Engaging with Dolls and with Play in Rural Southwest Angola

Inês Ponte (University of Manchester)


An intricate relationship between ritual and play has dominated the topic of African dolls in museums on account of historical field ethnographies based on a ritual-focused agenda. My doll-focused research agenda, grounded on eight months of fieldwork in an agro-pastoralist highland village in Namibe, proposes an alternative engagement to that of earlier scholars working in Southwest Angola. Engaging with dolls handmade by children in the said fieldwork setting, I combine information about dolls housed in museum collections with the social and material life of these artefacts in the region of their making. Privileging play, I discuss children’s engagements with dolls made from corncobs, thus offering a critical review of ethnographic approaches that constrain these dolls under differing concerns with ritual. Attending to young children’s participation in the dynamics of a subsistence domestic economy, I reconnect the material and social dimensions of doll use through children’s voices and practices, and argue that we may link children’s engagements with available materials and devices to the imaginative worlds they develop while growing up in a context organised around subsistence livelihood concerns.

Keywords: artefact-based research, ethnography, play, domestic economy, material culture, rural, Angola, agro-pastoralism


Five-year-old Kabuka was one of the youngest children living in the highland farm that hosted me in Namibe province, southwestern Angola. During that period, Kabuka allowed me to participate with her in the sporadic and spontaneous corncob-doll play that occurred within her day-to-day life. A variety of similar dolls are housed in Western museums, some of which I researched through short-term visits, but my fieldwork simultaneously focused on understanding the significance and social roles of handcrafted dolls through engagement with their local making and uses, in particular play. Omphunda, the agro-pastoralist village where Kabuka lived, was my main fieldsite for eight months in 2012. No museum collections have registered dolls produced in Omphunda, and the common practice of labelling museum artefacts by ethnicity to the detriment of other identifying information makes it difficult to trace specific doll makers or their communities of origin. Kabuka’s spontaneous play allowed me to understand how these handmade dolls are created and used in day-to-day life.

African dolls are a well-developed topic in museum publications such as exhibition catalogues, yet for certain regions, such as Southern Africa, these publications often describe exhibited specimens using historical ethnographies in the present tense, creating an ahistorical perspective. There are a few exceptions, such as Margo Timm (1998), who explores doll play and use through childhood memories of adult Ovakwanyama women in Northern Namibia, or Frank Jolles (2010), who analyses a contemporary private South African collection of dolls through the biographies of the artisans’ who sold them. Another strand of artefact-based research looks into the political economy of doll-making and use. For instance, Norma Wolff (2004) discusses the appeal of African craft by exploring wax candles sold in museum gift shops in the United States that are produced on an industrial scale in China, copying the form of wooden handmade Ghanaian ‘fertility dolls’.

In this article, I frame my contemporary data within the evolving history of regional ethnographic understandings of doll play. My ethnographic research traces continuities and discontinuities in representations of Southwest Angolan dolls over time by examining artefacts in museum collections and historical ethnographic descriptions in parallel with current local crafting.1 I thus follow an artefact-based approach (Henare 2003), combined with a broader stance similar to Steven Rubenstein’s (2007) contemporary approach to Shuar shrunken heads and their local significances. This leads me to focus on children’s voices and interpret the notion of make-believe as a way in which children imaginatively engage with broader livelihood concerns. As I will show, play and labour are intimately connected in this rural context.

In Omphunda, villagers’ livelihoods mainly depended on activities linked to changing seasonal demands, such as rain-fed agriculture, cattle raising, and charcoal making. On my arrival, the existing members of Kabuka’s farmstead consisted of her parents, a twenty-year-old sister in an advanced stage of pregnancy, a fifteen-year-old brother, and a toddler sister who still demanded a great deal of attention from their mother. This household composition made Kabuka’s collaborative participation in the domestic economy both accepted and expected. By then, Kabuka’s life experience had provided her with a certain amount of manual skills and autonomy, and she attentively followed and participated in many of the domestic tasks that her experience allowed her to. At such an age, children’s cooperative stand was a highly salient feature in the household dynamics, a feature shared by the other households in the village and mediated by adults according to each child’s developmental phase.

One of Kabuka’s favourite tasks was to tend cattle. In fact, the rare situations when I saw her having a tantrum were related to the few occasions on which an adult would not let her go with Kaundende, her older brother, and Basukuka, a neighbour who was about Kabuka’s age, on their nearby herding chores. However, during these rare tantrums, she managed to get her way in the end. Kabuka did not need to help with taking care of cattle, as there were boys available to do it. Rather, it was her personal will and her pleasure to do so. Helping in that daily task and her sporadic corncob-doll play are practices that combine different dimensions of labour and play in this rural context. Their evocation allows me to point out coexisting dimensions in everyday life between adults and children, the nature of which varies according to age relations. Previous approaches to handcrafted dolls in this context fail to consider how both adults and children pursue practices demanding skill development, as well as how seasonal and daily activities combine to inform doll making and doll playing practices.

In this article, I use the notion of engagement through a focus on play to explore the movement from villages where dolls are made to museums where dolls are displayed. With that movement, I situate how different people engage with dolls – previous researchers, children, and myself. I first discuss two 1930s ethnographic vignettes of corncob-doll play in rural Southwest Angola, which to my knowledge are the earliest descriptions of these dolls in the hands of young girls. Considering descriptions of such situated practice in the region arising from extensive field experiences, I underline the similarity of these vignettes and look at how these scholars conceptualised doll play in relation to ritual. In contrast to previous ethnographic accounts, I aim to explore an alternative interpretation of young children’s doll play. Rather than constraining doll play to children’s development (Lang and Tastevin 1938) or to ritual (Hambly 1934), I develop a discussion connecting play in rural Southwest Angola to a domestic mode of agro-pastoral production influenced by seasonal change. Approaching daily work and ordinary play under such circumstances, I suggest that this alternative interpretation explores what doll play says about the context in which it takes place. Thus, I explore children’s engagement with doll play to show how they socialise in their day-to-day life and imaginatively improvise toys from their surrounding materials, particularly at the age in which they occasionally decide to turn corncobs into temporary dolls.

