journal of cross-cultural psychology

http://jcc.sagepub.com

Journal of Cross-Cultural Psychology

DOI: 10.1177/0022022107305237 2007; 38; 539 Journal of Cross-Cultural Psychology

Yumi Gosso, Maria de Lima Salum e Morais and Emma Otta Pretend Play of Brazilian Children: A Window Into Different Cultural Worlds

http://jcc.sagepub.com/cgi/content/abstract/38/5/539 The online version of this article can be found at:

Published by:

http://www.sagepublications.com

On behalf of:

International Association for Cross-Cultural Psychology

can be found at:Journal of Cross-Cultural Psychology Additional services and information for

http://jcc.sagepub.com/cgi/alerts Email Alerts:

http://jcc.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://jcc.sagepub.com/cgi/content/refs/38/5/539 Citations

at UNIV DE SAO PAULO BIBLIOTECA on January 14, 2009 http://jcc.sagepub.comDownloaded from

http://www.iaccp.org/

http://jcc.sagepub.com/cgi/alerts

http://jcc.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://jcc.sagepub.com/cgi/content/refs/38/5/539


539

PRETEND PLAY OF BRAZILIAN CHILDRENA Window Into Different Cultural Worlds

YUMI GOSSOMARIA DE LIMA SALUM E MORAIS

Health InstituteEMMA OTTA

Institute of Psychology, University of São Paulo

This study compared pretend play of Brazilian children (41 girls and 35 boys, 4 to 6 years old) of five dif-ferent cultural groups: (a) Indians, (b) seashore, (c) low socioeconomic status (SES) urban, (d) high SESurban, and (e) mixed SES urban. This cross-cultural investigation was based in naturalistic observationand conducted within a methodological framework that considered the content and the meaning of sym-bolic transformation and its structure. Although pretend play occurred in all groups, high and mixed SESurban children engaged in more pretending than the others. In accordance with the theoretical claim thatmake believe is a universal phenomenon, the differences among cultural groups were expressed more inthe content than in the structure of pretend play.

Keywords: make believe; play; cultural differences; universals

In recent years, developmental psychology has gained new impulse with cross-culturalstudies, which have contributed to increasing our understanding both of cultural peculiaritiesand of universals of human behavior, exposing the psychological unity of humankindbeneath superficial differences (Berry, Poortinga, Segall, & Dasen, 1996; Bornstein, Haynes,Pascual, Painter, & Galperin, 1999; Eibl-Eibesfeldt, 1972; Farver, Kim, & Lee-Shin, 2000;Gaskins, 2000; Göncü, Mistry, & Mosier, 2000; Haight, Wang, Fung, Williams, & Mintz,1999; Moura, 2003). In Brazil, a team of researchers has been studying children’s play in dif-ferent cultural groups living at various regions from the north to the south (Beraldo &Carvalho, 2003; Bichara, 1994, 2002; Carvalho & Pedrosa, 2003; Gosso & Otta, 2003a,2003b; Gosso, Otta, Morais, Ribeiro, & Bussab, 2005; Magalhães, Souza, & Carvalho, 2003;Morais & Otta, 2003a, 2003b; E. C. Santos & Koller, 2003; J. E. F. Santos, 2003). Our inter-est has been focused on the comparative study of symbolic play of Parakanã Indians (fromParanowaona at the north of Brazil, in the Amazon region), seashore children (from Ubatuba,a small town of the southeast of Brazil), and urban groups of different socioeconomic status(SES) from São Paulo, a metropolis of the southeast of Brazil.

The Indians were from Paranowaona, one of five villages of the Parakanã IndianReserve located in Novo Repartimento and Itupiranga municipalities, at Pará State.

AUTHORS’ NOTE: This research was supported by a Conselho Nacional de Desenvolvimento Científico e Tecnológico grant.We would also like to thank Luis Fernando de Oliveira Saraiva, Izabel Almeida, Daniele Carli Licciardi, Márcio Berber DizAmadeu, Caroline Toshie Hara, and Fernanda Cazelli Buckeridge for their technical help and Gisele Zago for her assistance inanalyzing the data. We would especially like to thank Peter K. Smith for providing helpful comments on the article. Thanks arealso due to anonymous reviewers’ comments that greatly improved our analyses and text. The flip of a coin determined the orderof the first and second authors.

JOURNAL OF CROSS-CULTURAL PSYCHOLOGY, Vol. 38 No. 5, September 2007 539-558DOI: 10.1177/0022022107305237© 2007 Sage Publications



According to the census of the Parakanã Program (2000), 502 Indians distributed in five vil-lages live in the reserve area of 21,858 square miles. During data collection, 86 Indians livedin Paranowaona. They were first contacted in the 1970s and have been assisted by theParakanã Program since 1987, a contract agreed between the National Agency for Assistanceto Indigenous People (FUNAI) and the Northern Electric Company (ELETRONORTE). TheParakanã Program offers health assistance, education, support for farming, and protection ofreservation boundaries. The Indians do not have television or telephone. From 3 years old on,children play in groups without adult supervision.

Seashore children lived at Itamambuca, one of 92 beaches of Ubatuba, at São PauloState. According to the last Brazilian census (IBGE, 2000), this municipality had 66,861inhabitants and an area of 711 km2, of which 78% was composed of virgin forest. In thepast, Ubatuba was a harbor of mineral drainage and sugar cane mills. Nowadays, most ofits economy is based on tourism. Itamambuca has a low population density and a lot of nat-ural resources, such as native forest, rivers, and sea. The children of our sample lived inpoor areas near the school and its surroundings. Most of the parents were housekeepers,gardeners, artisans, or workers in civil construction. The public school attended by thechildren had good installations, a football field, and two playgrounds.

Urban groups were from São Paulo, a city located in São Paulo State in the southeastof Brazil, with an area of 1,525 km2 and 10,434,252 inhabitants (IBGE, 2000). Threegroups of different SES compositions were studied. Low SES urban children lived in ashantytown near the public school where the observations were made. Most of theirparents were excluded from the formal work market, living from occasional work on build-ing sites, from gewgaws and food sales, as washerwomen or housekeepers, and so on.Their monthly income was about 1-2 times the minimum Brazilian income (aboutUS$90.00 to US$180.00). High SES urban children lived in comfortable, large houseswith access to all contemporary electronic technology. Their school was located inside aset of these houses, and the whole complex was surrounded by walls and had a privatesecurity system—a solution found by Brazilian high SES families to protect themselvesagainst increasing urban violence. The school monthly fees were about 5 times the mini-mum Brazilian income. Their parents were autonomous professionals and executives, withmore than 30 times the minimum Brazilian income per month. Finally, the mixed SESurban sample was composed of children who attended a preschool at a public university.Their parents worked or studied at this university as professors, undergraduate or graduatestudents, and white-collar or blue-collar workers in various sectors of the university. Theirmonthly income varied from 3 to 20 times the minimum Brazilian income. We considerthis group different from both high and low SES because it was composed of children fromdifferent socioeconomical and cultural backgrounds. Some of them were children ofparents of high cultural level (e.g., professors and graduate students), others from mediumcultural level (e.g., administrative professionals, laboratory technicians), and others frompoor cultural and academic level (e.g., security guards, manual laborers). As a function ofthe companionship of peers from different cultural backgrounds, there were mutual influ-ences on their play behavior that cannot be neglected and that make this group unique.

THEORETICAL FRAMEWORK: THREE DIMENSIONS OF CULTURES

To understand the influence of culture on symbolic play we used a theoretical frame-work that distinguishes cultural groups on three dimensions: (a) collectivism versus indi-vidualism, (b) proximal versus distal styles of upbringing, and (c) simple versus elaboratedcommunication codes. Although they are highly interrelated, these dimensions are helpful

540 JOURNAL OF CROSS-CULTURAL PSYCHOLOGY



tools in the analyses of cultural differences. Given the range of complexity involved in thevalues and behaviors existent in all societies, it must be taken into account that they haveto be considered along continua. We cannot classify each group as exclusively located inone of the poles of the dimensions mentioned, but there are trends toward predominanceof one of the characteristics (e.g., the seashore culture is more collectivistic, more proxi-mal, and uses simpler communication codes than high SES and is less collectivistic, prox-imal, and uses more elaborated communication codes than Parakanã Indians).

