Friday, 22 July 2011

Gdynia unmasked

I was in my teens when I first became aware of places in Poland called Gdynia and Gdańsk. I remember wondering how they could be pronounced without omitting the initial g. The nearest thing in English seemed to be the ɡəd- sequence in Lady Godiva ɡəˈdaɪvə and Godolphin (nowadays we have another example in Gaddafi), but I knew that wasn’t right. Yet as an initial cluster gd- seemed impossible to pronounce.

Then I realized that English does have the articulatory sequence gd, but in final position. We get it in the past tense of verbs ending in g, thus for example bagged bægd, hugged hʌgd. We get it medially, too, in Ogden ˈɒɡdən. All I needed to do was to transfer this gd to syllable-initial position.

That still seemed very difficult to do. The reason (I know now) is that in English we normally pronounce these plosive sequences as overlapping articulatory gestures. You can’t hear the release of the g in hugged, or for that matter in Ogden, because it is ‘masked’ by the concurrent hold phase of the d. What you hear is a velar approach, a long hold, and an alveolar release (if you’re lucky). And we don’t ever have this sort of thing at the beginning of a syllable.

It wasn’t until I first visited Poland, when I was twenty, that I discovered that in Polish the initial plosives in Gdańsk ɡdaj̃sk and Gdynia ˈɡdɨɲa are not like the English gd in hugged. Rather, when my Polish friends demonstrated the pronunciation to me, they released the velar plosive BEFORE completing the approach for the dental. This inevitably gave rise to a tiny transitional vocoid between the two hold phases, but it was not long enough to count as a separate schwa segment (compare the definite schwa in English Godiva).

If you can play .ogg sound files, Wikipedia has one of Gdynia here.

Despite my A level in Greek, somehow I’d failed to realize that classical Greek has an exactly parallel, but voiceless, cluster in words such as κτείς kteís ‘comb’ and κτίσις ktísis ‘foundation’. (What we did to pronounce those words in the Classical Sixth I can no longer remember.)

The stem of κτείς is κτεν- kten-, and from this is derived the modern zoological Latin name Ctenophora, the phylum of marine animals also known as ‘comb jellies’. In English we abandon any attempt at the initial plosive cluster, and pronounce them simply as tɪˈnɒfərə.

Modern Greek has dissimilated the plosive sequence, making the first element fricative. The modern word for ‘comb’ is χτένι ˈxteni.

So there are two attested escape routes: deletion or dissimilation.

But the Poles succumb to neither of these tempting articulatory simplifications, and persist with a sequence of plosives. Unmasked.


  1. Yes, very right. The problem with those solutions ("look, you have that in English, too, just in a different context"), which I use occasionally in explaining, too, is that often enough, they aren't really identical. Suddenly, syllable boundaries etc. matter when you look at those nice graphs.

    Still, in this pticular case, native speakers of English might very well realise the schwa in Godiva as a tiny transitional vocoid, but not everybody does that, and it's only one out of several options even for those who do, while there's no free or register-related variant ɡədaj̃sk in Polish.

  2. Forgot to ask: is there an API convention for this, a very short schwa or the like?

  3. I've had a similar experience with Polish clusters - pronouncing them too 'well'. A slight vowel between certain consonants in clusters isn't a problem. I remember I used to pronounce 'mnie' (me) mɲɛ by fully forming the second nasal before opening my mouth and releasing the first. With time I realised that this was quite un-Polish and that letting a slight schwa slip in was much better.

    It's led me to think at times 'when is a cluster a cluster?'. Is it worth making a difference between a phonemic cluster and a phonetic cluster? In that case many Polish phonemic clusters are not usually pronounced as phonetic clusters and might not be heard as clusters by foreigners.

  4. You could use a schwa with the "non-syllabic" underbreve: [gə̯daj̃sk].

    Or you could leave it out and instead add the "unreleased" diacritic in the English clusters instead: [bæg̙̚d].

