Odd natlang features thread

[Yng]

- Voiceless stops cannot occur word-finally: e.g., arag 'he saw' ~ arkay/arkey 'I saw'.

Not that strange, is it?

- The language does not have independent articles or demonstratives, but attaches these determiners to the end of the noun. Further, two kinds of definite article exist: the first "used to mark new information or something just being introduced in a conversation" (-ka for masculine nouns, -ta for feminine nouns), the second used after "something has been introduced or if it is generally known" (-kii for masculine nouns, -tii for feminine nouns) (p. 31).

Not that weird at all - the only odd thing is that they call the first definite article, which is pretty much the textbook definition of indefinite, definite.

The phonology stuff is cool, though.

[Ser]

The link Trebor gave on Somali vowel harmony doesn't work but this one does, if anybody was interested.
https://ryantxanson.com/media/doc/somal ... armony.pdf

The harmony is described as +ATR /i e { } 3\ / spreading with priority over -ATR /I E a u O/, particularly leftward rather than rightward.

---

Also, I find it interesting that Mandarin doesn't actually have basic nouns meaning 'man' and 'woman'. Rather, it uses the adjectival verbs 'to be male' and 'to be female' along with the noun rén 人 'person', or standing by themselves in headless relative clauses followed by the relativizer ("REL") de 的 (the closest thing the language has to our typical European substantive adjectives).

男的 ~ 男人
nán de ~ nán rén
male REL ~ male person
'(a/the) man/men'

女的 ~ 女人
nǚ de ~ nǚ rén
female REL ~ female person
'(a/the) woman/women'

The headless relative clauses belong to a more familiar register than the other option.

[Nortaneous]

It's not strange for voiced plosives to be restricted from word-final position; it is strange for voiceless plosives to be restricted from word-final position if they can occur word-initially.

[Nortaneous]

In modern NWC, nearly the reverse has occurred: the high central unrounded vowel*ï has become a syllabic liquid [l] (which is retroflex [L] after retroflex initials). See Coblin (1994: 108).

[Trebor]

It's not strange for voiced plosives to be restricted from word-final position; it is strange for voiceless plosives to be restricted from word-final position if they can occur word-initially.

The latter is precisely the case in Somali: tag 'go (imperative)' and kari 'cook/boil (imperative)' are perfectly acceptable example lexemes.

[Trebor]

Not that strange, is it?

Nortaneous has clarified this issue.

Not that weird at all - the only odd thing is that they call the first definite article, which is pretty much the textbook definition of indefinite, definite.

But Zorc and Issa expressly state:

Note that there is a distinction between definite (with an article) and indefinite (without an article) equivalent to the English articles the versus a. Remember that there are two kinds of definite information: new (just being introduced in a conversation) which is marked by -a; and given (known by the speakers or already discussed) which is marked by -ii.

(pp. 39-40)

I'm still puzzled about this matter: your point is important. Maybe a linguist has done some work to figure out exactly what roles unmarked nouns versus -ka/-ta nouns play in Somali discourse.

[clawgrip]

One thing I've always found a bit weird about Old Japanese is that there is a verb that seems to lack any stem...it's just the verb endings alone.

Compare the conjugations of the following verbs, all of the shimo-nidan class:

Code: Select all

受く:     uk-e     uk-e     uk-u     uk-uru     uk-ure     uk-eyo
混ず:    maz-e    maz-e    maz-u    maz-uru    maz-ure    maz-eyo
答ふ:  kotah-e  kotah-e  kotah-u  kotah-uru  kotah-ure  kotah-eyo
得:          e        e        u        uru        ure        eyo

the verb u "get; attain" seems to have no actual verb stem. I can't determine what could have caused this, for example any sound change that could have eliminated the original stem. It's just an anomaly.

[Melteor]

Note that there is a distinction between definite (with an article) and indefinite (without an article) equivalent to the English articles the versus a. Remember that there are two kinds of definite information: new (just being introduced in a conversation) which is marked by -a; and given (known by the speakers or already discussed) which is marked by -ii.

