Legal Formalisation of Land Rights and Local Subsistence Security: Matrilineal Land Institutions in Northern Mozambique

Abstract

While the legal framework in Mozambique in principle provides equal rights to land for women and men, its underlying assumptions imply that customary tenure is patrilineal, that women get access to land through their relationships with men, while men in practice own and control land. This article focuses on matrilineal land institutions in Makhuwa communities in northern Mozambique. It argues that local matrilineal institutions have provided women with transferable rights to land, while men get access to land in various ways within the matrilineal institutional framework. Based on the Land Law of 1997, the Mozambican government in 2015 launched a large-scale land tenure formalisation programme, Terra Segura (“Secure Land”), with World Bank funding. The Mozambican Land Law recognises local community rights to customary land, while Land Law Regulations define the requirements for both community and individual formalisation of such rights. Field data collected in Makhuwa communities, where individual titling was carried out in the period 2019–2023, indicate that both women and men received titles. But what is secured through these individual titles? The article discusses to what extent formalised “secure land tenure” for individuals can weaken women’s land rights and traditional rights to subsistence—provided by matrilineal land institutions over time, across generations.

1. Introduction

At Mozambican Independence in 1975, the revolutionary nationalist movement Frelimo (the Front for the Liberation of Mozambique) took power in the country after a decade of armed struggle. Already in 1977, armed conflict resumed, resulting from internal insurrections supported by external—primarily South African and Rhodesian—destabilisation policies and subversion tactics [1]. The result was a devastating civil war which lasted until 1992. For many rural people in the north, however, the armed struggle for independence and the following civil war constituted a continuous period of armed conflict, referred to as “the war”.1

When the Mozambican Parliament in 1997 passed a new Land Law (Law 19/97), Frelimo was still in power after the first multi-party elections in 1994. It was “a period of uneasy peace” [1] (p. 49), marked by incipient post-war reconstruction and large numbers of displaced people returning to the rural areas. There were also new initiatives to address the almost overwhelming problems in a political context of a newly established multi-party democracy. It was furthermore a situation characterised by competing claims and struggles for land [2,3]. The new Land Law was part of a comprehensive legal package recognising occupation according to customary norms and practices—provided it did not conflict with the Mozambican Constitution’s principle of equal rights for men and women. The development of this “legal package”, starting with a new Land Policy in 1995, was the result of a broad and quite inclusive process, involving national and international experts, representatives of different sector-interest groups, politicians and civil society organisations, as well as broader public hearing events [3].

During the 1990s, the international context was being reshaped after the end of the Cold War. With a renewed version of a neoliberal world order, the Bretton Woods institutions, the IMF (International Monetary Fund) and the World Bank, to some extent revised their policy priorities for Africa. There was a shift from Structural Adjustment to Poverty Reduction and Good Governance as overarching policy priorities [4]. Formalising land tenure through the registration of land and issuance of titles to individuals became part of this renewed agenda to eradicate poverty and bring about development [5]. The neoliberal land agenda in Africa was met with critical arguments among many scholars and civil society organisations [6,7,8]. Diverse arguments were in fact represented in the process of developing the Mozambican Land Law package. There were strong voices favouring exclusive private property rights in order to attract commercial investments and bring about modernisation, but there were also active proponents, backed by social-science field research, who defended the need for a legal protection of community-based land tenure [3]. Other voices advocated the need to strengthen women’s rights to access and use land [9] (pp. 50-51). What came out of this complex process was, according to Tanner [3] (p. 53), «a modern law in which sound legal concepts and mechanisms reflect an underlying reality that is genuinely African” (italics in original).

The 1997 Land Law had no reference to either patrilineal or matrilineal customary land tenure. There were several initiatives to secure women’s access to land in the law-making process, and the provision that only those customary norms and practices that do not conflict with the Constitution could form a legitimate basis for legal land rights was included in the law to secure women’s equal rights to land [9] (p. 49). Based on available sources, I would claim that the general assumption—also for women activists in this process—was that customary tenure would basically be patrilineal, and even if customary tenure should be matrilineal, women were likely to access land through their relationships with men—such as brothers or uncles. In this way, the new law also reflected dominant ideas and approaches from previous legislation, including the first land law in the country from 1979, together with colonial legislation and perceptions deriving from Mozambique’s colonial legacy. Among urban and political elites, there was considerable resistance to accept tradition, which was seen as a “return to the past” [3] (p. 19). The 1997 Land Law represented a political compromise between groups supporting collective land tenure and interests seeking to facilitate individuals’ and business enterprises’ access to land for agrarian development and modernization. Nevertheless, State ownership to land was maintained and legal restrictions on the development of land markets were upheld2. The political support to collective land tenure was in this context based on the socialist legacy of the Frelimo revolutionary state rather than on a recognition of the fundamental role of lineages and descent groups in customary land tenure in Africa [10] (pp. 33–34).

In this paper, I will argue that the dismissal of lineage-based social institutions in practice implied an acceptance of the assumption that customary land tenure is patrilineal.3 Furthermore, the dismissal of what I will call matrilineal land institutions within the legal package on land rights developed during the 1990s also meant that there has been no officially accepted policy framework in Mozambique, be it to document and analyse existing diversity in gendered land rights, to understand the nature of men’s land rights in matrilineal groups, or to protect women’s de facto land rights under matrilineal tenure. One may ask how matrilineal land institutions are relevant today. If we look at the last national census in Mozambique [11], it indicates a total population of approximately 28.86 million. The population in 2025—the year of writing—is estimated to approximately 34 million. According to the 2017 census, 25.9% had Emakhuwa as their first language, while the closely related Elomwe language was spoken by 7% [12]. Using first language as an indicator for belonging to an ethnic group, the Makhuwa and the Lomwe stand out as the two major matrilineal ethnic groups in northern Mozambique. Together they constitute close to one third of the Mozambican population. In recent decades, several important academic works have been published dealing with Makhuwa life cycle rituals, gender, sexuality, and religion [13,14,15]. Few studies have, however, aimed to document how Makhuwa land tenure institutions work [16,17].

In this paper, I will discuss new empirical material on Makhuwa land institutions collected during fieldwork in two rural communities in the district of Ribáuè in the interior of Nampula Province in northern Mozambique. With 25.9% of the Mozambican population identifying as Emakhuwa speakers, the Makhuwa numbered 7.47 million in 2017. Given a stable percentage of Emakhuwa speakers, an estimate for 2025 would be approximately 8.8 million. This is a substantial number of people, which necessarily implies considerable internal diversity. The Emakhuwa language is known to have seven varieties, in addition to Elomwe, which is normally seen as a separate language. Thus, field material from one district, Ribáuè, cannot necessarily represent Makhuwa institutions throughout the territory. The findings presented here do, however, reveal clearly recognisable patterns of matrilineal land tenure [10,18]. By showing how Makhuwa matrilineal land institutions work in these two Ribáuè communities, I hope to broaden the discussion on gendered land rights and tenure security at a time of rapid agrarian transformation in Mozambique. I have in this context chosen to use the concept land institutions to refer to commonly accepted norms that structure social relations, interactions, and practices related to land in rural communities. On the one hand, I want to communicate new insights on matriliny in Mozambique to a broader interdisciplinary field of land studies. On the other hand, with individual titling, holding a legal title document as a material object or being able to access to the State’s judicial system can, especially in situations of conflict or competing claims to land, become crucial to “enforceable recognition” of access to land [19]. Local matrilineal land institutions can in this way become subject to transformations that challenge established norms and practices. The result may be a significant shift in the balance of individual and collective entitlements, and a change in the nature of women’s and men’s rights to traditional lineage land.4 Thus, my claim is that recent land tenure formalisation programmes in Mozambique have the potential to transform matrilineal institutions.

