Sociolinguistic ISSN: 1750-8649 (print) Studies ISSN: 1750-8657 (online)

Article

In the name of the father-in-law: Pastoralism, patriarchy and the sociolinguistic prehistory of eastern and southern Africa

Luke Fleming, Alice Mitchell and Isabelle Ribot

Abstract In a range of eastern and southern African language communities, stretching from Ethiopia to the Cape, married women are enjoined to avoid the names of members of their husband’s family as well as (near-)homophones of those names, and to replace tabooed vocabulary with substitute words. Although in-law name avoidance is a global phenomenon, the daughter-in-law speech registers thus constituted are unusual in their linguistic elaboration: they involve avoidance not only of names and true homophones of names but also an array of words whose only relation to tabooed names is phonological similarity. We provide an overview of the distribution and convergent social and linguistic characteristics of these registers and then examine one register more closely, namely, that of Datooga of . To tease apart the layers of causality that converge upon this particular sociolinguistic pattern, we consider archaeological, ethnological, sociolinguistic and genetic lines of evidence. We propose that any partial diffusion of in-law avoidance practices has been complemented by a complex of sociocultural factors motivating the emergence of this pattern at different times and places across the African continent. These factors include pastoralism, patrilineal descent ideologies and norms of patrilocal post- marital residence paired with cattle-based bridewealth exchange.

KEYWORDS: DAUGHTER-IN-LAW REGISTERS, NAME TABOOS, SOCIOLINGUISTIC PREHISTORY, HLONIPHA, KINSHIP, SOCIAL PRAGMATICS

Affiliation

Université de Montréal, Canada email: [email protected]

Universität zu Köln, Germany email: [email protected]

Université de Montréal, Canada email: [email protected]

SOLS VOL 13.2-4 2019 171–192 https://doi.org/10.1558/sols.37860 © 2020, EQUINOX PUBLISHING 172 SOCIOLINGUISTIC STUDIES

1 Introduction

In-law avoidance is a widespread practice cross-culturally, often associated with avoidance of physical contact, eye-contact or the direct transfer of possessions. It also commonly has sociolinguistic correlates. The most conspicuous and widespread of these is name avoidance and taboo. In this paper we look at the distribution, in eastern and southern Africa, of linguistic registers built up around such practices of name taboo. In language communities throughout this region, daughters-in-law are proscribed against uttering the names of their fathers-in-law. Furthermore, this name taboo often conditions avoidance of lexical forms iconic of the name of the father-in-law and other senior kin of the husband. These daughter-in-law registers are found across a wide geographic expanse from Ethiopia to the Cape, in languages representing all three of the largest African language families – Nilo-Saharan, Niger-Congo and Afroasiatic. In an important article on the Kambaata daughter-in-law ballishsha register, Treis (2005) underlined the striking parallels between the Eastern Cushitic and Southern Bantu daughter-in-law registers:

Why is such a similar type of ‘respect through avoidance’ (Herbert, 1990:455) practiced by women geographically so far apart? Kambaata society is patrilineal and strongly patriarchal, marriage is virilocal and forbidden among members of the same clan, and a married woman and her in-laws live in close proximity – i.e., there are major similarities [between Kambaata and] Nguni society. But these social similarities surely cannot be considered a sufficient reason for the parallels […]; otherwise, one would expect this special type of word taboo to be attested in many more languages and societies. (Treis, 2005:317–318)

The similarities between Eastern Cushitic and Southern Bantu daughter-in-law registers could be understood in at least three ways: (1) as surface similarities attributable to divergent sociocultural, demic, ecological, etc. causes; (2) as convergent (or ‘analogous’) sociolinguistic forms arising from homologous nonlinguistic factors; (3) as ‘derived’ sociolinguistic patterns resulting from horizontal cultural transmission. In this paper we seek answers to Treis’ question by drawing on data from all four anthropological subdisciplines. By using archaeological, ethnological, sociolinguistic and genetic lines of evidence we begin to tease apart the layers of causality that converge upon this particular sociolinguistic pattern. Though this article is question-driven, we also hope to make a methodological contribution: our exposition illustrates that sociolinguistic registers can be used to reconstruct information about social organization and migrations in the past, suggesting that sociolinguistic reconstruction should be

IN THE NAME OF THE FATHER-IN-LAW 173 integrated into the toolbox of linguistic evidence like cultural vocabularies, relationships and linguistic substrates that archaeologists employ to reason about prehistory. The article begins with an overview of registers built out of name and homophone avoidance in the African context. It then presents a case-study of the Datooga daughter-in-law register, based on field research by Mitchell, to exemplify name and homophone avoidance as a contextualized social practice. Finally, it seeks to probe diffusionary and convergent sources of similarities in sociolinguistic patterning by drawing on archaeological and genetic evidence concerning human migrations and cultural diffusion in eastern and southern African prehistory.

2 Name taboos and in-law avoidance registers

In-law avoidance is practiced all over the world, and in many places is associated with linguistic markers. In some languages an alternation between honorific and non-honorific pronouns is keyed to particular affinal relationships. For instance, in Djaru (Pama-Nyungan) the plural functions as an honorific in 2nd and 3rd person pronominals, but its use is restricted to avoided in-laws classified as majili (‘mothers-in-law and their siblings’) (Tsunoda, 1981). In many Australian com- munities, large lexical repertoires (so-called ‘mother-in-law languages’) substitute for everyday words proscribed in referring to, addressing or while co-present with, certain categories of affines (Dixon, 1990; Haviland, 1979). But the most common sociolinguistic marker of in-law avoidance relationships cross-culturally is name avoidance. Name avoidance, like the avoidance of touch or mutual eye- gaze – two other classic nonverbal emblems of in-law respect relationships – seems often to be experienced as an avoidance of direct contact with the taboo relation, as Stasch (2003) argues for Korowai (Papuan) mother-in-law name taboos. Name-use and -avoidance are often enregistered as an honorific distinction indexing deference-entitlements of the name-referent relative to the speaker. Cross-linguistically, junior kin are often enjoined to address seniors with kin terms, eschewing personal names (Fleming and Slotta, 2018). What makes in-law name taboo distinctive with respect to widespread practices of respect-based name avoidance is that it is not restricted to person reference. In languages across the world, in-law name avoidance is linked to word tabooing that extends beyond names to affect nouns, verbs and adjectives whose only relation to the name is their similar phonological shape. Most famous from the African continent are the hlonipha registers of southern Africa (Zulu [Luthuli, 2007]; Xhosa [Herbert,

174 SOCIOLINGUISTIC STUDIES

1990; Finlayson, 1995]; Ndebele [Davies and Quinche, 1933]). In these languages, in-marrying women avoid the names of ascending generation agnatic kin of their husband, as well as a large repertoire of words similar in sound to those names. Teferra’s (1987) description of the Sidamo in-law avoidance register, part of the Eastern Cushitic ballishsha-register complex, exemplifies how the pragmatics of name and homophone registers exceed the bounds of person reference (see Table 1).