Defining the Fieldsite and Reflecting on Methodology

Omphunda was an agro-pastoralist highland village where the majority of inhabitants self-identified as Ovamwila, a Bantu ethnic grouping fitting under the umbrella of the Ovanyaneka peoples, while also sharing cultural features with neighbouring Ovahelelo pastoralists. Omphunda had one permanent waterhole and a few transient ones situated on different mountain slopes. Its twenty-two occupied farmsteads were spread throughout five different areas along the mountainsides with seasonally dry rivers. The village population consisted of approximately 100 people, distributed along what demographers call an expanding pyramid; though the number of adults and children was similar, there were many more babies and toddlers than mature and elderly people.

During my eight months of fieldwork in Omphunda, I lived in the farmstead of Madu and her husband Pedro, Kabuka’s parents. I first visited Omphunda during two useful one-week stays at the community-built wooden school accompanying an Italian NGO trainee surveying the village with the local teacher as translator.2 I requested to stay there proposing to teach youngsters and adults to read and write their spoken Olunyaneka language at the school while learning to speak it myself. My teaching role lasted about five months, at which time the villagers started to feel the implications and pressure of the dry season on their domestic economy. During my stay, my social role included that of a teacher, but also of a visitor, and a researcher of the villagers’ language and practices, associated with my belonging to my host household.

After an initial short period in which we got to know each other, the children never seemed to consider my activities, whether related or not to local practices, uninteresting. The children either attentively observed me or enthusiastically collaborated in whatever I was doing. At the level of my host household, what I had to negotiate was the perception of my social role in relation to participation in the adults’ world while being able to observe children’s interactions among themselves and with adults, including myself. In one of the earliest disciplinary reflexive discussions about cross-cultural socialisation in informal contexts, anthropologist Jean Briggs (1978: 2008; 2008: 450) argues that during her fieldwork among the Utkuhikhalingmiu, an Inuit people in Northwest Canada, she was childlike. Upon her arrival, she neither knew the language nor the expected social behaviours for a local woman of her age, nor had the skills that an Inuit woman would have (Walton 1993: 388). My arrival in Omphunda shared similarities with Briggs’ case. From the beginning of my stay, my host Madu, for instance, clearly considered me culturally and linguistically incompetent in many ways.3 Moreover, I shared with Lipuleni, Madu’s toddler, a lack of skilful practice and formal language. But in contrast to Briggs’ case, neither Madu nor the other household members ever considered me a child or treated me as one. Instead, my hosts considered me as an adult who could perform certain social roles, such as child-minding, but who was unskilled in relation to certain tasks. For example, Madu’s initial assumption that I lacked local cultural competency in many dimensions, contrasts with how she expected me to care for her youngest children, actively promoting this from my arrival.

One of the outcomes of frequently caring for Madu’s youngest children was the opportunity to socialise extensively with them. Throughout my stay, both two-year-old Lipuleni and five-year-old Kabuka were great teachers of the local language, namely through the tasks we carried out and the games they played by themselves, with other children, and with me. Even if my hosts did not consider me a child (ol. omona) in our shared everyday life, for adults I progressively evolved from omuhikwena (ol. young woman) to omukai (ol. woman), according to their perception of my evolving skills in certain local practices. From the perspective of children, however, my progress from omuhikwena to omukai was much less relevant. Both of those roles implied often being already a more experienced practitioner of adults’ daily tasks than them. From early on, small children such as Kabuka or her toddler sister, perceived me as a potential role model for the local skills they saw me perform.

The first time I wanted to go to the waterhole by myself, afraid that I might lose my way, Madu requested Kabuka go with me. Paying respect to her mother, but also to adults in general, she did it out of choice, leading the way and not turning back once. Subsequently seeing me preparing to go to the waterhole, she often requested to accompany me. On one of our first journeys to the waterhole, Kabuka turned to me and looked at how I had made my temporary head ring from an extra piece of cloth in order to carry water, a practice I learned from locals. After attempting to do one for herself, and failing twice, she took mine as a valuable model. This episode also shows how we related to each other in meaningful exchanges from early on. Carrying out and mastering daily tasks during my stay played a relevant role in my research because, from the perspective of the household, including both adults and children, we all shared the same principle of contributing to the success of their subsistence domestic economy.

My last methodological reflection concerns the use of visual methods. My aim of producing photographs of everyday life in my fieldsite was linked to the scarcity of images showing dolls in the hands of local people; more often, drawings or studio photographs reproducing only the dolls are available – an historical legacy of their museum existence. In creating contemporary images of dolls, I wanted to produce intimate portraits based on the everyday life of the people collaborating in my research. These pictures came into being through the combination of our shared day-to-day life while informally supervising the young children of my host household. I am aware of a certain irony in freezing into permanent visibility events that I aim to characterise as spontaneous and volatile. I filmed Kabuka’s sporadic play with corncobs, but here I combine it with animated sequences of photographs to invoke the sense of Kabuka’s engagement with her dolls in her day-to-day life (see video 1

Corncobs-Turned-Into-Dolls: The (Un)Sacredness of Play

Anthropologist Wilfrid Hambly (1934) describes a doll made from a corncob that he gathered during his year-long fieldwork in Angola for the Field Museum of Natural History in Chicago, in a chapter dedicated to education modes practiced by Ovimbundu people in the Angolan Central plateau:

There is but one example of children’s dolls in Field Museum’s collection. This was obtained from a little Ocimbundu girl of Elende.4 The doll is made from a corncob which is draped in blue cloth of European origin such as women wear. My interpreter said that the dressing of dolls made from corncobs is a general custom at the time of cutting the corn. In view of the frequent occurrence of agricultural rites among Bantu Negroes, the use of dolls led me to inquiries into the possibility of there being some kind of fertility cult associated with a corn-mother concept, but my inquiries met with negative results (1934: 219).

