Help talk:IPA/Portuguese

missing symbol

The i with tilde (ĩ) is missing under vowels. I came here looking for it and thus do not know how it should sound, or I would fill it in myself. 174.19.239.54 (talk) —Preceding undated comment added 11:17, 21 December 2016 (UTC)

It's in the "diacritics" section. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:56, 21 December 2016 (UTC)

Move discussion in progress

There is a move discussion in progress on Help talk:IPA which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 16:18, 15 July 2017 (UTC)

ɐ sound

The character ɐ is listed as having a sound approximately similar to the 'a' in 'father' or the 'ur' in 'purse.' I'm not sure in which geographic area these two sounds are similar, but in US English they are very distinct and I've also not heard a similarity in British English, either. — Preceding unsigned comment added by 174.46.221.17 (talk) 14:18, 11 April 2018 (UTC)

They're more similar in British English, where the purse vowel. But you are right that this isn't helpful to American speakers. According to the vowel charts at Portuguese phonology#Vowels, I'm thinking cup and commA would be more accurate, the latter especially because this is an unstressed vowel. What do people think? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:25, 11 April 2018 (UTC)

In unstressed vowels section, the English example is actually about. The anon user is probably referring to the row with the words andaime and falámos/falamos (past tense of first person in plural of the verb falar), which is an open a in most European dialects, while it is closed and generally nasalized in Brazil (as virtually all stressed pre-nasal a, see Fonética e Fonologia do Português Brasileiro. p. 37).

As for the samples, maybe cup could be more helpful to American speakers.--Luizdl ^Talk 01:11, 12 April 2018 (UTC)

I've had the same doubt as the original commenter ("'father' and 'purse' don't sound similar"); I'm native to Brazilian Portuguese, fluent in American English and fairly used to British English (but no experience with European Portuguese). I agree that cup and commA are better examples of stressed and unstressed /ɐ/. Monstrim (talk) 00:42, 14 June 2018 (UTC)

ʁ and ɾ

I don't understand how these symbols are compared to the US latter. What am I missing? — Preceding unsigned comment added by 2605:A000:BFC0:21:1432:4CB9:846A:C1B8 (talk) 02:10, 26 August 2018 (UTC)

[ɾ] is the sound of latter in the United States. [ʁ] is a guttural r. Is that not clear in the table? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 04:35, 26 August 2018 (UTC)

The letters 'tt' are placed in bold in this example, implying that they are significant. It is the 'r,' of course, that should be the focal point. 2605:A000:BFC0:21:1432:4CB9:846A:C1B8 (talk) 01:24, 7 September 2018 (UTC)

No, it is actually the tt in North American latter that is pronounced [ɾ]. The er part is pronounced [ɚ]. — Eru·tuon 04:06, 7 September 2018 (UTC)

I'm sorry. I hear no connection between por and the 'tt' of latter (ˈlæt̬.əɹ.) There are some, perhaps many, Americans who pronounce 'latter' like 'ladder,' but how does this mean anything to our discussion about the 'r' sound of por? Are there other examples I might see and hear? Many thanks for your help with this. Incidentally, I am a native English speaker, raised in the Northeast United States, and reasonably proficient in Brazilian Portuguese.2605:A000:BFC0:21:1432:4CB9:846A:C1B8 (talk) 18:59, 7 September 2018 (UTC)

There are audio files of example words available at Dental and alveolar flaps. Most salient to this issue are the audio files for English better and Spanish caro. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:49, 7 September 2018 (UTC)

I think I can hear this in the Portuguese word 'prato.' Thank you. I'll keep studying. 2605:A000:BFC0:21:1432:4CB9:846A:C1B8 (talk) 23:35, 11 September 2018 (UTC)

Maybe your confusion is because por actually has a guttural r ([ʁ] or the like) in the dialect of Brazilian Portuguese that you are familiar with, while prato and caro have the flap [ɾ] in all dialects; see the Wiktionary entries for por and prato. — Eru·tuon 23:47, 11 September 2018 (UTC)