Table 1. Degrees of linguistic avoidance in Sidamo (source: Teferra, 1987:53–55).1

Origo – Target homophones tokens of the pronominal of the avoidance index of the name name reference

♀ – husband’s father * * V

♀ – husband’s mother; ♀ – husband’s elder sibling - * V

♀ – husband - * T

♀ – husband’s younger sibl. - - T

Different in-law relationships condition different degrees of linguistic avoidance: a wife is not expected to avoid her husband’s younger siblings’ names, but she does avoid her husband’s name. Importantly, this avoidance is not limited to those speech acts where she refers to her husband. Teferra provides the following example:

My father’s name is Teferra and I have a friend who is named Teferra Eshete, who was a student in the University. Whenever my step-mother wanted to refer to my friend, she spoke of him as ‘Eshete’s son who is in the University’. This is because she did not want to mention a name identical to that of her husband. (Teferra, 1987:58)

We can properly call this a name taboo, since the name is avoided in all contexts regardless of its discourse reference. What is the proper characterization of the social pragmatics of names in cases like these? In some dialects of American English, license to employ a first name indexes the speaker’s intimate relationship to the name’s referent. In Sidamo, however, the person honoured through name avoidance needn’t be the referent or the addressee, or even a bystander, but rather a nonparticipant in the interaction. Since the target of deference is another possible referent of the name Teferra, homophone avoidance may be conceptualized as involving a kind of ‘allusive reference’ (Mitchell, 2018).

IN THE NAME OF THE FATHER-IN-LAW 175

Returning to Table 1, we see that linguistic avoidance in the case of the husband is limited to tokens of the name type Teferra. With respect to the father- in-law’s name, however, married women avoid not only the name but also other lexemes whose phonological form is iconic with the name. The first syllable of the name has particular salience in Eastern Cushitic homophone avoidance such that, for instance, the name Marufa conditions avoidance of the lexeme matʔine ‘salt’, which is substituted by a special avoidance form, tʔeʔame ‘salt’ (Teferra, 1987:47). Here again, we must analyze the honorific function of tʔeʔame as indexing respect towards a nonparticipant (i.e. the speaker’s father-in-law). Speakers are expected to employ honorific substitutes in all contexts regardless the referent or the identity of participants in the interaction. That these are indeed honorific distinctions is evident not only from explicit ethnometapragmatic reports, but also in the co-occurring use of polite (‘V’-form) pronouns for those individuals whose names are subject to most disciplined avoidance.

2.1 Distribution of name-based in-law registers in Africa and beyond In Table 2 we present a list of attested cases of name and homophone avoidance registers in Africa. The rightmost column indicates the avoidance relationships that condition the use of these registers (e.g. ‘♀ → HF’ means that a woman avoids the name and homophones of the name of her father-in-law). As the table illus- trates, there is remarkable consistency in the patterning of these sociolinguistic registers in terms of the kinship relationships around which they are organized. Elsewhere in Africa, name and homophone registers do occasionally occur, although in those cases they are not (exclusively) associated with in-law avoidance. In Madagascar, name and homophone avoidance was practised with respect to deceased individuals and paramount rulers (Dez, 1965) (see also Huntingford [1969:116] on Maasai post-mortem name and homophone avoidance). In some languages of northeastern Nigeria (the Adamawa languages Longuda, Tso and Cham and the Chadic language, Tangale), names of deceased individuals were historically tabooed along with homophones (Kleinewillinghöfer, 1995). The closest parallel case to the in-law avoidance registers of eastern and southern Africa is that of central Fulani as described by Kintz (1986:38); names and homophones of the names of the mother, father, eldest son and daughter, mother-in-law, father-in-law, husband and wife are all avoided. Even though in- laws’ names are affected, the pattern diverges from that attested in eastern and southern African languages: there is no gender biasing (i.e. both men and women practice avoidance, and the names of males and females are equally subject to avoidance) and, more importantly, it is the names of consanguineal rather than affinal relatives that are most subject to avoidance.

176 SOCIOLINGUISTIC STUDIES

Table 2. In-law name and homophone avoidance registers in Africa.2

Language Family Location Source Kin relation Kambaata Highland E. Ethiopia Treis, 2005 ♀ → HF, HeB, HM Cushitic, Afroasiatic Sidaama Highland E. Ethiopia Teferra, 1987 ♀ → HF, HeB Cushitic, Afroasiatic Haddiya Highland E. Ethiopia Adane, 2014 ♀ → HF Cushitic, Afroasiatic Oromo Lowland E. Ethiopia, Mbaya, 2002 ♀ → HF, HeB Cushitic, Kenya Afroasiatic Kerewe Great Lakes Lake Victoria Komori, 1999 ♀ → HF Bantu, (Kenya/Tanza Niger-Congo nia) Kamba Northeast Mt. Kenya Lindblom, 1920 ♀ → H, HF, HeB, Bantu, (Kenya) HM, HeW Niger-Congo ♂ → WM, WMeD3 Nyakyusa Corridor Tanzania Kolbusa, 2000 ♀ → HF, HFB, Bantu, HFBW, HM, HMB, Niger-Congo HZ Datooga Southern N Tanzania Mitchell, 2015 ♀ → HF, HFF, Nilotic, HFFF, HFB, HFFB, Nilo- HFFFB, HM, HMZ, Saharan HFZS Ndebele Nguni, South Africa Davies/Quinche, ♀ → not specified Bantu 1933 S. Sotho Nguni, South Africa Kunene, 1958 ♀ → HF Bantu Zulu Nguni, South Africa Raum, 1973 ♀ → HP, HFF, Bantu HFSib Xhosa Nguni, South Africa Herbert, 1990; ♀ → HF, HFB, Bantu Finlayson, 1995 HM, HFF, HFFF

Beyond Africa, name and homophone registers are broadly if sparsely attested. This sociolinguistic pattern is most widespread in Austronesian and Papuan languages of eastern Indonesia and Melanesia (Simons, 1982). There are also attested cases from Asia and the Americas (see Fleming, 2014). In comparing the African daughter-in-law registers with other affinal name and homophone registers, one notices the strongly asymmetric, gender-stratified and generationally stratified character of avoidance in eastern and southern Africa. Women avoid

IN THE NAME OF THE FATHER-IN-LAW 177 men, but not vice versa, and junior affines avoid senior affines, but not vice versa. (The one exception is the case of Kamba, discussed below.) This gendered and generationally stratified pattern is associated with strong patrilineal descent ideologies, norms of patrilocal residence and agro-pastoralist economic systems. Similarly gendered patterns of avoidance are also found in patrilineal societies in Eurasia. In the areally connected Mongolian, Oirat and Kazakh language communities, women employ (or historically employed) name and homophone registers keyed to ascending generation affinal males. And in South Asia, avoidance of the name of the husband is the most conspicuous name avoidance practice (see Fleming, 2014 for references). Nonetheless, nowhere is this asymmetrical avoidance pattern so elaborated as in the African context.