The above description is the only specific remark about girls’ play in Hambly’s comprehensive ethnography, and appears in sharp contrast with the following pages about boys’ games and play, or with his previous detailed description of music making accompanied by singing and dancing as a playful activity.

Missionary Alphonse Lang’s 1930s study of the Ovanyaneka people also includes a vignette on girls’ use of dolls made from corncobs. Lang shares with Hambly a preponderance of interest in describing dance and music, but his observations on play are much shorter. In a descriptive chapter on what Lang calls ‘the normal evolution of a person’s life’, the section dedicated to ‘leisure and play’ mentions:

[A] corncob dressed as the [girl] serves the purpose of a doll. Depending on the seasons, the available fibres or the corn beards configure the doll’s hair. This doll is carried on the back as if carrying a baby. Among their group of companions, they lend them and receive them back with a small gift. Whoever picks up the doll in her hands, even just to look at it, should return it with a gift (Lang and Tastevin 1938: 23; my translation from the original French).

For Lang, ritual is the central dimension that constructs the worldview of Ovanyaneka people, while for Hambly ritual is the dimension through which he measures Ovimbundu people’s worldview against. This can be seen regarding Lang and Hambly’s interpretations of doll-making practices for young children’s socialisation, even if this is where the ethnographers’ data appear to diverge. Lang describes a particular game in which corncob dolls become precious toys worth a gift in return, thus ritualising the game while universalising it into a general Ovanyaneka custom. To a certain extent, ritualisation appears as a way to legitimise his interest in the children’s practice, while generalising it implies that all (or many) Ovanyaneka children play the same game, the same way. Hambly also attempts to find a ritual reason for the crafting of corncob dolls, placing this activity as relevant in young Ovimbundu children’s socialisation but not delineating it any further. Though with different results, both these vignettes assess doll-making as a socialisation activity of young children and aim to link play with ritual.

In contrast to these earlier ethnographic accounts of corncob dolls, present-day anthropologist Rosa Melo did not focus on doll play precisely because she did not view it as relevant to her research on efiko (2001), a constellation of ceremonies and rituals celebrating girls’ passage to womanhood that is marked by various ethnicities throughout rural Southwest Angola.5 In late 2011, I met Melo in Angola. At my request, Melo told me about a recent ethnographic fieldwork encounter with corncobs turned into dolls by young children. Melo has an Ovahanda background, a Bantu ethnic identification that scholars commonly include in the broader designation of Ovanyaneka-Ovankhumbi peoples, and her research among rural and urban Ovahanda is part of the recent scholarship on contemporary Southwest Angola.6 She told me how she had come across young Ovahanda girls making dolls from corncobs in her rural research setting in the north of Huila province, while studying efiko. This struck her as just children playing with dolls, and she did not give them much research attention. Unlike Lang, the children’s ritualised games with dolls that Melo encountered did not grab her attention; efiko did. Contrary to Hambly, Melo was not discouraged because the children’s play with dolls was not ‘ritualistic’; rather, it was because it was children’s play that she did not pay much attention to it, concerned as she was with other rituals rather than with play. In her first monograph about the Ovahanda (2007), she addresses this activity in a field photograph showing small girls playing with dolls made from corncobs. Describing the photograph, she indicates that ‘while playing with dolls, children take care of each other’ (2007: vii; my translation from the original Portuguese). Years later, she regretted not being able to take a better photograph of her encounter with the girls playing.

The ethnic and livelihood differences between Ovimbundu agriculturalists and Ovanyaneka agro-pastoralists in the previous historical descriptions, or of contemporary Ovahanda, are consistent with the diversity and porosity of Southwest Angola. Thus, rather than paying attention to such differences, I underline here the similarities between Hambly and Lang’s narratives. They both outline how dolls are made out of the same resources (a corncob dressed with fabric, to which further accessories can be added), their seasonal existence (in accord with the corn harvesting season), and their playful nature. Through the lens of ritual, both ethnographers attempt to point out the interconnection between the seasonally available materials and the social world, while Melo’s social explanation of the girls’ doll play attempts to locate a more complex understanding of ritual’s social role. Yet our encounter underlined another dimension besides ritual overshadowing an ethnographic understanding of doll play. It was actually Melo’s explicit research agenda on ritual, in contrast with the more tenuous approaches of previous ethnographers, that showed me that something else was undermining a fuller understanding of the social and material dimensions of children’s toys and play in this context. Previous studies have been characterised by a politics of disengagement in relation to young children’s activities.

In our conversation, trying to provide me with more information about doll play, Melo regretted her lack of attention to the practice. She had not asked the girls what they were doing, and assumed that they were ‘just playing’. The conversation led her to consider her own disengagement from doll play and led me to think about the lack of children’s voices with regards to doll play, and play with toys more generally. Curator Elisabeth Cameron (1996: 19, n.1) expresses the same regret of not seeking out children’s views during her fieldwork in Zambia when she was trying to enlarge the view on African dolls to account for play as well as ritual.