Collectivism versus individualism. In individualistic cultures, people value independence,privacy, competition, achievement, selfishness, open expression of negative emotions, closerelationships with only a few others, and self-promotion. On the other hand, in collectivisticcultures, people value affiliation more than achievement, a characteristic frequently associ-ated with a paced rhythm of life, closer relationships, restrained expression of negative emo-tions, emphasis on harmony among people, on we instead of I (Andersen, 1999; Conway,Ryder, Tweed, & Sokol, 2001; Kitayama, Markus, Matsumoto, & Norasakkunkit, 1997;Price & Crapo, 1999; Takahashi, Ohara, Antonucci, & Akiyama, 2002).

Proximal versus distal styles of upbringing. Underlying the collectivism and individu-alism dimension are different styles of upbringing. In a nuclear family structure, childrenspend most of the day with their caregivers, whereas in an extended family structure care-givers and children are inserted in a supportive network of relatives and friends. Accordingto Keller’s model (Greenfield, Keller, Fuligni, & Maynard, 2003; Keller, 1998; Keller,Borke, Yovsi, Lohaus, & Jensen, 2005), Western cultures tend to promote early auton-omy and a dyadic–distal structure, aiming at environmental control, emphasizing the selfas agent, whereas non-Western ones tend to promote warmer and closer multiple inter-actions, resulting in a concept of self as coagent. For example, Keller et al. (2004),studying five cultural communities in West Africa, Gujarat in India, Costa Rica, Greece,and Germany that differed in the model of parenting, found support for two styles of par-enting (distal and proximal) related to the sociocultural orientations of independence andinterdependence.

Simple versus elaborated communication codes. There are communities with suchharsh socioeconomic conditions that the needs of survival lead to the development of intel-lectual skills related to physical abilities, concrete thinking, and the search for immediatesolutions for practical problems that their members—be they children or adults—face intheir daily activities. Lexicons with relatively few terms (simple communication code)tend to occur in association with simple cultures and technologies, whereas lexicons withmany terms (elaborated communication codes) tend to be correlated with complex tech-nologies and cultures (Bernstein, 1970). According to Lock (2000), nowadays in Westerncultures middle and high SES people master a metalinguistic symbolic system of abstractconcepts that characterize the elaborated codes.

Nicolaci-da-Costa (1988) called the simple code shared identity model and the elabo-rated code differentiated identity model. In stratified society, the first tends to be typical ofworking classes and the second of middle and high classes. According to her point of view,the shared identity model emphasizes the similarity among the members of a particular socialgroup, reducing the need for verbal elaboration of experiences, intentions, or individual moti-vations. Thus, some areas of the self would have a lower probability of becoming theperson’s focus of attention. The social identity promoted by the simple code tends to bestrengthened by a type of familiar control in which the socially defined position occupied

Gosso et al. / PRETEND PLAY OF BRAZILIAN CHILDREN 541



by a particular member of the family is the major feature. A set of unquestionable rights andduties characterizes this peculiar position in the family. The use of several types of punish-ments on someone that infringes rules helps to guarantee the maintenance of this position.Greenfield et al. (2003) also stated that parents from lower socioeconomic backgroundsemphasize more obedience and conformity than those from higher SES. On the other hand,the differentiated identity model turns attention to internal processes and individual differ-ences and emphasizes planning. It enlarges the distance among the members of a social group,with language as the fundamental vehicle to communicate experiences, intentions, and indi-vidual motivations. Whole areas of the self may become visible both to the own person and toothers. The individual characteristics of family members are taken into account, and there isgreater flexibility in the application of rules and definition of roles. Authority is exercisedthrough verbal mastering of feelings and the child’s responsibility assumption for their acts.

Within this framework, Indian and seashore children tend to be more collectivistic, hav-ing a proximal upbringing style and a simple communication code. They have a slow paceof life, belong to extended families in which several relatives are included in children’sdaily life, and traditional values in which definite roles are expected of family members.In addition, these children are less restrained and less supervised by adults, being allowedgreater access to their physical and social world, adopting adult roles earlier.

By contrast, high SES urban children tend to be more individualistic, having a distalupbringing style and a more elaborated communication code. Upbringing of low SESurban children could be considered intermediate between high SES urban and both Indiansand seashore children. Their families have peculiar structures in which the mother is gen-erally the head of the family and works outside the home during long periods. The childrenstay part of the day with neighbors or in day care centers. When the father is present, hisrole is clearly defined as the head of the family; usually children must obey their parentsand stay in the role attributed to them by adults. Their communication occurs according toa simple code, although the elaborated code is highly valued. On one side, harsh survivalconditions lead the poor to have a more affiliative social life; on the other, loneliness andcompetition of big cities press in the opposite direction, leading them to develop somecharacteristics of dominant individualistic cultures.

Gender differences on play behavior are generally consistent across studies of differ-ent times and cultures (Else-Quest, Hyde, Goldsmith, & Van-Hulle, 2006; Fein, 1981;Morais & Carvalho, 1994; Pellegrini & Smith, 1998; Smith, 2005). Usually male playoccurs in wide spaces, with larger bodily movements, associated with more fantasticthemes, actions that demand greater physical strength, and violence in comparison withfemale play. Female play occurs in more restricted spaces, reveals greater awareness ofothers, and enacts daily life themes, closer to reality. We predict analogous results in thepresent study comparing boys and girls in the various cultural groups. However, it isexpected that the gap between male and female play will be wider in Indians, seashore,and low SES urban children than in high and mixed SES urban children, based onCarvalho, Beraldo, Santos, and Ortega (1993). These authors found lower sexual stereo-typy in a modern urban high SES group (São Paulo) in comparison with low SES groupsand traditional urban high SES (Recife).

METHODOLOGICAL FRAMEWORK FOR ANALYZING MAKE BELIEVE PLAY

A cross-cultural investigation of make believe play demands a methodological frameworkthat, besides taking into account the content and the meaning of symbolic transformations,




considers its structure. We think that, as discussed in the following, our approach to pre-tend play structure is a major contribution of the present study.

Symbolic play attains its peak in frequency and complexity at around 4 to 6 years (Fein,1981; Piaget, 1945/1975), when children also acquire the understanding that others alsohave their own desires, intentions, and mental representations (Harris, 1989; Lillard,2001). This consciousness transposed to play allows the development of collective sym-bolism that represents the more elaborated phase of pretense. In sociodramatic play, actorsand authors coact in a fictitious field shared by all—a hypothetical, ideal, or virtual levelbecause by consensus it is independent from the world of real objects.

Make believe is a voluntary, intentional transformation of meaning attributed to objects,persons, and the here and now. We adopt Lillard’s (2001) conception of six components ofmaking believe: a pretender, a reality that is pretended about, a mental representation of analternative situation, the projection of mental representation onto the reality, awareness ofboth the present and the imagined situation, and finally, an intentional projection of mentalrepresentation onto reality.

We have restrictions to the frequently used methodological framework to analyze makebelieve play proposed by Matthews (1977), who distinguished six categories: (a) substitu-tion, object is given new identity (e.g., wooden block is put on the head and called a hat);(b) attribution of function, object is given a functional property that it does not possess(e.g., a toy camera is used to take a picture); (c) animation, an inanimate object is treatedas if it were animate (e.g., talking to a stuffed bear); (d) insubstantial material attribution,reference to materials that do not exist in the present situation (e.g., picking up a nonexis-tent food from a plate and putting it to mouth); (e) insubstantial attribution, reference to asituation that is not actually occurring, word play, rhymes, jokes (e.g., a child announceshe or she will have a birthday party); (f) character attribution, portrayal of qualities of acharacter (e.g., a child announces “I will be the nurse”).

