Toomas Kivisild, Siiri Rootsi, Mait Metspalu, Ene Metspalu, Juri Parik, Katrin Kaldma,
Esien Usanga, Sarabjit Mastana, Surinder S. Papiha & Richard Villems, “The Genetics of
Language and Farming Spread in India,” ch. 17 in Examining the farming/language
dispersal hypothesis, eds. Peter Bellwood & Colin Renfrew (Cambridge: McDonald Institute
for Archaeological Research, 2003), pp. 215–222.
can be found on the web.
Quote:
Second, great caution is required when interpreting the dates deriving from Y-chromosomal STR coalescent calculations. Table 17.3 reveals that profoundly inconsistent time estimates can be reached when different calibration methods are used. Hence, it seems safer to operate with raw diversity estimates to determine the polarity of the movement leaving the time of origin question unanswered until reliable dating methods for Y-chromosomal STR diversity are worked out.
Yet, even if time estimates are avoided, there are some problems introduced by sampling strategies and differences in demographic histories. For example, in the study by Quintana- Murci et al. (2001), a decline in diversity stretching from Iran to India was observed in haplogroups 3 and 9 and the authors rushed to interpret this empirical observation in favour of directional gene flow to India during Neolithic period (haplogroup 9). They linked this finding to the introduction of Indo-Euro- pean languages (haplogroup 3) to India. Time estimates for their spread were derived from the STR clock.
Yet, even if time estimates are avoided, there are some problems introduced by sampling strategies and differences in demographic histories. For example, in the study by Quintana- Murci et al. (2001), a decline in diversity stretching from Iran to India was observed in haplogroups 3 and 9 and the authors rushed to interpret this empirical observation in favour of directional gene flow to India during Neolithic period (haplogroup 9). They linked this finding to the introduction of Indo-Euro- pean languages (haplogroup 3) to India. Time estimates for their spread were derived from the STR clock.
Here, however, the clock is just a secondary problem, the first being the Indian reference sample used. Indeed, the Indians included in this study consisted of a (limited) sample from Gujarat one of the western maritime provinces of India. When extending the Indian sample with colIections from different states, a quite different, even opposite, pattern emerges (Table 17.3). Indians appear to display the higher diversity both in haplogroups 3 and 9 even if a pooled sample of eastern and southern European populations was considered.If we were to use the same arithmetic and logic (sensu haplogroup 9 is Neolithic) to give an interpretation of this table, then the straightforward suggestion would be that both Neolithic (agriculture)and Indo-European languages arose in India and from there, spread to Europe. We would also have to add that inconsistencies with the archaeological evidence would appear and disappear as we change rate estimates (Table 17.3).
