#indo-uralic

Eastern Khanty has an ablaut system with the following paradigm:

alta ‘to extend’ ~ ultăm ‘I extended’ (perfect) ~ ïltï ‘extend!’ (imperative)

kat ‘house’ ~ kutăm ‘my house’ (possessed noun)

This is just one of the 7 possible ablaut patterns in eastern Khanty, which are: a- u- ï;  ä- i- i; ɔ- u- u; ɔ̈- ü- ü; o- ă- o; ö- ě- ö and e- ě- e

I know there is very little reason in Uralic to reconstruct any ablaut based on this at the proto-Uralic level. And several attempts have been made to explain this phenomenon as an umlaut in Ob-Ugric by some unattested suffixes.

However, within an Indo-Uralic framework, it makes a lot of sense to link this to PIE ablaut and to reconstruct ablaut at the Indo-Uralic level. I have come up with the following scheme:

A-grade corresponds to PIE E-grade

I-grade (imperative) corresponds to PIE zero-grade

U-grade (perfect, possessed noun) corresponds to PIE O-grade

The o-grade perfect has been attested in numerous branches of Indo-European. And a zero-grade imperative has been attested in Ancient Greek:

ḗimi ‘I go’  ~ ithí ‘go!’ (2nd person singular)

phḗmi ‘I speak ~ phathí ‘speak!’ (2nd person singular)

Also, the PIE optative (which became the imperative of Balto-Slavic) takes a zero-grade. In general one can suppose that the zero-grade was future-oriented. This can also be seen in the paradigm of the night words

nekʷ ‘time of shimmering’  → Hittite nekuz ‘evening’

nokʷ ‘time the shimmering is over’  → PIE  * nokʷts ‘night’

n̥kʷ ‘time it will shimmer’  → Germanic *unhtwo ‘last part of the night’ (Dutch ochtend, German Uchte, Icelandic ótta)

Ancient Greek also shows an interesting parallel for the possessed nouns:

patḗr ‘father’ ~ eu-pátōr ‘having a good father’ ~ a-pátōr ‘not having a father, orphan’

phrḗn ‘mind’ ~ sṓ-phrōn ‘having a sound mind, sane’

Due to these semantic parallels between the ablaut in Eastern Khanty and Indo-European, I am looking into if it is possible to reconstruct this to Proto-Uralic and even Proto-Indo-Uralic. In order to do that, I need to find how the sounds of the Eastern Khanty ablaut system relate to the Proto-Uralic sounds.

For example, the ablaut pattern a-u-ï is fed from the following Proto-Uralic vowel combinations:

PU *a-a

PU *o-a

PU *ï-a

PU *ï-ï

Now it is tempting to match these ablaut grades to the closest vowels in Proto-Uralic. That would give the following scheme:

Proto-Khanty *a ~ Proto-Uralic *a

Proto-Khanty *u ~ Proto-Uralic *o

Proto-Khanty *ï ~ Proto-Uralic *ï

But such a scheme is nothing more than a guess. Fortunately, I don't have to guess. There are plenty of words where Khanty has an u or ï in the first vowel. By looking at those correspondences, it is possible to get a better sense of what the real correspondences were. And for this ablaut class, much of the work has already been done by Mikhail Zhivlov [Zhivlov2014].

He reconstructs the following correspondences that do not show an A-grade:

Proto-Uralic *a-a2  → Proto-Khanty *ï, sometimes u

Proto-Uralic *ï-a2  → Proto-Khanty *ï, sometimes u

Now there is one issue with PU *a-a2  → PKh. *ï, and that is that there are more examples with PU *a-a2  → PKh. *u. So maybe the main correspondence here is PU *a-a2  → PKh. *u. So let's look at the examples with PU *a-a  → PKh.* ï. These are:

1. Palatized by č, ń

PU *čača 'to be born' → Kh. *čïčVm 'bear cub', Ms. šūšəm 'bear cub' [Zhivlov2014 p.127; Honti1982: p.132 #93; UEW: p.52 #95; HPUL: p.552; Aikio2014a: p.14–17; Aikio2015: p.54]

