(L&N) Q&A Thread - Quick questions go here

[Omzinesý]

What is the word for a language without a continuous VP?

[Creyeditor]

SPlit-VP language sounds about right.

[Sequor]

I wonder when this change happened. It looks like it should have been quite early, give that it gave rise to a Merovingian Latin spelling mistake. Intervocalic /g/ > /w/ before rounded vowels is regular and part of the Western Romance lenition, it is attested in Occitan as well where fagus > fau. So if we suppose this contraction followed lenition, we'd have /agos/ > /aɣos/ > /awos/ > /owos/ > /ows/ in paroxytones and /agos/ > /aɣos/ > /awos/ > /owos/ > /oos/ > /os/ > /s/ or maybe /awos/ > /aos/ > /os/ > /s/ in proparoxytones. But it may have been earlier, since there's words where intervocalic /g/ was lost already in Vulgar Latin Proto-Romance whatever the preferred term for the ancestor of Romance languages is.

I wish we had cognates of cercueil, especially Occitan cognates, but at least Wiktionary doesn't know of any. Anyone knows of anything that may help?

I don't know if the /g/ > */w/ > zero change is really part of the Western Romance lenition: Spanish sometimes retains intervocalic Latin /g/ as /g/ [ɣ], e.g. plagam > llaga and jugum > yugo, or Augustus > agosto and augurium > agüero. However, these exceptions can be possibly explained as an avoidance of phonologically weird *llaa > *lla, *yúo ~ *yuyo, and less confidently an effect of the diphthong [aw] (cf. the maintenance of voiceless /p/ in capuī > *[kawpe] > Old Spanish cope > modern cupe). Less explainable still would be arrugar, from ad + rūgāre.

Survivals of -agus in Spanish may be found in pelagus > piélago and asparagus > espárrago; possibly magus > mago too but that one is hard to state confidently. However, it seems weird not to consider piélago a semi-learned word, considering pelagus is a very lofty and poetic word in Latin, even if on the other hand, piélago is exactly what you'd expect from regular sound changes (intertonic syncope doesn't happen with /a/ in Spanish, cf. Malaca > Málaga, stomachum > estómago, passerem > pájaro, sabbatum > sábado, raphanum > rábano, although note colaphum > golpe, comparō > compro). Meanwhile, asparagus shows up with an unetymological trill in Spanish, and with syncope in Portuguese: Portugal espargo, Brazilian aspargo.

[All4Ɇn]

Are there any particular reasons why the most common title for the Virgin Mary in Italian ended up as My Lady (Madonna) while in other European languages it ended up as Our Lady (Notre Dame, Nuestra Señora, Unsere Liebe Frau, etc.)?

[Sequor]

Are there any particular reasons why the most common title for the Virgin Mary in Italian ended up as My Lady (Madonna) while in other European languages it ended up as Our Lady (Notre Dame, Nuestra Señora, Unsere Liebe Frau, etc.)?

I wouldn't think there are any particular reasons, madonna being simply what became popular in Italian. Compare English "oh my God!" and Spanish ¡Dios mío! ("my God!"), even though in these languages people normally say "the Lord, our Lord" and el Señor, nuestro Señor.

I love your observation though, and it reminds me of some similar cases:

Western European God/Gott/Dios/Dieu/Dio (used with no article)
Greek ὁ Θεός, Arabic allāh (reduced form of al-ilāh, lit. "the-god")

(for the respectful term of address, used for a person)
English m'lord (from "my lord"), Spanish señor mío ("my lord", archaic), French monsieur /məsjø/ (reduced form of /mɔ̃ sjœʁ/ "my lord")
Portuguese o senhor (lit. "the lord"), Spanish el señor (not that common but sometimes used)

[sangi39]

Are there any particular reasons why the most common title for the Virgin Mary in Italian ended up as My Lady (Madonna) while in other European languages it ended up as Our Lady (Notre Dame, Nuestra Señora, Unsere Liebe Frau, etc.)?

I wouldn't think there are any particular reasons, madonna being simply what became popular in Italian. Compare English "oh my God!" and Spanish ¡Dios mío! ("my God!"), even though in these languages people normally say "the Lord, our Lord" and el Señor, nuestro Señor.