In recent decades, the issue of children’s voices has been much debated within the anthropology of childhood (James 2007; Spyrou 2011). The common ground in this discussion acknowledges the importance of child-centred research when focusing on children’s practices. Anthropologist Camilla Morelli (2014), for instance, argues the need to look at the imaginative dimension of children’s lives, or their ‘world-making’, in order to understand and value children as autonomous subjects and to place them in the social world that surrounds them. Within this child-centred approach, it is relevant to combine the adult ethnographer’s engagement with children’s practices and perspectives. I follow this approach below, engaging with the social world of small children in a rural context in Southwest Angola where play is connected to the domestic mode of production in which they are immersed. I aim to suggest that the world of work and play are interconnected and provide fertile dimensions young children use to act imaginatively.

I also add, however, a word of caution in carrying out this child-centred approach. David Lancy (2015) suggests that a cross-cultural perspective on childhood would benefit from taking into account local understandings of children’s play in contrast to conceptions of play in ‘dominant cultures’, defined as Western, educated, industrialised, rich, and democratic (WEIRD), which while being a minority, are the subject of many studies. As Lancy underlines, given that different cultural conceptions of children’s development are a decisive variable in the evaluation of the social roles of play, it is relevant to consider the position of local adults who are usually not active in engaging with children’s play. Instead, they often observe child’s play from the periphery, occasionally interfering, and rarely participating. Similarly, as I have described above, adult researchers concerned with play or dolls in rural Southwest Angola have maintained a peripheral engagement with children’s activities, centred on observation rather than participation. What can an anthropologist actively engaging with children’s play in rural Southwest Angola today add to this debate? Drawing on the connective flux between children’s and adults’ worlds (see figs. 1 and 2), I sketch an answer to this question below. Fig 2

Fig 1

Fig. 1 + Fig. 2. Youngster Kaundende brings to his younger siblings and peers a mid-afternoon snack, making Kabuka and Basukuka take a break during playing with dolls and houses. These photographs were taken during the rainy season, March 2012.

Engaging with Everyday Life: Play and Labour

Originally, my concern with social uses of handcrafted dolls led me to prioritise engaging with Madu – Kabuka’s mother. This decision to focus on women and older girls was grounded in the dolls present in existing museum collections and the available literature citing regional ethnographic examples of dolls. Such collections contain more dolls made of logs than corncobs or other locally available materials, and these log dolls often demand greater crafting skills than Kabuka, for instance, had. Similarly, the sparse existing literature consisting of publications based on historical ethnographies and exhibition catalogues, largely refers to dolls made of logs. My fieldwork in Omphunda, however, initially led me to focus on dolls made of corncobs crafted by much younger individuals – Kabuka turned corncobs into dolls soon after my arrival. Moreover, during the beginning of my fieldwork, Madu’s oldest daughter delivered a premature baby girl that led her to pursue traditional treatment in another village. This event had several implications for the labour dynamics of the household, which I will now discuss.

The division of labour around childrearing in Omphunda fits a description by Lancy, who states, ‘[w]hoever can most readily be spared from the most important tasks will take care of the child[ren]’ (2008: 123). In Omphunda, adults were busy during much of the day, and their tasks usually consisted of a combination of domestic chores and others tasks that required going into surrounding areas in the bush. During my fieldwork, I realised that I often accepted the informal role of minding Madu’s small children and ultimately spent a large amount of time on the farm with them. When both Madu and I planned to carry out different tasks outside of the farmstead, Madu would often ask the children – if they themselves did not ask to do so first –, if they wanted to go with me on my elected mission, such as firewood collecting, fetching water, or visiting nearby neighbours. Moreover, if I decided to go with Madu on her task, the children would often join us. It was through situations such as these that I came to understand Madu’s rationale for taking advantage of my presence to try to involve the children in less onerous tasks while ensuring they were responsibly supervised.

Further, there was the issue of timing, linked to anthropologist Sylvia Yanagisako’s conceptualisation of the ‘development cycle of domestic groups’ (1979; 2001). Laboriously crafted dolls, such as the ones made from logs, demand advanced skills, including a mastery of basketry that young children do not have. For instance, Kabuka’s attempts in basketry were limited to occasionally practicing sewing a small base for a basket. Moreover, these log dolls are also rare because, beyond requiring advanced skills and skilful practitioners, they demand other valuable things requiring long-term planning. These include time to gather external resources such as beads and other accessories to make a beautiful doll, as well as the time to craft them into a doll. These constraints also add value to the doll: to find the necessary time and resources to make such a skilfully demanding doll results in people’s reluctance to give it away to strangers. Here I should also note that a child showing the desire to have an elaborate log doll is often not the only prerequisite for the making of this kind of a doll. For an adult to attend to such a request, the child has to have achieved greater social participation than that of a child of Kabuka’s age – the child has to be able to skilfully collaborate in the making of the doll and to give adequate care to an object which demands significant labour to produce. The importance of these two factors meant that adults rarely crafted toys for young children. For toddlers, adults often provided something simply crafted from improvised materials. As such, for children of Kabuka’s age, growing up meant that they were able to make their own toys using their own evolving skills with available devices and materials they found.

I started this article by framing how Kabuka engaged in doll play in relation to her participation in the productive endeavours of the household. Here it is relevant to return to how Kabuka’s everyday life combined play, whether alone or with other children, and collaborative participation in the household economy – a form of participation that children often associated with entertainment. Lancy (2012) underlines how anthropologists’ attention to children’s play appears quite unbalanced when compared with attention to children’s labour: researchers more often engage with their play than with their labour. He suggests that in contexts in which the productive role of children is significant for the household’s domestic economy, mandatory and optional tasks on the one hand, and informal learning on the other, define the ‘chore curriculum’. In comparison to Lancy’s overview in which chores play a major role in gender differentiation (2012: 33), Kabuka’s chore curriculum was defined by a lack of gender division that made her an active helper in herding rather than cooking, for instance (her older brother could also cook when their mother was away, or help her cook when she was there). Even though adults had substantial tasks to carry out daily, which made children’s collaboration welcome, Kabuka’s young age allowed her to explore and increase her skills – what Lancy calls a ‘non-mandatory skill’ (2012: 24).

