Notes :Burrow & Emeneau tentatively suggest an IA origin for the root (Old Indian sañcaka-, Panj. sañcā, saccā "mould"). This is highly unlikely, because inlaut -c- suggests a recent borrowing from one of IA dialects where *-ñc- > *-cc-; on the other hand, the ubiquitous dropping of the initial sibilant suggests a very old borrowing. Proto-Dravidian origin (with an old meaning like "mould" - not necessarily for casting iron) is thus more probable at the moment. // Present in 2/6 branches.
Notes :The original form of this stem is seen only in Tulu (aḍē-vuni). Most of the time, however, it is met with consonantal suffixes: PDR *aḍá-ŋ-, caus. *aḍá-p- with expected aphaeresis of the first syllable throughout Telugu and Central Dravidian languages. // Present in 5/6 branches.
Notes :A Tamil-Kurukh isogloss that is nevertheless quite plausible. (More widespread are derivatives like *aḍai-pa- "betel pouch" and *aḍai-kāj "betel nut"). The PNDR form would reflect the variant *aḍa[i]-kái. // Present in 2/6 branches.
Notes :Monosyllabic forms are also found but are in the minority and may be secondary. PNDR *aṛg- "spleen" can go back to *aḍV́K- and fits in rather well phonetically; as for the shift in meaning, it is commonly observed in South Dravidian: cf. Tam. kaṭṭi "clod, lump, enlarged spleen", Kod. gëṇḍe "spleen", etc. // Present in 4/6 branches.
Notes :Joined by Burrow & Emeneau with *adír- 'to shake, tremble'. If we think of the original root as *ad- with later suffixal additions, this is quite probable. However, the fact of there being two phonologically different stems cannot be denied (cf. Tamil atir vs. ataṭṭu < *ataḷ-tu; Tel. adaru vs. adalu), and although the meanings are not always differentiated (cf., however, the differentiation in Telugu), this is sufficient evidence for at least granting them two different entries. // Present in 2/6 branches.
Notes :A rather odd bunch of forms. First of all, Kan. nīr-aṛke is not easy to trace back to nīr-aḍike; its second component should better fit in with such forms as aṛipu 'desire, affection', aṛkar_ 'love' (see PSDR *aṛi-, *aṛk-). Second, Kolami aḍḍ- and Parji/Gadba anḍ- cannot go back to one and the same prototype either (the former suggests *aḍḍ-, the latter *anḍ- or *and_-). We are obviously dealing with different etymons here - or at least a mix of original and contaminated forms. Dubious. // Present in 3/6 branches.
Notes :A cultural term with an extremely low probability of being PDr; in fact, most of the forms can be explained as borrowings from Telugu (cf. also Tamil attam id., explicitly marked as a borrowing in DEDR). Nevertheless, the origins of the Telugu form are obscure; Burrow & Emeneau compare Prakrit addāa- 'mirror', but the word does not seem to have an IE etymology and may be of Dravidian origin. // Present in 3/6 branches.
Notes :The best evidence for the inlaut voiced geminate comes from South Dravidian languages (Kannada, Kodagu, and Tulu; in the Nilgiris gemination was lost due to the geminate finding itself in the absolute auslaut position after the syncopation of the last syllable) and Telugu. As for Central Dravidian forms, many of them may actually reflect old re-borrowings from Telugu, due to both the irregular -m in the auslaut and the scarceness of the forms (in many languages, only *aḍḍa-m "obstacle" is attested without the corresponding verbal stems). However, Konda aṛk and aḍi- are probably original. So are the NDR forms. Cf. also cases like PDR *mok- ~ *mogám "face" in which the original *-m is almost certainly preserved in Central Dravidian languages. // Present in 6/6 branches.
Notes :The few Malto verbal stems in -yar- that there are do sometimes correspond to SDR stems with -i- in the second syllable (cf. tanyare 'to become rich' - PSDR *taṇi- 'to thrive, abound', etc.); whether the resonant -r- is the same etymologically is, however, unclear. Nevertheless, all the other forms are more or less in agreement. // Present in 4/6 branches.
Notes :Essentially a Telugu-Parji isogloss. Burrow & Emeneau also add Kui jāpa (which is hardly possible since Kui j- may be < *(-)d_-, but hardly < *(-)ḍ-, and, under question, Tam. aḷavu "to inquire after one's health" and Tulu naṭṭuni "to beg" (both these forms have to be left out as well due to phonetic incompatibility). // Present in 2/6 branches.
Notes :The auslaut -m- in this root is only "detached" in Tamil, and even there in most forms it can be explained as the result of a secondary loss; its status as part of the root is quite obviously Proto-Dravidian. // Present in 4/6 branches.
Notes :Possibly derived from PDR *aḍV́- "to be near, close". The NDR parallel is quite dubious due to both the semantics and phonetics of the stem, but PNDR *aṭṭ- has no better etymology. // Present in 3/6 branches.