Saturday 30 June 2012

Italian triphthongs?

In a discussion on the history of the terms smoothing and compression, Jack Windsor Lewis, in an article published on the 25th of June 2012, said:

“Sequences like /aɪə/ and /aʊə/, [Daniel] Jones pointed out, because they were not single syllables, were strictly not properly termable as ‘triphthongs’. In the first edition of his Outline of English Phonetics (1918 §107) he instanced the Italian word buoi as containing a true triphthong.”

The last sentence in the quote above I find a bit baffling. Does Italian really have triphthongs?

According to a traditional view shared by the majority of linguists in Italy, a true triphthong is a sequence of three vowel sounds belonging to the same syllable. If we take this definition and apply it to the word buoi (‘oxen’) proposed by Daniel Jones, we get something like (ˈ)buːɔi, a pronunciation I've never heard from Italian native speakers. Buoi is normally pronounced (ˈ)bwɔːi

Buoi is not a word 'containing' a true triphthong but just a sequence of sounds consisting of a semivowel plus a falling diphthong. This claim is also supported by the fact that when Italians want to put emphasis on this word, or when in school teachers ask their pupils to syllabify it, the pronunciation that is often produced is buˈɔːi. This realization suggests that buoi is underlyingly disyllabic although Italians normally pronounce it compressed and monosyllabically as a sequence of w + ɔːi

Like buoi, terms such as miei, suoi, guai, aiuola, continuiamo, and quiete all exhibit similar phonetic features. In the case of continuiamo ('(we) continue') or quiete ('quietness'), we can even get, often in fast speech, the change of w → ɥ because of regressive assimilation: ˈkɥjɛːte/ˈkɥjeːte, (ˌ)kontiˈnɥjaːmo. Why then do most Italian writers consider these good examples of words 'containing' a triphthong?

The answer, I think, is to be found in the fact that grammarians and linguists in Italy have always been traditionally too much 'attached' to the orthography of words and only rarely have they paid attention to the actual phonetic reality of the language. A term like buoi DOES, of course, have three vowels but it's three vowel letters not sounds! The same can be said of all the other examples I gave above. 

It is perhaps not surprising to hear that some primary school teachers, when wanting to explain to their pupils how many vowels we use in Italian, often provide the noun aiuole ('hedges') as an example. This, they claim, is an 'excellent' word as it contains all the vowels of Italian. But again they don't realise that they're talking about letters, not sounds! Aiuole normally contains three vowels: a, ɔ and e.

As I see it, triphthongs in Italian are a no-no, but you might disagree with me.


  1. As they are attested as occurring in Spanish, Portuguese and Romanian it wdnt be at all surprising if triphthongs were found to occur in some dialects of Italian but as far as one knows, none have been claimed to exist other than the one attested by Jones. He no longer included that comment after his second edition of 1922 but we dont know why. He woudnt've been likely to take it on trust and the most likely person for him to have discust the matter with wd've praps been Amerindo Camilli (a Marche native) who published in 1921 under Jones's editorship An Italian Phonetic Reader. I feel no reason to dou·t the evidence of a scholar of Jones's calibre in this matter that some Italian has been observed as capable of using a true triphthong in 'buoi'.

    1. Spanish has as many triphthongs as Italian does: none. Spanish words like buey contain the sequence of a consonant and a diphthong: [wei], which -- just like in Italian as Alex pointed out -- might become disyllabic in slow speech: [u'ei], which isn't a triphthong, either.
      The idea of analyzing a phonetic sequence of sounds that never make up a true triphthong as a phonemic triphthong doesn't sound reasonable to me.

    2. There is something I don't understand here: if "uoi" and "uey" (in "buoi" and "buey") consist of a CONSONANT followed a diphthong when they are spoken as a single syllable, how come they consist of a VOWEL and a diphthong when that syllable is broken into two pieces? Is it because the vowel turns into a consonant in order to avoid a triphthong?
      In any case, I have always been told that we have a triphthong in Spanish "buey" -Could we say, then, that a triphthong would be a sequence of a consonantal vowel followed by a diphthong?

