The Great Proto-Indo-European Thread's Sequel

[Richard W]

The root *k^wel- 'to turn' seems to have a pretty good Mitian pedigree, with Turkic *Kul-, Mongolic *kok-ki-, Tungusic *xol/xul-, all with meanings related to 'to turn' (according to An Etymological Dictionary of the Altaic languages by Starostin et al., p. 850; I know this is not all too reliable), and perhaps Uralic *kulke- 'to go'.

It's not the root *k^wel- 'to turn' that's at issue; it's the apparently derived word meaning for 'wheel'. The connection to the root may be folk etymology, and the variations in precise formation makes one wonder if the word had spread through (P)IE. It could even be a dying word that has been repurposed, revived and spread.

[cedh]

English "turn up" [(unexpectedly) end up being at] seems like a reasonable semantic intermediate for both "go" [end up being at] and "live" [habitually be at] - but it's not clear to me how to get from "turn" to "take care of" or "serve"

Maybe something like "turn" > "turn to" > "look around for" > "look after" > "take care of" > "serve"?

[Ephraim]

But there's apparently a homonymous verb 'to dwell', attested in Latin colō 'to live, to inhabit', in-cola 'inhabitant', in-quil-īnus 'tenant'.

De Vaan derives colō from *kʷelH- "to turn", but annoyingly doesn't give the semantic motivation. I'd say a development "to turn" > "to go in circles" > "to stay in the same place" > "to inhabit" isn't unreasonable. But note that the Latin verb also means "to take care of", and Latin anculus "man-servant" can be derived from Pre-Latin *ambi-kʷolos, with *kʷelH- as its second element, which is identical to Greek ἀμφίπολος "(female) servant", as well as very close to Sanskrit abhicara- "servant".

English "turn up" [(unexpectedly) end up being at] seems like a reasonable semantic intermediate for both "go" [end up being at] and "live" [habitually be at]

For the meaning ‘inhabit’, an alternative route may be something like this:
turn (> go) > become > stay > inhabit
or perhaps
turn (> go) > be (in a location) > stay > inhabit

I'm not saying this is the correct answer, because how would we really know... But note for example that Greek πέλω, which is probably from the same root, has the following meanings from what I can tell:
1. to be in motion, to go (the oldest sense which may be rare but according to Middle Liddell appears in Homer)
2. to be
3. to become
https://lsj.gr/wiki/πέλω
https://en.wiktionary.org/wiki/πέλω#Ancient_Greek

Sanskrit cárati, also probably from this root, also has ‘to go’ as one of its core meanings.
https://en.wiktionary.org/wiki/चरति#Sanskrit
https://www.sanskrit-lexicon.uni-koeln. ... p?page=389

For the semantic shift ‘to turn’ > ‘to go’, also note that English suppletive past went is derived from a verb that used to mean ‘to turn’. The semantic shift ‘to turn’ > ‘to become’ is likewise found in in English turn as in “turn green”.

It makes some intuitive sense for a verb meaning ‘to be (at a certain location)’ (see πέλω above) to develop the sense of staying at a location for a certain time, and then staying at a location habitually. It's less intuitive for a verb meaning ‘to be’ to develop this sense, but note that Portuguese ficar means both ‘to become’ and ‘to stay, remain’, as well as ‘to be (in a location’). Swedish bli (older bliva) likewise has both the sense of ‘to stay, remain (at a location)’ (perhaps even to the point of dwelling there) and ‘to become’, although only the latter meaning is common in modern standard Swedish (the former is found archaically and dialectally). It is worth noting, though, that in both Swedish and Portuguese, ‘to stay’ is the older meaning, so this is the opposite shift from the one suggested above.

but it's not clear to me how to get from "turn" to "take care of" or "serve"

One route may be to derive the sense of ‘serve’ from the idea of literally turning one's face towards a thing or a person, and metaphorically turning one's mind or attention to a task, thing or person. I think it's pretty common for words for ‘servant’ or ‘to serve’ to be derived from words for watching, guarding or generally giving attention to. This may be similar to what cedh suggested above.

Another route may be to derive it from the idea that the servant is someone who follows a person around, or generally someone who is staying or being around. This might make some sense in this case if the root developed both the sense of ‘going’ and ‘being’.

For a somewhat parallel formation, see Gaulish ambaxtos, borrowed into Latin as ambactus ‘vassal, dependent; servant, retainer’, as well as into Germanic giving Icelandic ambátt ‘female slave, bondwoman, handmaid’. The amb- is the same element as in anculus, while the second element may be from PIE *h₂eǵ- ‘to drive’.

https://en.wiktionary.org/wiki/ambaxtos#Gaulish
https://en.wiktionary.org/wiki/Reconstr ... c/ambaxtos
https://en.wiktionary.org/wiki/Reconstr ... c/ambahtaz
https://en.wiktionary.org/wiki/ambactus#Latin
https://en.wiktionary.org/wiki/ambátt
https://en.wiktionary.org/wiki/Reconstr ... pean/h₂eǵ-

[WeepingElf]

I wish to address a different topic.

