A Bayesian investigation of the origin hypotheses of the ......Dravidian family Dravidian language...

A Bayesian investigation of the origin hypotheses of the

Dravidian family

Dravidian language family has about 81 languages according to Glottolog1. The Dravid-ian languages are well-studied [Krishnamurti, 2003] from a historical linguistics viewpoint. Re-cently, phylogenetic methods originating from bioinformatics have been applied to infer absolutechronologies of language trees for range of language families such as Indo-European [Rama, 2018,Chang et al., 2015], Pama-Nyungan [Bowern and Atkinson, 2012], and Dravidian [Kolipakamet al., 2018] based on lexical cognate data.

The recent study by Kolipakam et al. [2018] infers a time depth of 4500 years for theroot age of the Dravidian family (for 20 languages) based on calibration points such as theantiquity of Old Tamil (2100 Before Present [B.P]) and the first attestation of Telugu inscriptionsdating to 1300 years B.P. The inferred tree does not group South-Dravidian I and II groupstogether under a single node. Although, the authors infer an age that matches with the ageproposed by Krishnamurti [2003], the location of Proto-Dravidian has not been inferred throughphylogeography techniques. Apart from assigning an age, Krishnamurti [2003] links Proto-Dravidian to Indus Valley Civilization. On the other hand, based on linguistic innovationsand archaeological record Southworth [2004], associates Proto-Dravidian to be spoken in LowerGodavari Basin.

We test the statistical support for both the origin hypotheses by applying phylogeographytechniques to the lexical cognate dataset under different assumptions. The results of the firstset of dating experiment where the phylogenetic tree is topologically constrained according tothe findings of the comparative method is given in table 1. The inferred ages do agree with theages posited based on linguistic and archaeological evidence.

Constraint Southworth [2004] Median Inferred age

Root 4500–4000 4284PSD 4000–3000 3205PSDI 2600 (latest) 2515PSDII – 2407CD – 2141ND – 3063

Table 1: Comparison of inferred ages of major subgroups with expansion dates from linguisticand archaeological evidence. Missing ages are shown by ‘–’.

As the next step, we infer the possible homelands for the internal nodes of the chronologicallydated tree using a random walk model. The result shows that the Dravidian homeland would belocated in peninsular India as postulated by Southworth and not . The migration of Brahui isa debated topic [Elfenbein, 1987] and is generally believed to have migrated from central India[Elfenbein, 1998] as opposed to be a remnant from original migration into India. The randomwalk model supports migration from central India hypothesis as shown in the figure 1. The

1https://glottolog.org/resource/languoid/id/drav1251. Accessed on 12-14-2019

1

next steps would consist of joint inference of phylogeny and geography and expansion of thelexical word lists to include more number of languages.

Figure 1: Geographical spread of Dravidian family. Blue lines/dots show the dispersal ofsynchronic languages and red lines/dots show reconstructed nodes.

References

Claire Bowern and Quentin Atkinson. Computational phylogenetics and the internal structure of Pama-Nyungan. Language, 88(4):817–845, 2012.

Will Chang, Chundra Cathcart, David Hall, and Andrew Garrett. Ancestry-constrained phylogeneticanalysis supports the Indo-European steppe hypothesis. Language, 91(1):194–244, 2015.

Josef Elfenbein. A periplus of the ‘Brahui Problem’. Studia Iranica, 16(2):215–233, 1987.

Josef Elfenbein. Brahui. In Sanford B. Steever, editor, The Dravidian languages. Routledge, London,1998.

Vishnupriya Kolipakam, Fiona M Jordan, Michael Dunn, Simon J Greenhill, Remco Bouckaert, Russell DGray, and Annemarie Verkerk. A Bayesian phylogenetic study of the Dravidian language family. RoyalSociety open science, 5(3):171504, 2018.

Bhadriraju Krishnamurti. The Dravidian Languages. Cambridge University Press, 2003.

Taraka Rama. Three tree priors and five datasets: A study of Indo-European phylogenetics. LanguageDynamics and Change, 8(2):182–218, 2018.

Franklin Southworth. Linguistic archaeology of South Asia. Routledge, 2004.

2