In our point of view, Matthews’s (1977) classification system has two disadvantages:On one hand there is an overlapping of categories, and on the other hand some categorieshave low discriminative power. For instance, animation can be considered as an attributionof function, and the category insubstantial material attribution comprehends elements ofdifferent complexities such as word play and creation of scenes. For these reasons, Morais(1980) and Morais and Otta (2003b) proposed an alternative classification system. Theyconsidered that the make believe act involves an implicit or explicit attribution of meaningto an object, situation, or living being that is not its real or conventionally accepted mean-ing. Two types of ideational schemes are distinguished: creation of an element without realexistence and modification of meaning of a pivot (identity, property, or condition). In ouropinion, these two schemes have to be distinguished because they involve different degreesof complexity. When a child creates a situation or an object, he or she is trusting more inhis or her imagination than in the real properties exhibited by the situation or object (e.g.,the child pretends to have wings). He or she is relatively less restrained by the here andnow. In its turn, the symbolic modification is an operation through which the child modi-fies the meaning of an existent object at a particular moment.

Partly based on Garvey and Kramer (1989), Jensvold and Fouts (1993), and Matthews(1977), we distinguished three types of change of meaning based on differences incomplexity: creation, identity, and property modification. Creation is considered the moreelaborated process of symbolization because it happens without any material support pres-ent in the environment. Identity transformation is regarded as an operation of intermediatecomplexity because it involves a change of meaning that is not immediately given by the




pivot (e.g., a stick is treated as a gun). Property modification is considered the less com-plex operation because it is usually applied on a miniaturized toy; that is, the meaning issuggested by the pivot (e.g., a toy gun is treated as a gun). In addition, children’s actionswere categorized to define the content of pretending.

GOALS AND PREDICTIONS

This study addressed three major issues: (a) situating make believe play in relation toother kinds of play, (b) introducing a new classification system of make believe play char-acterizing it according to both structure and content, and (c) applying this classificationsystem to compare five groups of children as a function of culture and gender. We chose4- to 6-year-old children to participate in our study because this is a peak age for makebelieve play. Our approach is based on naturalistic observations of children’s behavior.Children need to be studied in the world in which the development takes place and in thecontexts in which researchers can be sensitive to the subtleties of the social understanding(Bruner, 1990; Dunn, 1988; Pellegrini & Smith, 1998; Smith, 1982).

In our opinion, the naturalistic study of play behavior, allied to a new classification sys-tem that takes into account the structure of making believe, a neglected aspect in the liter-ature about pretending, and its content, will allow a better understanding of differences andsimilarities in this kind of play among children reared under different life conditions.Concerning make believe structure, we predict that given the universality of the process ofthinking and symbolization, there would not be many differences as a function of cultureor gender. Greater differences were expected to occur in the content of play because theyreflect more directly the cultural issues experienced or observed by the children.

Children from cultures that use simpler communication codes—Indians, seashore, andlow SES urban—were expected to engage in less symbolic play than those from culturesthat use elaborated communication codes. We also predict that they would enact fewerroles in sociodramatic play because they have contact with less diverse adult roles. In con-trast, these children were predicted to engage in more physical play.

METHOD

PARTICIPANTS

In all, 76 children were observed, 41 girls and 35 boys, 4 to 6 years old, from five dif-ferent groups: Indians (6 girls and 6 boys), seashore (7 girls and 6 boys), low SES urban(10 girls and 10 boys), high SES urban (13 girls and 7 boys), and mixed SES urban (5 girlsand 6 boys). The Indians were from the Parakanã group living at Paranowaona village, ina reserve located at Pará State. Seashore children were from a small coastal town from SãoPaulo State. The urban samples were from São Paulo, the biggest city of Brazil.

PROCEDURES

The study involved naturalistic observations, using the focal sample method (Altmann,1974) during free time intervals at school for all groups except for Indians, who wereobserved playing freely in several locations in their village. Each observation sessionlasted 5 minutes, divided into 1-minute periods. The mean number of sessions was six per




child. Following Haight et al. (1999), pretend play was defined as a subcategory of play inwhich “actions, objects, persons, places, or other dimensions of the here-and-now aretransformed or treated nonliterally” (p. 1480).

Six trained research assistants collected the observational data.1 Following an extensive train-ing period using videotapes, reliabilities were obtained for all the behavioral codes. Interraterreliability was computed for pairs of researchers on the basis of 30 minutes of randomly chosenvideotape segments. Kappa indices between pairs of judges ranged from .81 to .97.

As Indians and both high SES and low SES urban schools did not allow videotape use,our observations were recorded by hand. Data collection was preceded by a 10-day periodof the children’s acquaintance to the observer presence and also to the video recorder inthe other samples. A clear habituation process occurred in both cases so that at the end ofthis period we noticed almost no reactions to the observers/cameras. Comparing codingsfrom written records with those from Göncü et al. (2000) reported the same finding.Besides that, the written records allowed a broader vision of the context in which the playepisodes occurred, whereas the videotaped records permitted a more detailed observationof nonverbal aspects of the interactions. However, we believe that these differences did notessentially modify our results such as presented here.

Although this study specifically examined pretend play, it is part of a larger research proj-ect examining other kinds of play (Gosso et al., 2005; Gosso & Otta, 2003b; Morais & Otta,2003a). Besides make believe, we used the following categories of play: physical exercise—play that involves various types of movements requesting gross motor coordination (e.g., run-ning, jumping, and swimming) and activities that produce action-contingent effects (e.g.,throwing or pushing objects); social contingency—games apparently motivated and rein-forced by pleasure in producing contingent responses in others and to respond contingentlyto others (e.g., peek-a-boo, tickling, imitating gestures or verbalizations); rough-and-tumble—play that involves both vigorous physical exercise and social contingency but is distinctfrom these categories because it simulates aggression at the same time that it is accompa-nied by playful signals (e.g., play face) that distinguish it from aggression; construction—physical transformations in objects such as sand, clay, or blocks are produced, includingmolding, arranging objects in piles or rows, and making small baskets; and games withrules—those guided by explicit rules, often involving at this age range sensory-motoraspects, such as volleyball and soccer.

General playfulness and pretend play indices were obtained by dividing the number ofoccurrences of each categorized behavior by the number of observation periods. Additionalproportions were calculated dividing the frequency of each category by the total involvementin play (e.g., frequency of rough-and-tumble/the sum of all kinds of play).

Because the focus of our study was on pretend play, this category was analyzed ingreater detail. Each symbolic transformation was analyzed considering its content andideational scheme (Morais & Otta, 2003b).

Pretend themes were categorized according to the predominant play content: work, adultproductive activities in the culture (e.g., hunter, fisher, housekeeper, physician, teacher);transportation, children take a vehicle toy from one place to another or behave as passengersor drivers; take care, caretaking behaviors such as feeding, dressing, bathing, or taking toschool (this category also includes disciplinary actions such as reprehending, sending tobed, ordering silence); animal actions, the child behaves as an animal (e.g., barking, neigh-ing, galloping); play fighting, includes threat, persecution, fight, flight, death, and revival;daily life activities, behaviors pertaining to the daily routine (e.g., eating, sleeping, tele-phoning, leaving home); entertainment, leisure activities (e.g., singing, dancing, going to




parties); fantastic themes, actions related to cartoon, legend, or fairy tale characters and toanimals that do not exist (e.g., Pokemon, Digimon, dragon, dinosaur, vampire, bad wolf,little mermaid, witch, princess).

Implicit or explicit changes of meaning were classified into creation or symbolic mod-ification schemes. Creation is the scheme through which the child behaves as if there weresome elements in the absence of any material noticeable substratum to the observer. Forinstance, a girl uses a pestle for crushing and grinding nonexistent manioc in a mortar, pre-tending to be a woman preparing manioc flour. Modification consists of the followingtypes of meaning change of real substrata:

Identity—The child attributes to a peer or assumes a different identity from the real one or trans-forms an object into another one in such a way that the main characteristics of the new iden-tity are incorporated. The following examples are illustrative of identity modification:object—a packing case is used as a car; a plastic bag is treated as a dressing suit; person—the child behaves as a turtle or as a teacher.

Property—Characteristics not really possessed are attributed to objects (e.g., a toy gun is treatedas a real gun; a doll is treated as a baby) or to living beings (e.g., the child pretends that firejets out from one finger; he or she pretends to fly).

