Talk:Spanish phonology

From Wikipedia, the free encyclopedia

Spain Portal
This article is within the scope of WikiProject Spain, an attempt to build a comprehensive and detailed guide to Spain on Wikipedia. If you would like to participate, please join the project.
??? This article has not yet received a rating on the quality scale.
??? This article has not yet received a rating on the importance scale.
This article is part of WikiProject Phonetics, an attempt to build a comprehensive and detailed guide to phonetics and descriptive phonology on Wikipedia. If you would like to participate, you can edit the article attached to this page, or visit the project page, where you can join the project and see a list of open tasks.
Other languages WikiProject Echo has identified Spanish phonology as a foreign language featured article. You may be able to improve this article with information from the Hungarian language Wikipedia.

Contents

[edit] Phonetic transcription of "muy"

Isn't muy' better described as /mwi/ rather than /muj/?

No, because it's clearly a descending diphthong, /muj/ (one of the few instances of this), unlike, say, ruido (/'rwido/). --Pablo D. Flores 11:32, 2 May 2005 (UTC)
No, when emphasized it gets lengthened into "muuuuuuuuy" and when clipped it becomes "mu'". This would make no sense if the peak of the diphthong were on the /i/. Thus, it is clearly [uj] (phonologically /úi/) and not [wi] (phonologically /uí/). Uaxuctum 04:42, 7 May 2005 (UTC)

[edit] Confusion over rhotic and nasal archiphonemes, and about the sound of Castilian "ch"

I think there is a lot of confusion in the article regarding the two rhotics and the three nasals. Those only contrast in certain positions (intervocalic for the rhotic, non-syllable-final for the nasals); in other cases, they get merged into the archiphonemes /R/ and /N/, whose allophonic pronunciation depends on the context. For example, the rhotic archiphoneme /R/ is pronounced as a tense trill (like that of "perro") when syllable-initial (such as in "ratón" /Ra.tóN/, "subrayó" /sub.Ra.yó/), and as a tap (like that of "pero") in intra-cluster ("sabroso" /sa.bRó.so/) and in syllable-final ("arder" /aR.déR/) positions (in the latter case, it may also sound as a very lax trill). The nasal archiphoneme /N/, for its part, is pronounced a homorganic nasal in preconsonantal position, and as a mere nasal release when absolute final (though it may become a default alveolar for emphasis).

As for the sound of Castilian "ch", I can attest from my own speech that it is indeed an affricate and not a plosive, but instead of the typical postalveolar [ʧ], it is an alveolopalatal [ʨ] or palatal [cç͡], which is why it may sound similar to a [tʲ]. Uaxuctum 04:42, 7 May 2005 (UTC)

