18

I do not speak a word Italian but I wanted to solve the obligatory dispute amongst ignorants about the pronunciation of gnocco and gnocchi by looking up the IPA pronunciation on the Internet. Unfortunately my results were non-satisfactory (for reasons I list below), so I ask you:

  • What is the pronunciation of gnocco and gnocchi in IPA? (Or is this a case of allophones with multiple “correct” pronunciations?)
  • Unless the answers are [ɲɔko] and [ɲɔki], how can my issues (see below) be explained?

How is cch realised?

  1. [ɲɔkki] or [ɲɔk.ki] – Found in the the Italian Wiktionary, the Langenscheidt dictionary and the English Wiktionary. Contains probably the most reliable sources on the list. However the [kk] sound seems like something to me that would be quickly replaced by [k] unless it distinguishes between meanings. I also fail to hear a [kk] in all¹²³⁴⁵ but one⁶ of the pronunciation samples I found – but then again, I know how ears can be deceiving you with such a thing. Finally, no non-IPA pronunciation instruction I found contains this, though it could have been easily represented (e.g., nok-key instead of nokey).
  2. [ɲɔki] – I couldn’t find this in dictionaries but it has surprisingly many findings on the Internet and would probably be en par with 1., if it weren’t for Wiktionary clones. Also what I hear from most audio samples¹²³⁴⁵, though I know that I am fallible here. Finally, what seems most plausible to me linguistically, since it seems to be the way of least resistance.
  3. [ɲɔkːi] – Found on some Wikipedias, amongst others German. Lengthening the [k] does not make sense to me, as it is by nature a punctual sound and thus cannot be lengthened (it’s like lengthening the sound of a clap). When I try to lengthen it, it becomes something like [χ].

How is gn realised?

For the singular form, the Italian Wiktionary claims that it is pronounced [ɲɲɔkko] and not [ɲɔkko], while the pronunciation of the plural form is given as [ɲɔkki]. As pluralisation happens at the end of the word orthographically as well as in all Indoeuropean and in particular all Romance languages that I am aware of, this seems rather unlikely (though not impossible) to me, moreover so as there seem to be other clear means to distinguish the plural from the singular form. I also do not hear anything like this in the pronunciation samples¹ ² ³ ⁴ ⁵⁶.

Wrzlprmft
  • 283
  • 2
  • 7
  • 1
    I don't know the answer but I envy your ability to make some sense out of the IPA pronunciation. I would be completely lost without an audio file providing me with the actual sound of a word. I can't memorise the symbols and even when I do, I end up making unintelligible sounds. If I were you, I would ditch IPA for an audio file. – gd1 Feb 24 '15 at 23:12
  • 3
    The “doubled consonants” are a distinctive phenomenon of Italian that's absent from most other European language. The amount of “doubling” differs from region to region. The sound [ɲ] calls for “raddoppiamento sintattico”, so “lo gnocco” is indeed (in central Italy), [lo'ɲːokːo]. In Veneto it's more commonly il gnocco, pronounced [il'ɲoko]. – egreg Feb 24 '15 at 23:19
  • 3
    @gd1: I did listen at audio files (and reported the results). The problem with audio files is that your ears (or more precisely your brain) are not trained to distinguish sounds you are not familiar with and thus will “autocorrect” certain phenomena to some approximate in languages you speak well, in particular your native language. We once had somebody on German Language SE who could not distinguish [e] and [i] (the e and i sounds in mise in Italian, if I am not mistaken), which was almost unimaginable to me. – Wrzlprmft Feb 25 '15 at 07:24
  • 4
    @gd1: As strange at it may appear to someone, there are some of us who actually consider it clearer and easier to have something explained to them by means of IPA transcriptions than by audio files. – DaG Feb 25 '15 at 09:30
  • Fair enough :) I wasn't making any point, it was just a side comment. – gd1 Feb 25 '15 at 12:15
  • there are two phenomena that are not your friend here:
    1. doubled consonants in Italian ortography produce very different audible results depending if they are voiced or not. So "palla" actually has a long L sound, longer than the L in "pala", because the L is voiced. So does the R sound in "carro" / "caro".

    But the T sound in "fatto" / "fato" is not voiced: so what you (should) produce is a small pause. So "fatto" -> "fat.to". 2) regional pronounciations differ: very roughly, Northern Italians tend to ignore some doublings while from Rome downwards doublings increase.