I love your observation though, and it reminds me of some similar cases:

Western European God/Gott/Dios/Dieu/Dio (used with no article)
Greek ὁ Θεός, Arabic allāh (reduced form of al-ilāh, lit. "the-god")

(for the respectful term of address, used for a person)
English m'lord (from "my lord"), Spanish señor mío ("my lord", archaic), French monsieur /məsjø/ (reduced form of /mɔ̃ sjœʁ/ "my lord")
Portuguese o senhor (lit. "the lord"), Spanish el señor (not that common but sometimes used)

It does seem to be somewhat a matter of popularity. Nostra Signora is still used in Italian for the Virgin Mary, apparently, but much less so than Madonna. I'd wonder if there might be a sort of contextual difference as well, e.g. is Nostra Signora used in more formalised contexts while Madonna is a more general-use title.

[KaiTheHomoSapien]

So if Latin had gotten its word for "daughter" from *dʰugh₂tḗr, would the word have been "fūter"?

I'm just thinking about alternative Latin again :P

[Salmoneus]

So if Latin had gotten its word for "daughter" from *dʰugh₂tḗr, would the word have been "fūter"?

I'm just thinking about alternative Latin again :P

It survived apparently only in Oscan, where it's "futír". De Vaan says that this probably meant a long /u:/. He gives a Proto-Italic fuxte:r. However, his little list of sound changes appears to suggest it 'ought' to be fuxate:r (where /x/ is probably voiced allophonically). In the note on the Oscan word, he says "apparently the PIE laryngeal remained unvocalised in Sabellic" (so supporting the idea that it 'should' have an /a/ in the middle, but oddly doesn't); he suggests that this is because of the following /tr/ cluster in the oblique forms.

But whether that's in Proto-Italic, Prot-Sabellic, or just unique to Oscan (presumably via analogical extension from the oblique forms), there's no way to tell.

I can't find any good sound changes that would tell us what happened to -xt betwenn PIt and Latin, unfortunately. ght in general seems to have hardened to ct... I THINK de Vaan would say that this was via xt, but it's possible it's via another route. He gives for example PIE forms like 'legh-to' (or maybe 'legh-tu') becoming Latin 'lectus'.

So, the Latin probably "should" be s. fuhater, pl. fuctre:s. But 'fuhater' may well have been replaced with analogical fucter. Alternatively, the /h/ in 'fuhater' would be gone in the speech of many people, giving fuater.

And if you want a wild, irregular derivation: -ghu- often got reanalysed as -xw-, which gives /v/ instead of /h/. That shouldn't happen here, but it's conceivable that the preceding /u/ might irregularly have rounded the following /x/ (particular as it's a rare sequence), in which case you'ld end up with fuvater.

I think?

[KaiTheHomoSapien]

Thanks for the answer. fuhater/fuctres seems a little unsual, though. Are there any examples of Latin nouns whose nominative singular contains an intervocalic /h/ and the other forms show /k/?

[Salmoneus]

Thanks for the answer. fuhater/fuctres seems a little unsual, though. Are there any examples of Latin nouns whose nominative singular contains an intervocalic /h/ and the other forms show /k/?

I doubt it - there's not that many with intervocalic /h/ to start with. It's a naturally bizarre word, thanks to the CHC cluster and the ablaut pattern. Hence many languages generalising the oblique form (where CHCC > CCC regularly). In fact, most branches did this: at least Iranian and Balto-Slavic, possibly Italic (certainly Oscan), possibly Celtic (certainly Gaulish), Armenian, and I'm not sure whether Germanic did or not.

So this analogical "fucter" form could have arisen in northwestern PIE, in Italo-Celtic, in Italic, in Latin-Faliscan (in parallel to to Sabellic) or in Latin itself. So yes, it probably would, one way or another, have ended up as "fucter", even though it "should" (i.e. by regular sound changes) be "fuhater".

