Same difference, different sameness: Gender-navigating the denominational maze in a Christian African context

Abstract

The large and growing denominational diversity of sub-Saharan Christianity has attracted considerable scholarly attention, but its implications for individual experiences and well-being remain poorly understood. I examine women’s perceptions, assessments, and practices in everyday construction of church idiosyncrasies, clustering, and hierarchies and women’s abilities and propensities to navigate along and across the corresponding organisational boundaries in a predominantly Christian setting in rural Mozambique. The analysis uses a combination of data from a census of religious congregations, a household survey, in-depth interviews, and participant observation conducted over a decade. It demonstrates that women’s church involvement is socially indispensable yet also subjectively non-binding. Accordingly, church boundaries, while clearly delineated and recognised, are also seen as highly permeable. This permeability facilitates the crossing of those boundaries but may also disincentivise such crossing. Although rural women are increasingly capable of navigating through the religious space, they are generally precluded from exiting that space altogether, as secular options for non-family social anchoring remain very limited in this and other similar low-income patriarchal contexts.

Introduction

In this study, I seek to contribute to cross-national research on gendered religious identities, belonging, and agency (e.g., Avishai 2008; 2016; Bartkowski and Read 2003; Burke 2012; Mwaura 2005; Olajubu 2003; Soothill 2007) by examining women’s religious affiliation and involvement in a predominantly rural Christian patrilineal sub-Saharan context, where considerable denominational diversity has been shaped by enduring colonial legacies and dramatic post-colonial transformations. Engaging with the scholarship on the religious tradition vs. modernity dichotomy in sub-Saharan Africa (Gifford 2016; Oduyoye and Kanyoro 1992; Wariboko and Afolayan 2020), and its strong emphasis on the denominational diversification of sub-Saharan Christianity as part of its ‘modernisation’, I approach this dichotomy through the prism of neoliberal societal transitions characterised by increased inequalities and precarities (Gifford 2016; Newell 2007; Pfeiffer et al. 2007).

Africa’s burgeoning and seemingly diverse Christian religious industry is a vital socio-cultural component of this neoliberal reality, and its functioning and implications are distinctly gendered. Church involvement promises women unique psychosocial cushioning against the multiple and severe challenges they face; yet this cushioning is generally ineffective in practice, and in fact, it may further exacerbate these challenges. Its centrality, and its contradictory implications, are largely determined by the scarcity of viable and sustainable secular complements and alternatives to religious involvement. Women’s quest for betterment is therefore limited to choosing within the range of available religious options. These aspirational options in the neoliberal religious market are numerous and nominally diverse; yet they are also subjectively similar, as their distinctiveness is usually defined through their comparable unachievability.

In this study, I examine the construction and contestation of cross-denominational and intra-denominational distinctions and similarities as part of women’s everyday religious experiences that are conditioned and constrained by gendered norms and ideologies in this post-colonial context, where traditional patriarchalism is fused into and enhances the neoliberal social order. I employ a combination of rich quantitative, in-depth interview, and participant observation data collected over almost a decade in a largely rural area in southern Mozambique.

Background and approach

In sub-Saharan Africa, as in many other parts of the world, religious involvement has been a socially and psychologically central aspect of women’s life (Oduyoye and Kanyoro 1992; Olajubu 2003). However, as cross-national scholarship shows, women’s religiosity and religious engagement is shaped by a broader societal context (Baker and Whitehead 2016; Schnabel 2018; Trzebiatowska and Bruce 2012; Walter and Davie 1998). Western-focused religious scholarship has devoted much attention to the traditionalism vs. modernism debate and its implications for various dimensions of individual views, preferences, and behaviours, including politics (Hoffmann and Miller 1997), gender ideologies and identities (Ruiz et al., 2017; Schnabel 2016), family and romantic relationships (Gallagher and Smith 1999; Irby 2014), among others. Engaging with this debate, research on Christianity in sub-Saharan Africa has paid considerable heed to ‘new’ and ‘modern’, typically Pentecostal-type churches, and their roles in transforming the religious and broader societal landscape, including in the sphere of gender relations (e.g., Manglos 2010; Soothill 2007; Van De Kamp 2016; Van Dijk 2003; Wariboko and Afolayan 2020). However, the modernisationist potential and impact of the ‘new’ churches in post-colonial neoliberal African contexts have been questioned. In fact, as Gifford (2016) argued, neo-Pentecostalist promises of prosperity are detached from, and even contradict, the premises and reality of modernity. ‘New’ churches build their status on the alleged genuineness of their connection with ‘authentic’ Christian teachings yet also on their presumably high positioning on the ladder of modernity. Hence, a ‘true’ church is packaged as one that best adheres to the ‘essential’ Christian doctrinal tradition (however nebulously articulated) but also one that is most ‘up-to-date’ in its organisational style and performative practice. This apparent conflict is reconciled by contrasting the Christian tradition, old and modern, with pre-Christian (ancestors’ spirits-based), i.e., truly old, ‘obscurantist’ beliefs and practices. ‘New’ churches’ promise of miracle healing of both physical and psychosocial ailments is seen as their key feature and a reason for their success (Blanes 2017; Manglos and Trinitapoli 2011; Newell 2007). Yet, even in the seemingly prominent area of healing, the traditional-modern scale and corresponding church categorisations and distinctions are not straightforward. Church-based healing parallels and even interconnects with pre-Christian healing practices as both are engaged as spiritual resources for dealing with insecurities and precarities of the neoliberal socioeconomic reality (Pfeiffer et al. 2007; Newell 2007; Blanes 2017; Schuetze 2023), thus further illustrating the fluidity and conditionalities of religious ‘modernity’ in today’s sub-Sahara.