ð in European Portuguese

I have never heard any Portuguese person pronounce the d like a ð. Definitely not in words like cedo. Is there any source for this? — Preceding unsigned comment added by Cuddleiro (talk • contribs) 06:16, 3 February 2019 (UTC)

@Cuddleiro: There are sources listed in Portuguese phonology#Consonants, under "Phonetic notes" after the table. — Eru·tuon 12:53, 3 February 2019 (UTC)

@Cuddleiro and Erutuon: I'm pretty sure that [ð] in European Portuguese is an approximant as in Spanish and Catalan, not a fricative as in English (though word-initial /ð/ in English might be an approximant as well).

The reason for calling this sound a fricative is that the approximant allophone of /ɡ/ isn't a consonantal variety of the close back unrounded vowel but a frictionless [ɣ] which is unspecified for rounding - see voiced velar approximant for more information. This is the sole reason for calling those fricatives as the distinction is meaningless for bilabial and dental approximants (though perhaps not for the former if you consider the bilabial approximant to be the consonantal variety of the close central rounded vowel). Kbb2 (ex. Mr KEBAB) (talk) 13:43, 5 February 2019 (UTC)

Iberian [β, ð, ɣ] are on a continuum, ranging from a (weak) fricative to an approximant. They can thus be classified as lenis fricatives, if we want to borrow that term from Germanic phonologies. Human speech is not an IPA chart, there's some free variation in it. Sol505000 (talk) 11:15, 24 January 2024 (UTC)

L sounds

The sounds of L are currently wrong! I tried to fix them, by adding the velarized allophone for European Portuguese, but user:Kbb2 reverted it and made a comment implying that I added a clear L. I'm not sure whether he didn't see what I edited or he is misunderstanding something. --Mahmudmasri (talk) 20:28, 24 April 2019 (UTC)

@Mahmudmasri: Maybe the phrasing of my edit summary was somewhat unfortunate.

[ɫ] is the only allophone of /l/ in EP. There are no neutral nor palatalized varieties of it. I believe that the same is true of BP, as far as the lateral allophones are concerned ([w] is labio-velar). Our readers can read all instances of ⟨l⟩ with ⟨ˠ⟩ implied. Kbb2 (ex. Mr KEBAB) (talk) 20:53, 24 April 2019 (UTC)

Accordingly, this only means that the normal L should not be used at all for European Portuguese. The page is currently confusing for our readers, since it lacks the velarized L in the table, yet, I find it used habitually all over Wikipedia, like there before you removed it. The velarized L is normally written with the precomposed ɫ rather than adding a diacritic to a normal L. --Mahmudmasri (talk) 23:26, 24 April 2019 (UTC)

@Mahmudmasri: ⟨l⟩ does not denote a normal L (however you'd describe that, there are at least 3 varieties of what you could call a normal L) but whatever coronal lateral approximant happens to occur in any given language. It can be dental, denti-alveolar, alveolar or postalveolar and apical or laminal, it can also be palatalized, neutral (perhaps schwa-colored would be another term for that), velarized or pharyngealized or variable between any of those. As long as there are no two or (rarely) more contrastive coronal laterals in the language, there's no problem with writing the sole coronal lateral with ⟨l⟩.

⟨ɫ⟩ is a precomposed symbol, that's true. It's a combination of ⟨l⟩ and an obsolete diacritic that could be used to denote velarization or pharyngealization, therefore it's equivalent to ⟨lˠ⟩ or ⟨lˤ⟩. ⟨ˠ⟩ or (in the case of languages other than Portuguese) ⟨ˤ⟩ can be implied when you write ⟨l⟩ the same way the dental and laminal diacritics are (Portuguese /l/ is, AFAIK, laminal denti-alveolar).

On Help:IPA/Catalan, Help:IPA/Dutch and some other pages we use ⟨l⟩ for a sound that is either [ɫ] or an [ɫ] that is in a free variation (more or less) with a clear [l]. As hundreds of milions of native speakers of English (from Canada, United States, Scotland, Northern England, Australia and New Zealand) use [ɫ] for /l/ in all positions, I don't think that we need to complicate our transcriptions with this additional detail.

