Talk:Portuguese phonology/Archive 1

Introduction

The introduction needs some work. The criticism of foreign misinterpretation of the vowel distinction needs to be toned down quite a bit. Peter Isotalo 11:51, Mar 18, 2005 (UTC)

Not only that. Phonology, allophones and dialectal differences are all mish-mashed into one big issue of "having many vowels". --Pablo D. Flores 12:50, 2 May 2005 (UTC)

Rodizio

Hi, I was working the article rodizio and I was thinking it might be nice to put it in IPA. Unfortunately, I don't speak Portuguese. The article says the pronunciation is "ro-DEE-zhyoo". This seems to conflict with the note in the orthography section here that "z" between vowels is pronounced [z]. Any advice? Lesgles (talk) 05:41, 26 February 2006 (UTC)

Leave it to me... Ciacchi 16:45, 17 March 2006 (UTC)

Thanks! Lesgles (talk) 04:04, 24 March 2006 (UTC)

Edits to the paragraphs on diphthongs

I've made some considerable changes to the paragraphs on oral and nasal diphthongs:

I deleted all the increasing diphthongs because I don't think they're usually counted as diphthongs per se, and because increasing diphthongs in Portuguese can be treated as hiatuses.

I deleted the statement "[ej] is only pronounced as [ɐj] in Lisbon. In the rest of Portugal, it is pronounced like common [ej]", because it's not accurate. This pronunciation is quite common in Portugal, nowadays.

I moved up and expanded a note on the pronunciation of the digraph ou which was at the bottom of the page. 30 Mar. 2006.

â ân

i've never seen a circunflex accent on a. only on e and o. can anyone give examples? --itaj 11:26, 26 May 2006 (UTC)

Câmara, gândara, tâmara, trâmite, Cândido, Tâmega. Joaopais 16:45, 26 May 2006 (UTC)

is it like that in BP too or just EP? --itaj 14:57, 27 May 2006 (UTC)

The examples that Joaopais gave are written the same way in BP and EP. Are you sure you haven't seen the word câmara before? FilipeS 17:58, 27 May 2006 (UTC)

P.S. Or how about Bethânia? :) FilipeS

OK, thanks for the answers. well i have been in brasil six months and i speak portuguese quite fluently for day to day conversations, and pretty much i can write (correctly) everything i can say. but i've really never seen that before. every day is good for something new...

i only once heard the name tâmara of the fruit, but never seen it written, so i supposed it was támara. actually i think my brasilian friend pronounced it like á.

the other words i don't know at all. are most of them names?

is tâmara imported to portuguese? i know it's an old semitic word. is there some "original" or common vocabulary with â?

how common is â in general? --itaj 19:25, 27 May 2006 (UTC)

The word câmara (city hall) is quite common. (I'm not sure the same term is used in Brazil, though.) The others are indeed uncommon, and the two last ones that Joaopais wrote are proper nouns. For the origin of tâmara, see here. All words with orthographic â are proparoxytones, and since Portuguese words are usually stressed on their last syllable, or on the one before the last, it's not surprising that you haven't noticed them yet.

On another note, I took a peak at your website, and I think here's a little incorrection in it. I would rather say página de capoeira. ;-) Regards. FilipeS 20:17, 27 May 2006 (UTC)

thanks. and then it's not just prepatoxytone. i know many prepaxoytones in portuguese, but they use á é í ó ú ê ô - like página :) it's just â that are too rare to have crossed my way until now.

and thanks for helping with the site. i need to ask something then: i wanted to specify that it is "the Capoeira" not just "capoeira", thus i wrote "página da capoeira" does it make no sense?

Why would you want to say "the Capoeira"? Isn't the site devoted to capoeira in general? FilipeS 21:51, 27 May 2006 (UTC)

oh, got you. thanks. --itaj 11:13, 28 May 2006 (UTC)

Incidentally, aren't the stressed vowels in "página" and "câmara" two different phonemes in Brazilian Portuguese ? At least, my English/Brazilian Portuguese dictionary gives two different transcriptions for them. 161.24.19.82 21:02, 7 June 2006 (UTC)

Yes, the former is an oral vowel, and the latter is nasal. FilipeS 16:15, 12 June 2006 (UTC)

all â examples given above are nasal vowels (come before m or n). does â exist as oral vowel (non-nasal)? examples? --itaj 15:45, 20 June 2006 (UTC)

Only in European Portuguese, as far as I know. FilipeS 15:50, 20 June 2006 (UTC).

what is only in EP (as far as you know)? which of the following are true:

1. â comes only before m or n and pronounced nasal in BP but oral in EP. as in the above examples: Câmara, gândara, tâmara, trâmite, Cândido, Tâmega.
2. there are words with â that is not before m or n (thus never nasal), but there words are used only in EP. anyone has examples?

--itaj 02:05, 21 June 2006 (UTC)

Nr. 2 is not true. Nr. 1 is half right, and half wrong: â does always come before m or n, and is always pronounced nasal in BP (as far as I know). In EP, â is not nasal if the m/n is followed by a vowel, but it is nasal if the m/n is followed by a consonant. This is explained in the article, if you read closely. FilipeS 14:11, 21 June 2006 (UTC)

thanks much for your help. you mean what's written after the table in the section of nasal monophthongs? i've seen it but could not derive that â never comes without a following n or m. also because â is listed with the oral monophthongs with no special comment. --itaj 18:00, 21 June 2006 (UTC)

It's true that the article does not state that â always occurs before m or n. However, I think it does make clear that, whenever it occurs before m or n, it is nasal in BP. Given the written form of a word, this should be enough information to know how to pronounce it. On the other hand, given the spoken form of a word, you already know whether it's nasal or not. FilipeS 21:47, 22 June 2006 (UTC)

exactly that i understood. but i was curious to know about the always part because then i can think of â as the nasal allophone of á. so they're much like the regular tonic oral/nasal a which differ in pronunciation about the same way. though as writing now i realise that i actually also need to know that á never comes before m or n. --itaj 22:41, 22 June 2006 (UTC)

Fricative allophones of the plosives in EP: citation needed

I'm going to rephrase the article, and ask for a citation for the statement that the plosives [b], [k], [g] have the fricative allophones [β], [ð], and [ɣ], respectively, in European Portuguese. The reasons are as follows:

• While I have heard people pronounce words with these allophones, in my experience it is not that common, and most people do not use them consistently. They seem to be used only in some regions, or in high registers, and in songs like the fado. I would venture to say that in their everyday life even the people of Lisbon do not use these allophones consistently. I may be wrong, but if so I am not the only native speaker who has this impression. See this earlier discussion.
• Mateus and d'Andrade (see references) state that there is no difference between the pronunciation of the plosives in Portugal and Brazil. FilipeS 13:40, 27 May 2006 (UTC)

Actually, they do. Footnote 2 to page 11. FilipeS

e in BP

sometimes unstressed e is pronounced /i/ (i think). like berimbau, tesoura. in tesoura it also palatalizes t like usualy i does, for being pronounced /i/.

is that correct? are there rules to know by the spelling of the word when e pronounced like that? --itaj 10:46, 2 June 2006 (UTC)

Your observations are correct. The Wikipedia article just gives general guidelines. For more specific information regarding the phonology and other linguistic aspects of Portuguese, you should consult the literature, or specialized websites. ;-) FilipeS 21:15, 3 June 2006 (UTC)

BTw, unstressed "o" is also sometimes pronounced /u/ in Brazil like in Portugal, e.g. in words like "dormir", "cobrir", "bonita", "João", "doente", etc...

Palatalization of /t/ and /d/ before /i/

A user changed the comparison of this phenomenon of Brazilian Portuguese with the very similar one in Quebec French into an analogy with Japanese, claiming that Japanese is a better analogy. I would like to know why, before accepting the change. FilipeS 21:19, 22 June 2006 (UTC)

i think you should keep both comments even if you accept the new one. french is closer to portuguese so it makes it interesting in a different way. also someone with interest in portuguese is more likely to know some french than japanese. --itaj 22:36, 22 June 2006 (UTC)

Both comparisons are indeed interesting, although neither of them coincides completely with the palatalization seen in Brazilian Portuguese: Japanese palatalizes all consonants before /i/, not just /t/ and /d/; and while Quebec French palatalizes only the consonants /t/ and /d/, it does so not just before /i/, but also before /y/ and /u/.

In any event, there's already a reference to Japanese at Brazilian Portuguese. FilipeS 12:18, 1 August 2006 (UTC)

I reverted changes made by Ciacchi, asserting that idade, sorte, etc. are palatalized rather than affricated; this is simply wrong. It is contrary to numerous sources (e.g. Routledge's "The Romance Languages") and is contrary to what my ears tell me. Benwing 07:13, 20 August 2006 (UTC)

I strongly agree with Benwing. 13:16, 20 August 2006 (UTC)

Please note that "palatalized" and "palatal" are two separate things. FilipeS

and neither of them are correct. "palatalization" is ok as a general description of the process which produced the affricates, but not as a description of their phonological manifestation. Benwing 06:36, 24 August 2006 (UTC)

delete the section on phonotactics?

someone recently added a section on "phonotactics". i see little point in the section as-is -- there is really no purpose in all the examples, and the basic idea could be summarized in a couple of sentences, but even then i don't see what is gained by the section. a real section on the phonotactics of portuguese would not try to reduce its phonotactics down to CCVC and such but would describe the actual restrictions. i.e. yes you can say "frustrar" but not "rfusrtar", nor "ftustrar", nor (in BP) "frustrac". also, EP and BP differ -- EP allows "advogado" but BP wants "ad/i/vogado". descriptions in terms of CCVC and such only work for languages like arabic, where the phonotactics actually work based on number rather than type of consonants. Benwing 07:24, 20 August 2006 (UTC)

• I agree with you. José San Martin 13:16, 20 August 2006 (UTC)
• I agree as well. The section is not informative in its current form, and it is unnecessary. FilipeS

incorrect BP pronunciation

Discussed at Talk: Portuguese Language.

Pronunciation of "ou"

The following has recently been added to the article.

Most speakers nowadays pronounce the digraph ou as a monophthong [o] when there's no possibility of confusion, although in some regions of Brazil and northern Portugal it is still pronounced as the falling diphthong [ou̯].

It's not clear to me what this "danger of confusion" might be. There are a few minimal pairs for the diphthong versus the monophthong, true, but most people in Portugal get along fine pronouncing both the same way. Context clears any doubts. Moreover, this pronunciation "to avoid danger of confusion" seems to be only done in Brazil (perhaps out of linguistic purism). Unless this is clarified, I am going to remove the remark from the article. FilipeS 14:44, 30 December 2006 (UTC)

Placement of Diphthongs section

This is not a big deal, but just for the record the reason why I had moved the Diphthongs section down was that they are not phonemes. I thought it was neater to discuss all phonemes first, and then list the diphthongs (which have a very straightforward correspondence with spelling, anyway, so they could just as well be listed at Portuguese orthography). FilipeS 15:25, 12 February 2007 (UTC)

New reference

What is the author, title, etc., of the newly added reference? FilipeS 19:20, 22 February 2007 (UTC)

Right, sorry, I should have clarified it. It is an excerpt from The Phonology of Portuguese , Maria Helena Mateus and Ernesto d'Andrade, published by Oxford University press. See it here on Google Books. Macgreco 23:52, 26 April 2007 (UTC)

I suspected as much, but didn't have the time to check. Thanks. FilipeS 12:41, 27 April 2007 (UTC)

Occitan influence

Before this escalates anymore, and since User:AnPrionsaBeag seems to be a newcomer to Wikipedia, I should perhaps give my reasons for reverting his edits, which were the following:

Palatal lateral approximant. In some Brazilian dialects, this phoneme is realized as palatal approximant [j], but this is not as widespread as yeísmo in Spanish. It came to Portuguese through Occitan in 1047, just like nh [1]. See also this and that.

[...] See lh for origin.

This edit is nonsensical. It's true that the digraphs "lh" and "nh" were adopted from Occitan, and that fact is mentioned at the Portuguese alphabet article, but this article is about phonology, not orthography. The phonemes /ʎ/ and /ɲ/ were not "borrowed" from any language; they developed in parallel in several Romance languages. User:AnPrionsaBeag has confused phonology and orthography completely, as I have already pointed out in one of my edit summaries.

With as much respect as I can muster, I have to say that it's particularly grating to be allegedly "corrected" and accused of "vandalism" by someone who is so clearly and utterly misguided. FilipeS 20:34, 28 June 2007 (UTC)

Learning Portuguese

To user FilipeS, I am an American female college student. If I ever learn Portuguese, it will be my second language. Should I rely on Wikipedia to get started? Which version will I be learning? I want to learn both European and Brazilian Portuguese if I can so I can be most effectively understood throughout the Portuguese-speaking world. I listen to bbcbrasil.com and bbcparaafrica.com, which are both media in Portuguese. I want to be a radio journalist.