2.2 African daughter-in-law registers: Sociolinguistic patterning The contemporary understanding of register is as ‘a linguistic repertoire that is associated, culture internally, with particular social practices and with persons who engage in such practices’ (Agha, 2000:216). Avoidance registers offer an interesting case for this formulation. Since different individuals’ in-laws have different names, speakers are obliged to avoid different words – what we call the NEGATIVE REPERTOIRE of the register. Though elements of the negative repertoire are phonologically iconic with the names of taboo relations, the set of forms which individuals avoid is not based upon absolute criteria of phonetic similarity but is conventionally determined (Mous, 2001). The negative repertoire is different for different speakers, even though each individual uses ‘the same’ register (e.g. Kambaata ballishshaa, Xhosa hlonipha). Name and homophone avoidance register repertoires are, in this sense, idiolectal (Fleming, 2014:123). Name and homophone avoidance registers vary widely in the range of forms that speakers avoid. The most minimally invasive kind of avoidance is found in languages spoken in the Brazilian Amazon – the Carib language Kalapalo and the two Arawakan languages, Wauja and Mehináku. Among these groups, names come from everyday words for ‘animals, household objects and natural events’ (Gregor, 1977:284), and speakers avoid only the true homophone of the name, e.g. Yapu ‘Stingray’. African daughter-in-law registers, contrastingly, represent a maximally extended pattern of linguistic avoidance. Indeed, these registers are notable for the sheer range of everyday lexemes that are subject to avoidance. Here it is not only homophones but a range of lexemes phonologically iconic with the name of the taboo relation that are affected. In Kambaata and Sidamo, daughters-in-law avoid words that have the same first syllable as the taboo names. In Xhosa and Zulu, iconicity between any syllable in the name of the taboo relation and another lexeme may be sufficient to condition avoidance. Married

178 SOCIOLINGUISTIC STUDIES women are thus expected to avoid a large negative repertoire of words phonologically iconic with the name(s) of their elder in-law(s). The substitutes for these avoidance targets constitute the idiolectal avoidance repertoires.

2.3 Gender and the differential enregisterment of avoidance forms A distinctive feature of daughter-in-law avoidance repertoires is their differential enregisterment as avoidance forms. This differs from many languages with name and homophone avoidance registers where substitutes for tabooed lexemes are simply synonyms that occur in the everyday talk of other speakers. For example, Alune (Austronesian; Seram) speakers normatively avoid the names and homophones of the names of all of their direct in-laws (Florey and Bolton, 1997). Avoidance substitutes are everyday words distorted in their semantic extensions. So, for instance, ala ‘rice’ is replaced by ume ‘sand’ in the idiolect of a woman whose sister-in-law is named Alaya; ni’wele ‘coconut’ is replaced by wa’ile ‘dried remnant of coconut’ in the idiolect of Ni’wela’s in-law. Avoidance substitutes are not exclusive to avoidance repertoires, but are simply everyday lexemes employed creatively by particular speakers. In the African daughter-in-law registers, the substitute avoidance vocabulary is often specific to the speech of married women – they are differentially enregistered. In Kambaata, Treis elicited 48 words making up this ‘core’ set of terms (2005:300). All but one are nouns, and the ‘semantic fields [center on] terms for food and drink (…), enset (…), body parts (…), and household goods’ (2005:301). The use of these forms unambiguously indexes the speaker’s marital status and gender identity. Diachronically, male speakers may have avoided these forms, a process which catalyzed the development of speaker-focal values of avoidance repertoire items as indices of married women. This sociopragmatic markedness of avoidance forms informs and is informed by speakers’ metapragmatic (or ideological) conceptualization of register variation; these registers are ubiquitously characterized in terms of speaker attributes as ‘women’s speech’ (e.g. Xhosa isihlonipho sabafazi ‘women’s respect language’ [Finlayson, 1995]). This gendered conceptualization of avoidance speech simultaneously links up with the gender and generational stratification that characterizes the social matrices within which these registers are used – that is, their use by generationally junior women as emblems of respect for generationally senior men. In some Southern Bantu speech communities, married women avoid certain everyday words with no phonological likeness to taboo names, instead employing a shared repertoire of substitute forms. These sociolectal repertoires are biased towards the same feminine-gendered semantic domains as the core set of ballishsha alternants in Kambaata, suggesting that differential enregisterment may

IN THE NAME OF THE FATHER-IN-LAW 179 be a diachronic pathway for sociolectal repertoires. Xhosa varieties spoken around Cape Town are attested as having at least 55 ‘core’ hlonipha terms, 19 of which involve foods or the activity of eating (Finlayson, 1995:146–147). At the time of Finlayson’s study, idiolectal repertoires were no longer being employed, suggesting that the register system is shifting away from name-based tabooing towards a shared honorific vocabulary. Luthuli’s questionnaire eliciting Zulu hlonipha forms (2007:9–10, 48–49) produced a list of 34 terms. Again, these terms relate to feminine-gendered activities (‘spoon’, ‘broom’, ‘skirt’, etc.). The semantic domains targeted by these differentially enregistered forms in Kambaata, Xhosa and Zulu appear to relate to the social conceptualization of avoidance practice. Avoidance speech is thematized not only as a relational index of affinal respect but also of speaker gender and marital status. Avoidance speech thus functions not only as a first-order respect-indexical – it has also developed a second-order indexical function of marking speaker identity (Silverstein, 2003). As Treis observes with respect to the Kambaata core vocabulary, ‘[f]ood preparation and enset processing are typical female activities among the Kambaata and this is probably why words from these semantic fields are dominant’ (2005:301). The differential enregisterment of ballishsha or hlonipha vocabulary items drawn from these domains serves to iconically endow social indexical form with stereotyped cultural values. The semantic selectivity of those elements of register repertoires that are exclusively employed by married women essentially provides a set of ‘predicates’ that characterize the individuals indexed by use of the speech register. There seems to be strong motivation for native speakers to telescope pragmatic function (the indexing of married women) and semantic sense (the denotation of domesticity) in explicitly reflecting on these registers. This tendency seems to have channeled the development of differentially enregistered repertoires (Kambaata, Zulu) and of sociolectal repertoires (Xhosa). In sum, the gendered and generationally stratified character of linguistic avoidance is a common feature of in-law name and homophone avoidance throughout eastern and southern Africa, being reflected not only in the patterning of language variation but also in linguistic ideologies and the semantic composition of register repertoires. There is, however, one register that is not exclusively employed by married women to index respect towards senior male affines – that of Kamba (Bantu; Kenya). Though women do avoid the names and homophones of the names of their husband’s kin, Lindblom (1920) states that the most constrained mosi ‘in-law avoidance’ relationship is actually between a man and his mother-in-law. In Kamba, name and homophone avoidance are not primarily conceptualized as women’s linguistic labour.