(pp. 39-40)

I'm still puzzled about this matter: your point is important. Maybe a linguist has done some work to figure out exactly what roles unmarked nouns versus -ka/-ta nouns play in Somali discourse.

I am going to go out on a limb and guess that it might be possible to introduce an additional informational structural notion here, of introduced definite nouns in the recent discourse. You might call this the "common ground" or "active topics". The new-definites would need to be distinguished from referents already mentioned, kind of like a switch reference system. Then the unmarked nouns would be nonce one-offs that either don't have specific referents i.e. generics, or aren't deemed particularly relevant. I would guess that there might be an additional morpheme for referring back to something mentioned by another person, i.e. pulling something out of irrelevance back into the conversation. That's my working theory.

[----]

One thing I've always found a bit weird about Old Japanese is that there is a verb that seems to lack any stem...it's just the verb endings alone.

Compare the conjugations of the following verbs, all of the shimo-nidan class:

Code: Select all

受く:     uk-e     uk-e     uk-u     uk-uru     uk-ure     uk-eyo
混ず:    maz-e    maz-e    maz-u    maz-uru    maz-ure    maz-eyo
答ふ:  kotah-e  kotah-e  kotah-u  kotah-uru  kotah-ure  kotah-eyo
得:          e        e        u        uru        ure        eyo

the verb u "get; attain" seems to have no actual verb stem. I can't determine what could have caused this, for example any sound change that could have eliminated the original stem. It's just an anomaly.

Almost the same thing occurs in Ket, with the verb meaning 'to eat'. The typical root is -a, but this is elided for a significant number of speakers:
ku-b-il
2.S-INAN.O-PST
"you ate it"

as well as in Ojibwe, with the verb izhi 'to say'. the verb itself is actually just a hanging 'initial', that would normally modify a root but none is actually present here. A couple different forms of the verb elide even this:
nindig 'he says to me' is analyzable as ni-ig '1-INV' with no phonological material present as the root at all.

[Nortaneous]

^ where's the -nd- come from?

Hongyan Qiang has four vowel POAs, but 24 vowels:

[----]

^ where's the -nd- come from?

Just epenthesis, also depending on the dialect the first person marker inserts a nasal before the nearest following consonant if it's any of /b d dZ g z Z/.
so, the same thing occurs with typical verbs:

ni-bizindaw-ig >> nimbizindaag 's/he listens to me'

ni-agwamo >> nindagwam 'I'm floating on the water'

[Chengjiang]

^ where's the -nd- come from?

Just epenthesis, also depending on the dialect the first person marker inserts a nasal before the nearest following consonant if it's any of /b d dZ g z Z/.

Ah, so it's epenthetic insertion of /d/, followed by insertion of a nasal before that /d/ by the first person marker?

[hwhatting]

the verb u "get; attain" seems to have no actual verb stem. I can't determine what could have caused this, for example any sound change that could have eliminated the original stem. It's just an anomaly.

Almost the same thing occurs in Ket, with the verb meaning 'to eat'. The typical root is -a, but this is elided for a significant number of speakers:
ku-b-il
2.S-INAN.O-PST
"you ate it"
as well as in Ojibwe, with the verb izhi 'to say'. the verb itself is actually just a hanging 'initial', that would normally modify a root but none is actually present here. A couple different forms of the verb elide even this:
nindig 'he says to me' is analyzable as ni-ig '1-INV' with no phonological material present as the root at all.

Something similar holds for the Russian verb vynut' "take out", where the non-present stem forms only consist of affixes around a zero stem:
Present / future stem: vy-jm- (vy- is a prefix meaning "out", -jm- is the verb stem).
Infinitive stem / past stem: vy-nu-, with prefix vy- and semelfactive suffix -nu-
But at least the root shows up in the present stem.

[----]

^ where's the -nd- come from?

Just epenthesis, also depending on the dialect the first person marker inserts a nasal before the nearest following consonant if it's any of /b d dZ g z Z/.