Children of Kabuka’s age helped daily with tasks such as milking the cattle (Fig. 3), directing the cow and calf herds into different pasture areas, giving water to both herds at different times, lending a helping hand in the garden, collecting water, firewood, and so on (Fig. 4). The tasks they collaborated in mostly took place at locales near the farmstead or on a couple of neighbourly farms. When adults were busy with their daily occupations, children would often be given a few tasks to do, such as carrying messages or goods to nearby neighbours. For Kabuka and her age mates, visits to more distanced locales, such as the waterhole, would often involve accompanying an adult or an older child in order to help perform a particular task such as washing or providing the calves with water. Nonetheless, beyond requested activities and willing participation in social and productive tasks, children often spent long periods on the farm engaging in activities they determined for themselves.

Fig. 3. Kabuka and Kaundende attending cattle in the domestic cowshed. February 2012.

Fig 3 Fig 4

Fig. 4. Mixing labour with play, as well as different degrees of practice and expertise, all the children weed the farmstead during the rainy season. Late March 2012.

It was during these moments of leisure in that year’s short rainy season that Kabuka occasionally turned corncobs into dolls. It is relevant to underline that Kabuka engaged in several forms of play during the eight months we lived together, which took place in a variety of spaces around the farmstead area (see Map 1). While instructive, note that the map does not show the impact of seasonal differences on Kabuka’s activities.

Map 1

Map 1. Spatial sketch of the farmstead at my time of arrival (January 2012, author’s sketch). My host Pedro was still fencing a part of the farmstead at this time.

Keeping children’s practices at the centre of my analysis, in the following section I want to suggest that when considering the place of young children crafting dolls in rural Southwest Angola, it is useful to establish a relationship between toys and play. This relationship involves a range of materials and creative practices according to the development stage of the child in question. Though I am interested in underscoring young children’s independent play, given that the following examples occurred while I was the adult informally supervising their activities, I also include our mutual interactions in the analysis, specifically the effects of my photography.

Kabuka’s Dolls: Temporary Play, Available Materials, and Imagination

It was during the rainy season that I had the privilege of seeing Kabuka spontaneously turning corncobs into dolls. The long dry season that followed it explains why, with hindsight, I take this period as a privilege.7 At the time, I observed Kabuka attempting to dress some corncobs she had found around the farmstead in a fashion similar to the way people dressed locally, as well as playing games with these corncobs. During this period, in her free time, I saw Kabuka creating make-believe constructions similar to the buildings we occupied, either alone or with Basukuka, the neighbourhood boy of a similar age and her often inseparable peer. In various instances when Kabuka was playing by herself, I engaged with her and asked her what the corncobs were. She replied either ovana (ol. for children) or ovimphumpho (ol. for corncobs), as well as giving me more complex narratives regarding the role her corncobs-turned-into-dolls played in the particular game taking place during our interaction. While moving the dolls into the structure she had skilfully rebuilt several times (given the fact it often collapsed, an outcome that amused her greatly), she would tell me ‘the children are going home’. When the structure collapsed, she often called to me, emphasising ‘Ines, the house fell down!’. She entered and exited from imaginative play very easily, adjusting, reacting, and discovering other ways to engage with what was happening around her. Seeing me engaged in taking photographs of her, she would, sooner or later, interrupt her activity to come and see the result (see video 1

During the same time period in which European and North American ethnographers and missionaries were measuring African children’s activities in relation to ritual, philosopher and literary critic Walter Benjamin (1996a[1928]; 1996b[1928]) discussed the relationship between European children’s play and toys. From a discussion about technological change, in which Benjamin examines the passage from handcrafted to industrially produced toys for children in the German context, he emphasises the importance of attending to the full spectrum of toy-making by children, suggesting that anything can become a toy in the hands of children, and a toy can become many different things, that is, can acquire multiple meanings. Kabuka’s dual understanding of the corncobs as corncobs and as dolls made me realise what the available ethnographic literature appeared to lack: a fluid awareness of corncobs as material devices on the one hand, and make-believe on the other, dynamically linking the material and the social worlds.

Descriptions of the combined material and make-believe interpretations of young girls playing with corncobs are difficult to find in the published literature on rural Southwest Angola. Furthermore, other published field sources, such as the work of Carlos Medeiros (1972), consist of ambiguous data focused on the interplay between toys and ritual. Medeiros’ work concerns information gathered before Angolan independence while he was an anthropology student, and some of it is available only as internal museum documentation, namely through the accession files of the dolls he collected (1972AM). While Medeiros’ unpublished information is ambiguous, it is revealing of children’s voices regarding the practice of turning corncobs into dolls.

In the early 1970s, during 14 months of fieldwork, Medeiros collected materials on dolls in the rural northern part of Namibe province (Medeiros 1984: 80).8 Afterwards, he donated a set of seven dolls, some of them made out of corncobs and others made from logs, to the National Museum of Ethnology (MNE) in Lisbon, Portugal.9 As I mentioned earlier, in Southwest Angola log dolls tend to be more skilfully crafted than dolls made of corncobs.10 Dolls made of logs, termed oxilolo (ol. for doll) in his ethnography (1972), are briefly mentioned in a section dedicated to play when Medeiros mentions the case of a young woman using one doll as a fertility amulet. It is only in internal museum documentation that one may find more information about two of the corncob dolls he gathered from small girls. In their accession file, he describes them as toys and mentions the word ovatita as the local name of those dolls. Without explaining its meaning, Medeiros briefly describes a game he had observed being played by their owners: ‘these dolls were found in a make-believe cowshed, and the little girls said that they [the dolls] were their children’ (Medeiros, 1972AM, my translation from the original Portuguese).