I'm not too familiar with Spanish phonology, but I'm a bit skeptical that you actually pronolunce it with a [ç]. That's the /x/-realization used before frontal vowels in German and is quite different from anything I've ever heard in Spanish.
There are samples of how the voiceless palatoalveolar and palatal fricatives sound in a neutral context here and here if you want to compare. [ʨ͡] sounds more likely to me, though I'm not going to try to sound authoritive about a subject I don't have intimate knowledge about.
Peter Isotalo 00:52, May 25, 2005 (UTC)
I mentioned a palatal affricate, not a fricative ([cç͡], that's the plosive [c] followed by the fricative [ç] and joined by a tie bar; not [ç] alone). Uaxuctum 06:05, 15 Jun 2005 (UTC)
The fricative/affricate wasn't really the issue. What I was questioning was the presence of a [ç] in any form as I'm very familiar with the sound in German and Japanese, but can't recall ever hearing it in any variant of Spanish. In fact, I have a small section on Spanish phonology in Svenskans fonologi by Claes Garlén (mainly a book on Swedish phonology, but which also includes many comparisons to other languages) that seems to describe it as a [ʧ]. There is, however, a mention of [ç] as an allophone of /x/ before /i/ and /e/ in some (unspecified) South American variants. Could you by any chance provide some sort of reference to support your claim?
Peter Isotalo 16:52, 25 July 2005 (UTC)
I am a native speaker of Castilian Spanish. I support my claim in that I am describing my own speech. Castilian Spanish "ch" is most definitely not the postalveolar [ʧ]. I can perfectly recognize that sound and it is not the one I regularly pronounce and hear pronounced for Castilian Spanish "ch". Instead, I clearly notice that postalveolar sound in the accent of Italians speaking Spanish. It sounds distinctly "harsher" than our Castilian Spanish "ch". Our "ch" is usually the palatoalveolar [ʨ], almost identical to the one in Japanese "chi", but I notice that before back vowels (as in "chu") it tends to be pronounced more backwards, nearly or fully as a palatal. It's not surprising that you did not recognize [ç], simply because it isn't really there. The affricate [cç͡], despite its notation, sounds distinctly different from both the fricative [ç] and the plosive [c]. To clearly perceive the characteristic accoustic effect of a fricative, you need it to last for a certain amount of time, longer than what it lasts as part of an affricate. Likewise, many people would find it hard to recognize there's a "sh" within "ch". The phonetic notation, thus, is somewhat misleading. Uaxuctum 21:40, 27 July 2005 (UTC)
If it doesn't sound like [cç͡], it might actually not be one. How about convincing us with a recording or a reference?
Peter Isotalo 11:19, 31 July 2005 (UTC)
I didn't say it doesn't sound like the affricate [cç͡], I said it doesn't sound like the fricative [ç] pronounced alone, which is quite a different statement. I even went into length explaining what I meant and likening it to the fact that many do not immediately recognize there's a "sh" within "ch". I'll make recordings of my own speech as soon as I can (which will take a while because I need to buy a new microphone and OTOH I'm now on holidays away from home). Nonetheless I'm starting to feel rather annoyed that you stubbornly insist in distrusting someone knowledgeable in phonetics describing the pronunciation of his native dialect. Meanwhile, you can listen carefully to this recording I've found on the web "vino a echarlo". Uaxuctum 09:59, 1 August 2005 (UTC)
You can get a second opinion if you want, but I hear a [ʨ] or [ʧ] and I'm leaning towards the latter. If you really want to confirm this, you should consult a Spanish phonology and quote it here.
Peter Isotalo 22:31, 13 August 2005 (UTC)
I listened to that recording, and it sounds like [ʨ] to me; it is quite similar to the sound written 'j' in Chinese. A few points of reference that may help clear this up:
  • The Spanish/Italian difference could involve aspiration as well as place. The Spanish recording I heard has almost no aspiration, which makes the fricative portion extremely short. If there is aspiration, that will lengthen the frication, which can account for some of the heaviness; this is the case in English, e.g.
  • The sound of [cç͡] is not a sibilant, whereas [ʨ] is. Uaxuctum, check whether your teeth are together when saying 'ch'; if so, it's probably the latter, if not, the former.
Benwing 03:25, 15 August 2005 (UTC)
The sound in that recording is actually halfway between alveolopalatal and palatal, because in that word "echarlo" the "ch" happens to be preceded by vowel "e", which makes the "ch" slightly more front than it would be otherwise. The range of allophones for Castilian "ch" is a continuum going all the way from [ʨ] to [cç͡], and the exact pronunciation will depend on the phonetic environment, particularly the degree of frontness or backness of surrounding vowels, and may also fluctuate in the speech of the same speaker (you usually get the extremes of the allophonic range in words with all-front or all-back vowels like "chichi" and "uchu" respectively, and the difference in sibilance is quite distinguishable between the high-pitched sound you get in the former and the 'flat', non-sibilant one in the latter). When I can, I will record myself saying words with only back vowels like "mucho", where the sound is more clearly palatal and non-sibilant. I repeat I'm a native speaker of the Castilian dialect, which means I've been pronouncing and hearing pronounced the sound I'm describing during all my life, and I'm well acquainted with phonetics to be sure what I'm talking about. Also, I'm directly describing the live speech coming out of my own mouth and not merely the semidistorted sound one hears in a recording. Thus, I can pinpoint my tongue into the exact point of articulation of Castilian "ch" with natural ease, and then experiment with it to see what sound it turns into when I progressively make it more front or back, or more plosive or fricative, than it should be, which indirectly confirms me what the actual point of articulation of my native "ch" is. If I stop the pronunciation of "mucho" before the fricative part, I get the palatal plosive [c] (very recognizably, I can feel the top of my tongue pushing against the front of the soft palate in the same position as when I'm about to pronounce Spanish voiced palatal plosive/affricate/fricative "y"), and if I lenite the affricate into a fricative, what I get is not [ɕ] (as in Japanese "shi"), let alone [ʃ] (as in English "sh"), but IPA [ç] (note that the German ich sound is actually more of a palatovelar than purely a palatal, so don't take it as a model for the sound of IPA [ç]). OTOH, both standard Italian and Castilian "ch" are most definitely unaspirated (aspiration in Spanish only occurs in certain contexts in Andalusian and closely related dialects, particularly when a syllable-final [s], turned [h], precedes a plosive as in "hasta" [ˈast̪a] > [ˈɑht̪a] > [ˈɑ̤t̪ːʰa], but Andalusian pronunciation is starkly different from that of Castilian in many aspects). And I repeat Castilian "ch" is most definitely not postalveolar [ʧ] and is never pronounced that way in this, my native dialect. I can instantly recognize the postalveolar being pronounced not only in Italian, but also for example in Mexican Spanish, and I can perfectly notice and reproduce the different point of articulation of postalveolar English "ch" and "j" with respect to Castilian Spanish "ch". Uaxuctum 11:47, 9 September 2005 (UTC)
My dialect (Rioplatense Spanish) has that, /xi/ = [çi] (/xe/ is not [xe] but it isn't so markedly palatal either). I believe most dialects have that; one of the reasons that the Spanish of Spain sounds distinctly harsh to my ears is the preservation of the guttural character of /x/ before front vowels. --Pablo D. Flores 17:53, 25 July 2005 (UTC)
In our Spanish of Spain, "j" is usually velar [x] before front vowels and uvular [χ] before back vowels. Uaxuctum 21:40, 27 July 2005 (UTC)
You're mistaken, "j" in Spanish is normally a postvelar sound (before front vowels), and not at all a velar one. Before back vowels, it tends to be uvular like you said. There's plenty of bibliography about this and many aspects of the Spanish language both Latin American and Castilian from very well-known linguists, starting with Tomás Navarro Tomás from the "escuela cervantista" to many other modern experts. You should back up your arguments with a citation from one of these sources and not rely on your intuitions, which may very well be wrong. You can use spectrograms to backup your arguments, some of which I know will not stand.
Regarding the sound of "ch," I agree with you in that it is more to the front of the mouth in Spanish than in English or Italian, but don't make generalizations for other dialects of Spanish. Many dialects of Latin American Spanish have this fronted "ch" sound, and as a matter of fact, it is far more common than the Italian one, which also exists. I have evidence for this taken from palatograms that I've performed on many speakers as part of my ongoing research, as well as spectrograms.
Regarding your comment about the German "ch" being more palato-velar, it is wrong! I suggest a citation if you want your statement to stand as true, for the vast bibliography on the subject says otherwise. --KY
It may not be quite wrong to say that the ich-Laut is palatovelar rather than palatal. If you listen closely (much more closely than necessary to be understood, however), the German [x ~ ç] continuum phoneme has a lot more than two realizations. While clearly closer to [ç] than to [x], the actual sound in ich /ɪx/ is not quite as fronted as the one in Mönch /mœnx/, which in turn is a little less fronted than the extreme which occurs in Milch (because /l/ is laminal-alveolar and has a very restricted allophony range). BTW, the allophony continues in the other direction as well; behind /uː/ (Buch) it reaches the Russian /x/ which seems to be what is called "postvelar". Or maybe pharyngealized or something. (Not uvular, however.)
While I've never heard, or heard of, a palatal affricate in Spanish (I've heard more Mandarin than Spanish…), I can imagine that it exists, considering how fronted the ch I've heard so far is. Such sounds exist in Mandarin. Pinyin x is probably best described as the voiceless palatal sibilant. (There's no IPA symbol for it, but [ç͡ɕ] or [ɧ̟] are probably appropriate enough.) Pinyin q is the aspirated affricate of [c] and x, and j is the finally voiced (and unaspirated) counterpart of q.
(And no, x is most emphatically not [ɕ], even though almost everyone transcribes it that way. It fills the space between [ç], [ɕ] and [sʲ], but almost never reaches any of these 3 extremes, even though different speakers don't pronounce it exactly the same. Citable source: [well… to be added later when I find it. Thought I'd find it faster. It's the website of someone who has been living in China for years, and linked to from some Wikipedia page on Chinese.])
David Marjanović | native speaker of German, armchair linguist | david.marjanovic_at_gmx.at | 19:20 CEST | 2006/5/14
I agree with you, but when it comes to the allophonic description of a language, we have to be aware of the fact that allophone are always going to be part of a spectrum, especially due to its neighboring sounds, which affect its formants. In that case, the most common and stable (articulatorily and acoustically) allophone must be chosen.--KY
I just had a discussion over at Talk:Irish phonology concerning palatalized velar fricatives, and just today I saw a mention of the realization of Mexico as [mexjiko]. I think this might be a description of the /x/-allophone that is more accurate than [ç].
Peter Isotalo 21:44, 2 September 2005 (UTC)

