Projects
Quantitative historical linguistics
Ideas
Languages are rich and highly integrated systems, which share many patterns and constraints, but also permit huge diversity from one language to another in their uses of sound, lexicon, and grammar. Regular relations among languages often result when they are descended from some common ancestor as a result of fission of a speech community. These relations can in principle be used to assess relative degree of similarity and relative recency of shared ancestry.
Historical linguists have developed a system known as the comparative method (see R.L. Trask’s Historical Linguistics) to determine which shared features among languages indicate common descent (versus cultural contacts), and to reconstruct plausible ancestral forms from these and from known patterns in language change. The comparative method aims to be formal and rigorous, but the complexity of language change processes, together with difficulties of principle in typology and dynamics (phonological, semantic), have limited its application as a quantitative method.
Rapid improvements in the scope and precision of what is known about language change, and in methods for probabilistic analysis, make it possible to study the properties of the comparative method as a quantitative system for historical reconstruction. Several conceptual advances are within reach of current methods. 1) Quantification will clarify the status of the comparative method as a formal system. 2) The scope of the method itself may be extended to include new quantitative regularities not previously known. 3) Quantitative models validated against history will more tightly constrain knowledge of language change processes, with implications for cognition and sociology. 4) Historical inference from languages may be brought together with inference from other domains such as molecular systematics, to reconstruct multi-factor pictures of human populations and migrations.
Links
The Evolution of Human Languages project at SFI
Mark Pagel’s page on linguistic phylogeny
Russel Gray’s Austronesian page
Quentin Atkinson’s research page
Mahé Ben Hamed’s DiaDM project
Erez Lieberman’s language page
Collaborators
George Starostin
