This blog is about a weird linguistic feature that I noticed many years ago and in the good old days of Orkut discussion groups (now defunct), this topic generated a major discussion thread in the Linguistics community. I was reminded of it after a recent discussion on Arabic cardinals and thought I should re-post it for everyone’s benefit and interest.
It is a widely accepted (and experimentally confirmed) theory that human languages evolve (drift) according to well-understood rules that are objective and general in application. Given the obvious Darwinian characteristics of language evolution, linguists classify languages in “families” of kinship (much like biological species). However, one must always remember that biological species is a much better defined (and testable) term than a language family and also better motivated. Part of the reason for linguistic classification had as much to do with missionary zeal of the (primarily Anglophone) Bible preacher and the reductionist zeal of the Enlightenment philosopher as genuine bipartisan search for objective truth.
Nonetheless, whatever their original motivation, linguistic families are good phenomenological models with some predictive power (e.g. Avestan grammar reconstruction using Paninian grammar is an obvious case in point). One of the easy tell-tale signs of two languages belonging to a single language family is the correlation between the sets of most frequently used words used in either language, because such words are most prone to change by natural linguistic drift than lexical borrowing. One such typical subset of frequently used words in a language consists of its words for cardinals used for counting, i.e. the usual ones, twos, threes etc.
It turns out that for some reason, still unfathomable to me, the words for cardinal numbers six (6) and seven (7) show uncanny correlation across West-Eurasian language families. Obviously, the in-group correlation in, say, the Indo-European group is well accounted for by the natural evolution of Proto-Indo-European. Compare Latin “sex” to Sanskrit “SakS”, where /S/ is the “sh” sound pronounced retroflexively (i.e. by touching the tongue to the palate). However, out-group correlation is quite surprising and seems rather tight across not just Semitic and Austro-Asiatic families, but even Kartvelian, Finno-Ugric and isolates such as Etruscan. And the structure is deep, as Akkadian (the earliest known Semitic language) is attested from the 2nd-3rd millennium BCE in Mesopotamia and Old Egyptian from a thousand years earlier in the Nile Valley.
Note the correlation disappears with Turkic, Tibetan, Dravidian, Malay etc – languages situated in or originating from the South/East of the Eurasian landmass. What makes it even more surprising is that this really holds for numbers 6 and 7 only, but not 5 nor 10 or higher cardinals. Does it hint at a deeper structure in West Eurasian languages (as opposed to Eastern ones or Sub-Saharan)? Are these hints of some ancient lexical borrowing of names for cardinals to count beyond 5? Or is this correlation across families just pure co-incidence? I do not know, but your comments and thoughts are very welcome.
(Disclaimer: This blog is moderated by its owners. The views expressed are my own and may not reflect the views of the blog owners.)