Kinship Terminology Research Paper

Kinship terminologies often reflect the culturally patterned ordering of social categories of descent, marriage, and alliance, and this has ensured their enduring interest to anthropologists. It is now clear that the diversity of kinship systems around the world boils down to a restricted set of variants on a few basic types. These typologies of kinship systems are discussed in Sect. 3.

1. Semantics Of Kin Terms

The challenges posed by kin terms can be illustrated by comparing English and Marri-Ngarr, an Australian language (Scheffler 1978). Fig. 1 illustrates the MarriNgarr system schematically, while Table 1 lists the full set of denotata for selected terms. In analyzing kinship, we distinguish kin terms, which are emic (languageinternal) categories, from etic (universally discriminable) kin types, the set of denotata a particular term includes. Thus the English kin term ‘uncle’ includes (at least) the kin types father’s brother (FB) and mother’s brother (MB). Other abbreviations we will use are C (child), S (son), D (daughter), H (husband), W (wife), Z (sister), m (man speaker), w (woman speaker), e (elder), y (younger).

To avoid clutter, many kin types known by the same kin term are omitted from Fig. 1. For example, in Marri-Ngarr same-sex siblings always receive the same kin term: F and FB are both known by it:a; but only the first is given in Fig. 1, though both are listed in Table 1. Further, since a FB is equivalent to a F, his sons are equivalent to father’s sons, in other words to brothers, meaning that this type of cousin (FBS) also falls into the ‘brother’ category, ngauwe.

The first dimension on which these systems differ, then, is that Marri-Ngarr but not English is ‘classificatory’—by using certain principles of extension, it extends indefinitely outwards to encompass the entire social universe. A corollary is that the distinction between consanguineal (or ‘blood’) and affinal relatives (related by marriage) is hard to maintain, since any potential in-law is already related as a blood relative: the rule favoring marriage to certain cousins (FZC) means that a man’s FZH who, as in English, is a type of uncle (kaka), is at the same time a potential father-in-law. In English, on the other hand, kin terms take in only a fraction of the social universe, and in general consanguineal and affinal terms are distinct.

The second striking difference is in how the denotational space is carved up. We focus on just one of the many differences that exist. Both languages have four terms to cover parents and their siblings, but in English father and mother, who are ‘lineal’ relatives lying directly on the descent line, are distinguished from the ‘collateral’ relatives uncle and aunt, which take in any siblings of parents, whether same sex or opposite sex. However, in Marri-Ngarr, parents’ same-sex ( parallel) siblings—FB and MZ—are known by the same term as the parents themselves, in contrast to parents’ opposite-sex (or cross-) siblings, who receive distinct terms (Fig. 2).

The cross vs. parallel distinction, although irrelevant to English, pervades the Marri-Ngarr system. The distinction between ‘cross-aunts’ and ‘cross-uncles’ (ngaia and kaka), related to one’s parents by an opposite-sex sibling link, and ‘parallel’ aunts and uncles (kela and it:a) is the key to understanding how ‘cousins’ get divided into two categories: cross-cousins (manggen if female, do’goli if male)—children of cross-aunts and cross-uncles and therefore potential spouses—vs. parallel cousins, the children of one’s parallel aunts and uncles, who are treated as siblings and are not possible marriage partners. The cross vs. parallel distinction is reflected in kinship terminologies around the world and there are good social reasons for its importance. If membership in social groupings is passed down a patrilineage or matrilineage, then you are in the same group as your parallel relatives but in a different group from your cross-relatives.

The complexity and cross-linguistic variability of kin terms makes them a fertile testing ground for theories of semantics. During the structuralist semantics period of the 1950s and 1960s a common approach was to represent meanings as clusters of abstract features, a technique known as componential analysis (see D’Andrade 1995 for a recent evaluation). The appeal of these analyses lay in their economical use of small numbers of components which combine to generate large numbers of terms, and in the way in which particular terms with wide reference could be defined through the use of abstract components which range over a large class of kin types.

Burling (1970) illustrates the application of componential analysis to Njamal, whose kinship system closely resembles the Marri-Ngarr system above. The Njamal term mama, for instance, which takes in a comparable range of kin types to the Marri-Ngarr it:a, is defined as G⁺¹ M^eS^m, i.e., as a relative in the first ascending generation who is male and belongs to the same patrimoiety as ego. (Patrimoieties divide Njamal society into two halves, with membership in each patrimoiety inherited from one’s father; marriage must be with a member of the opposite patrimoiety, which means that, for example, your mother and her brother will be in the opposite patrimoiety to you.) Abstract features such as ‘same/opposite patrimoiety’ enable the analyst to account for the wide denotational range of such terms as Njamal mama or Marri-Ngarr it:a.

