Good to hear from you, Madhav, 

A check of Mayrhofer’s 1992 and 2001 etymological dictionaries shows that most of the forms with the “wrong” sibilant are of uncertain or controversial origin; some may be borrowings from other languages or from Prakrit (with hypersanskritization); and only bhāṣ ‘speak’ may be inherited (if the explanation that s à to avoid homonymy with bhās ‘shine’. Forms like paryaṣasvajata obviously result from analogy (based on forms like pari-ṣvajati). (See details below.)

As for Fortunatov’s Law, I have discussed this in an earlier paper, as well as the assumption thata there was a variety of Vedic that retained the distinction between *r and *l; Mayrhofer further concludes that there is no evidence for such a variety in all of early Indo-Iranian. (See details and references below.) 

I also have a discussion of the issue of saṇakāra vs. aṇakāra, where I argue that external-sandhi retroflexion gets degeneralized in the history of Vedic and that it the sandhi (as in RV rājati ṣṭúp) must go back to pre-Rig Vedic. (Again, a reference is found below.) 

Mayrhofer’s and my papers contain extensive references to and discussion of earlier literature. Unfortunately, I’m not aware of any more recent detailed discussions of these issues. 

All the best,

Hans Henrich

Mayrhofer (to the extent that he has anything in his 1992 and 2001 dictionaries)

áṣatara: evidently corrupted

kavaṣa: Uncertain

cā́ṣa: No convincing etymology

jálāṣa: Uncertain meaning; no etymology provided

caṣā́la: Perhaps dissimilated from *carṣā̆la (see carṣ/karṣ)

váṣaṭ: No etymology given; I have speculated that ritual distortion has affected the form)

bhāṣ-: Controversial. Some derive this from *bhels- (cf. Lith. bal̃sas ‘voice’) ± Fortunatov’s Law; others think of homonym differentiation from bhās- ‘shine’

mā́ṣa: Problematic; compare perhaps MPers./NPers. maš ‘legume’, Shughni max̌ ‘pea, bean’, hence Proto-(Indo-)Iranian *marṣ̌a ?

jhaṣá: Uncertain; probably loanword

ṛbī́sa: Uncertain

kīstá: Loanword (unlikely) or hypersanskritization of *kīrtha > MIAr. kittha, with tth reinterpreted as corresponding to Skt. st.

kúsindha: Uncertain etymology

kaṣ: a colloqial form of karṣ- [i.e. a borrowing from Prakrit?]

kusuma: uncertain

kisalaya: see kiśalaya

bisa: Uncertain; note Pkt. variant bhisa.


Additional comments

None of these forms has a clear, uncontroversial etymology. Even though Mayrhofer tends to rule out borrowing in most cases, such a possibility should not be dismissed (except, probably, for bhāṣ); consider especially the word for ‘bean’.

mā́ṣa: Could this be one of those Central-Asian substrate words that Lubotsky has talked about? Lubotsky himself classifies it and its Iranian counterparts as a “wanderwort”. (The Indo-Iranian substratum. Early Contacts between Uralic and Indo-European: Linguistic and Archaeological Considerations. Papers presented at an international symposium held at the Tvärminne Research Station of the University of Helsinki 8-10 January 1999. (Mémoires de la Société Finno-ougrienne 242.) Chr. Carpelan, A. Parpola, P. Koskikallio (eds.). Helsinki 2001, 301-317.)

 paryaasvajat: Analogical extension of from forms like pari-ṣvajati

 Beyond aṣṭā(u), see also naś- ‘reach’ : aṣṭá, naś ‘perish’ : naṣṭá, vaś ‘wish’ : váṣṭi, sah ‘be powerful’ : sāḍhá, spaś ‘look’ : -spaṣṭa


Some further literature:

Kobayashi, Masato. 2004. Historical phonology of Old Indo-Aryan consonants. Tokyo University of Foreign Studies, Research Institute for Languages and Cultures of Asia and Africa. – §98-99; the latter on the controversial Fortunatov’s Law, but not referring to relevant discussion by Hock

Hock, Hans Henrich. 1991. Dialects, diglossia, and diachronic phonology in early Indo-Aryan. Studies in the historical phonology of Asian languages, ed. by W. G. Boltz & M. C. Shapiro, 119-159. Amsterdam/Philadelphia: Benjamins. – On Fortunatov’s Law see especially §4.9.

Hock, Hans Henrich. 1993. A critical examination of some early Sanskrit passages alleged to indicate dialectal diversity. Comparative-historical linguistics: Indo-European and Finno-Ugric: Papers in honor of Oswald Szemerényi III, ed. by B. Brogyanyi & R. Lipp, 217-232. Amsterdam/Philadelphia: Benjamins. – §5 on saṇakāra vs. aṇakāra

Mayrhofer, Manfred. Zur Vertretung der indogermanischen Liquiden in den indo-iranischen Sprachen. Indologica Taurinensia 149-161.  – Like Hock for Sanskrit (and referring to him), finds that there is no evidence for retention of contrast between r and l in Indo-Iranian as a whole.


