User talk:Srtª PiriLimPomPom

Portuguese pronunciation

Hi. Thanks for your contributions. In general your changes look good although I reverted one of your changes because you put the Pronunciation section in the wrong place. I think your intention was to indicate an open first 'o' in abobrinha. I made that change using 'ọ', which is how you do it using the automatic pronunciation template. I have a question, though ... under what circumstances do diminutives in '-inho'/'-inha' preserve the original vowel quality of the formerly stressed vowel? I think this may happen e.g. in certinho and velhinho but I don't know for sure whether this always happens, and I don't know of any good online dictionary containing Brazilian Portuguese pronunciations (Infopédia has European Portuguese pronunciations but this phenomenon isn't the same there, see [1] for an example). Benwing2 (talk) 22:02, 15 October 2022 (UTC)

Thank you very much for your help! That was indeed my intention.

And it's in all cases, in all dialects. Sozinho, cafezinho, canequinha, sogrinha and fezinha (betting on a lottery) all have open-mid pronunciations, while forcinha (support, often implied as a small financial contribution), corzinha (healthy blush, food browning), coxinha and forminha (baking mold) all have close-mid pronunciations.

In general, even the vowel reduction patterns (that tend to be more accentuated in longer words) are relatively unchanged, though people are more likely to pronounce the e in queridinha as [ɪ] than to pronounce it that way in querida. (Speaking of which, Portuguese /ĩ/ and /ũ/ are really [ɪ̃] and [ʊ̃]. We only have [i] and [u] in stressed oral monophthongs and diphthong nuclei. Shouldn't the Wiktionary IPA key for the brackets version be radically changed? In truth, there are many other ideas of true Portuguese narrow transcription I'd like to suggest, such as /l/ velarizing before /i/ in most of Brazil. Also, how to include [ɰ̃] – I would be more likely to interpret canto [ˈkɜ̃tʊ] as a corner and [ˈkɜ̃ɰ̃tʊ] as that OR singing. I would try to promote that if I still had a PC...)

We don't have a culture of using pronunciation keys in dictionaries, and even though that could be useful, I didn't find any dictionary that indicated whether an e or o should be pronounced open or closed in a cursory Google search. Srtª PiriLimPomPom (talk) 10:06, 16 October 2022 (UTC)

Oh, and speaking of narrow transcriptions, can you take a look at what happened to coroa? About everyone would have at least one glide in natural speech in that word. How can this be indicated? Srtª PiriLimPomPom (talk) 10:17, 16 October 2022 (UTC)

Thanks so much for the comments. I have plenty more questions, can you take a look at the section labeled "Questions for Ungoliant" in Module talk:pt-pronunc? These are questions for another Brazilian Wiktionary editor, who unfortunately hasn't responded. Some of those questions concern South Brazilian pronunciation specifically because I'd like to clean up the 2,000+ manually annotated South Brazil pronunciations, but some of them concern BP more generally. Some of the most pressing questions concern how to correctly handle e/i and o/u in hiatus in the module (e.g. in distância, adicionar, balancear, atual and a bunch of other words). I have no online source for how to pronounce these vowels other than the Collins Online Dictionary e.g. [2], which I don't trust very much. As for narrow transcriptions, please make a list of changes you want implemented and I'll try to implement them. I have definitely heard the velarized /l/'s, for example; I was taught that only people in Portugal velarize their /l/'s, but I remember noticing it fairly obviously when listening e.g. to Timbalada singing (although maybe they were exaggerating the velarization for effect, I don't know). As for pronunciation dictionaries for BP, I purchased the Houaiss dictionary, we'll see how helpful it is once it arrives. As for coroa, I currently add a glide in Portugal speech labeled 'regional', but not for Brazil; how should it be notated? Benwing2 (talk) 18:40, 16 October 2022 (UTC)

I am going through and trying to update all the -inho/-inha pronunciations correctly. Note that the module automatically handles the -zinho/-zinha endings correctly (or at least, it should; see alicatezinho for an example). It seems that even the epenthetic /(j)/ before final /s/ applies in diminutives, as in arrozinho; I added a special hack to handle this case (which looks like it's not quite right yet, because it generates /a.ʁo(j).zˈĩ.ɲu/ instead of /a.ʁo(j)ˈzĩ.ɲu/ [UPDATE: fixed]). I also imagine that aninho (diminutive of ano) has a nasal a, different from aninho (first-person verb form of aninhar). What about amarelento, amarelão, amarelado (in the meaning "somewhat yellow", not in the past participle of amarelar), amarelaço? Do these all preserve the open e that is found in amarelo? Benwing2 (talk) 19:47, 16 October 2022 (UTC)