Condition—The child attributes to a peer or assumes a particular state that he or she does not pos-sess at the moment, like pretending to be dead, sick, or mad. This category includes condi-tions that do not exist in the environment in the present situation, such as pretending that itis windy, raining, or cold.

RESULTS

PLAYFULNESS AND SYMBOLIC PLAY RELATIVE TO OTHER KINDS OF PLAY

To provide a broader framework in which to interpret pretend play, we initially exam-ined differences between cultural groups and sexes in the proportion of playtime. TheANOVA conducted for the effects of these variables on playtime yielded main significanteffect for cultural group, F(4) = 20.39, p < .001. Tukey post hoc tests revealed, p < .001, thatIndians played less (56% of time) than the other groups (79% of playtime for seashore,77% for high SES urban, 76% for mixed SES urban, and 66% for low SES urban) and thatlow SES played less than seashore.

As shown in Table 1, in all the five cultural groups the three most frequent forms of playwere pretend, physical, and construction play. Five ANOVAs were conducted for theeffects of cultural group on physical exercise, construction, social contingency, rough-and-tumble play, and games with rules. There were found significant effects on construction,F(4) = 5.48, p < .001; social contingency play, F(4) = 3.73, p < .01; and games with rules,F(4) = 5.26, p < .001. Tukey post hoc tests showed that low SES engaged in more con-struction play than high and mixed SES children and Indians (p < .05), seashore childrenengaged in more games with rules than high and mixed SES urban children and Indians(p < .05), and Indians played more social contingency than high and low SES urbanchildren (p < .05).

Regarding make believe, the ANOVA yielded significant effects of cultural group,F(4) = 8.41, p < .001, and gender, F(1) = 5.34, p < .05. Tukey post hoc tests revealed thathigh SES children played more than low SES (p < .001), Indians (p < .001), and seashorechildren (p < .01); mixed SES played more than Indians (p < .01) and low SES children(p < .05).




Regarding gender, girls showed a higher proportion of make believe than boys in allgroups save Indians. On the other hand, boys were higher than girls in physical exerciseplay except among low SES children. In all groups except Indians, boys also engaged morethan girls in rough-and-tumble.

At the seashore, girls exhibited more social contingency play than boys. Furthermore,mixed SES, high SES, and Indians girls engaged in more construction play than boys,whereas the opposite was found in the low SES.

SYMBOLIC TRANSFORMATIONS: CREATION VERSUS MODIFICATION

Table 2 shows the contribution of cultural groups on some pretend play indices (totalof transformations, creation, and symbolic modifications and the relative creation index[creation/(creation + modification) × 100]).

Two ANOVAs were conducted to examine the effects of gender and cultural group onfrequency of total symbolic transformation and on symbolic modification. The resultsshowed significant effect of cultural groups on total symbolic transformation, F(4) = 9.80,p < .001, and on the score of symbolic modification, F(4) = 10.30, p < .001. Although allgroups engaged in social pretend play, Tukey post hoc tests indicated that high SESchildren showed greater proportion of total symbolic transformation and symbolic modi-fication than seashore, Indians, and low SES children, p < .001, and mixed SES groupshowed higher frequency of total symbolic transformation and symbolic modification thanIndians (p < .05). In all groups, symbolic modification exceeded symbolic creation.

The ANOVA conducted to examine the effect of gender and cultural group on propor-tion of creation in the absence of any material noticeable substratum yielded a main


TABLE 1

Percentages of Various Kinds of Play by Culture and Gender

Indians Seashore Low SES Mixed SES High SES Mean Kinds of Play Gender (%) (%) (%) (%) (%) (%)

Physical exercise Girls 36 19 29 21 25 26Boys 40 38 21 29 36 33Total 38 28 25 25 29 29

Construction Girls 20 17 23 19 12 18Boys 9 20 41 4 7 16Total 14 19 32 11 10 17

Social contingency Girls 11 15 2 4 2 7Boys 9 2 4 5 3 5Total 10 9 3 4 2 6

Rough-and-tumble Girls 7 4 3 3 0 3Boys 7 9 4 14 7 8Total 7 6 4 9 3 6

Pretend Girls 25 32 42 53 60 42Boys 28 15 20 48 46 31Total 26 25 31 51 54 37

Games with rules Girls 0 13 2 0 3 4Boys 7 16 10 0 1 7Total 3 14 6 0 2 5

NOTE: Proportions were obtained dividing raw data by the total play score of each child. SES = socioeconomic status.



significant effect of cultural group, F(4) = 4.83, p < .01. Tukey post hoc tests showed thathigh SES urban children engaged in more creation than low SES and Indians (p < .01).Mixed SES children were in between. Some examples of creation are pretending to be ina spaceship, to be in a car and fasten a seat belt, swimming in a river, using a makeup kit,or taking a shower with soap in the absence of any substratum.

The ratio between creation and modification was greater in high than in low SES group(t = 2.64, p < .05), and there was a tendency toward significance between high SES andIndians (t = 1.85, p = .07).

Different kinds of creation were found among the various groups. Excepting Indians,children created mainly objects (e.g., cars, food, drinks, cleaning items, makeup items).

SYMBOLIC MODIFICATION: IDENTITY, PROPERTY, AND CONDITION MODIFICATION

The most frequent form or modification was in object property; second, in object identity;and third, in character identity. The least frequent form of modification was condition (Table 3).

Among Indians, object property modification (e.g., a toy wooden boat for a real boat)was more frequent in boys, whereas object identity predominated in girls (e.g., lemon forsoap, leaf for washbasin). A similar pattern was observed among seashore and mixed SESurban children, but character identity modification (e.g., pretending to be mother, child,horse, dog) was more common in this group than object identity modification. In both lowand high SES urban children, object identity predominated over object property modifica-tion in boys, whereas the opposite occurred in girls.

As these categories refer to structure of symbolic play, we might expect that there wereno differences among cultural groups. We observed a great variation among groups and


TABLE 2

The Mean Number (and Standard Errors) of Total of SymbolicTransformations, Creation, Symbolic Modification, and Relative

Creation Index Across the Five Cultural Groups

Indians Seashore Low SES Mixed SES High SES Mean

Measures Gender M SE M SE M SE M SE M SE M SE

Total of symbolic Girls 0.12 0.12 0.36 0.11 0.44 0.09 0.70 0.13 0.77 0.10 0.48 0.05transformationa Boys 0.15 0.12 0.24 0.12 0.26 0.10 0.39 0.12 0.72 0.08 0.35 0.05

Total 0.14 0.08 0.30 0.08 0.32 0.07 0.54 0.09 0.74 0.07 0.42 0.04Modificationa Girls 0.12 0.09 0.29 0.09 0.40 0.07 0.59 0.10 0.61 0.08 0.40 0.04

Boys 0.12 0.09 0.21 0.09 0.24 0.08 0.33 0.09 0.58 0.07 0.30 0.04Total 0.12 0.07 0.25 0.06 0.29 0.05 0.46 0.07 0.59 0.05 0.35 0.03

Creationa Girls 0.00 0.04 0.08 0.04 0.05 0.03 0.12 0.04 0.16 0.04 0.08 0.02Boys 0.04 0.40 0.02 0.04 0.01 0.04 0.05 0.04 0.14 0.03 0.05 0.02Total 0.02 0.03 0.05 0.03 0.03 0.02 0.08 0.03 0.15 0.02 0.07 0.01

Relative Girls 0.00 6.30 15.50 5.80 13.00 4.80 13.90 6.90 20.30 5.40 12.50 2.60creation indexb Boys 17.60 6.30 13.00 6.30 2.80 5.40 15.90 6.30 21.20 4.40 14.10 2.60

Total 8.80 4.40 14.30 4.30 7.90 3.60 14.90 4.60 20.70 3.50 13.30 1.80

NOTE: SES = socioeconomic status.a. Raw scores were divided by the number of observation periods.b. Relative creation index = (creation/modification + creation) × 100.



genders. Except for the low SES group, the unique clear gender difference that occurredwas that girls showed more character identity modification than boys. We suppose that thedifferent tendencies in object identity and property modification among the groups weredue to the kinds of objects available and the theme of play preferred.

