Re: Alphasyllabaries, phonology, etc. (correction)
- Just a correction in my statement about bhAguri: al-lopa in bhAguri's statement
should be read at+lopa, and not as containing the pratyAhAra aL. This is what
happens when a grammarian is too sleepy while typing an email.
--- In Indo-Eurasian_research@yahoogroups.com, "Madhav M. Deshpande"
> In addition to George Cardona's demonstration that the ZivasUtras were fullyand
> presupposed in Panini's sUtras, one may also note the strong possibility that some
> earlier forms of these sUtras were part of pre-Paninian grammarians like Apizali
> gAlava. A statement attributed to Apizali, vaha-vyadha-vRdhAM na bhaS, uses the
> pratyAhAra bhaS, in the same sense as Panini's use of it, and the statement ikAM
> yaNbhir vyavadhAnaM vyADigAlavayoH at least suggests the possibililty that the
> pratyAhAras iK and yaN were used by vyADi and gAlava. Another citation: vaSTi
> bhAgurir al-lopam ... ApaM caiva hal-antAnAm..., creates the likelihood that the
> pratyAhAras aL and haL were used by bhAguri. Thus, one needs to be cautious in
> claiming Panini to be exceptional in using a catalogue of sounds like the ZivasUtras.
> This type of use of a catalogue of sounds and short-forms based on it most likely
> predate Panini.
> Madhav Deshpande
> --- In Indo-Eurasian_research@yahoogroups.com, george <cardonagj@e...> wrote:
> > Dear Steve, Thanks for your prompt and detailed answer. I'll answer your
> > two questions in reverse order, for reasons that will become clear.
> > A.
> > SF: The transmission history of the 'received texts' in the Indian
> > linguistics tradition is obvious complex, especially given the
> > complications involved in the interactions between oral and written
> > traditions in India. This leads to real difficulties, as I see it --
> > e.g., in your reference above to the Mahabhasya in respect to
> > traditional sound inventories or the so-called Sivasutras. Does your
> > reference to Keilhorn's edition of the Mahabhasya suggest that our
> > first direct testimony to those tables doesn't come before 'Patanjali'
> > = c. 2nd century BCE?
> > GC: In terms of the actual textual records, yes, Patanjali's MahAbhASya is
> > the first record we have of all these materials put down in manuscripts.
> > However, as with the Vedic materials, it is clear that there was an oral
> > tradition. Moreover, since the ZivasUtras are part and parcel of an overal
> > package, PANini's ZabdAnuZAsanam, which includes also his catalog of verbal
> > bases (dhAtupATha) and sets of nominal/verbal items (gaNapATha), the
> > interdependence shows clearly that the sUtras require the sound inventory in
> > the order known to all students. In my earlier post, I cited the sUtra Gamo
> > hrasvAd aci NamuN nityam as evidence that PANini had reordered an earlier
> > set of nasal stops in the usual order (from velar to labial). The
> > abbreviatory term (pratyAhAra) Gam (in the abalative sg. GamaH and nom. sg.
> > GamuT, which specifies that the sounds in question are introduced as the
> > initials of a word following one that ends in a sound denoted by Gam)
> > presuposses the ZivasUtra set of nasals. Similarly abbreviatory terms like
> > ac (for vowels), ik (i u R L), hal (for consonants), yaN (for semivowels),
> > Zar (spirants) presuppose the same inventory. All such terms are formed by
> > a rule that serves to interpret terms of the type IM: a first elemement I
> > together with (saha) a final (antyena) marker (itA) denotes itself and all
> > intervening elements. So there can be no question but that Patanjali
> > received the list as it has been handed down traditionally. He also
> > discusses possible reorderings, e.g. of reordering the semivowel sequence of
> > h y v r T, so that we know that, like a good discussant he was (along with
> > KAtyAyana), he subjected the received list to considerations of alternatives
> > in view of the target matter to be accounted for. In brief, I don't think
> > we can entertain the thesis that what Patanjali attests to is the result of
> > any invention on his part.
> > B.
> > SF: So I was basing this part of my summary on this kind of data, which
> > does suggest to me that the sound tables were segmental/phonemic and
> > not syllabic. The contrast between them and the segmentally
> > _untransparent_ alphasyllabic structure of early Indic scripts -- the
> > rather jerry rigged nature of the system I refer to in the quote you
> > give from me above -- is another part of the preliminary data I've
> > cited questioning claims that those tables existed before the scripts
> > began to evolve. (Hence the hypothesized importance of the introduction
> > of Aramaic in the NW and the development of Indic scripts.)