Similar questions about bolada (“blow with a ball”), bolão (“large ball”), boleiro (“soccer player”), bolim (“type of ball”), bolita (“marble”), bolica (“(same)”), and similar words based on bola (“ball”), which seems to have /ɔ/ in it. Benwing2 (talk) 20:14, 16 October 2022 (UTC)

OK, a few more questions: (1) Is the first vowel of fininho, fominha, hominho, etc. nasalized? (I'm guessing yes.) (2) What about the first vowel of camarinha? (I'm guessing yes.) (3) What about the first 'a' in escrivaninha? It is pronounced nasal and closed? (I'm guessing yes.) (4) Plurals like arezinhos, florezinhas: Is the 'e' pronounced like /i/? (I'm guessing yes.) Thanks. Benwing2 (talk) 00:47, 17 October 2022 (UTC)

For the hiatus vs rising diphthong, it's difficult. Most words really are actually variable, but way more likely to trigger one or the other. With that said, in almost all cases where European Portuguese has a diphthong, in Brazil it may occur in very fast speech, particularly that of men. This is the case for balancear.

I would pronounce coroa as [koˈɾoʊ̯.wə], but [koˈɾo.wə] is also common and [kŏˈɾoʊ̯.ɘ] or [kə̆ˈɾoʊ̯.ɘ] (yes, last vowel closer than schwa) is a possible pronunciation if someone is speaking really fast and with a really low voice, but I wouldn't really transcribe it here, otherwise we should probably honor the way I pronounce -inho as [ɪ̃ɥ̃] (eye dialect -im, but it's definitely not the same generic nasal vowel m as [ɰ̃] for me, and I really doubt it's rare for other people to also maintain a purinho-Purim distinction here) in the same contexts. And what I mean is that [ɪ̃ɥ̃] is more common than [koˈɾo.ɐ], I can't imagine even [ˈleɪ̯tɪ ˈkẽ̞tɪ] or ['leɪ̯te ˈkẽnte] ([ˈɫeɪ̯tɕ 'k̟ʰẽ̞ɨ̞̯̃tɕ] for me for reference) dialect speakers pronouncing it that way, for the same reason hiatuses are dysphonic in English and people pronounce paella as [paɪ̯ˈ(j)ɛlə] or [paɪ̯ˈ(j)eɪ.(j)ə] (I have seen epenthetic /h/ be used instead for hiatuses in Brazilian acronyms in international media). The only exception is when people pronounce their surname "Corrêa, com acento/circunflexo" out loud, while someone else is writing it down, to distinguish it from Correia.

.

Lack of L velarization (or li lheísmo?) in Brazil before /i/ corresponds very neatly to people who have [ni] for ni instead of [n̠ʲi], and who sometimes fail to produce [tʃ ~ tɕ] and [dʒ ~ dʑ] before /i/. Most Brazilians are condistent ti/di palatalizers. alíquota is a common word in TV news that would probably be indicative of that, a pity we can't get an extensive list of it being pronounced. But even without that, a file of someone who uploaded their own pronunciation of Lituânia to Commons was added to the Wikipedia IPA dark L article (later merged with the other [l]) a few years back, but people keep deciding that it's actually clear. They agree that the pronunciation is dental, but the file does not sound like a clear dental L at all to me, I can actually notice the difference [3] I don't think the sources disagree with dark L, though, the PDF on Portuguese and Brazilian accents in Italian does discuss velarization in the case of Brazilian accents as well. Wayback Machine PDF Maybe some/most people actually have a dento-alveolo-palatal pronunciation that is mistaken for a dark L, but it's not a clear dental L regardless.

.

Now for the easy part, 1. you're right about alicatezinho, arrozinho and aninho.

2. I think we need to ask a sulista and a paulista here, I have vowel harmony in pororoca and free variation between close-mid, mid and open-mid in perereca and fofoca, and a mid > close-mid pattern in peteca and coroca unless I am emphatic (mid <<< close-mid for the first vowel in fofocar, second is always [o]). I would guess amarelão always gets an [ɛ], but about amarelado and amarelaço have my type of free variation; amarelento could have a lot of people pronouncing it close-mid [e] due to vowel harmony with close-mid [ẽ], I have open-mid > mid but my /ẽ, õ/ are almost always mid.