As I see it, this transformation may take place even before the nature of Makhuwa land institutions in Mozambique are well understood and documented. A pervasive bias against matriliny in relation to land governance in Africa has been well documented in the literature, especially in works dealing with former British colonies [18,20,21]. Elsewhere, I provide a further contribution to this debate.5 The same bias can be found in knowledge production on land tenure and (what was called) “matriarchy” in the Portuguese territories in Africa. It should be noted that available colonial-time Portuguese publications on these topics appear much more outdated in a post-colonial setting than British sources from the same period. I would claim that in Mozambique there were virtually no “domestic” elements to build on in developing a policy-relevant scholarly discourse on matrilineal land after independence in 1975. As a prelude to my empirically based arguments on matrilineal relationships and rights to land in Makhuwa communities, I will in the next section briefly refer to some recent publications by Portuguese scholars who have studied knowledge production on land in the former Portuguese colonies (O Ultramar Português). This is complemented with a brief review of works published during the time when Mozambique was a Portuguese colony in the 20th century, focusing on references to indigenous land tenure, matrilineal institutions, and Makhuwa people.6

2. Colonial Sources on Matrilineal Land Tenure and Makhuwa People

There are few publications from colonial times that provide background information on the Makhuwa, and available published material rarely covers the topic of customary land tenure. The absence of the land question in the descriptions of the Makhuwa seems to be part of a more general trend of overlooking African use and access dating back to the time of the Portuguese colony [22]. Joana Pereira claims that: “Portuguese research and discourse on traditional African land tenure held the essential purpose of legitimising a legal framework denying colonised people’s rights over their land» [23] (p. 337). This legal framework was revised and adapted through a series of provisions during the first half of the twentieth century. Though different perspectives on “the natives” were voiced in the colonial metropolis, in part under the pretext of “protecting indigenous customs and traditions” [23] (p. 339), the various legal provisions primarily worked to advance the interests of the colonisers. Bárbara Direito holds that the widespread “communal” forms of land holding were seen as “an expression of the backwardness of the populations” [24] (p. 356). This conception was supported by the belief that “indigenous populations did not grasp the concept of individual property” [23] (p. 344). The result was that legal provisions allowed Africans to occupy and use land that was not required for other purposes but did not accept them as owners of the land they used.

Conceptions based on 19th-century social evolutionism no doubt played a significant role in shaping and legitimising both legal policies and knowledge production in the Portuguese colonies. This influence is documented in a textbook for future colonial officers in Mozambique, Elementos de Antropologia Geral, Etnografia e Etnologia [25]. Here we learn that to pass the exam for the lowest level of colonial officer, Chefe de Posto, only basic knowledge about “racial differences” was required. Higher levels of colonial officers should have knowledge about “human progress” and know the “modern classifications” such as Henry Morgan’s scheme of evolutionary stages; that is, Savagery, Barbarism, and Civilization [25] (p. 142). In this context, matriarchy is presented as the most primitive form of kinship organisation, associated with group marriages assumed to be prevalent during the early stages of Savagery.7 As patriarchy is considered to be developed during Barbarism, it is also seen as more advanced, “semi-civilized”, and associated with a certain knowledge of property rights [25] (pp. 184–186). Property rights in the context of Portuguese colonial knowledge production does, however, not refer to rights in land.

The 19th-century evolutionary perspective is also reflected in a “Definitive Project” on what is commonly called “customary law”; a project ordered by the Governor of Mozambique in the 1940s. The idea was that a designated group headed by a lawyer, José Cota, should travel throughout the territory of present-day Mozambique and document the customary laws, norms, and mentalities of all the ethnic groups in the Colony [26]. However, the group—or Cota himself—decided to classify all ethnic groups into two major categories: matriarchal and patriarchal [26] (p. 12). They justified this decision by the observation that in the Colony, the evolution of the family seemed to follow fixed laws. These laws of evolution implied that matriarchy gradually develops into patriarchy, and that all groups will go through the same phases [26] (p. 10). The “patriarchal” ethnic groups were thus considered to have reached a more advanced stage than the “matriarchal” ethnic groups. It is further claimed that patriarchal norms on marriage and property are “closer to our laws” [26] (p. 75). Along these lines, Cota provides a brief, clearly biased, but interesting reference to land rights among the members of the “matriarchal group”:

“…instead of the domus being the [male] head of the family, it is the elder woman of a settlement in whose name the lands are held. It is with her, instead of with the father, the right to distribute the fruits [of the land] is vested.”

[26] (p. 45)8

In a section on rules that should be applied in dealing with customary rights across all ethnic groups in the Colony, Cota repeats the claim that the indigenous population “does not have what we can call consciousness as owners” [26] (p. 67).9 In situations where different customary norms enter into conflict, a judge should give preference to the customary norms that appear most logical or “most evolved and just” from the perspective of Portuguese Civil Law [26] (p. 74). By favouring “patriarchal” norms, a judge would be able to facilitate the transition from matriarchy towards patriarchy, and in this way assist native people in their process of social evolution [26] (p. 76).

For social scientists working in the British colonies, the “Definitive Project” published in 1946, must have appeared outdated. Since the 1920s, the social evolutionism that informed Cota’s work had been contested and later abandoned, especially in British social anthropology [27]. From the 1930s onwards, British anthropologists produced a series of detailed ethnographic studies based on field data collected in neighbouring colonies. They used and further developed structural-functionalist categories and concepts in their analyses. These included concepts such as lineage and descent groups, with analyses indicating how patrilineal and matrilineal systems of kinship were associated with norms of access to land and succession to positions of power and authority.10 The anthropologists further focused on “systems” of marriage, and how these built on norms of patri-/virilocal or matri-/uxorilocal post-marital residence (cf. Pauline Peters for an updated overview of kinship and descent in the anthropology of Africa) [10]. Nevertheless, matrilineal land tended to be treated as a non-issue, in colonial Nyasaland—and even in contemporary Malawi [29].11

As a colonial power in Africa, Portugal was aware of both policies and knowledge production relating to the territories controlled by the British and the French. This may have been one of the factors motivating the establishment, in the 1950s, of a new framework for scientific research in the Portuguese colonies. In Mozambique, the well-informed colonial officer and self-taught anthropologist Antonio Rita-Ferreira was appointed as responsible for the social sciences, including ethnology and ethnography. Drawing on available studies, he proceeded to compile an overview classifying all ethnic groups in Mozambique, but now based on the criteria provided by the British International African Institute [30] (p. 351). In 1975, the year of Mozambican independence, a revised version of this overview was published under the title Povos de Moçambique: História e Cultura [31]. This is a work that has been widely read and distributed.

In his overview, Rita-Ferreira categorised the Makua-Lomwe as a single ethnic group [31] (p. 205); a group which despite being the most numerous, was said to be the least known in Mozambique. During colonial times, no in-depth studies on the Makhua-Lomwe had been carried out by professional anthropologists, and according to Rita-Ferreira, the available information on these groups was full of contradictions [32] (p. 209). In his book there is no reference to knowledge on land tenure among Makhuwa or Lomwe people.

3. Mozambican Land Rights in Shifting Policy Contexts

The independent socialist People’s Republic of Mozambique passed its first land law in 1979, enacting the principle that all land is property of the State.12 The 1979 law stated that land could not be sold, rented, mortgaged or in other ways alienated.13 In line with the dominant revolutionary discourse at the time, the introduction to the 1979 land law emphasised that Frelimo’s aim in the armed struggle for national liberation had been “to liberate the land and the people”.14 Thus, after the “usurpation and plunder of the best land during five hundred years by Portuguese colonialism”, the goal was to “wrest the land from foreign subjection and exploitation and devolve it to the Mozambican People…”.15

The new Land Law enacted in 1997 retained from the previous law the principle of land as state property, and it maintained restrictions on the use of land as a commodity in a market economy. In contrast to the previous law, the political compromise expressed in the 1997 Land Law meant that it would, in principle, provide legal protection of existing land-use rights of the majority of the population. Tanner [3] (p. 22) describes the approach chosen as treating areas “where local rights existed as self-contained land management units within which the prevailing local land customs could and should apply” (italics in original). The concept of “local community” was chosen to denote the units where “…customary norms were acknowledged as the legitimate way in which local residents acquired and managed their land rights” [3] (p. 28, italics in original). The tenure security provided through the 1997 Land Law was in principle equally valid for customary rights without a formalised title and for individual or private land rights with legal titles. They were all defined as legal rights to “use and benefit from the land” (DUATs).16 According to the law, a DUAT could be acquired through three channels: (a) occupation by individual persons and by local communities according to customary norms and practices that do not conflict with the Constitution; (b) occupation by national individual persons who, in good faith, have been using the land for at least ten years—a provision taking into account the large number of refugees and internally displaced people after the civil war; and (c) authorization of a petition presented by individual or collective entities—a provision aiming to provide opportunities for both national and foreign commercial development interests and investors.