However, componential analysis of kin terms has been attacked from a number of quarters. Kay (1975) points out that the use of semantically uninterpreted features without a predicate-argument structure removes kin terms from standard logical relations such as entailment, and obscures the fact that kin terms are two-place predicates (and in many languages kin terms are actually verbs—see Evans 2000). He proposes an alternative logic-based system of representation.

From another angle, Wierzbicka (1986) attacks the hidden circularity of using terms such as ‘ascending generation’ to define a word such as mama without mentioning ‘father,’ since ‘ascending generation’ can only be defined with respect to parenthood. She proposes a definition of ‘father’ which refers outside the kin domain, drawing on his role as begetter (at least for English and Njamal). By allowing the analyst to spell out the relevant cultural components in the definition, this approach has the advantage of showing exactly how a concept such as ‘father’ may vary across cultures—for example, is the father the begetter, or merely the man married to the woman at time of giving birth to the child? Non-kin primitives are also useful in predicting metaphorical extensions of kin terms, such as the use of ‘father’ to mean ‘priest’ (Kronenfeld et al. 1985).

An attack from a third direction came from extensionist approaches (Lounsbury 1964, Scheffler 1978), which heralded a paradigm shift away from ‘categorical’ theories of meaning, which hold that concepts can be defined satisfactorily in terms of necessary and sufficient conditions applying equally to all members, toward prototype theories which allow for varying degrees of membership from focal to peripheral. For example not all ‘uncles’ in English are equally good exemplars—all speakers will agree that MB and FB should be included, while for FZH and MZH speakers may disagree, and some will prefer hedging expressions such as ‘sort of an uncle’ or ‘uncle by marriage.’ MB and FB are therefore focal members of the uncle category, while FZH and MZH are peripheral members.

Scheffler still uses componential analysis for focal category members: the primary denotatum of Marri- Ngarr it:a, F, is defined as K.L.G1. +.♂ i.e., parallel lineal male kin of first ascending generation. However, to get the remaining members of the category, he employs various extension rules. For example, a same- sex sibling merging rule, which places same-sex siblings in the same category, extends it:a from F to FB. Half a dozen extension rules, chained together if necessary, can generate all the extensions in column 5 from the primary denotata given in column 4 of Fig. 1.

2. Kin Terms In Interpersonal Context

So far we have been maintaining a traditional idealization, stemming from the use of the ‘genealogical method,’ which has long been used to elicit kin terms while obtaining genealogies: one asks how the informant addresses and is addressed by each person in the genealogy (Barnard and Good 1984). In fieldwork this method is a necessary step in obtaining an initial set of terms, but it has the drawback of examining kin terms in isolation from their actual use, thus bracketing out the contributions of politeness, conversational footing, and discourse goals to usage (Zeitlyn 1993). In Ungarinyin, for example, which normally distinguishes even-and odd-numbered generations, some terms are extended to take in all generations in a lineage when discussing patrilineal clan estates (Rumsey 1981). Within the functionalist program in linguistics, which sees language structure as resulting from the pressures of actual use, context holds the key to understanding changes in the system itself.

Once kinship terms get used in address, further factors come into play. Often languages have distinct kin terms for address (‘Dad!)’ and reference (‘My father was wise’), although the line may be hard to draw. For reasons of politeness, kin terms often have a wider range when used as address terms, extending to nonkin, as with Indonesian (i)bu ‘mother,’ which in address may extend to older females more generally (then dropping the first syllable).

With its long documented heritage, Chinese offers an unparalleled opportunity to trace how address and reference terms change over time. Kryukov (1998) shows how certain innovations start among address terms, later spreading to reference terms. In the Book of Song (first millennium BC), both father and his siblings were referred to by the general term fu, but in address the distinct term shufu could be used for FyB. By mid-millennium, however, shufu also appears as a reference term. This is typical of how, because they respond to interpersonal context, address terms are usually the more dynamic segment of the kinship system, giving them a special importance for understanding how kinship terminologies change.

Address, however, is not the only way interpersonal context impacts on kinship terms. As mentioned, kin terms are fundamentally two-place predicates, and in many languages there is obligatory marking of the reference point or anchor by a possessor suffix: for example, Dalabon bulu-ngan ‘father-my.’ In languages such as English, by contrast, the anchor is typically implicit and must be recovered by pragmatic inference. Speakers often have the choice of whether to take themselves, the hearer, or some third person as anchor. According to context, the English word ‘Mum’ can be interpreted with an anchor that is first-person (e.g., said by a child to an adult), second-person (e.g., said by an adult to a child), or third-person (e.g., said by a child-care worker taking the child’s perspective).