[edit] epenthetic g in huevo?

This article claims that there is an epenthetic g in the beginning of huevo. I know this is true for some accents, but for all of them? I doubt that.

Benwing 2 July 2005 23:35 (UTC)

In the most careful pronunciations, that word may indeed be pronounced ['weβo], but I think that virtually all natural accents have ['ɣweβo]. — Chameleon 2 July 2005 23:58 (UTC)


I have not heard it in Valencia or Aragon - but my experience of other accents (except Madrileño) is limited.

I'm not sure about the view that a hard 'g' in the middle of a word becomes an an approximant - I thought, and my Oxford and Collins dictionaries agree, that it is a voiced velar fricative (ɣ) - but I am not an expert on phonetics and they may be rather prescriptive rather than descriptive. User:Wee Jimmy 29 August 2005 22:32 (UTC)

Generally, they seem to be only slightly fricated or not at all. The velar fricative symbol is very often used, but they're usually explained to represent approximants and the recordings I've heard (with spectrograms) support this.
Peter Isotalo 21:36, 2 September 2005 (UTC)

kit In Puertorriqueño it's definitely [weβo], but the g must have been around for quite some time, as even my Sefardi speaking grandmother from turkey says [gwevu]! It must have been in use by 1500.

They are approximants or very slight fricatives, as mentioned above. As for the epenthetic g, I speak Rioplatense and it's definitely there (as a velar approximant) in un huevo. Moreover, there's also something epenthetic and palatal going on when I say un hielo. --Pablo D. Flores (Talk) 02:20, 7 October 2005 (UTC)

[edit] s/z aspiration

In my opinion all the s entry should be completely remade. It speaks about the s aspiration as if it were a group of independent, isolated phenomenons (Colombian Caribe, Madrid, Southern Spain) when in fact it is one and the same thing, with slightly different results. Please someone with better knowledge than me do it. Jotam 01:24, 11 January 2006 (UTC)

[edit] The trill is a phoneme?

I don't think the trill is a phoneme. The trill has a more limited distribution than the tap, and they share similar phonetic properties. The following two rules summarize the trill's production in terms of the tap, giving a more parsimonious phoneme space with just /ɾ/:

ɾ-STRENGTHENING: /ɾ/ --> r / #_ or C._
TAP DELETION: ɾ --> Ø / _r

So, a word like ron would derive as follows:

/ɾon/ --> (ɾ-STRENGTHENING) ron --> [ron]

A word like carro, which contrasts minimally with caro, would have two taps at the phonemic level:

/kaɾɾo/ --> (syllabification) kaɾ.ɾo --> (ɾ-STRENGTHENING) kaɾ.ro --> (TAP DELETION) ka.ro --> [ka.ro]

Malandi 10:40, 5 March 2006 (UTC)

Got a source for that? -Iopq 15:36, 2 October 2006 (UTC)
This is a long-made mistake. Spectographic evidence clearly shows that a trill is not a sequence of two taps, but consistently a sequence of 3 or 4 taps average. How can the trill be derived from a tap, then? As a matter of fact, perception experiments have shown that a trill equivalent to 2 taps is perceived as ill-formed by native speakers.--KY
If [tʃ] can be an allophone of /k/, I doubt the difference between two or four taps is too much for allophony. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 20:47, 23 October 2006 (UTC)
But taps and trills are made completely differently on the physical level. Trills use the force of air to move the tongue, while taps are just using the muscles of the tongue to "tap" the roof of the mouth. --Iopq 15:14, 2 December 2006 (UTC)
It's not much a leap anyway. Heck, [ɾ] is an allophone of /t/ in English and those are produced differently. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 21:12, 2 December 2006 (UTC)

Since Iopq mentioned the articulatory difference between trill and tap (there can be one-contact trills), maybe this is the place to bring up my question:

The article states that other than intervocalically, [ɾ] and [r] are "in complimentary distribution with the trill appearing in the word onset and after /l/, /n/, or /s/ and the tap appearing elsewhere." However, I frequently hear trills word finally, and over at talk:IPA someone brought up perla as having a trill as well, albeit a short one. Without a contrast, a trill may be shortened to 1-2 contacts, but isn't it still articulated as a trill? Wouldn't Spanish be better described as having a trill where there is no contrast? kwami (talk) 16:53, 23 January 2008 (UTC)

According to Martínez-Celdrán et al (2003) in Castilian Spanish, "The trill appears in word onset and after [l], [n] and [s] and in intervocalic position where it contrasts with the tap... In other contexts, the tap is usual." (258) You could be hearing the speech of other dialects, but we'd need a source that argues what you're proposing, otherwise it's OR. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 20:01, 23 January 2008 (UTC)

"The tap is usual": I wonder if this means tap and trill are in free variation, if it depends on dialect, or if the trill is often reduced to a single contact. kwami (talk) 20:26, 23 January 2008 (UTC)

I wonder that as well. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 21:34, 23 January 2008 (UTC)

[edit] "F" as voiceless bilabial fricative

I feel sure that I hear a voiceless bilabial fricative for the letter "f", at least in certain contexts ("café" and "Porfirio" but not "fuerte" or "Alfonso"), but I find no mention of this here. Am I mistaken?

