Linguistic Miscellany Thread

[Kuchigakatai]

Re: irregular oblique stems in Latin. The old IE pattern of nominative -r and oblique -n survives, kind of, in three Latin nouns. They're... a bit weird.

One of them is iecur iecin-oris 'liver'. You'd expect the oblique to be *iecin-is from comparative IE, instead of having that -or- attached that looks like it's an analogical from the -ur of the nominative/accusative, supported by the unstressed -or- of a number of other neuter nouns, like pectus pectoris, frīgus frīgoris, aequor aequoris, rōbur rōboris, fulgur fulguris. This extended oblique -oris ended up affecting a lot of neuter nouns and some feminine nouns in Romanian, so that Latin (accusative singular + plural) vīnum vīna > vin vinuri, locum locōs/loca > loc locuri, rīvum rīvōs > (gender changed to neuter) râu râuri. Also borrowings: radio radiouri, taxi taxiuri.

Another is iter itin-eris 'journey, trip; road, path'. You'd think it's the same story as the previous one and it largely is: it was supposed to have oblique *itin-is judging by its IE cognates, but has the -er of nom./acc. iter, supported by the likes of cicer ciceris, genus generis, opus operis. But funnily in Old Latin the nominative/accusative itiner is well attested, alongside iter. Lewis & Short give citations from Naevius, Ennius and Plautus, and then in the Classical period further archaizing ones from Lucretius' and Martial's poetry.

Finally, the third noun does have the old IE pattern: fēmur fēmin-is 'thigh'. However, unlike the previous two, it has no certain IE cognates. De Vaan thinks it might be a borrowing. But if so, how the heck did the one word that was allegedly borrowed retain the old pattern, when the ones that have cognates didn't? Only the Heavenly Gods (dī caelestēs) know, I suppose. It would be hilarious if true.

[Richard W]

Re: irregular oblique stems in Latin. The old IE pattern of nominative -r and oblique -n survives, kind of, in three Latin nouns. They're... a bit weird.

This "nominative v. oblique" categorisation grates. It's nominative singular v. the rest, and 'nominative singular' embraces the vocative and accusative singular as appropriate.

Now, there is an IE nominative v. the rest pattern, which is strong stems versus weak stems in non-neuter nouns. This is still not entirely case-based, as the accusative singular (and originally the locative singular) patterned with the nominatives.

Finally, the third noun does have the old IE pattern: fēmur fēmin-is 'thigh'. However, unlike the previous two, it has no certain IE cognates. De Vaan thinks it might be a borrowing. But if so, how the heck did the one word that was allegedly borrowed retain the old pattern, when the ones that have cognates didn't? Only the Heavenly Gods (dī caelestēs) know, I suppose. It would be hilarious if true.

In Middle English, the verb that would have become 'to latch' was replaced by the French borrowing 'to catch', which then adopted the irregular weak conjugation of the potential *latch. (The modern weak verb 'to latch' is a more recent derivative from the noun.) Such a pattern might apply to the development of Latin _fēmur_.

[Kuchigakatai]

This "nominative v. oblique" categorisation grates. It's nominative singular v. the rest, and 'nominative singular' embraces the vocative and accusative singular as appropriate.

Maybe I should do like the ancients and talk about the direct vs. the obliques.

(Well, (cāsus) rēctus doesn't quite mean 'direct', does it? Maybe the straight case... Or right-angled case... Or just right case... Or rect case. Wrecked case.)

In Middle English, the verb that would have become 'to latch' was replaced by the French borrowing 'to catch', which then adopted the irregular weak conjugation of the potential *latch. (The modern weak verb 'to latch' is a more recent derivative from the noun.) Such a pattern might apply to the development of Latin _fēmur_.

Yeah. It may have helped that iecur already referred to a large-ish body part, and maybe ūber ūberis 'udder' was still something like *oudər *oudənes at some point too.

[Moose-tache]

It's a result complement, which for some verbs can be truncated.

"We broke it [so that it became] open."

Much like complements of verbs of thought or speech, there are special rules for this sort of thing. The reduction can only happen to adjectival clauses modifying the thing acted upon (*"We broke it smithereens."), and it only works for verbs that describe a telic physical action (*"That news depressed me sad.").

[Pabappa]

as soon as i saw this i thought immediately of PIE's laryngeals, particularly the theory stating that they alternated between vocalic and consonantal pronunciations based on the surrounding environment. If they can do it, so can we.

the main obstacle i can see is that the PIE layngeals are usually reconstructed as mostly or entirely voiceless, whereas these are all voiced. and of course, possibility is not proof ... we have some evidence that shows the allophone theory COULD be true but none that it is.

[WeepingElf]

Well, nobody knows whether the PIE laryngeals were voiceless or voiced. The opinions of the scholars are divided on this matter. The Hittite reflexes seem to indicate voiceless fricatives, but that may have been an Anatolian innovation. Or Late (Post-Anatolian) PIE changed them.