Should this article be expanded further? I wouldn't know what to say 'cause I'm not a native speaker like you. Do you speak EP or BP or both? Does it help to know both versions to be understood in the Portuguese-speaking world?

learnportuguese 20:47, 23 August 2007 (UTC)

Hello. I've replied to you in your UserTalk page. :-) FilipeS 22:23, 25 August 2007 (UTC)

[s] vs. [ʃ] dialects?

Portuguese phonology#Sibilants says:

As was mentioned above, the dialects of Portuguese can be divided into two groups, according to whether syllable-final sibilants are pronounced as alveolar /s/, /z/, or as postalveolar consonants /ʃ/, /ʒ/.

But it isn't mentioned anywhere above!--Imz 00:08, 6 October 2007 (UTC)

I think I've addressed the issue now. For future reference, comments within the article should be hidden like this <!-- hidden comment-->. Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 01:13, 6 October 2007 (UTC)

Hmm, I think that above reference was reasonable in that that this information doesn't belong to Portuguese phonology#Connected speech section, at least in a straightforward understanding of what connected speech is. So, I thought someone could give full details about this fact in a right section.

As for how to format comments: I didn't want to hide that comment. Some editorial comments are admitted in articles, like those asking for references and factual accuracy. There are special templates for them. My comment was of that kind. It brought attention of a reader to an inconsistency in the article, which could be fixed by a reader with enough knowledge. So, I wouldn't use <!-- hidden comment--> for that. Thank you for your reaction!--Imz 17:57, 23 October 2007 (UTC)

Lenition of b, d, and g in European Portuguese

Several sources state that a recent sound change in European Portuguese as lenition for b, d, and g similar to what occurs in Spanish. Yet, there is no mention of this in the article. Anyone want to chime in? Azalea pomp 01:57, 9 November 2007 (UTC)

Yes, there is. Read it more carefully. FilipeS 21:27, 11 November 2007 (UTC)

Unstressed vowels

The table at "unstressed vowels is a total mess, some things are right but some others are false (either in portuguese portuguese or in brazilian portuguese). A correction would be welcome. —Preceding unsigned comment added by 85.242.251.79 (talk) 16:24, 26 November 2007 (UTC)

Please be more specific on what you believe is incorrect. FilipeS (talk) 19:18, 28 November 2007 (UTC)

Northeastern Brazillian phonology

This article simply ignores the phonology of Northeastern Brazillians, regarding the Southern accent as "Brazillian Portuguese". They are very similar, I know, but Northeasterns don't say "pegar" with a high /e/ but with a low /ɛ/, and there is no palatalization. Considering that 28,9 % of the Brazillian population lives in the Northeast, added with the great number of Northeastern immigrants in other regions, I think Northeastern accent should be considered for this article. 189.70.55.139 (talk) 02:54, 15 December 2007 (UTC)

You did not read the article very carefully:

[...] there is some dialectal variation in the unstressed sounds: the northern accents of BP have low vowels in unstressed syllables, /ɛ, ɔ/, instead of the high vowels /e, o/.

However, as the article adds, this accent is not regarded as the standard in Brazil. FilipeS (talk) 15:09, 15 December 2007 (UTC)

Not true that there is no palatalization in the Northeast. It's universal in Salvador (and is presumably spreading northward). Benwing (talk) 07:29, 1 September 2008 (UTC)

Pronunciation of "de"

According to the article, one would think that the Rio-area pronunciation of "de" (as in "Rio de Janeiro") would be [dʒi], yet I hear it as [d̪i]. What is the correct pronunciation here, and should it be mentioned in the article? (Disclaimer: I am only just learning Portuguese!) Thanks! Grover cleveland (talk) 18:48, 20 December 2007 (UTC)

By "here" do you mean in Rio? Rio de Janeiro is a big city, it must have people from all over Brazil. It's not surprising that you hear different accents. All of them are correct. However, I hear that the most widespread pronunciation in Brazil is indeed [dʒi]. FilipeS (talk) 22:29, 21 December 2007 (UTC)

"n" at end of a syllable

The current article says that /n/ never occurs at the end of a syllable, "except in rare learned words, for some speakers". However, Thomas, Grammar of Spoken Brazilian Porguguese, p. 8 says that "in final position in the word [the written letter 'n'] may be pronounced as a consonant or it may merely nasalize the preceding vowel" (with no restriction to "rare learned words"). He also says that in participles such as "falando" the "n" is realized as [n]. Could this be a EP vs. BP thing? Grover cleveland (talk) 07:50, 13 January 2008 (UTC)

The first part of Thomas' statement is not inconsistent with the article's phrasing. The article is just being more specific. As for his second statement, that in participles (and participles alone??!) /n/ can be realized as a consonant at the end of syllables I find it very strange and highly dubious, frankly. He may be thinking of some regional pronunciations, but in any case these would not be restricted to participles, or any other word class. Phonetics is not normally conditioned by grammar -- this is phonetics 101! FilipeS (talk) 20:16, 13 January 2008 (UTC)

Here's the Thomas quote in its entirety:

At the end of a syllable, but not at the end of a word, ["n"] is not pronounced as a consonant, but merely nasalizes the preceding vowel. Canta, onze. However, it is usually pronounced in the verbal ending -ndo. Falando. p. 8

And later on, when discussing the present participle, he repeats the same point:

The present participle is regular in all verbs, and simply replaces the -r of the infinitive with -ndo. Contrary to usage in other situations, the letter n is pronounced as a consonant here. p. 17.

If I have time, I'll try to find a reference to this in other sources. Grover cleveland (talk) 05:34, 14 January 2008 (UTC)

Words beginning with "nh"

Nhoque ("gnocchi") is one example. Grover cleveland (talk) 01:48, 2 February 2008 (UTC)

About the phonemes /l/, /n/ and /ʁ/

Wait a minute, Aeusoes, before going on in an edition war, let’s talk a little.

What do you mean by “dental and velarized”?!? You have cited me a source claiming that the Brazilian /l/ is alveolar and the (European) Portuguese /l/ is dental, I respect that source, but I don’t believe it’s true! We, Africans, alltogether with Brazilians, pronounce the /l/ in Portuguese as dental, very differentely than the Portuguese that pronounce the /l/ as alveolar. Besides, dental /l/ means not velarized, while the alveolar /l/ is velarized!!!