Perhaps @Aeusoes1: will explain it better to you. Kbb2 (ex. Mr KEBAB) (talk) 15:37, 25 April 2019 (UTC)

It makes more sense to avoid special diacritics like the tilde on velarized/pharyngealized l when there's no meaningful contrast in a language between clear and dark versions. Doubly so if there's no allophony between velarized and non-velarized (as is the case in English). It seems like l is velarized in all positions in Portuguese, so if there are transcriptions of Portuguese with dark l, we should fix that. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:48, 25 April 2019 (UTC)

Why is that discussion getting that long? Shouldn't it have been simpler for others to understand what we are talking about? Thanks for your explanations, but I really didn't need any of them. It felt like you are inviting someone who agrees with you to weigh upon your opinion. I only had one point, which is: using the correct, common symbol for the pronunciation, because it's not English, so a diaphonemic notation is unacceptable. And again, based on your claim, European Portuguese speakers only have one pronunciation for L [ɫ], not even in free variation with [l], which only makes ɫ the right choice. --Mahmudmasri (talk) 19:44, 25 April 2019 (UTC)

The transcription of Portuguese is by-and-large diaphonemic whenever we can make it. It's the same with Spanish. When we can't make it dialect neutral, we pick the dialect that is more strongly associated with the topic.

Are you also of the opinion that we should transcribe a language's single lateral as [ɫ] whenever it's more phonetically accurate? We would need some consensus on that before we changed the transcription of Catalan, Dutch, etc. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 21:02, 25 April 2019 (UTC)

The transcription convention in European Portuguese is [ɫ] in coda and [l] in onsets. That's it, I am a native speaker and a professional linguist, I can cite as many references as you wish. A major difference between European and Brazilian Portuguese is the coda lateral. [l] (or clear L) is never the sound that occurs in coda in EP. This transcription convention is also the same as in English and Catalan. @Kbb2 please change it. — Preceding unsigned comment added by Roy Aleksis (talk • contribs) 15:45, 11 April 2020 (UTC)

You say that "the convention in European Portuguese is [ɫ] in coda and [l] in onsets". But in a simple Google search I've found a site that transcribes European Portuguese using the dark L for syllable onset https://european-portuguese.info/pt/ipa.

Syllable onset /l/ is also dark in both European and Brazilian Portuguese, the European coda /l/ is only darker than the onset one. We're using /l/ for broad transcription and it is a simple symbol, we're doing the same for English and Catalan.--Luizdl ^Talk 19:13, 11 April 2020 (UTC)

I'm very late to this, just came across it as a curiosity - but as a native European Portuguese speaker, I can tell you it's almost impossible for me to start a word with the dark L. It feels horribly unnatural and requires focus so I go against my instinct. I wouldn't put it past some regions sticking to a dark L throughout as "regional accent", but "[ɫ] in coda and [l] in onsets", on a syllabic sense, is absolutely the most commonly accepted way.

"Lado" is absolutely a clear L. "Lordelo", two clear Ls. "Punhal", dark L. "Desfalque", dark L. "Lateral", clear L then dark L.

I see the table now reflects this, which I agree with. 81.109.71.137 (talk) 21:46, 7 November 2022 (UTC)

I have reverted the addition. In other guides (such as Help:IPA/Turkish and Help:IPA/Norwegian), transcribe the velarized-unvelarized distinction only when it is phonemic. In other cases, we don't bother with any special symbols and use ⟨l⟩, as in Help:IPA/Catalan and Help:IPA/Serbo-Croatian. Per Portuguese phonology, The consonant /l/ is velarized [ɫ] in all positions in European Portuguese, even before front vowels. In Portugal, the unvelarized lateral appears only in non-standard dialects. This is sourced to Fonética do português europeu: descrição e transcrição, ISBN 978-972-665-614-2. We follow the sources, not the opinions of native speakers which may or may not know whether they velarize their laterals in syllable onset. Sol505000 (talk) 19:45, 15 December 2022 (UTC)

j̃

This was added as the Brazilian variant of the palatal nasal, which Portuguese phonology backs up, but we should probably discuss it first. What do people think? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 22:52, 2 February 2021 (UTC)

