Module talk:it-pronunciation

Geminate affricates

 * Wouldn't it be a good idea to represent the geminate affricates as /at.t͡sa/, /ad.d͡za/, /at.t͡ʃa/, /ad.d͡ʒa/, etc. instead of /ˈat͡s.t͡sa/, /ˈad͡z.d͡za/, /ˈat͡ʃ.t͡ʃa/, /ˈad͡ʒ.d͡ʒa/, etc.? The latter transcriptions suggest that the geminates actually sounds like "stop + fricative + stop + fricative", which I don't think they do. They sound like "long stop + normal fricative", don't they? —Mahāgaja · talk 09:24, 8 February 2019 (UTC)
 * I agree. Benwing2 (talk) 09:29, 8 February 2019 (UTC)
 * Yes, I noticed this as well but wasn't sure how to fix it. – Jberkel 10:37, 8 February 2019 (UTC)
 * Wow, I didn’t know this module was created! But yes, it’s technically wrong to represent them that way because the stop onset of the affricate is the only part that geminates. Otherwise it would mean the same sound is repeated twice, which is not what occurs in standard Italian. [ ˌiˑvã̠n̪ˑ ˈs̪kr̺ud͡ʒʔ ˌn̺ovã̠n̪ˑˈt̪ɔ̟t̪ːo ] (parla con me) 10:51, 8 February 2019 (UTC)
 * I represented geminate affricates with doubled affricate symbols in the phonemic transcription (/.../) to indicate that like other sounds, affricates are geminated. But phonetically ([...]) a doubled affricate is pronounced as a stop with a longer hold followed by a fricative release so the stop–stop–fricative representation is appropriate there. But the Italian Wiktionary uses the stop–stop–fricative representation in phonemic transcriptions, and no doubt people will be confused by the doubled affricate representation, as you all are. — Eru·tuon 11:57, 8 February 2019 (UTC)
 * So does that mean you're willing to change it? —Mahāgaja · talk 12:04, 8 February 2019 (UTC)
 * Done. I was going to put that representation in the phonetic transcription, but never finished that part of the module. — Eru·tuon 22:05, 8 February 2019 (UTC)
 * It would be a great idea. — GianWiki (talk) 13:54, 9 February 2019 (UTC)

fortissimo
I tried to add the pronunciation but the stress was on /ˈsi/. Ultimateria (talk) 18:31, 25 March 2019 (UTC)
 * Did you add an accent mark on the antepenult? The module puts the stress on the penultimate syllable unless there is an accent mark on another syllable. — Eru·tuon 18:46, 25 March 2019 (UTC)
 * I didn't. I guess I assumed it would eventually be possible to infer the stressed syllable but now that I think about it I understand why it's not feasible, in comparison to other Romance languages. Thanks. Ultimateria (talk) 18:51, 25 March 2019 (UTC)
 * ca:Mòdul:it-general has a lot of rules that could be applied to help to determine stress (and distinguish between the pronunciations of o and e) if someone wants to figure out how to integrate them into this module. I didn't because I am not very knowledgeable about Italian. — Eru·tuon 19:43, 25 March 2019 (UTC)
 * As far as I know there are no real rules to determine the pronunciation of stressed o and e, except that it often follows some patterns in the Latin etymon. Curious, but unfortunately can't check the module code right now, the Catalan Wiktionary is blacked out. – Jberkel 22:07, 25 March 2019 (UTC)
 * Rules is sort of a loose term for it. It might be based on looking for particular morphemes that have a certain pronunciation of e or o or a certain accent position. For instance, maybe the module could automatically add an accent to the first vowel of  if no words with those endings are accented on the second vowel. — Eru·tuon 22:21, 25 March 2019 (UTC)
 * Ah ok, I remember looking at a list of suffixes with fixed pronunciation. Maybe this could be added, yes. – Jberkel 09:07, 26 March 2019 (UTC)

The syllable division looks wrong. I know it's not a common consonant cluster, but surely there are enough cases of -sts- to change the module. Ultimateria (talk) 23:55, 26 September 2019 (UTC)

IPA
Hi,.

Thanks for the hard work you've put into programming this module. I noticed that when I used it for the word salumeria, an implausible IPA was produced: /sa.luˈmer.ja/ instead of /sa.lu.me'ri.a/. I just thought I'd bring it to your attention.