[Pabappa]

i think the Latin reflex of intervocalic /gh/ was /g/, not /h/.

https://en.wikipedia.org/wiki/Proto-Ita ... velopments

unless the largyneal somehow intereferes with it
edit: its hard to say, actually .... youre right according to the first line, but then why are there so few (nearly zero) words with medial /h/ other than compounds? also, it would have been a voiced /h/, which i'd think wouldnt have survived.

also: i think italic had its own version of Grassmann's law such that diaspirates always lost one of the two asps. perhaps there was a subordinate law that made medial /gʰ/ always lose it when a root had two asps, thus accounting for the surprisingly low number of roots with medial /h/? if not, i think the second of two aspirates always deaspirated.

---
wait: the PIE root wasnt even a diaspirate. i think the medial consonant would be /g/ regardless of any other factors, even the largyneal.

[Salmoneus]

edit: its hard to say, actually .... youre right according to the first line, but then why are there so few (nearly zero) words with medial /h/ other than compounds?

First the root has to end with gh. Then the root has to NOT have a medial /l/, /r/ or /n/ preceding the final. Then the root has to NOT have a consonant-initial suffix attached. Then the word has to survive down into Latin.

De Vaan lists only 9 PIE roots that meet all those criteria:
- dheigho- gives Oscan feíhúss. Other derivations in Latin have /g/; de Vaan explains these as analogies from 'figula' (where an /l/ strenghtens the /G/ and 'fingo' (where the /n/ does likewise).
- ghe-ghou- does not give an /h/, because the reduplication has been lost, giving fu:di:, rather than reduplicated *fehu:di:
- hleghu gives levis, because the u-stem *lexu- has been reassigned to the i-stems (which was common), and the resulting /xw/ regularly becomes /v/
- hmeghio gives the dative pronoun mihi:, as expected
- hmeigie- does indeed seem to be an exception: we get meio, rather than expected meiheo. De Vaan doesn't comment, but we can guess that the presence of /j/ on both sides of the /h/ caused its loss, either through regular cluster simplification, or through haplogy, or through /xj/ becoming its own phoneme that merged with /j/.
- leghe- only survives in its participle form, lectus
- seghuero- gives seve:rus - again, /xw/ regularly gives /v/
- sm-ih-ghes-l-ih gives mi:lle. Apparently one theory involves loss of /h/: *mi:hi:lle > mi:lle.
- uegho- gives veho, as expected

To these we can add traho - De Vaan is perplexed by the PIE origin, if there was one, but it must have had -gh, and gives /h/ as expected.

So there's only seven words that 'should' have this gh > h change, and five of them do, and there's another word that does in Oscan (but doesn't survive in Latin). Meanwhile two of them don't, and one of them is maybe understandable (/hj/ > /j/ would parallel the established /hw/ > /w/ - although then there might be a question about why mihi isn't *miii or something), and the totally unexplained one also has other irregularities anyway (the last vowel is wrong).

also, it would have been a voiced /h/, which i'd think wouldnt have survived.

It's generally best to go on the basis of the actual evidence, rather than gut feelings about what would or wouldn't have survived. [anyway, who says it was voiced? We've no way to know. Maybe it merged with regular /h/. And in any case, it didn't survive, it was generally lost in vulgar latin]

also: i think italic had its own version of Grassmann's law such that diaspirates always lost one of the two asps.

De Vaan calls such an idea 'unwarranted', and only mentions one word where it might be invoked anyway, and that's a big stretch (trying to link the Latin for 'fat' to the Greek for 'thick', which he says cannot be seriously considered.

i think the second of two aspirates always deaspirated.

Again, mainstream linguistics relies on evidence. And the Oscan evidence of feíhúss shows that this is not true. Similarly in Latin, the word foveo (instead of *fobeo) demonstates this (and we know that /w/ must be from /gwh/, because otherwise the /o/ would have derounded). For obvious reasons there are few words with which your theory could be tested, but they seem to show it's wrong.

Of course, if you really think you've discovered a new fundamental sound law in Latin, do please submit a paper on it to the appropriate channels!

i think the medial consonant would be /g/ regardless of any other factors, even the largyneal.

And again, this isn't how historical linguistics works - you don't get to just decide on gut instinct what "would" happen, you have to look at what actually DID happen!

