How to Design a European Phonology (in Interesting Ways!)

[Skomakar'n]

http://en.wikipedia.org/wiki/Pirah%C3%A3_language
The Piraha phonological inventory on wikipedia highlights something I think is pertinent here: Piraha is often claimed to have "no nasals", but as Wikipedia points out, you could reanalyse it as having "no velars" instead, because it has [k g] and [m n] all as allophones of something or other. I think both analyses are very misleading because the language phonetically has both sounds, and it's just in phonological/theoretical terms that it doesn't have one or the other.

I thought the tilde marked a nasal vowel... Especially since the Wikipedia page says an alternative spelling is "Pirahán". S:

[Astraios]

Nasal consonants.

[Aurora Rossa]

http://en.wikipedia.org/wiki/Pirah%C3%A3_language
The Piraha phonological inventory on wikipedia highlights something I think is pertinent here: Piraha is often claimed to have "no nasals", but as Wikipedia points out, you could reanalyse it as having "no velars" instead, because it has [k g] and [m n] all as allophones of something or other. I think both analyses are very misleading because the language phonetically has both sounds, and it's just in phonological/theoretical terms that it doesn't have one or the other.

Yeah, and that reminds me of the debate over how many vowels the Northwest Caucasian languages have. From what I have read, they actually have a full set of vowels phonetically speaking but linguists analyze them as only two vowels.

[Skomakar'n]

Nasal consonants.

Ah. Doesn't sound too implausible to me, if a language never had/lost its initial nasals, and nasals elsewhere yielded nasal vowels and then disappeared..?

[ná'oolkiłí]

Nasal consonants.

The point's moot— "Pirahã" isn't their endonym.

[finlay]

Nasal consonants.

Ah. Doesn't sound too implausible to me, if a language never had/lost its initial nasals, and nasals elsewhere yielded nasal vowels and then disappeared..?

Piraha has one phoneme that is [n] initially and [g] (or something phonetically weird) medially. To say it's one or the other, and that therefore it carries all the characteristics and features of one but not the other, is unhelpful. And we should remember that it is this exact fact that keeps us from analysing [h] and [ŋ] as one phoneme in English, although I suspect that in Piraha there is evidence because the two alternate or something.

And as far as I know, it doesn't have nasal vowels either, because as someone else says, it's an exonym. They're known for having only 10 or 11 phonemes.

[MisterBernie]

Piraha has one phoneme that is [n] initially and [g] (or something phonetically weird) medially. To say it's one or the other, and that therefore it carries all the characteristics and features of one but not the other, is unhelpful. And we should remember that it is this exact fact that keeps us from analysing [h] and [ŋ] as one phoneme in English, although I suspect that in Piraha there is evidence because the two alternate or something.

Now, if I had a time machine I would go back to a few hundred years earlier and see if the distribution was something like
/b/ = [m] initially, elsewhere
/g/ = [ŋ] initially, [g] elsewhere
Because symmetry is awesome.

[Ollock]

Piraha has one phoneme that is [n] initially and [g] (or something phonetically weird) medially. To say it's one or the other, and that therefore it carries all the characteristics and features of one but not the other, is unhelpful. And we should remember that it is this exact fact that keeps us from analysing [h] and [ŋ] as one phoneme in English, although I suspect that in Piraha there is evidence because the two alternate or something.

Now, if I had a time machine I would go back to a few hundred years earlier and see if the distribution was something like
/b/ = [m] initially, elsewhere
/g/ = [ŋ] initially, [g] elsewhere
Because symmetry is awesome.

Buuut ... allophony is synchronic. Native speakers usually don't know the details of their language's historical development.

[MisterBernie]

Buuut ... allophony is synchronic. Native speakers usually don't know the details of their language's historical development.

What I meant with that was that if we could do a diachronic analysis of Piraha, the synchronic situation might make more sense. Personally, I don't really care whether [g]-[n] is /n/ or /g/ or /ŋ/, but how that allophonic contrast came to be would be interesting for me.

[Ollock]

Buuut ... allophony is synchronic. Native speakers usually don't know the details of their language's historical development.

What I meant with that was that if we could do a diachronic analysis of Piraha, the synchronic situation might make more sense. Personally, I don't really care whether [g]-[n] is /n/ or /g/ or /ŋ/, but how that allophonic contrast came to be would be interesting for me.

Hmm, yes, yes, that makes sense.

[brandrinn]

Guys about the nasal vowels: vowels are phonetically nasalized after glottal consonants. So whether it's an exonym or not, it still fits Piraha phonotactics, and yes Piraha words do have nasal vowels phonetically.

[Herr Dunkel]

Buuut ... allophony is synchronic. Native speakers usually don't know the details of their language's historical development.

For some wierd reason, Serbian kids can make pre-Slavic-Palatalisation sounds... And retrace [ɔØ] to [ lØ], when [ɔØ] < [ lØ] happened in the fifteenth (?) century.
And they sometimes can replicate nasal vowels What the Angstscheiße?

