Rare/unusual natlang features

[Nortaneous]

North Mekeo has six consonants: /b m v k ŋ ɫ/.

Source on that? Wikipedia lists additional /t n s ʔ/ for just Mekeo.

Jones 1998. /t n/ are loan phonemes, and I don't know where they're getting /s ʔ/. Mekeo has an epenthetic consonant of some sort in certain vowel sequences, which could be what that /s/ is supposed to be.

[Darren]

North Mekeo has six consonants: /b m v k ŋ ɫ/.

Source on that? Wikipedia lists additional /t n s ʔ/ for just Mekeo.

According to Jones' Towards a Lexicogrammar of Mekeo, it has seven - /b g v z m ŋ l/. There's a lot of difference between the dialects though; the inventories are

North: /b g v z m ŋ l/
Northwest: /p k β g m ŋ w y~e/
West: /p k b d̠ g m ŋ w l/
East: /p k ʔ~Ø f s̠ m ŋ l/

[Nortaneous]

North Mekeo has six consonants: /b m v k ŋ ɫ/.

Source on that? Wikipedia lists additional /t n s ʔ/ for just Mekeo.

According to Jones' Towards a Lexicogrammar of Mekeo, it has seven - /b g v z m ŋ l/. There's a lot of difference between the dialects though; the inventories are

North: /b g v z m ŋ l/
Northwest: /p k β g m ŋ w y~e/
West: /p k b d̠ g m ŋ w l/
East: /p k ʔ~Ø f s̠ m ŋ l/

Those inventories include the epenthetic consonants that developed between certain vowels from fortition of transitional glides. In Appendix 1 (p. 559), Jones gives:

North: /p β g k w m ŋ y/
Northwest: /p β g k w m ŋ y/
West: /p b k g m ŋ l/
East: /p f k ʔ m ŋ l/

The correspondences are:

Code: Select all

N  b v v g g m ŋ l
NW β p w g k m ŋ y
W  b p w g k m ŋ l
E  p f f k ʔ m ŋ l

Jones isn't entirely consistent here, but who cares.

[Pabappa]

The lack of coronals in some dialects is even more amazing to me than the small inventory. Maybe there's allophones, though, like how in Hawaiian /k/ sometimes surfaces as [t].

Reminds me of


Though I suspect the language has few if any closed syllables.

[Nortaneous]

Maybe there's allophones, though, like how in Hawaiian /k/ sometimes surfaces as [t].

ŋ > n adjacent to /i/, k > ts before /i/, /t ts/ have been introduced through loanwords

Though I suspect the language has few if any closed syllables.

(C)V

[bradrn]

Iwaidja apparently uses verbs as kin terms:

The Iwaidja languages are nearly unique among the languages of the world in using verbs for kin terms. Nouns[*] are used for direct address, but transitive verbs in all other cases. … An Iwaidja speaker … says I nephew her to mean "she is my aunt". Because these are verbs, they can be inflected for tense. In the case of in-laws, this is equivalent to my ex-wife or the bride-to-be in English. However, with blood relations, past can only mean that the person has died, and future only that they are yet to be born.

[*] NB: From the example, it appears that these nouns are derived from the verbs using a nominalisation.

[Kuchigakatai]

The Iwaidja languages are nearly unique among the languages of the world in using verbs for kin terms.

It sounds like the authors of that Wikipedia article haven't heard of Iroquoian languages, which are often mentioned to have the same thing. Cedh gives a few examples from Oneida, Seneca and Mohawk in this post. I'll edit the article later today...

[bradrn]

The Iwaidja languages are nearly unique among the languages of the world in using verbs for kin terms.

It sounds like the authors of that Wikipedia article haven't heard of Iroquoian languages, which are often mentioned to have the same thing. Cedh gives a few examples from Oneida, Seneca and Mohawk in this post. I'll edit the article later today...

Well, they did say nearly unique…

[Birdlang]

Oirata has an unique consonant set
/p t d ʔ/
/f v h/
/m n/
/w l j r/
According to Wikipedia. Apparently a lot of Papuan languages have weird phonologies, noticed this with Bunak too.

[bradrn]

Oirata has an unique consonant set
/p t d ʔ/
/f v h/
/m n/
/w l j r/
According to Wikipedia. Apparently a lot of Papuan languages have weird phonologies, noticed this with Bunak too.

Proto-Lakes Plain is even worse:
/p t k b d/

[mae]

-

[bradrn]

I don't think a proto language for a family that has very little work on it should be taken as solid/reliable.

Good point — I probably should have been more sceptical of that phonology when I first saw it.

[Darren]

I don't think a proto language for a family that has very little work on it should be taken as solid/reliable.

Good point — I probably should have been more sceptical of that phonology when I first saw it.

I was also sceptical, but having researched it I'm pretty sure the inventory is mostly accurate. Given that:

• The most common consonant inventory for the modern descendants is something like /b t d k ɸ~h s/ (e.g. Obokuitai, Kirikiri, Iau); those with extra phonemes mostly clearly derive from lenition in that one branch

• In Papuan languages p → ɸ~f~h is common, explaining the lack of /p/ in the inventories; /f/ also has a [p] allophone in most descendants

• In some descendants there is a wierd distribution of /ti/ or /si/ sequences, i.e. all Ci sequences other than si occur, and if more work is done on them it may turn out that [s] isn't necessarily phonemic

• Of the (admittedly limited number of) reconstructed terms which give /s/ in some languages, there is always /t/ in another branch; *t(+i) → s is much more common than *s → t, so it's logical to assume that there was some kind of *t-sound to begin with

I think it's pretty likely that the proto-language had something like */p b t d k/, although the paper I have notes that */ɾ w j/ might also have been phonemes. Duane Close, who I think proposed this inventory to begin with, says

...it is inconclusive at this point, whether a flap *ɾ was really a separate phoneme from *d...