PU *kačka 'to bite' → Kh. *kïč 'to ache, nettle', Ms. kūš [Zhivlov2014 p.128; Honti1982: p.145 #218; UEW: p.641 #1258; Aikio2014a: p.5–8]

PU *ńanča 'to stretch' → Kh. *ńïṇč­ 'to stretch', Ms. ńūnš [Zhivlov2014 p.128; Honti1982: p.172 #466; UEW: p.323 #637; Aikio2014b: p.84–85; HPUL: p.546]

PU *sańśa 'to stand' → Kh. *Lańć, *Lïńć 'to stand', Ms. tūńć [Zhivlov2014 p.129; Honti1982 p.138 #151; UEW: p.431–432 #873; HPUL: p.549; Aikio2015: p.55]

2. Ablauting/Alternating

PU *aŋa 'to undress, to open' → Kh. *ïŋk, aŋk, aŋăk 'to undress', Ms. uuŋk [Zhivlov2014 p.128; Honti1982 p.127-128 #47; UEW: #16 p.11; Aikio2002: p.50; HPUL: p.542; Aikio2015: p.54] 

PU *kad₂a 'to leave' → Kh. *kïj 'to leave', *kăj 'to leave behind', Ms. kuuľ [Zhivlov2014 p.128; Honti1982 p.146 #227; UEW: p.115–p.116 #221; HPUL p.537-538; Aikio2015: p.54]

3. Only attested in Western Khanty

PU *kara 'to dig' → Kh. *kïr 'to dig' [Zhivlov2014 p.128; UEW p.221–222 #427, #428; Aikio2015: p.55]

Compare this to the examples with PU *a-a -> PKh. u

PU *amma  'to scoop' → Kh. *um 'to scoop', Ms *ūm [Zhivlov2014: p.127; Honti1982: p.126 #32; UEW: p.7–8 #10; Aikio2015: p.53]

PU *aŋča  'nelma' → Kh. *uṇč, Ms. *ūš [Zhivlov2014: p.127; UEW: p.339 #669]

PU *ćarma 'wild animal' → Kh. ćurăm 'hermelin, weasel' [UEW: p.490 #986]

PU *kala 'fish'  → Kh. *kul 'fish', Ms. *kūl [Zhivlov2014: p.128; Honti1982: p.148 #245; UEW: p.119 #228; HPUL: p.538; Aikio2015: p.54]

PU *pala  'bit, to gobble' → Kh *puḷ 'bit', Ms. *pūl [Zhivlov2014: p.128; Honti1982: p.176 #512; UEW: p.350 #695; HPUL: p.540; Aikio2015: p.55]

PU *palwa 'village'   → Kh *puγăl, Ms *pï̄wəl [Zhivlov2014: p.128; Honti1982: p.175 #502; UEW: p.351 #697; HPUL: p.548]

PU *panča 'to open' → Kh. *puṇc 'to open', Ms. *pūnš [Zhivlov2014: p.128; Honti1982: p.177 #523; UEW: p.352 #701; HPUL: p.548; Aikio2015: p.55]

PU *pata  'pot' → Kh. *put 'pot', Ms *pūt [Zhivlov2014: p.128; Honti1982: p.180 #550; UEW: p.358 #710; Aikio2002: p.50; Aikio2015: p.55]

PU *śarta 'deer, young reindeer' → Kh. *surtï, Ms. *sūrtï [UEW: p.464 #936; Honti1982: p.186 #607]

PU *tarka 'twig, conifer' → Kh. *tuγăr 'conifer' [UEW: p.510 #986]

PU *wanča 'to step over' → Kh *wuṇč 'to cross a river', Ms. *wunš [Zhivlov2014: p.128; Honti1982: p.194 #689; UEW: p.557 #1115; HPUL: p.551; Aikio2002: p.36–38; Aikio2015: p.56]

PU *wara  'edge' → Kh *wur 'edge, mountain ridge', Ms *wur [Zhivlov2014: p.128; Honti1982 p.196 #708; HPUL: p.551; Aikio2006: p.27–28; Aikio2015: p.56]

Also, this aligns nicely with the Mansi sound-correspondence PU *a  → PMs. ū.