180 SOCIOLINGUISTIC STUDIES

The continuities between Ethiopian and southern African registers in terms of syllable-based lexical avoidance and in the differential enregisterment of avoidance vocabularies might suggest a direct historical connection between the two. However, gender and generationally stratified patterns of in-law avoidance are also found in Eurasian societies, suggesting the importance of an agro- pastoralist mode of production, as well as patrilineally biased descent ideologies and marriage practices, in conditioning these sociolinguistic patterns. We discuss this further in section 4, but first present a more in-depth case study of an East African daughter-in-law register.

3 Giing’á wêakshó odà : In-law name avoidance among Datooga

This section provides an ethnographic close-up of the daughter-in-law register of Datooga of Tanzania (see also Mitchell, 2015, 2018). Our intention is to paint a richer picture of these registers for readers unfamiliar with this phenomenon, and to explore the sociocultural context of one particular daughter-in-law register. Our comparative survey relies on normative stereotypes of in-law name avoidance which may diverge from actual sociolinguistic practice: this section serves as a counterbalance by discussing women’s everyday avoidance practices. Traditionally semi-nomadic pastoralists, today’s Datooga are agro-pastoralists, typically growing maize alongside their main occupation of cattle herding. The Datooga language is the southernmost member of the Southern Nilotic family and is spoken by around 140,000 people (Muzale and Rugemalira, 2008). The ethnic category ‘Datooga’ comprises numerous ethnic subgroups, most of which speak their own dialect. Although in-law name avoidance is found in all subgroups, details differ from group to group. This section focuses on the Gisamjanga and Barabaiga subsections of Mbulu District, Manyara Region. The daughter-in-law register described here is perhaps one of the last to still be in vigorous use in Africa, although there are already clear signs of shift as educated and Christian Datooga women abandon the custom.

3.1 Kinship, marriage and in-law relations A major organizing principle of Datooga society is patrilineal descent; a child is born into their father’s clan and retains life-long membership. As Klima (1970:39) puts it, clans function as ‘mutual-aid societ[ies]’. Clan membership also determines whom one may marry; rules of clan exogamy prohibit marriage between individuals whose ancestors, going back three generations, descend from the same clans. People determine their relationships to acquaintances by naming their male ancestors, though they may also refer to maternal links. Patrilineal genealogies are a basic component of social knowledge: most eight-year-olds can

IN THE NAME OF THE FATHER-IN-LAW 181 give the name of their father, father’s father, father’s father’s father and of their clan. Patrilineality is evident in the inheritance of both property and magico- religious power: cattle are transferred from father to son; the role of spiritual chief among Datooga (ng’ùttámìida) passes from father to (eldest) son; and healing abilities (qáwóoda) are inherited through the male line. The daughter-in-law register thus operates in an environment in which the patriline is of major sociocultural significance. Datooga post-marital residence is patrilocal. A man’s first wife is brought to live in her father-in-law’s compound, where she will stay until her husband has the means to build his own. When a bride first arrives at her father-in-law’s compound, she is expected to stay completely out of sight of him. Soon after her arrival, a ceremony called the bíiyúudá èanóoga ‘feeding of the milk’ takes place, in which the bride is offered livestock from her husband and his relatives and given new ‘domestic’ and ‘marital’ names. The next day, her father-in-law presents her with the déedá gíiwálòoda ‘cow of respect’ (or, literally, ‘of fear’), an event that relaxes the prohibition on co-presence. All other proscriptions of the avoidance relationship continue to hold, however: no eye contact; no contact with his possessions; no sitting on his bed; no eating in his presence; and no uttering of his name and homophones of the name. In turn, a man avoids the birth name of his daughter-in-law, calling her by her ‘domestic’ name instead. If a woman gets divorced, which is not uncommon, she will continue to practice name avoidance for her ex-husband’s family only if she has children from that marriage. This is one of several pieces of evidence relating avoidance practices to the protection of the descendants of a patriline in Datooga thought.

3.2 The Datooga daughter-in-law register Upon marriage, a Datooga woman is expected to avoid the names of her husband’s classificatory mothers and fathers, as well as his male ancestors of at least one ascending generation. Figure 1 diagrams the main categories of kin subject to name avoidance. Women married into Bajuuta, an important healing clan, avoid up to five generations of their husbands’ patrikin. Name avoidance targets a person’s birth name, though women also avoid the ‘marital names’ of their mother-in-law and her co-wives. (The marital name is a name given to a bride at the bíiyúudá èanóoga ceremony described above and used exclusively by her husband and his brothers.) Unlike Kambaata avoidance, there is no ritual that marks the onset of avoidance among Datooga. A woman will be informed of the important names of the household by her mother-in-law and/or senior co-wives (who use the avoidance equivalents of the names), and she will learn to avoid other names through observation or by correction.

182 SOCIOLINGUISTIC STUDIES

Figure 1. Kin relations avoided by Datooga married women.

Datooga daughter-in-law registers involve avoidance not only of in-laws’ names, but also of the lexical roots and near-homophones of these names. For example, Gídábárda’s daughter-in-law will avoid: (i) his name; (ii) the lexeme bárda ‘knife’ from which the name derives; (iii) other lexical items containing the syllable bar, such as the verb bar ‘beat’ and the noun bàréanyèeka ‘journey’. Words beginning with ba- may sometimes qualify as near-homophones; as discussed in Mitchell (2015), there is some degree of idiosyncrasy in decisions about what constitutes a near-homophone. In place of these taboo words, Gídábárda’s daughter-in-law will use words from the conventionalized avoidance repertoire. This vocabulary is derived by means of various linguistic strategies. The most common is lexical substitution, e.g. saying ‘loosen’ instead of ‘open’. Less common strategies include consonant replacement, e.g. dápta for bárda ‘knife’, and borrowing, e.g. sátàyda ‘knife’, borrowed from the neighbouring . A set of words also appear to have been derived uniquely for purposes of avoidance (i.e. they are differentially enregistered). The Datooga call this elaborate name-avoidance practice gíing’áwêakshòoda. Speakers unfailingly explain gíing’áwêakshòoda in terms of múréeda ‘respect’: one must pay respect to one’s husband’s forebears, and especially his father.