In several branches, there is *-ɾ- → Ø but *-d- → -ɾ-, although there aren't many proto-forms to test this with.

The semivowels *w and *y were included in the PLP consonant inventory above, though parenthetically, to reflect the word list transcriptions in Appendix 3 [...] Several phonologies have been done in Lakes Plain languages and in none of them have *y and *w been posited as phonemes nor have they had a bearing on determining the syllable structure.

It's certainly true that PLP is understudied and the reconstruction definitely isn't 100% reliable, there's a good case for an inventory of */p b t d k/, and most of the regular correspondences can be derived from this. Personally I suspect that there was some kind of liquid~approximant~rhotic phoneme because of the irregular *-ɾ- vs *-d- correspondences; the major argument against this is that I don't think *ɾ occurred word-initially or finally. It's a very interesting family none the less.

[caedes]

Russian too. I think it was pan-Slavic at some point but that many of the other Slavic languages added person markers later on.

No, it's the other way round, Russian and Ukrainian lost the markers. As äreo says, it originally was a periphrastic perfect formed from the perfect active participle and the copula (e.g. (ty) bilъ esi you have beaten). The East Slavic languages stopped using the copula, nit only here, but they generally stopped using the present tense forms, resulting in the current situation where you just have the former participle without personal endings. Some languages (e.g. Polish) fused the forms of the copula with the participle, resulting in a synthetic past ( (ty) biłeś). In other languages, like Serbo-Croation, the tense is still formed periphrastically ( (ti) si bio).

I think even weirder at first sight is that Polish started to use those participles not only in the past tense but also instead of the infinitive in the future tense:

czekamy we wait vs czekaliśmy we waited vs będziemy czekali we will wait

-------------------------

Alemannic dialects of German, at least before consonants, still differentiate between Proto-Germanic *s and /s/ that derives from PG *t via the High German consonant shift. PG *s is realized as something along the lines of [ʃ] ~ [ʂ] ~[ʐ̥], while *t remains [s] ~ [z̥]. E.g. Swabian [iʂ] is vs [iz̥d̥] eats, both [ʔɪst] in NHG.

The fun part is that this shift s -> ʃ is still productive, and [s] in modern-day loanwords is realized as [ʃ] ~ [ʂ] ~[ʐ̥] before consonants, too. E.g. pasta in Swabian is [ˈb̥aʐ̥d̥aː].

[mae]

-

[Darren]

"Mostly accurate" means "partially wrong"--being unable to account for certain regular correspondences means that there are phonemes that existed in the proto-language that have not been accounted for.

It *can* mean that there are extra phonemes, but it's not a certainty as you're implying. If you assigned every correspondence which isn't 1 to 1 in Indo-European languages an ancestor in PIE, there would be another series or two of stops, more vowels, probably another set of labials or velars and other products of minor inconsistencies. After extensive study, it became clear that, for example, the voiceless aspirated stops could be analysed as products of pharyngeals, or that the reason *p in Germanic languages sometimes became -f- and sometimes -v- was stress-related. I think that as *d became /ɾ/ intervocalically in effectively all the LP languages, and that all instances of possible *ɾ in PLP were intervocalic, that it might be some word-structure or prosody-related thing.

And as for potential *r only appearing between vowels as a 'suspicious' fact: this is true of many languages that exist right now, for instance virtually all Japonic languages. PLP apparently only allowed open syllables, and a constraint against initial rhotics is typologically common.

To begin with, I never called it "suspicious". I said that it was the major argument I could see against *ɾ. I'm not sure about the details of /r/ in Japonic languages as I can't find any papers on it, but I'll take your word for it. I think that the limited occurrence is still a good argument against it - as I mentioned earlier,*ɾ shows similarities to *d, which likely also had a *[ɾ] allophone; and I'd be very surprised if a protolanguage had an inventory of six phonemes and one of them completely lost phonemicity in every branch. Looking at the data I can find, my opinion is that /ɾ/ was phonemic at a surface level in PLP, but could be analysed out in some way.

I understand your points about not trusting reconstructions too much, but just because something was reconstructed doesn't mean it's wrong. If there was more research on the LP languages, the phonology would be revised a lot and *ɾ would be cleared up one way or the other; but regrettably there isn't.

[Qwynegold]

I just learned that Yaqui has the same word for "quickly" and "slowly", and that you use hand gestures to differentiate between the two meanings.

[Zju]

I just learned that Yaqui has the same word for "quickly" and "slowly", and that you use hand gestures to differentiate between the two meanings.

I'd like a source on that one.

[Gulliver]

I just learned that Yaqui has the same word for "quickly" and "slowly", and that you use hand gestures to differentiate between the two meanings.

I'd like a source on that one.

It's unsourced on Wikipedia and Encyclopedia.com (and one probably got it from t'other) and nothing's coming up on Google Scholar, so I would suggest it's to be taken with a pinch of salt.

[Qwynegold]

Aha, it was indeed Wikipedia where I read that.