I don’t know if it is you who made the changes about the phonemes /l/ and /n/ at the end of syllables, but that’s not quite so. The letter <n> does appear at the end of words, but it is pronounced differentely: in European Portuguese words like gérmen or cólon are pronounced ['ʒɛɾmɛn] and ['kɔɫɔn] while in Brazilian Portuguese they are pronounced ['ʒɛʁmẽj] and ['kɔlõ]. With “laterals” I think the author meant the phoneme /ʎ/ and not the phoneme /l/. It is so frequent at the end of syllables that I won’t even give examples... :-)

About the phoneme /ʁ/, I think that it is not quite so neither. By my own observations, the Africans don’t pronounce it uniformally (there is a widespread myth about something called “African Portuguese”). Cape Verdeans pronounce it like the Portuguese, i.e., [r], [ʀ], or [ʁ], in a free variation according to the speaker. I think that it also hapens with São Tomean. Some speakers of Guinea Bissau pronounce it as [ɾ] (emphasis on some, I do not have accurate data). But Angolans and Mozambicans only pronounce it as [r] (and I think that also the Timorese). That may have an substratum explanation: in Bantu languages (as far as I know) there is no [ʀ] or [ʁ] sounds. However, everything that I have said in this paragraph are due to my personal observations through many years, so do not cite me as a source, please...

Ten Islands (talk) 04:14, 8 February 2008 (UTC)

What do you mean "dental /l/ means not velarized"? Technically, they're not mutually exclusive so you can have a velarized dental lateral or a velarized alveolar lateral. As velarization states, it's actually more likely that a velarized /l/ is dental. I understand if you disagree with it, but please don't change it until you find a source.

I think the phrase "Nasal and lateral consonants do not normally occur at the end of syllables" means phonetically rather than phonemically. We can take it out, though, because that's kind of unclear.

As for /ʁ/, I didn't add any information on Africa but it seems to be sourced (Mateus & d'Andrade (2000)). Have you checked that source? — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 05:19, 8 February 2008 (UTC)

Wow! 24 hours of absence and a lot of changes! But let’s go to what matters.

After what you’ve said I remained in doubt for a few seconds, but after that (and after doing silly phonetic exercises in front of a mirror...) I reached to the conclusion of what I thought before: dental consonants and velarization are mutually exclusive. Why? When someone articulates a dental consonant the tongue stretches slightly to the front in order to reach the back of the upper teeth, and therefore, the tongue remains more and less horizontal. When someone articulates an alveolar consonant, the tongue slightly curves up to reach the alveolar ridge. In doing that, the tongue makes a concavity facing up and lifting the back part of the tongue to get close to the velum, hence the velarization. That’s why, in the particular case of the /l/, the dental /l/ is not velarized, while the alveolar /l/ is. The article velarization, does not say that anything about dental consonants, but the article velarized alveolar lateral approximant does mention European Portuguese, citing a source. This subject bothers me because, as a native speaker, saying that Brazilian Portuguese /l/ is alveolar and European Portuguese /l/ is dental is... is... it’s just wrong! I don’t have any other words to say it! That’s why I find very strange the source you cite (Mateus & d'Andrade (2000)) say that European Portuguese /l/ is dental, in later works of Mrs. Mateus she only says that European Portuguese /l/ is velarized.

Regarding the phoneme /ʁ/, again, I don’t have that source (yet), but as far as I know there aren’t yet studies about the phonology of the Portuguese spoken in the African countries. Be carefull in citing sources because there is a widespread myth that the 5 varieties of Portuguese spoken in Africa are exactly the same that the one spoken in Portugal, which is not true. I am pretty sure that most of Angolans and Mozambicans pronounce /r/ and not /ʁ/.

Ten Islands (talk) 17:43, 10 February 2008 (UTC)

That was a nice bit of original research. Unfortunately, you can't make sweeping statements about all speakers simply by your personal difficulty in pronouncing a consonant. With that logic, I could argue that dental trills are impossible because I can't do them. The source cited for EP in velarized alveolar lateral approximant is Cruz-Ferreira (I should know, I added it) and there is no article for a "velarized dental lateral approximant". That article (in the lead section), and Velarization state: "For many languages, velarization is generally associated with more dental articulations of coronal consonants so that dark l tends to be dental or dentoalveolar while clear l tends to be retracted to an alveolar position." This information comes from Recansens & Espinosa (2005). Of course, this isn't universal and Angr has mentioned that some varieties of Irish contrast a dental velarized lateral with an alveolar velarized lateral. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 20:44, 10 February 2008 (UTC)

Well, now you are starting to be aggressive, there’s no need for that if I’ve done nothing to deserve it. If an other user said that there exists velarized dentals, that’s O. K., I believe it. But that’s not the case of the Portuguese /l/. I am aware of the Wikipedia’s policy about verifiability, but one has to have a common good sense to and be critical. One cannot put a wrong statement just because there is a source saying this or saying that, we’ve got to have some critical sense also. Saying that Brazilian /l/ is dental and Portuguese /l/ is alveolar is not my original research. Just listen! I don’t know in which part of the globe you are right now, but if you have the chance to speak with native Brazilian speakers and Portuguese speakers, just listen, and then ask how do they articulate the /l/. I am sorry to insist on this subject, but the informtion as it is right now does not correspond to the truth... Ten Islands (talk) 05:36, 12 February 2008 (UTC)

Having me listen to a Brazilian speaker and a European speaker won't work, I personally can't tell the difference between dental and alveolar pronunciations. I tell you what, once (and not before) you find sourcing that says that it's alveolar in Europe and dental in Brazil, you're welcome to change it. It's really a minor phonetic note and nothing to get worked up about. Wikipedia won't break until then and there's not much more I can say other than "find a source." Regards. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 06:28, 12 February 2008 (UTC)

I finally had the chance to get my hands on Mateus & d’Andrade’s The Phonology of Portuguese (you’ve put me to delay my trip in order to get some time to visit a library... :-( ). I am sorry but, as I suspected, I didn’t find any reference about the articulation point, if it is dental or alveolar. Also, the book only talks about European and Brazilian Portuguese. It doesn’t mention the varieties in Africa. But perhaps you’re right, “it’s a minor phonetic note”, and talking about the velarization in European Portuguese is enough. Also, since we haven’t a solid source yet, it is better not to mention anything by now. I happen to have two sources (besides my awareness as native speaker), but I’ve been reluctant to mention them. I’ll keep on looking. Regards. Ten Islands (talk) 05:49, 13 February 2008 (UTC)

The statement that "lateral consonants do not normally occur at the end of syllables" is correct phonetically, though incorrect phonemically. However, it applies only to Brazilian dialects. It is totally false in European Portuguese. FilipeS (talk) 19:13, 9 February 2008 (UTC)

Converting away from tables, why?