CHARACTERS ENACTED IN SOCIAL PRETEND PLAY

In their sociodramatic play, children assumed mainly family members’ identity; second,animals; and third, fantastic characters (Table 4). In the five cultural groups, girls tendedto assume family roles. Maternal role predominated among girls in all groups, whereaspaternal role was exclusively enacted by high SES children, predominantly by boys.Animals were represented by all cultural groups but prevailed among seashore childrenwho played dogs and horses. Police and robber characters prevailed among high SESurban, mixed SES urban, and seashore boys. It is notable that among low SES urbanchildren, only girls exhibited identity modification.

Some examples are reported to illustrate these statements. The following episodes arerepresentative of playing mothering during daily activities. The different cultural groupsshowed peculiarities of their way of life.

Akwawia (6-year-old Indian girl) wrapped the ends of a piece of cloth around two tree trunks pre-tending to be a hammock. Then she takes her little 2-year-old brother and sits him down inthe hammock swinging him back and forth, pretending he is a baby.

Jade (5-year-old seashore girl), pretending to be Miriam’s (5-year-old peer) mother, orders her to buymilk, candy, chocolate, and cake in the grocery. Miriam pretends to go to the grocery (a bench)and shows an imaginary bag to Jade and asks: “Was it just that that I was supposed to buy?”

Clara (4-year-old mixed SES urban girl), pretending to be Letícia’s (4-year-old peer) mother, dri-ves her daughter to a “shopping mall.” They are sitting on the chairs of a merry-go-round asif it were a car. Clara says: “Look at that car stopped in the middle of the street!”


TABLE 3

Percentage of Identity, Property, and Condition ModificationAcross the Five Cultural Groups

Indians Seashore Low SES Mixed SES High SES MeanMeasures Gender (%) (%) (%) (%) (%) (%)

Object identity Girls 54 24 39 24 16 32Boys 22 6 61 22 43 31Total 38 15 50 23 30 31

Character identity Girls 11 43 5 48 33 28Boys 7 27 5 19 19 15Total 9 35 5 34 26 22

Object property Girls 31 25 47 23 50 35Boys 70 63 22 52 38 49Total 51 44 35 38 44 42

Condition Girls 3 8 0 4 1 3Boys 2 4 0 6 0 3Total 3 6 0 5 1 3

NOTE: SES = socioeconomic status.



Some examples of playing animals are:

Saying she is a turtle, Wewe (5-year-old Indian girl) jumps into the river and swims.Igor (5-year-old seashore boy) barks inside a covered plastic box. Iago (5-year-old peer) comes

near him also barking and uncovers the box. Igor goes out, walking on four limbs. Mateus(4-year-old peer) imitates Igor saying: “I am an angry dog.” The barking boys on four limbspursue some girls.

Luiza (5-year-old high SES urban girl) pretends she is a wolf and Nicolas (5-year-old peer) pur-sues the “wolf.”

Finally, some examples of fantastic characters:

Pedro (4-year-old high SES urban boy), pretending to be a Tentomon shooting a blast of electricenergy at his opponent (Kuwagamon, Gabriel, 4-year-old peer), says: “I knew everything, Ievoluted. I am going to kill you now.”

Yan (5-year-old mixed SES urban boy), wearing an orange cloth on his shoulders, invites Letícia(4-year-old peer) to play vampire saying: “I have been wearing this cloth for more than athousand years. I could borrow it to you.” As Letícia can’t wear the cloth properly, Yan takesit back and starts to run, pretending to fly. During the whole episode, Yan empowers himselfand suppresses Letícia’s power.

SOCIAL PRETEND PLAY THEMES

Table 5 shows the percentages of make believe play contents across the five groups. Inall cultural groups, the most frequent themes observed in girls’ play were work and takingcare, followed by daily life activities. Among boys, transportation themes prevailed, fol-lowed by play fighting, work, and entertainment.


TABLE 4

Percentages (and Absolute Figures) of Characters Assumed or Attributed inPretend Play by Children as a Function of Group and Gender

Indians Seashore Low SES Mixed SES High SES Total

Characters Gender % n % n % n % n % n % n

Mother Girls 44 4 22 7 16 6 38 3 49 25 33 45Boys 12 2 11 3 8 5

Father Girls 2 1 1 1Boys 21 6 10 6

Other adults Girls 7 2 5 2 38 3 4 2 7 9Boys 32 6 14 4 15 10

Baby/child Girls 44 4 10 3 24 9 33 17 24 33Boys 11 2 4 1 5 3

Police/thief GirlsBoys 25 4 42 8 11 3 23 15

Fantastic characters Girls 3 1 24 9 12 6 12 16Boys 16 3 29 8 17 11

Animals Girls 11 1 58 18 32 12 25 2 24 33Boys 100 3 63 10 11 3 24 16

Total Girls 9 31 38 8 51 137Boys 3 16 19 28 66




Similarities among cultural groups were found mainly for taking care and play fighting.Some cultural group peculiarities are notable for work, transportation, animal actions,daily life activities, and entertainment. Regarding fantastic themes, the noticeable datum isthe absence of this category in both Indians and seashore groups. In contrast, more enter-tainment themes were observed in these groups, for instance:

Tapiawa (4-year-old Indian boy) pretends he is dancing as an adult at the White Urubu (a kind ofSouth America vulture) Festival.

Levi (5-year-old seashore boy) acts as if he were taking a bath in a pretend waterfall.

Regarding work, domestic tasks prevailed among girls in the five groups, although ineach sample peculiarities emerged. Girls reproduced female adult activities prevailing intheir cultural group. These activities turned mainly around the preparation of manioc flour(e.g., grounding, making fire, toasting) among Indians and around house cleaning (e.g.,sweeping, dusting) and cooking (e.g., making coffee, cakes, and typical Brazilian candies)in seashore and urban groups. In contrast with the indoor activities of the girls, the boysperformed mainly outdoor activities like hunting and fishing among Indians and profes-sional roles such as physicians, car/ambulance/truck/bus drivers, and policemen in theseashore and urban groups. Some examples of girls’ work are:

Akawia (6-year-old Indian girl) pretends making manioc flour.Gabriela (5-year-old seashore girl) sweeps the ground with a toy broom.Talia (4-year-old low SES urban girl) pretends she is a seller. She is selling pastel (a kind of fried

pastry, a typical Brazilian snack).Rafaela (5-year-old high SES urban girl) pretends she is manicurist polishing the nails of her friend.

Regarding boys’ work, some examples are:

Suruapa (4-year-old Indian boy) pretends he is fishing. He throws a fishing net on the ground atcenter of the village.


TABLE 5

Percentage of Make Believe Play Contents Across the Five Groups

Measures Gender Indians % Seashore % Low SES % Mixed SES % High SES % Mean %

Work Girls 21 28 42 5 40 27Boys 4 12 17 6 42 16

Transportation Girls 18 4 9 6Boys 39 19 54 29 4 29

Taking care Girls 29 12 19 42 20 25Boys 8 11 4

Animal actions Girls 5 27 13 17 12Boys 6 37 1 8 11

Play fighting Girls 8 15 5 6Boys 2 12 8 54 8 17

Daily life activity Girls 26 19 19 23 10 19Boys 11 7 4 1 11 7

Entertainment Girls 1 5 1 6 2Boys 21 14 3 2 8 10

Fantastic themes Girls 1 3 4 2 2Boys 3 7 9 4




Pedro (4-year-old mixed SES urban boy) pretends being a doctor. He calls the first child who iswaiting in the line.

Lucas (5-year-old low SES urban boy) pretends he is a bricklayer building a house.

Pretended means of transportation were mainly canoes for Indians boys, trains and trac-tors for seashore boys, and buses and cars for urban boys. Concerning animal actions,Indian children played turtle, deer, and fish; seashore children played mainly dogs andhorses; and urban children used mainly toy animals (rabbits, horses, etc.) as pets. Finally,concerning daily life activities, Indian children played activities characteristic of their cul-ture such as hanging a hammock and taking a bath in the river (out of it), seashore childrenplayed going to the grocery, mixed and high SES urban groups pretended driving theirchildren to school and going to the shopping mall, and the low SES urban group pretendedriding a bus and some activities also played by other urban groups such as eating andspeaking by phone. Regarding gender, whereas girls engaged more than boys in conversa-tion, hygiene/beauty activities, rest, and shopping, boys were more interested than girls inmoney (e.g., counting) and sex (e.g., genital naming and copulation faking).