Women’s agency, and its ramifications, expressions, and limitations, are central to my conceptualisation. In dialogue with feminist scholarship on gendered doing and redoing religion (Avishai 2008; 2016; Darwin 2016; Schnabel 2018), I look at how women produce, reproduce, and exercise their religious self in everyday life. I argue that although women’s agency in religious membership choices is formally limited, especially in patriarchal contexts where such choices are usually made by men, it is nonetheless practically important and tangibly impactful for women’s quotidian religious experiences. Yet, as vast cross-national and cross-denominational evidence suggests, women’s religious organisational and social involvement may be both empowering and oppressive (e.g., Agadjanian 2024; Avishai 2008; Bawa 2019; Burke 2012; Darwin 2018; Ebaugh and Chafetz 1999; Maddox 2013; Sullivan and Delaney 2017). Similarly, women’s rise to church leadership roles may not necessarily challenge gender inequities and women’s dependencies (e.g., Agadjanian 2015; Mwaura 2005; Soothill 2007). While women’s church participation and roles may vary across denominations, there is no convincing evidence in the extant scholarship that membership in ‘new’ churches is associated with greater agency, autonomy, and decision-making power, compared to membership in ‘old’ churches (e.g., Agadjanian and Yabiku 2015; Gifford 2016; Pfeiffer et al. 2007; Soothill 2010). More generally, this seemingly conflicting evidence illustrates how women’s rising agency is intertwined with neo-patriarchalism in contemporary Africa.

Guided by this conceptual vision and empirical evidence, I use a combination of institutional-level and individual-level quantitative, in-depth interview, and participant observation data collected over almost a decade in a largely rural predominantly Christian and denominationally diverse setting in Mozambique to analyse women’s religious experiences across different denominations and churches. I start by examining women’s everyday involvement and engagement with their religious communities. I then look at women’s cross-church exposure through their kin and other personal network partners and reflect on how this exposure may contribute to the subjective construction of inter-denominational distinctions and compatibilities. Next, I consider church rules and regulations, such as dietary and vestimentary prescriptions and proscriptions, which are organisationally salient, yet barely consequential subjectively. Because of the centrality of marriage to women’s status and well-being in patriarchal contexts, I also focus on church norms pertaining to marriage building and maintenance, stressing both their regulatory importance and their practical unenforceability across the denominational spectrum. Given the mentioned prominence of church-based miracle healing, I show that such healing, while distinctively shaping church organisational image, has little bearing on individual religious identity and self-location. I then examine how women navigate these conspicuous yet permeable church boundaries.

As elsewhere in Christian sub-Sahara, church switching is common in the study site, but many of women’s cross-church transitions are normatively determined, such as conversion to the husband’s church upon marriage. Voluntary, woman-initiated church exiting is also frequent, yet leaving a church is almost inevitably followed by entry into another church, even if after a brief churchless hiatus to lessen the social and psychological strains associated with that transition (Agadjanian 2017). I argue, however, that religious conversion, whether normatively prescribed or personally motivated, while an expressively salient act, is only modestly transformative experientially. Such relatively unimpactful consequences may facilitate church switching but may also discourage it.

Setting

The data were collected in the Chibuto district of Gaza province in the Republic of Mozambique, an African nation of 32 million inhabitants with a GNI per capita of 480 USD (World Bank 2023). The district, with a current population of 220,000, is largely rural but also includes the town of Chibuto, which is the district’s administrative centre. Local society is traditionally patrilineal and highly patriarchal. The economy of the district, including much of its town segment, relies heavily on subsistence agriculture, with most farm work performed by women, as well as on male labour out-migration, mainly to the neighbouring Republic of South Africa. As much of the sub-continent, the study setting is characterised by considerable and aggravating economic precarities and social uncertainties.

The study area is overwhelmingly Christian, with considerable and growing denominational diversity. The Roman Catholic Church was the quasi-official church in the Portuguese colonial empire of which Mozambique was part until its independence in 1975. The Catholic Church lost much of it colonial-era clout, especially during the early, ‘socialist’ phase of Mozambique’s independent history, but has nonetheless remained a prominent player in the local religious market (Morier-Genoud 2023). The colonial era also saw the arrival and spread of Protestant missions, such as Presbyterian, Methodist (of different varieties), Nazarene, Anglican, Baptist, Reformed, among others (Cruz e Silva 2001). These mission-based, or ‘mainline’ Protestant churches, highly active in the colonial era, continue to have a noticeable presence in the country, including in the study site. Also, beginning in the late colonial era, Mozambique saw a large-scale proliferation of African Initiated Churches (AICs) of Evangelical and Pentecostal types that first arrived in Mozambique from South Africa. Zionist churches constitute by far the largest group among AICs. Another category of AICs is made up of Apostolic churches, which, in contrast to the ideologically and organisationally amorphous Zionist churches, typically have a very rigid and purist ideology and insular organisational structure. Finally, Mozambique, like much of sub-Saharan Africa, has also seen an expansive growth of neo-Pentecostal churches, mainly originating from outside the sub-continent (Pfeiffer et al. 2007; Schuetze 2023; Van De Kamp 2016). Yet, their presence in rural areas has remained relatively limited due to rural residents’ low capacity to meet these churches’ expectations of financial contributions. This denominational expansion and multiplication was catalysed by the abandonment of the early socialist (and largely secularist) ‘orientation’ of the Mozambican government in favour of neoliberal ideologies and policies since the late 1980s. Faced with the rapid proliferation of new churches, the local authorities, as well as the central government, have sought to regulate this burgeoning religious industry by specifying and enforcing registration requirements; yet, despite these efforts, at least some, smaller churches often function informally or semi-formally.

Women’s religious involvement in Chibuto, as in other parts of Christian Africa, is very common. Although as in other similar patriarchal contexts, women’s church membership is, at least nominally, determined by their husband or another male household head, women greatly outnumber men among church active participants (Agadjanian 2024), reflecting the universal gender differences in Christian religiosity and religious engagement (see Pew 2016; Trzebiatowska and Bruce 2012). Women’s large presence and active involvement in church life increasingly translate into women’s ascent to congregational leadership roles (see Agadjanian 2015).