IN THE NAME OF THE FATHER-IN-LAW 183

Avoidance is bound up with ideas about the link between morality and procreation, as mediated by the spirit world (Blystad, 2000). By showing respect to her affinal kin, a woman earns the blessings of the spirits, who then protect her during pregnancy and childbirth. The ideological link between avoidance and procreation is apparent in beliefs about the consequences of violating the name taboos: women say that repeated failure to avoid will cause problems in pregnancy and childbirth and will also result in the hánáng’wéenda – the sacred leather skirt of Datooga women, strongly associated with fertility – catching fire. Women are expected to practice name avoidance throughout their married lives, regardless of interactional context. Revealingly, the single exception to this is during a difficult childbirth, when women are encouraged to call out the taboo names of their husband’s ancestors. In these life-threatening circumstances, one woman explained, women may call on the aid of the ancestral spirits of their husband’s clan. Datooga characterize gíing’áwêakshòoda as a permanent feature of married women’s speech, and by and large women do adhere strictly to the name taboos in practice. Avoidance words often come to entirely replace their ordinary equivalents in women’s active lexical repertoires. Nonetheless, Mitchell observed deviations from speakers’ normative ideals in everyday linguistic practice. One way in which actual use of gíing’áwêakshòoda deviates from reported usage is in variability in the range of near-homophones avoided for different kinship relations. In some households, more words are avoided on account of an actual father-in-law’s name than as a function of the names of more distant relations. The degree of name avoidance is thus an iconic-index of the deference entitlements (Agha, 1993) owed different categories of kin. A second dimension of variation is the extent to which women use avoidance words to replace ordinary words. Evidence from a conversational corpus shows that, in many cases, a taboo lexeme is avoided with total consistency, although words considered taboo do get uttered occasionally. An example of this is given in (1), where a middle-aged woman utters both the ordinary word qéeda and the avoidance word múrgéenda for ‘house’ in succession (avoidance words are highlighted in bold):

(1) náhíittá qéedà (.) náhíittá múrgéenda ear house ear house ‘news of the house, news of the house’

Here, the short pause followed by self-repair suggests that the speaker noticed this particular ‘slip’ in avoidance.

184 SOCIOLINGUISTIC STUDIES

From a normative perspective, the avoidance vocabulary appears exclusively in women’s speech. However, avoidance language is occasionally used humorously by speakers who do not fall into the category of married woman. One morning, a young man walked past Mitchell’s host compound, and, as is customary, called out a greeting to the people inside. Someone returned his greeting, and he replied qámádìiw, an emblematic avoidance form for the ordinary greeting phrase qwásìindà. This response was interpreted by his interlocutors as a joke.

4 The prehistory of eastern and southern Africa: Sociocultural, genetic and archaeological evidence

Having summarized the distribution of daughter-in-law registers in Africa, and delved into more detail for the case of Datooga, we now turn to historical questions. Why do these sociolinguistic practices of avoidance parallel one another so closely within geographically distant language groups representing distinct language families? Does this distribution reflect the direct cultural transmission of the sociolinguistic pattern, a common population history socioculturally undergirded by pastoralism and strongly patriarchal cultural emphases, or is it due to chance? Two sociocultural profiles seem to be important here: patriarchy (i.e. patrilineal clans, patrilocal post-marital residence, the elaboration of an agnatic residence pattern organized around a literal patriarch) and pastoralism. Pastoralism is a crucial component of the bridewealth exchange system that symbolically underscores the value and circulation of women through marriage relations across eastern and southern Africa (Kuper, 1982). Meanwhile, it is patriarchal heads of polygynous family settlements who most often control the flow of women and cattle. A tight linkage between pastoralism, bridewealth and the patrilocal incorpora- tion of the bride is observed throughout the region. Ritually, these connections are often transparently symbolized in wedding rites: the theatrical transport of the reluctant bride to the husband’s parents’ village (where she may even be secluded in the kraal) and the corollary herding of the bridewealth cattle to the bride’s parents’ village (Kuper, 1982; Persse, 1934; Treis, 2005; Leslau, 1950). Indeed, pastoralism may have more profound connections with patrilineal descent systems. In a major cross-cultural comparative study, Holden and Mace (2003) showed that Bantu-speaking populations that acquired cattle shifted away from matrilineal descent reckoning. Because of the connection between cattle and bridewealth in sub-Saharan Africa, the male-biasing of inheritance in the pastoralist societies ‘may allow men to support several wives, a human form of resource-holding polygyny’, thereby motivating patrilineal descent ideologies (Holden and Mace,

IN THE NAME OF THE FATHER-IN-LAW 185

2003:2430–2431). Because of the important linkages between pastoralism and the sociocultural parameters which appear to subtend daughter-in-law registers, we review some of the archaeological and genetic evidence for population movements and the diffusion of pastoralism from East Africa towards the south.

4.1 Origin and spread of pastoralism The ‘cattle before crops’ model of food production systems in Africa (Marshall and Hildebrand, 2002) sees the early development of pastoralism as being driven by climatic instability which encouraged Saharan hunter-gatherers to find adaptive solutions to access more predictable resources, leading to the African domestication of cattle around 10,000 years before present [bp]. Pastoralism spread across the continent in a discontinuous and irregular manner according to climatic and ecological conditions which favoured herding practices (e.g. the tsetse fly-free corridor in East Africa [see references in Marshall and Hildebrand, 2002]). The south-eastern migration may have been encouraged by a shift to more humid conditions (10,500 bp). According to archaeological and linguistic data, groups in the north of Sudan migrated even further northwards where they practised cattle domestication, while others migrated westwards at a later date (c. 7,500 bp) (Blench, 2006). Numerous genetic signatures of these early prehistoric migrations have been observed in patterns of genetic variation present in modern African populations (e.g. Tishkoff et al., 2009; Berniell-Lee et al., 2009). These support both archaeological and linguistic data suggesting a common origin for pastoralism in Sudan among Nilo-Saharan speaking populations with subsequent migrations to East Africa (Ehret, 2002). Population history became increasingly complex in Kenya and northern Tanzania within the last 3,000 years. These originally Eastern Sudanese Nilo- Saharan pastoralists interacted and mixed with other groups possessing similar cattle-herding practices but a different language, such as the Cushites (Blench, 2006). Modern genetic data suggest ancient admixture between these groups; an East African-specific mutation associated with lactose tolerance has been identi- fied amongst both Nilo-Saharan and Cushitic speakers (Tishkoff et al., 2009). The diffusion of pastoralism further south within the last 2,000 years is a hotly debated subject amongst archaeologists. Although it is known that it was primarily delayed in tsetse-fly infested areas (Sadr, 1998; Smith, 2005), there is nevertheless debate regarding whether the transmission of pastoralism was primarily demic or cultural in character. Small groups of pastoralists with sheep and pottery might have successfully entered parts of southern Africa, independently from the migration of Bantu speakers, but archaeological data associated with herding remain sparse. Nevertheless, similarities in pottery styles

