Template talk:hrx-pr

Feedback

@Stríðsdrengur, Trooper57 The template is mostly ready, but there are a few details that might need some discussion. The module works with the orthography used in {{R:hrx:Boll}}, and the generated pronunciations are closely aligned with the ones from the dictionary. The main difference is that the dictionary shows some unstressed long vowels and diphthongs (albeit rarely), while the template always adds stress to them. Other than that, my main concerns are the following:

• The differences between /x~ç/ and /r~ɾ/ are not really phonemic, as they are in complementary distribution. Should we stick to /x/ and /r/ in the phonemic transcription or use all four symbols?
• The dictionary uses /m̩n̩ŋ̍/ instead of /əm ən əŋ/. Should we use the former even immediately after another vowel?
• Should the template support /t͡ʃ/ as an independent phoneme? It seems to be marginal, with just a few occurrences in the dictionary.
• Should the sequence qu be rendered as /kv/ or /kw/?

Right now the template generates a single phonemic transcription, but maybe adding a phonetic one for cases like these could be a good idea. Also, it would be easy to hide the phonetic transcription when it doesn't differ from the phonemic one, in order to keep the pronunciation section less cluttered. There's an example of what I mean here (showing /kv/ and [kw] in the first word, but in the second word only the phonemic transcription is shown).Santi2222 (talk) 16:11, 16 September 2024 (UTC)

Hi Santi, sorry for the late delay, I've been out of time all day, maybe @Piterkeo and @Iohanen'll be interested. Stríðsdrengur (talk) 23:19, 16 September 2024 (UTC)

Hi, Santi. My suggestions:

• Keep /x/ and /ç/ as separate. They are in complementary distribution (except for diminutives) but just like in Standard German it seems better to let them separate in my opinion. Regarding /r~ɾ/, sticking to /r/ is fine.
• In more recent versions of my dictionary (not publicly available yet), I do not use /m̩n̩ŋ̍/ anymore and opted for /əm ən əŋ/ instead. I also changed the semivowels from /ɪ̯/ and /ʊ̯/ to /j/ and /w/, thus /aj aw oj ɔj uj/ instead of /aɪ̯ au̯ oɪ̯ ɔɪ̯ uɪ̯/.
• I don't see much of a problem in replacing /t͡ʃ/ for /tʃ/, but it has about the same frequency of occurrence as in standard German.
• I prefer /kw/ instead of /kv/ since no speaker pronounce it as [kv].
• I also abandonded the use of /ɐ/ in more recent versions, using /a/ instead. Although the phone is [ɐ], the speakers perceive it as simply the same phoneme as /a/.

Piterkeo (talk) 02:56, 17 September 2024 (UTC)

Thanks four your suggestions! Now the sequences qu,tsch yield kw, t͡ʃ by default (if needed, writing t.sch will generate tʃ instead). Regarding the vowel /ɐ~a/, the template adds a non-syllabic diacritic (ɐ̯) when it is part of the word for (shown as fɔɐ̯), in the sequence äer (as ɛːɐ̯) and in word-final stressed er (as ɛɐ̯). However, it appears without a diacritic in word-final eer (shown as eːɐ) and word-final oer (shown as oːɐ). Regardless of the letter used for this sound, should the template add the non-syllabic diacritic in these two cases too? About ɪ̯ʊ̯~jw, both would probably work fine (the former makes the transcription more similar to that of Standard German, the latter makes the transcription somewhat cleaner, with less diacritics), so I have no strong opinions about which option we should use. Santi2222 (talk) 15:20, 17 September 2024 (UTC)

I suppose you could leave it with or without the diacritic, regardless of it being ɐ or a. I would transcribe Äerd as /ˈɛːat/ [ˈɛːɐt]~[ˈɛjɐt], for instance. Pinging @Piterkeo because he might've not seen your answer. Iohanen (talk) 22:39, 21 September 2024 (UTC)

The non syllabic diacritic should not be used for [ɐ] in any case. The speakers pronounce it as a full vowel in a separate syllable. So I would transcribe for, er, wer, Äerd, Meer, Hoer phonologically as /ˈfɔa/, /ˈɛa/, /ˈvɛa/, /ˈɛːat/, /ˈmeːa/, /ˈhoːa/. Sorry for the delay in replying! Piterkeo (talk) 15:29, 4 October 2024 (UTC)

No problem! I've changed the module so now it outputs /a/ instead of /ɐ~ɐ̯/. I've also fixed a bug involving the transcription of er and wer. I'd say the module is ready for deployment, unless there's something I've missed. Santi2222 (talk) 22:10, 8 October 2024 (UTC)

It seems that everything is now fine, thank you very much for your work and effort🙏 Stríðsdrengur (talk) 00:44, 9 October 2024 (UTC)

Semivowels

@Santi2222, is there a way to respell falling diphthongs? I'm having problems with some terms from Portuguese like Kui (/ˈkʰuːi/ instead of /ˈkʰuɪ̯/) and Pnëu (/ˈpneːu/ instead of /ˈpneʊ̯/). I needed to type the IPA directly because W and J give different phonemes. Trooper57 (talk) 21:56, 23 October 2024 (UTC)

Currently these non-native falling diphthongs are not supported. I'll take a look into it in a couple days. Santi2222 (talk) 15:56, 25 October 2024 (UTC)

@Trooper57 Done, I've updated the documentation and the two examples you provided. Santi2222 (talk) 22:22, 9 November 2024 (UTC)

Thanks. Trooper57 (talk) 22:25, 9 November 2024 (UTC)