3. Bolada as in blow with a ball or being hit by a ball has open-mid, but it doesn't in irritated (f) (Rio slang, maybe popularized nationwide with Perlla - Tremendo Vacilão). Bolão as in large ball or bumboozer has an open-mid, but bolão and bolada in a jackpot sense have free variation, with vowel harmony perhaps getting a weak tendency for open-mid in Rio de Janeiro, Brasília, the Northeast and other places with an accent that does that. Boleiro (a particularly good professional soccer player, someone who thinks they are that, or man who has a fascination with soccer to a degree deemed annoying) conversely is probably always [o] in our dialects due to vowel harmony (go figure) but idk about the rest. Bolita I would have [ɔ], bolim is not a part of my vocabulary but it's pronounced with [o] here YouTube 1:05, and I would pronounce bolica as [boˈɫikə ~ bʊˈɫikə] by analysis with pelica [peˈɫikə ~ pɪˈɫikə].

4. Fininho [fɪˈnɪ̃j̃ʊ] > [fɪ̃ˈnɪ̃j̃ʊ] but the latter isn't odd. Fominha "ball hog" more likely to be less nasalized than fominha "slight hunger" because of the intonation we use (ball hog, an insult, is pronounced more emphatically), but it's really free variation to the most part. In Rio I think women and adults are more likely to nasalize it in both cases than men and children, but I don't think I have ever heard the ball hog sense in other dialects. I'd say it's best transcribed with [õ] or both. likewise for hominho "weak man/amarelão", slightly more likely not to be nasal than hominho "small man" (we are taught that homenzinho is the correct form for the latter) or hominho "action figure, male doll in general". Camarinha is unique in that it could have [a] in the second a, unlike almost any natural pronunciation of câmara, which would get [ɐ ~ ɜ ~ ə]. Correct for escrevaninha. I have /i/ for the plurals but I can totally see people from southern Brazil and Mato Grosso do Sul having /e/ instead.

As for that discussion, I think 1. it's hard for a Brazilian Portuguese speaker to differentiate between /k/ [k̟ ~ k̟ʰ] not followed by a vowel, followed by a voiceless [ɪ] or followed by a very short voiced [ɪ]. This is probably accentuated before sybilants and affricates.

2. Paulistas and sulistas do lack most of the nasalization in words like cama, pronounced [ˈkəmə] in the São Paulo metropolitan area (derisively called the rinite) and [ˈkəmə ~ ˈkɜmə ~ ˈkɜmɜ] elsewhere. But words like agente always have at least some nasalization, albeit it is minor. I'd compare their /ẽ/ in agente to my /õ/ in fominha and hominho, it's there but not like THERE like in fã for them and lenha for me. With that said, I think transcribing sulista homem as [ˈome] is an exaggeration, it sounds totally different from an imaginary word hôbe.

3. I don't know why, but it's very difficult for Brazilians to perceive a coda [z] as a [z] rather than an [s]. This is a real struggle for me as an English learner, a lot of those coda /s/ vs coda /z/ minimal pairs don't sound like different words to me at all. It's definitely [izmo ~ izmʊ] but most of us can't hear it as such (no one spells the z in eye dialect, like in English), it's also produced very differently from an onset [z]. The same doesn't apply to telling [ʃ,ɕ] apart from [ʒ,ʑ] in chiado speakers.

4. Also, it's funny that ramen is mentioned above. Ramen is not nasalized in the a when pronounced rámen. Miojo can be pronounced as [m(ʲ)ɪˈj̆oʒʊ ~ m(ʲ)ɪˈj̆oʑʊ] and [ˈmi̯oʑʊ ~ ˈmi̯oʑʷ]. I think it would be adequate to transcribe these as [ɪ.(j)V] [Vɪ̯.(j)V] [ʊ.(w)V] [Vʊ̯.(w)V] and disregard the pure rising diphthong version as just fast prosody.