    3. Those consonants necessarily become vowels, because [w] or [j] couldn't constitute a syllable nucleus: [bw'ei] is ill formed, so it turns into [bu'ei].
      A diphthong by definition can consist of nothing but two vowels. Likewise, a triphthong can consist of nothing but three vowels.
      Since [w] and [j] are clearly not vowels (and thus not at all interchangeable with [u] and [i], respectively), no sequence containing them should be called legitimately a diphthong or a triphthong as doing so would violate the definitions. Under no circumstances can we consider sequences like [we, je, aj, aw; wei, jei, aja, awa] valid diphthongs and triphthongs. At least we can't if we keep the definitions in mind.
      As a consequence, words like Spanish bueno or Italian buono do not have phonetic diphthongs, either. A valid triphtong would be e.g. [uei] with [u] being the most prominent (and possibly even elongated) part of it.
      It is always the first member of a diphthong or triphthong that is the most prominent, since if the second or third vowel were more prominent, they'd be more stressed than the previous one(s), in other words: they'd start a new syllable ([u'e, u'ei]), which is obviously incompatible with the one-syllable concept of diphthongs and triphthongs.
      Unfortunately common practice too often disregards definitions...

    4. I see what you mean, yes. In the Spanish tradition, the vowels i and u are called "semiconsonantes" (but still regarded as vowels) when they appear "at the beginning of a diphthong or triphthong, as in piedra, hielo, huerto, apreciáis". On the other hand, they are referred to as "semivocales" when "at the end of a diphthong: aire, aceite, causa, feudo". (Translated from Diccionario de la lengua española. Real Academia Española):

    5. I think these articles use the word vocal in an orthographic sense rather than a phonetic one, because this word can be used also as an abbreviation for letra vocal 'vowel letter', i.e. a letter that represents a vowel sound.

      I guess that's exactly the case of being too much "attached" to orthography, as Alex referred to it above. Misleading and confusing.

      Besides, semi-vowels and semi-consonants don't even make sense at a phonetic (= articulatory) level, since any sound is either a vowel or a consonant. Semi-vowels and semi-consonants can be used only in a phonological (= structural) or orthographic (= spelling) sense -- neither of which is related to actual speech sounds.

    6. Well, I understand that the mark "FON." (in the links I copied above)makes it clear that the sense is phonetic -just another of our typically Spanish traditions. (Maybe the gentlemen were referring to the fact that the segment begins as something like a consonant and ends as a vowel, and vice versa; couldn't this make sense at a phonetic level?)

    7. Yes, you're right: I overlooked "FON.", which just makes it even worse, because it means that they truly mean and believe the nonsense that a consonant sound can be a part of a diphthong, contrary to its definition.

      If "the segment begins as something like a consonant and ends as a vowel", it's exactly a sequence of a consonant and a vowel, or vice versa, neither of which is a valid diphthong at a phonetic level.

    8. So you don't believe in triphthongs. What about diphthongs? I suppose that the weaker element must come after the stronger one, and that it must not be too close (I'm thinking of the diphthongs in RP "bear", "beer", "tour" etc.)

  2. To be precise, I don't believe in triphthongs that contain consonants. But a sequence such as [ɛɪə] you might hear in bed affected by Southern Drawl is a true triphthong.
    Diphthongs are analogous: they must not contain a consonant, but they can contain any of the vowel sounds, including [i]. For example, Spanish vais is [bais], which doesn't sound the same as [bajs] does.
    The weaker element must follow the stronger one, but weaker and stronger are not to be meant qualitatively, but rather quantitavely, meaning the weaker one is less prominent/stressed, so both [ai] and [ia] are valid diphthongs, whose first members are more energetic and possibly even longer than their second ones.
    Traditional RP bear, beer, tour all meet these criteria, having diphthongs of [ɛə], [ɪə], [ʊə], respectively.

    1. Thank you, p.c.
      Regarding [ɛɪə], wouldn't the closer middle element act as a sort of syllable boundary, so that we wouldn't have a triphthong even in this case ([ɛɪ-ə])?
      Somebody find us a valid triphthong!

    2. Not necessarily, it depends on how it's pronounced. BrE has a diphthong followed by a monophthong in words like layer, because the third vowel in the row lasts longer than it would if it were part of a triphthong. But in case of bed as exemplified above, also the third vowel is short.
      Any combination of three vowels can be a valid triphthong, it may not exist in any language, though, so [uoa], [eou], [æɑo] are all possible valid triphthongs.

    3. Well, thanks again!
      Hey! what about the initial u in [uoa]? Couldn't we substitute that for the [w] in Italian "buoi" -Alex just said that this word is "normally" pronounced [bwɔːi]. Then we would have a perfectly valid diphthong in a beautiful real language.

    4. Sorry, I mean "a perfectly valid TRIPHthong".

    5. "Couldn't we substitute that for the [w] in Italian "buoi""
      We could, but that wouldn't sound Italian. ;)

    6. Ok, you win, my dear fellow! :)