[edit] "Assimilation" of /s/

"In parts of southern Spain, the only feature defined for /s/ appears to be voiceless, adjusting point and mode of articulation to the surround. This explains the observed assimilations (/peskao/: [pexkao], /fosforo/: [fofforo]) in Madrid and (/est̪os/: [eht̪oh])."

This isn't any different from what the article says immediately above about /s/ not having a defined place of articulation (being defined only as a voiceless fricative), because none of the "assimilations" cited involves /s/ not being a fricative. Also, the last is not an assimilation, because the place of [h] is glottal, not dental. It's simply a lenition, very common in many Spanish dialects and already mentioned in the previous paragraph.

This section should be deleted, and the examples (if they are valid -- I can't say because I have no experience with Madrid Spanish) should be retained as examples of the Madrileño process.

--Armchairlinguist 00:44, 2 December 2006 (UTC)

Actually, [h] is phonetically placeless (so it's not lenition, it's debuccalization). It is not a fricative in the strict sense of the word so Southern Spain contrasts with Madrid in not requiring /s/ to be a fricative. The grammar of that paragraph is pretty sketchy though, it could use some retouching. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 01:06, 2 December 2006 (UTC)
Oops, sorry, you're right for this case, although I don't see how someone not experienced in this particular distinction can immediately understand that the [h] referenced here is a placeless non-continuant, not the glottal fricative of the IPA. It seems like this needs to be simplified a bit or explained better. --67.109.56.162 00:36, 5 December 2006 (UTC)

I'm now wondering why it has changed "it may lose its oral articulation entirely to become an aspiration (usually transcribed as /h/)" to "become /h/". Since /h/ in IPA is the symbol for a glottal fricative, if it's not really a glottal fricative (phonetically placeless, as claimed) then it's not accurate to say it's really /h/. I'd like to see this changed back. Likewise I am not really a fan of the wording change made about /z/. My choice of "voiced or lightly voiced /z/" meant that it might change to become fully voiced, which would be written /z/, or lightly voiced but still written /z/. "Change to become voiced or a lightly voiced /z/" makes it sound like vooiced doesn't mean /z/ when it actually does.

--Armchairlinguist 09:45, 10 December 2006 (UTC)

I think you misunderstand. The sound [h] is phonetically placeless. Its place in the consonant chart as glottal and as a fricative is due to the limitations of the chart. Likewise, the only difference between [z] and [s] is voicing and when one wants to represent an alveolar sibilant that is partially voiced, either symbol can be used [s̬] [z̥]. I'll reword it though. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 19:12, 10 December 2006 (UTC)

-- The examples are not valid for Madrileño Spanish. I live here and the s is not voiceless or fricative at all (maybe for some specific parts of Andalucía, but I can't say how common that it). That should be removed.Oconel 12:15, 20 January 2007 (UTC)

What is it instead? And do you have a source? Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 21:08, 20 January 2007 (UTC)
As you can see here, aspiration of the /s/ only happens in the south (you have to scroll down a bit) http://www.proel.org/lenguas.html Oconel 13:35, 24 January 2007 (UTC)
The question of the Madrilenian dialect is contentious. People from Madrid have been taught that they speak 'dialecto castellano', this is part of their ideological identity. But the language really spoken in Madrid has, and has always had, lots of southern traits. Madrilenians often deny this fact and cling to what they were told in school. Jotam 01:12, 24 January 2007 (UTC)
Well, since Madrid is in the middle of Castille it makes sense that they speak 'dialecto castellano'. Please specify what southern traits are those. As far as I know it's characterised by 'laísmo' and 'lleísmo'Oconel 13:35, 24 January 2007 (UTC)
It would make a lot of sense that madrilenians spoke 'dialecto castellano', if such thing existed. But it is just a political construct. Anyway, Oconel is quite right in pointing out that the /s/ aspiration is the main marker of dialectal boundaries for Spanish. This is a map with the /s/ aspirating areas in Spain : http://jotamartin.byethost33.com/alpi_aspira_e.php . Madrilenians aspirate their s's just like all southerners, but also they (and many other spaniards) are taught in school to repress their natural pronunciation and adopt another one based in 'pronouncing everything exactly as it is written'. Jotam 23:08, 24 January 2007 (UTC)
So, we have two sources with completely opposite results (and as you can see, Madrid is not at all in the south). Anyway, if Madrilenians are taught to "repress their natural pronunciation" and they don't aspire the s, doesn't it mean "they don't aspire the s" whatever the reason?. Why are they able to tell when someone aspire the s if they aspire it themselves as you said? Also, if that were due to schooling, people from the south of Spain wouldn't have seseo or ceceo. Just go to Madrid and hear people talk.

Another link (and I hope the RAE is good for you) http://buscon.rae.es/dpdI/SrvltGUIBusDPD?lema=s (it says the s is only aspired in Andalusia, Extremadura, Murcia and Canary Islands) Oconel 12:49, 28 January 2007 (UTC)

I don't think we have opposite results. The www.proel.org map looks very conservative, it probably uses a rather contrived criterion to make sure that the southern area is as small as possible. Besides, that site doesn't explain how the line between northern and southern dialects has been drawn. Keep in mind that no studies about madrilenian pronunciation exist, as far as I know, and I've even read otherwise serious articles where the fact that Madrid speaks 'standard Spanish' is simply taken for granted. I'd almost say that stating that Madrid speaks 'southern' is political taboo in Spain. Jotam 23:27, 28 January 2007 (UTC)
Moreover, when I said that many people in Spain are taught to repress their pronunciation, I didn't imply that they are successful in doing so. I reckon madrilenians aspirate some 10-20% of s's in very formal registers, some 40-50% in 'intermediate' registers, some 80-90% in relaxed speech. This a generalisation, just for foreigners to get an idea. Besides, probably most madrilenians are not conscious of how often they drop their s's, as most people they meet speak just like them. I think we can use the word Diglossia for this situation. Jotam 23:27, 28 January 2007 (UTC)
About going to Madrid : I often do. That's why I'm so sure about what I say. Jotam 23:27, 28 January 2007 (UTC)
About the RAE link : NO, the RAE is not a good source of information about dialectal differences. The mere fact that /s/ aspiration is confined to exact administrative boundaries speaks for itself. Jotam 23:27, 28 January 2007 (UTC)