During my fieldwork in Omphunda, Kabuka’s father Pedro told me that ovatita is the plural for shepherd in the local language, Olunyaneka. This small detail helped me widen Medeiros’ description. While playing in a make-believe cowshed, the Ovakwandu girls from whom he had gathered the dolls during his stay in an agro-pastoralist mountain village of Southwest Angola were playing with make-believe shepherds, thus naming them as such. This made it possible to look at Medeiro’s description as one of a game, which in fact implies two children’s games: one in which the girls were playing in a make-believe cowshed with their make-believe shepherds, and another game in which the girls were playing with their ‘make-believe’ children. Or was it a game in which their ‘make-believe’ children happened to be shepherds? Adding my own local knowledge to the field-collected information about the girls playing with their ovatita children, increases the possible interpretations of the little girls’ game. It resembles the moments in which Kabuka replied to me that her corncobs were her children, rather than corncobs. What is more, we notice the appearance of the interplay between material and make-believe dimensions (Benjamin 1996a[1928]), even in the case of this set of dolls now housed in a museum. This underlines how material and make-believe descriptions and subsequent interpretations of young girls playing with corncobs-turned-into-dolls are difficult to find in published field descriptions about rural Southwest Angola, and difficult to interpret without further local knowledge in the case of this unpublished field source.

Following Benjamin’s ideas about the dynamic interplay between toys and play for children, it seems that the initial impulse for these games lay in Kabuka’s willingness to play with the available corncobs, turning them into dolls. Kabuka’s doll play is an excellent example of the interplay between their material and make-believe dimensions, between their literal and figurative features. This interplay also resonates with other games that Kabuka engaged with during the eight months we lived together, and I will now discuss a few examples of play involving improvisation with available materials and devices. Linked to transitory games and play, I underline the particular choices arising from Kabuka’s personality traits that are also dependent on the development stage of her peers. For this purpose, I add a few examples of other games during my fieldwork, which include events taking place in both seasons.

fIG 5

Fig. 5. Kabuka playing with a glass bottle. February 2012, during the rainy season.

Before observing Kabuka turning those corncobs into dolls, I had already seen both her and her toddler sister, Lipuleni, playing in a similar way with glass and plastic bottles, and even stones, sometimes putting them on their backs with the help of a shawl to emulate a baby or a doll (see Fig. 5). After playing with these found objects in that manner, these artefacts could easily be used for something else, or disposed of.11 Moreover, while Kabuka sometimes used her corncobs-turned-into-dolls to imitate a baby or a doll on her back, she preferred to play with them in association with make-believe constructions, as I have discussed above and shown in previous images (Fig. 5 and 6), and in the video.

Fig 6

Fig. 6. Kabuka playing with her shawl, using it as a dress for an older woman. Early rainy season, cloudy day with no rain. Early March 2012.

Both the spontaneity and volatility of children’s activity is perhaps the most important point to take from Benjamin’s reflections with regard to these examples. In another temporary game, Kabuka used her usual shawl to fashion herself a skirt as might be used by an older woman (Fig. 6). In another moment, very early in the morning while the adults were at work somewhere else, Kabuka and her toddler sister played for a while in our sleeping area, and in this photographed scene, Lipuleni acts as Kabuka’s child – or should I say as her temporary living doll (Fig. 7)? Though adults rarely intervened or showed interest in these activities, I should note that they prohibit children playing with some material sources, such as with water stored in the farmstead or with fire – the former due to the daily labour of carrying water from the waterhole and the latter for the children’s safety.

Fig 7

Fig. 7 Kabuka and Lipuleni playing mother and daughter, mid dry season. Late May 2012.

A final word about the use of my field images throughout the paper besides the examples of young children’s autonomous play I have provided. In the region, the rainy and dry season take many different shapes (see Fig. 8 and 9). During the time of my fieldwork, that year’s rainy season was the most relaxed period in terms of people managing to make a living in day-to-day life. Living the long dry season that followed made the easier nature of existence in the previous season obvious.

Fig 8

Fig. 8. Partial view of my host farmstead before the rainy season, February 2012.

Fig 9

Fig. 9. Same partial view taken during the rainy season, late March 2012.


My account of Kabuka’s appreciation of her corncobs-turned-into-dolls and other playing practices is based on the idea that children perceive toys in two ways, appreciating both their material and imaginative elements. I have shown that previous scholars privileging a concern with ritual has prevented them from grasping this double dimension. Specifically, earlier accounts engaging with this practice on the one hand, have not engaged with children’s voices or with children’s wider worlds, on the other hand. My analysis points to the need for an alternative engagement. I suggest that, in settings driven by a domestic mode of production such as the one in Omphunda, it is important to locate children’s doll-related activities in terms of the wider dynamics of play and labour – a perspective that takes as central the combination of material and social dimensions, and children’s and adults’ modes of socialisation.

While early ethnographic data comes from researchers whose work is readily available in English, literature and ethnographic data in Portuguese is still difficult to access. This is seen in museum exhibitions whose contemporary curatorial research mostly involves the use of historical data seen as valid in contemporary times. I have contrasted the ethnographic awareness I developed relating to children’s leisure activities with simple props with a scholarly discussion on the same topic. I have examined how field-collectors, missionaries, and ethnographers have framed simple toys, such as corncob dolls autonomously made by children, and children’s play in Southwest Angola. Children’s voices are not the only absence in earlier approaches – a wider understanding of toy making is also neglected. My approach allows for children’s creativity and volatility. Children play in many other contexts, yet here, through focusing on corncob doll play, I have discussed play in the domestic context only. Focusing on dolls and doll play, I have widened the perception of children’s engagement with the productive domestic economy – an interplay that might be helpful for both ethnographers of childhood and museum curators.