A decade after the enactment of the Mozambican Land Law, the global financial crisis of 2007–2008 led to a dramatic increase in world food prices. One effect was a new “land rush” in Africa, as investments in the productive and financial value of land in the global South became an attractive option for international financial capital in the wake of this crisis. Investors’ interests were usually based on the assumption that there were “marginal, under-utilized, empty and available lands” [33] (p. 4). This interest was often welcomed by domestic power elites in countries such as Mozambique. The province of Nampula in northern Mozambique (where fieldwork for this article was carried out) became one of the hot spots for planned large-scale investment initiatives to promote commercial agriculture through the so-called ProSavana programme, in combination with extractive mining and infrastructure development [34]. It was clear that ProSavana would imply a massive increase in the pressure on smallholders’ land along the infrastructure corridor in the province, and the initiative became highly contested both at national and international levels [35]. At the same time, “land grabs” in the global South became a paramount topic in international studies of land tenure and development [36,37,38]. By 2015, however, the most fervent critique and discussion of this land rush had waned [39]. Meanwhile, the problem of land grabbing had created a new context for debate on land tenure in Africa. In this situation, powerful international actors—with the World Bank and FAO in key roles—were, according to Laura German [40], able to create a relatively broad consensus around a common discourse on “land governance” in development. German describes the consolidation of this new “land governance orthodoxy”, where notions of “tenure security” and “women’s empowerment” were used to: “… redirect the 2007 outcry over ‘global land grabs’ to garner support for land titling and procedural forms of rights recognition, while obscuring the relationship between these instruments, the commodification of land, and growing transnational interests in Africa’s farmland” [40] (p. 3).

In pre-land-grab debates on formalisation of land tenure, it had been argued that if there were “rights on the ground” then they did not need “formalisation” [8] (p. 21). This view was also reflected in the Mozambican Land Law of 1997, which aimed to provide sufficient security for customary tenure without formal land titling. The Mozambican Land Law Regulations did, however, open up for formalisation of existing informal rights through land titling if deemed “necessary” or requested by “interested parties”. A Technical Annex to the Law described how community land could be formally delimited.17 The legal package furthermore provided for individual titling through demarcation. Increasingly, individual titling has also been carried out within delimited local communities.

It has been argued that the implementation of the 1997 Land Law has been conditioned by the Frelimo government’s emphasis on the role of large-scale agriculture in economic development [42]. During the revolutionary socialist era, the major emphasis was on state farms. With neo-liberalisation, the priority shifted towards large-scale commercial agriculture. Smallholders have in practice received little support, “…despite the significance accorded to their land use rights in the law” [43] (p. 324). When formalisation of land rights in rural areas has been carried out under the 1997 Land Law Regulations, either through community delimitations or through registration of individual plots, funding has primarily come from international donors, while Mozambican civil society organisations or NGOs have taken on the role as implementers. In this context, the Mozambican NGO called ORAM, now Amder, has been a central actor from the time of the making of the Land Law.18 Initially, ORAM worked with delimitation of community land, and was able to raise funding from international donors to carry out this work. The Community Land Initiative (iTC—iniciativa para Terras Comunitárias) was a separate land formalisation initiative funded by six European donors, initially led by the British Department for International Development (DFID).19 ITC started up its work in three Mozambican provinces—Manica, Gaza and Cabo Delgado—in 2006.20 In 2009, with funding from the US-based Millenium Challenge Account (MCA)/Millenium Challenge Corporation (MCC), iTC expanded its work into three further provinces, all in the northern part of Mozambique—Nampula, Niassa, and Zambezia.21 While iTC above all worked with community delimitations, MCA also had a separate initiative to pilot individual land registration [42].

A significant further step towards more generalised land rights formalisation was taken when the Mozambican Government in 2015 launched an ambitious World-Bank supported programme called Terra Segura (“Secure Land”), aiming to strengthen formalised rights recognition, without—necessarily—changing the 1997 Land Law.22 Terra Segura represented a clear shift towards scaling up and prioritising individual land titling within the existing legal framework in Mozambique.23 The programme objectives originally included issuing 5 million individual and 4000 community land titles [44], but was considerably down-scaled during the programme period. In addition, a central objective was to modernise and digitalise national land administration.

The World Bank committed to fund large parts of the Terra Segura programme through a 100 million USD loan. The World Bank project called “Mozambique Land Administration” (Mozland) had explicit objectives of strengthening land tenure security and facilitating the development of commercial agriculture. Tenure security-related project objectives have included “raising awareness of the importance of real property rights [my italics] particularly at the community level”, as well as what German [40] refers to as “women’s empowerment”, in terms of providing “opportunities for women to express their concerns and learn about mechanisms available to present any claims…” [45]. Project targets in terms of the total number of community and individual DUAT titles issued had been significantly reduced by the closing date in 2024 [46]. The gender target of, “at least 40% of the project beneficiaries who receive DUAT titles are women,” on the other hand, was upheld [46] (p. 3). When we look into the indicator for this gender target, we find that the target of 40% of DUATs to women in World Bank accounting includes both DUAT titles issued to women and DUAT co-titles to husband and wife together.24

While other donors during the same period have supported titling projects under the Terra Segura umbrella, these have been at a more limited scale than the World Bank’s Mozland. After the finalisation of the Terra Segura programme, the end result remains far from comprehensive coverage of formalised land tenure in rural areas. In the region where the fieldwork reported in this article was carried out, what we found in 2024 was a mix of formalised and non-formalised land tenure. After highly contested general elections in Mozambique in October 2024, in January 2025 a new Frelimo administration took power in Mozambique. This government did not start out with any official statement on its land policies. So far, however, Terra Segura has been the major initiative in Mozambique, informed by what German [40] calls the “new land governance orthodoxy”. The programme has no doubt been instrumental in substantially scaling-up individual titling, especially within already delimited local communities. In the context of Terra Segura, tenure security has meant formalised tenure, either through individual titling or co-titling based on Western assumptions of conjugality. Through identifying, geo-referencing, and registering land holdings, and then issuing title documents (DUATs), tenure is considered “secured”. This implies a conception of security as primarily inherent in a particular type of legal procedure, that is, titling.

In the project documents published by the World Bank [44,45,46], there is no reference to the role of lineages in land tenure in Mozambique, either matrilineal or patrilineal. Terra Segura was a programme that operated with a gender target, but it had no reference nor any approach to how matrilineal land institutions should be taken into account in individual land titling processes, or what “tenure security” implies for women and men in matrilineal settings. As part of the more comprehensive process of neoliberal reform and modernisation of land governance under the umbrella of Terra Segura, a new Land Policy was also approved in Mozambique in 2022 [47]. The objectives of the new Policy are stated in terms of land tenure security for all, while the sub-objectives include the facilitation of both national and international capital investments in land development. The Policy at the same time recognises the diversity in land tenure arrangements in the country, including the existence of matrilineal land tenure; with §50 stating that the Policy provides a highly flexible approach, and considers:

“…local realities and objective social differences between rural and urban spaces, between spaces predominantly matrilineal and patrilineal, between spaces structured predominantly in relation to traditional authorities and spaces structured in relation to other forms of socio-cultural organisation or community authorities”

[47] (p. 14, italics added).25

For the realities of “predominantly matrilineal spaces” to be taken into account in the implementation of any land governance procedures in northern Mozambique, there is a real need for a better knowledge base on how matrilineal land institutions work.

