complementary distribution

[Herr Dunkel]

Yeah, my bad about <au>, kinda slipped in; I was thinking of [œʏ]. As for voiceless segments I think they're also somewhere analysed as /h+C/. What about palatals? AFAIR there's words like <gjóla(n)> <(að) góla> that contrast them but it's only like before back vowels. I also remember the pair <hljóð(ið)> <ljóð(ið)> etc.

[Richard W]

Maybe it's dialectal. I definitely have /n/ in "include" with a strong syllable break between the prefix and the root; but I have occasionally heard /ŋ/.

Idiolectal, in my opinion. Probably subject to tempo and an individual's stylistic variation as well.

[vokzhen]

Maybe it's dialectal. I definitely have /n/ in "include" with a strong syllable break between the prefix and the root; but I have occasionally heard /ŋ/.

Idiolectal, in my opinion. Probably subject to tempo and an individual's stylistic variation as well.

Context as well. Afaict "it includes" varies between [ɪʔtnkludz] [ɪʔŋkludz] and [ɪɾɨŋkludz] for me, for example, depending on whether the /t/ is unreleased, replaced, or flapped, which itself seems dependent on prosody.

[Qwynegold]

I'm not sure I've read this correctly (plus it's late) but if I have... are you people seriously proposing that /h ŋ/ are actually one phoneme?

/ꜧ/

[hwhatting]

I'm not sure I've read this correctly (plus it's late) but if I have... are you people seriously proposing that /h ŋ/ are actually one phoneme?

This has been seriously considered for German as well; there, /h/ only occurs syllable-initially in stressed or pre-stress syllables, while /ŋ/ only occurs syllable-final and syllable-initially in post-stress syllables. In the end, it depends of your definition of phoneme - if its only a question of contrasting or if you also expect the allophones of a phoneme to share some features.

[Seirios]

Well in the case of Standardized Mandarinandarin jqx I personally tend to think them as neutralized form before high-front approximants (I analyze i u y as syllabic approximants and actually I find forgetting the idea that a syllable must have main vowel really explains more things in Mandarin in a more concise, systematic way).

There is actually another example in Standardized Mandarin: [e] and [o] actually do not contrast phonologically and can be analyzed as a schwa-ish vowel surfaced beside an approximant j/w ([ie] [ye] [ei] [uo] [ou]), but nobody (i.e. normal non-phonologically-sophisticated people/crowd) actually consider [e] and [o] to be interchangeable in any way -- change [ie] to [io] people would think you're speaking [iau] or [ye] (or [iou]) or with accent, [ye] to [yo] features unambiguously a northern-to-middle accent, [ei] to [oi] is strange, [uo] to [ue] features mainly a middle-to-southern accent (if outside certain sets of characters, strange accent) -- unless when speaking of [uei] and [iou], when the difference of e and o is less pronounced due to the influence of adjacent approximants (it does not interfere communication and many people probably don't notice that in some cases, though if noticed people would still think that's an accent), but I tend to think them as with schwa.

Btw [ien] and [yen] are analyzed as with /a/.

[vec]

Yeah, my bad about <au>, kinda slipped in; I was thinking of [œʏ]. As for voiceless segments I think they're also somewhere analysed as /h+C/.

The stops are a bit of a mess. Word initially, aspirates contrast with tenuis: tá /tʰau:/ 'toe', dá /tau:/ 'coma'. Intervocalically and in codas, they do not contrast. In Southern Icelandic, they are tenuis intervocalically, in Northern Icelandic they are aspirated. However, whenever preceded by a short vowel (i.e. "geminated" in writing), there is a kind of contrast, but the fortis stops are realized as preaspirated rather than postaspirated: hitta /hɪhta/ 'hit', Hidda /hɪta/ (short for the female name Hildur). It is uncertain whether the latter has a delay /hɪt:a/ or /hɪtʔa/ or something similar – it depends on context, speed, mood etc.

To complicate matters, in Southern Icelandic, when stops are syllable-initial but word-medial, followed by a nasal or lateral, the fortis ones devoice the continuant while the lenis ones don't: geltu [cɛɬtʏ] 'bark (imperative or past 3rd person pl.)', geldu [cɛltʏ] 'geld, castrate (imperative or past 3rd person pl.)'. In Northern Icelandic, this doesn't happen. Instead the fortis-ness and lenis-ness appears on the stop and the continuant is always voiced: /cɛltʰʏ/ vs. /cɛltʏ/.

The Northern pronunciation is certainly older, and moving the /h/ segment into the preceding phone is an innovation, but it's an innovation that has become the majority way of speaking. The Southern system seems to lend itself to the analysis that aspiration is only phonemic word initially, and that intervocalically, we are dealing with clusters of h + stop, and voiced or unvoiced continuants + stop. This would imply all the unvoiced continuants have phonemic status. But some, at least /m̥/ only contrast intervocalically preceding /p/, and in the interjection hm /m̥/, vs. the interjection mm /m/, which mean the same as in English. So it's a bit of a stretch.

And the final complication is the possibly hypercorrective, possibly incredibly conservative, Southern pronunciation of ng clusters pefore t and d that I have, pronouncing langt /lauŋ̥t/ 'long' and lengd /leiŋt/ 'length', which further supports possible phoneme-status of at least /ŋ̥/. Northerners, at least older ones, will say /launt/ and /leiŋt/ – the dorsal nasal preceding the fortis stop will assimilate, but the other one will not.

So it's a messy mess.

What about palatals? AFAIR there's words like <gjóla(n)> <(að) góla> that contrast them but it's only like before back vowels. I also remember the pair <hljóð(ið)> <ljóð(ið)> etc.