Data and method

I use a combination of diverse and rich quantitative and qualitative data collected in Chibuto over almost a decade. The data collection was approved by the Institutional Review Boards of Arizona State University, the University of Kansas, and the University of California – Los Angeles. The quantitative data include two components: the Census of Religious Congregations and the Women’s Religious Survey, both conducted in 2008. The Census consisted of interviews with leaders of all the religious congregations registered at the District Commission on Religious Affairs (which, admittedly, may have excluded a few informal or semi-formal churches and congregations). In all, the Census covered 1,122 Christian and 3 Islamic congregations (the latter are not considered in this analysis), which, given the population of the district at the time, means approximately one congregation per 130–150 adult residents, on average.

The Women’s Religious Survey was a district-representative household-based survey that collected diverse information on religious membership and involvement, as well various other socio-demographic, economic, and cultural characteristics. The survey used a multi-stage sampling design: first, 82 residential clusters were randomly chosen within the rural and town parts of the district. Then, a sample of households was randomly selected within each cluster. Finally, one woman aged 18–50 was randomly chosen in each household and interviewed using a standardised questionnaire. The response rate was nearly 100%, resulting in 2,019 completed interviews. After excluding Muslim respondents (n=11) and respondents reporting no current religious affiliation (11.6% of the sample), the analytic sample consists of 1,775 women.

The qualitative data originate from 43 in-depth interviews, conducted in Changana, the main local language, or Portuguese, Mozambique’s official language, with female and male leaders and rank-and-file congregation members from different churches that were carried out between 2013–2019 as part of a larger project on religious aspects of household and family well-being. The interviews explored participants’ religious identities, experiences, perceptions, and attitudes.

I use both the Census and Survey data to create a statistical profile of women’s religious membership, identities, and involvement. In doing so, I focus, in particular, on denominational patterns, as they represent various dimensions of organisational and normative distinctiveness and resemblance. To classify and compare the denominations, I follow previous research in the study area (e.g., Agadjanian 2017; 2020; Agadjanian and Yabiku 2015) and define five denominational clusters: Roman Catholic (8.2% of congregations in the Census and 14.3% of affiliated Survey respondents); mainline Protestant, to whom I also refer as Protestant hereafter (15.8% and 11.0%); Zionist (48.2% and 48.7%); Apostolic (9.5% and 13.6%); and Neo-Pentecostal (18.3% and 12.5%). I acknowledge that this classification does not fully capture the religious diversity and complexity of the study site.

The qualitative data analysis builds and expands upon the quantitative foundation. In particular, this analysis explores how participants define and negotiate their religious membership and belonging, perceive and evaluate cross-church distinctions and similarities through their individual and collective everyday experiences, and ponder and decide to (not) exercise their church switching options. Selected quotes from in-depth interviews are included as illustrations of these patterns and processes (all names used here are pseudonyms). Finally, I infuse the analyses with insights from extensive participant observation inside and outside various religious congregations.

Findings

Scale and nature of religious belonging and engagement

Women’s church attendance is a normative and socially sanctioned and encouraged behaviour. Table 1 illustrates this by comparing church-affiliated married Survey respondents’ answers to three questions on whether they needed permission from their husband (or from his adult relative if the husband is not around) to attend church, to take a sick child to a health facility, and to visit a friend/neighbour in the community. Of the three scenarios, church attendance is the least likely to require husband’s permission. Table 1 shows some variation across denominational categories, but in all denominations, church attendance connotes greater autonomy, compared to the other two scenarios.

Table 1.

Ego needs her husband’s/his family’s permission for certain actions, Religious Survey (percent of currently married respondents)

	A. Going to church	B. Taking sick child to hospital *	C. Visiting friend/neighbour
Catholic	6.7	14.3	12.3
Protestant	8.1	16.4	14.8
Zionist	12.2	20.4	21.2
Apostolic	11.6	14.9	23.7
Neo-Pentecostal	9.3	17.3	20.9
All	10.5	18.0	19.6
N	1361	1231	1365

Note:

^*Respondents with at least one coresident minor child.

Overall, 74% of affiliated Survey respondents went to church at least once in the two weeks preceding the survey interview; among those women, the average two-week number of church attendances was 3.1. There is some variation by denominational category, which likely reflects church-specific expectations of attendance beyond the main Sunday service, but they are, in general, relatively modest (see Table A1 in the Appendix for denomination-specific statistics).

One possible indicator of women’s engagement in religious communities is their connectedness with members of their congregations. As a proxy for such connectedness, Figure 1 displays the distribution of responses to the Survey questions on the approximate share of adult women and men usually attending the respondent’s congregation whom the respondent knows by name. Reflecting the gendered nature of within-church social interactions, the share of respondents who know no or few male church attendees, 45%, was significantly higher than the corresponding share of female attendees, but even the latter share was quite high, 29%. Interestingly, restricting the sample to respondents who attended church at least once in the past two weeks (on the right of Figure 1) does not change these percentages much. I acknowledge, that these percentages may also vary by congregation size and overall number of regular attendees, which cannot be obtained from the Survey.

Figure 1.

Approximate share of male and female regular church attendees whom respondent knows by name – knows all/almost all, knows about half, knows few/almost none, can’t estimate, Religious Survey (percent)

The patterns in Figure 1 imply, even if indirectly, considerable social distancing and divides within congregations. Qualitative explorations suggest that variation in the level of financial contributions to the church is one important divisive factor, regardless of the denominational identity or congregation location. As Aurélio, a Methodist leader, summarised it, ‘What many churches value these days is money. They aren’t really churches, they are mini enterprises.’ Church members’ financial input is commonly perceived as a sign of an individual’s organisational worthiness, often fuelling tensions and distrust among congregation members with different levels of such input. These tensions are further aggravated by the perceptions that church leaders favour those members who contribute more resources and also by stigmatisation, even if not always explicit, of those who contribute little or nothing as uncommitted free riders. Importantly, although for married women regular or occasional contributions are normally family-based (except for any direct, and generally very small, contributions to church women’s group activities), these negative labels are typically attached to women, as men’s presence in the church is very reduced. These labels add to the cacophony of incessant inter-women gossiping and bad-mouthing that characterise the congregation members’ everyday life and relationships and are often construed as stemming from gender-intrinsic propensities (Agadjanian 2024), further undermining women’s emotional and social attachment to their church.