May I ask, Aeusoes1, why you're converting the tables into text? Is there a Wikipedia directive about this? FilipeS (talk) 23:35, 9 February 2008 (UTC)

No. But there's no reason to have all the information in tables. I've been doing this for a while on the phonology pages and Portuguese is (I think) the last phonology page with all of its information in tables. There shouldn't be a loss of information, although admittedly I have omitted some dubious information (such as mention of the nasalized quadphthongs). — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 00:31, 10 February 2008 (UTC)

Nasalized quadphthongs? Dubious? I'll have to see what you've been doing with that.

In my opinion, the information was more user-friendly when it was placed in tables. For example, I much preferred the table version of Spanish phonology. FilipeS (talk) 21:24, 10 February 2008 (UTC)

Another problem was the overemphasis on orthography, which in my opinion is too close to the letter x is pronounced as y in dialect z. As for the bit about Catalan, please see Catalan phonology and show me how standard Catalan uses vowel height to contrast stressed and unstressed vowels. If you're referring to the reduction of a seven-vowel system to a 3 vowel system, that's technically vowel reduction, not vowel height differences and is too dissimilar to the Portuguese vowel reduction system to make the statement "Like standard Catalan, Portuguese uses vowel height to contrast stressed syllables with unstressed syllables" accurate, meaningful, or helpful. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 21:49, 10 February 2008 (UTC)

Because the reduced vowels are higher than the unreduced ones. Like in Portuguese. FilipeS (talk) 22:48, 10 February 2008 (UTC)

I still say that the processes are too different. Catalan neutralizes vowel differences while Portuguese maintains most of them. In Portuguese, the vowels simply get a little higher while in Catalan /ɔ/ reduces to [u], which is considerably higher. It's not that big a deal I guess, but I don't think readers are going to get a lot out of that comparison. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 23:05, 10 February 2008 (UTC)

"Portuguese maintains most [vowel differences]" --> This is not quite true. In unstressed syllables, the vowels in the three groups (a, ɐ), (e, ɛ, ɨ), and (o, ɔ, u) are very nearly present in complementary distribution. There a couple of exceptions, but they are an insignificant percentage. This was said in the previous version of the article. FilipeS (talk) 23:14, 10 February 2008 (UTC)

Well that's even more at variance with Catalan. The only sound that doesn't appear in stressed syllables in standard Catalan is [ə].

Pretty much like [ɨ] in European Portuguese...

Are you saying that, for example, [ɐ] appears in unstressed syllables while [a] appears in stressed syllables? — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 23:27, 10 February 2008 (UTC)

It's not as simple as that, but, roughly, [a] tends to be predominant in stressed syllables, while [ɐ] is predominant in unstressed syllables -- this for European Portuguese. In Brazilian Portuguese, the two are allophones. FilipeS (talk) 00:34, 11 February 2008 (UTC)

Okay, I just wanted to make sure I understood what you were saying about the distribution. It's odd that you're saying that the exceptions are insignificant. Complementary distribution is one of the major measures of allophony. That and if one sound replaces for another. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 00:53, 11 February 2008 (UTC)

This was explained with some caution (though perhaps not explicitly enough) in the previous versions of the article. I haven't checked if it's still there. The thing is that [a] and [ɐ] are allophones in Brazilian Portuguese, to the best of my knowledge with no exceptions. (But unstressed [o ~ ɔ] and unstressed [u] do have rare minimal pairs.)

In European Portuguese, the situation is more delicate. There are not minimal pairs between unstressed [a] and unstressed [ɐ] made up of polysyllabic words, but there are many minimal pairs where one word is a monosyllabic clitic and the other is "stressed" monosyllable, such as da [dɐ] vs. dá [da], or ma [mɐ] vs. má [ma]. I guess it's arguable whether they are distinct phonemes in unstressed syllables... FilipeS (talk) 02:12, 11 February 2008 (UTC)

About /i/ /u/ and /ɪ/ /ʊ/ and other subjects

Here we meet again Aeusoes, for the third time... Since I’ve seen that you make serious contributions, you deserve an explanation for my edits.

The sounds /ɪ/ and /ʊ/ do not exist in Portuguese, what exist in Portuguese are the closed vowels /i/ and /u/ (at least phonologically, I don’t know if there is an obscure guy in a remote area that pronounces /i/ and /u/ as /ɪ/ and /ʊ/... :-) ). But most important, I haven’t seen any author claiming that the sounds /ɪ/ and /ʊ/ exist.

Portuguese (I mean Portuguese language) phonologists represent the semivowels either with /i̯/ /u̯/ or with /j/ /w/. What I’ve done was simply put the inverted breve mark below the symbols where they were missing. Not putting it will cause confusion with hiatuses (for instance, BP pais /pai̯s/ (parents) versus país /pais/ (country).

I am very sceptical about a difference between /ɐ/ in Brazil and /ɐ/ in Portugal. The French word de /də/ (of) and the (European) Portuguese da /dɐ/ (of the) sound different.

The text was not clear to say that the sound /ɐ/ is the most common occurance in unstressed syllables in EP. In BP it only appears before nasal consonants.

The word dá is not a compound word. The text is making confusion when (in EP) there is an etymological crasis of two /ɐ/ (for instance caveira /ka'vejɾɐ/, in medieval Portuguese caaveira */kɐɐ'vejɾɐ/) or when there is an /ɐ/ at the end of a word and another at the begining of the next word (minha /'miɲɐ/, amiga /ɐ'miɡɐ/ but minha amiga /ˌmiɲaˈmiɡɐ/).

The EP unstressed “e” is a central vowel, not a back vowel.

The words ao and aos are pronounced /au̯/ and /au̯ʃ/ in EP. I am not aware of any circumstance where the diphthong /ɐu̯/ exists.

At last “Central Portugal” can be a bit confusing. Coimbra, for example, is also considered Central Portugal, but dialectically it is a different area than Lisbon and surroundings.

I am not reverting anything by know, I would be glad if you could explain why you once claimed that /i̯/ /u̯/ is incorrect.

See you. Ten Islands (talk) 07:11, 4 March 2008 (UTC)

I apologize about the diphthongs. I don't actually oppose your edits to that table but I was being lazy. It's good that you are able to articulate the justifications and I wonder now if /j/ and /w/ might not be better. It is, after all, how Portuguese is represented at our article on diphthongs.

