The oddities of Basque

[bradrn]

photos like the one at https://florapix.nl/tropical/?gal=bsi&g ... 58#current show what i believe the ancestor of pineapples must have looked like, and although i admit it's hard to get a feel for the size of the plant when no scale is provided, i would assume that the grass is at least within the size range of that of the pineapple and therefore that the fruit is much smaller. and even that seems to be one of the larger ones.

I can’t really reply to the rest of your post, but that doesn’t look all that much smaller than a pineapple flower. From what I can tell, I think it’s safe to say pineapples haven’t changed much in domestication.

[Linguoboy]

I’m with Pabappa here. The pineapple has been domesticated for over 5000 years. There’s zero reason to think it hasn’t been altered as much during that time as other fruit crops of similar long standing, like peaches or melons, and that pre-domestication its fruits looked similar to those of Ananas macrodontes, a close relative.

[bradrn]

I’m with Pabappa here. The pineapple has been domesticated for over 5000 years. There’s zero reason to think it hasn’t been altered as much during that time as other fruit crops of similar long standing, like peaches or melons, and that pre-domestication its fruits looked similar to those of Ananas macrodontes, a close relative.

Well, I’m basing my argument off the fact that, as far as I can tell, A. macrodontes fruit looks roughly comparable to modern pineapples. Of course it’s somewhat thinner and smaller, but that’s only to be expected from domestication; I mostly objected to Pabappa’s description of it as ‘a tiny seed-bearing structure found on the top of a single blade of grass’, which it clearly is not.

[Talskubilos]

This isn't a PIE-native word but a Paleo-European substrate loanword,

How can you tell?
(and is it possible to answer that question without saying "because PIE doesn't exist"?)

Because it has a limited attestation within IE and *b is incompatible with Hittite šam(a)lu, which would point to a protoform *h₃m(V)l- instead. This is the same "laryngeal" found in IE 'eye' and 'nail' and it's compatible with the Uralic and Basque words.

so there's no point in reconstrucing *h₂ here.

why not?

Because this is another case of a-Ablaut, which is rather common in suspected non-native words. I once read a textbook (of which unfortunately I don't keep the reference) where the author(s) used ɑ to represent this vowel.

[Talskubilos]

Discussing historical linguistics with Talskubilos is about as meaningful as discussing evolution with a hard-boiled creationist.

This is why I keep you in my ignored list.

[Talskubilos]

Wiktionary suggests a different 'apple' Wanderwort as well:
Armenian xnjor "apple"
Hurrian ḫenzūru
Akkadian ḫinzūru
Aramaic ḥăzzūrā
Classical Syriac ḥazzūrā
Sumerian ḫašḫur

To which one could probably add Archi änš, Chechen ʿaž, etc.

In fact, Starostin reconstructs a NEC protoform *ʕämćō 'apple: medlar', but excludes the Nakh word you quoted.

[Talskubilos]

Getting back to *porḱ-o- 'piglet', we've got Basque bargo id., apparently a loanword from an IE-centum language which merged /o/, /a/.

[Nortaneous]

PIE *h2ebol- 'apple'

This isn't a PIE-native word but a Paleo-European substrate

How can you tell?
(and is it possible to answer that question without saying "because PIE doesn't exist"?)

It has *b and doesn't fit the usual PIE word structure, unless there's some way to analyze it as *h2eb-ol-.

loanword, so there's no point in reconstrucing *h₂ here.

why not?

It's not reflected in Anatolian in that form, so was likely borrowed after Brugmannization and was originally just *abol-.

It's possible that these tasty, fist-sized apples we love so much in fact arose before human contact, and got so big because they attracted squirrels or some other animal, but even if this is true i suspect they've grown quite a bit larger in human history even so.

An apple is about the size of a pawpaw, which evolved to be eaten by mastodons and/or giant ground sloths, so you'd need megafauna - something at least the size of a horse. But crabapples are pretty small, so it seems likely that the size of the apple is a product of domestication.

Medlars and tejocotes (a type of haw) are smaller than apples, but bigger than crabapples.

But language can tolerate a lot of phenotypic variation. Guavas come in a lot of sizes and colors, but they're all guavas. (I'd expect there to be some languages that merge apples and pears, and I think I've seen some that merge apples and quinces - which is especially easy to do if the local apples are better cooked than eaten raw.)

