Pabappa wrote: ↑Sat Sep 04, 2021 2:18 pm
People seem to like postulating metathesis all over the place in PIE ... in the word for wolf, the word for wheel, one of the words for nudity (in pre-Greek), etc.... I dont believe in any of it myself, but if we can establish that metathesis really did happen in PIE, it would be at least fair enough to apply it to case like this since there are other arguments pointing towards a sound change of /ml/ > /bl/ in Greek and perhaps in other branches.
Metathesis rarely is a regular sound change (e/aRC/# > Re/a(:)C/# in Common Slavic is one of the few I can think off), so each example needs to be looked at on its own. It's good when it can be directly observed (like the variation between
bird / brid in different varieties of English) or if it concerns a widely-attested etymon, without too many other exceptional developments thrown in. With the apple word, there are too many other things on top to make it likely.
Skookum wrote: ↑Sat Sep 04, 2021 3:36 pm
I know Burushaski
báalt "apple, apple tree" is sometimes connected to the IE forms, but this gets us even further away from Western-IE so its probably just a coincidence.
It could possibly belong to the group of words we discuss - words with a structure mVl-/bVl- in the wider vicinity of the area of original domestication of the apple (Central Asia).
Talskubilos wrote: ↑Sat Sep 04, 2021 11:53 pm
hwhatting wrote: ↑Sat Sep 04, 2021 2:09 pmTalskubilos wrote: ↑Fri Sep 03, 2021 12:03 pm
I don't see the link between "warm" and "apple".
'warm (season)' > 'ripe (fruit)'.
hwhatting wrote: ↑Sat Sep 04, 2021 2:09 pmIs actually a word meaning “apple” or “fruit” derived from the Nakh-Dagahestani lexeme attested in Eastern Caucasian?
Not necessarily so, but it would provide the "missing link" for IE
*mah2l-o-
A potential semantic development is not much of a missing link.
Talskubilos wrote: ↑Sat Sep 04, 2021 11:53 pm
hwhatting wrote: ↑Sat Sep 04, 2021 2:09 pmThe Hittite form could belong if we assume the
ša- is another prefix from a substrate.
Not exactly. The Uralic form points to a "laryngeal"
*h3 here, just like
sākuwa- 'eye' < IE
*h3ekʷ- 'to see' and
sankuwāi- 'nail; a unit of linear measure' < IE
*h3n(o)gh-
I see; this would add one more word to a small and problematic group (
š is not a regular continuation of laryngeals; Kloekhorst assume an s-mobile here, which is also ad hoc and doesn't seem to be attested in other branches). As I don't know much about Proto-Uralic, how does the Uralic form show evidence for a laryngeal? And how does the Uralic r/n-stem relate to the -l- in other instances of this wanderwort?
Talskubilos wrote: ↑Sat Sep 04, 2021 11:53 pm
hwhatting wrote: ↑Sat Sep 04, 2021 2:09 pmI also don't see how the Basque belongs here - how do you get from something like
mVl- to the Basque forms?
Apparently,
m disappeared and was replaced by a (seemingly) prosthetic
d, as in
ahari 'ram' (with the dialectal variant
adari) <
*a-mari or
adar 'horn' <
*kam-ar-. On the other hand,
udare, udari would be related to
uda 'summer'.
That looks basically like "everything can come from everything". It would be better if you applied the rigour and skepticism you apply to other people's theories also to your own theories.
bradrn wrote: ↑Sun Sep 05, 2021 1:14 am
Talskubilos wrote: ↑Sun Sep 05, 2021 1:10 am
As I said before, there's no such verb 'to dig'.
I was referring only to the semantic development. I know too little about PIE to properly assess your claims about the existence of a verb ‘to dig’.
Talskubilos is on relatively good ground here. The main support for the root is indeed nouns meaning “furrow” etc. (the English word is one of these), while a basic verb is only attested in Lithuanian (LIV 475), with a meaning “to hurt (with a stinging pain)” – something which might go back to “dig”, but it would be good if one had more examples of that verb in other branches, with a meaning closer to “dig”. Still, there is also Lithuanian
pra-paršas “ditch”, which doesn’t represent the zero-grade formation and needs to be explained if there wasn’t originally a verb
*perk’- “dig”; likewise for Vedic
párśāna- “depression, trough”.
@Talskubilos: On pigs - if you google “traditional pig breeds”, you’ll find a lot of multi-coloured and mottled sorts (I see now that this point has already been made by keenir). That’s more original than the uniformly pink pigs that have become ubiquitous nowadays. It’s also reasonable to assume that the further you go back, the more similar domesticated pigs were to wild boars.
AFAIK, pigs were domesticated in Anatolia about 13,000 years ago. That’s long before PIE is assumed to have been spoken, so there is no reason to believe that to the IEans, pigs were a novelty. But new words being derived especially for young animals and then becoming the word also for the grown animal is a well-documented process. So as long as we cannot document the chain of transmission between the origin of this supposed wanderwort and the instances in IE and Austrobesian, I find an internal derivation more convincing. That’s different to the “apple” word, because there is a range of similar words around the known origin of its domestication, the untypical structure of PIE *abVl-, and the fact that it has a /b/, which natively normally comes only from *-pH3- and is often a sign of loans.
Talskubilos wrote: ↑Sun Sep 05, 2021 8:36 am
The thing is the +2000 lexical items reconstructed for PIE haven't got the same Ablaut nor derivative patterns, IMHO because they belong to different linguistic strata.
On derivational morphology – the laws of IE derivation are well-known. There is actually a lot of regularity and semantic patterns can be reliably recognized. That said, if you look at derivational patterns in existing languages, you often find parallel formations using different patterns, and it’s also not always clear why this or that pattern was chosen.
Also, one shouldn’t be get too hung up on roots. What we have in the IE languages are words, from which roots can be abstracted; ablauting root languages work in that certain developments have lead to different ablaut patterns in different words derived from an original base, and the speakers abstract a core from this (the “root”) and apply the patterns to it. For example, spoken Arabic has abstracted a root
t-l-f-n from the loan “telephone” and derives verbs from that by inserting the corresponding vowels. So, having various ablaut or derivative patterns to form words doesn’t make PIE a “conlang”, it just behaves according to its type.
(So, when you are making the point that certain words only or mostly are attested as derived zero-grade nouns and take that as an indicator that these are loans, you are basically using the absence of derivations with a wide range of derivative patterns as an indicator - but that is only a valid point if using a wide range of derivations is a characteristic of inherited lexicon, which I think is true.)