Barbosa & Albano 2004, p. 228 says "The palatal nasal very frequently reduces to a nasalised palatal approximant" then it's just an allophone.--Luizdl ^Talk 00:20, 3 February 2021 (UTC)

Well, an allophone would have a phonetic context tied to it. But if even if the palatal approximant variant is what appears "very frequently" it's the same phoneme with a different phonetic attribute, which we have tended to gloss over here (especially with the rhotics). — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:38, 3 February 2021 (UTC)

I've removed all instances of that symbol, replacing it with either ⟨ɲ⟩ or ⟨j⟩, depending on the context. Editors who transcribe Portuguese into IPA have long disregarded this guide (not through malevolence but simple ignorance), so it was time to fix this and much more (e.g. replace other symbols not found in the guide such as ⟨h⟩ or ⟨ɫ⟩, remove redundant diacritics and non-standard pronunciations). Sol505000 (talk) 21:04, 31 March 2023 (UTC)

Several issues

Let me begin by saying that I know I'm fighting an uphill battle when I manifest all my objections. The first of which is that I think that, even though most portuguese language linguists keep using these symbols (w, j) when describing the portuguese language, they are incorrect because Portuguese doesn't have semivowels. I understand the confusion, it took me a while to realise the difference between the two, so allow me to explain: semivowels (/w/, /j/) simply behave as consonants, for instance, in dialects of english that are non-rhotic, that is to say, they don't pronounce the sound 'r' if it's at the end of a syllable, or in other words, before a consonant or at the end of a word. So, for instance , in the words 'Darwin' and 'are you', the 'r' isn't pronounced, proving that in those words the sounds spelled 'w' and 'y' are indeed consonants. But let's see for instance the portuguese word 'iate', which is currently transcribed in wiktionary as /ˈja.tɨ/; if that first sound was indeed a consonant, then the plural 'os iates' would be pronounced /uʒˈja.tɨʃ/, which isn't the case (except for dialects that pronounce final 's, z' as [ʒ] before vowels, which is irrelevant for my argument), instead, it is pronounced /uzˈja.tɨʃ/, or more propperly /uzˈi̯a.tɨʃ/. So, to end this part, my point is that portuguese doesn't have semivowels, only asyllabic vowels, and all instances of /w/ and /j/ should be replaced with /i̯/ and /u̯/, or, in the case of nasal diphthongs, with /ĩ̯/ and /ũ̯/.

My next complain is the insistance in usin the symbols [β], [ð], and [ɣ], wich are dialectal an allophonic and shouldn't be usen in phonemic transcription.

Regarding the rolled 'r', I feel like I'm going insane every time I read someone saying that it's a 'rural' feature; that's absurd, I'm from a rural background and I use a guttural 'r', on the other hand, many people from Lisbon still use a rolled 'r' (including our fucking president if I'm not mistakened), so I have no idea why people insist with this argument; in my view, using a rolled or a guttural 'r' is mainly a generational feature; younger people tend to use the guttural one while older people the rolled one, but it's not in any way an indiocator of 'rurality'.

I think that's all for now; I'm sure I have a lot more complaints but that's for another post. Also, is there a way for the pt IPA module to output phonemic notation (/.../) instead of phonetic ([...])? The ideal would be for it to output phonemic notation followed by optional phonetic notation. Sérgio R R Santos (talk) 17:09, 16 August 2024 (UTC)

Approximants are non-syllabic vocoids. Cross-linguistically/purely phonetically, there is no difference between ⟨j, w⟩ and ⟨i̯, u̯⟩. The latter are sometimes used e.g. to signal connection to /i, u/ in the same language or when /j, w/ exist separately, but there's no definitional difference, and all differences are conventional. So we should use whichever are more common in reliable sources.