I would assume those are all cultivars as well. i am skeptical that nature would ever evolve a fruit that grows on a plant that can't physically support that fruit.

Pawpaws are one of the largest fruits native to North America. They grow on tiny, frail understory trees whose branches struggle to support them. But pawpaws aren't very relevant to the reproductive cycle of pawpaw trees - they mostly propagate in clonal colonies, but aren't self-fertile, and it's rare for a flower to be successfully pollinated and even rarer for a tree to bear fruit.

We cant know for certain what went on before the advent of writing, but I think it would be highly unusual for a fruit the size of a pineapple to evolve in nature since the small, seed-pod version would propagate itself just as well and be a million times more efficient at growing without human intervention. Its the same story as with pumpkins, etc .... why would a pumpkin evolve in nature, when a smaller gourd-like fruit could spread its seeds just as quickly with a lot less calorie consumption?

Because bigger fruits attract bigger animals. I don't know what the largest wild fruit is, but Osage oranges are pretty big - and completely inedible. Keep in mind that you come from a continent that underwent an early megafauna extinction event, so the few remaining native large fruits are all evolutionary relics. All other edible fruits in North America outside the historical range of the moose evolved to be eaten by birds, rodents, deer, etc.

There are also a lot of imported fruit-bearing plants, but smaller fruits are preferred for landscaping. (In more recent times, landscapers have preferred to plant only male trees to avoid fruiting altogether - which is good for the stock price of whoever makes Claritin.)

photos like the one at https://florapix.nl/tropical/?gal=bsi&g ... 58#current show what i believe the ancestor of pineapples must have looked like

That's a flower. Ananas sagenaria is an alternate name for Ananas macrodontes.

[Talskubilos]

It's not reflected in Anatolian in that form, so was likely borrowed after Brugmannization and was originally just *abol-.

Would you be so kind as to explain us what do you mean by "Brugmannization"?

[Nortaneous]

It's not reflected in Anatolian in that form, so was likely borrowed after Brugmannization and was originally just *abol-.

Would you be so kind as to explain us what do you mean by "Brugmannization"?

The development from "Proto-Indo-Hittite" of the sort reconstructed by the Leiden school (esp. Roland Pooth) to post-Anatolian "Proto-Nuclear PIE" of the sort reconstructed by Brugmann, with loss of most if not all laryngeals.

The laryngeals actually have to be reconstructed for Nuclear PIE since they're necessary for Indo-Iranian (Iranian preservation of initial *h2-, some details in Vedic that I can't remember), Greek (the triple reflex), Balto-Slavic (tone register), and maybe Germanic (Cowgill's Law), Armenian (double reflex?), and Tocharian (high vowel breaking before *h2/*h3) - but a-coloring had already happened, and without Indo-Iranian there's no reason to reconstruct initial prevocalic *h2 except forward projection of the system of pre-Anatolian PIE.

[Raholeun]

Just to interject, why are reconstructions like *(mb,p)(i,u)t(iu)C 'fingernail', *k(a,o)nd(a,o)[C] 'foot, lower leg', *[si]si, *siti, *pisi 'urine' entertained or even accepted for Proto-Trans–New Guinean, but we are nitpicking Talskubilos for some metathesis here and there? Put differently, why are Pawley and his Papuanist buddies not laughed out of the room for suggesting just underdefined etyma?

NB: I am particularly interested in bradrn's take on the standards of Papuanist diachronic studies and the validity of TNG as a whole.

[bradrn]

Just to interject, why are reconstructions like *(mb,p)(i,u)t(iu)C 'fingernail', *k(a,o)nd(a,o)[C] 'foot, lower leg', *[si]si, *siti, *pisi 'urine' entertained or even accepted for Proto-Trans–New Guinean, but we are nitpicking Talskubilos for some metathesis here and there? Put differently, why are Pawley and his Papuanist buddies not laughed out of the room for suggesting just underdefined etyma?

NB: I am particularly interested in bradrn's take on the standards of Papuanist diachronic studies and the validity of TNG as a whole.