First of all, there is no palatal lateral. Hljóðið is /ɬjouðɪð/, ljóðið is /ljouðɪð/.

Whether the palatal stops are phonemes in their own right or whether they are underlyingly sequences of velar stop + /j/ is up for debate. But I think near minimal-pairs like gæi /kai:jɪ/ 'guy', and gægist /caijɪst/ 'peek' would suggest they are phonemic. Typically, the velars palatalize before í, i, ý, y, æ, e, é. And before all of them they can never be velar, except for æ. But I don't know. It's contentious.

[Pabappa]

I'm not sure I've read this correctly (plus it's late) but if I have... are you people seriously proposing that /h ŋ/ are actually one phoneme?

This has been seriously considered for German as well; there, /h/ only occurs syllable-initially in stressed or pre-stress syllables, while /ŋ/ only occurs syllable-final and syllable-initially in post-stress syllables. In the end, it depends of your definition of phoneme - if its only a question of contrasting or if you also expect the allophones of a phoneme to share some features.

What about Ahorn ("maple") ?

[TaylorS]

Well in the case of Standardized Mandarinandarin jqx I personally tend to think them as neutralized form before high-front approximants (I analyze i u y as syllabic approximants and actually I find forgetting the idea that a syllable must have main vowel really explains more things in Mandarin in a more concise, systematic way).

There is actually another example in Standardized Mandarin: [e] and [o] actually do not contrast phonologically and can be analyzed as a schwa-ish vowel surfaced beside an approximant j/w ([ie] [ye] [ei] [uo] [ou]), but nobody (i.e. normal non-phonologically-sophisticated people/crowd) actually consider [e] and [o] to be interchangeable in any way -- change [ie] to [io] people would think you're speaking [iau] or [ye] (or [iou]) or with accent, [ye] to [yo] features unambiguously a northern-to-middle accent, [ei] to [oi] is strange, [uo] to [ue] features mainly a middle-to-southern accent (if outside certain sets of characters, strange accent) -- unless when speaking of [uei] and [iou], when the difference of e and o is less pronounced due to the influence of adjacent approximants (it does not interfere communication and many people probably don't notice that in some cases, though if noticed people would still think that's an accent), but I tend to think them as with schwa.

Btw [ien] and [yen] are analyzed as with /a/.

Wait, so you are proposing that Mandarin has a vertical vowel system? That... actually makes a helluva lot of sense!!!

[Herr Dunkel]

@Vecfaranti, I think there's pretty much no distinction medially anyway because <b d g> /p t k/ medially become <v ð g> /v ð γ/ except after /l n (m?)/ as far as I've noticed where they're /p t/ and <ng> is /ŋ/. The guy I've been skyping with pronounces <hlj lj> as /ʎ̝̊ ʎ/ so I kinda thought the rest of Iceland does it as well.

@Publipis, there's also "anhalten", "anhören" and lots of other verbs with a stressed prefix, then things like "Urheimat" and other nouns with stressed prefixes, and derivations with "-heit", "-haft", "-haltig" (Dunkelheit, dauerhaft, alkoholhaltig), probably some others as well.

[vec]

@Vecfaranti, I think there's pretty much no distinction medially anyway because <b d g> /p t k/ medially become <v ð g> /v ð γ/ except after /l n (m?)/ as far as I've noticed where they're /p t/ and <ng> is /ŋ/. The guy I've been skyping with pronounces <hlj lj> as /ʎ̝̊ ʎ/ so I kinda thought the rest of Iceland does it as well.

Are you sure you're hearing him right? I've never read that this exists and I don't hear it myself but it's kind of hard to judge for myself, I guess. I start overthinking it.

[Herr Dunkel]

I solicited these from him:

ljóð
hljóð(ið)

Kinda sounds palatal to me but I'm not really sure; what would you say?

[hwhatting]

@Publipis, there's also "anhalten", "anhören" and lots of other verbs with a stressed prefix, then things like "Urheimat" and other nouns with stressed prefixes, and derivations with "-heit", "-haft", "-haltig" (Dunkelheit, dauerhaft, alkoholhaltig), probably some others as well.

Yes, I forgot about these things - basically, what I said is only true for non-compunds and non-suffixed words, in compounds and suffixed words, the second part / the suffix behave like a simple word phonetically.

[Pabappa]

That's good, but anyone arguing for true equivalence would have to show that there are no counterexamples at all. Unless Ahorn is somehow a compound of A and Horn, it kills the argument by itself.

[hwhatting]

That's good, but anyone arguing for true equivalence would have to show that there are no counterexamples at all. Unless Ahorn is somehow a compound of A and Horn, it kills the argument by itself.

Etymologically it's not, but people treat it that way, creating joke formations like "Behorn" etc. So it's a borderline case. A good counterexample would be a word with a post-accent syllable beginning in /h/ that is not at the same time a valid noun (like Horn) or suffix.

[Herr Dunkel]

"Ahorn" looks terribly odd. The etymology I've seen derive it from "*axurnaz" which gave OHG "āhorn". PGmc. <*h> usually gave German <ch> /x/, mostly before consonants ("Tochter", "Tucht") or <h> in a stressed syllable (like "Behuf") or generally dropped ("Schuh", "ziehen")

[hwhatting]

"Ahorn" looks terribly odd. The etymology I've seen derive it from "*axurnaz" which gave OHG "āhorn". PGmc. <*h> usually gave German <ch> /x/, mostly before consonants ("Tochter", "Tucht") or <h> in a stressed syllable (like "Behuf") or generally dropped ("Schuh", "ziehen")

Exactly, Ahorn is an oddity. I assume the only reason why it didn't end up as *Achern, *Ahern ['a:?ern] or so is exactly that people re-analysed it as a (however nonsensical) compund with Horn as the second element.