Reconstructing the History of Marriage and Residence Strategies in Indo-European—Speaking Societies

Abstract

Reconstructing the History of Marriage and Residence Strategies in Indo-European—Speaking Societies Laura Fortunato Keywords Indo-European, Cultural Phylogenetics, Marriage, Monogamy, Polygyny, Affinal Terminology, Residence, Neolocality, Uxorilocality, Virilocality This file provides additional information on the data and methods used in Fortunato (2011a,b), and discussion of the results of the fossilization of nodes Proto-Indo-Hittite (PIH) and Proto-Indo-European (PIE) for marriage and residence strategies. Data and Methods Below I provide details on the criteria used to collate the cross-cultural sample, with the cross-cultural data in table form, and information on the procedure used by Pagel et al. (2007) to infer the posterior probability distribution of trees on which I mapped the cross-cultural data. Finally, I provide a detailed description of the method used for the comparative analyses. Cross-Cultural Data. Variable identifiers in this section follow Gray's (1999) Ethnographic Atlas (EA) codebook. I collated the cross-cultural sample by matching societies scored as speaking Indo-European (IE) languages (based on EA variable 98) with speech varieties in Dyen et al.'s (1992) IE basic vocabulary database, where needed using information from additional ethnographic and linguistic sources (e.g., Gordon 2005; Levinson 1991-1996; Price 1989; Ruhlen 1991). I also checked for correspondence between speech varieties in the linguistic database and the 62 societies in the EA with linguistic affiliation unknown and located in East Eurasia or in the Circum-Mediterranean region (based on EA variable 91). In some cases, more than one speech variety in the linguistic database could be matched with the same society in the EA. For example, Dyen et al. (1992) include five entries for Greek: three for dialectal forms (Greek D, Greek K, Greek ML), one for modern Greek (Greek Mod), and one for modern spoken [End Page 129] Greek (Greek MD), the latter compiled from dictionary data. In these cases, where available I selected the variety derived from dictionary data, which is likely to be less specific than other entries; alternatively, I selected the variety with data for the greatest number of meanings, or the first variety listed in Dyen et al. (1992, pp. 99-101). The phylogenetic tree model used to represent how societies are related captures the process of diversification of taxa from a common ancestor; therefore, I included in the sample only societies located in Eurasia, corresponding to the geographic range of IE languages before 1492 CE (Diamond and Bellwood 2003). I excluded the Icelanders because the EA description for this society refers to 1100 CE, while the descriptions for the 27 societies included in the sample refer to the "ethnographic present," with dates ranging from 1880 to 1960 CE, and median 1945 CE (Murdock 1967). Table 1 includes the recoded data on marriage strategy and residence strategy (prevailing and alternative modes) for the 27 societies. Tree Sample. Pagel et al. (2007) inferred the posterior probability distribution of trees from Dyen et al.'s (1992) IE basic vocabulary database, using the Bayesian Markov chain Monte Carlo (MCMC) phylogenetic tree-building method developed by Pagel and Meade (2004). The linguistic database includes word forms and cognacy judgments for 95 modern IE speech varieties (languages, dialects, and creoles) across the Swadesh 200-word list of items of basic vocabulary; two or more word forms are cognate if they share a common origin. Swadesh lists consist of cross-culturally universal items of vocabulary such as pronouns, body parts, and numerals, which are less prone to innovation and borrowing (i.e., horizontal transmission) than other meanings (Swadesh 1952). The tree-building analysis was performed on a data matrix obtained from the linguistic database as follows. First, Pagel et al. (2007) excluded eleven speech varieties suspected of methodological bias by Dyen et al. (1992) and added data for three extinct varieties (Hittite, Tocharian A, Tocharian B) to be used as "outgroup" taxa. Outgroups provide information on the direction of change in the data by virtue of being distantly related to the groups under investigation, the "ingroup" taxa; they are used in tree-building for determining ancestor-descendant relationships (Felsenstein 2004, p. 6). As discussed in Fortunato (2011a), Hittite belongs to the extinct sister-group to the IE languages, the...