Alas, I fear you will be disappointed with me. I haven’t looked terribly deeply into TNG languages, or even Papuanist diachronics more generally. (If anything, Nortaneous might know more.) But from what I’ve heard, I’d say that at least Nuclear TNG (as circumscribed by Glottolog) is likely to be valid. At least the pronouns (free and bound) seem to match up, there are some cognates in basic vocabulary, and the family is fairly typologically coherent. The real problem is lack of data: Pawley mentions that ‘The largest number of [Proto-TNG] reflexes so far noted for any one language is around 40, for Kalam’, and I think it’s no coincidence that Kalam also has one of the most comprehensive dictionaries of any Papuan language. (Available online, by the way; I think it should be required reading for all conlangers.) I strongly suspect we could far more easily confirm the TNG hypothesis if only we had proper dictionaries across the whole family.

(Also, I think there is a basic methodological difference between Pawley and Talskubilos. Pawley compares lexemes across many languages at once, and thus is forced to state the ambiguity and uncertainty in his reconstructions. Talskubilos only ever seems to compare single word pairs at a time, and these consistently have rather strange correspondences, but it is impossible to demonstrate certainty or lack thereof from only one word pair. Still, I think it would be enlightening to see if Pawley has supplied any regular sound correspondences: I would surely reject TNG if they were consistently as strange as Talskubilos’s.)

EDIT: Clarify some minor points

[Nortaneous]

Just to interject, why are reconstructions like *(mb,p)(i,u)t(iu)C 'fingernail', *k(a,o)nd(a,o)[C] 'foot, lower leg', *[si]si, *siti, *pisi 'urine' entertained or even accepted for Proto-Trans–New Guinean, but we are nitpicking Talskubilos for some metathesis here and there? Put differently, why are Pawley and his Papuanist buddies not laughed out of the room for suggesting just underdefined etyma?

NB: I am particularly interested in bradrn's take on the standards of Papuanist diachronic studies and the validity of TNG as a whole.

Alas, I fear you will be disappointed with me. I haven’t looked terribly deeply into TNG languages, or even Papuanist diachronics more generally. (If anything, Nortaneous might know more.)

I don't know anything because there's no solid work and no dictionaries, which is why Wurm compared pronouns in the first place. But comparing pronouns gets you Mitian, so the error bars are huge. It's pretty hard to find Rotokas-Konua lexical correspondences.*

Talskubilos only ever seems to compare single word pairs at a time, and these consistently have rather strange correspondences

Loans don't necessarily exhibit regular correspondences, so the criteria have to be different. For a Wanderwort, you need a chain of transmission; for a standard loan from A into B, you need a bulk of loans that all have native derivations or regular correspondence in A but not in B. The chart needs to be either wide or tall - single words that look alike in A and B won't cut it.

Sometimes loans do exhibit regular correspondences, though - correspondence strata are useful for Bai, IIRC.

* But there are some:
peruva ~ heruba 'three' (Askopan pɛnuma)
opita ~ ohita 'coconut'
kara ~ hara 'leaf' (Askopan kunuara?)
opoa ~ ehia~eiha 'taro'
puki ~ hukihuki 'mountain' (Askopan pusiko)

visuriko ~ pisirue 'star'
kiuvu ~ kipu 'wind'
-vira ~ -pita 'ADV'
vara ~ para- 'down'
gavuta ~ kaputao 'ashes' (Askopan kautato)
tavauruva ~ kupairipa 'girl'

uvareoua ~ uvateo 'ear'
oira ~ oita 'man' (Aita oida-to)

vuurapa ~ aripu 'good' (Askopan avɛrɛ-pato)
airepa ~ rupa 'new' (Askopan ayɛ-pato)
vurei ~ bopuure 'feast' (bo- could be a definite article)

toua ~ ato 'house' (Askopan tunato)
taavoto ~ tahavu 'shoot'
orupa ~ ori 'feather'
virao ~ virao 'moss'
karoto ~ karo 'rafter'
rui ~ ruisae 'spit out'
ivooru ~ ivoru 'turtle'
ava ~ avau 'go'

aruve-a ~ aruvei 'yesterday' (Askopan anumɛ)
voa ~ ivo 'here'
ivarai-a ~ ibalai 'above'

There are probably some suffixes in here that should be filtered out.

[Creyeditor]

Just to interject, why are reconstructions like *(mb,p)(i,u)t(iu)C 'fingernail', *k(a,o)nd(a,o)[C] 'foot, lower leg', *[si]si, *siti, *pisi 'urine' entertained or even accepted for Proto-Trans–New Guinean, but we are nitpicking Talskubilos for some metathesis here and there? Put differently, why are Pawley and his Papuanist buddies not laughed out of the room for suggesting just underdefined etyma?