Just a note that the map Jotam provided is from the ALPI, which was a linguistic survey of the Iberian Peninsula conducted in the 1930s. So although it's still a valuable source of information, it may not be an accurate picture of the modern situation. But I'm not from Spain, so I can't comment on the validity of it either way. --Miskwito 18:50, 28 January 2007 (UTC)

You're right, language is changing quickly in Spain these days. And overall the change can be described as a 'madridization' of the language, due to most media being based in Madrid. Another reason why madrilenians can't hear the differences in other dialects, many people are adopting their speech. Jotam 23:27, 28 January 2007 (UTC)
So when a friend and I were told in Bilbao that we didn't say "ej que" (/s/ aspiration for "es que"), what happened was that in the north they are starting to aspire the /s/ too, and they can't see the difference anymore? Are you sure you talk to madrileños when you go to Madrid? Also, what part of Madrid you visit (maybe there are differences I'm not aware of between the different towns)?Oconel 21:58, 3 February 2007 (UTC)
In 1930 the population of Madrid was 1.290.445, nowadays is 5.964.143. Since Jotam doesn't like the links I provide and insist everything is biased, there's very little I can do to prove my point. Also, some people think Madrilenian speech is posh speech (similar to a drawl and which doesn't seem to be the case here) and others think it's what could be called macarra (hoodlum according to my dictionary) who do aspire the s, and I'm starting to wonder if that's where Jotam got his idea. None of those Spanish is the average in Madrid.Oconel 22:18, 4 February 2007 (UTC)
That the RAE source is biased is absolutely obvious, and very easy to prove too. That the www.proel.org map is biased is impossible to prove, because we don't know how that north/south line has been drawn. The people I usually meet in Madrid are middle-class, from all districts in the city and most of them have university degrees. When they meet a stranger, they adopt a pronunciation based on 'written spanish' (not on northern spanish), similar to TV newscasters. But I think we can all agree that the question is : how do they speak with friends and family ? I found out long ago. Jotam 11:34, 5 February 2007 (UTC)
I'd like Oconel to answer this : to work in sports programmes on the radio, is it compulsory to be a macarra ? Jotam 11:34, 5 February 2007 (UTC)
So, Madrilenians only aspire the /s/ at home and not in front of you? Right. Then why doesn't the acting hide the yeísmo or the laísmo? Or how did you manage to find out the "secret"? Don't answer, I'm not sure I want to read whatever new theory you're going to make up next. I give up, let's have the Wikipedia say they do aspire the /s/ and if you want to add something else, feel free to do it.
I haven't listened to a football radio program in a long time, but I've caught glimpses of those so called "debates" on TV, and yes, there are always a couple of macarras.Oconel 07:31, 8 February 2007 (UTC)

I'm pretty sure that most of what is applied to "Southern Spain" on this page really only applies to certain areas such as Cadiz. Could we maybe be more specific as to what pronunciation applies to which area? I know a lot of people tend to lump all southern Spaniards (rather, Andalusians) together as having the same manner of speech, but this is far from true.

[edit] allophone tables set up sloppily

The "Sound" column in the allophone tables is misleading and reflects a fundamentally incorrect understanding of phonology (or else the tables are just sloppily set up). A phoneme represents an abstraction of sounds, so to speak. For example, the Spanish phoneme /b/ comprehends the plosive [b] and the approximant [β̞] in their respective environments. The phoneme /b/ is therefore not a "voiced bilabial plosive" as the table would indicate; this PHONETIC term refers to [b], not /b/. If such technical phonetic terms need to be included, they should be incorporated into the text of the Allophone column, or the phoneme-rows with two or more allophones could be divided into subrows for each allophone, and the "Sound" description added for each allophone. Tawagoto 02:41, 16 April 2007 (UTC)

Most sources consider the plosives to be the underlying phonemes. This can be disputed, but it's a reasonable and traditional analysis. Notes on the allophones and their manner of articulation are included in the sound description. I don't see any need for more than this. --Armchairlinguist 23:29, 16 April 2007 (UTC)
Nevertheless, /b/ simply isn't a "voiced bilabial plosive". Being misleading and inaccurate, the "Sound" column isn't necessary. I would recommend deleting it. Tawagoto 01:18, 17 April 2007 (UTC)
I think a better way to set this up would be to list the phoneme column with the underlying phonemes, then an allophone column with the sound symbols and descriptions, then spelling, then change the name of the allophone column to Distribution. So you'd have
Phoneme: /b/
Allophones: [b], voiced bilabial plosive (newline) [β̞], voiced bilabial approximant
Spelling: "b" burro (newline) "v" vaca
Distribution: [b] appears initially (in some accents) and after nasals (bomba, envidia), approximant [β̞] elsewhere (nube, la bodega). In rapid speech, [β̞] can replace [b] in the initial position. After /l/, there is variation among speakers (el burro can be either [elˈburo] or [elˈβ̞uro]).
This would make things clearer. --Armchairlinguist 00:10, 11 May 2007 (UTC)
The nasal table seems to be quite sloppy. What are the numbers (1), (2), (3) for? The nasal archiphoneme is called /N/ in the section on /m/, but nowhere else. And somewhat contradictorily, various allophones are listed in the section on /n/ that (it seems to me at least) would more properly belong to the archiphoneme than to /n/. (And does réquiem really have [ɛ]?) 71.82.214.160 23:24, 6 May 2007 (UTC)
I worked on the table per my suggestion above. I think it's much improved but perhaps still needs some further work. I removed the following from the list of nasal allophones:

/ʎ/ (conllevar, un llavero); /x/ (enjambre, un jarro) or the semivowel /w/ (enhuesar, un huevo, but not nuevo)

These realizations may occur, but they are not nasal, so it would require a lot of explaining as to why they can be considered part of a nasal archiphoneme. If someone else wants to do that explanation, awesome. I don't feel adequate to the task because it would require me to check out some real sources for a good phonological analysis. --Armchairlinguist 04:57, 11 May 2007 (UTC)