[KathTheDragon]

This looks more like epenthesis of schwa followed by loss of the dorsal sonorants with vowel colouring - remarkably parallel to the PIE laryngeals' development.

[Kuchigakatai]

I remember years ago I was once telling a Latin learner I had found it very amusing to have recently read that Romanian may have dropped currō 'to run' because the 1SG form would've become a homophone of cūlus 'ass', so only the latter was preserved, as cur. The guy found it too facile an explanation, and said the homophony was more likely irrelevant and therefore the explanation was false.

But I can't help but wonder. I'm reminded of the weird outcome in Old Spanish of cum nōscum > connusco 'with us', cum vōscum > convusco 'with you guys' (cf. cum mēcum > conmigo 'with me'). By regular sound hanges these should've ended in -osco. Were the pronoun morphemes replaced by some dative like vōbīs > *vois > *vuis > -vus-? Or, assuming the -nn- of connusco represented /ɲ/ (the orthography is ambiguous in this case), was the vowel modified to prevent such a basic grammatical word as *connosco 'with us' from containing coño, a vulgar and offensive word for the female genitals, afterwards affecting convusco by analogy?

Relatedly, I've noticed that a lot of people who learn Spanish whose L1 is (Canadian) English or Mandarin have problems distinguishing /mp nt nk/ and /mb nd ng/ before an unstressed vowel, because phonetically they sound similar (although never the same), and naturally their L1 doesn't help. I've been amused by how hard it is to think of any minimal pairs, as very often such words are distinguished by not one but two segments if I limit myself to commonly-known vocabulary:

anca *(h)anga *(h)onco hongo
limpio *limbio *limpo limbo
lenta *lenda *linta linda
renta *renda *rienta rienda
pinto *pindo *panto pando
tonta *tonda *tunta tunda
campo *cambo *compo combo, compa *comba
acampar *acambar *acapar acabar
ancla *angla *ánculo ángulo

canto *cando quinto *quindo
yunta *yunda llanta *llanda
tinto *tindo tanto *tando, tiento *tiendo, tonto *tondo

It feels as if the language has been avoiding such minimal pairs to some extent... But of course, a few minimal pairs involving common words exist:

cuánto 'how much?', cuándo 'when?'
siento 'I feel', siendo '(while) being'
venta 'event of selling, sale', venda 'a blindfold; may s/he sell'
rompo 'I break, tear', rombo 'rhombus'
tanta 'such many Xs, so much [mass noun]', tanda 'turn of batch of activities in a sequence, row in a series, group'

Besides some involving one or two uncommon words, like cantor 'cantor' vs. candor 'candour' (neither of which is too uncommon really, but anyway...).

[Kuchigakatai]

https://youtu.be/FRe5K1Am4cI

Ulster English, man.

(5:19) anniversary period [ˈpɹ̥i.əd]

(6:06) UK [ˈjʉː ˈkʰeː]

(6:25) Today, [tʰʉˈdeː] the British presence remains with us, alive and kicking and going to work. [əˈlɛːɪv ən ˈkʰɪgɪn ən ˈgoːɪn du ˈwɵɹ]

(7:48) by the events [ɪˈvæ̃nts] of today

(8:03) I look at the faces [ˈfeə̯sɪz]

(8:13) So as we stand here, on the ground [gɹɐːʏ̯nd] trod by Ulster's first volunteers, ...

The last one sounds a lot like North American English "grind" [gɹaɪnd]...

[Pabappa]

Spanish has two words for fruit: fruto and fruta. The first is scientific, the second is culinary ... so a tomato is a fruto but not a fruta. This might be happening in other Romance languages too, since Italian has both words also, though i note that the icecream flavor is the masculine tutti frutti.

[Raphael]

German has Frucht and Obst, with the former referring to individual fruits and the latter to the concept of fruit, though you can use the plural of Frucht - Früchte - for the latter, too.

[Travis B.]

tutti frutti.

In this case, could it be though that it is rhyming which is responsible for this?

[Pabappa]

As far as I know, it'd still rhyme in the feminine ... tutte frutte. Italian has some gender-variant plurals but I dont think this is one of them.

[Kuchigakatai]

Like its English "cognate", Italian frutta is mostly a mass noun. So the normal expression would be "tutta la frutta".

Maybe tutti (i) frutti originates in a literal translation from an Italo-Romance variety where the countable frutto and its plural frutti are normal for fruit as food. Just a guess, I don't really know. Neapolitan would be my first object of suspicion, since it doesn't clearly distinguish gender a lot of the time. The cognate of il frutto would be /o ˈfɾuttə/ (plural /e ˈfɾuttə/), and that of la frutta /a ˈfɾuttə/ ((mass noun?) plural /ef ˈfɾuttə/).

However, I notice that even standard Italian has "frutti di mare" for 'shellfish' (that is, seafood other than fish).

