The Historical Development of Numeral Systems: Insights from Hunter-Gatherers

While few robust correlations between linguistic systems and sociocultural patterns such as subsistence have been identified, numeral systems appear to constitute one example of such a correlation. Unlike most linguistic subsystems, numerals are highly prone to cultural elaboration (Andersen 2005), and tend to develop systemic complexity in response to socioeconomic need (Winter 1999, Heine 1997). Accordingly, societies having little need for keeping track of exact quantities are likely to have incipient or low-limit ('restricted') numeral systems. Such systems are apparently more prevalent in languages spoken by hunting/gathering peoples, who tend to live in relatively small-scale, egalitarian societies (Hammerström 2008).

            This talk reports on a comparative investigation of the numeral systems of hunter-gatherers and their neighbors in three regions, South America (principally Amazonia), Australia, and North America (principally California). The investigation draws from an extensive survey of numerals and quantifiers in the languages of these regions, involving 186 languages from South America, 121 languages from Australia, and 48 languages from North America.

            We find support for the generalization that hunting-gathering subsistence correlates with relatively low-limit numeral systems, and likewise with low (or no) bases. This is true for both Australia and South America, in which systems surveyed rarely go above 20, and bases range from 2 to 5 (see Zentz and Bowern forthcoming, Epps and Hansen 2009). While Australian groups are (or were) exclusively hunters/gatherers, we find a significant correlation between a relatively strong hunting/gathering subsistence focus and lower numeral limits within South America. However, the languages in the North American sample generally tend to relatively higher limits (with terms for 100 frequently attested), and numeral limits do not show a robust correlation with subsistence pattern.

            In South America, we also find a high degree of etymological transparency of low-level numerals (1-4). The lexical sources of these numerals, and the layering of bases over time, provide a rare glimpse into the genesis and development of numeral systems. In Australia and North America, on the other hand, etymological sources for most low-level numerals are not easily identified.

            We conclude that hunter-gatherer numeral systems exhibit extensive variation, but that they nonetheless provide important insights into the historical processes that shape the development of numerals in language.