Alexander Akulov. 2015. Prefixation Ability Index (PAI) as a powerful typological tool of historical linguistics. Lingua Posnaniensis, Volume 57, Issue 1 (Jun 2015); pp.: 7 - 24

At present there are many odd hypotheses about the genetic affiliation of certain languages. Most such hypotheses are invented without any serious examination of the structural differences between the languages being compared. The PAI method was inspired by ideas of A. P. Volodin, who noticed that there were two types of languages, one type has prefixation and the other does not. Actually, there is no sharp divide between the two types, it is more precise to use a coefficient (i.e. PAI) rather than simply ask “does a language allow prefixation?” The PAI theory supposed there was correlation between values of PAI of genetically related languages. Tests of PAI on the material of well assembled stocks prove that such correlations exist. Being applied to Ainu and to languages that are possibly supposed to be related to it. The PAI shows that Ainu that is not related to either Altaic or Nivkh, while a search for relatives of Ainu to the south shows potential. Also PAI can be useful in the case of other unsettled questions of language affiliation in North America, New Guinea, Southeast Asia, Africa and other places.

http://www.degruyter.com/view/j/linpo.2 … format=INT