Ah, so it's epenthetic insertion of /d/, followed by insertion of a nasal before that /d/ by the first person marker?

exactly!

Hwhwhatting: is there really a separable semelfactive in Russian? neat

[hwhatting]

Hwhwhatting: is there really a separable semelfactive in Russian? neat

Semelfactive is one of the main uses of the suffix -nu-. It's productive, i.e. it can be used to form new semelfactive verbs, although some of theses coinages have spread beyond semelfactive use (e.g. colloquial faksanut' "to fax" is quite often used generally, not just for single instances of sending a fax.

[Nortaneous]

Here are some papers about aspirated fricatives:
http://www.academia.edu/968778/A_panchr ... _from_Pumi
http://www.academia.edu/6380291/The_ris ... fricatives
http://www.academia.edu/2118759/Proto-Q ... nt_Changes
http://www.academia.edu/5944858/The_pro ... Sgaw_Karen
http://www.academia.edu/3318754/A_contr ... fricatives

[Nortaneous]

Consonant clusters in a rGyalrongic language.

nvse 'morning', nvtsæ 'meal', ʁmno 'awl', χsmɑr 'wheat', ʁvdʑvər 'sprout', çɲɕə 'sweat', χscər 'danger', jmbjəmpɑ 'bird', ʁvrdʑɣə 'geminate', ʁjnzdəjnzdə 'cause to buy each other things (for their own benefit)'

Also: diachronics of sonority hierarchy change.

[KathTheDragon]

Are those all (apart from bird and the last one) monosyllables?

[2+3 clusivity]

Consonant clusters in a rGyalrongic language.

I know the paper is about phonology . . . but . . . that language has crazy causative formation.

[CatDoom]

Are those all (apart from bird and the last one) monosyllables?

Skimming the paper, it looks like they are, but the initial consonants tend to be re-syllabified as coda consonants when they follow a word ending with an open syllable.

[Nortaneous]

They're monosyllabic if you don't consider ʁ a vowel even though it's pronounced about as long as a vowel.

[Nortaneous]

Lower Xumi contrasts all of /ts tʃ tʂ tɕ/ (though retroflex affricates are only from Tibetan loans), and has eight stop/affricate POAs total (adding /p t k q/). Japhug rGyalrong also has eight stop/affricate POAs: /p t ts tʂ tɕ c k q/. (Yanyuwa only has seven.) Also:

Japhug has an intriguing five-way contrast between /velar stop+j/ as in kjo ‘to cause to glide’, palatal stops co ‘valley’, /dental affricates+j/ ɯ-mtsjoʁ ‘beak’, alveolo-palatal affricates as in tɕoʁtsi ‘table’ and /uvular stop+j/ as in qjoʁ ‘to vomit’. Palatal stops can be followed by -ɣ-, -r- or -l- as in ɲcʰɣaʁ ‘birch bark’.

[din]

Consonant clusters in a rGyalrongic language.

nvse 'morning', nvtsæ 'meal', ʁmno 'awl', χsmɑr 'wheat', ʁvdʑvər 'sprout', çɲɕə 'sweat', χscər 'danger', jmbjəmpɑ 'bird', ʁvrdʑɣə 'geminate', ʁjnzdəjnzdə 'cause to buy each other things (for their own benefit)'

Also: diachronics of sonority hierarchy change.

And if you ever wondered what L'internationale sounded like in this language, here you go: https://www.youtube.com/watch?v=gFHGr3eNaL4

(Basically: tsəχtəsəskəftsətətɕə)

[CatDoom]

It sounds like they're dropping a lot of consonants in the song, which wouldn't really be that unusual, I suppose. It would be interesting to compare it to a sample of everyday speech.

[Chagen]

Rabbits in Japanese may take the same numeral classifier as birds. This is because Buddhist monks in Japan used to be restricted to only eating birds, but they loved rabbit meat so much that they bullshitted up a justification that rabbits were actually birds, and that their ears were actually wings. Somehow people believed it.