Talk:Haida language/Archive 1

The consonant table

Well, I thought this page was very empty. There was only the introduction which said "at least 46 consonants" but didn't mention one of them, and the external links. That was a pity. So I copied the table from the Navajo language page, went to www.haidalanguage.org, compared the orthography page and the sounds page there, and then modified the table accordingly.

First of all I don't speak Haida, nor have I heard any Haida other than the sound files mentioned above.

I am an armchair linguist – no training in linguistics other than what I have read on my own –, so clearly it's possible that I've made all manner of mistakes. Apart from that, the quality of the recordings is not breathtaking, and the examples are not enough to do phonology from them.

My guess that the plain stops are voiceless when word-initial and voiced when word-internal is probably defensible, but I don't here pure (!) voicing contrasts well. Maybe a Spanish speaker could help. (The example sound file for d probably has a voiced [d], but it's prenasalized, so maybe it counts as word-internal and preceded by /n/ – what do I know!)

My claim that x c c is palatal rather than velar is certainly correct for the example sound file, but there it precedes an [i]. Maybe there's allophony like in German?

The same holds for l being velarized-or-whatever and 'l not being such a thing.

Based on the plosives and on many other Native American languages, I would have expected an aspiration contrast in the affricates. Surprisingly, ch is not aspirated. In spite of this, j is not voiced, at least in its example sound file (where it's word-initial). Still they don't sound exactly the same. :-o This must be a pure fortis-lenis contrast. If so, it is the first example of such a phonemic contrast outside of my native southeastern German (but there the contrast is in the plosives and maybe in /s/).

(Ts isn't aspirated either, but there's nothing to contrast it with – and, based on the plosives, it isn't supposed to be aspirated in the first place because, in the example sound files, it only occurs word-finally. Does anyone know if it occurs word-internally as well?)

I wrote the glottalized nasals and approximants as ejective. While common, that's nonsense, except perhaps from a phonological point of view. Those for which there are example sound files are, to my ear, just consonant clusters that begin with a glottal stop (or perhaps more precisely glottal stops with all manner of funny releases?), except for 'y. From what I've read on glottalized sonants, I guess that's creaky voice – but I don't recognize creaky voice when I hear it.

Update: I'll replace the ejective diacritic with the general glottalized diacritic in a few minutes. That doesn't mean much, but at least it can't be wrong.

David Marjanović | david.marjanovic_at_gmx.at | 13:50 CET | 2006/11/9

In the example sound file of k' q' q' I hear a simple glottal stop. Is that just me, or is it the quality of the recording? (Of course I'm not used to hearing uvular ejectives!)

I had to guess which dialect has the famous epiglottal trill. Alaskan Haida does not, as shown by the sound files; Skidegate Haida does not, because it doesn't have any epiglottal consonants; that leaves Masset Haida… among the dialects I've ever read of! If you know better, please "be bold".

BTW, I wonder how the figure of 46 was arrived at. :-)

Now... does someone understand the vowels? Apparently there /a i u/ in short and long versions plus high and low tones, but www.haidalanguage.org doesn't even try to explain that.

David Marjanović | david.marjanovic_at_gmx.at | 2:35 CEST | 2006/5/4

Addendum: I can't quite hear whether the lenis affricates are voiced word-internally and whether tl is aspirated.

David Marjanović | david.marjanovic_at_gmx.at | 23:11 CEST | 2006/5/5

According to Enrico (2003, p. 12), the Masset dialect has all of the Skidegate consonants as well as "pharyngealized ħ and...a pharyngealized glottal stop such as is also found in Nootka...not to be confused with the Arabic voiced pharyngeal fricative (ʿain)." Thus, I think [ʔˤ] is the proper representation for <r>, and [ħ] the proper form for <x> (rather than epiglottals). As far as I can tell, Enrico makes no mention of an epiglottal trill. neatnate 02:59, 17 October 2006 (UTC)

Well. I'm willing to bet Enrico simply didn't get the idea these sounds might be epiglottal – according to the Epiglottal consonant article, many pharyngeals all over the world, explicitely mentioning those of Haida, have turned out to be epiglottals. Maybe Enrico didn't even know the term "epiglottal" – if you read Starostin's paper on Dené-Caucasian comparative phonology (external link at the bottom of the Dene-Caucasian languages article), you'll find that Starostin scrupulously distinguishes pharyngeal, epiglottal, and glottal consonants, but never uses the term "epiglottal", instead you have to burrow your way through several confusing terms like "emphatic glottal" and the like.