[edit] /x/ allophones

I think this page should say much more about different allophones of /x/. In particular, I think it's uvular in some dialects. Also, would it be worth saying that it is spelled <x> in a few Mexican Spanish words derived from Native American languages? 71.82.214.160 23:12, 6 May 2007 (UTC)

[edit] Trill challenged

Should the fact that some people are unabled to correctly pronounce the trill (saying, for example, something like pego instead of perro, slightly sounding like a fake French accent) be mencioned in the article? It is a fairly common difficulty in Spanish. Helios 03:52, 11 May 2007 (UTC)

[edit] Allophones of /k/

My impression is that /k/ is palatalized to a voiceless palatal plosive [c] before front vowels, that is casa ['kasa] but queso ['ceso]. But maybe I just have a Greek accent in Spanish? Comments? --Macrakis 16:48, 21 July 2007 (UTC)

That certainly doesn't happen in any of the Spanish I've heard. I think languages like Spanish that have /tʃ/ resist diachronic palatalization of /k/. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 17:58, 21 July 2007 (UTC)

[edit] Stress

The article says:

Stress falls on the penultimate syllable unless the word ends in a liquid or -z, in which case it falls on the ultimate syllable. Exceptions are marked orthographically with an accent mark over the vowel (i.e, comunicación, Lázaro).

Not exactly. Stress is on penultimate syllable if the word ends in 'n', 's', or a vowel. It's on last syllable if the word ends in any consonant other than 'n' or 's'

penultimate syllable stress: canten, regalo, regalos

last syllable stress: verdad, reloj, cantar, capaz, cartel

Dab14763 06:51, 12 October 2007 (UTC)

According to Hochberg in "Learning Spanish Stress: Developmental and Theoretical Perspectives" (Language, 1988), "Stress on the penultimate syllable of vowel-final words and on the final syllable of consonant-final words is most frequent, and is therefore traditionally considered regular... However, stress can actually fall on any of the last three syllables of a word, regardless of the final segment." (p 684) He has the following table to demonstrate this:
stress Vowel-final words Consonant final words
final mamá tenedór
penultimate cuchára cadáver
antepenultimate teléfono hipótesis
Eddington in "Spanish Stress Assignment within the Analogical Modeling of Language" (Language, 2000) seems to repeat Hochberg: "Stress may fall on any of the last three syllables of a Spanish word. In general, penult stress on vowel-final words is the norm... while consonant-final words with final stress are considered regular... antepenult stress is always considered irregular... since it runs counter to the first two more general tendencies. Preantepenult stress is rare, and occurs only when certain verbal forms are followed by two clitic pronouns (e.g. guardánoselos 'saving them for him/her')." He also states, though, that "penult stress is viewed as the norm for words ending in -s or a vowel, while final stress is considered regular for words ending in all consonants except s." He has this table of stress placement in the most frequent Spanish words:
Word ending Final Stress Penult Stress Antepenult Stress
Vowel 178 2494 178
Consonant 798 1085 96
/s/ 20 909 94
Consonant (except /s/) 778 176 2
He also states that because so much stress contrast is much more evident in verbal forms, morphology is often considered in underlying stress rules. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 07:33, 12 October 2007 (UTC)


"while final stress is considered regular for words ending in all consonants except s."
No, it's all consonants except 'n' or 's'; that's why words ending in 'n' that have stress on the last syllable are marked with an orthographic accent: razón, andén, canción Dab14763 18:52, 12 October 2007 (UTC)
There are so many words that end in n that have an accent marker that that can be considered an orthographic convention. I can't even think of a word that ends in n that isn't stressed on the last syllable... except maybe in verb conjugations. Besides, this is two published sources by presumably reputable linguists in a scholarly journal. If you don't back your claims up there's no reason why we can't go with this explanation. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 19:36, 12 October 2007 (UTC)
'n' is treated differently from other consonants precisely because so many verb conjugations end in 'n' which means that the majority of words that end in 'n' are penultimate stress, as are the majority of words that and in 's' or a vowel. Though there are some words that end in 'n' which are penultimate stress and are not verbs: crimen, virgen, polen, germen, Carmen, canon, orden, origen, abdomen, acumen, examen, volumen, imagen, resumen, etc
http://www.ati.es/gt/lengua-informatica/estilo/acentos.html
http://www.dat.etsit.upm.es/~mmonjas/acentos.html
http://www.estu.info/index.php?topic=1157.0;wap2
http://www.pastranec.net/acentos.htm
etc, all treat final 'n' with 's' and vowels and separate from other consonants for stress purposes Dab14763 05:20, 13 October 2007 (UTC)

The first, second, and last are talking about orthography, the third does indeed say "la mayoría de las palabras que acaban en n o en s son también llanas" (that is, they behave like words ending in vowels) but it (and all of the other sources) seem to be a random website rather than a scholarly source. I'm presuming that this third source is a quote from something published but I have no idea what it is. I'm really looking for linguistic sources. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 08:28, 13 October 2007 (UTC)

These "rules" about stress are orthographic conventions, not phonological facts. They belong in the article Spanish orthography, not here. FilipeS (talk) 01:22, 30 November 2007 (UTC)

[edit] Velar approximant

Before anybody reverts me, this is from Martínez-Celdrán et al:

the symbol [ɰ] is not adequate for Spanish because it represents an unrounded semivowel that requires spread lips. This semivowel does not exist in Spanish. There are plenty of words in Spanish with a velar approximant as in la gorra 'the cap', algo 'something', este gusano 'these worm', el guante 'the glove', etc. Before back vowels or the semivowel [w], the sound is rounded through coarticulation. Therefore it is impossible to pronounce Spanish algo as *[ˈalɰo], [ˈalɣ̞o] being more phonetically feasible because the approximant [ɣ̞] is not marked for the rounded/unrounded feature, like the approximants [β̞] and [ð̞].