On 29 Aug2021, at 19:49, Madhav Deshpande <> wrote:

Hello Hans,

     I have lost track of some of the relevant old publications, but I remember that some of the occurrences of ṣ in Sanskrit were accounted for by Fortunatov's law regarding the IE l+dental changing to retroflex in Sanskrit, and some others may be what Thomas Burrow called spontaneous retroflexes. Are some of your examples [other than ruki and oḱtō > aštā ‘eight’, covered by these theories?
     The other indication to suggest the instability of ṇ/ṣ is the discussion in the Aitareya-Āraṇyaka about whether the RV Saṃhitā was aṣakāra/aṇakāra or saṣakāra/saṇakāra. The Āraṇyaka says that the Māṇḍūkeya version of the RV was saṣakāra/saṇakāra, and that Śākalya followed Māṇḍūkeya in this respect. But the discussion itself indicates that there may have been other reciters whose Saṃhitā was aṣakāra/aṇakāra. 


Madhav M. Deshpande
Professor Emeritus, Sanskrit and Linguistics
University of Michigan, Ann Arbor, Michigan, USA
Senior Fellow, Oxford Center for Hindu Studies
Adjunct Professor, National Institute of Advanced Studies, Bangalore, India

[Residence: Campbell, California, USA]

On Sun, Aug 29, 2021 at 1:55 PM Hock, Hans Henrich via INDOLOGY <> wrote:
Dear Colleague,

Even as early as the Rig Veda there is evidence, both for  occurring after a-vowels and for s occurring after i- and u-vowels. See the evidence further below.

What made the distribution of s and  unpredictable is the fact that Proto-Indo-Iranian š, the source of Skt.  is of two sources. One if the development of earlier s to š after “RUKI” (i.e. r-sounds, u-sounds, velars, and i-sounds; in the case of the vocalic sounds, both syllabic and nonsyllabic); the other was the development of PIE *ḱ to š before obstruent. Examples are nis- > niš ‘down’ and oḱtō > aštā ‘eight’. 

As the second example shows, the second of these changes introduced š after a-vowels and thus made the RUKI outcome of s opaque and hence contrastive (consider e.g. Skt. asta- ‘thrown’ beside aṣṭā(u) ‘8’, with s and  contrasting after a-vowel. 

This contrastiveness, in turn, made it possible for analogical processes to extend  into contexts after a-vowels (as in pary-a-ṣasvajat) as well as for borrowings and the like with  after a-vowels and s after “RUKI” to be adopted without further adjustment.

All the best,

Hans Henrich Hock
Linguistics and Sanskrit (emeritus)
University of Illinois

Contrastiveness of retroflex sibilant in Sanskrit

Unpredictable occurrences after a-vowels in the RV 

áṣāḍha ‘invicible’

áṣatarā ‘more beneficial’ (1.183.4)

kaváṣa (PN) (534.12)

cā́ṣa ‘Häher’ (923.13)

jálāṣa ‘healing’ (1.43.4 in compound)

caṣā́la ‘Knauf der Opfersäule’ (1.162.6)

váṣaṭ (ritual call) (passim) 

Note also

paryaasvajat (pluperf.) ‘embraced’ 

Contrastive and unpredictable examples after a-vowels in later Vedic

mā́ṣa ‘bean’

mā́sa ‘moon, month’

bhāṣ- ‘speak’

bhās- ‘shine’

jhaṣá ‘large fish’ 

Some Post-Vedic examples after a-vowels

kaṣ- ‘rub, scratch’

kas- ‘go, move’ (DhP)

laṣ- ‘desire’ (MBh etc.)

Dental sibilant (s) after i- and u-vowels in Vedic

ṛbī́sa ‘cleft, gap’ (RV)

kīstá ‘singer’ (RV)

kúsindha ‘trunk’ (AV)

Some examples of ental sibilant (s) after i- and u-vowels in Post-Vedic

kisalaya ‘sprout, shoot’

kusuma ‘flower’

bisa ‘shoot, sucker’








On 23 Aug2021, at 14:11, Jim Ryan via INDOLOGY <> wrote:


A question: I go back to a memory (possibly incorrect) of hearing from a linguistics teacher at UW (long ago) that the retro-flex "ṣ" in Sanskrit was "barely phonemic." A  former student who had studied, through his Ph.D. exams, historical linguistics at UCLA focusing on Indo-European (maybe also Indo-Aryan) insisted that this sound was not phonemic. From time to time I'd encounter the issue in articles/books and found that the consensus seemed to favor this understanding. I used to challenge my student from time to time to test this, somehow, I suppose, wanting to vindicate my long ago teacher's position (or at least what I thought I recalled it to be). I've thought recently of two examples: the verbal root bhāṣ - “to speak.” and ṣaṣ (six). In neither case is there a "non-vowel" preceding the sibilant, which would ordinarily condition retroflexion. In the case of "six,"  the ṣ is initial also.  How do we explain these instances in accord with the non-phonemic nature of ṣ?

Jim Ryan

Asian Philosophies and Cultures (Emeritus)
California Institute of Integral Studies
1453 Mission St.
San Francisco, CA 94103

INDOLOGY mailing list;!!DZ3fjg!pkLQA4HtQOusbNyeaxRGvL3ZJLP3OBpn0ZKjsaLFQaEFKjusvnVZNc-NFX4aZ_0mmg$

INDOLOGY mailing list