> > GC: I agree that sound inventories (akSarasamAmnAya) existed before the
> > scripts known to us from Ashoka and later were devised. I also think there
> > is evidence showing that PANini attests to the syllabic references of the
> > type Ca: 'ka' (with default 'a') and so on. To begin with, note that,
> > although I cited the ZivasUtras in the manner you have quoted, in the work I
> > alluded to I also gave the DevanAgarI version, which has ha ya va raT laN
> > jha bhaJ ... The reason for this is fairly straightforward. As the
> > TaittirIyaprAtiZAkhya makes explicit (TPr. 1.16: varNaH kArottaro varNAkhyA,
> > 1.17: akAravyaveto vyaJjanAnAm [AkhyA 16], 1.21: akAro vyaJjanAnAm), a term
> > naming (-AkhyA) a sound can consist of that sound (varNaH) followed by
> > 'kAra', as in ekAra- 'e'; in a term naming consonants (vyaJjanAnAm 'of
> > consonants'), on the other hand, the sound in question is separated
> > (-vyavetaH) from 'kAra' by 'a', as in kakAra 'k'; and a term consisting
> > simply of a consonant followed by 'a' also serves to name that consonant,
> > e.g. ka 'k'. That is, when the ZivasUtras list ha ya va ra la ..., they are
> > referring to the consonants h y v r l. In the ASTAdhyAyI, PANini clearly
> > makes use of such names of the type Ca. For example, the genitive singular
> > of 'sa' used to name the consonant 's' is sasya, and the ablative rAt refers
> > to 'r' named by the nominal ra, as in the sUtra: rAt sasya (8.2.24);
> > similarly in the sUtra: suT kAt pUrvaH (6.1.135), a heading which provides
> > that, under conditions stated in following rules, the initial augment s
> > (suT) is introduced preceding (pUrvaH) k, the term kAt, ablative singular of
> > 'ka' is used to refer to 'k'. Things are a bit more complex, since PANini
> > also attests---in forms where the consonant is followed by a vocalic
> > ending---a usage such that a consonant by itself refers to a consonant, as
> > in saH sy ArdhadhAtuke (7.4.49), where the genitive singular saH and
> > locative singular si refer to the consonant 's'; these are forms of the type
> > vAcaH, vAci to the consonant-stem vAc 'speech'. Moreover, he also attests
> > to an apparent dialect variation, such that terms of the type Ci are used to
> > refer to consonants. For example, aseH in adaso'ser dAd u do maH (8.2.80)
> > is the genitive singular of asi ('not ending in s'), in which 'si' refers to
> > 's'; in the same rule the abl. sg. dAt (from da-) refers to 'd' and the
> > genitive singular daH (sandhi form 'do') also refers to 'd'; but the
> > nominative sg. maH uses 'ma' to refer to 'm'. All this can hardly be
> > accounted for except under the assumption that PANini too was conversant
> > with the convention of using syllables of the type Ca and so on to refer to
> > sounds. In other words, we have here the same convention that prevails in
> > the usual alphasyllabaries known in India, though with added complications,
> > as noted. Such conventions are clearly attested in early prAtiZAkhya work,
> > and, as I said in my earlier post, it is quite possible that works like the
> > TaittirIyaprAtiZAkhya were pre-PANinian. In turn, given all this, the fact
> > that KharoSThI and BrAhmI also are alphasyllabic system then does not, I
> > think, serve to give any probative evidence of Armamaic influence: by itself
> > this does not establish that Aramaic was the source of either, though I
> > agree that both doubtless were indeed adapted from an Aramaic model; nor
> > does this give any evidence, at least that I can discern, for the alphabet
> > development being key to the development of 'sophisticated' phonology and
> > grammar.
> > On the topic of Aramaic again, let me note one point that has not been
> > brought up so far, I think. In the usual Indian varNamAlAs consonants are
> > referred to generally by syllables of the type Ca (+ Ci). This is true of
> > consonants that can be pronounced with a syllable nucleus following them.
> > On the other hand, some sounds that do not meet this condition are given
> > special names: thus 'anusvAra' and 'visarjanIya' to name 'M' and 'H'. Here
> > too, there are complications of detail, which I omit for brevity's sake.