    I'll be amazed if all those symbols come out correctly but I'm sure you know what I mean!

  5. Then there's Russian гдеgdʲe. To my ear, it's like Gdynia and Gdansk, but my wife swears that there's no sound between the first two consonants.

  6. I still find it amazing how strongly the phonotactics of our native language constrain our pronunciation of second languages.

    English contains some fantastically complicated consonant clusters in perfectly common words like widths wɪdθs and sixths sɪksθs (although I suppose most people normally say sɪkθs) but a simple word like gdʲe is a challenge.

    And then plenty of Italians struggle with words like wounded 'wuːndɪd because Italian doesn't have w immediately followed by .

  7. I, native of Gdańsk, seem to pronounce both 'gdańsk' and 'gdynia', or 'gdakać" (to cluck), or even 'gwda' (a Polish river) without any vocoid between the 'g' and the following consonant. Neither can I detect any in (my pronunciation of) 'bagdad' or 'agda' (a kind of computer software). Or for that matter in that of 'abd(ul)', 'servant of', a frequent element of Arabic names.

    I'd think David Crosbie's Russian wife is right in swearing there is no sound (vocoid?) between 'g' and 'd' in где — gdʲe, the Russian for 'where'. As little as there is any between the two consonants of the Polish 'gdzie' (the same).

    I remember there once was a funny song, in which the Polish word 'ptak', or 'ptaszek' (bird, little bird), cognate probably to the Greek root pet-, pt- (fly), was pronounced with a kind of schwa between 'p' and 't' --- it was a satire on what we call 'wymowa zmanierowana', affected pronunciation or pronunciation with a strong mannerism. There is always a chance that a foreign linguist mistakes this kind of pronunciation for normal, standard, neutral ...

  8. One more reflection: in sloppy speech, we Poles delete the 'g' of 'gdzie' (where), what remains is simply 'dzie' [d͡ʑe], a pronunciation considered vulgar.

    Interestingly, the German name of Gdańsk is 'Danzig', which seems to reveal that the German settlers in the XIII century simply did not hear the 'g'. Does this make the case for a vocoid more or less probable? I dunno. Judge for yourselves.

    Ad Paul Carley

    can you give me an idea of what your 'first' pronuciation of 'mnie' sounded like? I am hard put to imagine an un-Polish pronunciation of this innocuous word.

  9. @Wojciech

    I don't think I could add anything that would improve on previous description of 2 overlapping nasals compared with non-overlap and a brief obscure vowel.

    As for imagining un-Polish pronunciations, I can imagine lots of ways a native English speaker might mangle the word 'mnie'. My original over-careful phonetician's pronunciation is just one example.

  10. I was delighted to read a note about the place where I was born and lived my life. I was born in Gdańsk and now I live in Sopot, a smaller town squeezed between Gdańsk and Gdynia; together they form a metropolitan area which we call Trójmiasto (Tricity). The toponimies go back to the Proto-Slavic times: Sopot ˈsɔpɔt comes from *sopotъ, a Proto-Slavic word for brook, while both Gdańsk and Gdynia stem from *gъd, meaning wetlands, possibly woody wetlands (and such are the surroundings here).

    As you can see, between the plosives there was originally a hard yer, an extra short close back vowel, apparently rounded: ŭ. Being in a weak position, it eventually disappeared, but as I now realise, not completely. There has to be a discreet moment when the vocal tract is open, a tiny transitional vocoid, as John has put it; something I imagine as an ultrashort, non-syllabic, close and voiceless vowel.

    But as Wojciech observed, it's virtually invisible for the native speakers, as we are naturally accustomed to stack our consonants in ways that may seem mind-boggling for L1 English speakers. And speaking of consonant clusters, greetings to fellow gdańszczanin :-) On the other hand, having a native inventory of just six simple vowels, I always found the number of John's lexical sets overwhelming...

