LanGo Institute

It looks as though the only unaccented vowel that may surface after [y] (from *n-y) is [a] ; the others, namely high /i u/, are syncopated

*p+mo > bo : Neg. apó
(Pres.)

At the left end of a verb, sequences of consonants are avoided via vowel copying,

while at the right end, they are avoided by deletion--at least synchronically.

Forms in Yásin suggest the 3s ending was simply *-i (identical to the 3s prefix), and -imi arose later.

3s séymi écyi
pl-subj séyman (?)écyan

Endings -umo (3f), -imi (3s), -uman (pl) appear to have been truncated: their initial vowel is removed.

When these abbreviated forms are added to an obstruent (here cy), they further lose their /m/:

*écy+mo > écyo
*écy+mi > ecyi

This parallels 'be' 3f ...

2s sénuma éttuma
3f sénumo éttumo
3s sénimi éttimi
pl-subj sénuman éttuman

The Future adds these nonpresent endings to the Imperfective stem, characterized by *y-suffixation. Note *n-y yields [y] and *tt-y yields [cy] (palatal affricate).

1s séyam écyam
2s séyma écyuma
3f séymo écyo
...

#Burushaski
Húnza dialect

has related sets of endings for (i) the Preterite and (ii) the Future tense, for which a good label is "nonpresent."

I'll illustrate with the verb stems 'say' sén- and 'do (3s-obj) étt-. The Preterite adds these endings to the Aorist (unmarked stem):

1s sénam éttam
...

I'm curious re the Neg. forms to these stems for 'make/cause to be'

2s ddukósman-
3f ddumósman-
1p ddimésman-

#Burushaski
Húnza dialect

We find modern sequences like -rk- alongside -rg-, but I don't know of *rkh.

In my view the protolanguage had only voiceless stops (+asp and –asp), so my new hypothesis is that
*-rk- > rg (voi assimilation)
*-rkh- > rk (deasp. in a cluster)

#Burushaski
Yoshioka 2012:ccxviii

Support for the idea that Pl. ending [-nc] is a syncopated form of [-muc]:

'glacier' ghamú has the pl. forms ghamú-nc ~ ghamú-muc

#Burushaski
Yoshioka 2012:ccxvi

'scrape off, gnaw'
gukór- (Type 1 stem)
{i}khókur- (Type 2)

Suggesting an underlying form *khu(Q)kuQr-

or perhaps the *Q metathesizes; or the /o/ is underlyingly a mid vowel.

Here too deaspiration kh>k feeds the initial voicing rule k>g, suggesting Húnza forms

Compare 1s

'I will say' séyam

'I will come' zrúcram

where [+am] is invariant

#Burushaski
Húnza dialect, Future tense

There is, iirc, a discrepancy in the endings for e.g. 3f subj:

'she will say' séymo
< *sén-y-(u)mu, perhaps with syncope after the palatalized nasal

'she will come' zrúcro
< *zrú-cy-mu

Whether the /m/ of *-mu surfaces seems to depend on the Ipfv. form
<

Compare, from root 'be/do' man-,
meíme (or meéyme)

for
*man-y-úm-e

Maybe a better spelling for 'drinking' would be *mií(y)me

... accent on mora 2:
*mi[n]-y-úm-e

-cy/y marks Imperfective

-um Adnominal (adj.)

-e likely Essive

There's also 'eat (y obj)' syé- (Ptcp. nusyén ?), notably lacking a coda.

A common formula concluding a narrative is 'I came (back) eating and drinking'

je(-0) syécyume míyme ddáayam

syé-cy-um +e

exactly parallels míyme:
*min-y-um +e

whose accent tho defies explanation atm--would expect ...

But note that in
sén-as, sén-am etc. /n/ is also onset, not coda!

The verb 'know' has unprefixed hén- : Neg. akhén-, with the same pattern, therefore likely *khan- at an earlier stage.

What is the participle? Would expect (*)nukhén.

Are there unprefixed (Type 1) verbs with initial /kh/ ?

*-lttélan- may be ruled out simply because the /l/ is not a coda but the onset of the following syllable.

What sén/san, barén/paran, ttalén/lttálan share is that a>e only occurs between coronal consonants...

so why not *-lttélan- ?

Does it have to be between a coronal onset and coda /n/ ?

Shâri Bâno # 131 has
'do not say' oósan Prohib. (neg. imper., zero ending, but no deletion of final coronal)

[oó] is the Húnza leveling of /a+ú/, Neg. prefix plus Null element

'say' therefore has allomorphs sén- (tonic), -san- (atonic), parallel to

*paran- 'look'
barén-as : apáran : nupára(n)

#Burushaski UNT Archive, Shâri Bâno # 127f

'she arrived late' 3f Aor.
mu-kháran-umo

Ptcp. numúkara
i.e. /n+ mu+ kharan/

vowel copying -> accent mora 2 -> deaspiration (kh -> k) in an unaccented syllable (regular for Húnza verbs);
final coronal deletion (n -> 0)

#Burushaski
Húnza dialect

Aside from /ltt/ which behaves like a unitary phoneme, in the verb system clusters are broken up via vowel copying.

But this is only true in the domain of moras 1 and 2. E.g. 'arrive/reach'

3s ddésrqalttas
<sr> retroflex fricative
<dd, tt> dental stops (not geminates)

Always happy to talk about Pānini!

Been a few years since I've studied it, but he's been tremendously helpful in his approach to grammar

On verb roots, indicatory Ñ shows that both active (Parasmai) and mediopassive (Ātmane) endings are used, rather than just one set

zero grade for the root, rather than the default Guņa

So from the root svap- 'sleep' the participle (suffix -Ktá) is suptá.

Another IT marker is P, which says "this affix shall not bear the accent."

'press' Pres. active 3rd dual is sunutás, vs 3rd sg (suffix -tiP) sunóti

It would indeed. In my attempt to model the alternation Q started out as purely a lowering element; then I realized it would tidy things up if it could itself bear the accent

Pānini uses segments with indicatory purposes in his grammar. When an affix is marked with K for instance it selects ...

Note the variant forms for '3'

iskí : uskó

Likely another instance of vowel copying, this time with a cluster intervening. Underlying forms perhaps

/ sk.í : sk.úQ /

I regard the ó as derived from /u/, since we don't have *oskó

Though Burushaski has ergative marking, 1s-Agent is marked absolutive in future clauses

Transcription: Retelling of Shari Bano from Fairyland Transcription of the recording of Laila Khan, a Hunza Burushaski speaker, reciting the story of "Shari Shaskulum Shari bano" ('Shari Bano of Fairyland'), a popular story, with translations in English.

digital.library.unt.edu/ark:/67531/m...

#Burushaski archive, Shâri Bâno #106

'I will marry (take) you if you make me say 3 words within 3 days'

iskí-kuc xáa
three-day until

uskó bar-ing
three word-PL [accent uncertain]

dd-áa-gus-uma
you caused me to utter

je-0 gu-chú-cy-am
I[-ABS] will take you

This may not be the best approach to the forms of 'say' sén- : Ptcp. nusé(nin)

iirc the HB Neg. is oósan-. This e/a alternation occurs also in 'look' barén- : Neg. apáran- : Ptcp. nupára(nin)

and 'appear' (? iirc) ttalén- : Ptcp. nulttálan

In all these cases accented é alternates with [a]