Notes :The PK form may be borrowed from Telugu dumārumu, which makes the comparison less trustworthy. In any case, vowel length in PK is secondary, probably compensatory (from *dumb-aru).
Notes :Gondi *ḍum- probably goes back to *dumḍ-; else it is hard to explain the initial retroflex (if we do not assume the reconstruction *ḍum- as PGn).
Notes :Various assimilations, progressive as well as regressive, tend to obscure the root's consonantism. The form dupa by itself is only fixed in Konda by Burrow and Bhattacharya and in Kuwi_Isr, but it is the different types of assimilation that give away the original root structure.
Notes :This pronoun was preserved in PK and PPM, but lost almost everywhere else, presumably because of its superfluous character (PGn possessed at least three 'that' pronouns for various degrees of distance).
Notes :One of the main PGn interrogative bases, formed on the basis of PDR *yā-. This one had a quantitative sense ('how much', 'how many'), preserved in most dialects. In Gondi prothetical b- as in all interrogative bases, and the vocalism is secondary, probably by analogy with other bases.