You say that /ɪ/ and /ʊ/ don't exist in Portuguese. This is true, but unstressed /i/ and /u/ are pronounced as [ɪ] and [ʊ] in Brazilian Portuguese; remember the distinction between phonetic [brackets] and phonemic /slashes/! If you looked at the vowel charts you would see that Barbosa & Albano (2004) make this claim (they also transcribe unstressed /a/ as [ɐ] You'll also notice that the two vowel charts show that there is a height difference between [ɐ] of the two dialects. I mean, c'mon. Use your eyes, not your ears. Especially if you're going to compare it to French. French /ə/ is rounded and often very front in comparison to other languages.

As for the rest, you might want to talk to FelipeS. I believe he can back up those claims better than I can.

I also recommend that you find and read Cruz-Ferreira (1999) and Barbosa & Albano (2004). That might help us get on the same page. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 07:46, 4 March 2008 (UTC)

I disagree with several things that both of you have said.

Ten Islands:

• Actually, diphthongs and hiatuses can normally be distinguished in Portuguese by stress alone. For example, pais, [pais] or [paiʃ], a monosyllable stressed on the "a", versus país, [paˈis] or [pɐˈiʃ], with two syllables and stress on the "i". It is not necessary to use the non-syllabicity diacritic.

• There is most definitely a difference between /ɐ/ in Brazil and /ɐ/ in Portugal. You just have to look at the vowel charts in the article to see that the Brazilian vowel is more open (though I suspect that the pronunciation of this vowel varies considerably within Brazil itself). Nevertheless, I agree that the difference is minimal, and does not justify using different symbols for the two; they are dialectal variants.

• The unstressed "e" in Portugal is a near-back vowel, not a central vowel. Again, just look at the chart, where it's transcribed as [ɯ].

• The words ao and aos are pronounced /ɐu̯/ and /ɐu̯ʃ/ by most EP speakers. This is an exceptional pronunciation (it's allophonic), but it is the standard in Portugal.

• For the purposes of this article, "central Portugal" should be interpreted as the estremenho group of dialects, which includes both the dialects of Coimbra and of Lisbon. The difference between these two are frequently overstated. FilipeS (talk) 14:33, 7 March 2008 (UTC)

Aeusoes1, in addition to seconding all the staments made by Ten Islands which I did not contradict above, I would like to say that, like him, I do not like the use of the symbols ɪ and ʊ in an article on the phonology of Portuguese, for two main reasons:

• First, they are merely allophones, not phonemes.
• Second, they do not seem to be present in European Portuguese. Lord knows Portuguese phonology is complicated enough already without us making it even more messy by listing dialectal phones that aren't even phonemes in any dialect!

I'd be curious to see, though, a few words where the author claims these phones are used by Brazilians. Best regards. FilipeS (talk) 20:03, 8 March 2008 (UTC)

Here I am back again. Answering each point of both of you:

• Sorry, I didn’t pay attention to difference between brackets and slashes. Even so, I keep on skeptical about the existence about the phones [ɪ] and [ʊ] in BP (never heard it!). But since it is referenced, I won’t revert it, I admit that some speakers may use the phones [ɪ] and [ʊ] as allophones of /i/ and /u/.
• Yes, the words pais and país can be distinguished by the stress alone. But if we want to be precise and use correct phonological representation, the distinction between vowels and semivowels should be made.
• I don’t think there is a difference between BP ɐ and EP ɐ (either phonemic either phonetic). Besides the charts, do you guys have another source for that?
• Sorry to disagree with you again, Aeusoes, but I think you are mistaken. The sound /ə/ in French is neither rounded neither front. Aren’t you by any chance influenced by the Parisian dialects? People in Paris do not make anymore the distinction between /ə/ and /ø/, some minimal pairs like de \ deux or je \ jeu are not distinguished anymore. But perhaps this should be discussed elsewhere…
• Talking about the charts, I don’t know who came up with them. Yes, I know that they are from Cruz-Ferreira and Barbosa & Albano, but what is their source? The phoneme /ɐ/ is located in different places by these authors, but does any of them explicitly say that they are different in BP and EP? But, if there is a difference, as Filipe says it is minimal, perhaps it should not be pointed out at all.
• Regarding the sound called “mute e” it is a bit problematic. Yes, I can see the chart, and yes, the sound /ɯ/ is (kind of) at the correct place. The trouble is, it’s not the correct sound. The representation of that sound have always been problematic among Portuguese linguists, being sometimes represented by /ɨ/, sometimes by /ɯ/ and sometimes even by /ə/. But the most recent representations tend to use /ɨ/, Mateus consistently uses /ɨ/. I have seen some charts where that sound appears vertically above /ɐ/ and /a/, and not together with /u/, /o/ and /ɔ/. I mean, c’mon. Use your ears, not your eyes! The unstressed e in EP is definitely a central vowel! But how to overcome this? Perhaps citing that different authors use different symbols for that sound.
• Thank you for suggesting me to talk to Filipe, I know as well as you that he has made good contributions regarding the Portuguese language. But either you are forgetting that me too I’m a native speaker, or either you are not giving me too much credit, what makes me sad…
• Sorry but the word ao is pronounced /au̯/. The same goes to its plural. I will insist on that point.
• At last, I never liked the dialectal division of Portuguese presented here in Wikipedia. It looks like more an administrative division (in Brazil, states, in Portugal, regions). It is a common and popular division used mostly by people that are not linguists. For instance, what exactly is the difference between the Alentejo dialect and the Algarve dialect? I prefer the division made by Cintra (check page 17), it is more scientific and more concise. Regarding the usage of the word “Central” I still think that it might lead to misinterpretations. The center of Portugal is a vast, not well-defined area that may comprise places like Coimbra, Guarda, Lisbon or Portalegre, each one of these four belonging to four different dialectal subgroups. Using “Lisbon area” or “Estremenho” would be less misleading.

Ten Islands (talk) 20:56, 8 March 2008 (UTC)

TenIslands, first of all you are not the only native speaker here. I am one too.