5. Verbs are mentioned. I would say vozear, prosear and golpear can be pronounced in free variation [ɪ ~ e ~ e̞ ~ ɛ], and chimarrear [ɪ ~ e], but the others probably are reduced to [ɪ] in most natural speech. Also, for most of those verbs, I think they should be transcribed as [Cɪˈah] rather than [ˈCjah], even though Ungoliant transcribed them otherwise, based on the miojo observation. Srtª PiriLimPomPom (talk) 05:04, 17 October 2022 (UTC)

Thank you so much for your detailed comments! It's bedtime for me now but I'll take a look in more detail tomorrow. I will need to sort out what can be automated and how to do it. As for the hiatuses/rising diphthongs, it's always possible to generate two outputs (I do that in several cases currently, as described in the documentation for Template:pt-IPA), so maybe it makes the most sense to generate both by default, maybe labeling the hiatus variant as "careful pronunciation" or "slow speech" or something and the glide variant as "fast speech" or something. These can always be overridden in specific cases as described in the "Vowels in hiatus" section of the Template:pt-IPA documentation. Benwing2 (talk) 05:58, 17 October 2022 (UTC)

One more question ... about words in -io: rio, fio, escorregadio, arredio, mulherio, sadio, vadio, etc. These are often given with Brazilian pronunciation /-iw/, but the module currently generates two syllables: /-i.u/. What is correct? Does it differ from word to word, and if so what do you think the default should be? Benwing2 (talk) 05:42, 19 October 2022 (UTC)

Free variation between [iʊ̯], [i.ʊ], [i.jʊ], [iɪ̯.jʊ], [iɪ̯.ʏ], [i.ʏ], [iʏ̯]. Sulista tends more towards /iw/, carioca /i.u/, but you will probably find all of these in any dialect. I would say careful pronunciation /i.u/ [i.(j)ʊ], fast speech /iw/ [iʊ̯]. But in truth it's not even a standard vs nonstandard thing, yeísmo of -ilho will also produce all of those. Srtª PiriLimPomPom (talk) 13:09, 20 October 2022 (UTC)

Thanks very much! In general, to try to reduce the multitude of possible pronunciations, I want to display only what educated speakers produce when speaking relatively carefully (but still naturally), and only for certain representative dialects. I gather there is somewhat of a cross-regional standard that newscasters produce (e.g. which avoids tu entirely, but still might use te, and in phonology presumably uses some variant of [h] or [χ] for coda-final 'r' in place of e.g. [ɾ] or [ɻ]); I would like to record that, along with the Rio and São Paulo accents, plus maybe some representative Sulista accent since there are so many of these manually added. But this is really more than enough information; I'm concerned about information overload. I also would like to avoid the phonetic/bracketed version having so much detail that it becomes unreadable to all but a specialist in phonetics. So I think your careful pronunciation /i.u/ [i.(j)ʊ], fast speech /iw/ [iʊ̯] sounds like a good compromise. I will implement this in the next day or so. Benwing2 (talk) 03:53, 21 October 2022 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ I implemented handling of hiatuses involving i to have either /i./ or /j/ and u to have either /u./ or /w/. Phonetic representation of the hiatus variant is [ɪ] and [ʊ] respectively. I didn't make hiatuses involving e and o behave similarly by default because I imagine there are words where an e in hiatus is only /e/, and likewise for o, but you can get the e/i/y and o/u/w behavior using e^ and o^ respectively. In these cases, the e^ is converted to either e or i internally, and then the i in hiatus becomes either i. or y so you get three outputs; similarly for o^. Let me know if this makes sense. Benwing2 (talk) 00:56, 23 October 2022 (UTC)

Another question about stressed /ẽ/: I was under the impression that pronouncing this as a diphthong /ẽj̃/ in Brazil is restricted to word-final positions, optionally followed by an -s, hence in garagem, garagens, bem, além, tu tens etc. but not in lento. So this is how the module generates its pronunciation. But I see manual pronunciation sometimes written with the diphthong in other positions e.g. epêntese is written |/eˈpẽ.te.zi/|[eˈpẽj̃.te.zi]. Is this standard, regional, ...? Benwing2 (talk) 17:11, 23 October 2022 (UTC)

BTW also in unstressed positions, hence epentético written /e.pẽˈtɛ.t͡ʃi.ku/|[e.pẽj̃ˈtɛ.t͡ʃi.ku]. Benwing2 (talk) 17:12, 23 October 2022 (UTC)

Also pronunciations like /pe.ˈt͡ʃi(j)ʃ.ta/ for petista, identified in this case as "Rio de Janeiro"; the corresponding manual pronun for "São Paulo" is /pe.ˈt͡ʃis.ta/. Can you explain the epenthetic (j) here? I thought this only occurred word-finally in words like faz and Jesus (and not in any case after /i/). Thanks! Benwing2 (talk) 18:01, 23 October 2022 (UTC)