Talk:Spanish phonology

Describing the relationship between /ɛ, ɔ/ and Spanish /ie/, /ue/

I just noticed that the edit I made in regards to this is related to a reverted previous edit I made in September. I find it unclear to describe the use of Spanish /e/, /o/ and /ie/, /ue/ as " similar to the distinction between the close /e, o/ and the open /ɛ, ɔ/" of other Romance languages: the connection is historical (the vowels /ɛ, ɔ/ developed by sound change into Spanish diphthongs) but the current phonetic outcome of those sound changes is not especially similar phonetically to /ɛ, ɔ/. However, I see that my previous wording was described as "an utterly wrong sweeping statement" in this January revision. While the correspondence is not perfect, due to separate sound changes in separate languages, I don't understand exactly why that makes it preferable to use the wording "similar to" instead of "correspond to". @Sol505000 are there alternative wordings that would address my concern while avoiding the inaccuracy that you found with my wording? Urszag (talk) 16:26, 9 January 2024 (UTC)

/dl/ does exist

regarding Spanish phonology: Difference between revisions - Wikipedia

@Nardog, the words from Listado de lemas que contienen «dl» | Diccionario de la lengua española | RAE - ASALE can be either /dl/ or /d.l/, just as intervocalic /tl/ can be /t.l/. this is also noticeable in foreign words or names like sandler or chandler where sometimes it's /d.l/ and other times it's /dl/ Brawlio (talk) 05:45, 5 March 2024 (UTC)

What makes you conclude those are not /d.l/? Nardog (talk) 07:09, 5 March 2024 (UTC)

it's not that they aren't /d.l/, it's that they can also be /dl/. ¿what makes you conclude that they're not /dl/? Brawlio (talk) 00:01, 6 March 2024 (UTC)

/tl/ is considered a possible onset in certain varieties of Spanish because they have words beginning with /tl/. Nobody is going to regard sandler or chandler as having /dl/ as an onset unless there are Spanish words that begin with /dl/, just as nobody analyzes English indict to have the onset /nd/ because there is no English word that begins with /nd/. Nardog (talk) 08:12, 6 March 2024 (UTC)

The /ndl/ sequence in Sandler and Chandler has to be divided as /nd.l/ or /n.dl/. In native vocabulary, Spanish has neither words that start with /dl/ nor words that end in /nd/. Spanish does have borrowed words that end in /nd/; however, on the other hand, there doesn't seem to be a theoretical reason why it couldn't have syllables that start with /dl/ in borrowed words (and an absence of word-initial examples could be accidental). Ultimately this is an empirical question (as much as matters of syllabification can be) where we should follow what reliable sources say.--Urszag (talk) 19:15, 6 March 2024 (UTC)

I see. Then let's remove the claim of nonexistence, which seems unsourced anyway (not to mention proving a negative). Nardog (talk) 00:19, 7 March 2024 (UTC)

I agree that we should generally say nothing rather than making claims of nonexistence that are of dubious accuracy, but the problem with just not mentioning /dl/ is that then the introductory sentence ("If and only if the first consonant is a stop /p, t, k, b, d, ɡ/ or a voiceless labiodental fricative /f/, a second consonant, either /l/ or /ɾ/, is permitted") makes it sound like "dl" is permitted (even if it doesn't quite say that). Furthermore, at least some reliable sources do seem to say that /dl/ is unattested as an onset (I added one citation now; I'm not sure whether the sources already cited at the end of the sentence about /tl/ were also supposed to apply to the claim about /dl/). So I think that in this case, what would be ideal is to find a source that discusses examples like those that Brawlio brought up and addresses if they can be analyzed as marginal exceptions to the general absence of onset /dl/ in Spanish.--Urszag (talk) 20:34, 7 March 2024 (UTC)

Phonotactic "restrictions" (not strict) on word-initial/post-consonantal palatals, etc.

Hi @Brawlio, I appreciate your contribution regarding the occurrence of /ʎ, ɲ/ after other consonants. However, I think we'll both agree that the article shouldn't be left in this state, so let's discuss what the best end state for the article would be. Do you think nothing of this kind should be mentioned, or do you think the current statements are indefensibly broad but have some kernel of truth to them?