Because the character <ɣ> already has a descending part that would obscure diacritics, I've removed the lowering diacritic from [β̞] and [ð̞] as well to be consistent (and not redundant). Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 20:54, 9 November 2007 (UTC)

Sorry, but Martínez-Celdrán is nonsense. ɰ is no more specified for rounding than ɣ is. Either ɰʷ or ɣ˕ʷ could be used for w (unless you want to argue that /w/ is closer to [β̞ˠ] in a particular language). If you want to argue for <ɣ˕>, better to say that it's graphically analogous to β̞ and ð̞, or that all three of these are less than fricatives but more than approximants. I've seen both arguments, but I suspect the main reason is that people are simply more familiar with the fricative symbols, the same reason people stick to ʕ for 'ayin. kwami (talk) 00:58, 24 January 2008 (UTC)
There was also the oddity of writing [θ̬] for [ð]. I could see that for a phonemic description, in an attempt to remain faithful to the orthography, if we also wrote [s̬] for [z], but otherwise it just seems arbitrary. kwami (talk) 01:29, 24 January 2008 (UTC)
Is it really nonsense? Martínez-Celdrán points to Pullum & Ladusaw (1986) (page 98) as one of a number of linguists who argue that ɰ has an "unrounded" feature. I guess if the official IPA description is neutral, we can still go with the "graphically analogous" argument (I like the way you think) and, per extra research into the notability of the position that ɰ is specifically unrounded, velar approximant can mention the issue in greater depth.
As for [θ̬], that's what the source uses and I'm willing to keep it that way since 1) it's used only once, 2) this distinguishes it from the approximant allophone of [ð] (that is, as long as we continue to omit the undertack in general transcription) and 3) may actually indicate some sort of voicing difference (not sure of this). — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 05:55, 24 January 2008 (UTC)
Yes, ɰ is unrounded, but so is ɣ, and so are β and ð, so either they're all acceptable or none are. Sometimes ɰ will be specifically described as unrounded, in order avoid confusion with w, but the other consonants aren't contrastively ambiguous as to rounding the way ə is. I just took a look at P&L, and they only use the word "unrounded" when referring to ɰ being used for a vocoid, since rounding is generally specified for vocoids, but not when referring to it being used for a consonant, as is the case in Spanish. (As a consonant, they are more concerned with laterality and call it a "median approximant consonant"; you wouldn't argue from that that the vowel ɯ was incompatible with the consonant l, because ɰ is defined as median and therefore its syllabic equivalent ɯ can't have lateral assimilation!) Back in the days when ɣ was used for both fricative and approximant (as ʁ still is), you would also have had to describe ɣ as 'a semivowel (glide) with the properties "high," "back," and "unrounded",' but no one objected that that made it somehow inappropriate for Spanish. So it's a non-issue, based on a overly literal reading of some descriptions of the IPA. However, since ɰ is by definition ɣ˕, and all the Spanish voiced obstruents are commonly called fricatives in the lit, that works too.
BTW, I deleted the comment in the intro of velar approximant about ɰ being unrounded, as it's not only obvious from the text, but the case for all consonants bar the specifically rounded ɥ, w, ʍ. It would be like going on about how ʃ and ʒ are non-palatalized postalveolar fricatives just because ɕ and ʑ are palatalized postalveolar fricatives, or about n being a non-syllabic nasal stop, just because some people still use a special symbol for a syllabic nasal stop, about ɪ and ʊ being syllabic … vowels, just because in the English tradition they're also used for the non-syllabic off-glides of diphthongs. Secondary articulation is normally only mentioned when it's present, or when it's somehow notable by its absence, such as when contrasting two IPA letters.
I assume Martínez-Celdrán invented the odd amalgam <θ̬> for just the reason you mention, to avoid confusion with his inaccurate use of fricative symbols for approximants: He dispensed with the diacritic in the one instance, and to make up for it had to artificially introduce one in another, less common instance. Unless he actually states that the voicing of <θ̬> is somehow different than that of <z> (which he should, considering in the journal he's writing in!), I'd prefer we simply say what we mean, and not ask our readers to try to decipher an idiosyncratic orthography. kwami (talk) 06:59, 24 January 2008 (UTC)
Ahh, I've been thoroughly schooled in the ways of approximants, thanks for the detailed response. I assumed that, because Martínez-Celdrán argued as such, that the IPA itself specified "unrounded" for ɰ like it might for a vowel (JIPA occasionally has pieces with linguists advocating new methods of transcription or representing the IPA). As for the dental fricative, let's see, the actual quote is "[θ]... like [s], becomes voiced before a voiced consonant." The voiced allophone of /s/ is represented as [z] so I imagine that he does indeed wish to indicate a difference between the voiced fricative and the voiced approximant. In addition to this being "idosyncratic," the two appear in complimentary distribution anyway so it shouldn't matter with or without the undertacks. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 10:41, 24 January 2008 (UTC)
Sounds like it's just [ð] then. As for [ɰ], there's the added complication of whether it's being used as an approximate consonant or semi-vowel. Not all phoneticians distinguish the two. kwami (talk) 18:28, 24 January 2008 (UTC)

[edit] This article is vague and incorrect in places

I have a problem with some of the claims made in this article. For example, it fails to mention that words ending in /n/ are also accented on the penultimate syllable (I fixed that). Spanish was my first language, and some of what the article says here makes absolutely no sense to me. I have never in my life used [ɟ], [θ], [ð], [ɣ], [ɱ], [ç]. I understand that this article pertains mostly to European Spanish, but nowhere in the intro does it say so, and someone unfamiliar with the language (i.e. someone doing research) may never realize it. I propose that we either create a separate article on Latin American Spanish phonology or seriously revamp this one. I would also like to point out (but have no idea where to put this in the actual article) that in some regions of Spain, /s/ before consonants or at the end of a word is pronounced somewhere between [s] and [ʃ], probably retroflex ([ʂ]). Diego Argueta

