BEHAVIORAL AND BRAIN SCIENCES (2010) 33, 367–416 doi:10.1017/S0140525X10000890 Human kinship, from conceptual structure to grammar Doug Jones Department of Anthropology, University of Utah, Salt Lake City, UT 84112 [email protected] Abstract: Research in anthropology has shown that kin terminologies have a complex combinatorial structure and vary systematically across cultures. This article argues that universals and variation in kin terminology result from the interaction of (1) an innate conceptual structure of kinship, homologous with conceptual structure in other domains, and (2) principles of optimal, “grammatical” communication active in language in general. Kin terms from two languages, English and Seneca, show how terminologies that look very different on the surface may result from variation in the rankings of a universal set of constraints. Constraints on kin terms form a system: some are concerned with absolute features of kin (sex), others with the position (distance and direction) of kin in “kinship space,” others with groups and group boundaries (matrilines, patrilines, generations, etc.). Also, kin terms sometimes extend indefinitely via recursion, and recursion in kin terminology has parallels with recursion in other areas of language. Thus the study of kinship sheds light on two areas of cognition, and their phylogeny. The conceptual structure of kinship seems to borrow its organization from the conceptual structure of space, while being specialized for representing genealogy. And the grammar of kinship looks like the product of an evolved grammar faculty, opportunistically active across traditional domains of semantics, syntax, and phonology. Grammar is best understood as an offshoot of a uniquely human capacity for playing coordination games. Keywords: cognitive anthropology; conceptual structure; coordination games; evolutionary psychology; grammar; kinship; kin terms; language evolution; Optimality Theory (OT); recursion 1. Introduction itself in a marginal position in anthropology and cognitive science. Cultural anthropology, with some exceptions, has This article addresses a longstanding puzzle in the human grown increasingly committed to cultural particularism, sciences: Kinship and language, very different in content, and increasingly divorced from developments in cognitive are intriguingly similar in form. Kinship has to do with science. The minority of anthropologists who apply evol- aunts and uncles, matrilineages and patrilineages, and utionary theory to kinship mostly have not investigated ascending and descending generations, while linguistics its cognitive aspects. (On the “fall of kinship” in anthropol- has to do with phonemes and syllables, morphemes and ogy, see Sousa 2003 and commentators, and Shapiro word classes, and heads and phrases. But kinship and 2008.) Meanwhile, as linguistics has grown increasingly language are similar in their combinatorial structure, specialized – and especially as syntax has moved to a pointing toward general principles of cognition or com- dominant position in the field – some earlier connections munication at work in both cases (Kroeber 1909, Le´vi- with other areas of inquiry have weakened. Strauss 1963). In this article, however, I argue that the classic topic of The parallels between kinship rules and language rules kinship and language deserves a fresh look in the light of were the subject of pioneering work several decades ago, recent progress in linguistics and cognitive science – in the heyday of structural linguistics and structuralism. and, in turn, that the study of kinship is relevant to Especially in the area of kin terminology, and especially current debates about mind and language. The rest of from the 1960s to the 1970s, researchers demonstrated this section introduces advances in two areas – the that the classification of kin reflects not just social facts, exploration of conceptual structure at the interface but also cognitive and linguistic principles. Cross-cultural between semantics and cognition, and the development regularities in kin terminology were documented, compar- able to regularities in color terminology (D’Andrade 1971; Greenberg 1966; 1975; 1990; Nerlove & Romney 1967), and an assortment of formal methods was devised to DOUG JONES is Associate Professor of Anthropology at analyze systematic variation and universals in kin terminol- the University of Utah and author of Physical Attrac- ogy (Buchler & Selby 1968; Goodenough 1965; Lounsbury tiveness and the Theory of Sexual Selection (1996, Uni- 1964a; 1964b; Romney & D’Andrade 1964; Scheffler 1968; versity of Michigan, Museum of Anthropology) and co- Scheffler & Lounsbury 1971; Wallace & Atkins 1960). editor (with Bojka Milicic) of Kinship, Language and But intellectual