About the Author

Inês Ponte received her PhD in Social Anthropology with Visual Media at the University of Manchester. Her research focused on the social, material, and visual lives of handcrafted dolls from Southwest Angola by examining their local making and transnational collecting. In addition to the thesis ‘Crafted ‘Children’: an ethnography of making and collecting dolls in Southwest Angola’, she also made the film ‘Making a Living in the Dry Season’. Email:


An earlier version of this article was presented at the Royal Anthropological Institute Postgraduate Conference, Anthropology and Engagement, in Manchester in June 2015. I thank Jessica Symons for her useful and challenging feedback, as well as the anonymous reviewers and Anthropology Matters editors for their helpful and enriching comments. I am also grateful to Theodoros Kyriakides for all his efforts and support in the editing of this article, Gloria Dominguez for her fast and attentive feedback, Zoe Goodman for the meticulous copy-editing as well as Mary-Anne Decatur and Siobhan Magee for further helpful critical engagement with it. My research was made possible through the funding provided by the Portuguese Foundation for Science and Technology (FCT), grant SFRH/BD/69805/2010.


Benjamin, Walter. 1996a [1928]. ‘The Cultural History of Toys.’ In Selected Writings, 113–16. Cambridge, Mass: Belknap Press.

———. 1996b [1928]. ‘Toys and Play.’ In Selected Writings, 117–21. Cambridge, Mass: Belknap Press.

Briggs, Jean. 2008. ‘Daughter and Pawn: One Ethnographer’s Routes to Understanding Children’. Ethos 36, 449–456. https://doi.org/10.1111/j.1548-1352.2008.00026.x

———. 1978. Never in anger: portrait of an Eskimo family. Cambridge: Harvard University Press.

Cameron, Elisabeth. 1996. Isn’t S/He a Doll?: Play and Ritual in African Sculpture. Los Angeles, California: UCLA Fowler Museum of Cultural History.

Carvalho, Ruy Duarte de. 2008. ‘Sobre a ‘Composição étnica’ de Angola.’ In A Câmara, a Escrita e a Coisa Dita: Fitas, Textos e Palestras, 64–68. Lisbon: Livros Cotovia.

Estermann, Carlos. 1960. Etnografia do Sudoeste de Angola: o Grupo Étnico Nhaneca-Humbe. Vol. 2. Memórias. Série Antropológica e Etnológica. Lisbon: Junta de Investigações do Ultramar.

Hambly, Wilfrid. 1934. ‘The Ovimbundu of Angola.’ Anthropological Series, Field Museum of Natural History XXI (2).

Henare, Amiria. 2003. ‘Artefacts in Theory: Anthropology and Material Culture.’ Cambridge Anthropology 23: 54–66.

James, Allison. 2007. ‘Giving Voice to Children’s Voices: Practices and Problems, Pitfalls and Potentials.’ American Anthropologist 109 (2): 261–72. https://doi.org/10.1525/aa.2007.109.2.261

Jolles, Frank. 2010. African Dolls: The Dulger Collection. Stuttgard: Arnoldsche Art Publishers.

Lancy, David. 2008. The Anthropology of Childhood: Cherubs, Chattel, Changelings. Cambridge; New York: Cambridge University Press.

———. 2012. ‘The Chore Curriculum.’ In African Children at Work: Working and Learning in Growing Up for Life, edited by Gerd Spittler and M. F. C. Bourdillon, 24–57. LIT Verlag Münster.

———. 2015. ‘Mapping the Landscape of Children’s Play.’ In Handbook for the Study of Play, edited by James Johnson and Scott Eberle, 431–39. Lanham, MD: Rowan & Littlefield.

Lang, Alphonse, and C. Tastevin. 1938. Mission Rohan-Chabot: Tome V, Ethnographie, La Tribu Des Va-Nyaneka. Paris: Ministère de l’Instruction Publique, Société de Géographie.

Medeiros, Carlos. 1972. Vakwandu: Análise de Uma Vivência. Lisbon: ISCSP.

———. 1972AM, AV.839 Ficha de Inventário (inventory file), Museu Nacional de Etnologia (MNE) Archive, Lisbon, Portugal.

———. 1984. ‘Kwandu Law: The Evolution of a Juridical System Among an Herero People of South-West Angola.’ Journal of African Law 28 (1-2): 80–89. https://doi.org/10.1017/s002185530000526x

Melo, Rosa. 2008. Crenças, Poder e Práticas Medicinais entre os Handa. In Para lá da Manipulação dos Espiritos. Crenças e Práticas de Cura, 1–76. Dakar: CODESRIA.

———. 2007. Homem é Homem, Mulher é Sapo: Género e Identidade Entre os Handa no Sul de Angola. Lisbon: Colibri.

———. 2004. ‘´Nyaneka-Nkhumbi’: Uma Carapuça Que Não Serve aos Handa, Nem aos Nyaneka, Nem aos Nkhumbi.’ CEA 7–8 (May): 157–78. https://doi.org/10.4000/cea.1074

———. 2001. ‘Efuko: Ritual de Iniciação Feminina Entre os Handa (Angola).’ Unpublished PhD thesis, Lisbon: ISCTE_IUL.

Morelli, Camilla. 2014. ‘The River Echoes with Laughter.’ Unpublished PhD thesis, Manchester: University of Manchester.

Ponte, Inês. 2015. Dolls of Southwest Angola. Lisbon: INCM & MNE.

———. 2015a. ‘‘Crafted Children’: an Ethnography on Making and Collecting Dolls in Southwest Angola.’ Unpublished PhD thesis, University of Manchester.

———. 2012. Okalivulu okatutu kokuatesako okilongesa okutanga nokuhoneka melaka liolunyaneka: liapangyilua ovanthu veliongia mohika potchilongo tchivahiua pokati koNamibe noHuila. Namibe: COSPE & Direcção Provincial da Educação, Ciências e Tecnologia do Namibe (available on-line: www.academia.edu/3372157/2012, accessed 23 May 2014).