Cross-church exposure

In this dense and diverse religious environment, women’s exposure to other churches is very likely. Such exposure is arguably most impactful when it involves relatives. Church-affiliated Survey respondents were asked whether their parents currently belonged to their churches (for dead parents, the question referred to their religious affiliation at the time of their death). Mothers of 38% of respondents and fathers of 36% of them belonged to a church that was different from their daughter’s current church. The shares of other-church parents were understandably higher among respondents who had changed churches at least one once, with 64% and 60% of them, respectively, belonging to a church other than their mother’s and father’s, (see the breakdown of ego’s vs. parents’ affiliation in Appendix, Figure A1).

Church difference between parents and their adult children is normatively acceptable. However, even for younger, co-resident children, alternative church choices are tolerated. For example, Victória, a life-long Catholic, whose co-resident adolescent granddaughter attended a nearby neo-Pentecostal church, did not see it as a problem. ‘Everyone does what they want’, she reasoned. And notably, Victória was not sure to which church her daughter, the girl’s mother, and her husband, who lived in Maputo, Mozambique’s capital, belonged.

Church-affiliated respondents of the Survey were also asked about religious affiliation of their living brothers and sisters (if they had any), relative to their own. Overall, only 32% of them had all their siblings belonging to their churches. The shares of respondents with all-same-church siblings were understandably much lower among ever-converts, 19% for either brothers or sisters. Yet, even among the respondents who had never changed churches, only 47% and 46%, respectively, had all their brothers and sisters, affiliated with their church, reflecting the religious mobility of the siblings (see Appendix, Figure A2).

Church-affiliated Survey respondents were also asked about religious affiliation of up to three most trusted non-relatives and up to three closest neighbours. Overall, 43% of most trusted non-kin were not part of the respondent’s church, and notably, this share was higher among those who had switched churches at least once (45%) than among those who had never changed churches (36%). In comparison, 62% of closest neighbours did not belong to the respondent’s church, with generally minor differences between ever-switchers and never-switchers (see Appendix, Figure A3).

Whereas the diversity of women’s religious personal networks and their contact with members of other churches per se cannot measure their exposure to those churches, the Survey data on visits to churches other than their own offer a more direct indication of such exposure. These data show that 15% of religiously affiliated survey respondents went to other churches at least once in the twelve months preceding the survey, typically to participate in inter-church group/organisation meetings, to attend baptisms, weddings, or funerals, or to join their relatives and friends belonging to those churches. I acknowledge that survey respondents may have underreported their individual visits to other churches, especially those that may be disapproved by their congregation leaders and co-religionists. But in general, these occasional interactions may bring to the fore procedural and performative inter-church differences, such as (not) crossing oneself, wearing certain types of clothes, using different psalm books, beating drums during services, etc., which are clearly noticeable and meaningful but are also easily comparable and translatable.

Rules, norms, and morals – distinctively compatible

Behavioural and dietary norms and restrictions are prominent external markers of religious identity and often convey both cross-church similarities and differences. For example, drinking alcohol and smoking tobacco are seen as serious vices and are explicitly prohibited across denominations even though these prohibitions usually target men. In contrast, prohibition to eat pork (and typically also the meat of any other animals with unsplit hooves) is more gender-neutral, and it shows considerable denominational variation, being distinctly more common in Zionist churches. Other dietary restrictions in at least some churches may be even more narrowly specific, such as bans on eating certain types of fish or elephant and hippopotamus meat. And in some churches, periodic fasting is an important element of religious normative repertoire. While these restrictions, and corresponding cross-church variations, are nominally prominent, for most rural residents they have little practical relevance as food insecurity is endemic and, especially, any meat or fish consumption is a rare luxury.

Individual moral quality, and especially its gendered emphasis, is understandably crucial to the construction of church female member’s identity. Women’s everyday appearance is a prominent visual marker of their moral worthiness. The Census questionnaire included an open-ended question asking participants to list any clothing restrictions for female members of their congregations. Only 3% of the participants stated no such restrictions. Most restrictions were similar across denominations and included prohibition of short skirts, shirts and blouses not covering shoulders, tight clothes, transparent clothes, and otherwise revealing garments.

The Survey included closed-ended questions on (un)admissibility of specific types of women’s clothes in respondent’s church, and Table 2 shows the corresponding statistics. Not surprisingly, short skirts are most commonly banned. In comparison, the prohibition to wear pants is generally less common and is also less consistent. Table 2 also shows percentages of responses to two identical woman-specific questions asked in the Census and the Survey – acceptability of wearing earrings and using lipstick. The results point to considerable variations across denominational categories, but also between the Survey and the Census, with Survey respondents generally being less restrictively minded (except for wearing earrings among Zionist church members). In-depth interviews confirmed that this type of requirement may indeed affect individual church preferences, especially among younger people: for example, as Josina, a Zionist pastor acknowledged, some young women may be discouraged from joining her church because of its requisites for female members’ appearance. Importantly, these norms regarding women’s appearance, including church-specific sartorial rules (e.g., the expectation to wear only white clothes to church in some Apostolic churches) are publicly prominent as they can be observed by outsiders, and therefore they are considered as key individual markers of women’s church belonging.

Table 2.

