1- Pronunciation of /x/
The sound written with J (and G before E and I) can be pronounced a number of different ways depending on accent. In the northern and central part of Spain, this sound is pronounced χ, a voiceless uvular fricative, also found in French in croissant. In Argentina, Uruguay, Peru, central Mexico, it’s pronounced x, a voiceless velar fricative, before /a o u/, and ç, a voiceless palatal fricative, before /e i/. These sounds are also found in German in ach and ich respectively. (You can also find a uvular χ variant in Argentina, especially before back vowels /o u/. Personally, I’m guessing this is either due to the influence of northern-central Spanish immigrants, Arabic speaking immigrants, or both.) In the south of Spain, the Canary Islands, the Caribbean (islands and coasts) and Central America, you find a h pronunciation, a voiceless glottal fricative, like in English hot.
2- Pronunciation of Y/LL
In most accents of Spanish, this pair of spellings is pronounced the same with no difference between them. In some accents, however, they are pronounced differently. For example, they are pronounced differently in Misiones (a northeastern province of Argentina), in some pockets in Spain, and in some Quechua-speaking areas in Bolivia. In the latter two cases, Y is pronounced ɟʝ (a voiced palatal affricate) after a pause and ʝ or j (a voiced palatal fricative or approximant, the latter found in English yet) in all other cases; and LL is pronounced ʎ (a voiced palatal lateral approximant, found in Italian aglio). In Misiones, LL is pronounced the same way, while Y is pronounced dʒ (a voiced palatoalveolar affricate like in English jet).
In accents that do not differentiate between Y and LL, pronunciation of the single sound varies regionally. In some areas, it’s pronounced dʒ. In most of Argentina, it’s pronounced ʒ (a voiced palatoalveolar fricative, like in English vision) or ʃ (a voiceless palatoalveolar fricative, like English show, except with no rounding). This sound can vary a lot even within this same dialect area. It can be rounded, and in the case of the voiced alternative, also fronted (pronounced alveolarly) or pronounced as a retroflex. Uruguay also shares this ʒ / ʃ pronunciation. In the case of the younger speakers in Misiones who are merging Y and LL, they prononce it dʒ. In most of the rest of the Spanish-speaking Americas, and in Spain, it is pronounced ɟʝ after a pause and ʝ or j in all other cases (the approximant j pronunciation is the most common one when the sound is between vowels).
3- Pronunciation of S, C (before E-I) and Z
In northern-central Spain and parts of southern Spain, S on the one hand and C/Z on the other represent two different phonemes. Everywhere else in the Spanish speaking world, S//C/Z represent one single phoneme. The differentiation of S from C/Z is called distinción. Let’s call the phoneme represented by S in accents with distinción (and the single phoneme in accents without it) /s/, and the one represented by C/Z in accents with distinción /θ/.
In the Castilian accent (most of northern-central Spain) /s/ is pronounced as a retracted apico-alveolar voiceless fricative s̺. To English-speaking ears, this will sound more like a SH than an S, that is, more like ʃ than s. If you want to try to pronounce this sound, I find that it helps to start from an unrounded ʃ, and then de-palatalizing it by lowering the back of the tongue. It’s not perfect, but it’s the closest I can get, anyhow. On the other hand, /θ/ is pronounced θ, like in think. Because of these particular sounds of the S and the C/Z, it is a common myth among English speakers that people with this Spanish accent speak with a “lisp”. It’s not a lisp, and it’s not even just one sound, it’s two different sounds – albeit two sounds that to English speakers don’t sound anything like the English s.
There are other accents with distinción besides the Castilian one. In Catalonia, for example, some speakers pronounce /s/ as a lamino-alveolar s (like English S), while still pronouncing /θ/ as θ.
In the rest of the Spanish speaking world, there is only the phoneme /s/, and it is pronounced in different ways depending on the region. The Argentinian, Uruguayan, Chilean, Mexican or Peruvian s (with some regional exceptions) is for the most part indistinguishable from the English s. In other places, like Bolivia, the quality of the /s/ can be a little different.
Aside from that, interesting things can happen to /s/ when it is syllable-final, and this brings us to section 4.
In many accents of what some dialectologists call “Atlantic Spanish” (southern Spain, Canary Islands, the Spanish-speaking Americas), /s/ has different allophones when it is syllable-final.
Let’s get down to specifics. In Buenos Aires, /s/ is “aspirated” (has allophones distinct from s) when it is syllable-final and another consonant follows. If the following consonant is a non-velar plosive, affricate, nasal, /l/ or /r/, it is pronounced as h (another post to come will discuss the allophony in detail in every case). In other accents further away from the Buenos Aires region, as well as in Uruguay, Chile, Paraguay, the Caribbean, Canary Islands, south of Spain, Central America, and regions of many other countries including (but not limited to) those along the Caribbean coast, S is also aspirated when it is syllable-final and a vowel follows.
To illustrate it with an example, the phrase dos ojos negros (two brown eyes, poetically; literally two black eyes) would be pronounced like this (word stress not marked):
dos̺ oχos̺ neʁɾos̺ (Castilian Spanish; uvular J, apical s)
dos oxoz neɣɾos (Mexican Spanish; velar J, voiced z before n)
dos oxoh neɣɾos (Buenos Aires Spanish; velar J, aspirated /s/ before n)
doh oxoh neɣɾos (Rosario Spanish; velar J, aspirated /s/ before o and n)
doh ohoh neɣɾos (Caribbean Spanish; glottal J, aspirated /s/ before o and n)
(In Paraguay, where syllable-final /s/ before a vowel is aspirated, sometimes intervocalic /s/ is aspirated even when not syllable final, so that casa can be pronounced ˈkaha – the Caribbean pronunciation of caja.)
