Why Thracian Should not be Indo‐European 1

Thracian, an extinct Paleo‐Balkan language, is traditionally considered to belong to the Indo‐European language family. About 25 Thracian appellatives are known from ancient sources, including basic concepts such as ‘meat’, ‘mouse’ and ‘animal skin’. None of the Thracian words has a direct and reliable Indo‐European (IE) etymology (i.e. with a direct sound match, semantic coincidence and comparanda in more than one language group). To roughly estimate the probability of such a result, two experiments were conducted. In the first experiment, wordlists with the same concepts were compiled for six IE languages (Sanskrit, Ancient Greek, Latin, Standard German, Czech and Albanian) and two non‐IE languages (Basque and Turkish). Unlike Thracian, the tested IE languages have between three and five words with a direct and reliable IE etymology, and even Basque and Turkish have a couple of parasitic (i.e. probabilistically conditioned) matches with the IE data. In the second experiment, Phrygian, Lydian, Carian and Galatian glosses were examined for IE etymologies; each language has words with transparent IE cognates, which contrasts with the Thracian case. This suggests that, first, the null hypothesis should be adopted—namely, that Thracian is a language of unknown origin. Second, Thracian is unlikely to belong to the IE family. A possible scenario is that Thracian is a linguistic remnant of pre‐Indo‐European Europe.