Selected women-specific prohibitions, Religious Survey and Religious Census (percent)

	Short skirts	Pants	Earrings		Lipstick
	Survey	Survey	Survey	Census	Survey	Census
Catholic	92.1	62.1	8.3	16.1	16.6	26.9
Protestant	96.2	84.1	48.7	54.2	50.3	62.2
Zionist	99.4	97.0	90.3	77.5	81.9	83.2
Apostolic	98.3	95.4	14.5	28.0	18.3	38.3
Neo-Pentecostal	91.9	80.5	61.1	68.6	61.5	73.5
All	97.0	88.3	60.1	62.4	57.9	69.2

Marriage largely defines women’s moral and social identity, and accordingly, pre/proscriptions surrounding marriage are among the most salient church-promoted norms. These norms are typically characterised by a very similar formal stance but also its similar unenforceability across churches. The contrast between rule and reality is most vividly illustrated in the matter of religious wedding: while at least 60% of the congregation leaders in each denominational cluster interviewed in the Census said that their church requires for a married couple to have a religious wedding, fewer than 3% of church-affiliated Survey married participants were married through church (see Table A2 in Appendix for denomination-specific statistics). Similar inconsistencies between formal rigour and actual flexibility were noted in a study on churches’ official rejection of polygyny yet its acceptance in practice (Agadjanian 2020).

Perhaps the most important, and universal, gendered taboo is on women’s casual sexual partnerships, invariably labelled by church leaders as ‘prostitution’ regardless of whether they involve any financial or other transactional benefits. Yet, the boundary between ‘casual’ partnership and non-formalised marriage, i.e., a marriage that is not contracted through bridewealth payment, is increasingly blurred. In fact, bridewealth-based marriages, while generally tolerated and even legitimised across religious denominations, have been on the decline in Chibuto as elsewhere in the sub-Sahara (see Chae et al. 2021). Thus, the unions of 51% of affiliated Survey respondents who considered themselves being married did not involve any bridewealth payment.

Marital stability is another prominent social concern for Christian churches: ‘to divorce is to die’ is a common refrain one often hears in conversations with church leaders and members alike. In general, considerable consistency between church position on divorce, as articulated by church leaders in the Census, and by respondents of the Survey can be observed: churches overwhelmingly do not accept divorce, 92% and 93%, respectively, based on the Census and Survey. Despite the explicit and vigorous rejection of divorce by churches, 22% of currently affiliated ever-married respondents of the Survey reported having experienced marital breakup at least once in their lives, which illustrates both the general consistency of church norms and real-life deviations from those norms.

What makes a difference is healing… Does it, really?

Church-based healing is an important marker of church institutional and functional identity, especially for women, who are usually in charge of caring for children’s and other family members’ health. Overall, 44% of the congregation leaders interviewed in the Census said that their churches could cure at least some illnesses. The corresponding share was somewhat higher among the Survey respondents, 52%. Not surprisingly, Zionist churches had the highest share of congregations providing healing (75% in the Census and 94% in the Survey), especially compared to Catholics (0% vs. 4%) (see denominational details in Appendix, Table A3).

The contrast between the Census and the Survey probably reflects the fluid nature and meaning of church-based ‘healing’, which may target a wide and intersecting variety of physical, mental, and psychosocial ailments. Among healing churches, a major distinction, however, is not in the type of conditions they heal but in how they exercise healing – by adjuration only or by complementing adjuration with other procedures and ingredients, with the ‘mixed-method’ approach being particularly common among Zionists. According to the Survey, whose respondents were asked an open-ended question about healing tools other than prayer used in their churches, the most common of such tools (mainly in Zionist churches) are water (reported by 78% of women whose churches used anything other than prayer to heal), salt (78%), ashes (76%), laces and ropes (76%), traditional herbal medicines (41%), and even modern drugs (26%). Other healing accoutrements mentioned in the Survey are candles, oil, flour, pigeons, etc. Again, the specific healing toolkits varied across churches, and as the field observations abundantly suggest, these church-specific toolkits (and what people know about them) are among the primary indicators of constructed cross-church differences.

Understandably, treatment by adjuration only also entails considerable variation, illustrated by qualitative evidence. For example, Hortência, a member of an Apostolic church, told us about her daughter, who had health problems that, according to Hortência, were caused by the girl’s deceased paternal great-grandmother after whom she was named but for whom appropriate naming-related ceremonies were not performed at the time. Hortência, used the example of her daughter’s illness to explain how her church handles such problems compared to the Universal Church of the Kingdom of God, a Brazilian-origin neo-Pentecostal church with rising presence in Mozambique: ‘So, in the Universal Church they don’t want to hear about those people and ceremonies, they just want to get the spirit out. But in the Apostolic church they say to the [grandma’s] spirit “Madam, don’t get mad, in the name of Jesus Christ calm down”’. To note, Hortência had a relatively good idea of how the Universal Church would handle her daughter’s problem because her mother, a former witchdoctor, had joined that church.

Prophet figures are closely linked to healing practices in many churches, especially in AICs (see Section B of Table A2 in Appendix). Within the range of churches that use prophesising for treatment of illnesses, the subjectively salient differences are about how prophecy is exercised. Clara, who belonged to an Apostolic church, thus described the difference between prophesising in her church and in Zionist churches:

A Zionist [prophet] at some point [of prophesising] says the name of the person who did it [caused illness] to you. But in our Church it is not acceptable to say that so-and-so did this to you. The prophet, who has that spiritual order, detects your problem and take you to the father [church leader] and let him know what is going on. And then the father will do the prayer with you. He will say ‘our brother has problems’, but he wouldn’t say what kind of problem because he cannot spread a personal secret. And then we go to your home and pray there because the devil is always afraid of prayer.

Although some prophets are former witchdoctors who reoriented their skills and energy to serve the holy spirit, stressing their distance from traditional healing aetiologies and methods is key to church’s identity and claim of Christian authenticity. Thus, Zionist churches’ excessive use of healing tools other than prayer (i.e., other than the power of God), is often construed by outsiders as a marker of their embeddedness in traditional practices, which questions their Christianness. Yet, these procedural variations in prophesising and healing techniques go hand-in-hand with the fundamental similarity across healing churches (including Zionists) – that, unlike witchdoctors, prophets use the holy spirit, rather than ancestors’ spirits, to perform their job. As Lurdes, a Zionist preacher, explained: ‘In my church, we don’t accept demons’ prophecy. We want the spirit of God. You pray well, you fast, until the holy spirit appears. But if you want to use the holy spirit along with native demons [mademonio ya ntumbuluku in Changana] – in my church we don’t allow it.’ Lurdes also admitted, however, that not all Zionist leaders share her attitude: ‘Some pastors of other Zionist churches have a little hut at home where they do those things. But for us it’s a ndhumba [a cabin where witchdoctors perform their activities]. We don’t accept it in our church. We only use laces [to tie them to different parts of patient’s body].’