  11. Ad Paul Carley

    OK, except that this gets me no forr'arder, your description seeming to be of an articulation yielding rather quite Polish, or Polish-ish, rather than un-Polish, results. Maybe the un-Polishness consisted in being over-careful, over-meticulous, over-something else? But such things are un-English, un-French, un-most other languages, aren't they?

    The only sound in 'mnie' I think likely to cause difficulties to an Anglophone is the palatal 'ni'. Or perhaps the vowel without a yota-like off-glide. (*mniey..). Like an American I once knew who kept saying (in Italy): toodou vah beyney (tutto va bene).

    When you say 'monotonous' do you let a schwa slip in between the two first consonants?

  12. @ Wojciech

    Monotonous, indeed!

  13. Ad nul

    so you believe in a vocoid in 'gdańsk' etc.? And if you do, did John's description convince you or did something else, such as your own observation? This would interest me for methodological reasons (I am a philosopher interested in, among other things, the methodologies of empirical sciences).

  14. On [sic] Paul Carley, to whoever might find this useful.

    The point is exactly that, when I hear the word 'monotonous' pronounced care-, but not over-carefully by native speakers, and when I try as hard as I can to pronounce this word myself in a 'correct' way (the way I am used to regarding as correct), I both hear a schwa between the 'm' and the 'n', and have an articulatory 'feeling' of it; when, by contrast, I either hear or pronounce the Polish word 'mnie' (mihi, moi, mir, dative of 'ja', I) I neither hear nor 'feel' any schwa. The same goes for 'gdańsk', 'gdzie', 'gdy', 'abdul', 'bagdad' 'boabdil' and others.

  15. Don't think of it as a vowel you ought to detect being there, think of the plosive release. Does it not ring true that you release the /g/ in the Polish cluster and the English cluster differently?

    I know that if I try to pronounce 'hugged' the way I would the Serbian 'gde' or 'Gdanjsk' it sounds distinctly non-native like.

  16. Ad gassalascajape

    Thank you my Serbian (?) friend. I am, alas, not expert enough, in phonetics or in English, to relate to what you say about the release etc. but if you say the John-C-Wellsian vocoid is not any kind of a vowel, however short, which I then ought to be able to detect in native Polish speech, then .... well... then I am released (and relieved).

  17. Ad Wojciech

    That's an interesting subject, I'll try to answer your question as extensively as I can.

    I can think of findings from three different areas on which I based this conclusion. The first one is of course, the opinion of an expert phonetician, namely John. It is pretty convincing, but I like to rely on my own judgement.

    The second one was my quick research in etymology, which I've described above – there was originally a short vowel between g and d, which was reduced in the process of disappearance of yers, a thousand years ago. Moreover, the two plosives belong to the single stem, unlike in the English hugged, where g is from the stem and d from the suffix. So, English and Polish gd clusters have entirely different provenance.

    The third and most important of my findings was the careful observation of my own pronunciation while simultaneously imagining the physical behavior of air flow in my vocal tract. I realised that whenever I pronounce the name of my hometown, I quickly release the air pressure (in a closed vowel-like manner) built up behind the tongue-velum blockade – and only then I shut the passage again by pressing my tongue against the alveoral ridge and start building up the pressure again – to be released into a full-blown proper open vowel this time.

  18. Greetings from Gdynia! ;-)

    In normal Polish pronunciation I can't hear nor see (waveform, spectogram) the vocoid you're talking about. I agree, however, that occasionally you can hear the vocoid when someone speaks slowly and thinks up of what he's going to say next.

    As for the consonant cluster, I remember my friend from high school, a railway enthusiast, was sometimes making onomatopoeia of a sound which could be heard from time to time in older EN57 trains. It went like this: dgdgdgdgdgdgdg... He pronounced it with full release of the plosives and an optional short schwa. We competed for making the sound last longer. That was fun ;-)

  19. No mention of g'day yet?

    Interestingly, the German name of Gdańsk is 'Danzig', which seems to reveal that the German settlers in the XIII century simply did not hear the 'g'. Does this make the case for a vocoid more or less probable?

