Some IPA News

[Taronyu]

So, I couldn't resist, and emailed Frommer with some questions. Here are the questions and the responses: they concern diphthong presentation, secondary stresses, and the u and ʊ distinction.

I have a small question about the diphthongs. (Taken from a thread on LN:) Kemeoauniaea raised some good points in this thread, stating that my IPA notation of the diphthongs may not, in fact, be accurate. He, from his admittedly not-extensive look into IPA, would submit that:

[ ɛj ] [ aj ] [ ɛw ] [ aw ]

be replaced with

[ ɛɪ̯ ] [ aɪ̯ ] [ ɛʊ̯ ] [ aʊ̯ ] .

Note: Wikipedia: The diacritic <  ̯> is placed under the less prominent component to show that it is part of a diphthong rather than a separate vowel, though it is sometimes omitted in languages such as English, where there is not likely to be any confusion.

I initially had brushed this off as just differences in notation, but I think he actually has a really good point. The way I'm telling it now, there's no difference between 'e.wll and new. I think that the diphthong ought to be reflected and different: also compare yem and 'eyng.

So, I figured I'd ask you before I go through changing them, if this is what you envision the Na'vi sounds to sound more like. I know that you've said that:

Open syllables always use a tense "u" pronunciation.  Closed syllables can be either tense "u" or lax "ʊ".  There is no rule for which to use in closed syllables.  So tsun, for example, can be either "tsun" or "tsʊn", but "lu" is always "lu", never "lʊ".

So I wanted to make sure before I go through and possibly mess things up.

And his response! Very detailed, as always.

Sìpawm asìltsan.

There's variation in how people use IPA to transcribe diphthongs.

Kemeoauniaea is right that most transcribers seem to use the vowel+vowel method (with or without the diacritic under the second element in the case of "falling" diphthongs). But [aw], [aj], etc. are also used. Note what the Wikipedia "Diphthong" article says:

"The less prominent component in the diphthong may also be transcribed as an approximant, thus [aj] in eye and [ja] in yard. However, when the diphthong is analysed as a single phoneme, both elements are often transcribed with vowel letters (/aɪ̯/, /ɪ̯a/)."

A well-known textbook for students in introductory linguistics classes, Edward Finegan's "Language: Its Structure and Use," uses the vowel+approximant method, as I tend to do myself. I spoke to Ed about this, and he said he had considered using vowel plus vowel but decided against it.

So the bottom line is that [aj], [aw], [ej], and [ew] are used by some if not all professional linguists in an IPA context; the way you currently have it is justified.

That being said, you and Kemeoauniaea raise a valid point about possible ambiguities in transcription in certain circumstances. A clear example, I think, would be:

maweya ikran

If you transcribe the first word as [maweja], without showing the syllable boundaries, it's not clear whether the structure is (with D=diphthong) CVCVCV, CVCDA, CDVCA, or CDDA. (It's intended to be CVCDA, of course.) If you include the syllable boundaries, however, the ambiguity disappears: [ma.wej.a] can only be CVCDA. Same for 'ewll. It's true that [?ewll] is ambiguous, but if you insert the correct syllable boundary--[?e.wll]--then it's clear that ew is not a diphthong.

So it's really up to you. Either method is, I think, fine, as long as you've indicated the syllable boundaries. Perhaps the easiest solution would be to include a brief note somewhere, saying something like: "An alternate method of transcribing diphthongs, not use here, is . . . " and then illustrate the 2-vowel method.

As for your statement about the transcription of u, it's correct. In a word like tsun, some speakers will use the tense and others the lax pronunciation.

In a follow up email, I asked him for stress information about the following words that have two stress marks, because I've been wondering for ages which was the secondary and which was primary.

Eywa Ngahu <<< I've included this, because it would presumably be said a single intonational phrase, and might have secondary stress. I may be wrong.
Eywa'eveng <<< Shoudn't this have two, being a compound polysllabic word?
iknimaya <<< Please, was there ever a derived form *ik sky or *maya heaven, or perhaps mountain (which we don't have a word for!)
meoauniaea
palulukan
talioang
tireafya'o <<< presumable main on RE
tireapängkxo <<< presumably main on RE as well
Uniltaron

And here is his response!