The PU *ï-a2  → PKh. ï change has less examples, but it is clear that PU *ï-a2  → PKh. ï is the main correspondence here:

PU *ïla 'below'  → Kh. *ïl 'below', Ms. *jal [Zhivlov2014: p.133; Honti1982: p.199 #737; UEW: p.6 #7; HPUL p.536; Aikio2015: p.59]

PU *ïkta 'to hang'  → Kh. *ïɣăt [Zhivlov2014: p.134; UEW: p.5 #6; HPUL: p.536;  Aikio2015: p.59]

PU *jïka 'to divide' → Kh *jïγăḷ 'fork', Ms. *jïɣtət 'neck' [UEW: p.87 #157; HPUL: p.552]

PU *kïčča 'smelly' → Kh. *kïč 'mould', Ms. *kïšγï [Zhivlov2014: p.133; Honti1982: p.146 #216; Aikio2014a: p.5–8]

PU *sïksa 'siberian pine'  → Kh. *LïγăL 'siberian pine', Ms *tït, *tajt- [Zhivlov2014: p.133; Honti1982: p.136 #130; UEW: p.445–446 #903; HPUL: p.540; Aikio2015: p.60]

PU *wïlka 'descend, go down'  → Kh. wïγăl ~ waγăl 'to descend', Ms. wï̄ɣəl ~ wajl [Honti1982: p.192 #676; UEW: p.554 #1110; HPUL: p.551; Aikio2015: p.61]

Versus PU *ï-a2 -> PKh. ï with m-:

PU *mïksa 'liver'  → Kh. *muγăL 'liver', Ms *mït, *majt­ [Zhivlov2014: p.133; Honti1982: p.163 #382; UEW: p.264 #519; HPUL p.538; Aikio2015: p.59]

PU *mi̮tka 'passage'  → Kh *muɣăt 'river branch' [Aikio 2015a: 13–14; Aikio2015: p.59]

So this gives us the following correspondences:

PU *a-a2  → PKh. u (U-grade)

PU *o-a  →  PKh. a

PU *ï-a2  → PKh. ï (I-grade)

PU *ï-ï  →  PKh. a

One thing that is unclear is what is going on with the PU *V-a1  → PKh. *a correspondence. Maybe this is a class of words that were ablauting in the second-syllable vowel and Khanty copied that ablauting vowel to the first syllable.

Also, the correspondences between the Khanty ablaut grades and PU vowels are different for each different ablaut class. This can be shown by the main correspondences for the I-grade ă of the o-o-ă ablaut class: PU *a-ï -> PKh. ă. Here the I-grade vowel is clearly PU *a(-ï). While for the a-u-i the I-grade vowel is clearly PU *ï(-a).

More speculatively, this seems to point to a system where you have multiple vowels. These vowels were modified according to the ablaut grade. An umlaut solution like Eugene Helimski proposed for Ob-Ugric [Helimski2001] might work for Indo-Uralic. I would like to note here that the ablauting suffixes proposed by Helimksi are present in PIE:

perfective *-w in the perfective participle *-wos 

possessed *-w in the possesive adjective *-went

future/imperative *-i in the optative *-yeh₁

I have been dabbling into the comparison of the Indo-European and Uralic language families for some time. But now I am at the point where I feel that I have scratched the surface, and there is so much below it. It has been said that Indo-Uralic is a project that could take a dedicated linguist more than 30 years to work on. I'm starting to see that this is right. There are so many ideas to work out, but as an amateur I can't work full-time on this. So here are of some ideas that I’m working on now or plan to work on in the future.