    Unfamiliar consonant clusters undergoing deletion is not unknown in borrowings, though—witness Finnish Ranska for France and Tukholma for Stockholm. Not sure how German would have treated consonant clusters in borrowings, since it had far fewer constraints on consonant clusters than Finnish did historically.

  20. John Wells: "This inevitably gave rise to a tiny transitional vocoid between the two hold phases, but it was not long enough to count as a separate schwa segment (compare the definite schwa in English Godiva)."

    Yes! It's an illusion of our writing system to hide these non-phonemic schwas. Non-phonemic schwas inevitably pop up too in Proto-Indo-European (PIE) where *dʰǵʰōm 'earth' requires a non-phonemic schwa between the two stops. In fact, the phantom sibilant that pops up between two dental stops in PIE (eg. *sedtos 'seat') must be caused by the sibilantization of, yet again, a former non-phonemic schwa that had intervened in the stop-stop sequence only to be crushed into a spontaneous /s/. This /s/ however continued to be non-phonemic in such instances.

    So a language can have both non-phonemic and phonemic schwas at the same time. I seem to recall Nenets having such a distinction too.

  21. Ad nul

    thank you. As for etymology I agree---see "Gydannyzc" in the Vita S. Adalberti of about 1000 AD.

    Re your careful observation: don't you think it applies to 'gdańsk' pronounced slowly, or even very slowly? (Otherwise the observation could not be so careful...). In very slow speech, 'gdańsk' indeed sounds sometimes like 'gydańsk'.

  22. Ad Jongseong

    Finnish tolerates, or used to tolerate, no initial clusters at all---hence ranta for strand, beach, from Old Germanic *strandaz, they say. Neither does Hungarian, compare the Hungarian weekday names with their Slavic prototypes.

    But the thing is that if the Germans of the 12hundreds heard 'gdańsk' as starting with a cluster, no matter how kindly or not they took to the latter, they did not hear any vocoid after the 'g', as little as do present-day Poles. This does not invalidate John's description of how 'gdańsk' is really pronounced in Polish, from the physiological point of view.

  23. 1. Pete, yes, non-syllabic schwa is a good way, thanks.

    2. Wojciech, are you simply contradicting yourself, or have you rethought what you had written further above?

    3. The etymology doesn't play the least role for today's language, might just explain (or not) how it came about.

    4. A factor, also in the issue of speaking slowly, seems to be whether the first part of a two-consonant cluster is a sibilant or a nasal rather than a stop. In the case of voiced stops, though, you can, after all, sound it in a limited way even if no air passes beyond the place of articulation. I suppose this happens by increasing the air pressure between the lungs and the place of articulation. Is that so?

  24. Ad Lipman

    I do not know on which particular point I might be contradicting myself, can you help me? I was not asserting anything, besides, just asking.

    I think nul meant not etymology but phonetic history, in this point there is a role to be played, take for instance what the Germans call 'Ersatzdehnung' (in English?). As I understood nul, he was hypothesising that the hard yer after the 'g' in 'gdańsk', 'gdynia' etc. did not disappear completely but left a trace, kinda. I trace, here is my problem, which I simply am not able to detect, regardless of how hard I try. Unless reference is being made to (very) slow pronunciation. Unfortunately, I know too little of the physiology of articulation to take a stance to all these descriptions and stuff.

  25. It's presumably possible to make simultaneous g and d closures, and to release then simultaneously — even if that's not what the Poles do. After all, isn't something similar at work in the name of the former Ivorian President Laurent Gbagbo?

    To distinguish simultaneous release from sequential release, surely the transcription should assume g release first? Any sound between that and the d release is involuntary. To me it would make more sense for the transcription to mark the consonants as simultaneous in the other pronunciation.

  26. Good news for the friends of a vocoid: I asked my wife, who speaks neither Polish nor any Slavic language, and she said she could hear a short 'e' (i.e. a schwa) between the 'g' and the 'd' in my pronunciation of 'gdańsk', 'gdynia' and similar words.