Listen to the sound files of Agul, the apparently only language that has both pharyngeal and epiglottal consonants. To me, the sound samples at haidalanguage.org (disclaimer – that's all Haida I've ever heard; I have no idea of dialectal differences or anything!) sound epiglottal to me, not pharyngeal, and the thing sold as a pharyngeal stop sounds like a voiceless affricate that ends in the voiceless epiglottal fricative.

What, BTW, is a "pharyngealized ħ" supposed to be? [ħ] is already pharyngeal by definition.

David Marjanović | david.marjanovic_at_gmx.at | 13:53 CET | 2006/11/9

Actually… wouldn't a pharyngealized glottal stop sound rather different? In the sound file for g, I can hear a loud and clear voiceless fricative following whatever the stop is. Generally, pharyngealized stops (or approximants) are not affricates, right? Wouldn't you at least expect a voiced fricative instead of a voiceless one? The Nuu-chah-nulth language article says the sound in question is a glottalized pharyngeal approximant, so I'd expect to hear a voiced approximant rather than a voiceless fricative.

Update to the above: Actually the Haida fricative doesn't sound very much like either of the Agul ones to me; not more like one than the other, at least when I take into account that the [ʜ] of Agul is partly trilled.

David Marjanović | david.marjanovic_at_gmx.at | 15:58 CET | 2006/11/9

OK. Since nobody has produced evidence for them being pharyngeal, I'll change them back to epiglottal in a few minutes hours. If you find evidence, please reverse this. David Marjanović 12:38, 2 March 2007 (UTC)| edit David Marjanović 13:08, 2 March 2007 (UTC)

Done.

Additionally, I have noticed that someone changed the plain consonants to [p], [t], [k] and so on. I know that the combination of [b], [d], [g] etc. with the voiceless diacritic is unorthodox. However:

• Removing it gave j and ch the same transcription, but they are not pronounced the same; they are even different phonemes (and therefore distinguished in all three orthographies). I have put a mention of this into the notes under the table. Since both are voiceless and both are unaspirated, I can't think of another way to transcribe them than the way I did.
• The same user transcribed dl as [tɬ]. This is wrong: the sound does not end in a voiceless fricative, it ends in a voiced approximant.
• While globally rare, a pure fortis-lenis contrast ( = between two sets of voiceless unaspirated consonants that do not differ in length) exists elsewhere, for example (though only for plosives) in most Bavarian-Austrian dialects of German (such as my native one) and in Austrian Standard German. The contrast seems to be the loudness of the release, produced by different amounts of (pulmonic) air pressure. Indeed, b, d, g sound like my own /b/, /d/, /g/, which are identical to the /p/, /t/, /k/ of Spanish, Hindi or Thai; they don't sound like my /p/, /t/, /k/, which are identical to the /p/, /t/, /k/ of languages like French or Russian but not Spanish. Dl starts with my /d/, not with my /t/, and I also hear the plain uvular plosive and j as lenes.

While I was at it, I restored the tie bars because there's no other way to write lateral affricates in the IPA ([tɬ] would be a consonant cluster, not a single affricate phoneme).

I used a more obscure apostrophe for the ejectives (U+0315). It fuses more closely to the preceding letter, which may mean it is the actual IPA diacritic. Unfortunately (like the tie bar) it doesn't display in IE 7 unless the font is specified, so I can only see it in the table and not in the notes.

To circumvent the bizarre bugs in Internet Explorer 7 (not 6!) and its treatment of Arial Unicode MS, I wrote a font family into the table itself. Now everything gets displayed correctly (except for the Arial Unicode MS bug that misplaces the tie bar), whether or not the IPA template is present. (Thinking the template would screw things up, I deleted it from the first line, but it doesn't.)