So far, I used to think of Archaic PIE (the pre-ablaut stage) as a language with three vowels (*a, *i and *u; ancestral to the ablaut series *e~o~0, *ei~oi~i and *eu~ou~u, respectively) and a penultimate accent as the main conditioning factor of ablaut. This, however, left quite a big messy residue: thematic nouns and verbs, acrostatic inflections, *o in unexpected places, etc. Now I think that there were three further vowels, *â, *î, *û, probably long counterparts of the others, which attracted the accent and surface in "Classic" PIE as non-ablauting *o, *i, *u. The thematic stems were originally *â-stems, while the acrostatic nouns had *â as the root vowel. I think this at least reduces the messy residue somewhat.

[Ares Land]

(Belated) thanks for the answers on PIE vowels.

I wish to address a different topic.

I'm intrigued in these ideas, which intersect a bit with my own questions... What's your own theory on the origin of ablaut?

[KathTheDragon]

I wish to address a different topic.

So far, I used to think of Archaic PIE (the pre-ablaut stage) as a language with three vowels (*a, *i and *u; ancestral to the ablaut series *e~o~0, *ei~oi~i and *eu~ou~u, respectively) and a penultimate accent as the main conditioning factor of ablaut. This, however, left quite a big messy residue: thematic nouns and verbs, acrostatic inflections, *o in unexpected places, etc. Now I think that there were three further vowels, *â, *î, *û, probably long counterparts of the others, which attracted the accent and surface in "Classic" PIE as non-ablauting *o, *i, *u. The thematic stems were originally *â-stems, while the acrostatic nouns had *â as the root vowel. I think this at least reduces the messy residue somewhat.

This is pretty broadly the view I've held for a while now, except that *o < *ā (your *â) actually ablauted (and further, all ablauting *o comes from *ā).

[WeepingElf]

I wish to address a different topic.

So far, I used to think of Archaic PIE (the pre-ablaut stage) as a language with three vowels (*a, *i and *u; ancestral to the ablaut series *e~o~0, *ei~oi~i and *eu~ou~u, respectively) and a penultimate accent as the main conditioning factor of ablaut. This, however, left quite a big messy residue: thematic nouns and verbs, acrostatic inflections, *o in unexpected places, etc. Now I think that there were three further vowels, *â, *î, *û, probably long counterparts of the others, which attracted the accent and surface in "Classic" PIE as non-ablauting *o, *i, *u. The thematic stems were originally *â-stems, while the acrostatic nouns had *â as the root vowel. I think this at least reduces the messy residue somewhat.

This is pretty broadly the view I've held for a while now, except that *o < *ā (your *â) actually ablauted (and further, all ablauting *o comes from *ā).

Maybe, though if ablaut was conditioned by accent and long vowels attracted accent, they should not ablaut. There may be other factors, though; and under such "secondary" ablaut rules, the "zero grade" of *â may be *a. But well, I don't seriously expect to crack that riddle - otherwise I'd have achieved a major breakthrough in IE historical linguistics, on a par with Saussure's discovery of the laryngeals! All I can do is to bang together a homebrew internal reconstruction just good enough to build my Hesperic conlang family on it.

[KathTheDragon]

Likewise! Imo I'd need a more comprehensive reconstruction to base things on (which I'm [very slowly] working on) before I could say with any confidence "this is what I totally believe and here's a 3 paper's worth of arguments why".

[Richard W]

Fellner has complained that the very flat grouping of core IE suggests the Indo-European spread in a 'big bang'. How unreasonable is a 'big bang' as a model of what happened? When I look at the time scales in Gray and Atkinson, they give dates of 7300 BP for Graeco-Armenian to split off, 6900 BP for Indo-Iranian-Albanian to split off, 6500 BP for Balto-Slavonic to split from Western IE, and 6100 BP and 5500 BP for Celtic to split off and Germanic and Italic to split. These dates match the Anatolian homeland theory; if we accept the Kurgan theory, the time scale must be compressed. My feeling is that there was rapid expansion, followed by clumping in the resultant dialect continuum. I also get the feeling that we can't resolve Western IE, and some of the other splits look doubtful to me.