In general, as the field observations also suggest, the boundaries between ‘truly’ Christian and traditional healing are often not very clear in the interpretation of disease aetiologies, especially when the ‘disease’ has both corporeal and psychosocial components. Unlike the choice of the healing gears, where the power of the holy spirit is seen as the most genuine and central instrument and the distinctions are typically made with respect to engaging auxiliary tools, demons are seen as illness vectors, but their ultimate sources and triggers may be humans (dead or alive), implying a clear parallel with sorcery.

Despite explicit and often expressive self-distancing from traditional practices, in everyday life, the exposure of church members, and even of church leaders, to them is unavoidable. For example, church leaders are normally invited to ceremonies of veneration of family ancestors’ spirits. And these interactions may be particularly explicit and even formalised when sanctioned by secular authorities. Thus, when the secondary school of Chibuto town started experiencing frequent fainting among students, the school administrators, supported by the district authorities, mobilised traditional healers (also asking students’ parents to chip in cash to pay for the healers’ services), along with various church leaders, to help rid the school premises from the evil spirits ostensibly causing students to faint.

Some traditional healers are church members, especially when their church is generally tolerant of or indifferent to traditional medicine. For example, Mónica, a career witchdoctor, was a life-long Catholic, and in fact, was the women’s group leader in her congregation. The witchdoctor side of Mónica’s life was well known to her co-parishioners (and in fact, some used her services, also benefiting from discounts), and, as she explained, the priest did not care about it as long as she did not make it noticeable on the church premises. Of course, somewhat greater tolerance of traditional medicine among Catholics can be at least partially explained by the fact that the Catholic Church generally does not compete with witchdoctors in the local healing market. In contrast, for healing churches, this competition is real: in fact, Mónica, was invited to convert by some of those churches, attracted by her highly reputed healing skills and experience. However, those churches conditioned her conversion on burning her witchdoctor’s tools and accoutrements – which Mónica did not accept, not least because her witchdoctor job provided her with a relatively stable income.

The financial costs of church-based healing are an important factor in how healing churches are subjectively categorised. Many Zionist churches charge non-members fixed fees for treatment (again, similar to witchdoctors). These fees are normally waved if the patient decides to join the church (and the long-term success of treatment may be conditioned on joining the church too), but patients are still expected to provide substantial contributions to the church coffers even after they decide to convert. Ultimately, of course, the effectiveness of healing is what matters to potential patients, but given financial insecurities, perceived effectiveness cannot be separated from costs. This combined cost-effectiveness is thus central to potential patients’ preferences and choices, further reducing the subjectivised weight of the etiological and therapeutic idiosyncrasies of different healing churches.

Finally, although healing is an important axis of cross-denominational and cross-church variation, the Census shows considerable consistency in the list of illnesses that churches are unable to cure. Those are mainly physical pathologies that are conventionally reserved for biomedicine, including malaria, TB, cholera, anaemia, as well as HIV/AIDS and other STIs. And when Census participants were asked what advice they give to church members whose health conditions their churches cannot treat, encouraging sick members to go to a health facility or even taking them there was an almost universally response. But even church-based healing, regardless of its denomination-specific packaging, is typically seen as connected to biomedical therapies. Most healing churches focus on exorcising and neutralising demons causing the illness but delegate the treatment of the consequences of demons’ destructive actions to biomedicine by encouraging their patients to go to health clinics for ‘complete’ recovery. And the church-clinic interaction may be bi-directional: as one local nurse confided to us, she occasionally sent her patients directly to a neighbouring Zionist church if she suspected a demonic cause of their symptoms. Of course, neither traditional medicine nor biomedicine in this and similar contexts make a clear diagnostic and therapeutic distinction between physical vs. mental ailments. Understandably, however, given the virtual non-existence of formal ‘modern’ mental health services, traditional and church-based medicine fully covers what in western contexts would be considered mental health problems, into which such psychosocial concerns as those pertaining to marital and social relationship malfunctioning, or even economic insecurities, are also naturally integrated.

To switch or not to switch – is that the question?

Church switching is common in Chibuto: 52% of currently affiliated Christian respondents of the Survey reported having had at least one previous religious affiliation and 9% had been affiliated with two or more churches before joining the current church. More than nine out of ten congregation leaders interviewed in the Census said that at least one adult from a different church had joined their congregation during the twelve months preceding the Census interview, with remarkably little variation across denominational categories. The share of congregations that had at least one member leaving the church voluntarily (i.e., who was not expelled from the church) during the same period was noticeably lower, 62%, probably reflecting a desirability bias. Yet, again, the denominational variations in the exit probability were not very pronounced. Apostolics were the least likely to have lost a member but even among them, 52% of the Census participants reported at least one exit.

In this patriarchal setting, women’s church re-affiliation is strongly determined by marriage; in fact, in some Zionist churches, members are expected to marry outside the church as within-church marriages are seen as almost equivalent to consanguine unions. Eighty-three percent of the congregation leaders interviewed in the Census stated that if a male member of their congregation wants to marry a woman who is not a member of it, the woman must convert. There was little variation across denominational categories, with the exception of Catholic leaders among whom only 68% asserted the bride’s conversion requirement. Accordingly, only 5% of the women interviewed in the Survey had husbands who belonged to a church other than theirs, even though 17% of participants had husbands who were not affiliated with any church. Yet, even if the husband is not church-affiliated, he still needs to provide his consent for his wife to join a church. Here is how João, a male Zionist evangelist described the process: ‘[If a woman wants to joint our church], we will first call her husband and ask him “Your wife wants to pray with us, do you agree or not?” If he says yes, we will start looking for a way to get him to join too. Because when he joins, he will bring another friend later.’ João’s words also illustrate the symbolic and practical centrality of male church membership, even though men’s actual church involvement is very low.