(It's not true allophony but it's a historical development. Srsly how can they retrace their linguistical roots 1k+ years back)

EDIT:

Guys about the nasal vowels: vowels are phonetically nasalized after glottal consonants. So whether it's an exonym or not, it still fits Piraha phonotactics, and yes Piraha words do have nasal vowels phonetically.

Isn't that rhinoglottophilia?

[Rodlox]

Piraha has one phoneme that is

If Piraha turns out to be a European (or IE) language or creole, I will be surprised. Pleasantly surprised, but still surprised.



(rampant speculation) On the other hand, maybe its just older than all the other IE creoles in its neck of the Americas. (/rampant speculation)

[finlay]

Piraha has one phoneme that is

If Piraha turns out to be a European (or IE) language or creole, I will be surprised. Pleasantly surprised, but still surprised.



(rampant speculation) On the other hand, maybe its just older than all the other IE creoles in its neck of the Americas. (/rampant speculation)

It's called a tangent. They happen a lot here. In this case I brought up Piraha because it illustrates the folly of calling a phoneme by one label or another and then proclaiming that a language "doesn't have" any phoneme that fits the description of the other allophone. The basic sense was that normally we wouldn't say that Icelandic has /g/, but the contrast between the two velar plosive phonemes, usually marked /k kʰ/ for Icelandic, bears a lot of similarity to that between /g k/, and in many ways, the aspiration contrast corresponds to the voicing contrast in other languages like French, and we would call them lenis and fortis (weak like g and strong like k) to emphasise this point. ie Icelandic's /k/ corresponds to /g/ in other languages (and this is upheld by the orthography – same for Chinese Pinyin although Icelandic's orthography is historical and Pinyin is invented)

[Rodlox]

Piraha has one phoneme that is

If Piraha turns out to be a European (or IE) language or creole, I will be surprised. Pleasantly surprised, but still surprised.



(rampant speculation) On the other hand, maybe its just older than all the other IE creoles in its neck of the Americas. (/rampant speculation)

It's called a tangent.

I know. I just typed down what I thought of when I saw the juxtaposition...that, combined with either in this or the prior European Phonology thread, about how not all European languages/conlangs seem particularly European. that gave me the idea that maybe Piraha was an extreme form of that.

[finlay]

if only

[Plusquamperfekt]

Once I wanted to create a Germlang based on Modern High German, but with a third consonant sound shift, ejectives and tones...

[WeepingElf]

Once I wanted to create a Germlang based on Modern High German, but with a third consonant sound shift, ejectives and tones...

I once attempted a Germanic conlang based on Modern High German with a third consonant shift (though not with ejectives and tones), but it never went far.

[Skomakar'n]

Once I wanted to create a Germlang based on Modern High German, but with a third consonant sound shift, ejectives and tones...

Scandinavian languages have tones (pitch accent and stød are tones, and I don't care if anyone disagrees!)!

[WeepingElf]

Once I wanted to create a Germlang based on Modern High German, but with a third consonant sound shift, ejectives and tones...

Scandinavian languages have tones (pitch accent and stød are tones, and I don't care if anyone disagrees!)!

Some German dialects are also pitch-accent languages, with two contrasting pitch accents called Stoßton and Schleifton. A nice minimal pair is found in the sentence

Sie liebte den Seher sehr. 'She loved the seer much.'

In this sentence, Seher has Schleifton and sehr has Stoßton, otherwise the two words are pronounced the same.

I have been inspired by this in Old Albic, which also has such a system.

[Ser]

But brandrinn claims there's "many", what are other examples?

Hungarian once had this distinction as well. That's why <s> and <sz> are the opposite of what you'd expect in modern Hungarian. I think some other Romance languages also had this feature, but I'm not sure how to find out.

In the case of the Spanish in central and northern Spain, Old Spanish /dz̻/ merged with /ts̻/, and /z̺/ with /s̺/. Then this new /ts̻/ became a fricative /s̻/ opposing the inherited /s̺/ only in place of articulation: /s̻/ vs. /s̺/. Then they underwent dissimilation and became modern /θ/ and /s/.

I know something similar happened in French, where the Old French of Paris had its /ts̻/ and /dz̻/ deaffricated to /s̻/ and /z̻/, merging with the preexisting /s/ and /z/. The old /s/ and /z/ could have been laminal as well like the fricative component in the old affricates, or not, having some apical-laminal distinction for some time before merging. No idea what it's usually thought to have happened.

In Spanish at least, it's well known because of how names of places in Nahuatl were transcribed: Nahuatl has [s̻], so at first, they started to be transcribed with ‹c/ç, z› /s̻/, but once the dissimilation happened and this [s̻] became [θ], places started to be transcribed with ‹s› /s̺/.

