Dear Hans, Nagaraj, George, Herman and others,

I am overwhelmed by the responses that I have been receiving offline and online. Having made a bold proposal,  I guess I have my work cut out. I will try to briefly respond to a few of the questions raised here in this email, before taking myself off to work on a comprehensive article with stronger evidence to back up my proposal.

Hans and George: Thanks for the kind responses. I will contact you privately on very interesting points you raised. I believe voicing was phonemic in the pre-Dravidian languages, and echo words and onomatopoeic words found throughout India are infact retentions from these Paleolithic Hunter-Gatherer languages (Niṣādic languages). For instance, in Telugu, the following onomatopoeic words, varying only in voice, have entirely different meanings: cakacaka, jagajaga; kaṇakaṇa, gaṇagaṇa; paṭapaṭa, baḍabaḍa; pakapaka, bagabaga etc. 

> I thought each reconstructions were done meticulously case by case choosing the most plausible root of phonetic change from among those required to account for all the available forms.

Nagaraj: On the question of Dravidian reconstruction, you are right about the need for meticulous reconstruction to account for all available forms. As you know, there have not been comprehensive reconstructions for Dravidian. DEDR only lists a set of cognate words, but provides no reconstructions. Krishnamurti proposed a few reconstructions for a few chosen lexical items and grammatical markers (bound forms) in his 2001 and 2003 books.  BhK, however, argued that under certain circumstances etyma found only in Proto-South Dravidian can be reconstructed for Proto-Dravidian. Here is his argument, and I don't agree with it:

"Only four of the Dravidian languages have recorded history and literature starting from pre-CE to the 11th century. The available dictionaries of the literary languages are extensive, running over 100,000 lexical items in each case. The [recorded] vocabulary of the non-literary languages is not commensurate . . . Therefore, most of the cognates turn up in the four literary languages, of which Tamil, Malayalam and Kannada belong to South Dravidian I and Telugu to South Dravidian II. The absence of cognates in the other subgroups cannot be taken to represent the absence of a concept or a term in Proto-Dravidian. The presence of a name (a cognate) in the minor languages and its exclusion in the major languages should lead to a significant observation that the cognate could be lost in the literary languages, but not vice versa."

Krishnamurti, as far as I know, has been very careful about positing a South Dravidian root for Proto-Dravidian. However, there have several others who applied Tamil words as a Proto-Dravidian reconstructions, without making a convincing argument for such use.

For example, several of the Dravidian proposals for Indus script equate the "fish" sign with "star", as the proponents consider the words for 'fish' and 'star' as homophones in Proto-Dravidian.  I had earlier argued with Parpola (who is a member of this list) and others on the folly of equating mīn- 'fish' [DEDR 4885] with miṇ-/min-(?) 'to glitter, star, sky' [DEDR 4876]. I believe these two entries are unrelated etymologically. If you look at the entries for DEDR 4876, you notice a retroflex ṇ in many entries, which is somehow absent from Tamil-Malayalam. De-retroflexion of a nasal is common in the South-Central, Central and North Dravidian languages. That means, an absence of a retroflex in those languages is not surprising, but the presence of a retroflex consonant in languages across subgroups certainly adds credence to its presence in the proto-language.

4876 Ta. miṉ flash, glitter, lightning; miṉmiṉi firefly; miṉṉu (miṉṉi-) to emit lightning, shine, glitter; miṉṉal lightning; bright coin; miṉuṅku (miṉuṅki-) to glitter, shine, appear bright; miṉukku (miṉukki-) to polish, brighten, beautify, make a display; miṉukkam, miṉukkal polish, brightness, excellence, showiness, show; mīṉ star. Ma. minnuka to flash, shine; minnal lightning; minni shining; a gem in ear-rings; minnikka to cause to flash or shine; minnulightning; minukka to be fine, glitter; minukkuka to polish, varnish, make glitter; minukkam shining, polish; minuṅṅuka to glitter; minuppu sparkling; mīn star. Ko. minc- (minc-) to flash, glitter, be dazzling, lighten; mi·n star. To.mic- (mi&cangle;-) to flash, lighten; mic lightning (in songs); mi·n star. Ka. miṇa glittering, sparkling; miṇaku, miṇuku to glitter; n. glitter; minu, mini sparkling, shining; minuku, minugu to shine, glitter; n. lustre, etc.; miñcu to shine, be bright, sparkle, glitter, flash, lighten; n. shine, lustre, brightness, glitter, lightning; mīn star. Koḍ. minn- (minni-) to lighten, flicker. Tu. miṇimiṇi twinkling, glistening, dimly shining; meṇů glitter, sparkle; miṇuku, meṇaků, meṇuku sparkling; miṇ(u)kuni, meṇ(ů)kuni, minukuni, meñcuni, miñcuni to shine, sparkle, glitter; meñci brightness, lightning; (B-K.) meṇkoḷi, menkōri glowworm. Te. miṇuku to glimmer, sparkle; n. glimmer, glimmering, sparkling; miṇũgu, miṇũguṟu, miḍũgu, miḍũguṟu spark of fire, firefly; min(u)ku twinkling, twinkle, glitter, flash, ray of light; (K. also) vb. to glitter, shine; minuku-minukum-anu, minukkuranu to twinkle; mincu a flash of lightning, shining, brilliancy; (K. also) vb. to shine as lightning, shine; minna a gem; minamina glitter, shining. Pa. minnal spark. Ga. (S.3munake firefly. Go. (Tr.) mīnkō the stars which a stunned, dazed, or liverish man sees; (W. Ph.) minko, (Mu. Ko.) miṛko firefly; (Mu.) miṛkom, (M.) miṛko, (L.) miḍkos star; (Ma.) minˀkonj(i) (pl. minˀkosku) star, firefly (Voc. 2842); (Tr.) miḍstānā, (W.) mirsālnā, (M.) miṛkānā, (Ph.) mirsīlnā, mirsīltānā to flash, of lightning; (SR.) miḍcānā to flash; (A.) miṛc-, (Ma.) miṛs- to lighten (Voc. 2844); (ASu.) miṛc- to glitter; miṛcval lightning; (L.) mīḍsā, mīrcā id.; (LuS.) meershinta to glitter; meersheetatta lightning; (Mu.) mirŋgul, (Ma.) miṛŋgor̥spark (Voc. 2837); (ASu.) minṅūṛ id. Konḍa (BB) mirs- to lighten. Kuwi (Su.) mṇih- (mṇist-), miṇs- to lighten; (Isr.) mṇīh- (mṇīst-) id., glitter; mṇispu lightening; (S.) mirsi mannai to scintillate; mrih'nai to sparkle; (Mah.) miṇig-to shine. Kur. bīnkō star; bincō firefly. Malt. bínḍke star.