Unfortunately I can't hear the sound files at The Sounds of Haida anymore. Windows Media Player 10 and 11 find something wrong with them and refuse to play them. I'd be grateful for any help…

David Marjanović 12:59, 3 March 2007 (UTC) | edit David Marjanović 13:08, 3 March 2007 (UTC)

Correction: I can still hear some of the sound files. Not those with j or ch, though. David Marjanović 14:41, 3 March 2007 (UTC)

Life article

Life magazine had an article on the language in its issue for the weekend of January 19, 2007. Perhaps that can be an article resource? Mapsax 20:42, 20 January 2007 (UTC)

Sounds great, but Life doesn't seem to have online archives, and I won't find it on paper either here in Paris. What does the article say? David Marjanović 12:38, 2 March 2007 (UTC)

The epiglottal trill

I just noticed that someone had removed the mention of the epiglottal trill. Please! Masset Haida is famous for its epiglottal trill! See:

• Epiglottal consonant
• Voiceless epiglottal fricative
• Epiglottal trill

So I have restored that. If you want to argue with this, please do so – here – and then remove it. David Marjanović 13:40, 3 March 2007 (UTC)

Sources?

Does anyone have a direct source from linguistic literature that Masset Haida has an epiglottal trill? Also where is the source that the Haida consonants are epiglottal and not pharyngeal. Thanks. Azalea pomp 21:15, 14 April 2007 (UTC)

• Ask the authors of the three articles mentioned above – I'm not among them. IIRC one of the articles cites a book by Ladefoged – I don't see why we shouldn't trust that.
• Ask yourself whether a pharyngeal affricate is physically possible… :-)
• Is there any source that the Haida consonants are pharyngeal and explicitely not epiglottal? So far, all claims that they are pharyngeal have – as far as I can tell – been made by people who didn't know that epiglottal consonants exist in the first place and therefore had to recognize any epiglottal consonants as the closest place of articulation they knew, pharyngeal. David Marjanović 23:37, 16 April 2007 (UTC)

Well, it isn't so much that I am disputing anything, but I want to read the sources for myself. :) Azalea pomp 05:40, 17 April 2007 (UTC)

Consonant table obscured?

I'm new to Wikipedia and don't have a firm enough grasp of the markup to try to fix this myself. A good bit of the consonant table is obscured by the info box (at least in my browser, Firefox).

Is there any way to fix this? I've noticed this problem on a couple of other language pages as well. Kutsuwamushi 22:40, 23 May 2007 (UTC)

I use Firefox as well, & I'm not experiencing this problem. Maybe it's caused by something else -- perhaps operating system issues? (On the machine I'm on right now, I'm using Windows XP.) --Yksin 22:44, 23 May 2007 (UTC)

Windows XP, IE 7 here – no problems, the consonant table is below the infobox (creating a large empty space which will hopefully encourage someone to fill it). David Marjanović 19:54, 3 June 2007 (UTC)

I'm using Firefox and am experiencing it. OS is irrelevant but perhaps the version of Firefox is relevant: the one I'm using right now is 0.9.2: old, but good enough for just about anything else that I want to do. -- Hoary 01:52, 27 October 2007 (UTC)

Consonant table again

I've restored the font-family specification because there's no other way to make the table legible in Internet Explorer 7 (though strangely not 6). I've also restored the tie bars.

I'd be grateful for references to certain other things. For example, someone moved [ç] to [x] – as mentioned above, the only x in the audio files is an unmistakable [ç], but it's in front of an i, so I wouldn't be surprised if there's allophony like in German, but that alone doesn't show that [x] actually exists in the language.

David Marjanović 21:43, 27 September 2007 (UTC)

Relationships

The word for language, "kil"?? Extremely similar to Finnish or Estonian. What related? —Preceding unsigned comment added by 219.173.119.57 (talk) 06:24, 9 November 2007 (UTC)

Well, I wonder what the Swadesh list would look like. Unfortunately, I have no Haida dictionary. And I doubt I will ever be able to obtain one here in the Czech Republic... :.( --Pet'usek ^{[petrdothrubisatgmaildotcom]} 15:41, 10 March 2008 (UTC)

You can download one over on ERIC: link --Friendly Cave (talk) 16:51, 8 September 2008 (UTC)