understanding of reciprocity to understanding of reversible relations (e.g., I am my older brother's younger brother), occurring in the same age range described by Piaget for Swiss children. Greenfield and Childs also found evidence of memory development and the influence of category size on memory retrieval, theorized to be universal processes.

The methodological point of this example is that maximum adaptation of methods to each particular culture in which a phenomenon is studied maximizes the possibility of uncovering nomothetic, even universal, processes at the same time as it obviates the possibility of direct, quantitative cross cultural comparison (cf., Enriquez, 1977). Ironically, the comparable methods required for direct comparison in this example would probably have reduced, if not eliminated, the manifestation of universal processes. For example, unfamiliar quantitative questions would not have been answered so well by Zinacantecos, and cognitive competencies in kinship terminology would have been masked. Similarly, the use of unfamiliar aggregates of siblings to conform to the French sibling terms used by Piaget would have been confusing and would have hidden the Zinacanteco child's mastery of a more complex sibling system. The paradox is that universal processes are revealed more by noncomparable methods and theoretical comparison than by comparable methods and quantitative comparison across cultures.

However, this study was not explicitly comparative. Is it possible to use noncomparable methods in an explicitly comparative study? Morelli, Rogoff, Oppenheim, and Goldsmith (1992) did exactly that in a comparative study of sleeping arrangements in Utah and in a Maya community in Guatemala. In each place, "the interview was tailored in ways appropriate to each community" (p. 606). The tailoring did not preclude basic quantitative comparisons, although the emphasis was on qualitative results. Of great pragmatic interest to those advocating noncomparable methods as a way of drawing conclusions both about cultural variability and possible universals is the fact that this article was published by a mainstream journal.
Cultural Generation of Procedures leads to Generalizabie Processes of Cognition and Culture Acquisition

This important point will be illustrated with the example of the cognitive appropriation of cultural tools. Processes of transforming cultural tools into mental tools (Bruner, 1964; Vygotsky, 1978.; Saxe, 1991) are part of universal cognitive equipment and operations. The exploration of these processes demands the identification of particular cultural tools, an understanding of how they function in cultural practices, and experimental methods by which the transformation and use of the tools can be studied. The important point is that, although a particular tool is rarely universal, its learning and use will draw upon and reveal universal processes in the cognitive appropriation of cultural tools (Saxe, 1991).

An outstanding example is the exploration of the "mental abacus" begun by Hatano, Miyake, and Binks (1977) and continued by Stigler and colleagues (Stigler, 1984; Stigler, Chalip, & Miller, 1986). The research begins with an analysis of the abacus as a calculation tool. Clearly, an understanding of the cultural tool in question is an important part of the methodology. This understanding must be gained either by being a cultural insider who has been exposed to and trained in the use of the tool, by ethnographic exploration, or by collaboration with a cultural insider.
A number of different methodological strategies have been used to explore the transformation of the physical abacus into a mental tool and the operation of the "mental abacus." One strategy (Stigler, 1984) is based on identifying steps in the arithmetic problem solving process that were specific to abacus use (i.e., would not be used in standard numerically-based calculations). When shown photographs of an abacus with its beads in various positions, experts were extremely accurate at distinguishing abacus states that were intermediate steps to the solution of particular problems from states that were not part of the problem's solution. In addition, response time was slower if the depicted abacus state occurred later in the problem-solving sequence.
A second methodological strategy was to do an error analysis to see if the nature of abacus user errors reflected a mental representation specific to abacus use (Stigler, 1984). Because the abacus has upper beads that represent the quantity five, errors that deviate by exactly five would reveal the use of an internal abacus representation in mental calculation. Abacus operators made significantly more mistakes that were off by exactly five, both in using the abacus and in mental calculation, than did American students.
Consequently, it could be concluded that (1.) abacus experts had internalized a representation of an abacus that they could operate upon mentally, (2.) they carried out operations on the mental abacus in the same sequence as would be done on the physical abacus, and (3.) their errors bore the trace of the mental abacas. The point is that the conclusions from this research are not specific to the abacus, a cultural tool particular to certain cultures. The point is not whether subjects could generalize their skills to other tools. In essence, the conclusions are about potentially universal processes of transforming cultural tools into mental tools (Gauvain, 1995).
By studying the effects on representation and problem-solving of activities using other culture-specific tools, further conclusions about the scope and nature of these cognitive processes of cultural appropriation (Saxe, 1991) can be drawn. For example, in the domain of math, the influence of money denominations on the representation of math problems was demonstrated by Saxe (1982a, 1982b) in New Guinea. Note that these findings about the representational impact of money generalize the conclusions of Stigler and colleagues about the representational impact of the abacus, even though Saxe used different procedures to study the effects of a different tool in a different society. Indeed, the cogrutive consequences of using a range of cultural tools in a variety of activities have been studied (Greenfield & Lave, 1982; Guberman & Greenfield, 1991): tailoring (Lave, 1977), pottery making (Price Williams, Gordon, & Ramirez, 1969), video games (Greenfield, 1993; Greenfield & Cocking, 1994b), candy selling (Saxe, 1991), money (Saxe, 1982a, 1982b; Guberman, in press a, b), weaving (Greenfield & Childs, 1977), and bartending (Beach, 1984, 1992), In each case the structure of the artifact is internalized as a partially isomorphic representation.
The method of using cultural tools embedded in cultural activities as a methodological strategy not only reveals information about how cognitive enculturation varies from culture to culture, depending on the tools available in a particular cultural niche. The method also reveals information about the universal role of tools in developing processes of mental representation. Only by generating particular studies out of the specific tool using activities that occur in a given cultural niche can the general question of the relation between cultural tools and cognitive processes be addressed. The ability to generalize cannot be based on using formally equivalent procedures across cultures. Instead, generalization is based on using common conceptual questions about situated cognition to generate a multiplicity of procedures that are appropriate to diverse cultural ruches.