NB: I am particularly interested in bradrn's take on the standards of Papuanist diachronic studies and the validity of TNG as a whole.

The problem with TNG is that low level reconstructions are missing. Pawley's work might actually stimulate work on low level reconstructions.

Also, I have noticed the strange habit with some Papuanists to only reconstruct phonemes that are robustly attested in daughter languages. So, whereas in other areas a regular correspondence /p/<->/mb/ might be reconstructed as *mp, they might write *p/mb, because no daughter language has voiceless prenasalized stops.

[Talskubilos]

**EDITED**

Getting back to the main topic, Proto-Basque had a limited number of labial phonemes in its inventory and certainly no /m/, so in most cases this consonant is a secondary development in Basque, namely before back vowels /a, o, u/, especially at word initial. To complicate matters, in the these contexts, original /m/ was apparently assimilated to /n/ in Paleo-Basque and so it further evolved to /h/ or zero, usually leaving nasalized vowels in the Easternmost dialects: Zuberoan and (the now extinct) Roncalese.

On the other hand, I already mentioned labials weren't allowed at word final in PB*, and sometimes (although not regularly) disappeared at word initial, being replaced by prosthetic liquids /l-, -r(-)/. Either prosthetic or not, /r/ was replaced by by /d/ (dissimilation) when followed by another rhotic, so in short this is the explanation why the later would correspond to an original /m/, as in uda, udare.

*Spanish phonotactics does the same, so in loanwords such as e.g. the toponym Catalayud < Qalˁat Ayyūb, /b/ was replaced by /d/.

[Talskubilos]

Sorry, I missed that. In that case it does seem a great deal more plausible.

Not really. (1) As has been pointed out, "horse" has a good internal etymology in PIE, (2) the missing nasal in IE is not accounted for and (3) while developments from velar -> to sibilant / affricate are well attested, the reverse development is not. So even if we assume the words are related and ignore the internal PIE etymology, the reconstructed "*ɦɨ[n]tʃwi (~ -e)" cannot be the original form, but must go back to an older preform with a velar.

Right. The 'horse'-word (just as the 'wheel'-word, BTW) has a good etymology within PIE - it is simply a thematicization of *h₁oḱus 'swift' -

Wait a moment! This is precisely the etymology proposed for Latin aqua 'water' and related words in Germanic meaning 'river', so both connections can't be possible at the same time.

[WeepingElf]

Not really. (1) As has been pointed out, "horse" has a good internal etymology in PIE, (2) the missing nasal in IE is not accounted for and (3) while developments from velar -> to sibilant / affricate are well attested, the reverse development is not. So even if we assume the words are related and ignore the internal PIE etymology, the reconstructed "*ɦɨ[n]tʃwi (~ -e)" cannot be the original form, but must go back to an older preform with a velar.

Right. The 'horse'-word (just as the 'wheel'-word, BTW) has a good etymology within PIE - it is simply a thematicization of *h₁oḱus 'swift' -

Wait a moment! This is precisely the etymology proposed for Latin aqua 'water' and related words in Germanic meaning 'river', so both connections can't be possible at the same time.

I'm sorry, but your objection is a non sequitur. A language may have multiple words with different meanings from the same root, especially if it has borrowed words from a related language (which may have been the case with aqua, which perhaps is from the hypothetical Southwest IE language of the Bell Beaker people which may underlie the Old European Hydronymy). This is called a word family, and every language has such.

[Travis B.]

Right. The 'horse'-word (just as the 'wheel'-word, BTW) has a good etymology within PIE - it is simply a thematicization of *h₁oḱus 'swift' -

Wait a moment! This is precisely the etymology proposed for Latin aqua 'water' and related words in Germanic meaning 'river', so both connections can't be possible at the same time.

I'm sorry, but your objection is a non sequitur. A language may have multiple words with different meanings from the same root, especially if it has borrowed words from a related language (which may have been the case with aqua, which perhaps is from the hypothetical Southwest IE language of the Bell Beaker people which may underlie the Old European Hydronymy). This is called a word family, and every language has such.

This is like how healthy and whole are closely related etymologically.

[Talskubilos]

This is like how healthy and whole are closely related etymologically.

Not so according to std etymologies.

[Zju]

This is like how healthy and whole are closely related etymologically.

Not so according to std etymologies.

what do you mean?