Church switching may also be bureaucratically complicated as church leaders seek to reduce the pervasive concern about ‘stealing believers.’ Typically, potential converts are expected to provide a proof that the previous church’s leaders authorised their exit: 78% of the Census participants stated this requirement (see the denominational breakdown in Table A4 of Appendix). Yet, not surprisingly, the field observations show that this requirement is often bypassed. And in some cases, it may be used to screen the potential converts for worthiness. As João, who was quoted earlier, explained: ‘We want to see why they left the other church. For example, it could be that they did not pay the tithe. [Interviewer: And in that case would you not accept them here?] No, because they are not going to pay it here either.’ Being possessed by traditional (and therefore unquestionably evil) spirits may be another barrier the applicant should clear. Again, according to João, in such cases, his church tries to take the evil spirit out of the applicant’s body and have the holy spirit enter their body instead.

At the same time, switching is facilitated by generally low re-baptism requirements, especially among Zionist churches (see Section B of Table A4 in Appendix). And if re-baptism may have a symbolic meaning, it typically entails relatively small emotional or financial costs. Such costs are even lower for collective switching, usually under the follow-the-leader scenario, as it involves mainly changing self-labelling and reduces the individual psychological and social burden typically associated with cross-church transition. This was the case of a Zionist congregation headed by Armando. He was approached by representatives of a neo-Pentecostal church who promised a new building for his congregation and a new suit and other gifts for him personally if he switched his entire congregation to that church – an offer that Armando could not resist. Almost all members of the congregation accepted his decision, embracing the new labels and procedural epithets, while essentially maintaining the old church routines.

Yet, doubts about the potential benefits of switching, relative to its expected social and financial costs, are common and consequential. Celina, a Presbyterian woman who had two children from two different men with whom she no longer had relationship, was thinking about trying a Zionist church to see if it would help her find a stable marital partner. In addition, as she struggled through the high school (in which she was still enrolled despite being in her twenties), she thought that a Zionist church could help her with her studies too. Yet, she also shared her strong doubts that a Zionist church could really help with these two challenges. And Celina would not consider a neo-Pentecostal route, such as joining the Universal Church, which might be more appealing in terms of potential effectiveness and more ‘modern’ connotations but would also be prohibitively expensive. In comparison, Admirança, a Catholic widow who also at times considered switching, was deterred by potential psychosocial penalties switching might involve. ‘I don’t disrespect the Universal Church or Zionists,’ Admirança said, ‘but you see, if you go to pray there, they will tell you that your father has this, that your brother causes you this [sorcery]. I don’t want to have all this in my head, as my head will get swollen. And I don’t want this. So, when I go to the Catholic church, I pray, get out, and leave.’ Notably, Admirança had relatively high exposure to both Zionist churches and the Universal Church: her 25-year-old daughter switched to a Zionist church under her friends’ influence and her cousin joined the Universal Church.

Ironically, while the switching decision may be seen as a sign of considerable resolve, especially given the organisational and social barriers involved, it may also be perceived as a marker of spiritual weakness and of focusing on externalities rather than fundamentals. When Gilda, a Catholic woman, asked about those who switch, she explained: ‘Those are persons of little faith. Because those persons ask God for something today and expect him to give it to them tomorrow. But it’s not like that. We need to be perseverant. Because nothing happens just like that, without sacrifice.’ And Ana Maria, a female Methodist evangelist, when asked why members of her church may switch to other churches, answered, using the example of the Old Apostles church: ‘They do it because they covet the [white] clothes of the [Old Apostles] church, not because of faith, not because they really want to be in that church.’

Given the risks and uncertainties involved in church switching, at least some women may choose to attend services in other churches (i.e., beyond simply ‘visiting’ them on some symbolic or organisational occasions), without changing their church affiliation. Although such double-dipping practices are typically frowned upon – only 5% of the Census participants stated that they would not say anything to a member who attended another church (as opposed to telling them to stop doing so or to choose one of the churches) – these practices are nonetheless quite frequent and are facilitated and legitimised by having relatives belonging to those churches. For example, Admirança, who was quoted earlier, said that she openly and rather frequently went to her daughter’s Zionist church as well as to her cousin’s Universal Church congregation. And, again, the exposure to and easiness of attending other churches may also discourage permanent conversion: although Admirança admitted that she enjoyed her occasional experiences in those other churches, she definitely did not want to convert. ‘Entering one church, getting out, then entering another, ah, that’s no good’, she reasoned.

Conclusion

Applying the gendered doing-religion lens (Avishai 2008; Darwin 2016; Olajubu 2003) and engaging with the debate on intersections of Christianity and modernity in the Global South (e.g., Clifford 2016; Oduyoye and Kanyoro 1992; Pfeiffer et al. 2007; Soothill 2007; Wariboko and Afolayan 2020) I examined the meanings and implications of church membership and belonging for women’s everyday lives in a rural/semi-rural Christian sub-Saharan setting. As my analysis demonstrates, church involvement there is intrinsic to women’s identity and status. However, while church is a crucial anchor of women’s social self, positioning, and relations, it is also a constraint on and even a potential challenge to their well-being. Within-church divisions, distances and frictions – both vertical and horizonal – create hierarchies of worthiness and make own-church identity less coherent and less distinct. This vertical and horizontal distancing, rather than the often-obscure doctrinal and scriptural tenets or highly negotiable and even ignorable rules and regulations, primarily define women’s church persona and corresponding aspirations, choices, and behaviours.