Linking the ablaut in eastern Khanty to the PIE ablaut.  Eastern Khanty has an ablaut system with the following paradigm:

alta ‘to extend’ ~ ultăm ‘I extended’ (perfect) ~ ïltï ‘extend!’ (imperative)

kat ‘house’ ~ kutăm ‘my house’ (possessed noun)

This is just one of the 7 possible ablaut patterns in eastern Khanty, which are: a- u- ï;  ä- i- i; ɔ- u- u; ɔ̈- ü- ü; o- ă- o; ö- ě- ö and e- ě- e

I know there is very little reason in Uralic to reconstruct any ablaut based on this at the proto-Uralic level. And several attempts have been made to explain this phenomenon as an umlaut in Ob-Ugric by some unattested suffixes.

However, within an Indo-Uralic framework, it makes a lot of sense to link this to PIE ablaut and to reconstruct ablaut at the Indo-Uralic level. I have come up with the following scheme: 

A-grade corresponds to PIE E-grade

I-grade (imperative) corresponds to PIE zero-grade

U-grade (perfect, possessed noun) corresponds to PIE O-grade

A zero-grade imperative has been attested in Ancient Greek:

ḗimi ‘I go’  ~ ithí ‘go!’ (2nd person singular)

phḗmi ‘I speak ~ phathí ‘speak!’ (2nd person singular)

Also, Ancient Greek does show an interesting parallel for the possessed nouns:

patḗr ‘father’ ~ eu-pátōr ‘having a good father’ ~ a-pátōr ‘not having a father, orphan’

phrḗn ‘mind’ ~ sṓ-phrōn ‘having a sound mind, sane’

The difficult part here is Proto-Uralic. If this is done right, it should be possible to find the correspondences between PIE ablaut grades and the PU vowels. So this is a key idea for me. 

For example, the Eastern Khanty, a-ï-u pattern probably goes back to PU a-ï-o. The simplest idea here might be a vertical vowel system (a ~ PIE e; ë ~ PIE o; ï ~ PIE zero grade) with front/back; labialized with 'w'/unlabialized; short/long variants in PU. But if you know anything about the vowel correspondences in PU, you would know what a huge task it is to sort everything out.

An ofshoot of that little project is the idea that disharmonic stems like Finnish likoaa 'he washes' should be reconstructed as lïuka (~PIE *lewh₃) in Proto-Uralic.

The correspondence between PU plural -t and PIE plural -s. Also PU 2nd person -t and PIE 2nd person -s. I think this will require its own correspondence set if it is ever going to work as a correspondence. I don't think that Kortlandt's idea of Finnic-like *ti -> *si assibilation holds. My idea that this goes back to an Indo-Uralic **z (or maybe **ð, but that is already taken in Uralic). This PIU **z would correspond to *t, *s and *r in both PIE and PU but under different circumstances.

Potential cognates with this set include:

PIU **zïwxa 'pig' ~ PU *tika 'pig' ~ PIE suH 'pig'

PIU **näz 'nose' ~ PU *näri 'nose', nistä 'pant, blow'  ~ PIE neh₂s 'nose’

PIU **maz- 'wet' ~ PU *mośkï(1) 'to wash', PUg *mar-/*mär- 'to dive' ~ PIE *mesg(1) 'to dip', *mori 'sea'.

PIU **z 2nd person ~ PU -t 2nd person marker, PSam -r-,-l- 2nd person marker ~ PIE -s- 2nd person marker, PIE th2- 2nd person marker

PIU **z plural ~ PU -t plural marker ~ PIE -s plural marker, ? PIE -r 3nd person plural marker

(1) These may also be linked to a PIE root *meh₂ 'wet'.

PIU consonant gradation (low prio) There are lots of roots in PIE where there are variants with different stop grades but similar meaning, e.g. *keh₂l ‘to call’, *gʰel ‘to shout, to yell’, *gels ‘to call, voice’ or *bʰer ‘to carry, to bear’, *per ‘to travel, to fare’. Could these be caused by a consonant gradation mechanism on the PIU level?