Primary vs. secondary stress:

In Eywa ngahu, the two stresses are equal.

Eywa'eveng: Yes, you're right: secondary stress on Ey, primary on 'e.

For the others:

In talioang and the two tirea- words, the first stress is primary.

In all the others, it's the second stress that's primary.

As for iknimaya, that's a James Cameron word, and I don't know its derivation

What a good day. ˆ____________ˆ

ta Taronyu

[Kemaweyan]

Irayo

['Oma Tirea]

So it's really up to you. Either method is, I think, fine, as long as you've indicated the syllable boundaries. Perhaps the easiest solution would be to include a brief note somewhere, saying something like: "An alternate method of transcribing diphthongs, not use here, is . . . " and then illustrate the 2-vowel method.

I agree with Frommer here.  Also if {aj}, {aw}, {ɛj}, and {ɛw} really mean {aɪ}, {aʊ}, {ɛɪ}, and {ɛʊ}, then you could notify the reader about that at the beginning of the dictionary of the transcriptions used so that phonetic misinformation is not given out.

[Taronyu]

So it's really up to you. Either method is, I think, fine, as long as you've indicated the syllable boundaries. Perhaps the easiest solution would be to include a brief note somewhere, saying something like: "An alternate method of transcribing diphthongs, not use here, is . . . " and then illustrate the 2-vowel method.

I agree with Frommer here.  Also if {aj}, {aw}, {ɛj}, and {ɛw} really mean {aɪ}, {aʊ}, {ɛɪ}, and {ɛʊ}, then you could notify the reader about that at the beginning of the dictionary of the transcriptions used so that phonetic misinformation is not given out.

Go look at the second page of the dictionary.

['Oma Tirea]

So it's really up to you. Either method is, I think, fine, as long as you've indicated the syllable boundaries. Perhaps the easiest solution would be to include a brief note somewhere, saying something like: "An alternate method of transcribing diphthongs, not use here, is . . . " and then illustrate the 2-vowel method.

I agree with Frommer here.  Also if {aj}, {aw}, {ɛj}, and {ɛw} really mean {aɪ}, {aʊ}, {ɛɪ}, and {ɛʊ}, then you could notify the reader about that at the beginning of the dictionary of the transcriptions used so that phonetic misinformation is not given out.

Go look at the second page of the dictionary.

Nang nga lu win!  Nice update

[Taronyu]

Nah, I did that this afternoon, before I knew if I could post this email or not. Hrh.

[Kì'eyawn]

Ma Taronyu, do we know if this stress information holds for the other "Unil-" compounds?  i.e., Uniltìranyu, Uniltìrantokx?

[okrìsti]

tsaw oeyä eltur tìtxen si
Added it to wiki.
There are no other single words matching to \[[^\]]*ˈ[^\]]*ˈ[^\]]*\] (IPAs with two primary stresses).

This unil-compound question is interesting, but compare for example way and telem becomes waytelem or both pongu-compounds, hm...

[Lrrtoksì nìhawng]

Tewti! Fì'u txantsana fmawn leiu!

By dumb luck, I was pronouncing these correctly the whole time. It gives me that warm fuzzy feeling. Since it's important to show the syllable boundaries anyway, I'd be inclined to go with that method along with Frommer's suggestion about having a note explaining the alternate method for treating dipthongs. As for the iknimaya question, I remember hearing it pronounced as iknimaya in an interview. I want to say it was James Cameron but it could have been someone else associated with the movie.

[Taronyu]

Ma Taronyu, do we know if this stress information holds for the other "Unil-" compounds?  i.e., Uniltìranyu, Uniltìrantokx?

No, we don't. But Frommer has never given us two stresses for those.

[kewnya txamew'itan]

I never said your notation wasn't accurate, just that it was the minority transcription from what I'd seen.

Whatever, we've got a Frommerian carte blanche so that's good.