• There is most definitely a difference between ao [ɐu̯] and au [au̯], at least for many Portuguese and in standard European Portuguese. This distinction may well be absent in other dialects, like in northern Portugal or in Africa, and I know there is no difference in Brazil. But we were talking about (standard) European Portuguese. You probably miss the difference because it's allophonic.
• My statement that there is a phonetic difference between BP [ɐ] and EP [ɐ] is based on the vowel charts in the article, but not only in them. It's also based in my experience of many years as a native speaker listening to both European and Brazilian Portuguese.
• According to the available sources, the unstressed "e" of European Portuguese is neither a central vowel nor a back vowel. It's a near-back vowel. And it is most definitely not a mid vowel like /ə/, even though many authors use that misleading notation because of its spelling. As an aside, I've often found the close back compressed vowel of Japanese (IPA /ɯ̫/) reminiscent of the Portuguese unstressed "e" (though they are not identical, to be sure).
• O óptimo é inimigo do bom. If you make the phonetic transcription too narrow, you will end up with an illegible mess of an article. Remember that Portuguese has many, widely divergent accents which need to be addressed together here.
• This article is not about dialects, and it should not attempt to become an article about dialects. I agree that Portuguese dialects currently leaves a lot to be desired, but let's improve it and discuss it there, not here.

Later. FilipeS (talk) 21:16, 8 March 2008 (UTC)

I only refered to FelipS because he knows better than I do on some of the points you brought up, not simply because he's a native speaker.

As for French /ə/ I was pretty much thinking you meant Standard/Parisian French. The sources for the charts, if I recall correctly, used native speakers. The Brazilian chart is that of a "female undergraduate student at the University of Campinas at age 21 at the time of the recording. She was born in the city of São Paulo, where she lived up to age 10. From then on she has been living in Valinhos, a city on the outskirts of Campinas." You seem to have better ears than I do at distinguishing between variations in vowel height and frontness.

The best thing to do in regard to mute e is to use the most commonly accepted representation. The Bulgarian vowel inventory has a vowel that many speakers and editors would consider is in the vowel space of [ə] and should be transcribed as such but the HIPA transcribes it as [ɤ] and that's how our article represents it.

You say that the Brazilian/European distinction is used mostly by people who aren't linguists but it seems as though the Journal of the International Phonetic Association uses that distinction. If you are suggesting we highlight distinctions between what one might call "sub-dialects" then I don't have a problem with that as long as we use sourcing. I can't read Portuguese but if you would like me to turn the maps in that pdf file into something Wikipedialific, then I can do it.

Finally, Barbosa & Albano include ɪ ʊ in the vowel chart, and throughout the article saying that "[i]n post-stressed position, the variety studied here contrasts four oral vowels, including a centralised version of [e]..." Why don't we simply use [i] [u] for pan-dialectal representations and ɪ ʊ for specifically Brazilian representations? Obviously for phonemic representations we would always use /i/ /u/. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 21:18, 8 March 2008 (UTC)

Hi guys! Answering to FilipeS, first:

• I know that you are a native speaker. That’s why it shocks me when you say that the word ao is pronounced /ɐu̯/. Where did you see that? Do you have a source for that? Perhaps the best solution is to ask to other (European) Portuguese speakers here in Wikipedia (for those who don’t understand IPA symbols, the symbols /ɐu̯/ represent what could be written âu in Portuguese, while the symbols /au̯/ represent what could be written áu).
• Perhaps it is a good strategy to compare to other languages’ sounds. No, I don’t agree. I believe that the unstressed “e” is closer to Romanian written “î” or Russian written “ы” than the Japanese /ɯ̫/. If we can’t overcome this, I repeat, perhaps the best solution is to say that different authors use different symbols for that sound.
• Do you really think that distinguishing vowels from semivowels is “an illegible mess”? Let’s hear the opinion of other people.

Answering to Aeusoes now:

• O. K. But which is the “most commonly accepted representation”? What do you think about what I said to Filipe about citing different authors?
• It was only yesterday that I stumbled in the article near-close near-back vowel. I don’t know if it was you who entered the info about Portuguese, but that’s completely misleading! The BP unstressed /u/ does not sound at all like the EP unstressed “e”! This means two things: 1st, the BP unstressed /u/ is not near back, but rather back; 2nd, the EP unstressed “e” is not near back but rather central (and not rounded). That’s why I prefer to use my ears rather than look at charts. Sometimes the symbols are misplaced or miswritten. Even if the BP /u/ may have several allophones, phonemically it is always identified with EP /u/ which contrasts (phonemically) with EP /ɨ/.
• As I said before, if the statement that BP /i/ and /u/ has the [ɨ] and [ʊ] allophones is sourced, I won’t revert it, but of course for phonemic representations we would always use /i/ and /u/.
• When I was talking about “mostly by people who aren’t linguists” I was referring about dialectal divisions rather than phonetics itself.

For both of you:

• Do you guys want to go for improving Portuguese dialects? I don’t know how to build maps in Wikipedia, but if you want, Aeusoes, I can help you with translating some info in this document, unless FilipeS would be so kind in doing it quicker than me.

See you! Ten Islands (talk) 06:45, 10 March 2008 (UTC)

• What I said was in response to what you both said. I'm not exactly sure what the most common representation is, but I think discussing the matter of different authors' representation a bit in the article would be appropriate.
• It's important to realize that near-close near-back vowel is talking about both a rounded and an unrounded vowel since no language is known to contrast the two. The EP vowel is unrounded and the BP vowel is rounded but the vowel space both occupy is roughly the same
• What I was suggesting is that even in phonetic transcriptions that we should use [i] [u] when making transcriptions that should apply to both dialects.
• As was I.
• I'm working on the map as we speak, if anything in the text can be used to add to Portuguese dialects then you and FelipeS are the ones to do it. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 20:26, 10 March 2008 (UTC)

Sorry for this week of absence but I’ve been busy. I believe that by now we are able to join our efforts to improve this page.

