#so like. the finnish front vowels have umlauts. for example

[elysianasterism]

the umlaut and the diaeresis refer to different kinds of sounds...

[ayearinlanguage]

A Year in Language, Day 34: Proto-Germanic

Proto-Germanic is the ancestor of all Germanic languages. The exact dates of its split with Proto-Indo-European (PIE) raises some issues with the nature of language change, but it probably began developing in the 2nd-1st millennium BCE and was spoken around 500 BCE.

Proto-Germanic is not a language we have any direct evidence of. The closest is the attestation of Romans, which would have been around the time it began splintering into the various Germanic branches and some ancient loan words in Finnish and other nearby language groups.

The defining feature of the Germanic languages is a sound shift known as Grimm's law, named for Jacob Grimm who formalized it, the same Jacob Grimm who compiled fairy tales with his brother. Here's the simplified version of it: in PIE there was a three way distinction in stop consonants. There were voiceless stops (/p/, /t/, /k/), voiced stop (/b/, /d/, /g/), and aspirated voiced stops (/bʰ/, /dʰ/, /gʰ/). Grimms law shifts all of these in turn; first the voiceless stops become fricatives instead (/p/ > /f/, /t/ > /θ/, /g/ > /x/ [note: /θ/ is like English "th" and /x/ is like German "ch"]). The voiced stops then are pulled by this vaccum and become voiceless themselves, and the aspirate voiced ones tumble after becoming voiced, losing the aspiration. Compare, for example, Latin "ped" to English "foot" or "dent" to "tooth"

Proto-Germanic did not undergo umlaut, another unique feature common in Germanic languages that causes front vowels to round, and is generally responsible for the large vowel inventories of those languages (compare English's 15 or so vowels to Spanish's 6). It also had nasal vowels, largely absent from modern Germanic languages.

[ululem]

Indo-Uralic, so many ideas, so little time.

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?

1st person absolutive -w versus oblique -m The idea is that the me-we split in PIE is not singular/plural but absolutive/oblique.  The original 1st person nominative singular pronoun would have been 'u' in early PIE, as attested by Hitt. u-uk (<- *u-eǵ), and various 1st person singular -w forms in ancient PIE languages. This would parallel the bi-/min- split from the 'Altaic' languages. Strangely, this split can even be found in Kartvelian (m- verbal prefix versus v-). In Uralic there is very little evidence of this.  Only some Samoyedic languages have verbal forms in 1st person -w.