Tuesday, 22 December 2009

deaccenting repeated words

As we know, there is an intonation rule in English and the other Germanic languages that when we repeat words that have just been uttered we normally deaccent them. That is, we signal the information status of items in our message by removing potential accents from “old” information.
Do you speak English?
— Of \course I speak English.

Clearly this principle is not universal. If it \/were universal, | NNSs would not get it wrong when speaking English, as many of them certainly do.
In some languages the situation hardly arises, because freedom of word order allows non-foregrounded items to be placed away from the end of the utterance (if that is where the intonation nucleus normally goes, as it usually does). Or you rephrase the sentence so as to avoid repeating words.
When listening to non-Germanic foreign languages I always try to watch out for examples or counterexamples. (For Spanish, Héctor Ortiz-Lira has listed plenty of them in his articles on the subject.) But this can be difficult to do in the middle of a conversation when one’s knowledge of the language in question is limited.
The other day I called in at a printshop here in London to order some work. The assistant I dealt with gave me his business card. I saw that his name was obviously Greek, so I asked him
Μιλάτ’ ελληνικά; miˈlateliniˈka ‘Do you speak Greek?’
—to which he replied
Βεβαίως μιλώ ελληνικά. veˈveozmiˌloeliniˌka ‘Certainly I speak Greek’.
At least, I think that’s the accentuation he used. But I could be wrong. I’m not sufficiently au fait with spoken Modern Greek intonation to be certain of having heard the tonicity correctly. Did he really remove the accents from the last two words?
With Germanic languages there seems to be no uncertainty: they follow the English rule.
Sprechen Sie deutsch?
—Ja, natürlich spreche ich deutsch.

What about Romance languages? And other languages?
Vous parlez français?
—Mais oui, bien sûr que je parle français. (tonicity?)

Parla italiano?
¿Habla español?
Czy Pan mówi po polsku?
Говорите по-русски?


  1. Same in all Romance and Slavic languages, as far as I can tell.

  2. Q. Parla ita/liano? A. \Certo che parlo ita\liano.
    In Italian the last word of the sentence tends to have always a pitch variation, anyway (usually a fall if a statement, a rise if a question). If there is "new" information in the sentence, the word that carries it would have pitch variation too, which would be regarded simply as "emphasis".
    In "certo che parlo italiano", "certo" would have equal/more/less pitch variation than "italiano", depending on how much emphasis one would want to place on it.
    I believe that the point is that there is no such a thing as a "nucleus" in Italian.

  3. Hablo espa\ÑOL. ¿Y usted? ¿También habla espa/ÑOL?

  4. Pues claro que hablo espa\ÑOL. Hablo y escribo espa\ÑOL. (But you can also hear "...Hablo y es\CRIbo español.")

  5. Doesn't that suggest that we're talking about different situations?

  6. The NNSs I've heard get this wrong in English tended to be from East Asia (China, Korea, Japan), so maybe we need to look there for counterexamples rather than to other European languages.

  7. Side note: I hear American-born NS actors get this wrong all the time - when they are working on text. Something about reading makes them sound like they are speaking EFL, though only in this respect.

  8. American-born NS actors often overdo the German accent.

  9. What do you mean by 'same', Lipman? What strikes me about the Germanic languages is that once a sequence loses all rhematicity it drops to a low monotone, except for any left-over bit of thematic intonation distributed over it, as in JW's example 'If it \/were universal'. Of the three Romance languages he mentions, I get the impression that in similar circumstances Italian preserves the accent and pitch almost intact, Spanish less so, and French retains some of the oxytonic accent, after a gradual descent to a mid-high or mid-low pitch. My impression of what happens in Polish and Russian is not worth the digital footprint, but I think I see what you mean, i.e. that they come in the same sort of range as the Romance languages.

    I got carried away trying to comment on the Japanese, but I think I have boiled it down enough to be of some use here. JW's example already has an explicit theme at the start, あなたは (As for you), which for the purposes of comparison should probably be left out, as in Italian and Spanish. It would unduly complicate this comment to deal at all with the ways in which that might be treated. The trouble is that we cannot really deal with 日本語を (Japanese+object marker) at all, as the parallel to the Germanic deaccenting would be to leave that out altogether as well. If however you leave it in it may still be deaccented (but not necessarily completely depitched, if the actual pitch-accent of the assertion 話します (hanashimásu - I speak) survives on the 'ma'. This would depend on how you said 'of course': a word like 当然  (tôzen) would be likely to preserve it, but 勿論 (mochíron) would have such a dominant accent that the whole phrase日本語を話しますmight well be both deaccented and depitched, which is presumably what JW is looking for. But there is probably always going to be a –yo particle on the end of it which is likely to have the same sort of rising intonation as the last syllable of 'universal' in JW's abovementioned example 'If it \/were universal'. Alternatively there is another enclitic which itself means 'of course', which would completely preserve the accent of hanashimásu and therefore the pitch of Nihongo (Japanese), if it's there.