4. Methodology

The regularisation of customary land tenure through individual titling includes mapping, registration of boundaries between family lands and individual plots within communities, and decision-making on whose name will be written on each title document. These processes can from an empirical researcher’s perspective provide opportunities to learn more about local land institutions. Land regularisation through individual titling requires that women and men in rural communities make explicit the norms that inform their practices related to access and authority over land when they make decisions on whose name should be written on a title document. Titling processes in local communities may bring up diverse perspectives on relevant norms regarding land tenure, but a titling process that provides real space for local men and women’s participation may also contribute to more awareness and possibly strengthen local consensus on relevant norms on land tenure. Such a process can at the same time provide arenas for strengthening the power of some categories of community members at the expense of others [41]. When it comes to matrilineal land tenure in Mozambique, these processes provide some unique opportunities for researchers to learn about local people’s perspectives and reasoning. Seeking to understand processes of individual titling, we also gain access to the norms that guide local land related practices and constitute local land institutions.

Funding from the Swedish Research Council during the period 2021–2025 and collaboration with Amder as a provincial project-implementing organisation, have provided a research team—where I collaborate with colleagues from Gothenburg University and Eduardo Mondlane University in Maputo. Fieldwork has been carried out in Ribáuè District in the northern Mozambican Province of Nampula. Our team has participated in numerous meetings with province and district officials, as well as with civil society representatives at the provincial level.26 The team members were introduced to the two communities in Ribáuè District where we wanted to carry out fieldwork through community meetings organised by local representatives of the NGO Amder. We had preliminary meetings in both communities in 2022, where the community representatives agreed that we could come back to do fieldwork. In 2024, we were able to return, and new community meetings were held in both field sites. With the consent of those present, the community meetings were followed by group interviews. Only then we started with individual interviews. In my research, I had by the end of 2024 carried out 19 individual in-depth interviews focusing on land, land use, and land titling in the two selected fieldwork communities “Muthoko” and “Canacura”.27 The interviewees, nine women and ten men, represent 19 different households. The interviewees were selected by helpful local leaders.

Each research team member has worked closely with a local interpreter to translate between Portuguese and Emakhuwa, which is the language spoken by all community members in the research area. The individual interviews were carried out based on a common framework, a joint but relatively open interview guide developed by the research team. My particular focus during individual interviews, in addition to basic information on family structure, access and rights to land, has primarily been kinship relations, life histories, and land-related practices such as cultivation and handling of crops.

5. Findings from Fieldwork

The group interviews carried out in Muthoko and Canacura focused on the level of local community as defined in the 1997 Mozambican Land Law:

“A grouping of families and individuals, living in a circumscribed territorial area at the level of a locality or below, which has as its objective the safeguarding of common interests through the protection of areas of habitation, agricultural areas, whether cultivated or in fallow, forests, sites of socio-cultural importance, grazing lands, water sources and areas for expansion.”

(Law 19/97, Article 1/1)

The social and territorial unit of local community is basic in the implementation of the Land Law, and the fundamental unit in regularising traditional land rights through community delimitation.28 The concept of community—comunidade—would also serve as a key concept in our conversation during the group interviews reported below. The territories of both Muthoko and Canacura had been formally delimited as comunidades, with collective land rights registered 8–10 years before our fieldwork.

In the individual interviews, when focusing on access and rights to land, I sought to understand men and women’s perceptions of the new land demarcations (i.e., the registration of individual plots) that had been carried out on community land. When discussing general access and rights to land, as well as the demarcation of individual land holdings, interviewees commonly referred to the nloko, the Emakhuwa term commonly denoting a descent group in the form of a minimum three-generation matrilineage.

Each nloko belongs to a larger matriclan, or nihimo. At the individual level, the nihimo serves as a unit of belonging, and a person belongs to the nihimo of the mother. The nihimos are named, with Alaponi, Amale, Amirasi, Amucovo, Aceleche, and probably others present in the field area. Unlike the nloko, however, the nihimo does not correspond to any territorial unit and has no leader. At present, the Makhuwa nihimo thus constitutes a social category drawn upon in some situations, rather than an organised social group. The different nihimos are spread over the territory where present-day Makhuwa people live. Thus, as a Makhuwa you can meet somebody from your own nihimo anywhere when you travel in the northernmost provinces of Mozambique, and you can ask somebody from your own nihimo if you need help. Nihimos are also in principle what anthropologists call exogamous clans, which means that the father and mother of a person belong to different nihimos, and you are expected to marry someone from a different nihimo than that of your mother—which is also your own in a matrilineal society.

5.1. Analysis of Group Interviews Focusing on Land at Community Level

As an introduction to our individual interviews in the community of Muthoko, our research team met with a group of approximately 20 local people, mostly male community leaders.29 The traditional male leader, the Mwene, and the corresponding female ceremonial leader, the Piyamwene, were both present together with the Cabo de Terra, who stands out as the most knowledgeable on land issues.30 From the group interview we learn that Muthoko covers a vast territory, reaching as far as several of the mountain peaks in the horizon. Such peaks and in between them large areas of cultivable land are characteristic of this landscape in the interior of Nampula Province, and Ribáuè District in particular. We further learn that Muthoko was established as a separate community “when Frelimo entered”.31 During colonial times, the current Muthoko area belonged to a larger Regulado, which was dissolved and subdivided into smaller administrative units by the Frelimo authorities when they came to power.32

Under the auspices of Terra Segura, the land belonging to Muthoko was delimited as a local community in 2017, with the NGO ORAM (now Amder) as implementing service provider. Responding to a new request from the community, Amder continued with family and individual land demarcations in 2022. Family and individual DUATs were later issued by government authorities and handed over to the owners, though this process of issuing DUATs has taken time. Still, people in Muthoko seemed confident that they would receive their individual DUAT titles. Among those present at the meeting there was general agreement that this regularisation of community land has been positive for the community. Earlier there tended to be discussions and even emerging conflicts over lands and boundaries. Now, they say, internal boundaries have been clarified, there is agreement on the extension of land holdings, and there are no more internal conflicts over land. So far.

In our introduction to the neighbouring community of Canacura, we met with a group of 16 people. Here, too, most of those present were community leaders at different levels; one third were women. The Piyamwene was present, together with the Cabo de Terra, and a well-informed Community First Secretary.33 In the group interview, we learnt that Canacura is a community with much land, 15,000 Ha in total, with a population of around 2500 people. Some of the mountain peaks we see from our meeting place mark community boundaries, while a hill closer by has given the community its name. During colonial times, the community was part of a larger unit with one leader, a Mwene, but also part of a Regulado that covered an even larger territory. It appears that historically, after the death of a Mwene, a new one was not appointed to take over and as a result several communities established themselves with their own local leaders. In Canacura, it appears that the Cabo de Terra has now taken over some of the ceremonial responsibilities of the former Mwene.

At the beginning of this century there were discussions with the leaders of neighbouring communities concerning boundaries, they say. At that time, ORAM was introduced to the community, and after some conversations it was decided to proceed with a delimitation of the community. This process was completed in 2014. Under the Terra Segura programme (with World Bank funding for the Mozland project), a first phase of individual land demarcations started in 2019, with a service provider I shall call “TG”. Among those present there was considerable frustration with how TG carried out these demarcations. They say representatives of the service provider just came and informed the community that the following day they would start demarcations. This meant that the work was not always well planned, those affected were not informed, and some interested parties were not present during the demarcation of boundaries between plots belonging to different nlokos. They say that TG wanted to work fast and preferred to register plots with easy access close to the road; they did not want to walk long distances. In 2022, Amder came back (with funding from another project) to register the land that was not demarcated by TG. According to those present in the group interview, all nlokos now have their land registered, but a number of people are still waiting impatiently for the DUAT titles to the land registered by the first service provider.