  27. I thought Gdańsk was ˈgdaɲsk, not ɡdaj̃sk. Wikipedia agrees.

  28. One more reflection: in sloppy speech, we Poles delete the 'g' of 'gdzie' (where), what remains is simply 'dzie' [d͡ʑe], a pronunciation considered vulgar.

    This is similar to Serbo-Croatian, where in everyday speech you may find ’de dê̞(ː) instead of gde gdê̞ and in dialects variations like di dî(ː) and đe dʑê̞.

  29. This comment has been removed by the author.

  30. Duchesse, Wojciech

    Perhaps one of you could set Wikipedia straight.

  31. Hello, I've just read your interesting conversation and would like to comment on the gdaɲsk vs. gdaj̃sk question:

    Based on my speech corpora analyses (and also some publications, like W.Jassem's 'Illustrations of the IPA: Polish') I'd say that both versions might be acceptable and correct in Polish. In very careful speech the first will probably be more often.

  32. This agrees with my knowledge and impressions.

  33. Though gdaɲsk without any nasalisation would be unexpected.

  34. Thank you, Klessa!

    David, no need for editing, it reflects the careful speech, as things should be.

  35. I am not a linguist, still less a normative linguist, just a native speaker, sensitive to language. In my correspondingly (un)informed opinion, the form gdaɲsk is not correct, not even in careful speech, but what we call 'hiperpoprawna', i. e. hypercorrect (not sure if tis word exists in English, sorry you native Anglophones), I think you call it 'spelling pronunciation' in English. The only correct form, I mean the really naturally-sounding form, is the one used by John, i. e. gdajsk with a nasal 'j' (yota). Of course Klessa is right in claiming that even the incorrect, unnatural form gdaɲsk is 'acceptable' in the sense that people will understand it, and, lacking linguistic expertise, only on reflection will find it funny, and only on second, third reflection will be able to say what exactly makes it funny.

    Polish Catholic priest are obviously taught, or used to be until recently, to employ this kind of hypercorrect spelling pronunciation in liturgical contexts --- presumably to sound solemn. Unnatural, _ergo_ solemn. 'Unnatural' does not mean 'careful' here. The word 'gdzie' pronounced with a 'g' does sound careful but not unnatural.

    Generally, in Polish a sequence like: a vowel+an n or m-like consonant+s or z does not exist. The middle element always dissolves into something vowel-like though not syllable-forming and nasal. Beg your pardon for my ignorance of the right terminology.

  36. For what it's worth, here's what's available on Forvo:

  37. The forvo-pronunciation of 'gdańsk' by macyisbetterthanyou strikes me as foreign, non-native. polka's is spelling pronunciation, I'd say, the way ingenuous people think words ought to be pronounced (as they are written). The other 3 are OK, the first (bossanova) is the clearest and the most natural --- imho.

  38. @Duchesse

    I, a Polish native speaker, agree completely with Wojciech on the pronunciation of "Gdańsk", etc. "Gdańsk" is pronounced ɡdaj̃sk in Polish. Some books write it ɡdaĩ̯sk, but it's the same.

    Polish orthographic "ń" when followed by "s" or "sz" is always pronounced as and makes a nasal diphthong with the preceding vowel.

    The letter "ń" in Polish has also a third pronunciation jn (or i̯n). I wasn't aware of it until recently I've read about it in a book. It occurs mainly before "c" and "cz". For example, the words "słońce", "wieńce", "kończyć", are realized as ˈswɔjnce, ˈvʲjɛjnce, ˈkɔjnt͡ʂɨt͡ɕ.

  39. On Polish nasals:

    I used to make the generalisation that Polish orthographic sequences of vowel + nasal consonant + fricative are pronounced as nasalised vowel + fricative.