None of our IPA keys except the one for English are phonemic, because this is English Wikipedia and we can't expect our readers to be familiar with the phonological system of any language besides English. Nardog (talk) 00:14, 17 August 2024 (UTC)

"Approximants are non-syllabic vocoids. Cross-linguistically/purely phonetically, there is no difference between ⟨j, w⟩ and ⟨i̯, u̯⟩. The latter are sometimes used e.g. to signal connection to /i, u/ in the same language or when /j, w/ exist separately, but there's no definitional difference"

You apparently decided to ignore my english example that proves that the two sounds are phonemicaly distinctive, one of the most basic distictions that there is in linguistics/phonetics, that is, between a consonant and a vowel. If there's really no distinction, why are there different IPA symbols for them; surely most linguists would be calling for the abolition of the redundant symbols, no?

"None of our IPA keys except the one for English are phonemic, because this is English Wikipedia and we can't expect our readers to be familiar with the phonological system of any language besides English."

What? That makes no sense whatsoever; that's exactly the reason transcriptions should be mostly phonemic, because they provide the most simple description of a language; that exception for english makes absolutely no sense for me. Sérgio R R Santos (talk) 17:56, 17 August 2024 (UTC)

Hello @Nardog. Let me just start by saying that a user recently admonished me for my tone in some of my posts and, while at first i disagreed and din't think i was in the wrong, as i re-read my response to you i did indeed realise that i was too brash, and honestly, a bit of an arsehole (and maybe that's why you haven't responded to me?), so i'd like to apologise to you for the tone I used.

That being said, I mantain my arguments, let me give some examples. For instance, the portuguese southern town of Vidigueira is currently given the pronunciation of [viðiˈɣɐjɾɐ] (with the "locally [vidiˈɡeɾɐ]" added by me), however, [viðiˈɣɐjɾɐ] isn't really the European Portuguese pronunciation, it's the Lisbon pronunciation. Would't it be much better to give the more neutral phonemic pronuntiation as /vidiˈɡei̯ɾɐ/ (or if you insist with the semivowels, /vidiˈɡejɾɐ/), followed optionally with the local pronunciation in phonetic notation, in this example [vidiˈɡeɾɐ]? Using the allophonic symbols [β], [ð], [ɣ] - who are dialectal and not universal in european portuguese - in giving the names of places where the people don't even pronounce them as such, doesn't make much sense to me. Conversely, most portuguese people pronounce the L as [ɫ] instead of [l], but we don't feel the need to indicate that in a broad transcription, because its just an allophone, so following the same logic the fricative allophones of /b/ /d/ /g/ (which is actually a feature less widespread than the [ɫ]) should also not be indicated, unless when using a more phonetic transcription to indicate the local pronunciation. To better illustrate my positions let me give you some examples of what it would be like:

In the North:

Braga /bɾagɐ/ - [bɾaɣɐ]

Miranda do Douro /mirɐ̃dɐ du dou̯ru/ - [mirɐ̃dɐ ðu ðou̯ru]

Viseu /vizeu̯/ - [viz̺eu̯] or [biz̺eu̯]

Gouveia /ɡou̯ˈvei̯ɐ/ - [ɡou̯ˈvei̯ɐ]? (this one i'm not sure if its north or center)

In the Centre: (by that I mean Lisbon area)

Penha de França /ˈpeɲɐ dɨ ˈfɾɐ̃sɐ/ - [ˈpɐɲɐ ðɨ ˈfɾɐ̃sɐ]

Loures /ˈlou̯ɾɨʃ/ - [ˈloɾɨʃ]

Oeiras /ɔˈei̯ɾɐʃ/ - [ɔˈɐi̯ɾɐʃ]

Torres Vedras /ˈtorɨʒ ˈvɛdɾɐʃ/ - [ˈtoʁɨʒ ˈvɛðɾɐʃ] (or [ˈtorɨʒ ˈvɛðɾɐʃ]; I don't understand the insistence in classifying the rolled r as a "rural" feature - there's no evidence for that)