[Pabappa]

the PIE root in the word for daughter didnt have /gh/ to begin with though, unless by /gh/ you mean /g/ + laryngeal. the medial consonant cannot possibly be /h/ even given all that youve said in that last post.

[sangi39]

What's going on with Latin fruor and fructus?

[Salmoneus]

the PIE root in the word for daughter didnt have /gh/ to begin with though, unless by /gh/ you mean /g/ + laryngeal. the medial consonant cannot possibly be /h/ even given all that youve said in that last post.

No, I meant /gh/, which I thought would be clear from the context of us discussing the reflexes of /gh/ and me listing examples with it in.

Sangi: De Vaan thinks they're both from a fru:g- root, but that the /g/ is eroded by a following /j/ in the first case (/ugj/ > /uj/ > /uw/ > /u/), but strengthened by the following voiceless consonant in the second.

The latter is certainly regular, and the former seems to be mostly so too - cf aio:, I say (alongside axa:re, to name), and ieienta:re, to have breakfast (alongside Sanskrit yájyu-, praiseworthy and Avestan yashtar-, worshipper*), and striatus, grooved, alongside striga:re, to halt, praesti:gia:trix, female juggler, strigor, man of solid strength, and of course Old English stri:can, to wander, and Khotanese pastramj, to oppress.

It gets obscured by the fact that -gi- sequences do still occur from older -gij- and -gej-, and in cases of later affixation (eg with -ia) and analogy.



*the very cunning semantic shift there being an exercise for the interested reader...

[Sequor]

*the very cunning semantic shift there being an exercise for the interested reader...

I imagine it has to do with food offerings and sacrifices. I vaguely remember a Germanic or Greek word that also showed a relationship between eating and the divine (do you happen to have any idea of what word I'm talking about? maybe it was a verb or a god's name, something like "to drink" > "to worship" or "the one people drink to" > "(name of a god)").

Also, man, these last two posts of yours in this thread have been the most word-focused posts I think I've ever seen you write in all these years I've been reading you... Whenever you talk about languages, you usually do it in very broad strokes, only mentioning one to three examples if any, like in this amusing post about the orders of vowel shifts in Romance. By the way, how are you confidently stating those numbers of De Vaan's book (like there being only nine relevant PIE roots with -gh- there)? There's an index of all PIE roots in pages 818-825, do you have a good digitized form of it?

[Salmoneus]

*the very cunning semantic shift there being an exercise for the interested reader...

I imagine it has to do with food offerings and sacrifices. I vaguely remember a Germanic or Greek word that also showed a relationship between eating and the divine (do you happen to have any idea of what word I'm talking about? maybe it was a verb or a god's name, something like "to drink" > "to worship" or "the one people drink to" > "(name of a god)").

Do you mean 'god'? (< 'pour out', via presumably 'one to whom libations are to be poured').

It is indeed to do with sacrifices, but De Vaan suggests it's not the simple connexion you might think (food > sacrifice, or even sacrificed thing > burnt food or the like). Rather, he says that the verb for 'to have breakfast' comes from the active participle of the verb 'to sacrifice' - so it literally means 'to be sacrificing'. Why? Because Roman sacrifices were conducted first thing in the morning, on an empty stomach (there's an adjectival form from 'to sacrifice' that's come to mean "fasting", or just "hungry").
[Whether the specific derivation is via "time of day when you sacrifice" > "meal taken in that time of day", or whether it's plain irony in the form "sacrificing" > "fasting" > "not fasting" (of the "aren't you meant to be fasting?" - "I am [glumph glomph] fasting! This is me fasting!" or "Do you want to join me for some 'fasting'?" kind), I guess we'll never know.]

Also, man, these last two posts of yours in this thread have been the most word-focused posts I think I've ever seen you write in all these years I've been reading you... Whenever you talk about languages, you usually do it in very broad strokes, only mentioning one to three examples if any, like in this amusing post about the orders of vowel shifts in Romance.