At first I intended to give a feedback about each question that was approached in this discussion, stating my point of view. But then I remembered one of the rules of Wikipedia about the verifiability and instead I am suggesting now some changes in this page (and other related pages) in accordance of what is verifiable or not. Here we go:

• “European Portuguese has also two near central vowels” — remove or change. I hadn’t the chance to check Cruz-Ferreira yet, but it would be better to certify if she uses the graphic symbol /ɯ/ or if she clearly says that that vowel is a back one (or near-back). I’ve seen the usage of the word “central” for describing /ɨ/.
• “European Portuguese has also two near central vowels” — again, remove or change; all the sources I’ve seen use the word “central” for describing /ɐ/ and not near-central.
• “The exact realization of the /ɐ/ varies (…). In Portugal, it is (…) higher (…).” — keep; I keep on skeptical but it is sourced.
• “In European Portuguese, the general situation is similar (with [ɐ] being more prevalent in unstressed syllables), except that for some regions (…)” — change; not “in some regions”.
• The inclusion of unstressed “e” among back vowels is completely wrong! Even if we use a non-rounding diacritic. That’s not sourced at all and it would imply serious errors:
  • It would imply that for a Brazilian speaker the words remar and rumar pronounced in EP would sound nearly the same. Even for a brazilian ear they sound different.
  • Some African speakers have some trouble pronouncing the EP unstressed “e”. They replace it by /i/ or /e/, never by /u/.
  • Historically, the sound of today’s unstressed “e” derived from /i/. Even today it may be found (dialectically) pronunciations with /i/ (for instance: chegar pronounced /ʃi'ɡaɾ/) but never with a back vowel.
• The phonological representations of Portuguese are always consistent in using different symbols for vowels and semivowels. If we choose to use /j/ and /w/ or /i̯/ and /u̯/ that’s an other discussion, but not /i/ and /u/.
• “Allophone [ɐi] in central and southern Portugal (…)” (for /aj/) — remove; not sourced.
• The pronunciation /ɐj/ for the written “ei” is exclusively from Lisbon, not from “central Portugal”. Let’s not make confusion with the fact that it was incorporated in Standard European Portuguese.
• The Estremenho dialect cited by Vasconcelos does not include Coimbra.
• “Allophone [ɐu] in Portugal (…)” — remove; not sourced and incorrect; I have sources that prove that the word ao is pronounced [aw].
• About the existence of the allophones [ɪ] and [ʊ] in BP — keep; I am still skeptical but it is sourced;

Ten Islands (talk) 10:08, 25 March 2008 (UTC)

• "The pronunciation /ɐj/ for the written “ei” is exclusively from Lisbon, not from “central Portugal”. " I disagree, and your claim is unsourced.
• "I have sources that prove that the word ao is pronounced [aw]." In which dialects of Portuguese?
• TenIslands, are we speaking Chinese to you? The unstressed "e" of EP is not a back vowel. Nobody here has made that claim. It is a near-back vowel. You seem to be under the misguided notion that vowels located near each other in a vowel chart must necessarily sound alike. This is not the case. Spend some time listening to the IPA cardinal vowels here. In many cases, vowels that are apart can sound more similar to each other than vowels that are next to each other.

BTW, I intend to come back to this talk page soon, when I have more available time. FilipeS (talk) 15:12, 27 March 2008 (UTC)

• “...your claim is unsourced” You are kidding, right?

Where's your source?

• As far as I know in all standard dialects. If there is a pronunciation [ɐw] it is either dialectal or idiolectal.

"As far as you know"?! Unsourced again.

• No need to be rude!

Likewise. FilipeS (talk) 00:03, 30 March 2008 (UTC)

• What I am questioning is if your sources clearly say that unstressed “e” is a near-back vowel (or back, or near close or whatever). As far you guys have only said that the symbol /ɯ/ is used or that the image shows that symbol in the position of a near-back vowel, and not that the authors clearly say that it is a near-back vowel. On the other side, my sources are unanimous in saying that it is a central vowel even if they are not unanimous in the symbology or in the graphic representation.

Ten Islands (talk) 15:47, 28 March 2008 (UTC)

Cruz-Ferreira (1995) says of the unstressed e: "...which occurs only inunstressed syllables, is often represented as /ə/ but does not correspond to the mid central quality associated with schwa. It is a fronted and lowered high back unrounded vowel; hence the symbol chosen here for its trnascription." (p.89) Even if she didn't say this, though, it's not going too far to look at the chart and describe their positions. It's possible that sources using the word "central" are either going with a shorter description because it's convenient, looking at the symbol used and not the formant values, or picking "central" because it's to fronted to be considered back. We can be more precise than them. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 21:51, 28 March 2008 (UTC)

TenIslands, can you mark the claims you dispute with {{fact}} tags? — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 21:51, 28 March 2008 (UTC)

Aeusoes1, do Barbosa & Albano give any examples of words where Brazilian unstressed /i/ and /u/ are pronounced as [ɪ] and [ʊ]? FilipeS (talk) 17:07, 4 April 2008 (UTC)

Mateus and d'Andrade on [β], [δ] and [γ] in Portuguese

The following quote is from Chapter 2, page 11, footnote 2 of the 2000 edition of the book.

As a matter of fact, in northern and central EP dialects, voiced stops may be realized as the correspondent non-strident fricatives [β], [δ], [γ], except in phrase-initial position and after an homorganic non-continuant.

(Emphasis mine.)

When they describe these allophones as "fricatives", I suspect that they're being misled by the ambiguity of IPA. Most likely, in my opinion, they are approximants, not true fricatives. If anyone still has any doubts, I can even show you several speakers who never pronounce the voiced plosives as anything other than true plosives.

The authors who claim otherwise are either ignorant foreigners who assumed wrongly that Portuguese is "pretty much like Spanish", or purist natives who find the approximant pronunciation more chic, and therefore fail to notice that many (perhaps most) speakers do not talk that way naturally. FilipeS (talk) 14:20, 7 March 2008 (UTC)

Is [δ] a variant of [ð]? I've never seen [δ] used in IPA before. Grover cleveland (talk) 16:21, 7 March 2008 (UTC)

They actually use the Greek letter delta in the book. Sometimes that's done when the proper IPA symbol is unavailable. FilipeS (talk) 16:25, 7 March 2008 (UTC)

Nice. I'm concerned though because Cruz-Ferreira (1995) (which is probably identical to Portuguese's HIPA entry) says these are fricatives as well. I understand your speculation that they could be approximants but we should find sourcing. I don't understand how speakers who don't exhibit this phenomenon of lenition are proof that these could be approximants--unless you're trying to illustrate that this isn't as common as Cruz-Ferreira state ("normally pronounced"), in which case you've certainly driven the point home quite thoroughly. — Ƶ§œš¹ _{[aɪm ˈfɻɛ̃ⁿdˡi]} 18:17, 7 March 2008 (UTC)

Yes, I was trying to show the latter. FilipeS (talk) 19:54, 8 March 2008 (UTC)

/oi/ vs /ɔi/

The article says: "There are very few minimal pairs for /oi/ and /ɔi/, all of which in oxytone words." There is at least "apóio" (verb) vs "apoio" (noun), so not all are oxytone. There are probably other words that follow the same pattern, but I can't remember any right now. 189.95.48.244 (talk) 02:11, 12 November 2008 (UTC)