During the group interviews in Muthoko and Canacura, we learnt about some of the complexity of tenure security and insecurity on the ground. The delimitation of the area belonging to each community was originally requested by community representatives. At the time of fieldwork, community delimitation appears to have been well received by community members, as it clarified the extent and boundaries of the territory over which each community has “enforceable claims” [19] (p. 155). These communities cover extensive areas of land, and the process of clarifying and agreeing on inter-community boundaries appears to have ended earlier discussions regarding community territories and boundaries. The community members interviewed have also been positive to the demarcation of individual land holdings within each community, as these have clarified the extension of the land belonging to each nloko. The identification and registration of individual plots also appear to have resulted in agreements on the boundaries between plots within nloko land, and have thus put a halt to intra-community discussions and possible emerging conflicts over land—so far.

In Canacura, however, a considerable number of people who believe they had their land registered still lack DUATs. Here it is especially in relation to outsiders that the perceived tenure insecurity among those without DUATs has risen, it seems. They do not know what has happened to their title documents or where they are. A dominant perception is that if potential “land grabbers” enter the community to claim or occupy land you need a DUAT document to prove that you have “enforceable claims” to your land. Showing the DUAT title will provide security when facing land grabbers, they say.34 In this perspective, people who have not received the DUAT documents feel less secure in a community setting where a considerable part of individual land holdings can now be documented with a formal title.

5.2. Analysis of Individual Interviews Focusing on Family Land and Demarcation

The individuals we interviewed in this project were selected—and probably motivated to participate in our research—by local leaders. On this basis, I cannot claim that I interviewed a representative sample. Many of those interviewed—but not all—are members of the dominant nlokos in each community. Some are local people who said they belong to nlokos with a long-time presence in the community; their ancestors arrived khalaí—a long time ago! Other interviewees had arrived more recently, during or after “the war”.35 When I have chosen to present a set of individual narratives here, it is with the aim of providing a description of matrilineal institutions through the way in which they structure women’s and men’s opportunities and choices in relation to land. I also aim to show how marriages are relevant to individual men’s and women’s access to and use of land. The individual narratives presented below are furthermore chosen to represent some of the variation and complexity in my material of recorded conversations.

The Cabo de Terra in Muthoko told that his nloko had its origin in a grandmother and a grandfather who “went the same way”.36 From them emanated various “ways” or branches of family currently living in the community. Some interviewees told a more elaborate story on how their nloko came to live in the place where it is now located. Here is one version of this story of origin, as told by both the Cabo de Terra and the Piyamwene in Canacura community. They belong to the same nloko, as their mothers were sisters. Here is the story of their nloko:

The ancestors originally lived in a place called Namuli. But then they left Namuli and went to places where nobody lived. Before they decided to settle down here, they brought a chicken and left it to see if it could survive on its own. They wanted to know if the land was good and see if the chicken could sustain itself on the land. If the chicken lived after one week, they would know that the land was good for living. When they returned to where they had left the chicken, they found that it had survived and grown. Then they knew that the land was good and settled here. It was “the father and the mother of the nloko” who brought this chicken.37

The Piyamwene in Muthoko tells that her nloko came to the community before the war, during colonial times:

The grandfather and the grandmother settled here, the grandmother got pregnant, and daughters and sons were born, and then grandchildren were born.

The Piamwene herself was born in Muthoko, and both her mother and father were also born here. Land for cultivation is distributed by the elders to the members of the nloko. She says that land is primarily given to the daughters, since they give birth to children. “The sons go to marry other nlokos.”

During the individual interviews, I discussed the Emakhuwa concepts of nloko (descent group) and erukulu (womb; sibling group) with most of my interviewees. I further sought to identify Emahuwa terms that could possibly correspond to “household”, a concept used in most interdisciplinary publications on land tenure. I could confirm that Emakhuwa-speakers do not use—or need, apparently—a term denoting household. A term that sometimes came up was amosi, meaning “kin” in general and not referring specifically to a social unit living together, sleeping under the same roof, and/or regularly eating together as “household”. A husband could mention that he had made an etthoko for the couple after moving to live together with his wife; but this is a word referring to the “house” or “home” as a built construction. Sometimes the etthoko has a storage room for harvested crops, sometimes the husband builds a separate storehouse on pillars, a nikhipirani. The woman manages and controls what is kept in the storage room or storehouse. One interpretation could be that the wife has the keys to “household resources”, while I think it is closer to local conceptions to say that the woman manages the resources that secure the subsistence of her erukulu, “her womb”, in the sense of the children born from her womb.

Most interviewees, both men and women, emphasised that husband and wife work together, and sit down to make decisions together on what to do with crops and income. The married couple appears to be a key unit of production in agriculture. Only a few interviewees mentioned that children living with them also participate in agricultural activities. Regarding decision-making, some interviewees say that the husband decides, usually after consulting with the wife. Here they seem to refer to the—fairly widespread—notion of “head of household”. However, the Cabo de Terra in Muthoko firmly held that: “The man who decides alone is an unmarried man.”

The term erukulu was used in addition to nloko when I discussed land and access to resources with my interviewees. Both women and men would first point towards their lower abdomen when trying to explain the meaning of erukulu. Then they would clarify that as a social category it refers to the sisters and brothers who have their origin in the same womb; that is, siblings who have the same mother, but not necessarily the same father. In my understanding, the erukulu constitutes the unit of reproduction and growth of the nloko through its female members. By belonging to an erukulu you have access to your mother’s nloko land for cultivation, with prioritised access for daughters. You can also claim to belong to the place where your mother was born, which is where her erukulu had its origin. As members of the same erukulu, a mother’s sister or mother’s brother may further take on specific roles and responsibilities, as illustrated below.

Avelina, a mature woman I interviewed in Canacura, illustrates some prevalent norms on women’s rights to nloko land, with a member of Avelina’s mother’s erukulu (her mother’s sister) giving her land when she came to ask for it:

Avelina was born as her mother’s only daughter in a community located at some distance from Canacura. Her mother had been born in Canacura, but she had moved to the community where Avelina’s father lived to join him when they married. Avelina, however, decided to go back to Canacura to live on the land of the nloko she belonged to through her mother. Avelina’s mother’s sister gave her a plot of land for cultivation, and Avelina says this is her “place of belonging”. Here she is kamumu.

In the fieldwork area people distinguish between kamumu and viente. As kamumu you belong to a place through birth or nloko membership; as viente you have come from the outside to live in the community, and your nloko may or may not have land elsewhere. This is typically the situation of a husband born in another community, who moves to the wife’s community to settle and work on her nloko land.38

Daniel is viente. He is a confident, relatively young man. He was born in a neighbouring community, where his matrilineal nloko has land.

When Daniel was young, he came to Canacura to look around, and he saw a girl he liked and wanted to marry her. Not just have an affair but marry. When he asked, she accepted. He returned home and told his parents, and getting to know which nloko she belonged to, they agreed. Then his family sent his mother’s eldest brother to ask for marriage, and the family of the bride accepted. They said: “He can come”.

After Daniel had come to live with his wife in Canacura, he asked her family for land to build an etthoko (house) for the couple, and then a nikipirani (storehouse) for his wife. They also got land for cultivation within his wife’s nloko land.

Two of my interviewees in Canacura were mother and daughter, Mariana and Elina, and their accounts illustrate how the concepts of erukulu and nloko are used to explain and justify their own relationships to land.

Mariana was born in a locality not far from Canacura, and married a man who was born in Canacura, she says. When marrying, he “followed his wife” and went to live with her on Mariana’s mother’s nloko land. But there was not much land for cultivation there, so Mariana’s husband asked her parents if he could take her to his nloko in Canacura, where there would be lots of land for cultivation. Mariana’s parents accepted, and the couple moved to Canacura. This was before the end of the war. Mariana’s husband is now the only one of seven siblings in his erukulu [sibling group] who is still alive. Mariana is the only one in her erukulu who lives in Canacura; she says she is the only one who lives here “with her daughters”. In total, she has seven living children. Three daughters and one son are married.