     – Walter A. Aprile Feb 25 '15 at 15:29
  • The problem is not in "voiced"/"voiceless", at least if the terms are taken for what they mean in a phonological context. For example, "d" is voiced, and "s" is voiceless, but "d" and "t" go together, as do "l" and "s", I guess, @WalterA.Aprile. – MickG Feb 25 '15 at 15:44
  • The issue concerns the type of consonants. We divide the consonants into "obstruents + affricates" and "the rest". "k" "t" "d" etc are obstruents, i.e. produced by blocking the airflow and releasing it with a burst. Obstruents are geminated by doubling (or lengthening) the time you keep the air inside before the release. A more precise IPA representation for that is [k̚k], where the [ ̚ ] above the first k signals it is unreleased, i.e. there is only the blocking, not the burst release. Affricates are combinations of an obstruent (henceforth stop) and the corresponding fricative. – MickG Feb 25 '15 at 15:49
  • More precisely, an affricate is when a stop is released through its fricative. That is probably what OP was getting by trying to "lengthen k", as he said he got "something like [χ]". Now the actual corresponding fricative for k would be [x], so the affricate would be [kx]. However, I see that anywhere there is a velar fricative [x], it always turns to [χ] for some reason: see German with Bach and similar, or Arabic's seventh letter kha, which would be read [xa] but I actually heard a lebanese guy name it [χɛ]. – MickG Feb 25 '15 at 15:52
  • Now Italian affricates are [ts], [dz] (the two sounds of z) and [tʃ] and [dʒ], [tʃ] being the sound of c before i and e (ci -> [tʃi], ciao -> [tʃao]) and [dʒ] that of g in the same contexts. THose affricates are doubled by doubling the stop and then releasing it as the fricative, so cci -> [t̚tʃi]. The rest of the consonants can really be "lengthened", so you can easily make sense of, e.g., [l:]. – MickG Feb 25 '15 at 15:56
  • Indeed, many of the other consonants can be used as vowels. r and l are in many Slavic languages (e.g. Trieste in Slovenian is Trst), [z] can be seen as used as a vowel in Mandarin, though some may view it as an [ɨ] in disguise, and the same goes for [ʐ]. But that goes out of the question's scope. – MickG Feb 25 '15 at 15:57
  • Hope this helps you @Wrzlprmft. The answers below already answer very well, so I decided to post this as comments, mainly just to offer the alternate IPA for geminate consonants. – MickG Feb 25 '15 at 16:00
  • 3
    And @DaG I am one of the "some" :). – MickG Feb 25 '15 at 16:01
  • I suggest you do not use the Italian Wiktionary, because it has little information and, probably, few editors, therefore may be inaccurate, as with the double ɲ here. Use the Enlish version instead. – MickG Feb 25 '15 at 17:05
  • FWIW, the only sound samples that sound correct to me are 2 and 5. – Matteo Italia Feb 25 '15 at 23:24
  • 1
    Nyoki is how you pronounce gnocchi nyoko is how you pronounce gnocco –  Feb 26 '15 at 01:19

4 Answers4

18

“Doubled consonants” are a distinctive characteristic of Italian; depending on the regional variant, this “doubling” can be more or less long. Typically, in Italian of the northern regions the amoung of “doubling” is usually weak, while it's more evident in central and southern Italy. The English term is geminate.

The digraph gn is almost always realized as [ɲ], except in some loan words such as gneiss. In central Italy the sound is usually geminated, so we hear [compa'ɲːo]. The same if the word is after the article and the phenomenon is made stronger by the usual raddoppiamento sintattico, so we have

[lo'ɲːokːo] and [ʎi'ɲːokːi]

(Note that sometimes instead of [kː] the geminated consonant is denoted by [kk], but it denotes the same: a single sound longer than the simple one).

As I said, actual pronunciation of the geminated consonants may vary; I find myself using different length depending on the context and where I'm speaking (I almost never do the raddoppiamento sintattico, however).

Note that while grammar imposes the articles lo and gli before gnocco and the other masculine words starting with gn, regional variants may well use il and i (albeit “ungrammatical”).

Note also that the h in cch is just a graphic device for stating the non palatal sound of c, that is, [k].

egreg
  • 17,991
  • 3
  • 28
  • 67
  • I would replace stronger with longer in your answer. A geminated consonant may be stronger, but it is longer for sure. – Walter Tross Feb 25 '15 at 14:58
  • @WalterTross I agree. – egreg Feb 25 '15 at 15:21
  • 1
    I'd only add that the gemination of a central /ɲ/ or an initial one after a word ending in vowel, beyond being a Central-Southern, is, for what's worth, “standard Italian”: see for instance the DOP, where the little +-sign before the pronunciation means the the initial letter is affected by raddoppiamento sintattico. – DaG Feb 26 '15 at 11:50
9

I may be able to shed some more light on the issue with geminated consonants. While I don't speak more than a few words of Italian, my native language is Finnish, which also makes a phonemic distinction between long and short consonants.

First of all, let me mention that I do hear a long [k:] in most of the pronunciation samples that you link to — except, curiously, the last one, where you claim to hear it.