186 SOCIOLINGUISTIC STUDIES were found between geographically distant early pastoralist sites from northern Namibia, northern Botswana and Zambia (e.g. Bambata-ware) and the border of Kenya/Tanzania (e.g. spouted pottery of Ngamuriak site), which might indicate some traces of cultural diffusion (Smith, 2005). The linguistic evidence is also debated. Ehret (2002) proposed that some Khoe-San languages (e.g. Khwe) present a vocabulary relating to pastoralism which is borrowed from an East Sahelian subset, offering linguistic evidence of this ancient diffusion. Haacke (2008), however, has called into question the validity of the proposed word borrowings (see also Güldemann, 2008). Whatever the validity of the linguistic evidence, recently discovered genetic markers (Y- chromosome specific polymorphisms such as haplogroup E3b1f and population- wide microsatellite diversity estimates) support Ehret’s scenario of an ancient and independent demic diffusion of pastoralists from eastern to southern Africa (Henn et al., 2008).

4.1.1 Bantu expansion and pastoralism The homeland of is located in the Grassfields (Cameroon and Nigeria) and it is assumed that at around 5,000 bp proto-Bantu speakers spread from this region throughout sub-Saharan Africa as far as the Cape region (Nurse and Philippson, 2003). Although the nature of this dispersal (demic and/or cultural) is still debated (e.g. Vansina, 1995; Robertson and Bradley, 2000; Ehret, 2001; Bostoen, 2007), there is consensus that Bantu languages and agricultural techniques spread together with people, and that Bantu speakers adopted pastoralism from pre-existing eastern African groups. This supports the thesis that pastoralism and patrilineality are crucial factors in the development of daughter- in-law registers; it is only among Bantu groups of eastern Africa – and not among any of the, often matrilineal, Bantu groups of central Africa – that in-law avoidance registers are attested. The spread of various forms of pastoralism in southern Africa was likely associated with independent migrations before and during the expansion of Bantu- speakers into the area (Sampson, Hart, Wallsmith and Blagg, 1989; Marshall and Hildebrand, 2002). For example, Iron Age pottery thought to be associated with an early dispersal of Bantu-speakers and dated to 2,300 bp has been discovered in southern Zambia, but without any remains of clearly domesticated fauna (e.g. sheep, cattle) (Mitchell, 2002). After 2,000 bp, early Iron Age sites in southern Africa clearly reflect a sedentary lifestyle based on small livestock species (e.g. sheep, goats) and African grains and pulses. At around 1,600 bp, remains of domesticated cattle become more abundant, although other food resources were exploited too (see references in Marshall and Hildebrand, 2002).

IN THE NAME OF THE FATHER-IN-LAW 187

In sum, the archaeological data support an East African origin for pastoralism with subsequent expansion southwards. The distribution of daughter-in-law registers in eastern and southern Africa is consistent with this model suggesting that pastoralism-centered subsistence practices, and a set of marriage practices intimately connected with them, motivated the development of specialized linguistic registers that mediated in-law relations throughout this region, and that these (or the sociocultural practices that motivate them) were diffused southward by agro-pastoralists.

5 Discussion

As we have seen, the distribution of daughter-in-law registers is tightly correlated with economic and sociocultural parameters: these registers are employed by pastoralists with patrilineal descent ideologies and norms of patrilocal post- marital residence paired with cattle-based bridewealth exchange. Ethnographic records and archaeology indicate that agro-pastoral settlements, past and present, in both eastern and southern Africa, share a common pattern although with some variations (Huffman, 2007; Whitelaw, 2013): villages are circularly organized around cattle kraals (the male domain, where high status men can be buried), with kitchen, storage and sleeping areas encircling the kraal, and where proximity to the center indexes status. This settlement organization materially reflects the patrilineal ideology of male hereditary leadership evidenced in male control of the cattle that are used as bridewealth payments to obtain wives for junior male members of the patriline. These correlations might suggest that daughter-in-law registers have deep historical roots, that they originated in East Africa amongst early pastoralist populations thousands of years ago and that the custom diffused with pastoralism, acquired by Bantu agriculturalists such as those in the Great Lakes and the Nguni further in the south. And yet despite the similarities between the daughter-in-law registers of eastern and southern Africa, it is important to underline that historical diffusion of non-linguistic practices would likely have been sufficient to produce the attested distribution of these registers. Name and homophone avoidance are found on all continents and are often associated with in-law avoidance relationships. Furthermore, in patrilineal societies across Central Asia, daughter-in-law registers are also attested. As illustrated in Fleming (2011), well-worn diachronic pathways seem to be responsible for the high incidence of name and homophone avoidance registers globally. This diachronic pathway is divided into three stages: (a) avoidance of a name in address and reference to a particular individual, which may lead to (b) avoidance of the name type in reference not only to the target of avoidance but also to namesakes, which may in turn give way to (c) avoidance of

188 SOCIOLINGUISTIC STUDIES homophonous (or near-homophonous) lexemes. Given this backdrop, the diffusion of a sociocultural assemblage including pastoralism and the patrilocal post-marital residence of women in, often, large, polygynous households headed by a patriarchal figurehead may have been sufficient to stochastically motivate the reemergence of the sociolinguistic pattern across the region. Rather than by direct cultural transmission of these language practices, it is more likely that partial diffusion of in-law avoidance practices has been complemented by a complex of sociocultural factors motivating the emergence of this pattern at different times and places in Africa. An important component of this complex is the patrilineal descent ideologies which motivate linguistic avoidance. Among the Datooga, linguistic avoidance is seen as quite literally undergirding the reproduction of the patriline, as not avoiding is thought to endanger a woman’s children (cf. Treis, 2005:312). Sidama women avoid not only the names of male members of their husband’s clan, but also the name of the patriclan itself (Teferra, 1987:44). Daughter-in-law registers index deference towards local patriarchs and toward the patrilines that they stand for, patrilines whose power manifests in their capacity to obtain ‘wives for cattle’ (Kuper, 1982). Whatever the precise historical mechanisms of their emergence, these sociolinguistic practices clearly fit hand-in-glove with the patrilineal agro- pastoralism of eastern and southern Africa.