In all these examples the kinship predicate is calculated with respect to just one anchor. However, kinship terms may also make two anchors simultaneously explicit in special ‘trirelational’ systems, such as the Mayali terms known as Gun-dembui: alongside basic terms like garrard, ‘mother,’ are terms such al-garrng, ‘the one who is your mother and my daughter,’ and its converse al-doingu, ‘the one who is your daughter and my mother.’ See Merlan (1989) for discussion of a similar system in Jawoyn.

The ability to take another’s perspective is a hallmark of mature ‘social intelligence,’ and studies of how speakers manipulate the choice of anchor have important implications for the development of social cognition. Concomitantly, the need constantly to shift away from an egocentric viewpoint in reckoning kinship makes it a difficult semantic field for children to master. Piagetian-style studies of how children acquire kin terms have examined the development of relational abilities (Danziger 1957, Luong 1986).

3. Typologies Of Kinship

The systematic comparison of kinship terminologies around the world, begun with Morgan’s Systems of Consanguinity and Affinity of the Human Family (1871), has made kinship vocabulary the most widely sampled and analyzed semantic field, leading to the finding that ‘the thousands of systems of kinship terminology in human societies around the world are but variants of a limited number of different types’ (Godelier et al. 1998, p. 5).

Consider sibling terminologies. Nerlove and Romney (1967) examined 245 languages for the patterning of terms across the eight logical kin types given by the product of sex of ego sex of referent x relative age. Common patterns are the use of a single term (‘sibling’) for all eight kin types, and the division into two terms based on sex of referent (as in English ‘brother’ vs. ‘sister’). Strikingly, they found that of the 4,140 logically possible types only 10 were attested in more than one language of the sample.

The dominant typologies of kinship terminology take the patterning of parent /aunt/ uncle terms to be the best predictive feature of the rest of the system. The relevant patterns are illustrated in Table 2, which gives the three commonest type-terms used in the literature.

While the first four types are each widely attested, the fifth, although logically possible, is unattested, presumably because it would require a disjunctive definition (F or MB), whereas the extended categories in the other types allow one to find a common factor (e.g., ‘male relative of one’s patriline’ for the Iroquois type). Here the systematic absence is probably attributable to cognitive rather than social causes, since disjunctive categories are widely dispreferred (Green-berg 1980, Hage 1997).

The central role of parent uncle aunt terms in typologies of kinship follows from their part in regulating marriage possibilities. The commonest type of marriage partner among relatives is some type of cousin, and the figuring of cousin relationships depends in turn on the classification of parents’ siblings.

The Dravidian system is a widely cited example of the relation between kin terms and marriage rules, since it is widely associated with exchange of spouses between descent groups, in the form of reciprocal cross-cousin marriage (i.e., between FZC and MBC), as in southern India (Trautmann 1981). If a man marries his cross-cousin (e.g., FZC), then his cross-uncle is simultaneously his father-in-law, and in fact the terminological equation of these two kin types is widespread in Dravidian terminologies.

This example shows how choices made in one part of the system—here parent uncle aunt terms— predict what choices will be made elsewhere. The development of implicational statements—cross- linguistically robust logical relations between one linguistic feature and another—is a central concern of linguistic typology, and one can revisit typologies of kinship in these terms, testing out formulations such as ‘if a language distinguishes parallel and cross-cousins, it will distinguish parallel and cross-uncles.’

Characterizing possible kinship systems by suites of implicational statements rather than simply by idealized types has the advantage that, although in ‘type cases’ large numbers of features co-occur, this is not always the case. First, not all kinship systems display full type consistency. For example, there is a natural correlation between a bifurcate merging pattern in the first ascending generation and a parallel versus cross- cousin distinction in ego’s generation. But not all languages display this: Pitjantjatjara has a bifurcate merging pattern (F=FB≠MB; M=MZ≠FZ) but does not distinguish cross-cousins from parallel cousins or siblings (Scheffler 1978). In terms of the typology in Table 2, we would have to say that Pitjantjatjara is ‘Iroquois’ in the first ascending generation, but ‘Hawaiian’ in ego’s generation. A comprehensive typology thus needs to allow for many parameters to be varied independently, rather than assuming they always co-occur; the sorts of extension rule outlined in Sect. 1, being logically independent, allow this.

Second, the types given in Table 2 are too crude. Murdock (1949), extending his focus into other generations, recognized a further two types, Crow and Omaha, based on cross-generational mergings (e.g., Crow: FZ =FZD=FZDD). The basic types also neglect the affects of relative age of the father’s sibling, yet Gui, for example, treats FyB as a type of F while lumping FeB and MB with FF (Ono 1997). The most detailed explorations of differences between systems supposedly of the same type have concerned bifurcate merging systems. In Kariera systems, for example, spouses in the ascending generation always differ in crossness, whereas in most Dravidian systems they do not (Viveiros de Castro 1998).