    Then one day I was sitting in a notary's waiting room in Bełchatów and my eyes came to rest on the notary's name written on the fancy glass door - Galimska. Then the penny dropped that my rule didn't work for bilabial nasals. There's no nasal vowel there, just vowel + nasal + fricative.

    That's very obvious, of course, but such combinations are rare, so I didn't have the chance to notice it earlier. Are there any Poles out there who can think of a word with such a combination that isn't a proper noun?

  40. komża (surplice); also various -ski-adjectives, such as karaimski, Karaite, chamski, chamstwo (churlish, churlishness...), zemsta (revenge),

    But the rule doesn't work with 'n' either, if the fricative is not a sibilant, e.g. kanwa --- canvas.

  41. This is hilarious. It's ˈgdaɲsk, full stop. We don't want to repreoduce peasant sounds from Polish villages, only the correct, standard Polish. Which is with a palatal nasal.

  42. @Wojciech - I was afraid you were going to say something like that. Because, whereas I agree with you about macyisbetterthanyou and Bossanova, the middle two, zic84 and Kamila, I hear as pronouncing the word with a ɲ.
    What do others say about these Forvo recordings?

    Also, Duchesse, with some of the attitudes you keep presenting in your comments, I can't help but wonder about your credentials, and general professional background.

  43. Late to the party, so just two and a half points:

    (1) If you have a problem believing there's some kind of vocoid in Polish /gd/, (1a) download the soundfile of Gdynia John linked to, (1b) open it up in e.g Praat, and (1c) you will see a rather clear vocoid (about 0.07 sec in length) between the two stops. (If you are a "native" speaker of Polish, try to decide if the whole pronunciation sounds "natural". It does to these "native" ears. Granted, it's slow and careful, but not "weird" or anything.)

    (2) I routinely use this as one of the ways of teaching the English schwa to Poles. Otherwise, more often than not, you tend to end up with full vowels as "cued" by the spelling.

    (3) Duchesse -- what Gassalasca said.

  44. @Wojciech

    Thanks for all those great examples. Makes me wonder why I needed a notariusz for it to click.

    You're right about specifying the type of fricative too. I'll have to update my wording.

  45. @Wjarek

    I've been thinking of a way to teach Poles schwa, but haven't had a chance to try it out.

    What about the contrastive/citation version of the prepositions 'w' and 'z'? Don't I hear a schwa there? It may be a close one, but not Polish /y/, surely?

  46. Ad Duchesse

    it is gdaɲsk, full stop.

    No, it ain't... . gdaɲsk is something like glow-sester, or wur-sester, or chawlmawndilay, or fethther-stone-haw, or bucking-ham, something like that. Priests say, perhaps, gdaɲsk, when they want to sound solemn. But they also say 'wziął' or 'wzięła' (took) with a nasal vowel, o horror.. Peasants (Kashubian) say 'gduńsk',any way.

    Ad Gassalasca

    Frankly, I can't hear clearly how no. 2 and no 3 on forvo pronounce gdańsk. I seem to---seem to---hear a nasalised yota. But folks reading on forvo are mostly linguistically ingenuous, not just in the sense of not knowing much terminology (neither do I) but in the sense of cherishing various naive ideas about language, such as that words' true sound ought to correspond to how they're written.

    Ad wjarek

    I of course don't question what youguys linguists say. How would I dare, not being one? I only establish that I cannot hear any schwa in natural pronunciation. The gdynia in John's posting strikes me as a bit 'wysilona' 'bemueht', as the German says (English?), not really natural. But if you say there is a schwa there, well, then there is surely one.

    Ad Paul Carley

    why notariusz? Probably you were bored of waiting and allowed your mind, unharnessed, to wander about for a while. Works with me too this way.

    I consider such locutions as 'z godziny na godzinę' or 'z wtorku na środę' or 'w domu lub w pracy' and I can't hear, sorry to say that, any schwa, or 'y'. You certainly know there are forms 'we' or 'ze', conserving possibly an old yer (nul? are you there?) such as 'we wtorek' or 'ze Zdzichem' or 'ze dworu' (two different 'z/ze's!) for euphonic reasons, with a full-blown 'e' (close to DRESS).