I know the Gray & Atkinson tree has some oddities - Anglo-Frisian disappears (probably because of Scandinavian vocabulary in English and the continental West Germanic dialect continuum). Also not that the position of Tocharian is imposed on the tree derivation; the vocabulary data doesn't seem to force the conclusion that it branched off early!

[KathTheDragon]

Gray and Atkinson can generally be discounted - most mainstream linguists do.

As for the rapid expansion, this is the most likely option. Late PIE was almost certainly a diffuse dialect continuum, meaning no hard boundaries can be dawn between neighbouring lects. The slightly sharper divisions we see today are then just the result of intermediate dialects disappearing.

[WeepingElf]

Gray and Atkinson can generally be discounted - most mainstream linguists do.

As for the rapid expansion, this is the most likely option. Late PIE was almost certainly a diffuse dialect continuum, meaning no hard boundaries can be dawn between neighbouring lects. The slightly sharper divisions we see today are then just the result of intermediate dialects disappearing.

Concurred 100%.

[Zju]

A question I thought of in relation to related discussions: did Ukrainian and/or Russian ever leave PIE homeland, or did Slavic languages at some point displace some other, likely IE, language? Have/has IE language(s) ever left the homeland, or was there at some point linguistic redisplacement, probably of Turkic languages?

[Skookum]

My understanding is that the Ukrainian/Russian steppe was originally home to Iranian languages, which were then replaced by Turkic languages, then by Slavic. This is an oversimplification, since Gothic, Hungarian, and Mongolian speakers also lived in this area at various points, but the general answer to your question is that Russian and Ukrainian represent a back migration into the IE homeland.

[Pabappa]

the Srubna and Sintashta people lived in what is now southwestern Russia before it became Slavicized. Looking up on Wikipedia, it seems that scholars consider them most likely to be Iranian, and it makes sense that they were satem speakers. and Iranian langauges such as Ossetian still survive in the northern Caucasus region.

Im not sure if Ossetian is actually a true survivor of the pre-Slavic population or if they, too, have moved around a lot throughout history.

[Nortaneous]

Would a PIE word for 'twenty' of the form *dḱ-dwoh1-, with the zero-grade of the *deḱ- root found in PGmc *teguz, be at all morphologically possible?

(Context: Albanian -zet 'twenty' is the only solid example of the purported sound law of *wiḱ- > z-, but *ḱ-dwō- < *h1ḱ-dwoh1- would give Proto-Albanian *dz-dwō-; from there, deaffrication would produce *zdwo-, which would regularly develop into ze-.)

[bradrn]

A questions for the Indo–Europeanists here: what happened to the clitics in IE? I know that PIE had a set of second-position (‘Wackernagel’) clitics, which have been preserved quite well in some families and language — Slavic comes to mind as the big one, but I know Pashto has some, as did Homeric Greek. But English, for instance, doesn’t have them. Neither does Latin (to my admittedly small knowledge), nor Celtic. So: what exactly happened to the PIE clitics? Did they simply fall out of use, or is there something more interesting going on?

(Also, related question: How reconstructable is the original clitic set? Do we know exactly what the PIE clitic inventory was?)

[Znex]

A questions for the Indo–Europeanists here: what happened to the clitics in IE? I know that PIE had a set of second-position (‘Wackernagel’) clitics, which have been preserved quite well in some families and language — Slavic comes to mind as the big one, but I know Pashto has some, as did Homeric Greek. But English, for instance, doesn’t have them. Neither does Latin (to my admittedly small knowledge), nor Celtic. So: what exactly happened to the PIE clitics? Did they simply fall out of use, or is there something more interesting going on?

(Also, related question: How reconstructable is the original clitic set? Do we know exactly what the PIE clitic inventory was?)

I'm pretty sure Celtic does though, especially Old Irish for instance. Otherwise, Hittite of course has those, as does Vedic Sanskrit I believe.

[Kuchigakatai]

A questions for the Indo–Europeanists here: what happened to the clitics in IE? I know that PIE had a set of second-position (‘Wackernagel’) clitics, which have been preserved quite well in some families and language — Slavic comes to mind as the big one, but I know Pashto has some, as did Homeric Greek. But English, for instance, doesn’t have them. Neither does Latin (to my admittedly small knowledge), nor Celtic. So: what exactly happened to the PIE clitics? Did they simply fall out of use, or is there something more interesting going on?

Latin had them.

-que 'and'
-ve 'or'
-ne '(introduces polar questions)'
enim 'for' (introduces a new sentence explaining / giving a cause of what precedes it, ~ "this is/was because...")
vērō 'in truth' (with very low semantic meaning really, often it just seems to mark a new sentence)
autem 'but'
igitur 'so, therefore'

And in "Silver Latin" (imperial Latin after Augustus' rule but before the 3rd-century crisis), itaque 'so', earlier an adverb+enclitic meaning 'and so, and therefore', was added to the inventory.