All kinship systems employ a limited series of kin terms for a potentially infinite set of kin types. But although the number of attested systems way exceeds the classic fourfold or sixfold typologies of Lowie and Murdock, it is nonetheless striking that most logically possible systems are not found. While extensionist approaches manage to parametrize this patterned diversity by not requiring groups of characteristics to go together, they do not make explicit the implicational relationships between one type of extension and another—here we need to bring in tools developed by linguistic typologists. Only by combining these approaches will we solve the puzzle of how human beings, despite being evolutionarily equipped for diversity in their constructions of social categories, repeatedly come up with similar solutions, due to a tangling of cognitive and social factors that we have yet to unpick.

Bibliography:

1. Barnard A, Good A 1984 Research Practices in the Study of Kinship. Academic Press, London
2. Burling R 1970 Man’s Many Voices. Language in its Cultural Context. Holt, Rinehart and Winston, New York
3. D’Andrade R 1995 The Development of Cognitive Anthropology. Cambridge University Press, Cambridge, UK
4. Danziger K 1957 The child’s understanding of kinship terms: A study in the development of relational concepts. Journal of Genetic Psychology 91: 213–32
5. Evans N 2000 Kinship verbs. In: Vogel P, Comrie B (eds.) Explorations in Word Class Typology. Mouton de Gruyter, Berlin, pp. 103–72
6. Godelier M, Trautmann T R, Tjon Sie Fat F E (eds.) 1998 Transformations of Kinship. Smithsonian Institution Press, Washington, DC
7. Greenberg J 1980 Universals of kinship terminology: Their nature and the problem of their explanation. In: Maquet J (ed.) On Linguistic Anthropology: Essays in Honor of Harry Hoijer. Undena Publications, Malibu, FL, pp. 9–32
8. Hage P 1997 Unthinkable categories and the fundamental laws of kinship. American Ethnologist 24: 652–67
9. Kay P 1975 The generative analysis of kinship semantics: A reanalysis of the Seneca data. Foundations of Language 13: 201–14
10. Kirchhoff P 1932 Verwandtschaftsbezeichnungen und Verwandtenheirat. Zeitschrift fur Ethnologie 64: 47–72
11. Kronenfeld D B, Armstrong J D, Wilmoth S 1985 Exploring the internal structure of linguistic categories: An extensionist semantic view. In: Dougherty J D W (ed.) Directions in Cognitive Anthropology. University of Illinois Press, Urbana, IL, pp. 91–110
12. Kryukov M V 1998 The synchro-diachronic method and the multidirectionality of kinship transformations. In: Godelier M, Trautmann T R, Tjon Sie Fat F E (eds.) Transformations of Kinship. Smithsonian Institution Press, Washington, DC, pp. 294–313
13. Lounsbury F G 1964 The structural analysis of Kinship semantics. In: Hunt H G (ed.) Proceedings of the Ninth International Congress of Linguistics. The Hague, Mouton, pp. 1073–93
14. Lowie R H 1928 A note on relationship terminologies. American Anthropologist 30: 263–7
15. Luong H V 1986 Language, cognition and ontogenetic development: A reexamination of Piaget’s premises. Ethos 14: 7–46
16. Merlan F 1989 Jawoyn relationship terms: Interactional dimensions of Australian kin classification. Anthropological Linguistics 31: 227–63
17. Morgan L H 1871 Systems of Consanguinity and Affinity of the Human Family. Smithsonian Institution, Washington, DC
18. Murdock G P 1949 Social Structure. McMillan, New York
19. Nerlove S, Romney A K 1967 Sibling terminology and cross-sex behavior. American Anthropologist 74: 1249–53
20. Ono H 1997 How do the Gui categorize people? Gengo Kenkyu (Journal of the Linguistic Society of Japan) 111: 42–58
21. Rumsey A 1981 Kinship and context among the Ngarinyin. Oceania 51: 181–92
22. Scheffler H 1978 Australian Kin Classification. Cambridge University Press, Cambridge, UK
23. Trautmann T H 1981 Dravidian Kinship. Cambridge University Press, Cambridge, UK
24. Viveiros de Castro E 1998 Dravidian and related kinship systems. In: Godelier M, Trautmann T R, Tjon Sie Fat F E (eds.) Transformations of Kinship. Smithsonian Institution Press, Washington, DC, pp. 332–85
25. Wierzbicka A 1986 Semantics and interpretation of cultures: The meaning of alternating generation devices in Australian languages. Man 21: 34–49
26. Zeitlyn D 1993 Reconstructing kinship or the pragmatics of kin talk. Man 28: 199–224