Have a good one. -- Dentonius (talk) 06:06, 25 September 2020 (UTC)

Please, disregard my previous comment,. User:Jberkel (thanks, once again), was kind enough to demonstrate the correct usage by adding the accent mark to the word. The module functions as intended. Now if only I had checked the documentation beforehand :-)

All the best. -- Dentonius (talk) 06:24, 25 September 2020 (UTC)
 * In case of this seems to be regular, so maybe it could be encoded in the module itself. BTW, MediaWiki seems to have problems detecting edit conflicts as of recent, this is the second time in 2 days. – Jberkel 06:34, 25 September 2020 (UTC)

S initially before voiced consonant
Hello,. I'm pinging you because you created the template (I wasn't sure where to post this, to be honest). Since the distinction between and  is neutralized when ⟨s⟩ is found in word-initial position, and is followed by a voiced consonant – that is, in clusters like ⟨sb, sd, sg, sl, sm, sn, sr, sv⟩ – I think this should be implemented in the template. For example, results in, while phonemically it should look like  (phonetically, of course there would be ). What do you think? — GianWiki (talk) 22:51, 22 March 2021 (UTC)
 * Well, it could be represented as either as there's no distinction. I have no strong attachment to but happened to decide to use the consonant that more closely matches the phonetics. — Eru·tuon 23:34, 30 March 2021 (UTC)

new features
I added some new features to it-pr and it-IPA. Most notably is the ability to specify the pronunciation of just a part of a word. For example, for, instead of having to write out the entire word to insert a syllable break after ri, you can write this:
 * [ri.a]

Essentially, this first replaces riabbandonare with ri.abbandonare, which then gets processed normally (i.e. it recognizes the suffix -are and puts a stress on it). I refer to this notation as substitution notation. If you need multiple such substitutions in a single word, separate them by a comma, like this for :
 * [ri.a,tts]

You can include multiple invocations of substitutions as well, just like you can include multiple pronunciations, and they can be tagged with references, notations such as  to indicate a traditional pronunciation, etc. For example,  can be written like this:
 * [re.i],[re.i,ri.a]

This specifies two pronunciations re.industrializzare and re.industri.alizzare, cited to DiPI. (-izzare is a recognized suffix so you don't need to explicitly indicate how the zz is pronounced.) You can also include accents as appropriate, e.g. can be written like this:
 * [ìo]

In this case, just [ì] would work as well since there's only one i in the word. This is similar to but more powerful than the notation using.

You can include as much or as little context as you want but it needs to match uniquely; if a given substitution matches in two or more places, you get an error. (Internally, the substitution is first mapped to the raw spelling by removing accents,  and * and mapping ts, tts, dz, ddz, [s], [z], etc. respectively to z, zz, z, zz, s, z, etc.) Benwing2 (talk) 20:33, 11 February 2023 (UTC)
 * BTW you can include things like  inside the brackets but the   can't come directly after an opening bracket because the sequence   confuses the MediaWiki parser into thinking it's the start of an internal link. Benwing2 (talk) 20:34, 11 February 2023 (UTC)
 * Sounds good. A somewhat unrelated question: I've seen cases where you (or your bot) have specified a pronunciation where the pagename would have produced the same one. To what extent should we be specifying pronunciations? Is that just your personal preference? Ultimateria (talk) 22:06, 11 February 2023 (UTC)
 * Good question. Maybe a year ago I went through the top 20,000 or so words by frequency and added pronunciations for them. I did this using a script that tried to generate a default pronunciation, which I then edited by hand as necessary and used another script to push the resulting pronunciations to Wiktionary. A side effect of the way I did this is that all such pronunciations were generated in full. Actually, in general my current preference (I've gone back and forth on this a bit) is not to do this, but use shortcuts when available, because IMO they express the most essential part of the pronunciation while defaulting everything else and also are less error-prone when people edit the pronunciations manually. The only potential concern is that too much defaulting can lead to problems and mistakes; with this in mind, when adding new pronunciations using the  syntax I usually include enough context to clarify what part of the word is being modified (e.g.   instead of just   or even  ). Benwing2 (talk) 22:23, 11 February 2023 (UTC)