Conforming to the cross-national evidence of gender differences in religious involvement (Pew 2016; Schnabel 2018; Walter and Davie 1998), the analysis illustrates the pressing need for women to be part of a church. However, it also shows the relatively little unique significance of which church to belong to. Accordingly, women’s exposure to different churches is frequent and continuous, yet it is also largely superficial. Cross-church differences, often adamantly depicted by church ideologues as indicators of their church authenticity, in practice are defined by what are essentially ornamental and performative features. These features may be subjectively meaningful and salient, but they are also easily comparable and negotiable. Even for such a seemingly central attribute as healing, its church-specific toolkits are constantly questioned, and so is, by extension, its effectiveness. Other church-specific rules, such as dietary restrictions, may be prominent markers of church uniqueness, but they have little practical relevance to everyday life. And more universal guidelines, like the requirement of church wedding, indissolubility of marriage, or ban on using witchdoctor services, may be forcefully articulated by church leaders but are rarely enforced and therefore are practically inconsequential. The pervasive moralisation of women’s image and behaviour across churches further diminishes the subjective salience of cross-church differences.

In sum, the inter-church boundaries are real and clear, yet these boundaries are also subjectively permeable. Crossing them is seen as common and normal, but partly for that very reason such crossing may not happen, especially with respect to conversions that are not normatively expected (e.g., those unrelated to marriage). ‘Voluntary’ church switching, often driven by health concerns or psychosocial challenges (which are often likened to health problems aetiologically), is typically under women’s soft control: it is a family-based decision that is supposed to be sanctioned by a male household head, but, in reality, it reflects mainly women’s preferences. Yet, such switching entails potential bureaucratic, procedural, and even financial costs. These costs may vary across churches but so may the perceived benefits of switching, as lower costs are usually associated with lower benefits.

The analyses demonstrate that in this context of rapid and multifaceted societal transformations, the traditional vs. modern scale may be subjectively very prominent. Yet, this scale, and the positioning of different churches on it, are defined primarily by the degree of constructed Christian genuineness rather than by the nature and quality of spiritual, social, and economic opportunities that church membership and involvement might offer women or allow them to engage in. In fact, normative and performative contrasts are often more pronounced not between the ‘old’ and ‘new’ churches but within each category. In particular, while the seemingly ‘modern’ performative flavours and accessories of neo-Pentecostal churches still have certain, even if fading, appeal, especially to younger (and generally better educated) women, they are hardly empowering in a broader sense.

I acknowledge, of course, that cross-church variations may still be consequential for women’s certain well-being outcomes, such as those related to their physical health, which may be affected by different churches’ doctrinal and organisational characteristics (e.g., Dodzo et al. 2016; Garner 2000), their leaders’ attitudes (e.g., Yeatman and Trinitapoli 2008) or their historical institutional ties with the modern health sector (e.g., Agadjanian and Jansen 2018). Yet, such variations and their potential benefits and penalties are generally unrelated to individual agency. The pressures on women to aspire and thrive, articulated most fervently by neo-Pentecostals, but increasingly also by older churches, are in stark contradiction with the lack of tangible opportunities to do so in the neoliberal context (cf. Sullivan and Delaney 2017) and, in fact, may have the opposite effect (cf. Gifford 2016). For religious membership and belonging to have a truly distinct emancipatory and empowering impact on women’s lives in rural and semi-rural Africa, women need to have viable alternative economic, social, and emotional choices outside of the organised religious realm.

APPENDIX

Table A1.

Church attendance in past two weeks, church-affiliated women, Religious Survey

Denominational category	A. Attended church at least once in past two weeks (percent)	B. Number of times attended church in past two weeks, mean (SD in parentheses)*
Catholic	68.0	2.3 (.09)
Protestant	76.9	2.4 (.15)
Zionist	73.4	2.9 (.08)
Apostolic	78.8	4.9 (.27)
Neo-Pentecostal	73.8	3.1 (.19)
All	73.8	3.1 (.07)
N	1775	1310

Note:

^*Respondents who attended church at least once in the past two weeks.

Table A2.

Requirement of religious wedding, Religious Census (percent), and actual experience of religious wedding, Religious Survey (percent)

Denominational category	Church requires religious wedding (Census)	Had a religious wedding (Survey)*
Catholic	60.2	2.6
Protestant	63.1	4.7
Zionist	61.8	1.4
Apostolic	72.6	6.4
Neo-Pentecostal	62.6	2.9
All	63.0	2.8
N	1114	1361

Note:

^*Among those currently married and affiliated with a church

Table A3.

Church-based healing (A) and prophets in congregation (B); Religious Census and Religious Survey (percent)

Denominational category	A. Church cures at least some illnesses		B. Congregation has at least one prophet
	Census	Survey	Census	Survey
Catholic	0	3.6	1.1	0.8
Protestant	1.7	16.9	.6	3.6
Zionist	75.1	94.1	74.7	93.0
Apostolic	32.7	50.6	63.6	85.1
Neo-Pentecostal	26.0	29.4	19.1	6.3
All	44.4	51.9	45.7	58.1
N	1122	1775	1122	1775

Table A4.

Requirements for potential converts, Religious Census (percent)

Denominational category	A. Potential convert must present an authorization document from previous church	B. Re-baptism requirement for converts
		Not required for any convert	Required for converts from some churches	Required for converts from all churches
Catholic	61.3	51.6	25.8	22.6
Protestant	76.3	81.3	11.9	6.8
Zionist	85.4	92.6	1.7	5.7
Apostolic	56.6	43.9	9.4	46.7
Neo-Pentecostal	76.4	66.0	23.2	10.8
All	77.6	78.0	9.9	12.1

Figure A1.

Mother’s and father’s church, relative to Respondent’s current church, overall and by Respondent’s church switching experience, Religious Survey (percent)

Figure A2.

Brothers’ and sisters’ church, relative to Respondent’s current church, overall and by Respondent’s church switching experience, Religious Survey (percent)

Figure A3.

Non-kin confidants’ and closest neighbors’ church, relative to Respondent’s current church, overall and by Respondent’s church switching experience, Religious Survey (percent)