When I interview Elina, Mariana’s eldest daughter, my impression is that she looks fairly young, but exhausted. During the interview I understand why she looks tired. I also get a different story about the origin or her nloko.

Elina says she is kamumu de Canacura, she was born here. Her nloko is located here, she says. She also says that her mother was born here [but see above]. Elina herself married a man from Canacura, a kamumu. Her husband moved to live with her when they married—he “followed her”, and he built an etthoko (house) for them to live in close to her parents. They worked together on her nloko land, and had five children, three daughters and two sons. Suddenly, a couple of years ago, Elina’s husband left and moved to another woman in the community to live on her land. Now, Elina is left to work alone in her fields. Though she has a 19-year-old son living with her, he does not help her in the fields, she says, tired and somewhat resigned.

What Mariana and Elina tell illustrates several interesting points concerning matrilineal land institutions. One is how the land allocated to Mariana and her husband from—what she claims was—his matrilineal nloko, with the husband/father as the only surviving member of his erukulu, in Elina’s account becomes the starting point of a new nloko “branch”. Mariana then becomes “the mother” of a new matrilineal nloko, which has three generations, including Elina’s daughters and sons. Elina adapts the story of her mother’s origin, affirming that her nloko is a matrilineal nloko with its basis in Canacura.

The other point in this case is that Elina’s former husband, who is kamumu and born in Canacura, went to live with her and work on her family’s land when they married. According to the interviews, matrilocality is clearly the prevalent pattern of residency in both Canacura and Muthoko. The exception seems to be men in nloko leadership positions. A new couple will normally settle on the wife’s nloko land. Most also marry within the same community, almost all will find a spouse from a different nihimo, and everybody will marry outside the nloko.

Mariana’s (and Elina’s) nloko land has been demarcated, and in total, this nloko had many plots registered. While their accounts, as presented above, describe the situation in matrilineal terms, both women and men had DUATs registered in their name within the nloko land. Mariana says she had three plots registered in her name, while three were registered in her husband’s name. Elina has four plots registered in her name, her two married sisters have two and three plots, respectively. Mariana’s married son, who lives with his wife and cultivates her nloko’s land in Canacura, also has three plots registered in his name. However, Mariana holds that she keeps the DUAT titles for all the nloko land with her.

In Muthoko, the Piyamwene tells that after the demarcations were carried out within her nloko’s land, it is her eldest maternal uncle, the Cabo de Terra, who keeps all the DUAT documents. She can ask him if she needs her documents. The Cabo himself is married for a second time—after his first wife died—to a woman from outside the community, and the DUATs for the land they cultivate together are issued in his name. He has seven daughters and three sons from his two marriages. In time (when he dies), he says, these DUATs will be passed on to members of his matrilineal nloko. However, he also has land that has not been registered. In order to provide for his wife (who is viente), he has talked to members of his nloko to get their acceptance that specific parts of this non-registered land can be passed on to his (living) wife and his daughters, while he expects his sons to cultivate their wives’ nloko land.

In Canacura, Avelina (see above) is married to a man from a distant community. The land they cultivate together was registered in her name. She is, however, one of those still waiting for the DUAT documents. She has five living children (of 12 born), two daughters and three sons. The two daughters are married, and each has their own DUATs. The sons do not have DUATs in Canacura. One of them married a woman in the community where Avelina was born, and “followed his wife” to live on her nloko land. At the time, Avelina’s parents were still alive, but when they died, her son took over the land his maternal grandparents had cultivated in that community. What will happen with this land in the generation of Avelina’s granddaughters is still an open question.

Daniel (see above) tells that most of the land belonging to his mother-in-law’s nloko was registered during the titling process in Canacura. In total there were four DUATs issued for his wife’s matrilineal nloko. One DUAT was issued in the name of his mother-in-law, one in the name of Daniel’s wife, and two DUATs for two (of her three) sisters. There was also a land titling process in the community where Daniel was born, and he got a DUAT there within his mother’s nloko land. But that land just “rests” as a reserve while he and his wife together cultivate her nloko land in Canacura.

Table 1. Interviewed men and women: Basic findings.

Interviewees	Number Inter-viewed	Living on Own Matrilineal Nloko Land	Living on Spouse’s Nloko Land	Divorced/ Single	Polygamous Union	DUATs in Community of Residence	Land Registered with Title Holder Waiting for DUAT
Men	10	7	3		3	7	2
Women	9	9 *	0	3		7	1
Total	19	13	3	3	3	14	3

* Number includes Mariana and her daughter Elina (based on Elina’s account above).

summarises my findings across all 19 interviewees. Among my interviewees, seven of nine women and six of ten men lived on and cultivated their own nloko land. Three men and no women lived on the spouse’s nloko land. The relatively low proportion of men living on the wife’s nloko land (3 of 10) probably reflects the high proportion of men in leadership positions among the interviewees. These are men who tend to live on their own nloko land.

Two of the nine women were divorced, living with their children. One middle-aged woman was single and without children. She was cultivating the “land of her erukulu” to which her (married) sister had a DUAT. Three of the ten men lived in polygamous unions, each with two wives. Of the wives of the ten men interviewed, seven had their own DUATs. Here the wives of the three polygamous men stand out, since two thirds (four of six) of these women had not acquired DUATs during the demarcation processes.

Among the interviewees selected for our study by local leaders, several of the men had leadership positions at some level within the communities. Men with a leadership position in a nloko occupying a lot of land—including the Cabos de Terra in both communities—would usually live on their own matrilineal nloko land. One of these nloko leaders, who was also one of the polygamous men mentioned above, said that his two wives cultivated his nloko land, not their own. When his nloko’s land was registered, he secured several DUATs in his own name, in addition to DUATs for his one sister and six brothers living in the community. But his wives did neither obtain DUATs within their husband’s nloko land, nor in their own nloko land. The reason, he says, was that they were not cultivating the land of their own nloko. However, two of his married daughters were cultivating plots within their mother’s nloko land. This land was registered with DUATs for the plots each daughter cultivates. These DUATs were issued in the name of their living maternal grandparents, but his daughters were confident that the DUATs in due time would be transferred to them as the title holders’ matrilineal granddaughters. In the interviews, most of the male leaders said that it would primarily be their daughters who would inherit their share of their own nloko land, and thus also their DUATs.

The situation regarding land for the wives of the polygamous men in my material seems to indicate stronger elements of a patrilineal/patriarchal pattern. This may be interpreted in terms of diverse norms and practices coexisting in the communities. While my material clearly reveals a dominant pattern of matrilineal norms and practices in the research area, constituting what I have called local matrilineal land institutions, the polygamous men to a larger extent appear to draw upon patrilineal—and patriarchal—norms. It is also well known from the literature on matriliny, including works dealing with northern Mozambique, that patrilineal/patriarchal norms can coexist with matrilineal norms [49].

An important question in relation to the present land formalisation initiatives is to what extent individual titling, with a significant proportion of titles in men’s name, will contribute to fundamental shifts in the balance of entitlements and power between women and men in their access to land. My field material is inconclusive on this point.39 A key issue is to what extent men with DUATs in the future will pass on the titles issued in their name to their matrilineal nlokos for inheritance by female nloko members. To make valid conclusions on transformation in the wake of titling would probably require a longitudinal study—for example along the lines of the research carried out by Pauline Peters in Malawi [32,50,51]. The field material does, however, provide a basis for a further discussion of the nature of Makhuwa land institutions, and the possible implications of formalisation of land tenure through demarcation of individual plots to both male and female members of a landholding nloko.