On an unrelated side-note, as regards u, closed syllables include those without an onset if I understand correctly so uvan and unil (and derivatives) could both take the lax pronunciation ko? The only reason I ask is that I had though that the pseudo-vowels were only allowed in open syllables but then found out that they had to have an onset so I'm just checking that when we say closed we actually mean closed and not something broadly similar as we did with open syllable ll and rr.

[Taronyu]

I never said your notation wasn't accurate, just that it was the minority transcription from what I'd seen.

Whatever, we've got a Frommerian carte blanche so that's good.

On an unrelated side-note, as regards u, closed syllables include those without an onset if I understand correctly so uvan and unil (and derivatives) could both take the lax pronunciation ko? The only reason I ask is that I had though that the pseudo-vowels were only allowed in open syllables but then found out that they had to have an onset so I'm just checking that when we say closed we actually mean closed and not something broadly similar as we did with open syllable ll and rr.

Sorry man! Must have misread.

And I don't know about open syllables. I always thought that onsets didn't matter, but I'm not sure.

[wm.annis]

And I don't know about open syllables. I always thought that onsets didn't matter, but I'm not sure.

I am.    Onsets don't matter for determining if a syllable is open or closed, only a coda.

[omängum fra'uti]

For rr and ll, frommer does not only explain their structure in terms of requiring an open syllable.  Rather he describes them as taking the place of the vowel in CV syllables specifically.  So the required onset is just the nature of the psuedovowel syllables.  (It just so happens that also means they must be open syllables, but it is not the defining feature.)

[kewnya txamew'itan]

For rr and ll, frommer does not only explain their structure in terms of requiring an open syllable.  Rather he describes them as taking the place of the vowel in CV syllables specifically.  So the required onset is just the nature of the psuedovowel syllables.  (It just so happens that also means they must be open syllables, but it is not the defining feature.)

Ok, so it was just me reading a flawed explanation. This would mean that uvan could be pronounced with a lax u then if Frommer's always used the terms properly.

[wm.annis]

Ok, so it was just me reading a flawed explanation. This would mean that uvan could be pronounced with a lax u then if Frommer's always used the terms properly.

No, no, uvan needs tense /u/.  Remember, /v/ can never be a coda in Na'vi.  It must be syllabified u.van.

[Taronyu]

For rr and ll, frommer does not only explain their structure in terms of requiring an open syllable.  Rather he describes them as taking the place of the vowel in CV syllables specifically.  So the required onset is just the nature of the psuedovowel syllables.  (It just so happens that also means they must be open syllables, but it is not the defining feature.)

Ok, so it was just me reading a flawed explanation. This would mean that uvan could be pronounced with a lax u then if Frommer's always used the terms properly.

No, because uvan splits u.van. Ul.te, on the other hand, works.

[kewnya txamew'itan]

Ok, so it was just me reading a flawed explanation. This would mean that uvan could be pronounced with a lax u then if Frommer's always used the terms properly.

No, no, uvan needs tense /u/.  Remember, /v/ can never be a coda in Na'vi.  It must be syllabified u.van.

Gah! Forgot that.

Unil though can be lax then?

For rr and ll, frommer does not only explain their structure in terms of requiring an open syllable.  Rather he describes them as taking the place of the vowel in CV syllables specifically.  So the required onset is just the nature of the psuedovowel syllables.  (It just so happens that also means they must be open syllables, but it is not the defining feature.)

Ok, so it was just me reading a flawed explanation. This would mean that uvan could be pronounced with a lax u then if Frommer's always used the terms properly.

No, because uvan splits u.van. Ul.te, on the other hand, works.

I really need to remember about v, s, f, z, h and ts.

[wm.annis]

Unil though can be lax then?

It looks like Frommer syllabifies this /u.nil/ (accent on the first syllable), so again tense /u/ here.

[Muzer]

U.nil does seem the most natural way of saying it. I think I read somewhere that most Na'vi tends to "want to" fall into a CV.CV.CV pattern, and usually does so unless the rules prohibit it.