> > Now, note, that in Aramaic (as in Hebrew and derived alphabets of Europe
> > generally), sounds are given names: alef, bet, gimel, dalet, he, vav, zayin,
> > het, tet, yod, kaf, lamed and so on. Such a naming system is, to my
> > knowledge, not a norm in Indian alphasyllabaries. The most straightforward
> > conclusion to be drawn from this discrepancy is that the naming conventions
> > using 'Ca' and so on in Indian alphasyllabaries is independent of any
> > Aramaic influence. This is reinforced by the fact that the orders of sounds
> > in the two systems are quite divergent.
> > And this brings me full circle back to the main difference I have with
> > Steve. As I indicated in my first post, in my view, which I think I share
> > with George Thompson, although KharoSThI and, doubtless, also BrAhmI, were
> > devised from an Aramaic model, this does not imply that deliberations over
> > developing such writing systems were the trigger for refining 'vague' early
> > Vedic linguistic thought into the 'sophisticated' though one sees in
> > prAtiZAkhyas and PANini. Those who devised these scripts could as well have
> > brought with them knowledge of syllabic terms of the type Ca and
> > phonolosgic/phonetic knowledge from much earlier Indian traditions. I say
> > this without going into details concerning what Steve said in his response
> > to my first post, to which I plan to return. On the other hand, before
> > finishing, I think it worthwhile touching on the point George Thompson
> > brought up: KharoSThI and BrAhmi being devised for Middle Indic. The fact
> > that these scripts are first attested in Ashokan inscriptions speaks in
> > favor of this view. It is certainly noteworthy that one has symbols for
> > initial 'e' and 'o' but not for 'ai' or 'au', dipththongs that were
> > monophthongized to e and o in Middle Indic, or that there is not attested,
> > until later in KharoSThI, a symbol which people conventionally consider to
> > be vocal R: Middle Indic eliminated this also. As for the Sanskritization
> > witnessed to by forms such as genitive singular in -sya instead of -sa,
> > which Richard Salomon refers, this does not necessarily mean that the early
> > developers of the script are to be placed in Achaemenid times. I cannot say
> > anything about the particular document, since I don't have Corpus
> > Inscriptionum Indicarum at home and have to wait until I get to Penn
> > tomorrow to check, I doubt that this is from such an early date; and later
> > Sanskritization in Middle Indic is well known.
> > That's it for the moment. Regards, George (PS This was written in some
> > haste, so please forgive typos and possible infelicities; I hope there are
> > no factual blunders.)
- Dear Nath Rao,
That is indeed the case, but then how do we interpret Katyayana's vArtika "yavanAl
lipyAm". P.4.1.49 (indra-varuNa-bhava-zarva-rudra-mRDa-hima-araNya-yava-
yavana-mAtula-AcAryANAm Anuk" has some clear, and other unclear, cases. By
allowing, for example, indrANI for Indra's wife, the rule seem's to be blocking the
derivation of indrA. However, the tradition tells us that while AcAryANI refers to the
teacher's wife, for a female teacher the word AcAryA applies. I have a feeling that by
the vArtika "yavanAl lipyAm", KAtyAyana is providing a similar clarification: yavanAnI
refers to the script of the Yavanas, while a female Yavana can then be referred to as a
YavanI. Now what would allow YavanI, rather than yavanA? That depends on how one
derives the word yavana within Sanskrit. The traditional way of deriving the word
yavana from the root yu by adding the affix -ana (= lyuT) makes it possible to predict
yavanI, rather than yavanA. The marker T of the affix lyuT is explicitly listed in P. 4.1.
15 (TiDDhANa...). The traditional derivation is seen in the ZabdastomamahAnidhi (p.
320) of Taranatha Tarkavachaspati Bhattacharya, 3rd printing, Calcutta, 1893. Of
course this does not provide 100 percent certainty about what Panini may have
intended, but I would put my money on yavanI for Yavana woman and YavanAnI for the
--- In Indo-Eurasian_research@yahoogroups.com, "nathrao" <nathrao@y...> wrote:
> --- In Indo-Eurasian_research@yahoogroups.com, "Madhav M. Deshpande"
> <mmdesh@U...> wrote:
> > The only thing one can say about Panini with certainty is that
> > he is aware of the Yavana script (lipi) in the NW.
> I thought that the restriction of 'yavanAni' to the script is in a
> vArtikA. Does Panini allow the derivation of yavanI?