In addition to what I described above, accents that feature S-aspiration can also drop (i.e. not pronounce at all) word-final /s/ in some situations. Typically, this happens when that word-final /s/ is also utterance-final (followed by a pause); and especially when that /s/ signifies a plural and the plural is redundant, because of a previously occurring numeral, article, etc. This would create the following pronunciations of the phrase dos ojos negros provided a pause followed:
dos oxoh neɣɾo (Buenos Aires Spanish)
doh oxoh neɣɾo (Rosario Spanish)
doh ohoh neɣɾo (Caribbean Spanish)
This is not to say that word-final /s/ is only dropped when utterance-final or when signifying a redundant inflection, or only when both things are true: just that these are the cases where /s/ is most likely to be dropped.
Regarding how “acceptable” this is, in the social sense: S-aspiration is largely invisible to people when it matches the pattern in their own accent. For example, people from Buenos Aires do not notice their own aspiration, and only notice the prevocalic S-aspiration of other accents because it doesn’t exist in theirs (but not the preconsonantal S-aspiration of other accents, because they do it too). S-dropping, on the other hand, is socially more complex and could be considered a sociolinguistic marker. But, it’s not an either-or thing where people will regard any dropped /s/ as a marker of a certain social status, but more of a matter of degree. A speaker’s occasional dropping of utterance-final /s/ which would redundantly signify a plural is largely invisible as well; while a speaker that consistently drops every utterance-final /s/ (or, more so, every word-final /s/) is more likely to be the target of linguistic pigeon-holing. Personally, and as a linguist, I try my best to stand against this kind of linguistic prejudice, but it is my duty to inform you of the fact that people are people, and these prejudices exist. You also have to remember that every region (and every person, really) has their own idea of what is “proper”.
6- Pronunciation of ɾ (as in cara) and l
In Caribbean Spanish, as well as in the Canary Islands and the south of Spain, in syllable-final position the distinction is neutralized between /l/ and /ɾ/, so that in for example the Dominican accent aeropuerto is pronounced ai̯ɾoˈpu̯elto, Puerto Rico as ˈpu̯elto ˈriko and amor as aˈmol, while in other accents the neutralization goes in the opposite direction and mal is pronounced maɾ and so on.
7- Pronunciation of TR
The cluster /tɾ/ is in some regions, for example La Rioja in Argentina, pronounced tʃɹ, like in English buttress. The funny thing here (for people who teach Spanish to English speakers) is that this pronunciation is basically the same that their English-speaking students produce when initially trying to pronounce tɾ.
In some regions /r/ is pronounced as a fricative, not a trilled r. In many provinces in northwestern Argentina, for example, this sound is pronounced ʒ, or ʒʷ (with rounding), or ʐ (a voiced retroflex fricative).
9- Voiced plosives
In all accents of Spanish, there is no difference between B and V, excluding some cases where the speaker is bilingual and the other language has such a distinction. In spontaneous, colloquial Spanish speech, the only case where one might encounter a voiced labiodental v is as a voiced allophone of /f/ in words like Afganistán. So, the letters B/V represent a single phoneme /b/, which along with /p t d k g/ constitute the plosives of Spanish. In all accents of Spanish, there is a process of lenition (weakening) that makes it so the voiced plosives /b d g/, otherwise pronounced b d̪ g, have a different pronunciation when between vowels, namely β ð ɣ (fricative) or β̞ ð̞ ɣ̞ (approximant), with the approximant allophones being more common.
(In the Castilian accents that have a uvular J χ, the fricative/approximant allophone of /g/ is also backed to a uvular place of articulation, thus being ʁ or ʁ̞; the former found in French rien)
In most accents of Spanish, this allophone is not only found intervocalically, but also in all other cases except after a pause, a nasal consonant and, in the case of /d/, after a lateral consonant. Only in those cases are these phonemes realized as b d̪ g.
An exception to this is the standard accent of Bogotá, Colombia. In this accent, the lenition only happens between vowels, and not in other cases. This is a very striking phonic difference that speakers with other accents notice right away. This kind of accent strikes those who do not speak it as similar to the ones found in Spanish dubs of foreign (usually American) movies and television. The reason for this seems to be that the dubbing companies (many of them based in Miami) seem to really like the accent of Bogotá, and because of this they ask their voice actors to mimic it, if they do not have it to begin with. I don’t have this accent, and in my opinion, it is a little foreign-sounding, because one of the telltales of a person whose native language is not Spanish is failing to apply the expected lenition to /b d g/. This is assuming their native language does not apply this lenition in the same way; English, for example, does not. This is, of course, just an impression, and not a judgement on Colombians from this particular region, or those who choose to sound like them.
10- Lenition of voiceless plosives
In some circumstances, /p t k/, otherwise pronounced p t̪ k, can become voiced and then lenited to β̞ ð̞ ɣ̞. This is more likely when the syllable they are in is unstressed, and when they are followed by a voiced consonant. So, for example, it is common in Spain to hear Atlántico as að̞ˈlantiko, with a voiced and lenited /t/, and in Argentina to hear porque as ˈpoɣ̞e, with a dropped /ɾ/ and a voiced and lenited /k/.
Comments are as always appreciated, and do tell if you know of a regional accent trait that I failed to mention.