DISCUSSION

In the present study, we found that Indian children engaged in less play than the othercultural groups. First of all, we have to consider that these children were observed in dif-ferent conditions than the remaining groups. Urban and seashore children were observedduring free time intervals between classes, whereas Indians had all the time to play, except-ing 1 hour dedicated to classroom activities. Furthermore, Indian children lived togethermost of the day in the same village, whereas the other ones met their peers predominantlyat school. Social peer deprivation or play deprivation could be responsible for increasedplayfulness among non-Indians. In addition, as Indian children work much earlier thanhigh and mixed SES urban groups (e.g., washing clothes at the river, taking care of theiryounger siblings, and cooking), work may fill the time that would otherwise be taken forplay or could partially accomplish its psychological function. Even though seashore andlow SES urban children also begin to work early, probably they played more than Indiansbecause they were observed only in their free time interval.

Concerning make believe, we observed that high SES urban and mixed SES urbangroups engaged in a greater proportion of pretending than low SES, Indians, and seashorechildren. Although this finding was predicted, it must be pointed out that pretend play wasnoticed in all cultural groups, supporting the theoretical claim that making believe occursuniversally. Similar findings are reported in the literature about socioeconomical differ-ences: High- and middle-class children play more make believe than low-class children(Fein, 1981). In addition, little symbolic play, characteristic in our study of Indians andseashore children, was also found by Martini (1994) in the Marquesas Islands in Polynesiaand by Gaskins (2000) among Mayan children. The low frequency of pretending in Indian,seashore, and low SES urban may be explained by the predominance of simple communi-cation codes in these societies. Insofar as they are more worried with their material sur-vival issues, their way of rearing emphasizes concrete and immediate solutions to theirdaily life problems instead of strengthening symbolic and abstract thought. This interpre-tation is in accordance with Keller’s (1998) statement that the societies with less role spe-cialization require fast learning and immediate solution to practical problems.




However, we have to consider that the emphasis of Indians and seashore cultural groupson practical solutions to daily life problems may be associated with precious developmentalopportunities: deep contact with nature and native forests, hills and high trees to climb, rivers,sea, and wild animals to master. Besides that, the children in these societies live under lessadult supervision, and almost all adult-world objects are available to them without prohibi-tions. Children’s play may be useful for parents by keeping children occupied in theirautonomous world and allowing parents to conduct their activities unimpeded by theirdemands (Gosso et al., 2005). The highly challenging environment in which they live mightreduce their tendency to look for desire fulfillment through fantasy. This interpretation isconsistent with the finding that more fantastic themes were practically absent in Indians andseashore play (Gosso et al., 2005; Gosso & Otta, 2003a; Morais & Otta, 2003b).

We can think that a setting more similar to the evolutionary adaptation environment pro-vides optimal developmental opportunities to the developing being. Our children who live inbig cities, even the rich ones, might benefit from greater contact with a more natural way ofliving. We noticed that seashore and Indian children showed less stress, more pleasure andhappiness in their play, and other behavioral manifestations. Teachers’ spontaneous com-ments on seashore children revealed that they considered them calm, curious, and clever, instriking contrast with urban teachers who regarded a great part of their pupils as behaviorallydisruptive, with learning disabilities and with social–affective and familiar problems. In addi-tion, the opportunity to live together with peers of different ages and several adults providesIndians and seashore children with several possibilities of interaction; for example, mother-ing (that involves the abilities to recognize and to assume others’ perspective) is more likelyto occur in interaction with younger children (Carvalho & Beraldo, 1989).

Regarding time dedicated to each type of play (Table 1), we observed that as predicted,in Indians, seashore, and low SES groups, the frequency of physical exercise play wassuperior or similar to that of make believe. In the low SES group, construction was as fre-quent as pretend play, and for seashore children, construction occurred almost as often aspretend play. Construction play might have a symbolic purpose similar to make believebecause it requires the establishment of a goal, planning, and execution. Thus, it could ful-fill a similar function to that of pretending: “For cognitive development, dramatic play isno more beneficial than constructive play” (Smith, 1988, p. 222).

Among seashore children, social contingency and games of rules followed pretendingand construction play in frequency. Among Indians, social contingency play also repre-sented 10% of play time, exceeding the frequency shown by urban children in this kind ofplay. We can interpret these data, taken as a whole, in the broader framework of culturaldimensions. We suppose that the proximal and collectivistic component of Indian andseashore groups underlies the finding of higher frequency of social contingency play inthese children. Because this kind of play allows knowledge of the effect of one’s behavioron another and is essential for the development of understanding his or her reactions andfundamental to a society based in cooperation and empathy, it is not surprising to find ahigher frequency of social contingent play in collectivistic and proximal societies. If weconsider rough-and-tumble as a kind of social contingency play insofar as it is motivatedby provoking the other’s reaction (Morais, 2004; Smith, 1988) and if we add these twokinds of play, this conclusion becomes even more evident.

Concerning the structure of pretend play, we found that the frequency of total symbolictransformation and symbolic modification was greater in high SES urban and in mixedSES children. In all groups, symbolic modification exceeded symbolic creation.




Considering roles played in make believe, children assumed mainly family members’identity; second, animals; and third, fantastic characters. In the five cultural groups, girlstended to assume family roles, predominantly mother and babies, sons, and daughters.Animals were represented by all cultural groups but prevailed among seashore childrenwho played dogs and horses. Police and robber characters prevailed among high SESurban, mixed SES urban, and seashore boys. In all cultural groups, we found that the mostfrequent themes observed in girls’ play were work and taking care, followed by daily lifeactivities. Among boys, transportation themes prevailed, followed by play fighting, work,and entertainment.

Concerning the structure of make believe play, as predicted, the similarity among thevarious cultural groups regarding the ratio of creation and symbolic transformations andidentity modification of object and characters as a whole leads us to conclude that thereare no differences in cognitive functioning of the children of various cultural backgrounds.

The differences among cultural groups were expressed more in the content than in thestructure of pretend play. The great frequency of character modification among high andmixed SES urban children may be due to their media exposition to fantastic models and tothe motivation to enact roles that they cannot perform in their real life. The predominanceof domestic animals (mainly dogs and horses) in seashore play may be due to the closecontact of the children with them in their daily life. The low frequency of role-playing byIndians and low SES urban children is understandable because they already perform someadult roles. In accordance with the interpretation of communication code models, high andmixed SES children might need to develop a greater variety of adult roles to cope with thedemands of their cultural world than low SES urban, seashore, and Indian groups.

We observed a high frequency of transportation themes among Indians and low SESurban. Insofar as Indians are concerned, we notice that although canoes were commonlyused, pickup trucks and motorboats were a sign of status because they were used to makelong-distance travels. Only Toria (White people) drove pickup trucks and only certain adultmale Awaete (Parakanã Indians) drove motorboats. Everybody including older children(otyaro) could row canoes up and down the river; that is why, in our opinion, rowing canoeswas not represented in make believe play. Regarding low SES urban children, driving is alsoa sign of status because having a car is an indication of heightening of social class.

Regarding gender differences, we found that girls represented a greater and more variednumber of roles than boys in all cultural groups except mixed SES urban. There are severalpossibilities to explain this finding: The girls are more focused on persons and social interac-tions than boys, as suggested by other researchers (Bichara, 1994; Morais & Carvalho, 1994;Tarullo, 1994); female adult roles are more varied and flexible in comparison to male adultroles (we suppose that this is a worldwide phenomenon); mothers are in closer contact withchildren than fathers. The private versus public spheres are represented in the sociodramaticplay. Both Awaete and Toria women—the mother or a substitute—stay at the home base,whereas both Awaete and Toria men go out to work (Gosso et al., 2005).

