Help talk:IPA/Czech and Slovak

transcription changes

Currently, this page offers too narrow a transcription convention. There are a number of things we should do to modify it

Use <o> for both long and short o. It looks like the vowel is actually mid to close-mid, and we might as well use the typographically easiest letter (this is how it's done with Spanish, Greek, and Hebrew).
Remove the diacritics for the diphthongs so that they are either [ou], [ɛu] and [au] or [ow], [ɛw] and [aw]. I'm leaning towards the former but if Czech makes contrasts between e.g. Cau and Ca.u, then the latter is better.
organize the symbols in alphabetical order like they are done in the other IPA for X pages. This also includes not separating sonorants and consonants into different tables.
remove the voiceless version of the fricative trill since it's an allophone and requires additional diacritics that we needn't require editors to use.
note somehow that /oː/. /f/, /g/, and the two voiced affricates appear mostly/only in foreign loanwords.

Then we can create {{IPA-cs}} and start using this page. — Æµ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 18:28, 14 October 2008 (UTC)[reply]

FWIW, I don't know whether you can find true minimal pairs, but there are certainly Czech words where adjacent vowels do not form a diphthong, e.g., použít [ˈpɔ.uʒiːt], nauka [ˈna.uka], neustále [ˈnɛ.ustaːlɛ]. These can be also pronounced with an intervening glottal stop, e.g., [ˈpɔʔuʒiːt]. — Emil J. 12:44, 15 October 2008 (UTC)[reply]

All right. That's reason enough to use w rather than u. — Æµ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 07:23, 16 October 2008 (UTC)[reply]

This needs to reflect what's used in the articles. Currently the voiceless and diphthong diacritics are being used.

I realize the English approximations I just added aren't always very good. Maybe we can add note to the ones that are particularly poor. But this key is for people who know neither the IPA nor Czech, so merely illustrating the IPA with Czech isn't going to do much good.

Some of the articles are going to need normalizing. For instance, Občanský průkaz was transcribed [ɔbtʃʌnskiː pɾuːkʌz]. (I otherwise corrected it but left the flap, which I wasn't sure about.)

I think obvious allophones such as devoicing should be left in. Better a little more work for the editor than to mislead the reader. After all, Czech consonants don't all devoice, and the reader shouldn't need to know which are which.

BTW, I've been through a hundred of these, and none have used w for diphthongs. (Not that that means we shouldn't switch.) I'd rather get the diphthong and devoicing diacritics added to the WP edit window, like on Wiktionary. kwami (talk) 08:34, 12 May 2009 (UTC)[reply]

Things I've noticed to to look for when auto-normalizing with AWB or whatever: "lax" vowels (ʊəʌɔɑ); also maybe e and short i; tie-bar affrics like t͡ʃ (with a space) and t͡ʃ (without); oddities for rz, such as r̩ʒ. kwami (talk) 09:29, 12 May 2009 (UTC)[reply]

English approximations are a good idea, but I'm still not sure about the voiceless fricative trill and the semivowel diacritics. These aren't so important when giving a general pronunciation on Wikipedia pages and there aren't so many pages using the IPA cs template that we can't fix them based on what we decide. — Æµ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 03:37, 13 May 2009 (UTC)[reply]

There are about a thousand articles, not all yet linked to the key. If we remove the diacritic from the articles, then we can remove it from the key, but I really think we need to explain the system we use, not the system we'd like to use. (We can always give deprecated alts in parentheses, as we do w IPA-en.)

We reflect voicing assimilation for all other consonants, so it would strike me as really odd to make an exception for rz just cuz it requires a diacritic. It's also a very salient allophone to English ears. kwami (talk) 08:04, 13 May 2009 (UTC)[reply]

Ahh, now I'm on board for the voiceless fricative trill. I think we should mark the semivowel diacritics as deprecated. It should still be pretty easy (though monotonously time-consuming) to change the deprecated forms. — Æµ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 13:27, 13 May 2009 (UTC)[reply]

I have no strong opinion on the diphthongs, though I don't see how the diacritic hurts. (I have, however, been removing the tie bars from the affricates, though just my luck these are phonemic in Czech.) We can set up AWB to take care of the diphthongs when we go through to fix up the monophthongs. With a thousand articles, there's probably only a hundred that would be a problem, except for ʊ, which is all over the place.

There are also a few articles which use this template for Slovak names. Can we adjust this to cover both? kwami (talk) 14:54, 13 May 2009 (UTC)[reply]