The first of your samples is particularly illustrative, since the same speaker says the word gnocchi three times, with different speeds and stress patterns. To my ear (which, of course, is calibrated for Finnish consonants, and therefore may not perfectly match what a fluent Italian-speaker would hear) the second (slow) pronunciation has a clear long [k:] (and a long [o:]), while the last (fast) one sounds like a short [k] to me, with the first one falling somewhere into the gray area between the two. (Of course, a native Italian-speaker might have a different threshold, and might hear them all as [k:].)

As for how to pronounce [k:], remember that [k] is a stop consonant: you pronounce it by raising your tongue so that it blocks the airflow, and then releasing it. The basic distinction between a short [k] and a geminated [k:] is simply the length of time that you hold the airway closed; in [k], the release is almost immediate, so that there's only a momentary "clap" in the airflow, whereas for [k:], the airflow is held closed for a noticeable time (comparable to the length of surrounding vowel sounds), producing an audible pause between the closing and the opening of the airway.

Of course, since the length of the stop is really a continuous variable, the line between [k] and [k:] is necessarily somewhat arbitrary, and may vary between different speakers and listeners, and also depending on context. Still, in languages that make a distinction between short and long stops, there is generally sufficient agreement between speakers to get the point across.

Ilmari Karonen
  • 191
  • 3
  • To clarify: I did not claim to hear a long [k] in the last example, but a double k ([k.k], i.e., two stops and claps), which is what I read from the IPA representations [ɲɔkki] and in particular [ɲɔk.ki]. – Wrzlprmft Feb 25 '15 at 13:34
  • Also, do I understand this correctly that there is no audible distinction between [kː] and [pause]+[k]? – Wrzlprmft Feb 25 '15 at 13:36
  • @Wrzlprmft: The closing of the airflow does also produce a sound, albeit a faint one; it may or may not be distinguishable from simply closing your vocal cords (a glottal stop) to produce a pause. One situation where most people can hear the difference is when the place of articulation changes between the closing and the opening of the airway, as in [tk] (vs. [k] / [k:] / [ʔk]); that said, many English dialects do tend to lose even that distinction, so it may be hard for native English-speakers to hear or pronounce. – Ilmari Karonen Feb 25 '15 at 14:39
  • As for an actual double stop like [k]+[k], I'm not aware of any language that has that, as it's really awkward to pronounce and sounds like you're stuttering. I'm not sure how you're hearing one in the sound sample you mentioned, either. – Ilmari Karonen Feb 25 '15 at 14:40
  • Nice analysis. We don't have a phonemic distinction between long and short vowels, like in Finnish. However, a vowel can be longer depending on the phonetic context: in poco the first o is usually a bit longer than in gnocco, where it has the shortest length. – egreg Feb 25 '15 at 22:09
2

It is pronounced as nyawk-kee. Gn is pronounced as an Spanish accent ñ. Cch is pronounced as a k.

Gnocco is pronounced as nyawk-koo

Hope this helps!!!

Tia27
  • 970
  • 4
  • 11
  • 29
  • 3
    nyawk-kee is not equivalent (it would be niocchi), if we don't want to use the IPA, i would say ñawk-kee. Because, as far as i know, english doesn't have gn sound. – Fabio Filippi Feb 25 '15 at 12:50
  • @FabioF: Actually, I'd say "nyawk-kee" can be a pretty good approximation, although of course it depends on which dialect of English you're speaking (and on how you choose to pronounce it, given that it's often not possible to unambiguously determine the pronunciation of an English word from its spelling). In most varieties of English, the "y" in non-word-final "ny" is purely consonantal (as e.g. in the word "canyon"), and the combination is pronounced fairly close (though not exactly like) the Spanish ñ or the Italian gn. – Ilmari Karonen Feb 25 '15 at 15:16
  • 1
    @IlmariKaronen: Something that “can be a pretty good approximation” when pronounced according to some specific English dialect might be good elsewhere, but here the OP described better how “gn” is pronounced in Italian than this answer does. – DaG Feb 26 '15 at 18:27
1

I am Italian. I have recorded a short video that shows exactly how to pronounce both gnocco and gnocchi. Gnocco is tentatively pronounced: ñawk-ko, gnocchi is pronounced: ñawk-kee.

I am sure a recording is going to give the best representation of the words involved, better than any graphing, Wiki and whatever could achieve.

Some additions: I have some additional answers to the original post.

How is cch realised? Your 1 and 3 findings are mostly correct. To get the exact sound, imagine saying: "Kentuky". Now take the second "k" and hold on releasing its sound for half a second. Visually speaking, imagine you "lag" while you are halfway done pronouncing the "k" so it doubles.

How is gn realised? This is harder. The closest way to express it is to pronounce the "incipit" of a "g" with an attached ñ. We Italians refer to two kinds of "g": soft and hard. What English gets by default is the hard version, whereas Italians by default use the soft version and write: "gh" for the hard version.

Charo
  • 38,766
  • 38
  • 147
  • 319
Dario Fumagalli
  • 119
  • 2