The girls enacted more family and domestic themes, whereas the boys played more fan-tastic and transportation themes, a fact widely documented in the literature (e.g., De Conti &Sperb, 2001; Fein, 1981; Morais, 2004; Nicolopoulou, 1997; Smith, 2005). Some exceptionsare notable: the absence of fantastic themes among Indians and low SES urban boys.

These boys had an especially low frequency of role enactment in general. In spite ofthat, when we examined the play themes, we found out that Indian boys played workthemes—basically hunting and fishing—whereas low SES urban boys played work andvehicle themes. These findings may suggest that in these two cultural groups, adult male




work is more accessible to their children, and in accordance with a simple communicationcode model, this role is more clearly defined, understandable, and expected to be followed.

In the present study, similarities among cultures were just as evident as differences.Because traditional gender stereotypes were present in role-playing of all five groups, andchildren most of the time reproduced real-life situations of their societies, all the culturalgroups seemed, in our opinion, more traditional and conservative than innovative. Somequestions deserve additional investigation, such as how do the adult models work in eachcultural context, which are the rearing and educational practices, and which are the socialworlds of conventions, rules, and role valorization. The structure and value transmission ofthese worlds have to be investigated more to base cross-cultural comparisons on a betterunderstanding of the social representations and messages that are transmitted to thechildren concerning their present and their expectancies about the future.

Besides the macro system to which the children belong, we should consider the micro sys-tem: the degree of interaction and the pattern of relationship among the children establishedby the history of their interactions, the influence of patterns of leadership and submission onplay behavior and group functioning, and the influence of individual characteristics—such ascognition, language, emotional development, and personality traits—on play preferences.

In spite of all the suggestions we gave to improve the understanding of our data, ourmethodology has proved to be a very promising approach to the cross-cultural study ofmake believe play regarding the comprehension and description of the characteristics ofcultural groups as well as the proposed perspective of analysis of pretend play structure.

NOTE

1. The first author (Y.G.) conducted the observations of Parakanã Indians. To be allowed to do this research,she was requested to give classes to children and adults in the Parakanã language at the local school. The secondauthor (M.L.S.M.) conducted the observations of seashore children and mixed socioeconomic status urban. Fourtrained research assistants observed the other two samples.

REFERENCES

Altmann, J. (1974). Observational study of behavior sampling methods. Behavior, 49, 227-265.Andersen, P. A. (1999). Nonverbal communication: Forms and functions. Mountain View, CA: Mayfield.Beraldo, K. E. A., & Carvalho, A. M. A. (2003). Na cidade grande [In the big city]. In A. M. A. Carvalho,

C. M. C. Magalhães, F. A. R. Pontes, & I. Bichara (Eds.), Brincadeira e cultura: Viajando pelo Brasil quebrinca (Vol. 1, pp. 157-186). São Paulo, Brazil: Casa do Psicólogo.

Bernstein, B. (1970). Class, codes and control, I: Theoretical studies towards a sociology of language. London:Routledge & Kegan Paul.

Berry, J. W., Poortinga, Y. H., Segall, M. H., & Dasen, P. R. (1996). Cross-cultural psychology: Research andapplications. Cambridge, UK: Cambridge University Press.

Bichara, I. D. (1994). Um estudo etológico da brincadeira de faz-de-conta em crianças de três a sete anos [Anethological study of pretend play in three- to seven-year-old children]. Unpublished doctoral dissertation,Institute of Psychology, University of São Paulo, Brazil.

Bichara, I. D. (2002). Crescer como Índio ás margens do Velho Chico: Um desafio para as crianças Xocó [Growup as an Indian around “Velho Chico”: A challenge for Xocó children]. In E. R. Lordelo, A. M. A. Carvalho,& S. H. Koller (Eds.), Infância brasileira e contextos de desenvolvimento (pp. 137-163). São Paulo, Brazil:Casa do Psicólogo.

Bornstein, M. H., Haynes, M., Pascual, L., Painter, K. M., & Galperin, C. (1999). Play in two societies:Pervasiveness of process, specificity of structure. Child Development, 70, 317-331.




Bruner, J. S. (1990). Culture and human development: A new look. Human Development, 33, 344-355.Carvalho, A. M. A., & Beraldo, K. E. (1989). Interação criança-criança: Ressurgimento de uma área de pesquisa

e suas perspectivas [Child-child interaction: Renewal of an area of research and its perspective]. Cadernos dePesquisa, 71, 55-61.

Carvalho, A. M. A., Beraldo, K. A., Santos, F., & Ortega, R. (1993). Brincadeiras de menino, brincadeiras demenina [Boys game, girls game]. Psicologia: Ciência e Profissão, 13, 30-33.

Carvalho, A. M. A., & Pedrosa, M. I. (2003). Teto, ninho, território: Brincadeiras de casinha [Shelter, nest, terri-tory: Playing with houses]. In A. M. A. Carvalho, C. M. C. Magalhães, F. A. R. Pontes, & I. Bichara (Eds.),Brincadeira e cultura: Viajando pelo Brasil que brinca (Vol. 2, pp. 31-48). São Paulo, Brazil: Casa doPsicólogo.

Conway, L. G., Ryder, A. G., Tweed, R. G., & Sokol, B. W. (2001). Intranational cultural variation: Exploring furtherimplications of collectivism within the United States. Journal of Cross-Cultural Psychology, 32, 681-697.

De Conti, L., & Sperb, T. M. (2001). O brinquedo de pré-escolares: Um espaço de ressignificação cultural[Preschoolers’ play: A space for cultural re-signification]. Psicologia: Teoria e Pesquisa, 17(1), 59-67.

Dunn, J. (1988). The beginnings of social understanding. Cambridge, MA: Harvard University Press.Eibl-Eibesfeldt, I. (1972). Similarities and differences between cultures in expressive movements. In R. A. Hinde

(Ed.), Non-verbal communication (pp. 297-314). Cambridge, UK: Cambridge University Press.Else-Quest, N. N., Hyde, J. S., Goldsmith, H. H., & Van-Hulle, C. A. (2006). Gender differences in temperament:

A meta-analysis. Psychological Bulletin, 132, 33-72.Farver, J. M., Kim, Y. K., & Lee-Shin, Y. (2000). Within cultural differences: Examining individual differences

in Korean American and European American preschoolers’ social pretend play. Journal of Cross-CulturalPsychology, 31, 583-602.

Fein, G. G. (1981). Pretend play in childhood: An integrative review. Child Development, 52, 1095-1118.Garvey, C., & Kramer, T. L. (1989). The language of social pretend play. Developmental Review, 9, 364-382.Gaskins, S. (2000). Children’s daily activities in a Mayan village: A culturally grounded description. Cross-

Cultural Research, 34, 375-389.Göncü, A., Mistry, J., & Mosier, C. (2000). Cultural variations in the play toddlers. International Journal of

Behavioral Development, 24, 321-329.Gosso, Y., & Otta, E. (2003a). Em uma aldeia Parakanã [At a Parakanã village]. In A. M. A. Carvalho, C. M. C.

Magalhães, F. A. R. Pontes, & I. D. Bichara (Eds.), Brincadeira e cultura: Viajando pelo Brasil que brinca(Vol. 1, pp. 33-76). São Paulo, Brazil: Casa do Psicólogo.

Gosso, Y., & Otta, E. (2003b, June). Parakanã Indian children at play. Paper presented at the 33rd meeting ofJean Piaget Society, Chicago.

Gosso, Y., Otta, E., Morais, M. L. S., Ribeiro, F. J. L., & Bussab, V. S. R. (2005). Play in hunter-gatherer society.In A. D. Pellegrini & P. K. Smith (Eds.), The nature of play (pp. 213-253). New York: Guilford.

Greenfield, P. M., Keller, H., Fuligni, A., & Maynard, A. (2003). Cultural pathways through universal develop-ment. Annual Review of Psychology, 54, 461-490.

Haight, W. L., Wang, X., Fung, H. H., Williams, K., & Mintz, J. (1999). Universal, developmental, and variableaspects of young children’s play: A cross-cultural comparison of pretending at home. Child Development, 70,1477-1488.

Harris, P. L. (1989). Children and emotion: The development of psychological understanding. Oxford, UK: BasilBlackwell.