    One way of teaching Poles schwa is reminding them of that funny song in which they sang: 'ptaszki, petaszki...'. Or making them recognise the difference between the onset of these two words: 'police please!' They can hear it, as experience teaches.

  47. Wojciech,

    peasants (Kashubian) - yes, and peasants (German) say dancyk.

    wysilona - forced, in this case possibly genteel

    z, w - wjat Paul means, if I understand him correctly, is the isolated form. What do you answer when somebody asks you "How do you say in in Polish?"? Syllabic [v], or [v] plus schwa?

    (As an aside, the quotation forms of the letters are the only case where a schwa appears in Czech.)

    Police please - unfortunate example for many speakers.

  48. Ad Lipman

    Thank you. Dancyk I dunno, haven't ever heard. Not -ich, _etwa_? Which German peasants are you referring too? Those of Żuławy, perchance?

    'In' in Polish is 'v', I'd say. Without a schwa. There is 'vy-', meaning either 'you (guys)' or 'out -of' as a prefix, Indoeuropean ud-, they say. In Polish, we are not even able to pronounce a schwa, I'd insist, without, in the middle of Polish speech, suddenly switching into a language that has it: German, Russian perhaps, English, French, Slovene, Hebrew (ancient, at least), Proto-Indoeuropean (if our command of the latter happens to be good enough).

    Why is 'police please' an unfortunate example?

  49. Many (most RP NSS?) have pliːs for police.

  50. @ Wojciech

    You really don't recognise an exaggerated contrastive pronunciation of 'z' and 'w' with a short schwa after? Maybe it only occurs in very specific contexts.

    This is how it came to my attention. I was at an optician's in the UK with a mother and daughter. The mother was having her eyes tested and because her English is very weak, she had brought her daughter, and I had come along because the daughter's English wasn't so hot either.

    So during the eye test both me and the daughter were interpreting. We came to the part where the optician was asking 'better with or without (the lens), with or without, with or without' which I translated as 'z czy bez'. I pronounced 'z' as a single consonant, which felt a bit strange in my mouth, but I thought that must be the Polish way. I noticed, however, that the daughter was pronouncing it with a following vowel. A very short vowel, but very distinct from a lone [z].

    I asked her about it later and confirmed that that was what she had said, that it was distinct from /z/ and /zy/, that for the context it was acceptable and that the same can happen for 'w'.

    Another thing, Wojciech. Since you seem to know your stuff, maybe you could solve a puzzle for me. I'm used to /e/ being inserted in Polish to avoid tricky (even for Poles) clusters, but are you aware of any explanation for the form 'ze soba'? That's the normal form, isn't it? Peculiar...

  51. Ad Lipman,

    So I have antiquated pronunciation in my ear, clearly. It's been ages since I visited any English-speaking country last...

    What about 'monotonous' though? Do the RP NSS say 'mnotnes' without any schwa at all? I was taught, like 40 years ago, to articulate 'm-schwa-no-t-schwa-n-schwa-s. Not _salonfaehig_ no more? Sorry this is getting monotonous indeed...

    Ad Paul
    Was the daughter saying 'ze' or what? Was it like the English schwa? I'd say 'z' without any vowel, personally. Maybe it's an articulatory habit of some persons, no part of the 'system' of the language.

    Ze sobą --- yes, it's common, it dissolves the 'ss' in to 'ze s', so it's easier, since Polish, tho' it knows geminated consonants, doesn't like them too much. Historically, it may be a left-over of a hard yer, a short back-ish vocoid, which dissapeared in very second syllable like in the 12th century. But I am not 100 p.c. sure here, ask nul perhaps. Originally, it was z-hard-yer, v-hard-yer, plus an n, cum/syn, in, or something like that, being what these words come from.