Sobral de Monte Agraço /suˈbɾal dɨ ˈmõtɨ ɐˈgɾasu/ - [suˈβɾal dɨ ˈmõt(ɨ) ɐˈɣɾasu] or [suˈβɾaɫ-

Gouveia /ɡou̯ˈvei̯ɐ/ - [ɡoˈvɐi̯ɐ]

In the South:

Alcácer do Sal /alˈkasɛɾ du ˈsal/ - [alˈkasɛɾ du ˈsal] or [aɫˈkasɛɾ dʷ ˈsaɫ]

Santiago do Cacém /sɐ̃tiˈagu du kɐˈsẽĩ̯]/ - [sɐ̃tiˈagu du kɐˈsẽĩ̯] or [sɐ̃ˈti̯agʷ dʷ kɐˈsẽĩ̯] ( [sɐ̃tiˈaɣu ðu kɐˈsɐ̃j] sounds like an auditory crime to my southern ears)

Ferreira do Alentejo /fɨˈrei̯ɾɐ du ɐlẽˈtɛʒu]/ - [fɨˈreɾɐ du ɐlẽˈtɛʒu] or -ɐɫẽˈtɛʒu]

Melides /mɨˈlidɨʃ/ - [mɨˈlidɨʃ] (I gave myself the liberty to add the local pronunciation of [mɫidʒ̊], since it's where i'm from; that could be even another further level of phonetic transcription, but lets complicate things for now)

I hope i didn't add so much information to this post that would dicourage people from reading it, but tell me if this does't look like a much more elegant system than the one currently being used - which again, is giving pronunciations from all over the country the way people from Lisbon pronounce it, because they're using phonetic transcription instead of the more preferable and neutral phonemic transcription, which is more appropriate for a pluricentric and multi-dialectal language like Portuguese. But that's just my opinion. Sérgio R R Santos (talk) 22:53, 21 August 2024 (UTC)

Woops, I forgot to ask if you're a native speaker of Portuguese, it doesn't seem so based on your user page, in which case most of this will be very confusing to you! Anyway, I'm posting this for anyone who can answer my inquiries. Sérgio R R Santos (talk) 22:59, 21 August 2024 (UTC)

My two cents on the issue. I am favorable to amending the help page with notes specifying that, while for simplicity we present two main standards in table, the same symbols, when possible, may be used to provide more accurate local pronunciations – in other words, instead of having a double phonemic vs phonetic transcription, a broad phonetic transcription should suffice to our readers (let’s remember this is not Wiktionary); e.g. [fɨˈʁeɾɐ du ɐlẽˈtɛʒu] instead of the current [fɨˈʁɐjɾɐ ðu ɐlẽˈtɛʒu] or the proposed [fɨˈreɾɐ du ɐlẽˈtɛʒu] because [e] and [ð] are listed in the tables, but [r] is not and is one of a long list of realizations that exist throughout the Lusophone area for the same phoneme, all of which we cannot list: an expansion to the related note on the key should be enough in my opinion.

As for the choice between ⟨j w⟩ and ⟨i̯ u̯⟩, let’s clarify that there is actually no phonemic distinction in Portuguese between the two pairs: /ˈaɡwɐ/ cannot possibly contrast with /ˈaɡu̯ɐ/, for example, contrarily to /ˈkaru/ vs /ˈkaɾu/, so it all comes down to which symbols we choose to transcribe the same phonemes. Personally, I have a preference for ⟨j w⟩ since they are the most common choice in the other IPA keys, especially for Romance languages. ~ IvanScrooge98 (talk) 14:49, 26 August 2024 (UTC)

"a broad phonetic transcription" - did you mean "broad phonemic transcription"? Because the term "broad transcription" is usually associated with a less specific phonemic transcription, as opposed to the much more specific phonetic transcription, which is precisely the opposit of "broad". I myself when writing a post, have sometimes written phonetic for phonemic, and vice-versa, and only notice the error when re-reading before posting. They should've come up with more different terms!