Rather interestingly, these adverb-like clitics could perfectly interrupt noun phrases, simply attaching themselves at the end of the first stressed word. Mīles glōriōsus 'a bragging/lying soldier', Mīlitem enim glōriōsum illum nescīs 'You don't know that lying soldier'. Even proper names can be interrupted, as in Marcus enim Tullius Cicerō, although they can also not be.

When these weren't present, the personal pronouns mihi and tibi could also be placed as Wackernagel position clitics, even if it meant interrupting a noun phrase too.

In Romance, every single one of them was abandoned. And mihi/tibi lost their ability to be in Wackernagel's position too.


Gothic also has at least a couple Wackernagel enclitics (-u '(introduces polar questions)' and -uh 'and'). Ancient Greek continued to have Wackernagel enclitics in the Classical, Hellenistic and late Koine periods. I don't know when they were abandoned in Greek (if we can know that at all, due to written language conservatism: written Latin never abandoned them!).

[bradrn]

A questions for the Indo–Europeanists here: what happened to the clitics in IE? I know that PIE had a set of second-position (‘Wackernagel’) clitics, which have been preserved quite well in some families and language — Slavic comes to mind as the big one, but I know Pashto has some, as did Homeric Greek. But English, for instance, doesn’t have them. Neither does Latin (to my admittedly small knowledge), nor Celtic. So: what exactly happened to the PIE clitics? Did they simply fall out of use, or is there something more interesting going on?

(Also, related question: How reconstructable is the original clitic set? Do we know exactly what the PIE clitic inventory was?)

I'm pretty sure Celtic does though, especially Old Irish for instance. Otherwise, Hittite of course has those, as does Vedic Sanskrit I believe.

True, I had forgot about Hittite (which indeed has a rather elaborate clitic system). I’m similarly unsurprised that Vedic Sanskrit has them as well — most of the conservative IE languages do. But I hadn’t known about clitics in Old Irish, which is rather divergent; do you have any more information on this?

A questions for the Indo–Europeanists here: what happened to the clitics in IE? I know that PIE had a set of second-position (‘Wackernagel’) clitics, which have been preserved quite well in some families and language — Slavic comes to mind as the big one, but I know Pashto has some, as did Homeric Greek. But English, for instance, doesn’t have them. Neither does Latin (to my admittedly small knowledge), nor Celtic. So: what exactly happened to the PIE clitics? Did they simply fall out of use, or is there something more interesting going on?

Latin had them.

-que 'and'
-ve 'or'
-ne '(introduces polar questions)'
enim 'for' (introduces a new sentence explaining / giving a cause of what precedes it, ~ "this is/was because...")
vērō 'in truth' (with very low semantic meaning really, often it just seems to mark a new sentence)
autem 'but'
igitur 'so, therefore'

And in "Silver Latin" (imperial Latin after Augustus' rule but before the 3rd-century crisis), itaque 'so', earlier an adverb+enclitic meaning 'and so, and therefore', was added to the inventory.

Rather interestingly, these adverb-like clitics could perfectly interrupt noun phrases, simply attaching themselves at the end of the first stressed word. Mīles glōriōsus 'a bragging/lying soldier', Mīlitem enim glōriōsum illum nescīs 'You don't know that lying soldier'. Even proper names can be interrupted, as in Marcus enim Tullius Cicerō, although they can also not be.

When these weren't present, the personal pronouns mihi and tibi could also be placed as Wackernagel position clitics, even if it meant interrupting a noun phrase too.

But how ‘2P clitic’-like are these really? =que, for instance, was inserted after the second conjuncted NP. (The famous example is of course Senatus Populus=que; 2P would be *Senatus=que Populus.) As for the rest, though they’re certainly 2P, the orthography seems to indicate phonologically independent words rather than clitics. (The situation seems reminiscent of Tagalog, which also has 2P ‘clitics’ which are really separate words.)

In Romance, every single one of them was abandoned. And mihi/tibi lost their ability to be in Wackernagel's position too.

Interesting! So, if the clitics were abandoned, then what constructions arose to replace them?

[Znex]

True, I had forgot about Hittite (which indeed has a rather elaborate clitic system). I’m similarly unsurprised that Vedic Sanskrit has them as well — most of the conservative IE languages do. But I hadn’t known about clitics in Old Irish, which is rather divergent; do you have any more information on this?

This paper by John Koch addresses early Celtic syntax more broadly, though quite thoroughly, including the development of preverbal elements and the emergence of the absolute/conjunct verb distinction, but does address the treatment of Wackernagel enclitics as well.