Furthermore, miṇ- 'to glitter, star, sky' is perhaps related to viṇ- 'sky' [DEDR 5396], which also shows a retroflex nasal.  I believe it is clearly wrong to rely on Tamil-Malayalam word miṉ 'flash'to equate 'fish' with 'star', which is fundamental to many of Parpola's readings of Indus seals. I think it is untenable, as the reconstruction for miṇ- 'flash', 'light'  should contain a retroflex consonant, and cannot be equated with miṉ 'fish' which has an alveolar nasal. As Hans Hock once said, if you want to represent a star, why draw a fish?

Similarly, Parpola says the following in his recent book:

 "The Dravidian languages appear to have spread to central and southern India from the area of the Indus Valley. Copper Age cultures of the Deccan, which derive from the Early and Mature Harappan cultures of the Greater Indus Valley, spread farming and animal husbandry to central and southern India, in place of hunter-gathering" (Parpola, Asko (2015). The Roots of Hinduism: The Early Aryans and the Indus Civilization. Oxford University Press).

If Parpola believes that the Dravidian languages moved recently into central and southern India, then, he must agree that they haven't moved into "empty" area. We know there have been Paleolithic population living in South and Central India for at least 20,000 years. Then, there must have been a Dravidian colonization of South India where the pre-existing hunter-gatherer population have gone through acculturation and language-shift.  Then, the Peninsular Dravidian langugaes must have acquired several non-Dravidian features (perhaps some of them are common across Indian linguistic area) through the influence of the languages of the HG (hunter-gatherer) population. That means, it is possible that several archaic features found in Tamil are in fact the archaic features of the aboriginal languages. As Herman Tieken had suggested earlier, Tamil cannot be the language of the southernmost region of Dravidian migration, and at the same time be the most conservative Dravidian language.

Furthermore, Parpola uses Tamil-based reconstructions to argue Dravidian roots for words with voiced consonants such as gardabha,  bhekuri/ bekuri etc. without convincingly explanation for the word-initial voiced consonant in such wide variety of languages. The same Tamil-centric argument is made for  kiyāmbu without explaining their presence in several Indian languages including East-Indian languages such as Assamese, Bengali and Oriya ( CDIAL 3465 kēmuka m. ʻ Costus speciosus ʼ lex., ʻ Colocasia antiquorum ʼ Bhpr. [Cf. kēvuka -- m. Car., kēcuka -- m. Suśr., kacu -- m., kacvī -- f. lex., kēlūṭa -- m. Npr., also ʻ a kind of potherb ʼ Car.: ← Drav. EWA i 266] B. kẽu ʻ Costus speciosus ʼ, Or. kaükaüā, H. keũākeuā m. kēmbuka -- see kr̥muká -- .2609 kacu f. ʻ the esculent root of Arum colocasia. ʼ [Cf kacvī --f., kēcuka -- m. ʻ Colocasia antiquorum ʼ, kēmuka -- ] A. kasu, B. kacu, H. kacūkaccū). He uses palatalization rule to explain kiyāmbu > cēmpu in Tamil, but by the same token he doesn't explain why kinnara doesn't become cinnara in Tamil.

  I believe many of the features of the Indian linguistic area, commonly believed to be originated in Dravidian, in fact belong to the languages of pre-Dravidian substrata (for ex: dative subjects, echo words, reduplication, compound verbs, datives in -k-/-g- etc.). I also believe all the words related to the flora and fauna of mainland India must also belong to the aboriginal languages of India (what I call Niṣādic languages). I hope to provide a stronger evidence for it in a comprehensive essay on this topic.

Regards,
Suresh.

Nagaraj gaaru: You said "The usages of రేవు rēvu are more with /r/ (ర్) than with /ṟ/ (ఱ్). As you know, a lot of medieval grammar books on Telugu discuss the rēpha-ṟa-kāra-nirṇayamu in great detail. They all include ṟēvu (ఱేవు) with ṟ, and not with r (including appakavīyamu). Here is the verse listing the words starting with /ṟ/ (ఱ్) by Tāḷḷapāka peda-Tirumalacāryulu, son of Annamācārya (15-16th century):

ఱట్టు, ఱవికె, ఱంకు, ఱాయి, ఱిక్కించుట
ఱేపు, ఱేవు, ఱేను, ఱేసి, ఱెక్క, 
ఱేడు, ఱెప్ప మూసి  ఱివ్వన నది దాటి 
రాచు డనగా పెద్ద ఱాలు కృష్ణ  (51)
By the way, ఱేవడి is the oblique form of ఱేవడు. "రెంటికి చెడిన ఱేవడి చందాన" is the complete proverb.