Regarding that specific example, the simplest form to indicate its pronunciation would be /fɨˈrei̯ɾɐ du ɐlẽˈtɛʒu/; all the diphthongs /ou/, /ai/, /ei/, should always be given as such in phonemic notation (with the slashes //), their specific pronounciations according do dialects should be left to an (optional) phonetic notation (with the square parenteses []).

Regarding the tables in this page, do they look straighforward and simple to understand to you? Because I find them extremely confusing, and I'm a native speaker, so I can only imagine how a non-native speaker with some curiosity about Portuguese phonology might find themselves discouraged and confused when looking at that chart.

The main reason for the table looking so confusing is the addition of several sounds that are mere allophones, and are not at all necessary for a brief description of portuguese phonology - the inclusion of the sounds [β], [ð], and [ɣ] (who are, mind you, not even a universal feature of European Portuguese dialects), indicating L-vocalization and the palatization of /d/ /t/ before [i] in brazilian portuguese; and that's just reffering to the consonants. I'm not saying that those allophonic pronunciations shouldn't be mentioned at all - they should, but only in the notes section, and not in the main table. Regarding the rolled R, just a few hours ago I accidentally turned my TV to a portuguese channel, which I endured for a few minutes for the sole reason that the president of Portugal was speaking, just so that I could confirm that he indeed speaks with a rolled R; meanwhile everytime I read anything about that it's constantly stated that the rolled R is relegated to a few rural speakers, whith is objectively false - I already vented my grievancies about that here in the final paragraphs of that post, if you want to be bothered to read it. I just think the strong R should be rendered as /r/, for simplicity and historical reasons; however if people wanna insist on keep using /ʁ/, the rolled pronuntiation should at least be included in the table - it's still relativelly common in Portugal, and even more in Angola - even though I have heard younger angolan people using a guttural r, so probably in 50 years from now no one will use it.

Also, and this is a very important point, what is usually refered to as "European Portuguese", when you look at the phonetic description of it, is actually "Lisbon Portuguese", with most linguists just assuming the Lisbon dialect to be the "standard" and taking little to no consideration for the other dialects, wich to me reveals a high degree of incompetence or simply lack of linguistic curiosity; for instance, as opposed to english, where you can find endless Wikipedia pages with detailed descriptions of english dialects, you wont find much on portuguese dialects. Just one example is the page for my dialect, Alentejano, which describes just half a dozen features, half of them actually wrong!

"As for the choice between ⟨j w⟩ and ⟨i̯ u̯⟩, let’s clarify that there is actually no phonemic distinction in Portuguese between the two pairs"

Let me try to explain this in the simplest way possible, hopefully without sounding too condescending: it's not a matter of Portuguese not distinguishing those pair of sounds, it's a matter of one of those pairs ⟨j w⟩ not existing in portuguese. The IPA symbols /j/ an /w/ represent consonants, and not vowels. It doesn't matter if they are pronounced exactly like the vowels [i̯] [u̯], it's a question of how they behave phonemicaly in the languages that have them. For example, the Finish language has both [j] as a consonant (spelled "j") and the assyllabic vowel [i̯] (spelled "i" after another vowel).

"Personally, I have a preference for ⟨j w⟩ since they are the most common choice in the other IPA keys, especially for Romance languages."

And that, my friend, is the main source for this confusion: most Romance languages don't have semivowels, having lost them from Latin, and speakers from languages lacking semivowels have difficult undestanding the difference between them and vowels - it took me many years until I finally understood, too.

Just one last note against using /j/ and /w/, especially to indicate falling diphthongs, is the wiktionary definition of a diphthong:

A complex vowel sound that begins with the sound of one vowel and ends with the sound of another vowel, in the same syllable

"A vowel followed by another vowel"; that's why it is, in my view, incorrect to represent, for instance, the portuguese word "pai" as /paj/, rather than /pai̯/ - [ai̯] is indeed a vowel followed by another vowel, per the definition, while [aj] is a vowel followed by a consonant. I know that having to use all these symbols with all these extra diacritics might feel a bit annoying and unconvenient for some people, but, as I've said before in some other post, what was the point of inventing an entire phonetic alphabet with all of it's specificities, to end up not using it propperly?