It's obvious that these consonants are not strictly prohibited in that position, but I wonder if you would concede that they are at least uncommon, or not found in native vocabulary (other than compound/prefixed words)? huaiño, aclla, lliclla are borrowings from Quechua: I assume that Baker 2004 intended to exclude words of this type. As a thesis, the paper by Baker is not a top-rate source, so I'd be fine with removing it, but I think Morales-Front 2018 deserves some weight as an academic source, although its coverage of this topic is brief and not very clear (All I found to go on was the statement that "[ɾ] and [ɲ] are restricted word-initially"). Urszag (talk) 07:00, 8 March 2024 (UTC)

there are words like panllevar which we could surmise is a compound word but I'm not sure whether we can say so for certain. or enllentecer which comes Latin with no "un-prefixed" form in Castilian. there's also piesllo from vulgar Latin, aullar from Latin, traillar which is just the verb form of another word, and maullar. but yes, i agree that they're uncommon. Brawlio (talk) 18:47, 8 March 2024 (UTC)

@Brawlio, what do you think of the current phrasing? It is not fully sourced, but I'd like to remove the "dubious" tags from the statements cited to Baker 2004, Hualde 2022, and Morales-Front 2018, given that I think the article now provides enough context that a reader is not in danger of misinterpreting these statements as referring to ironclad prohibitions on word-initial /ɲ/ or postconsonantal /ɲ/ or /ʎ/.--Urszag (talk) 03:47, 8 April 2024 (UTC)

muy and /uj/

The note for this diphthong is "Harris (1969:89) points to muy ('very') as the one example with [uj] rather than [wi]." Nevertheless, there are several (mostly) regional examples of its occurrence too, like the reginal "cuy" or the widely used interjection "¡uy!". Saviourofthe (talk) 22:06, 30 October 2024 (UTC)

I am convinced, Spanish /ie/, /ue/, /ui/ should not be described as rising diphthongs (as it is claimed everywhere), but as falling diphthongs or at best floating diphthongs. - And yes, there are more examples than "muy"; think of the name "Luis", which according to the rules should be stressed on the "i", but undoubtedly gets stressed on the "u" part! - Also think of "pie" (foot) which should be stressed on the "e" part, but which I only ever hear stressed on the "i" part! - There are more arguments that support my position: "hasta luego" can be heard all the time as if the stress were on the "u" and "e" becomes kind of a Schwa sound, some even say more or less "hasta logo" (and not "hasta lego" which would be more logical if the "e" were prominent). - When people are starting to curse (saying "mierda"), they can usually be heard saying "miiii-er..." (and not "mjer...). - Also compare French "pied" (foot) and Spanish "pie" (foot) which should sound almost identical, but indeed do NOT - French is like "pje", Spanish like "piie" (in contrast to Spanish pronunciation and spelling rules!). --- ManfredWV 94.219.14.220 (talk) 02:21, 26 January 2025 (UTC)

/s/ and /tʃ/

Are there accents or dialects of Spanish in which /s/ is some kind of retracted or apico-alveolar sibilant (i.e. the sibilants which are heard by English speakers as sounding "in between s and sh") and, moreover, the sibilant part of /tʃ/ sounds like this /s/, so the affricate sounds like /t/ + /s/? Or alternatively, even if the /s/ is a normal alveolar sibilant like in English, the /tʃ/ still sounds like /t/ + retracted or apico-alveolar /s/? I seem to distinctly remember /tʃ/ sounding that way often. --Florian Blaschke (talk) 15:24, 26 October 2025 (UTC)

Most Peninsular Spanish has a somewhat retracted /s/ and /tʃ/ can be somewhat fronted in some Peninsular varieties (including some Madrid speakers). I recall some IPA Illustration-like source talking about the fronted /tʃ/ but I can't seem to find it right now. /tʃ/ can also be pretty fronted in some upper & middle-class Chilean speech. Erinius (talk) 23:40, 26 October 2025 (UTC)

Just found the source I was thinking of which talks about fronted ⟨ch⟩ in Madrid - https://www.ipds.uni-kiel.de/kjk/pub_exx/kk2009_4/spanish/spa_madrid.html - it was cited on the [tʃ] article. Erinius (talk) 23:44, 26 October 2025 (UTC)

Thanks. Do you mean the column "Assimilation" and the partial row "Palatalized" in the table under the heading "Conventions", where it shows an allophone [tsʲ]? That's not really what I'm talking about. I mean varieties where /tʃ/ is consistently realised as [ts̠] or [ts̺], always, under all circumstances, never actually as [tʃ]. --Florian Blaschke (talk) 20:30, 7 November 2025 (UTC)

Yeah that's what I was referring to. I wouldn't know if it's always [ts] in some varieties. Erinius (talk) 11:46, 8 November 2025 (UTC)

To be clear, I do not mean the voiceless alveolo-palatal affricate [tɕ] (though it would also be interesting to know where this one occurs, if at all). I mean the retracted alveolar or apico-alveolar affricate (regardless of the realisation of the fricative sibilant). (Theoretically, it would make sense for the affricate sibilant to have the exact same place of articulation as the fricative sibilant at least in some varieties.) --Florian Blaschke (talk) 20:37, 7 November 2025 (UTC)

