Collins Conference Room
George Starostin (Higher School of Economics, Moscow / SFI)

This event is closed.

Abstract: From 2001 and all the way up to his recent passing, Dr. Murray Gell-Mann, in addition to his ongoing research in physics and other branches of science, was also the co-director for the "Evolution of Human Languages" project — an ambitious international program whose main goals aim at the establishment of a general historical classification of the worldʼs language families and reconstruction of ancestral language states that go as far back into the past as possible, way beyond the chronological limits typically asserted by mainstream comparative-historical linguists.

Although Gell-Mannʼs interest in linguistics was often regarded as an amusing hobby, it was actually quite consistent with the main theme of his primary research — disentangling the patterns, laws, and accidents that generate complexity from simplicity in all spheres of existence. One particularly strong issue that he had most of his life with contemporary historical linguistics was that the latter, in his opinion, placed too much emphasis on the "burden of proof" in relation to any far-reaching hypothesis of linguistic relationship: according to numerous scholars, common ancestry as the reason for similarities between languages and/or language families can only be claimed once all the alternatives (ranging from contact scenarios to sound symbolism to chance resemblances) have been definitively ruled out. In practice, this requirement usually boils down to refusing to accept or even seriously consider any hypothesis of distant linguistic relationship the evidence for which does not match the strong standards previously set by such theories as Indo-European, Uralic, or Austronesian.

In my talk, I shall discuss the validity of the "burden of proof" placed on long-rangers, including whether hypotheses of distant language relationship such as Altaic, Nostratic, or Dene-Caucasian can really qualify as "extraordinary claims", and also attempt to show that the current predilection towards areal / contact-based explanations of similarities between language families may be just as skewed and impractical as the opposite tendency to ascribe all of them to common descent. The general strategy of the Evolution of Human Languages project, well in line with Gell-Mannʼs own thinking, is to replace a simplistic yes / no dichotomy with a more complicated scale of probabilities when it comes to assessing various hypotheses of relationship — arguably the only way to resolve the long-standing conflict between "lumpers" and "splitters" in the field of comparative-historical linguistics.

Those unable to attend can stream the lecture from Twitter and our YouTube page.

Research Collaboration
SFI Host: 
Jennifer Dunne