5. The original alphabetic order of Indus, if read from left to right, was this:

a. It is presumed that the nasal group also contained a sound M. At least, the nasals were a group of six, M standing for all of them.

b. This was, needless to say, redrafted and grouped by the Sanskrit school as under: 

1. a i u,  2. R L ,  3. e o,  4. ai au,  5. ha ya va ra,  6. la 7. 5-na ma 3-na  Na na,  8. jha bha, 9. gha Dha dha,  10. ja ba ga Da da,   11. kha pha cha Tha tha ca Ta ta,   12. ka pa,   13. s'a Sa sa   14. ha.

6. About the Indus palatal stops (ca cha ja jha) something needs to be said specifically. It appears that the Indus palatal stops were dialectally pronounced also as affricates, where the stop and the following friction produced not only a single sound but also a cluster (cf. Gleason, Descriptive linguistics, p. 248 para 15.10). That is to say,/c/was/ts'/ and /j/ was /dz'/ in specific dialects of Indus.

a. Only /ts'/ value for /c/ can explain the Sanskrit ca (and) corresponding to the Greek te (and). It is supposed that the Indus clause ci ha, pronounced /ts'i ha/ was raised to /ts'e ha/ before being reduced to te (through *teh) in Greek and Punjabi, to just -h in Gothic (Benfey), though the Sanskrit ca and Latin que show /c/ as a stop liable to interchange with a velar. 

b. Though the Vedic jAmAtár (son-in-law) corresponds to the Av. zAmAtar, the modern Persian dAmAd presupposes an old Persian *dAmAtar. Here we have to assume an Indus jAmAtar pronounced *dzAmAtar. Similarly, Skt. jinAti (oppresses), corresponding to Av. zinAiti (harms) and OP adInam (took away) presupposes an Indus base ji-, pronounced dzi-.

(i) Corresponding to the Old Persian  daushtar `friend', Modern  Persian dost, there is joSTár `lover' in Vedic. Here the original Indus base jau seems to have an affricate, /dz / for / j/. At the agglutinative stage of Indus, there was a clause jau-Sa tara meaning `gives love'. But at the isolating stage, even jau Sa was a clause meaning 'gives joy'. This jau Sa became a phrase, then a lexeme, at the agglutinative stage, when it formed a new clause *jau-Sa tara (gives love friendship) at the agglutinative stage. Later, at the inflexional stage, it was reduced to *jau-s-ta'r, which became daustar in Old Persian. joSTa'r or jo'STar in Vedic.

(ii) The Modern  Persian mi-dAn-am (I know) and the Vedic jAnAmi presupposes an original Indus verb jAn, pronounced /dzAn-/ in the Iranian branch. Then the Vedic root jnA (know) appear to be the result of contraction of the original *jAn. This also agrees with our view that tan was variously extended as *tan-o, tan-A, etc., and the Vedic verbal group su-no-ti, lu-nA-ti, bhu-na-kti, etc., arose out of them analogically. Thus, jAnAti was actually jAn-A-ti rather than jA (for jnA) -nA-ti.

c. Even the aspirates /ch/ and /jh/ must have affricate values. That is why, cha is treated as a cluster in RV (e.g., ga-cha-ti, etc.) for which Panini had to insert a regular c (making ga-c-cha-ti). When the Vedic aha'm corresponds to Avestan azem, Old Persian adam, the basic form appears to be ajham, pronounced adzham. Similarly, the Vedic ha'sta (hand), Avestan zasta and Old Persian dasta appears to have a basic jhasta, pronounced dzhasta. In the western dialects jh interchanged with gh, behaving like a stop.

(I) These ch/ts'-h/ and jh/dz-h/ should have the same phonetic value as ch-h in church-hill and dge-h in judge-hat respectively as typical clusters. Later this cluster-value turned into a single integral sound, as in other voiceless aspirates (kh, th, ph) and necessitated the insertion of a c before ch.

(II) An example of /ch/ being a cluster is seen in the sandhi of t + s' = ch. Here a pure stop (t) followed by a friction (s') results into an affricate (ts') with aspiration. Therefore in the S'B (1,1,1,10), we see that yat + s'amI-dhAnyam = ya t-s'amIdhAnyam = ya chamI-dhAnyam. The interposition of a -c- was not felt to be necessary at all.

Even in the RV (5,78,9), mAsAn + s'as'ayAna (H) = mAsAn t s'as'ayAna(H) = mAsã chas'ayAna(H), t-s' becoming the cluster ch does not require the interposition of -c-. It appears that /ch/ was a cluster (aspirated t s') even as late as the S'B. But by the time of Panini  /ch/  had turned up a single affricate sound /ch/, which required the interposition of a -c. The affricate ch/ts'h/ is reduced to just s' before na in pra-s'-na and vi-s'-na. Often, however, ch = tsh, e.g. in u-cha-ti corresponding to which there is vas 'to shine'. 

(III) The Indus voiced aspirate jh / dzh / is represented mostly by h in Vedic, but very often it reverts to its original form in derivation: 

gAh	gAjh + ta dzh + ta dzh + dha gA-Dha
gUh	gaUjh + ta dzh + ta dzh + dha gU-Dha
vah	vajh + ta u dzh + ta u dzh + dha  U-Dha

The Vedic ahi is azi in Avesta. Its Indus form could beajhi   \adzhi\, meaning 'it obtructs', its vedic equivalent being vRtra (from vra-tra).           

d. The behaviour of  an Indus palatal stop is then manifold:

(I) Its c is a pure stop, when interchangeable with the corresponding velar in some dialects (cf. Latin que for ca). In Hindi, c is a stop, because in bec-nA (to sell): bik-nA (to be sold) it alternates with k.

(II) It is also an affricate when the two elements  (the stop and the following friction) make a cluster dialectally. But in the later dialects  only a single element has been left: As a rule, c  /ts'/ is represented by s in Av.  but it is a stop /c/ in Vedic: Av. hasi  for haci from sac (ved. sacate).

(III) The aspirate ch  /ts'h/ has been retained as a cluster in the RV. Thus there is only ga-cha-ti,not ga-ccha-ti as later.

(IV) The voiced j /dz/ is represented by d in OP, by z in Av. and by the stop sound j in Vedic.

(V) The aspirate jh  /dzh/ is represented by d in OP, by z in Av. and by h in Vedic.

 

Copyright: INDUS SCRIPT 2001-02
Site Concept By Sumit Mishra