IBGE. (2000). Anuário Estatístico. Brasil [Brazilian statistical yearbook]. Rio de Janeiro: Author.Jensvold, M. L. A., & Fouts, R. S. (1993). Imaginary play in chimpanzees (Pan troglodytes). Human Evolution,

8, 217-227.Keller, H. (1998). Diferentes caminhos de socialização até a adolescência [Different ways of socialization until

the adolescence]. Revista Brasileira de Crescimento e Desenvolvimento, 8, 1-14.Keller, H., Borke, J., Yovsi, R., Lohaus, A., & Jensen, H. (2005). Cultural orientations and historical changes as

predictors of parenting behaviour. International Journal of Behavioral Development, 29, 229-237.Keller, H., Lohaus, A., Kuensemueller, P., Abels, M., Yovsi, R., Voelker, S., et al. (2004). The bio-cultural of par-

enting: Evidence from five cultural communities. Parenting: Science and Practice, 4, 25-50.Kitayama, S., Markus, H. R., Matsumoto, H., & Norasakkunkit, V. (1997). Individual and collective processes in

the construction of the self: Self-enhancement in the United States and self-criticism in Japan. Journal ofPersonality and Social Psychology, 72, 1245-1267.

Lillard, A. (2001). Pretend play as twin earth: A social-cognitive analysis. Developmental Review, 21, 495-531.Lock, A. J. (2000). Phylogenetic time and symbol creation: Where do zopeds come from? Culture & Psychology,

6, 105-129.





Magalhães, C. M. C., Souza, A. R., & Carvalho, A. M. A. (2003). Piras no Riacho Doce [“Piras” at Riacho Doce].In A. M. A. Carvalho, C. M. C. Magalhães, F. A. R. Pontes, & I. Bichara (Eds.), Brincadeira e cultura:Viajando pelo Brasil que brinca (Vol. 1, pp. 77-88). São Paulo, Brazil: Casa do Psicólogo.

Martini, M. (1994). Peer interactions in Polynesia: A view from the Marquesas. In J. L. Roopnarine, J. E.Johnson, & F. H. Hooper (Eds.), Children’s play in diverse cultures (pp. 73-103). Albany: State University ofNew York Press.

Matthews, W. S. (1977). Modes of transformation in the initiation of fantasy play. Developmental Psychology,13, 212-216.

Morais, M. L. S. (1980). O faz-de-conta e a realidade social da criança [Make believe and child social reality].Unpublished doctoral dissertation, Institute of Psychology, University of São Paulo, Brazil.

Morais, M. L. S. (2004). Conflitos e(m) brincadeiras infantis: Diferenças culturais e de gênero [Cultural andgender differences in children’s play and conflicts]. Unpublished doctoral dissertation, Institute ofPsychology, University of São Paulo, Brazil.

Morais, M. L. S., & Carvalho, A. M. A. (1994). Faz-de-conta: Temas, papéis e regras na brincadeira de criançasde quatro anos [Make believe play of four-year-old children: Themes, roles, and rules]. Boletim de Psicologia,100-101, 21-30.

Morais, M. L. S., & Otta, E. (2003a). Entre a serra e o mar [In between the mountains and the sea]. In A. M. A.Carvalho, C. M. C. Magalhães, F. A. R. Pontes, & I. D. Bichara (Eds.), Brincadeira e cultura: Viajando peloBrasil que brinca (pp. 127-156). São Paulo, Brazil: Casa do Psicólogo.

Morais, M. L. S., & Otta, E. (2003b, June). Pretend play in two Brazilian communities. Paper presented at the33rd meeting of Jean Piaget Society, Chicago.

Moura, M. L. S. (2003). Cross-cultural studies of human development: Some theoretical and methodological con-siderations [Abstract]. In A. A. Dias & A. Roazzi (Ed.), Anais do IV Congresso Brasileiro de Psicologia doDesenvolvimento (pp. 132-134). João Pessoa, Brazil: Editora Universitária.

Nicolaci-da-Costa, A. M. (1988). Estruturas sociais e subjetividade [Social structures and subjectivity][Abstract]. In Sociedade Brasileira de Psicologia (Eds.), Anais da XVIII Reunião Anual de Psicologia(pp. 633-642). Ribeirão Preto, Brazil: SBP.

Nicolopoulou, A. (1997). Worldmaking and identity formation in children’s narrative play-acting. In B. D. Cox& C. Lightfoot (Eds.), Sociogenetic perspectives on internalization (pp. 157-187). Mahwah, NJ: LawrenceErlbaum.

Parakanã Program. (2000). População Parakanã [Parakanã population]. Unpublished annual report.Pellegrini, A. D., & Smith, P. K. (1998). Physical activity play: The nature and function of a neglected aspect of

play. Child Development, 69, 577-598.Piaget, J. (1975). A formação do símbolo na criança [Play, dreams and imitation in childhood] (A. Cabral,

Trans.). Rio de Janeiro, Brazil: Zahar. (Original work published 1945)Price, W. F., & Crapo, R. H. (1999). Cross-cultural perspectives in introductory psychology. Belmont, CA: ITP.Santos, E. C., & Koller, S. H. (2003). Brincando na rua [Playing on the street]. In A. M. A. Carvalho, C. M. C.

Magalhães, F. A. R. Pontes, & I. Bichara (Eds.), Brincadeira e cultura: Viajando pelo Brasil que brinca (Vol.1, pp. 187-206). São Paulo, Brazil: Casa do Psicólogo.

Santos, J. E. F. (2003). Brincando na maré: Infância e brincadeiras em novos alagados [Playing in the tide:Childhood and play in “Novos Alagados”]. In A. M. A. Carvalho, C. M. C. Magalhães, F. A. R. Pontes, &I. Bichara (Eds.), Brincadeira e cultura: Viajando pelo Brasil que brinca (Vol. 1, pp. 109-126). São Paulo,Brazil: Casa do Psicólogo.

Smith, P. K. (1982). Does play matter? Functional and evolutionary aspects of animal and human play.Behavioral and Brain Sciences, 5, 139-184.

Smith, P. K. (1988). The role of rough and tumble play in the development of social competence: Theoretical per-spectives and empirical evidence. In B. H. Schneider, G. Attili, J. Nadel, & R. P. Weissman (Eds.), Socialcompetence in developmental perspective (pp. 239-255). London: Khuver.

Smith, P. K. (2005). Social and pretend play in children. In A. D. Pellegrini & P. K. Smith (Eds.), The nature ofplay (pp. 173-212). New York: Guilford.

Takahashi, K., Ohara, N., Antonucci, T. C., & Akiyama, H. (2002). Commonalities and differences in close rela-tionships among Americans and Japanese: A comparison by the individualism/collectivism concept. Inter-national Journal of Behavioral Development, 26, 453-465.

Tarullo, L. B. (1994). Windows on social worlds: Gender differences in children’s play narratives. In A. Slade &D. P. Wolf (Eds.), Children at play: Clinical and developmental approaches to meaning and representation(pp. 169-187). New York: Oxford University Press.



Y. Gosso received her PhD in psychology from the Department of Experimental Psychology of theUniversidade de São Paulo (University of São Paulo), Brazil, in 2005. For several months she has livedin a Brazilian indigenous village to collect her data on children’s time budget and play. Her researchinterests include play and cross-cultural research.

Maria L. S. Morais, scientific researcher of the Health Institute, São Paulo State Department of Health,received her PhD in psychology from São Paulo University, Institute of Psychology, Department ofExperimental Psychology. In the child developmental psychology area, her research interests include playand conflicts among preschool children, focusing on gender and cultural differences. In the public healthfield, her work consists of investigations in Brazilian primary care issues, with special attention to thesymbolic dimension of access to health services.

Emma Otta received the PhD in experimental psychology from the Universidade de São Paulo, Institutode Psicologia, Programa de Pós-Graduação em Psicologia Experimental. She is professor of psychologyat the Universidade de São Paulo. Her research interests include human motivation, emotion, and devel-opmental psychology from a cross-cultural perspective.




journal of cross-cultural psychology

Documents