We can change the lead section to be clear that this is primarily about Standard or Castilian Spanish but I have some doubts about your knowledge of phonetics/phonology if you assert that you don't use [ð], [ɣ], or [ɱ] (The other consonants you mention are dialectal). The first is the d in cada, the second is the g in diego and the third is the n in ánfora. As far as I know, every dialect of Spanish uses these and certainly any form of Standard Spanish does. We don't need to create a separate article on any other dialects of Spanish because the differences are so minor that they can be accounted for in a single section in this article (this is the "dialectal variation" section) as well as in the Spanish dialects and varieties article. I agree that the portrayal of dialectal variation is pretty poor right now and we can probably find sources to corroborate and expand on that section.
Also, see #Stress above, I've provided some academic sources on how stress works in Spanish. If you dispute the claims then I propose you find other academic sources that argue that /n/ and /s/ are equal exceptions. Otherwise I'll have to revert what constitutes original research that is based, it seems, on an orthographic convention. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 03:30, 28 November 2007 (UTC)


[edit] Stress ortography

Hi. I'm a Spanish native speaker (from Spain) with no particular knowledge of phonology or phonetics, but I need to say that some excerpts of this article are simply wrong when they refer to stress ortography. The following paragraph in particular contains some errors:

  • Stress may be marked orthographically with an acute accent (mantél, distinción, etc). This generally marks exceptions to the rules above but may be redundant (as with the case of distinción). An acute accent may also differentiate homophones (such as té and te) and words which contrast solely on stress such as sabána ('savannah') and sábana ('sheet').

It's not that stress may be marked by the accent. If there's an accent, that's where the stress goes, of course; but some Spanish words do not need an accent at all. For the sake of simplicity, the accentuation system (that is, the way you should place or not the acute accent mark over a vowel) has been designed so that most of the times the accent is not necessary. Thus, this sentence, as it is, is not correct. Besides, the accentuation system, if correctly used, leaves absolutely no doubts about the syllable in which the stress is placed.

These accentuation rules are taught in primary school to all native Spanish speakers, so it will be difficult to find a paper to justify them. They're supposed to be known if you are in University. In the ortography related article at the Spanish Wikipedia the sources are the following: R.A.E., Ortografía de la lengua española, Madrid, Espasa Calpe, 1999. ISBN 8423992500. By the way, a normative text about Spanish orthography.

On the other hand, "mantél" and "sabána" are simply not proper Spanish words. The correct spelling removes the acute accent: "mantel", "sabana".

I understand it is NOT the purpose of this article to deal with orthographic issues, so I'm removing these errors and just linking to the related Wikipedia article. I'd recommend you to have a look at it to know when you should or shouldn't place an accent in a Spanish word. With regards, Cvalda (talk) 20:01, 15 December 2007 (UTC)

Thank you for pointing out some of the errors. I've replaced mantel with ácido as a better example. The problem with your edit is that the orthography differs from the phonology in that it treats n like s, that is, as an exception to the rule that words ending with consonants are stressed on the ultimate syllable. This is a minor but important difference between the phonology and the orthography so that we can't say that stress must be marked when it doesn't fit with the phonological rules laid out. It is, instead, marked when it doesn't fit with the orthographic rules.
People periodically edit this page to pair /n/ with /s/ in this exception but this is not what the sources on the phonology say. I hope I've clarified it for you. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 21:01, 15 December 2007 (UTC)
I understand why you insist, but if rules are different, why not keeping a link to the Orthography article so that people will see both rules? If the article stays as it is, you'll have to revert it every time a Spanish speaker shows around ;) And I think it's relevant to the article. But, anyway, you seem to be keeping this article very tightly, so I leave the choice up to you. Thanks for your cooperation, Cvalda (talk) 19:25, 16 December 2007 (UTC)
Oh I forgot to keep that link. I'll be sure to link to it. Regards. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 21:15, 16 December 2007 (UTC)
Hi again. I consider the following statement:
  • Stress only occurs in one of the last three syllables of a word.
...to be false, because there do exist some words that are sobreesdrújulas. I'd recommend to keep this other one:
  • Stress most often occurs in one of the last three syllables of a word; preantepenultimate stress occurs very rarely and only in words like guardándoselos ('saving them for him/her') where a clitic follows certain verbal forms.
As for the rest of the rewording, it's perfect. Keep up with your excellent work on this. Greetings, Cvalda (talk) 00:07, 17 December 2007 (UTC)
Good catch. I've changed the wording to be closer to that of the source ("Stress may fall on any of the last three syllables of a Spanish word.") Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 01:44, 17 December 2007 (UTC)
Even if most of the words have the stress at some of the last three syllables, please note that the contraexample given at that sentence is true and does exist (as do many other words that are sobreesdrújulas). Thus, even if your source is worth of trust, this particular sentence is just wrong, because a few words have the stress at the fourth last syllable! Even if they're few, this renders the sentence invalid! You can take a look at the Spanish Wikipedia on Accentuation or even Wordreference or any dictionary of your choice to learn more about this subject (assuming you can read Spanish). I'd recommend you to take this into account. But thanks again for your excellent work, I let you do the changes you see fit. Good work! Cvalda (talk) 01:54, 17 December 2007 (UTC)
I guess the sentence by itself would be incorrect, but I think the third bullet point that follows it clarifies this. I don't think people will have the misperception that there aren't words like guardándoselos. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 02:00, 17 December 2007 (UTC)

[edit] This article is all wrong

Has any one else noticed that this article is full of misinformation? But what should you expect from wikipedia? —Preceding unsigned comment added by 207.210.202.237 (talk) 22:41, 26 January 2008 (UTC)

[edit] I know some native Spanish speakers who pronounce the b/v as a voiced labiodental fricative (ie English v)

Perhaps we should include the fact that some native speakers, perhaps in an attempt to be "educated", pronounce both the b and the v in Spanish as a voiced labiodental fricative (English v). —Preceding unsigned comment added by 205.246.153.217 (talk) 04:04, 15 May 2008 (UTC)

If we can find a reliable source that says so, I don't see why not. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 05:05, 15 May 2008 (UTC)

[edit] Trapezium vowel diagram

Hi, can someone add a trapezium to the vowel diagram, please? Thanks. --Kjoonlee 02:12, 26 May 2008 (UTC)

Hmm, that's a toughie. The Journal of the International Phonetic Alphabet has a chart with formant values, but it's not the typical trapezium and I'm not sure how to convert it. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 04:12, 26 May 2008 (UTC)
If you post the table, I'll see what I can do. Maybe the table will work just as well, with some work. --Kjoonlee 06:21, 26 May 2008 (UTC)