Johanna Nichols

Does phoneme inventory size correlate with population size

Atkinson 2011 finds a significant positive correlation between population size and phoneme inventory size (confirming Hay & Bauer 2007) and explains it by migration: phoneme sizes are largest in Africa, and as societies spread out of Africa and around the world they went through population and cultural bottlenecks and underwent phonological simplification as a consequence. We believe the correlation is artefactual if it exists at all, and it probably does not exist. To test it we surveyed 1,350 languages with excellent genealogical and geographical coverage and distribution. We used the Autotyp areal breakdown (Bickel & Nichols 2002). Population size figures were taken, where possible, from grammars, ethnographies, and/or recent census data, and attempt to give population figures for the entire ethnic group (and not just speakers of the language, since most of the world’s languages are losing speakers to large national and international languages); where we did not have this information we used figures from Lewis (ed.) 2009. Language shift has increased rapidly in recent decades, so that the size of the ethnic group is a better measure of the size of the speech community in which the oldest and most fluent speakers grew up, and it is these speakers’ competence that grammars usually describe. The log of population size was coded for each language. Since ethnic groups and speech communities of under a few hundred individuals are ordinarily unstable (not impossibly, but the languages of such communities are frequently undergoing shift and death), all population sizes reported in units or tens were coded as 499 (treating these speech communities as though they still had the sizes reported for them in the early to mid-twentieth century). For each sample language we surveyed the total number of consonant phonemes (excluding phones found only in unassimilated foreign loans), the total number of phonemic vowel qualities, and the number of tone oppositions. AUTHOR’S COPY | AUTORENEXEMPLAR

Selection for m: T pronominals in Eurasia

Pronominal paradigms with first person m and second person t , s , etc. are common in Eurasia and only there. In language families with well-understood histories this type of system behaves as an attractor state, gaining tokens more readily than it loses them. Bound morphology stabilizes such systems, acting as repository of conforming tokens, source of new tokens, and source of pressure toward leveling. This does not explain why the paradigm forms a cluster in Eurasia, however. This chapter hypothesizes that this is due to the distinctive sociolinguistics of post-Neolithic steppe and steppe periphery expansions that maximized opportunities for selection of conforming tokens. The author surveys most of the language families of Eurasia that exhibit m-T systems, sampling densely so as to cover the entire geographical range and all but the lowest-level branches, and also surveying received view about the origins of and changes in person markers. Keywords:Eurasia; Eurasian steppe; m-T systems; pronominal paradigms