Stumbled on this: https://oxfordre.com/linguistics/display/10.1093 (can access with Wikipedia Library). It mentions [ts̺]-like realizations are noted "in coastal Colombia, Bolivia (Santa Cruz), and Chile" citing the Nueva gramática de la lengua española. I've checked and the gramática gives little more detail. And I'm not sure there's any variety where /tʃ/ is consistently, exclusively [ts̺]. Erinius (talk) 00:16, 12 November 2025 (UTC)

@Erinius: Thank you. Hm. It just seems to me, as I hear it in my mind, that /tʃ/ is commonly not actually pronounced with a [ʃ], but something else, which has really stood out to me. It could even be a realisation more like [tj], like ocho as [ˈotjo]. Or is that just the [tsʲ] allophone you've mentioned? But that one is only found in specific contexts. Or maybe what's throwing me off is simply the fact that in German and possibly also English, /ʃ/ tends to be rounded. --Florian Blaschke (talk) 01:56, 22 November 2025 (UTC)

Broad vs. narrow transcription of final obstruents

I wanted to ask who has decided to transcribe /peɾspekˈtiba/ as /peɾspeɡˈtiba/ in a broad transcription here and the Wiktionary? IMO a hard c is mostly associated to /k/ in standard Spanish transcriptions. I understand that the difference between (voiced–voiceless) obstruents is lost word-finally in narrower contexts, but not in broader ones as I mention (c.f. with betacism in broad vs. narrow transcriptions: /b/ and [b ~ β]). IMO this interpretation is not fully correct. — Jɑuмe ^(dis-me) 09:25, 11 November 2025 (UTC)

nm

The section on consonant assimilations covers all the possibilities with except as in inmediato, inmortal, un montón, en México. I couldn't find reliable information on that. Can someone who has reliable information add it to the article, please? —JerryFriedman (Talk) 16:00, 12 November 2025 (UTC)

It does cover it, before a bilabial it becomes [m], therefore (which is the Spanish modern less etymological form of ) is pronounced as [mm] or [mː]. — Jɑuмe ^(dis-me) 08:47, 16 November 2025 (UTC)

The sequence "nm", which appears only across morpheme boundaries, is fairly rare and unfortunately I think it's hard to find sources that discuss it in detail. I believe I've seen some article that mentioned that /nm/ does not invariably undergo regressive place assimilation (either across word boundaries or within a word), but I don't know if there is a good source of data on how common its various potential phonetic realizations are compared to each other.--Urszag (talk) 15:00, 16 November 2025 (UTC)

IMO it is not so rare, but it is mostly found in learned terms (inmortal 'immortal') or in compounds (conmigo 'with me'). The pronunciation I suggested is mostly accepted and is found in some dictionaries, other pronunciations could be [nm] or [ŋm]. Elision of n is not accepted in the standard as in Portuguese although it could be the choice for some accents. It is accepted a weakening or almost/quasi elision (therefore ⁿm) — Jɑuмe ^(dis-me) 08:59, 17 November 2025 (UTC)

And, third, assimilation of /n/ followed by a bilabial consonant in colloquial Spanish commonly gives rise to co-articulation, resulting in a nasal with a simultaneous occlusion at the alveolar and the bilabial PA: i[n͡m]móvil ‘motionless,’ co[n͡m] Pablo ‘with Pablo,’ etc. (Harris 1969: 14–16; Navarro Tomás 1977: 89, 113). Nasals realized with a simultaneous velar and bilabial or labiodental PA have also been reported for velarizing varieties https://doi.org/10.1002/9781118228098.ch6 citing Harris 1969 for the latter. Erinius (talk) 19:12, 17 November 2025 (UTC)

According to the CVC: Debilitamiento y "casi" desaparición de la «n» en el grupo «nm». En pronunciación lenta, percepción clara de las dos consonantes

inmóvil, con mucho gusto > in/m móbil, con/m mucho gusto.