The methodology for studying culture as process must elucidate the ontogenetic, sociohistorical, and phylogeneiic origins of psychological functioning in the enculturated adult human being (Vygotsky & Luria, 1993). But methodology must go farther. It must be adequate to deal with socially shared cognition (Resnick, Levine, & Teasley, 1991) and with the participation of the individual in social relations and in a cultural community (Rogoff, in press). The development of research methodology adequate to these requirements is an extremely demand­ing task. The suggestions in this chapter should be of practical utility to research­ers and students who wish to study culture as psychological process.
1.          The author wishes to thank Jennifer Jacobs,      scores in different cultural groups" fp. 225]). Even
Shinobu Kitayama, Ashiey Maynard, Joan Miller,       when insider judgments of the cultural appropri
Richard Shweder, and James Stigler for helpful              ateness of individual stimuli are used, the possi
comments on an earlier draft of this chapter.
               bility that the very method (e.g., opinion survey)
2.          Along with Jahoda, 1977, this chapter rejects          on which a particular instrument is based may
the etic/emic distinction as a useful way to talk              have different meanings in different cultures is
about methods for studying universal and cul-                                                                         not dealt with.
hire-specific patterns of thinking.                               4. An unanticipated methodological insight
3.   Often the problem of variable cultural inter-            from this historical study was that more open-
pretations is dealt with as a statistical issue inter-       ended, naturalistic methods are more sensitive to
nal to a particular instrument (e.g., Berry et al.,             sociohistorical change than are more controlled
1992; Van de Vijver & Poortinga, 1991). Statistical     experimental procedures. The latter provide too
problems then find statistical solutions. One ex   
much constraint on subject responses to be sensi
ample of such a solution is the measurement of              tive to future cultural change, unknown at the
bias ("all unwanted factors that unequally affect                time of the historical baseline observations.
Ainsworth, M. D. S. (1967) Infancy in Uganda: In-    Ainsworth, M. D. S., Blehar, M. C, Waters, E., &
fant care and the growth of attachment. Balti-
           Wall, S. (1978).   Patterns of attachment.
more: John Hopkins University Press.                          Hillsdale NJ: Erlbaum.

2 komentar:

  1. ane masih baca buku daniel goleman yang emotional intelligence, sama yang social intelligence, terima kasih banget share ilmunya..

  2. Jasa Pencarian Literatur / Pustaka / Jurnal untuk Penelitian (Skripsi, Tesis, Disertasi, Perusahaan)

    cepat, akurat, mudah dan murah