[bradrn]

A (hopefully) quick question: Some languages (e.g. English) use a verbal copula. Others (e.g. Hausa) use a non-verbal copula. But what exactly is the difference between the two? To me there doesn’t seem like much of a difference between an English sentence He is a boy and the Hausa sentence shi yaro ne (lit. ‘he boy COP’).

Also on the subject of copulae, it seems like there are only a few strategies which are used to give a copulative meaning: verbal copula (e.g. English), non-verbal copula (e.g. Hausa), zero copula (e.g. Tiwi), nonverbal person agreement (e.g. Turkish), personal pronominal copula (e.g. Motu), demonstrative pronominal copula (e.g. Kâte). Is that an exhastive list, or did I miss a strategy?

[akam chinjir]

Turkish is often described as having a suffixed copula, for what it's worth (which is at least an option you don't mention).

Sometimes a copula might be called nonverbal if it doesn't inflect the way verbs do (it might instead just be called uninflecting). There might also be obvious word order differences, I suppose, like maybe there are some adverbs that would follow a verb but precede the copula.

In fact another option is allowing zero copula unless you need tense inflection, in which case you use a verbal copula. (I think this is how it works in Russian, for example.)

I'm pretty sure I've seen English "as" compared to nonverbal/uninflecting copulas, the idea being that it can take something nonverbal and turn it into a predicate. With "as" this works only in very particular circumstances, though, like in certain so-called small clauses ("I see her as a hero").

In classical Chinese wei 為 gets used as a copula in certain deranked subclauses, where you sort of need something to fill a verbal slot, but it's rarely if ever used as a plain copula in main clauses, I suppose that's another pattern. (A famous example: 昔者莊周夢為胡蝶 "Once Zhuang Zhou₁ dreamed ∅₁ wei butterfly = dreamed he was a butterfly.")

[bradrn]

Turkish is often described as having a suffixed copula, for what it's worth (which is at least an option you don't mention).

I do mention that option, as ‘nonverbal person agreement’. (And I even give Turkish as an example of that!) Beja is another one with this pattern.

Sometimes a copula might be called nonverbal if it doesn't inflect the way verbs do (it might instead just be called uninflecting). There might also be obvious word order differences, I suppose, like maybe there are some adverbs that would follow a verb but precede the copula.

Thanks!

In fact another option is allowing zero copula unless you need tense inflection, in which case you use a verbal copula. (I think this is how it works in Russian, for example.)

I am aware of this option already — Modern Hebrew is another one which does this. (Although the situation in Hebrew is a bit tricky in that the copula is derived from a third person pronoun.)

I'm pretty sure I've seen English "as" compared to nonverbal/uninflecting copulas, the idea being that it can take something nonverbal and turn it into a predicate. With "as" this works only in very particular circumstances, though, like in certain so-called small clauses ("I see her as a hero").

In classical Chinese wei 為 gets used as a copula in certain deranked subclauses, where you sort of need something to fill a verbal slot, but it's rarely if ever used as a plain copula in main clauses, I suppose that's another pattern. (A famous example: 昔者莊周夢為胡蝶 "Once Zhuang Zhou₁ dreamed ∅₁ wei butterfly = dreamed he was a butterfly.")

Those are both pretty interesting patterns, although they weren’t quite what I was interested in. Another pattern along similar lines is to have one copula for nouns and another for adjectives, as seen in e.g. Eʋe and Vai.

[akam chinjir]

Turkish is often described as having a suffixed copula, for what it's worth (which is at least an option you don't mention).

I do mention that option, as ‘nonverbal person agreement’. (And I even give Turkish as an example of that!) Beja is another one with this pattern.

Right, I meant that as a possible correction.

Another pattern along similar lines is to have one copula for nouns and another for adjectives, as seen in e.g. Eʋe and Vai.

There are also copulas for locative predicates.

Among languages that have two copulas that are used with noun, adjective, and location predicates, I think there's supposed to be a pretty good generalisation that one of the copulas is used for both nouns and adjectives. (I guess if you're using a copula with your adjectives, it's a good bet your adjectives are kind of noun-y.)

[bradrn]

Another pattern along similar lines is to have one copula for nouns and another for adjectives, as seen in e.g. Eʋe and Vai.

There are also copulas for locative predicates.

Among languages that have two copulas that are used with noun, adjective, and location predicates, I think there's supposed to be a pretty good generalisation that one of the copulas is used for both nouns and adjectives. (I guess if you're using a copula with your adjectives, it's a good bet your adjectives are kind of noun-y.)

It’s interesting that you mention that, as both Eʋe and Vai use the same copula for adjectives and locations (and, as I said, a different copula for nouns).

[akam chinjir]

It’s interesting that you mention that, as both Eʋe and Vai use the same copula for adjectives and locations (and, as I said, a different copula for nouns).

Hm, I hope I haven't misremembered the generalisation!