Nasals realized with a simultaneous velar and bilabial or labiodental PA have also been reported for velarizing varieties - aren't these simply [mˠ, ɱˠ] (labials with velarization) then? Sol505000 (talk) 12:26, 18 November 2025 (UTC)

mm and nn

I just wanted to discuss the pronunciation of gamma (which contrasts with gama) and innato (which contrasts with y nato). How do you transcribe them? — Jɑuмe ^(dis-me) 09:30, 17 November 2025 (UTC)

Es interesante observar que, aunque el español no tiene consonantes geminadas, la secuencia [nn] puede aparecer en posición interior de morfema (analizable como /nn/) en dos o tres palabras, incluyendo perenne y pinnípedo. Con la excepción de estas, no existen geminadas en posición interior de morfema en español. Con el prefijo in-, hay ejemplos como innoble, innatural, etc. Hualde, José Ignacio (2014). "Nasales". Los sonidos del español (in Spanish). p. 173. Same page also mentions /mn/ as in columna as an exception to nasal assimilation (and obvio as having a geminated /b/). Erinius (talk) 19:29, 17 November 2025 (UTC)

mn

How about in alumno? — Jɑuмe ^(dis-me) 09:40, 17 November 2025 (UTC)

⟨-ms⟩

¿what is the realization of the ⟨m⟩ in words like "álbums" or "films"? ¿does it assimilate to the following consonant's place of articulation as the article seems to suggest happens with "álbum de fotos" or "film de suspense"? Brawlio (talk) 21:13, 14 November 2025 (UTC)

I don't think it does. The correct plural of álbum is álbumes. Film or filme are films or filmes. In my opinion if you don't pronounce final m as /m/ in careful speech you'd pronounce it as [n] or even [ŋ] and/or nasalization of a precedent vowel. Possible pronunciations of the word ítems ('items') are [ˈitẽms], [ˈitẽns], [ˈitẽŋs], [ˈitẽⁿs], etc. (excluding pronunciations where /s/ is omitted) — Jɑuмe ^(dis-me) 08:36, 16 November 2025 (UTC)

[ð]-devoicing outside Spain

Is there really nothing in the literature on this? You only need to watch one (random really) episode of Rebelde to hear many instances of (pre-pausal) [ð]-devoicing in Mexican Spanish. The actors that seem to do it pretty consistently are Ninel Conde and Enrique Rocha. Only yesterday I heard [usˈteθ] and [sjuˈðaθ]. OK, it might not be complete devoicing at all times, but sometimes it is (at other times it's just devoicing of the latter portion of the fricative, compare it with allophonic devoicing of the lenes in English which works just the same). The thing is, Mexicans don't seem to notice it because they lack the extra /θ/ phoneme present in Spain that would make them go "he said ustez" or "she said ciudaz". It's as unconscious as /ɾ/-devoicing, which coincidentally (or not!) occurs in the very same pre-pausal position. As ⟨z⟩ and soft ⟨c⟩ stand for the sibilant /s/ in Mexico, such respellings would make no sense unless you had a lisp (compare /ɑː/ respelled as ⟨aw⟩ in cot-caught-merged American and Canadian English, which makes no sense to Brits and Australians, who have that extra mid back rounded /oː/ vowel that contrasts with unrounded open back-to-central /ɑː ~ aː/).

It appears that MS (or some types of it at least) could have the most complex /d/-allophony of all accents, involving 3 phones in complementary distribution (and no phonetic overlap with other phonemes). Sol505000 (talk) 13:54, 20 November 2025 (UTC)

I share your frustration at not being able to find information on allophonic variation that is not allophonic to you. North Americans seem to have a hard time telling if a /d/ in their own speech is flapped, and I haven't had luck finding sources for the affricate /r/ in Indian English. Nardog (talk) 09:12, 2 December 2025 (UTC)

I've definitely heard that before. I'm also reminded of -d becoming [d] or even [t] which happens occasionally in some varieties of formal Colombian Spanish. I vaguely remember there being some sourcing on it. Example here: los dineros en su totalida[t]. Erinius (talk) 23:27, 2 December 2025 (UTC)

/ue/

Do we have something on the progression of that diphthong? Surely it didn't go from [ɔ] straight to [we], but probably from [ɔ] to [wɔ], unrounded to [wʌ], centralized to [wə] (probably through an intermediate [wʌ̈]) and then fronted to [we]. We have evidence of [wɔ] from Italian, compare Spanish fuego [ˈfweɣo] and Italian fuoco [ˈfwɔːko] with Portuguese fogo [ˈfɔɣu] (and Tuscan [ˈfɔːko], which seems to be more conservative than the Standard Italian opening diphthong [wɔ]). Sol505000 (talk) 21:34, 29 December 2025 (UTC)

There might be something on a page about Romance languages in general, I don't see anything on History of the Spanish language. Also I recall reading somewhere that some Astur-Leonese dialects show some different outcomes of o-diphthongization, like /ua/ or /uo/. In any case that's more diachronic stuff and I think this page's topic is more synchronic. Erinius (talk) 01:03, 1 January 2026 (UTC)