This is because I'm pig ignorant, particularly about languages. And what's more, when I do learn something about a language, because I'm bad at languages, and because I'm usually reading from the point of view of shaping my own conlangs, rather than actually learning a language, I don't mentally file it on a word-by-word basis, or often even a language-by-language basis - it's usually just a mental note "this sort of thing happens". So even when I have some sense of what happens, or a vague memory of a specific thing, I can't generally cite anything or go into details (also because most of my information comes from "a random page of an article on a totally different topic that I happened upon while searching Google for something else", which makes it hard to remember my sources).

By the way, how are you confidently stating those numbers of De Vaan's book (like there being only nine relevant PIE roots with -gh- there)? There's an index of all PIE roots in pages 818-825, do you have a good digitized form of it?

In this case, I'm able to cite specific words because I have a PDF copy of De Vaan's book. It's not html-ised, but adobe's search function works relatively well in it (I find it varies a lot between documents!), although it gives a bunch of false positives. But since De Vaan gives that index of roots (annoyingly, not actually all the ones he cites - eg the proposed root of 'traho' is missing, presumably because he didn't know the exact form), I just had to use search on those couple of pages.

I'm unusually confident with De Vaan's book, because a) I can actually remember where on my harddrive it's stored..., and b) I find his layout very intuitive and easy to navigate (now if only he had bothered to include the PIt > Latin changes!). I have a couple of other etymological dictionaries from the same series, which I consult much less often because they're so damn annoying. (although iirc the Celtic one's not bad).

[Pabappa]

Sal, the original poster wanted help deriving a word for daughter from the unattested PIE form, which had a /g/, not a /gʰ/. You made a silly mistake and brushed over it by talking at great length about something else. i wish you didnt need to always turn every discussion into a debate .... in this case about something that isnt what the original poster asked. i dont find the evidence for medial /gʰ/ > /h/ compelling, but i'll leave that debate to other people since it doesnt really interest me and, again , turns out to be not even relevant to the original question that was asked.

[Salmoneus]

Sal, the original poster wanted help deriving a word for daughter from the unattested PIE form, which had a /g/, not a /gʰ/.

Fair enough. I was mislead by De Vaan wondering why there is no 'h' in the Oscan; /gH/ doesn't normally yield /h/, and the relevant sound changes seem to predate the Latin/Sabellic split, so now I'm wondering why De Vaan thinks there should be an /h/ in the first place. In fact, De Vaan specifies in his reconstruction that the Proto-Italic form should have had /x/, not /g/, but that only makes sense if the PIE had /gʰ/. So did he make the same mistake? Or is there some other rule we're both missing?


I'd note, incidentally, that you made the same mistake when you looked up the soundchanges yourself. You were the one who expressed doubt about them, which is why I provided the evidence for them. Had you pointed out at the time that we should have been talking about /g/ instead, we wouldn't have had this discussion...

You made a silly mistake and brushed over it by talking at great length about something else. i wish you didnt need to always turn every discussion into a debate .... in this case about something that isnt what the original poster asked.

Again, I didn't 'turn it into a debate'. I gave what I thought /gʰ/ would yield; you told me I was wrong, that /gʰ/ would yield something else. I defended myself on the basis of the evidence. If you don't want a "debate", on a subject that "doesn't interest" you, don't challenge people's information! I'm not sure what you could have been looking for in disputing the facts I laid out, if not a discussion about those (purported) facts.

Likewise, when you say out of the blue that Latin had a reverse Grassman's Law... what did you expect to happen, if not for someone to say "no it didn't"?



In any case: Kai, since the etymon had /g/, not /gʰ/, and assuming De Vaan's just plain wrong for some reason, rather than knowing something we don't know, the Latin "should" have been fugater, fu:ctre:s. This parallels, for example, tegimen next to te:ctus. The long vowel is from Lachmann's Law, which I forgot about in my earlier answer. The /g/ form would be less likely to be analogised away, as it would be less weird, although it would still be a really weird inflection pattern, so analogy would still be possible.

[KaiTheHomoSapien]

I've found this discussion to be very interesting, just to be clear, even if it's veered away from my original question.

And fugater, that is so close to thygater. But yes, I do think it would likely not have preserved that unusual inflection.