Notes

1. Table abbreviations: Sib=Sibling; Sp=Spouse; V=honorific pronoun; T=non- honorific pronoun. A ‘*’ sign indicates the linguistic forms that should be avoided in the particular kinship relationship. 2. Table abbreviations: H=husband; W=wife; F=father; P=parent; M=mother; B=brother; Sib=sibling; Z=sister; D=daughter; e=elder. The arrow indicates the direction of linguistic avoidance (e.g., “♀ → HF” should be read ‘A woman avoids the name and homophones of her Husband’s Father’). 3. WMeD = Wife’s mother’s eldest daughter (cf. wife’s elder sister). A man must avoid his wife’s mother, but also the other wives of his father-in-law, and the eldest daughters of those wives (Lindblom 1920:89).

Acknowledgments

Thanks to Bernard Bernier for reading Komori’s Japanese-medium article on Kerewe. Library research for this project was supported by a generous grant from the Social Sciences and Humanities Research Council of Canada. Fieldwork on Datooga avoidance was supported by the NSF (BCS-1422677) and made possible

IN THE NAME OF THE FATHER-IN-LAW 189 with the help of many research collaborators in Tanzania. We are also grateful to the editors of this special issue, Kirsty Rowan and Eyo Mensah, as well as the anonymous reviewers.

About the authors Luke Fleming is Associate Professor in the Department of Anthropology at the Université de Montréal (Québec, Canada). Alice Mitchell is Junior Professor in the Institute for African Studies at the University of Cologne (Germany). Isabelle Ribot is Associate Professor in the Department of Anthropology at the Université de Montréal (Québec, Canada).

References Adane, D. (2014) Social deixis in Haddiya. International Journal of Language and Linguistics 2(5): 301–304. Doi: https://doi.org/10.11648/j.ijll.20140205.12. Agha, A. (1993) Grammatical and indexical convention in honorific discourse. Journal of Linguistic Anthropology 3(2): 131–163. Doi: https://doi.org/10.1525/jlin.1993.3.2. 131. Agha, A. (2000) Register. Journal of Linguistic Anthropology 9(1–2): 216–219. Doi: https://doi.org/10.1525/jlin.1999.9.1-2.216. Berniell-Lee, G., Calafell, F., Bosch, E., Heyer, E., Sica, L., Mouguiama-Daouda, P., van der Veen, L., Hombert, J. M., Quintana-Murci, L. and Comas, D. (2009) Genetic and demographic implications of the Bantu expansion: Insights from human paternal lineages. Molecular Biology and Evolution 26(7): 1581–1589. Doi: https://doi.org/ 10.1093/molbev/msp069. Blench, R. (2006) Archaeology, language, and the African past. New York: Rowman & Littlefield. Blystad, A. (2000) Precarious procreation: Datoga pastoralists at the late 20th century. PhD dissertation, University of Bergen. Bostoen, K. (2007) Pots, words and the Bantu problem: On lexical reconstruction and early African history. Journal of African History 48: 173–199. Doi: https://doi.org/ 10.1017/S002185370700254X. Davies, W. N. G. and Quinche, C. (1933) AmaNdebele taboos and etiquette. Bantu Studies 7(1): 277–284. Doi: https://doi.org/10.1080/02561751.1933.9676322. Dez, J. (1965) Le nom de personne dans la tradition malgache. Civilisation Malgache 1: 91–114. Antananarivo: Faculté des lettres et sciences humaines. Dixon, R. M. W. (1990) The origin of ‘mother-in-law vocabulary’ in two Australian languages. Anthropological Linguistics 32(1–2): 1–56. Ehret, C. (2001) Bantu expansions: re-envisioning a central problem of early African history. The International Journal of African Historical Studies 34(1): 5–41. Doi: https://doi.org/10.2307/3097285.

190 SOCIOLINGUISTIC STUDIES

Ehret, C. (2002) The civilizations of Africa: A history to 1800. Charlottesville: University of Virginia Press. Finlayson, R. (1995) Women’s language of respect: Isihlonipho sabafazi. In R. Mesthrie (ed.) Language and social history: Studies in South African sociolinguistics 140– 153. Cape Town: Philip. Fleming, L. (2011) Name taboos and rigid performativity. Anthropological Quarterly 84(1): 141–164. Doi: https://doi.org/10.1353/anq.2011.0010. Fleming, L. (2014) Australian exceptionalism in the typology of affinal avoidance registers. Anthropological Linguistics 56(2): 115–158. Fleming, L. and Slotta, J. (2018) The pragmatics of kin address: A sociolinguistic universal and its semantic affordances. Journal of Sociolinguistics 22(4): 375–405. Doi: https://doi.org/10.1111/josl.12304. Florey, M. J. and Bolton, R.A. (1997) Personal names, lexical replacement, and language shift in Eastern Indonesia. Cakalele 8: 27–58. Gregor, T. (1977) Mehinaku: The drama of daily life in a Brazilian Indian village. Chicago: University of Chicago Press. Doi: https://doi.org/10.7208/chicago/ 9780226150338.001.0001. Güldemann, T. (2008) A linguist’s view: Khoe-Kwadi speakers as the earliest food- producers of southern Africa. Southern African Humanities 20: 93–132. Haacke, W. H. G. (2008) Linguistic hypotheses on the origin of Namibian Khoekhoe speakers. Southern African Humanities 20: 163–177. Haviland, J. (1979) Guugu Yimidhirr Brother-in-law language. Language in Society 8(3): 365–393. Doi: https://doi.org/10.1017/S0047404500007600. Henn, B. M., Gignoux, C., Lin, A.A., Oefner, P.J., Shen, P., Scozzari, R., Cruciani, F., Tishkoff, S.A., Mountain, J.L. and Underhill, P.A. (2008) Y-chromosomal evidence of a pastoralist migration through Tanzania to southern Africa. Proceedings of the National Academy of Sciences 105(31): 10693–10698. Doi: https://doi.org/10.1073/ pnas.0801184105. Herbert, R. K. (1990) Hlonipha and the ambiguous woman. Anthropos 85(4–6): 455–473. Holden, C. J. and Mace, R. (2003) Spread of cattle led to the loss of matrilineal descent in Africa: A coevolutionary analysis. Proceedings: Biological Sciences 270(1532): 2425–2433. Doi: https://doi.org/10.1098/rspb.2003.2535. Huffman, T. N. (2007) Handbook to the Iron Age: The archaeology of pre-colonial farming societies in Southern Africa. Scottsville: University of KwaZulu-Natal Press. Huntingford, G. W. B. (1969) The southern Nilo-Hamites. London: International African Institute. Kintz, D. (1986) Ce que disent les anthroponymes peuls. Langage et société 36: 27–40. Doi: https://doi.org/10.3406/lsoc.1986.2052. Kleinewillinghofer, U. (1995) Don’t use the name of my dead father: A reason for lexical change in some Northwestern Adamawa languages (Northeastern Nigeria). Afrika und Übersee 78: 125–137. Klima, G. (1970) The Barabaig: East African cattle-herders. New York: Holt, Rinehart & Winston.