I think it is likely that [wɔ] was an intermediate. I think that information on the historical development could be put at History_of_the_Spanish_language#Diphthongization_in_open_and_closed_syllables. I don't think intermediate steps or the comparison to Italian are very relevant to this page, so I agree with Erinius's rollback.--Urszag (talk) 02:08, 2 January 2026 (UTC)

Ok, thanks. What about regional realizations? I'm watching a 2000-01 Champions League game Heerenveen vs. Valencia. The commentators (both of them, working for TVE1) keep saying something like [ˈxwɵɣo ~ ˈxwəɣo] for juego. Surely that's notable, if we had a RS? They're not likely to have a speech impediment, not both of them anyway (on the other hand, we had this guy Leszek Orłowski in Poland who worked for Canal+ and couldn't pronounce his /r/'s, which many people found a bit strange). Sol505000 (talk) 14:56, 5 January 2026 (UTC)

I'm not surprised. The idea that this Spanish diphthong ends specifically in [e] is the result of studies by philologists, who typically have trouble telling apart spoken language from written language. --Jotamar (talk) 00:57, 6 January 2026 (UTC)

It makes sense to have information about the current phonetic realization of "ue" in this article. I'm not sure how dialect-specific such pronunciations are: since this article focuses on describing a standardized form of the language, any pronunciation feature that is restricted to regional dialects might be best placed in the "Dialectal variation" section, or in the article Spanish dialects and varieties (which doesn't currently have any coverage of differences in the pronunciation of diphthongs). I have not yet come across a reliable source talking about variation in the pronunciation of "ue" in Spanish, but I see some discussion online, e.g. at fuego, luego (pronunciation) - ue > o (WordReference.com).--Urszag (talk) 10:21, 6 January 2026 (UTC)

cç

[c̠͡ç̠] as the voiceless countepart of ɟʝ? Compare mujer Voiceless_palatal_fricative and kilo Voiceless_palatal_plosive JMGN (talk) 16:48, 29 January 2026 (UTC)

Perhaps in Asturian? — Jɑuмe ^(dis-me) 23:15, 8 February 2026 (UTC)

[unstressed] allophones: tajada

As far as I know, the -ada /-ada/ suffix is pronounced [ˈɑðä̝] or [ˈɑ(ː)], so tajada ('chop') is [tɑ̈ˈχɑðä̝] (Northern Peninsular) or [tɑ̈ˈxɑ(ː)] (Southern Peninsular) rather than [tɑˈxa̠ða̠].

• ä̝ approaches to ɐ (like in Galician and Valencian), but it is still lower than the Barcelonan ɐ.
• x might be more emphatic χ in Northern Peninsular, whereas in some Southern Peninsular (but not in Murcia and parts of Eastern Andalusia) accents it is a soft glottal (h).

I believe in most cases these (unstressed) vowels are more centralised. In some varieties they might move slightly (like [ä → ɐ]) and/or further ([ä → ɐ → ə]). — Jɑuмe ^(dis-me) 23:15, 8 February 2026 (UTC)

@Nuvolet Ojo [ˈo̞ʀ̥o̞] https://en.wikipedia.org/wiki/Voiceless_uvular_trill JMGN (talk) 01:36, 9 February 2026 (UTC)

That's the right allophone 👍, thanks. How about [ˈo̞ʀ̥o̟]? — Jɑuмe ^(dis-me) 01:50, 9 February 2026 (UTC)

Per SOWL, the transcription ⟨χ ʁ⟩ doesn't rule out allophonic trilling. It's intermittent/more chaotic than in an underlying rhotic (unless, well, it's a rhotic itself - here it is not). We should differentiate between a fricative that's allophonically trilled (i.e. a phone that's primarily a fricative) vs. a trill with allophonic frication (i.e. something that's primarily a trill) at voiceless uvular fricative and voiced uvular fricative. They're not the same and uvular trills with frication should go to uvular trill and voiceless uvular trill. In diacriticless IPA, I think that the former are usually transcribed ⟨χ ʁ⟩ and the latter just ⟨ʀ⟩, even if it's voiceless (or just ⟨r⟩ - sadly this is still common). That's the usual practice, AFAIK. Sol505000 (talk) 17:20, 13 February 2026 (UTC)

I agree with you, Sol505000. — Jɑuмe ^(dis-me) 11:16, 16 February 2026 (UTC)