  10. Sorry, I meant to say "another enclitic, 'tomo'". Thus (/Nihongo-wo) /hanashimás'tomo.

  11. Oh dear, Christmas brain fog. /\hanashimás'tomo.

  12. My wife is a native Russian speaker and is sure she leaves the accent intact.

    Говорите по русский?
    Да, я говорю по руссцй.

    Тhe reply has a falling intonation, with the main fall over the accented final syllable of говорю 'speak'. However, to my ear there is a distinct intonation contour to the words по русскиы 'in Russian'. It's also a fall, but it starts with a slightly higher pitch to unstressed по 'by' before a resumed fall on the first syllable of русский 'Russian'.

    A tentative and non-phonemic transcription:

    gəvʌrʲitʲe pʌ russkʲij
    da ja gəvʌrʲu pʌ russkʲij

    Suggestions welcome for improvement.

  13. Sorry for the misspelled русский! it started with the all-too-common mistake of typing ц for с. (Wrongly using the C key). I partly corrected, then deleted the wrong letter.

  14. David C, your description (though I'm afraid not your transcription, which I won't attempt to comment on, as we have seen peculiarly strong opinions on the transcription of Russian on here) corresponds exactly with what I was hearing in my head for Russian, and what I thought Lipman meant. I wouldn’t be surprised if he feels as I do that there is some sort of general principle operating here for languages that don't throw the 'given' away as completely as the Germanic languages that JW lumps together here (and I have come dangerously close to lumping together with Japanese!)

  15. Russian:

    Вы говорите по-русски?

    Конечно, я говорю по-русски.

  16. I think in Polish (my native language), I would expect this:

    (Mówi Pan po polsku?)

    \Jasne, że mówię.
    'Of course I do.'

    Oczy\wiście, że mówię.
    'Of course I do.'

    \Jasne, \mówię.
    Of course I do.

    \Tak, \mówię.
    'Yes, I do.'

    \Jasne, \mówię.
    'Of course, I do.'

    So it seems (even though I haven't done quite enough thinking) that whether the latter part is deaccented depends on whether it's a subordinate that-że-clause or not.

    And English doesn't seem to be all that different. What intonation pattern you get depends on whether you echo the "speak English" clause or not. Yes I do would probably carry the accent on do, woudn't it? The thing is, most of the other languages discussed here (I can't speak for Japanese) do not have this "short answer" option.

  17. Sorry for repeating one pair of examples -- posting from a really small screen...

  18. If you include po polsku in the reply, it'll be as deaccented as in English, unless, and that's true for most of the languages above including English, you're implying "and not [other language]".

  19. @Lipman: Thanks for making me aware that I ignored the po polsku bit -- silly me. But I don't think it changes much in my pseudo-analysis above. In those cases where you do get the that-clause, the whole final bit, including po polsku would be deaccented, like in John's English, Greek and German examples. Through my NNS ears, I honestly can't hear much difference between the pattern in the English example and any of the Polish ones. Things change if the second clause is not a that-clause -- that's the point.

  20. As Sergei says, Конечно, я говорю по-русски. is a particularly natural reply (and he spells it properly!) It was one of the replies my wife suggested, but I chose to listen to the one with Да because it could only be spoken as a reply.

    Lena's first suggestion was

    Говорите поьрусский?

    The Russian way to de-emphasies the repeated phrase is ... not to repeat it.

  21. I refer you to my above observation that the Japanese way to de-emphasize the repeated phrase is ... not to repeat it, only more so - more NOT to repeat it, that is!

    Wjarek's "short answer" option is the English analogue, and as he says, "most of the other languages discussed here (I can't speak for Japanese) do not have this".

    And indeed, Japanese doesn’t either, and just as in the Russian Говорю, you can throw away the 'of course' and all the enclitics and particles, and say matter-of-factly,


    With the moras of ma and su distributed over the remaining final syllable, to finish up with a pitch-accent pattern as near as dammit to Говорю.

    A lot of this has to do with whether the language is supposed to have a pitch accent or a stress accent. For a matter of supposition is what it is. Some of the languages in this discussion have moved in the direction of a stress accent, including allegedly Japanese - but certainly not my wife's Kansai (western) Japanese! The extent to which the accent of a given survives deaccenting is the extent to which it is a pitch accent, if you ask me! Stress alone doesn’t seem to cut the mustard.

  22. This comment has been removed by the author.

  23. This is been so right that when you are speaking literally those sentences and ideas, the native speaker would almost get the right understanding to their minds that you would not get understand even. paragraph reworder