IN THE NAME OF THE FATHER-IN-LAW 191

Kolbusa, S. (2000) Ingamwana: Nyakyusa-Schwiegermeidung. MA thesis, University of Bayreuth. Komori, J. (1999) Analysis of women’s words in Kerewe (Tanzania): A language aspect of the avoidance relationship. Journal of Asian and African Studies 58: 343–363. Kunene, D. P. (1958) Notes on Hlonepha among the Southern Sotho. African Studies 17: 159–182. Doi: https://doi.org/10.1080/00020185808707057. Kuper, A. (1982) Wives for cattle: Bridewealth and marriage in Southern Africa. London: Routledge & Kegan Paul. Leslau, W. (1950) Ethiopic documents: Gurage. New York: Academic Press. Lindblom, G. (1920) The Akamba in British East Africa: An ethnological monograph. Uppsala: Appelbergs boktryckeri aktiebolag. Luthuli, T. (2007) Assessing politeness, language and gender in Hlonipha. MA thesis, University of KwaZulu-Natal. Marshall, F. and Hildebrand, E. (2002) Cattle before crops: the beginnings of food production in Africa. Journal of World Prehistory 16(2): 99–143. Doi: https://doi.org/10.1023/A:1019954903395. Mbaya, M. (2002) Linguistic taboo in African marriage context: A study of the Oromo Laguu. Nordic Journal of African Studies 11(2): 224–235. Mitchell, A. (2015) Words that smell like Father-in-law: A linguistic description of the Datooga avoidance register. Anthropological Linguistics 57(2): 195–217. Doi: https://doi.org/10.1353/anl.2016.0004. Mitchell, A. (2018) Allusive references and other‐oriented stance in an affinal avoidance register. Journal of Linguistic Anthropology 28(1): 4–21. Doi: https://doi.org/10. 1111/jola.12174. Mitchell, P. J. (2002) The archaeology of Southern Africa. Cambridge: Cambridge University Press. Mous, M. (2001) Paralexification in language intertwining. In N. Smith and T. Veenstra (eds) Creolization and contact 113–123. Amsterdam and Philadelphia: John Benjamins. Doi: https://doi.org/10.1075/cll.23.05mou. Muzale, H. R. T. and Rugemalira, J. M. (2008) Researching and documenting the . Language Documentation and Conservation 2(1): 68–108. Nurse, D. and Philippson, G. (2003) Towards a historical classification of the Bantu languages. In D. Nurse and G. Philippson (eds) The Bantu languages 164–181. London: Routledge. Persse, E. M. (1934) Ethnological notes on the Karimojong. Uganda Journal 1(2): 110– 115. Raum, O. (1973) The social functions of avoidances and taboos among the Zulu. Berlin: Mouton de Gruyter. Doi: https://doi.org/10.1515/9783110832884. Robertson, J. H. and Bradley, R. (2000) A new paradigm: The African early iron age without Bantu migrations. History in Africa 27: 287–323. Doi: https://doi.org/ 10.2307/3172118. Sadr, K. (1998) The first herders at the Cape of Good Hope. African Archaeological Review 15: 101–132. Doi: https://doi.org/10.1023/A:1022158701778.

192 SOCIOLINGUISTIC STUDIES

Sampson, C. G., Hart, T. J. G., Wallsmith, D. L. and Blagg, J. D. (1989) The ceramic sequence in the upper Seacow Valley: Problems and implications. South African Archaeological Bulletin 44: 3–16. Doi: https://doi.org/10.2307/3888314. Silverstein, M. (2003) Indexical order and the dialectics of sociolinguistic life. Language and Communication 23: 193–229. Doi: https://doi.org/10.1016/S0271- 5309(03)00013-2. Simons, G. F. (1982) Word taboo and comparative Austronesian Linguistics. In A. Halim, L. Carrington and S. A. Wurm (eds) Accent on variety: Papers from the Third International Conference on Austronesian Linguistics. Vol. 3 157–226. Canberra: Australian National University. Smith, A. (2005) The concepts of ‘neolithic’ and ‘neolithisation’ for Africa? Before Farming 1: 1–6. Doi: https://doi.org/10.3828/bfarm.2005.1.2. Stasch, R. (2003) Separateness as a relation: The iconicity, univocality, and creativity of Korowai mother-in-law avoidance. Journal of the Royal Anthropological Institute 9(2): 311–329. Doi: https://doi.org/10.1111/1467-9655.00152. Teferra, A. (1987) Ballišša: Women’s speech among the Sidama. Journal of Ethiopian Studies 20: 44–59. Tishkoff, S. A., Reed, F. A., Friedlaender, F. R., Ehret, C., Ranciaro, A., Froment, A., Hirbo, J. B., Awomoyi, A. A., Bodo, J.-M., Doumbo, O., Ibrahim, M., Juma, A. T., Kotze, M. J., Lema, G., Moore, J. H., Mortensen, H., Nyambo, T.B., Omar, S. A., Powell, K., Pretorius, G.S., Smith, M. W., Thera, M. A., Wambebe, C., Weber, J. L. and Williams, S. M. (2009) The genetic structure and history of Africans and African Americans. Science 324: 1035–1044. Doi: https://doi.org/10.1126/science.1172257. Treis, Y. (2005) Avoiding their names, avoiding their eyes: How Kambaata women respect their in-laws. Anthropological Linguistics 47(3): 292–320. Tsunoda, T. (1981) The Djaru language of Kimberley, Western Australia. Canberra: Pacific Linguistics. Vansina, J. (1995) New linguistic evidence and «the Bantu expansion. Journal of African History 36(2): 173–195. Doi: https://doi.org/10.1017/S0021853700034101. Whitelaw, G. (2013) Pollution concepts and marriage for the southern African Iron Age. Cambridge Archaeological Journal 23(2): 203–225. Doi: https://doi.org/10.1017/ S0959774313000279.

(Received 14th February 2017; accepted 7th December 2018; revision received 21st December 2018; final revision received 29th April 2019)