I don't know about Galician, but in Portuguese [ɐ] is significantly less prominent than [a]. I imagine there's a parallel difference in Catalan regarding [ə] vs. [a]. There's no such difference in Spanish, as far as I can tell. The "machine gun" (syllable-timed) effect is very audible, less so in Portuguese which is somewhat closer to English (i.e. reduced vowels are actually reduced in both articulation ([mid-]centralized) AND prominence (sometimes to pure voiceless aspiration - see Portuguese /ɨ/), something that IPA fails to convey - it's not just about quality. IPA fails miserably to convey the stress-timed effect vs. the syllable-timed one. I think ⟨◌͉⟩ (U+0349) could be used for reduced vowels, but that'd be an ad-hoc transcription.) It's one of the reasons European Portuguese "sounds like Russian" (which is stress-timed) and European Spanish "sounds like Greek" (which is syllable-timed), at least to people who speak none of these languages. Sol505000 (talk) 16:44, 13 February 2026 (UTC)

Catalan and Brazilian Portuguese could be intermediates between Spanish (syllable-timed) and European Portuguese/English (stress-timed). Centralisation and vowel reduction (along with vowel harmony, which extends the unstressed vowel system with extra allophones) is a key feature in many Iberian Romance dialects. — Jɑuмe ^(dis-me) 11:16, 16 February 2026 (UTC)

Did you check the link I provided with the usage of [ɐ ~ ə] in Bogotá? I believe inner (as well as Northern) Mexico and the highlands in South America tend to have some (sort of) reduced vowels which approach them to the intermediate systems between syllable-timed and stress-timed. — Jɑuмe ^(dis-me) 11:16, 16 February 2026 (UTC)

Caribbean debuccalization of /s/

It's mentioned as occurring at the end of words, but what about after vowels and before /k/, /p/ and especially /t/ (¿Cómo está Usted? > "¿Cómo ehtá Uhted?"), the pronunciation I picked up from my Cuban neighbors when I was young? Keith Ammann (talk) 14:46, 27 March 2026 (UTC)

Right now the article does mention debuccalization/aspiration before consonants but it might not be very clear Erinius (talk) 05:02, 28 March 2026 (UTC)

It happens in every syllable coda. However this trait often works as a sociolinguistic variable, so a person might pronounce [h] in an instance of a word and then [s] seconds later in another instance of the same word. The current text of this page explains it ok, I think, but the explanation is a bit lost in the middle of other considerations. --Jotamar (talk) 22:42, 30 March 2026 (UTC)

Voiced uvular fricative for /ɡ/ in Spain

Surely it appears at least allophonically? I've heard it many times whenever /ɡ/ is in contact with /o/. Algo [ˈalʁo], ¿Qué hago? [ke ˈaʁo] etc. It's probably an assimilation of place (/o/ is probably more back than cardinal velar consonants, being more post-velar or pre-uvular). I wonder if there are some minimal pairs with Portuguese /ʁ/ (which corresponds to an alveolar /r/ in Spanish) that one could give to a native European Portuguese speaker (which does have lenition of /b, d, ɡ/ at least in some dialects) and see whether they interpret this as their /ɡ/ or /ʁ/. Sol505000 (talk) 14:04, 2 September 2026 (UTC)

Remember Sol, a lenited /ɡ/ ranges from fricative ɣ to approximant ɰ (usually transcribed as a lowered ɣ̞), and gheada includes velar versions (found in Galicia, and maybe also in Asturias, perhaps, in different environments or similar, I don't know. Note someone added that the Asturleonese ḥ, found in different contexts, can be pronounced as [ħ, ʕ, ɦ, x, χ] without a source, see glottal fricative), so I wouldn't be surprised there is also uvular pronunciations of /ɡ/, especially in northern Spain, where these pronunciations are more frequently found (but personally, I wouldn't associate this with a rhotic, as the "French/German R", which might be the reason of IPA using an inverted ʁ for its representation). Therefore, I'm not sure whether /ʁ/ would be accurate for the representation of non-rhotic consonants in Spanish (a bit like ɰ is not popular in Spanish transcriptions, since its relation to [ɯ], which is non-existant in Spanish, except in Japanese-related content, like vocaloids), as it's regarded as either foreign, and for some others, defective (or better said, imperfect, since this could have a negative interpretation). Remember also, Spain (including some parts of Portugal) is an island between guttural France and guttural Portugal. — Jɑuмe ^(dis-me) 10:32, 12 September 2026 (UTC)

Regarding your second question, the "modern" Portuguese R (not sure about the Brazilian version of it), as I call it, might be pronounced with longer duration, since it can be pronounced with a trill (a repeated vibration), remember the Spanish initial r and rr are analysed as a long [with multiple batecs, "beats", or flaps] alveolar /ɾɾ/ (or /ɾː/). But it does not make sense the same analysis for the "modern" European Portuguese r, since initial r and rr are uvular, except for some of the guttural dialects, where [ʁ̆] (or [ʀ̆]) might also replace the alveolar flap, just as Recasens points that the flap can (also) be guttural in Northern Catalonia. Do you have a source for this phenomenon for European Portuguese? — Jɑuмe ^(dis-me) 11:48, 12 September 2026 (UTC)