[johanpeturdam]

Faroese has been a little nicer to it, keeping it initially before back vowels and /ø/, but getting rid of it in most intervocalic and final positions, usually by turning it into some other sound such as /v/ or /j/. It has usually become /dʒ/ before front vowels (except for /ø/), but also /j/ in some positions (generally in the same place as where Icelandic does it). It is generally kept before other consonants, though, such as the clusters /gv/ and /gd/ (but */gj/ is /dʒ/).

<gøta> /ˈgøːta/
<skuggi> /ˈskʊdʒɪ/ -› <skugga> /ˈskʊga/
<dagur> /ˈdɛavʊr/ -› <degi> /ˈdɛjɪ/, <dagar> /ˈdɛavar/ (I may be wrong about <degi> and <dagar>, though)
<eg> /ɛː/

Eg fleyg hagar í gjár.
/ɛːØ flɛi̯j ˈhɛavar ʊi̯ dʒɔɑr/ (I think)
I flew there yesterday.

Sorry to burst your bubble, we are just as bad as the Icelanders:

<gøta> /ˈkøːta/ (most dialects of Vágar, North Streymoy and Eysturoy preaspirate the t giving: /ˈkøːʰta/)
<skuggi> /ˈskʊtʃːɪ/ -› <skugga> /ˈskʊkːa/
<dagur> /ˈtɛavʊɹ/ -› <degi> /ˈteːjɪ/, <dagar> /ˈdę:aɹ/ (Note: ę is used in Faroese transliteration, but I'm not exactly sure what for. I asked my phonetics teacher, and he said it was some sort of lowered vowel... so it should probably be /ɛ:/. ǫ is used in a similar fashion.)
<eg> /eː/

Eg fleyg hagar í gjár.
/eːØ flɛi̯ːj ˈhę:Øaɹ ʊi̯ tʃɔa:ɹ/ (Corrected)
I flew there yesterday.

/ð/ went through the same changes, so <g> and <ð> are in many places homographs, especially word-finally and inter-vocalically. Which leads me to believe, that Faroese probably somewhere along the road experienced a sound shift similar to the Irish /ðˠ/ -> /ɣ/ one (or how it was, I can't find the post it was posted in). I'll have to ask about it closer though.

The preaspiration of non-geminated plosives in Faroese is as I said limited to a relatively small area of the Faroes (still it's present in roughly 1/4 of the population's dialects). I also think the 'intensity' of the pre-aspiration varies since sometimes I almost hear a sound more similar to /ˣt/ instead of the 'standard' /ʰt/. The pre-aspiration of geminated plosives is present in all of the islands so <-tt(-)> <-pp(-)> and <-kk(-)> are /ʰt:/ /ʰp:/ /ʰk:/ are everywhere.

I think that was all, I might have forgotten to comment on something, and if so, I'll comment on it later.

[finlay]

Faroese has been a little nicer to it, keeping it initially before back vowels and /ø/, but getting rid of it in most intervocalic and final positions, usually by turning it into some other sound such as /v/ or /j/. It has usually become /dʒ/ before front vowels (except for /ø/), but also /j/ in some positions (generally in the same place as where Icelandic does it). It is generally kept before other consonants, though, such as the clusters /gv/ and /gd/ (but */gj/ is /dʒ/).

<gøta> /ˈgøːta/
<skuggi> /ˈskʊdʒɪ/ -› <skugga> /ˈskʊga/
<dagur> /ˈdɛavʊr/ -› <degi> /ˈdɛjɪ/, <dagar> /ˈdɛavar/ (I may be wrong about <degi> and <dagar>, though)
<eg> /ɛː/

Eg fleyg hagar í gjár.
/ɛːØ flɛi̯j ˈhɛavar ʊi̯ dʒɔɑr/ (I think)
I flew there yesterday.

Sorry to burst your bubble, we are just as bad as the Icelanders:

<gøta> /ˈkøːta/ (most dialects of Vágar, North Streymoy and Eysturoy preaspirate the t giving: /ˈkøːʰta/)
<skuggi> /ˈskʊtʃːɪ/ -› <skugga> /ˈskʊkːa/
<dagur> /ˈtɛavʊɹ/ -› <degi> /ˈteːjɪ/, <dagar> /ˈdę:aɹ/ (Note: ę is used in Faroese transliteration, but I'm not exactly sure what for. I asked my phonetics teacher, and he said it was some sort of lowered vowel... so it should probably be /ɛ:/

You're the one who speaks the language... you tell us! Look up the IPA or something.

[johanpeturdam]

You're the one who speaks the language... you tell us! Look up the IPA or something.

But I'm no phoneticist (or phonetician, or how it's called) either.

In any case, I'm gonna go with /ę/ actually being /ɛ/. (Although it might just as well be /æ/, especially, if it's based on E caudata from Old Norse manuscripts.)