Wow, I didn't intend for this reply to get this big, maybe that's why no one listens to me on this website! Sérgio R R Santos (talk) 19:00, 26 August 2024 (UTC)

Yeah, I got a bit lost in your reply, sorry 😅 but I think I got your point. The thing is, if want to change this to a diaphonemic key like the English one, we would have to rework it entirely and amend all the transcriptions pointing here. Not that I’m against it, but we need to build a strong consensus before we take such a step. What I meant by “broad phonetic” is my proposal of a phonetic transcription that simplifies the most variable realizations using broader symbols, which is what this key mostly does already (let me remind you there are different levels of phonetic transcription in terms of narrowness); and also a good reason to choose a simpler ⟨j w⟩ symbol over ⟨i̯ u̯⟩—especially because there normally is no phonetic difference: I don’t understand your insistence on this point, there’s a reason why they are called semi-vowels after all. ~ IvanScrooge98 (talk) 19:21, 26 August 2024 (UTC)

Well, you can try to read it a few more times, although, from my own experience, it probably won't help! - I too have had some difficulty to fully grasp what a replier meant, especially when they're too verbose, which is often my case, unfortunately.

Regarding my issue with the semivowels, I honestly don't think that I can explain it more clearly and profoundly than all the ways I already did, with all examples that I provided. But I may try to provide - probably to no avail - one last example. You're Italian, right? The Italian word "uomo" is usually transcribed as /ˈwɔmo/, starting with a consonant: however - and correct me if i'm wrong - if I remember correctly the Italian definite article li becomes l' before vowels, which includes the word uomo; that seems like an exception if you analyse the word uomo as starting with a consonant, but, if you analise it as /ˈu̯ɔmo/ there is no exception to be accounted for; in other words li becoming l' before uomo to me proves that the word begins with a vowel, and not a consonant.

That's what I tried to explain to you in my previous post; the diference between semivowels and assylabic vowels is not necessarily in any difference in pronunciation, but in how they behave phonemically. For instance, if portuguese diphthongs were really composed of a vowell followed by a semivowel (functionally a consonant), then howcome these vowels don't suffer reduction when they're unstressed? That's exactly what happens with Russian й. Take the word for war, война́ (vojná), which is pronounced vajná, following the general rule of untressed o's being pronounced as a's, or the more transparent Belarusian вайна́ (vajná), which stupidly indicates the variation in the spelling. If that were the case in portuguese then unstressed /oj/ and /ɔj/ should become *[uj], since unstressed "o" is pronounced [u] in portuguese.

Anyway, I hope I made things a bit more clear instead of making you even more confused! Sérgio R R Santos (talk) 21:50, 26 August 2024 (UTC)

Ok, I got it. But still there is a reason why they are called semivowels, i.e. because they share features of both consonants and vowels and depending on the case they might “behave” like one or the other. The choice of the symbols ⟨j w⟩ does not necessarily mean we want to tell the reader “they are consonants”, because they are symbols and as such they merely represent a convention. ~ IvanScrooge98 (talk) 21:57, 26 August 2024 (UTC)

Well, there's also a reason semivowels are represented in IPA's consonant chart. Fun fact, I like reading ancient Spanish and Portuguese grammars from the 16/17th century, and they usually reffer to other Latin grammarians and their works, and apparently what the Romans reffered to as semivowels were the sounds l, m, n, r, s, z, completely different from the modern meaning. Sérgio R R Santos (talk) 22:22, 26 August 2024 (UTC)

That’s interesting! Seems a category akin to our sonorants/continuants. I guess what made those consonants similar to vowels for them is that their articulation permits to protract their sound indefinitely, unlike stops. But let’s stick to the modern concept :P ~ IvanScrooge98 (talk) 22:49, 26 August 2024 (UTC)

As for ⟨ʁ⟩, I find it a pretty fair compromise to represent a range that goes from alveolar [r] to glottal [ɦ]. ~ IvanScrooge98 (talk) 19:26, 26 August 2024 (UTC)