I asked few weeks ago what shall we do about some duplicates on some listings, like the [voiced] bilabial fricative/approximant, [voiced] dental fricative/approximant, and [voiced] velar fricative/approximant (see Duplicates). What would you do about this? List each language twice (i.e. fricative and approximant), or list them once (as variable pronunciation)? — Jɑuмe ^(dis-me) 10:46, 12 September 2026 (UTC)

[ʁ], not /ʁ/. The phoneme (in Spanish) is still /ɡ/. I'm not talking about a phonological rhotic and frankly I dislike that label a lot. It lumps together way too many different sounds just because they all happen to be written with the letter ⟨r⟩ in one language or another. It's a bit absurd to me. And on the official IPA chart, there's no such label as "rhotic", and that's what I'm going with.

⟨ɰ⟩ is not used in Spanish transcriptions because it represents a non-labialized sound, whereas in "algo", "Santiago" etc. the sound is labialized without becoming [w] which also exists in the language.

Belen Santo (who's from Vitoria-Gasteiz) definitely uses [ʁ] instead of [ɣ] at least some of the time. Her "[ɣ]" definitely sounds uvular at times (or even more often than not, I don't know) and definitely has some friction, even if not a lot. I don't know how to phonetically represent this better than [ʁ] which represents a sound that is voiced, uvular and a fricative. Its presence in a given Spanish accent/idiolect probably has a lot to do with whether /x/ is uvular, or in a free variation between uvular and velar. It'd be a leveling of POA of two very similar fricatives that differ by voicing and possibly by the amount of frication (which is nothing new or strange - see fortis and lenis in many Germanic languages). No [ɣ]-[ʁ] contrast needs to be maintained in Spanish, whereas in Portuguese the situation is very different and you need to make sure that your [ɣ]'s are velar (so this is a potential source of mispronunciations for learners of Portuguese of Spanish origin).

/ɾː/ makes no sense as a transcription. What would come out is [dː], i.e. [d̚d] - so not a rhotic at all (not that ⟨ɾ⟩ signifies a rhotic in IPA - think of /d/ flapping in English and Danish). You could transcribe it /rː/ and clarify in prose that /r/ has only one contact. This used to be a pretty common transcription, AFAIK.

The question was regarding Spanish [ʁ] (the uvular replacement for [ɣ]) vs. Portuguese [ʁ]. I'm sure we could find minimal pairs that sound almost the same, especially when you replace the last vowel ([a e o] vs [ɐ i u]), given the most likely syllable structure would be /ˈ(C)VCV/ (so [ˈ(C)VʁV]), with a reduced final vowel in Portuguese but not in Spanish. I'm really curious what the Portuguese would have to say about this [ʁ] realization of /ɡ/ in that context, and whether phonetic drills would have to be invented so that the Spanish don't "pronounce their G's as R's" when speaking Portuguese. Surely it would be a challenge to learn to control that.

The duplicates need to be listed once in the "variable" section, if we have that. It makes little sense to do otherwise. Sol505000 (talk) 14:32, 14 September 2026 (UTC)

If it's part of /ɡ/ it must be a type of gheada that is found outside of Galicia, I don't know other name to call this phenomenon. I've heard [χ] in the coda (due to final devoicing) in words like blog and Peppa Pig, but I haven't heard [ʁ] since ⟨ɣ⟩ is more frequently heard as an approximant than a fricative (in these days) in the environment you describe, but I don't doubt its existence and your assertions. Yes, definitely. It'd be good to hear the difference between that Spanish ⟨g⟩ and the Portuguese ⟨r⟩. Regarding /ɐ/, isn't it closer to [ə] in European Portuguese? In Valencian, at least, /a/ is near-open in many cases, same as /ɛ, ɔ/ (which are also transcribed as [æ, ɒ]), and I'd say there is also a tendency to use [ɐ] in Castilian, too. Mostly as an allophone of a slightly more open /a/. — Jɑuмe ^(dis-me) 18:32, 14 September 2026 (UTC)

Ok, I agree with you. I could add a section with variable pronunciation, it's not a problem. — Jɑuмe ^(dis-me) 18:32, 14 September 2026 (UTC)

It's not gheada because it appears only as a variant of [ɣ] in highly specific phonetic contexts (in contact with /o/ and possibly somewhere else). Gheada is more general, as far as I know.