fashions have changed since then. In Prehistory (2010, University of Utah). He is interested in how the social anthropology of kinship relates to spite of important advances (Gould 2000; Hage 1997; recent advances in the cognitive sciences and evol- 2001; Hirschfeld 1989; Kronenfeld 1996; 2009; Leaf utionary theory. Jones is also currently doing research 2006; Lehman 1993; 2001; Read 1984; 2001a; 2001b), on the cognitive anthropology of race in Brazil. the study of kinship, language, and cognition now finds # Cambridge University Press 2010 0140-525X/10 $40.00 367 Jones: Human kinship, from conceptual structure to grammar of a new approach to rules of language called optimality approach to rules of language, originally developed in pho- theory – and previews how they apply to kinship. nology but apparently of much wider application (Archan- Conceptual structure. The study of human cognition has geli & Langedoen 1997; McCarthy 2001; Prince & received a major boost from the realization that there are Smolensky 2004/1993). two distinct levels of language – ordinary and VIP, as it OT doesn’t say what the rules of language are – rules were – which differ both in linguistic behavior and in differ among linguistic domains – instead, it describes meaning (Grimshaw 2005; Jackendoff 1983; 2002; Pinker how rules interact. According to OT, rules, or constraints, 1989; 2007; Talmy 2000a). The first, open-class or don’t actively transform linguistic representations but act lexical, level includes most nouns and verbs. The as filters on randomly generated variation, with each con- number of open-class forms in any language is very straint weeding out variants that violate it. Constraints may large, and virtually any imaginable concept can receive be mutually inconsistent, so grammars manage trade-offs an open-class form. between conflicting constraints through constraint The second, closed-class or grammatical, level is a more ranking. Constraints are put in rank order, with each con- exclusive club. Forms at this level include prepositions like straint strictly dominating lower ranking ones: One viola- in and on, demonstratives like this and that, inflections like tion of a constraint outweighs any number of violations the plural -s and past tense -d, and word classes like mass of all lower ranking constraints. OT covers cross-linguistic noun and intransitive verb (but not specific instances of universals and variation in grammar in a unified frame- nouns or verbs). Within languages, closed-class forms work: To a first approximation, languages all use the are limited in number: Compare the number of preposi- same constraints but differ in their constraint rankings. tions, demonstratives, tenses, and word classes in OT has inspired large body of research: As of this English with the number of nouns and verbs. Closed- writing, one major online resource, the Rutgers Optimality class forms are also restricted in the range of meanings Archive (roa.rutgers.edu), has 1,069 articles on file. Yet in they bear. Across languages, some of the distinctions that some ways OT is still finding its place in linguistics. For regularly make it into closed-class forms are: singular example, recent theories of language evolution (Hauser and plural, bounded and unbounded, figure and ground, et al. 2002; Hurford 2007; Jackendoff 2002) have nothing near and far, present and past, unmovable/inalienable to say about the phylogenesis and adaptive significance and movable/alienable, and human and nonhuman. But of OT. We may get a better idea of the nature and poten- many more semantic distinctions – some quite impor- tial of the theory by comparing it to Utility Theory (UT) in tant – are seldom or never registered in closed-class economics. Both UT and OT are theories of constrained forms, including odd and even, legal and illegal, and absol- optimization. They posit that people mentally search a ute size or duration (e.g., more and less than x feet long or large space of possibilities to discover an optimum, t years old). either an optimal basket of goods or an optimal linguistic The distinction between two levels of language is not output. Optimal doesn’t mean perfect. It means doing as only interesting in its own right, but also potentially impor- well as possible in the face of trade-offs between conflict- tant for investigating cognitive universals. Suppose, as ing constraints. Yet optimization works differently in UT many lines of evidence suggest, that in learning about and OT (Prince 2007; Smolensky & Legendre 2006). In words and the world, people draw not just on perceptual standard UT, trade-offs and potential outcomes vary quan- data but also on an inventory of innate concepts. Then titatively, so economic optimization involves finding open-class