6. The Character of Makhuwa Matrilineal Land Institutions

The empirical material presented above has, I would say, two types of data. One is revealed in the elements indicating a pattern of land holding based on matrilineal nloko and erukulu membership. However, the recent processes of legal formalisation through individual plot demarcations appear to have resulted in an outcome with men’s names on a significant part (over 50%) of the DUAT titles. Whether such a pattern will be maintained or changed when new generations get access to land, will depend on a complex set of factors and be influenced by processes at both local and national levels. The other type of data in my empirical material is constituted by the explanations and narratives provided in the interviews, the local discourse on land institutions. The dominant narrative in both individual and group interviews is that land is accessed and passed on to new generations through matrilineal nlokos. Women are entitled to get land from their mother’s nloko, while men are expected to get access to land through marriage, and primarily work on the wife’s nloko land. Men who occupy leadership positions in dominant nlokos within the community do, however, tend to live on their own nloko land. Still, it is commonly expected that this land will be passed on to new generations through their matrilineal kin. The pattern of gendered landholding emerging from these institutions and how they shape access to land can serve to explain a key finding in Meinzen-Dick and Zavale’s recent survey-based study of matrilineal land rights and tenure insecurity in two Emakhuwa- and Lomwe-speaking districts in northern Mozambique [52]. Their quantitative study indicates that women in these districts experience less tenure insecurity than men do. In our research communities, however, there is not full agreement among community members on the norms and practices of local land tenure. In some contexts, at least some of the men—both elder and younger—will subscribe to more patrilineal or patriarchal norms. Since the current processes of institutional transformation through legal formalisation are to a limited extent—if at all—informed by a “sensitivity to the unique relationships” characterising matrilineal land institutions, these processes can also be used by men to strengthen their individual rights to land [52] (p. 17).

In my further discussion of the field material below, I will employ key concepts provided by Ribot and Peluso [19], Moore [53], Scott [54], and to Christian Geffray’s work to reconstruct “pre-contact” Makhuwa society in Erati District of Nampula [55]. I will also draw upon Peters and Kambewa [56] and Kingwill [28,57] in my further discussion of possible outcomes of matrilineal land formalisation in Makhuwa communities.

6.1. Elements in the Further Development of a Conceptual Framework

While Ribot and Peluso define “access” as “the ability to benefit from things” [19] (p. 153), I will restrict my discussion here to land as a cultivated or cultivable natural resource. According to Ribot and Peluso, access relations are always changing, “depending on an individual’s or group’s position and power within various social relationships” [19] (p. 158). This means that some individuals or groups are able to mediate others’ access and exercise what Ribot and Peluso call “access control”. Access must be maintained, thus the need for “expending resources or power to keep a particular sort of resource access open” [19] (p. 159). Ribot and Peluso further refer to access mediated by social identity [19] (p. 170), but do not discuss gender in this regard. Here I find it useful to introduce the gendered character of “transferable interest”, a concept used by Sally Falk Moore in her work on a century of “customary law” in the Kilimanjaro region of Tanzania [53]. Moore provides an account of a strongly patrilineal ethnic group where land has been held by men and passed on to their sons. A married woman would live on her husband’s land and could also be allowed to continue living on this land as a widow. If she had no sons, on her death that land would revert to her deceased husband’s localised patrilineage. Moore describes a social setting where “[w]omen cannot hold a transferable interest in land” [53] (p. 200). Based on my material from Ribáuè, I would say that it is precisely women’s transferable interest in and access to land as members of a lineage, descent group or sibling group (erukulu) that stands out as a key defining characteristic of local matrilineal land institutions.

As one of the few authors writing on Makhuwa society, Geffray [55] actually sought to avoid notions of matriliny or matrilocality in his analysis. His empirical material represents an effort to reconstruct social structures of production, distribution and consumption after the whole territory occupied by the Makhuwa in northern Mozambique had finally been “pacified” and come under Portuguese colonial control in the 1920s.40 Focusing on food as a social fact [58], Geffray provides an account of the processes through which the mothers’ authority was created and maintained. He claims that an overriding concern and mechanism of authority was feeding the children of the erukulu—the social unit that in my field area would be referred to as the nloko. In Geffray’s analysis, it is the mother’s mother who gains such authority through controlling the labour and the agricultural output produced by her daughters and sons-in-law [55] (p. 50). According to Geffray, the grandmother would collect a considerable part of the outputs of the nloko members’ subsistence production in her storehouse (nikipirani), and the food cooked in the grandmother’s house would above all be dedicated to the grandchildren’s “small stomachs”. Thus, in Geffray’s analysis, the women’s role in managing the food storehouses and their “power to feed” are key to the social recognition and power of the mothers within Makhuwa society [55] (p. 58).

Geffray’s focus on food production and consumption resonates with what I found to be a paramount interest among my interviewees in the field. It relates to what other authors have discussed in terms of subsistence, with Scott’s The Moral Economy of the Peasant [54] still standing out as a relevant contribution [59]. Based on empirical material from Southeast Asia, Scott describes rural people living “close to the subsistence margin and subject to the vagaries of weather and the claims of outsiders… [with] little scope for the profit maximization calculus of traditional neoclassical economics” [54] (p. 4). He claims that “the need for a reliable subsistence” will be the primordial goal of “the peasant cultivator”, and the social arrangements established imply that “all are entitled to a living…” out of the locally available resources [54] (p. 5). This entitlement, in turn, involves access to a minimum of land. As a moral principle, it constitutes in Scott’s view a right to subsistence. The fact that land is kept as a reserve for sons in case of, e.g., divorce in my field area, I see as constituting one of various mechanisms to secure subsistence for all members of Makhuwa erukulus.

6.2. The Character of Makhuwa Land Institutions in the Field Area

Interviewees in the field site said that before the demarcation process there were discussions over boundaries, as well as emerging conflicts. The individual land demarcations carried out in this situation served to clarify existing claims to specific plots of land within the community, they said. The claims sanctioned by the interested parties summoned to participate in each plot demarcation were, in principle, further strengthened as “enforceable claims” through State recognition and formal documentation in DUAT documents—though a significant number of people have not yet received these DUATs.

My findings, presented above, show that the matrilineal land institutions in the field area in principle provide access to land for women and men who are members of a matrilineal nloko with a basis in the local community. Such membership comes primarily from being born from the (physical) erukulu of a woman who is herself a member of a landholding nloko through matrilineal descent. I would hypothesise that the presence of a minimum of three generations of matrilineal descent is required to form such a localised nloko.41 The nloko thus provides what Ribot and Peluso would in principle call “rights-based access…sanctioned by custom” [19] (p. 161, italics in original).

However, it will be women who, according to local norms and practices, hold the “transferable interests” [53] in land based on their social identity as female members of a nloko. It is primarily the daughters in an erukulu who have access to their mother’s nloko land. Young men are primarily expected to get access to land through marriage, and a man will usually work together with his wife on his wife’s nloko land. They practice what anthropologists call uxori- or matri-locality. However, both men and women said that sons could ask their mother and father for access to land “if they needed it”. In Mariana’s account, this was the case for Mariana and her husband, when they found that there was not enough land for cultivation in the community of her mother’s nloko. The couple thus came to Canacura to settle and, with three generations living in Canacura, in Elina’s account, have formed a (new) matrilineal nloko with Mariana as the “mother of the nloko”. In cases of divorce, men can ask—and expect to be given—access to their mother’s nloko land for cultivation and subsistence. Another option is to move directly in with a new woman with access to her own nloko land, which was the case with Elina’s former husband.