It's hard to hear the difference, to be honest. You'd need to analyze the formants each time to be absolutely sure whether what you're hearing is a fricative or approximant. This is why the symbols ⟨β ð ɣ⟩ are perfectly sufficient and you don't need the lowering diacritic. The IPA specifies those sounds as "fricatives", rather than "fortis fricatives" or "strongly fricated fricatives" or similarly. Same goes for ⟨ʝ⟩ or ⟨ʁ⟩. It's actually perfectly logical to associate voiced fricatives of this type with the lenes of Germanic languages, with the caveat that those "lenes" in Spanish are even weaker phonetically, tending towards approximants. And in Spanish, "lenis" could only ever be used in the phonetic sense, whereas in Germanic languages it's much more of a phonological classification (sometimes being phonetically perverse, e.g. Dutch /v/ is phonetically fortis, not lenis).

The fricatives are more characteristic of deliberate speech, but I also hear a length difference there. "Algo" pronounced deliberately seems to have a slightly longer [ɣ] or [ʁ] (whichever it is) in addition to it being stronger/more fricative. That can also be a factor that helps you decide whether you hear an approximant or a fricative.

I was going for a more broad transcription but yes, final ⟨a⟩ is much higher in Portuguese. I'm not sure whether it's always mid central, I think [ɐ] is possible in stressed positions. Sol505000 (talk) 19:22, 14 September 2026 (UTC)

OK, I understand, so it's a sort of less frequent allophonic variant. I've never heard it in normal contexts, but I've heard some people who merge /r/ [ʁ] and /ɡ/ [ɣ], which occurs in speakers who find difficult or are not used to pronounce /r/, but this is not restricted like you say, it's just a general merger that occurs in most of the cases, like gheada.

As you already know, there are some compromises in the Standard Portuguese transcriptions, like the usage of /ʁ/ for all Brazilian dialects, and /ɐ/ in Portugal. You are probably right, it's more open in stressed position, not sure if it is [ɜ] or lower and if it's more rounded like in Mallorca. — Jɑuмe ^(dis-me) 20:15, 14 September 2026 (UTC)

Yeah, it's like the Ich- vs. Ach-Laut distinction in German, but voiced and more back. A pseudo-minimal pair from southern Dutch would be liggen [ˈlɪʝə] 'to lie' (with a post-palatal or pre-velar voiced fricative) vs. roggen [ˈrɔɣə] 'stingrays' (with a velar one). It's closer, but it still doesn't match the POA in Spanish, which is more back - the more front allophone is velar, whereas the backer allophone is uvular. I believe you can hear the same type of allophony when it comes to /x/, at least in some people (for others it sounds like the default sound is [χ] which is variably fronted to [x]. I think this gets parodied overseas as much as distinción or the retracted /s/.) I think we're not yet at the point in which [ʁ] becomes the default lenited allophone and [ɣ] only its less frequent variant. It could happen, as in Luxembourgish.

Voiced uvular fricative lists Spanish as spoken in the Asturias as having [ʁ] as the most common allophone of /ɡ/, which is a bit surprising. I think we definitely should put this in this article!

That merger with people with rhotacism is perfectly possible and I'd even say expected. The two approximants are really close to each other. I'd even say that the cause of the merger could be that they hear people around them using both for their lenited /ɡ/, causing even more confusion. After all - textbooks don't say that [ɣ] can really signify [ʁ] in European Spanish. They're just outdated or idealistic, making work harder for people like speech pathologists or language teachers. They need to rely on their ears alone for such errors, which should not be the case.

Portuguese transcriptions are much older than the IPA chart on which [ɜ] is specified as open-mid. Before that, it was just a "variant of [ə]". And ⟨ə⟩ is not used for Portuguese /ɐ/ because it was already chosen for what we commonly transcribe with ⟨ɨ⟩. pegar was transcribed [pəˈɣaɾ]. [pəˈɣar] or [pəˈɡar] (depending on the narrowness and conventions), rather than the modern [pɨˈɣaɾ]. Its closeness was specified in prose as very high and ⟨ɨ⟩-like, or not commented on at all. Sol505000 (talk) 08:04, 16 September 2026 (UTC)

Yes, definitely, I didn't expect to see a source for Spanish as spoken in Asturias having this sound. But is it phonemic? It seem for certain speakers, including Valencian, /ɣ/ (approximant) might be phonemic, but unlike Asturian, it does not become uvular in Valencia. Yes, let's add it to this article :)

It makes sense what you say about Portuguese, many of Wikipedia's IPA (like Catalan, Portuguese) guides are "conventional", rather than phonetic. If they were narrower, we'd probably have to use different diacritics. — Jɑuмe ^(dis-me) 19:19, 16 September 2026 (UTC)