Rubenstein, Steven L. 2007. ‘Circulation, Accumulation, and the Power of Shuar Shrunken Heads.’ Cultural Anthropology 22: 357–99. https://doi.org/10.1525/can.2007.22.3.357

Spyrou, Spyros. 2011. ‘The Limits of Children’s Voices: From Authenticity to Critical, Reflexive Representation.’ Childhood 18 (2): 151–65. https://doi.org/10.1177/0907568210387834

Timm, Margo. 1998. ‘Kwanyama Ovambo: Kwanyama Child Figures.’ In Evocations of the Child: Fertility Figures of the Southern African Region, edited by Elizabeth Dell, 207–18. Cape Town, Pretoria & Johannesburg: Johannesburg Art Gallery, Human & Rousseau.

Walton, Susan. 1993. ‘Jean Briggs’s Never in Anger as an Ethnography of Experience’. Critique of Anthropology 13, 379–399. https://doi.org/10.1177/0308275x9301300406

Wolff, Norma H. 2004. ‘African Artisans and the Global Market: The Case of Ghanaian ‘Fertility Dolls.’ African Economic History 34: 123-141.

Yanagisako, Sylvia. 1979. ‘Family and Household: The Analysis of Domestic Groups.’ Annual Review of Anthropology 8: 161–205. https://doi.org/10.1146/annurev.an.08.100179.001113

———. 2001. ‘Household in Anthropology.’ Edited by Neil J. Smelser and Paul B. Bales. International Encyclopedia of the Social and Behavioral Sciences. Oxford, England: Elsevier Science.


1 I underline that by looking from an historical perspective, I will examine ethnographic descriptions of corncob-doll use by children that appear in comprehensive monographs about rural populations.

2 When the NGO approached the village in the previous year asking which project they wanted to develop in collaboration, the villagers requested help with setting up a school. The community thus built a wooden school, and the NGO offered the metal thatched roof and paid for a teacher to come to teach children the national fast-paced literacy programme delivered in Portuguese. This was in partnership with the province’s government.

3 Madu and the children’s mother tongue was Olunyaneka, the widely spoken language in the village, which I increasingly mastered throughout my stay, namely by teaching villagers to read and write it – see Ponte (2012). Pedro, Madu’s husband, was proficient in both Portuguese, my mother tongue, and in Olunyaneka.

4 To avoid any confusion with the designation of a hairstyle associated with Ovakwanyama people inhabiting the Kunene plains, I note that the name of Hambly’s fieldsite is the mountainous Elende Mission Station. I also add that in Umbundu, elende means cloud. Hambly’s stay in Angola was from February 1929 to February 1930 (cf. 1934: 103).

5 Melo was raised in Namibe city (2001: 6–9). Concerned with the issue of ethnic identification and the contemporary role of efiko, Melo addresses its role in contemporary rural and urban relations among the Ovahanda, as descendants of people who have migrated to urban areas return to their relatives in rural areas to celebrate this rite of passage. Efiko’s current importance is further underlined by Melo’s research on Catholic celebrations in urban areas. Regional and ethnic diversity has produced a range of terms to describe these ceremonies; common variants with no strict geographical boundaries are efiko, efuko, ehuko, or ehiko. In Omphunda, people designate efiko differently even within the same household, the variation relating to personal trajectories.

6 Melo (2004; 2008) contests merging Ovanyaneka and Ovankhumbi ethnicities into one hyphenised category. The villagers in my fieldsite recognise themselves as Ovamwila, thus with an Ovanyaneka agro-pastoralist background, while sharing cultural features with neighbouring Ovahelelo pastoralists. Ethnic categorisations in multi-ethnic Southwest Angola are a complex issue, suffering from added confusions emerging from Portuguese colonialism that are often still not settled in current scholarship. See Carvalho (2008) for an illuminating discussion regarding the overall Angolan context.

7 Later during the dry season, I also engaged with other children turning corncobs into dolls for play in a village in the neighbouring Huila province. In the latter context, families’ livelihood/economy was less dramatically affected by the year’s dry season, see Ponte (2015a).

8 Medeiros worked with self-identified Ovakwandu people, who differ from the Ovamwila who constituted the majority of my interlocutors. The two locales, my fieldsite and Medeiros’, are relatively close in geographical terms; his interlocutors were affected by both Bantu Ovahelelo and Ovamwila regional influences, seemingly in a similar combination to that which I observed in my village fieldsite. Medeiros also visited neighbouring areas, from which he brought back field experiences and material culture.

9 Medeiros’ collection, gathered in Southwest Angola during this period and later donated to the museum, consists of 115 artefacts. For a published report on Medeiros’ collected dolls see Ponte (2015).

10 Though some of the corncob dolls collected by Medeiros challenge this principle, the two corncobs that I discuss below do not. Moreover, in regard to the general information Medeiros provided about the dolls, he appears to follow a distinctive principle based on the different material they are made of, that is, logs or corncobs.

11 Again, my ethnographic data stands in sharp contrast with published regional ethnography, such as Carlos Estermann’s (1960: 59–60) widely cited ethnographic example of girls playing with small stones in which fertility is the interpretative framework. Estermann mentions the permanent condition of such found artefacts as a legacy passing over to the next generation. It is worth noticing that Estermann’s ethnographic work on Southwest Angola was translated into English and French during the 1970s and widely reviewed at that time. Estermann’s first ethnographic writings date from 1935, yet his substantial research on Southwest Angola was conducted up to the 1980s.

Anthropology Matters Journal ISSN: 17586453 Publisher: Anthropology Matters url: www.anthropologymatters.com