In most nlokos, part of the land nloko members considered to belong to these nlokos had not been registered for individual titling in the 2019–2022 period. One reason was the nloko’s own decision to keep some of the more distant land as a “reserve” so that a plot could later be given to for instance a son who needed it. This opportunity to provide land for sons contrasts with the situation in the Zomba district of the Malawian Highlands described by Peters and Kambewa [56]. In a situation of general shortage of land, they report on few cases of men using land belonging to their own matrilineages, and in these cases tensions building up with matrilineal “sisters” tend to force the men to “to move to their wives’ villages or elsewhere” [56] (p. 454). In fact, Peters’ longitudinal study in the area shows that most of the recorded disputes turn on land, both “within matrilineal groups as well as between them…” [56] (p. 454). The communities of Muthoko and Canacura in Nampula, where fieldwork was carried out for this study, can even for the context of northern Mozambique be characterised as “remote”, with difficult access by road and apparently limited pressure on land to date. Local people often said that there was “enough land”. The situation in this respect contrasts with the conditions in the southern Highlands of Malawi as described by Peters [32,50] and Peters and Kambewa [56]. In southern Malawi there is high pressure on available land for cultivation.

The demarcation processes in my field area were informed by principles of gender equality—and gender mainstreaming—assuming that men were in practice controlling access to land. In the name of equality, individual land tenure formalisation provided an opportunity for both men and women who at the time were cultivating nloko land to have plot(s) of land registered in their name. Through our study we could only get some glimpses into the deliberations that had been taking place within the nlokos and between nloko members and external service-providers during the demarcation processes. The outcome seems to be that women’s “transferable interests” in land as nloko members giving birth to new erukulus in the demarcation process was not always sufficient to obtain DUAT documents. The findings presented above indicate that a further requirement could be that a woman also cultivated nloko land in order to have her access rights sanctioned by the State—with other nloko members’ consent—in a formal DUAT document.

A DUAT document should in principle imply a strengthening of a woman’s individual enforceable claims, both in the local setting and in a national legal context. To what extent this will be the case remains to be shown for women holding transferable rights in land as members of matrilineal nlokos. In part it depends on men’s access to land being formally sanctioned as “enforceable claims” through individual DUATs in this context. Here, I believe we see what may turn out to be a fundamental shift. That is, the difference in the nature of men’s and women’s access to land provided through local matrilineal land institutions has, in practice, been formally removed through the legal formalisation of individual titles for both men and women. Peters holds that: “The sets of ideas and practices constituting and enacting matrilineal organization… present a particular definition of reality for people…” [18] (p. 138). With individual DUATs, men and women in Muthoko and Canacura have secured the Mozambican state’s recognition of their Constitutional “equal rights” to land, and in principle equal tenure security for their plots with DUATs issued in their name.

Finally, it appears that the individual titling processes in the two communities provided new spaces for what Ribot and Peluso call access control [19]. As documented in an earlier study from central Mozambique [41], such land registration processes tend to open new opportunities for male community members, in particular. Men often have more experience in relating to external actors and are often better at presenting their concerns and claims in ways that resonate with external, male-dominated, interlocutors. Thus, the resource of enforceable recognition becomes mediated with a male bias under externally mediated institutional change—a finding that resonates with experiences elsewhere on the continent (German 2022). In the field area of the current study, we learnt that senior male members of a nloko will be consulted and can also have decision-making power in land-related questions. Though this “depends on the person”, people said.42 Such male authority within the matrilineal group (erukulu or nloko) appears to be based on personal attributes to a greater extent than what has been reported from matrilineal groups in Malawi [32]. Nevertheless, male authority may have come into play to exploit the indeterminacies of the situation in the demarcation processes, resulting in a redefinition—and strengthening—of men’s entitlements. Thus, this form of situational access control may have resulted in weakening “the particular definition of reality” [18] inherent in matrilineal land institutions, and contributing to undermining women’s prioritised access to transferable rights in nloko land as mothers of erukulus.

7. Conclusions

This article set out to describe matrilineal land institutions in Makhuwa communities in northern Mozambique in the context of initiatives to formalise community land rights through individual titling (in DUAT documents). Based on data collected through fieldwork with qualitative interviews in two fairly remote rural communities in the district of Ribáuè in the interior of Nampula Province, I have given an account of what I consider to be some of the key characteristics of matrilineal norms and practices related to land in the research area. I argue that local matrilineal institutions have provided women with transferable access and rights to land. Men have access land in various ways within this institutional framework but are primarily expected to get access through marriage and then work in their wives’ fields. In 2015, when the Mozambican government with World Bank support launched its large-scale land tenure formalisation programme, Terra Segura, the officially expressed goal was to strengthen land tenure security through legal formalisation of land rights, prioritising land titling with individual DUATs. While the programme operated with a gender target of 40% of DUAT documents issued to women, it had no reference nor any approach to take matrilineal land institutions into account in individual land titling processes.

The findings reported here indicate a general perception of increased tenure security in the two research communities as a result of community land delimitations carried out before and under the umbrella of Terra Segura. These findings are in line with results from Meinzen-Dick and Zavale’s recent survey-based study [52]. Community members in our field research area who had received individual DUATs also said they believed the title documents would strengthen their individual land rights. However, as Ribot and Peluso emphasise, rights-based access must be maintained, often requiring that rights-holders expend resources or power to keep their resource access open [19] (p. 159). Local people—both men and women—in our research area basically lack such resources or power, especially in potential confrontations with more powerful actors from outside the community. Non-registered nloko lands “resting” or kept as reserves for family members who need it at a later stage may be particularly vulnerable to external actors’ interests—or “grabbing”—in a local context where much of the land is formally registered with individual titles. In the context of a national neoliberal political economy supporting secure access to land for large-scale investments, legal formalisation through individual titling of parts of a community’s delimited land may in practice serve to make the rest of the land in this community more easily accessible to outside actors and commercial interests.

In the field area, I experienced an emphasis on subsistence production with a great variety of foodstuffs produced for household consumption. Geffray’s analysis provides an approach to analyse kinship arrangements working to secure what Scott calls “rights to subsistence” [54]. As social arrangements, these span several generations with the women’s responsibility and power to feed covering both children and grandchildren. In this perspective, the nloko has served to maintain matrilineal members’ access to land, to redistribute access to land in succeeding generations, and to provide a basis for subsistence for all members. In Nampula Province today, despite government priorities and World Bank projects seeking to promote commercial agriculture, living in rural communities for most people still requires a basis in subsistence agriculture and the social support provided through kinship. In this situation, tenure security in the form promoted and implemented through individual titling may serve to secure a basis for subsistence production in situations of increasing pressure and thus higher risks of internal conflicts over land, as described by Peters in Malawi [50,56]. However, it is perhaps even more likely that individual titling can serve to undermine an entire set of lineage-based support provided by nlokos and erukulus.

During the processes of legal formalisation of local land, community members worked with external project implementors who no doubt had knowledge of the legal package related to the Mozambican Land Law. As I have sought to document in this article, formal knowledge of the complex web of norms, rights, and relationships constituting matrilineal land institutions has been notoriously absent in Mozambican State contexts—starting in colonial times and continuing throughout 50 years of independence. Thus, knowledge about women’s roles as mothers and grandmothers and central actors in a moral economy of subsistence security cannot be expected to be part of the repertoire managed by the service provider agencies implementing Terra Segura. Matrilineal land institutions neither appear to resonate with current policy priorities in Mozambique, nor with the new land governance orthodoxy promoted by the World Bank, despite its proclaimed gender focus.

In the research area, legal formalisation through individual titling has in my view at best provided a one-dimensional form of tenure security. This has involved “freezing” the current distribution of access and land use in the communities into formalised individual DUATs, which to a considerable extent also formalise the present distribution of power and resources within the communities. We may see that nlokos in the future continue to regulate members’ relationship and access to land in terms of the customary concepts and norms of kinship and descent, including when it comes to managing land with individual titles, as Kingwill reports from her research area in South Africa [28] (p. 215). In our research area, however, when the difference in the nature of men’s and women’s access to land characterising local matrilineal land institutions has been formally removed through the legal formalisation of individual titles, this has most likely also weakened the definition of reality inherent in these institutions. The individual DUATs themselves come without the generational and collective orientations of the traditional institutions, where nlokos are expected to provide a minimum of livelihood opportunities and subsistence security through access to land for cultivation for all members of the matrilineage.