    But methinks that 'z sobą' ('ssobą') is equally correct. Or 'z soboty' alongside 'ze soboty'. 'W wtorek' ('fftorek') would not, strangely, nobody says so. The without-e forms are, possibly, more recent.

  52. "Mnotnes" is a different case, with or without vocoids. Police has the monosyllabic form firmly lexified, it's not an optional or contextual ad-hoc elision. (John might correct me if I'm wrong here.)

    Ze sobą - yes, it's historic, a leftover of ŭ. That still means, though, that at an earlier stage of the language, it wasn't elided between sibilants, just as it wasn't before clusters. (This doesn't necessarily mean the same phonotactic restrictions apply in today's Polish.)

  53. @ Lipman

    I wouldn't say lexified. I wouldn't be surprised to hear 'police' with two syllables. There may be a tendency to make it a single syllable, but the orthography will always be pulling in the other direction.

    The current situation with 'medicine' seems to be moving away from the traditional 2 syllable version, to my ears at least. I find 2 syllables recommended in LPD - time for a pronunciation preference poll?

    And did I imagine it, or isn't it said that the 2 syllable version is more used with the meaning of 'medical studies'?

  54. Regarding the citation forms of Polish z and w, I can imagine both a "long-plain-consonant" and "consonant-plus-tiny-schwa" versions. I can't really decide what I do myself. But "naive" Polish speakers will not recognise that mini-schwa as a bona-fide vowel.

  55. BTW, at one time, Polish kids were taught not to use "letter names" (such as be, de, ka etc.) to refer to letters during writing instruction in kindergarten and first year in the primary school. What they (and the teachers) did instead for plosives was consonant-plus-mini-schwa (voiceless in the case of the voiceless ones). No idea if that's current practice any more.

  56. Paul,

    sorry for the misunderstanding - I didn't mean to imply that pliːs (or mɛds(ɪ)n) were lexified instead of the "fuller" versions, but that they were lexified (next to the fuller versions) and not ad-hoc reductions. The latter are possible even if the resulting cluster isn't in the core repertoire. Those who say pliːs (or mɛdsɪn) might never reduce extraordinarily to anything less than seven syllables or monotonous to less than four.

  57. @Wjarek - what you describe is what is usually done in Serbian. There is also the option of saying be, de, ka etc. but it is rarely used. For all intents and purposes, what is called bi: in English is called in Serbian (unless we are talking about the insect, of course :).

  58. This comment has been removed by the author.

  59. Ad Wojciech

    Thank you for your earlier correction, phonetic (and phonologic) history is exactly what I meant by what I'd poorly chosen to call 'etymology'. In this regard I consider historical sound change to be an important clue – not as a direct evidence, but as a means to broaden my perspective, e.g. knowledge of Great Vowel Shift helps me understand the rich and complex English vowel system, and disappearance of yers might shed some light on Slavic consonant clusters.

    As for the form 'ze sobą', however, I can't confirm its historical origins. It's early form was *sъ soboją, and as *soboją begins with a full length vowel, the yer was in a weak position and disappeared altogether:
    *sъ soboją > z sobą

    However, if we chose the first person singular pronoun in place of the reflexive one, the case would be different:
    *sъ mъnoją > ze mną
    The yer in *mъnoją underwent deletion and, as we treat the phrase as a single unit, the yer in *sъ
    became a full length vowel in the process of compensatory lengthening.

    Nowadays 'ze sobą' is a more popular form than 'z sobą' (this indeed may be due to general lack of trust in gemination) and both versions are considered correct. In case of 'ze mną', however, it's the only correct form, and it's incorrect to say 'z mną'. And it's not the case of an easier pronunciation as we don't have any problems with saying words like 'zmniejszyć'.

  60. Ad wjarek,

    the practice you describe (calling characters by their phoneme, not their traditional name) is known to me as an erstwhile schoolboy only to well. It was well-meant: lest the name of the character should confuse the child as to the phonetic value of the character. But we schoolboys heard not a schwa, but a 'y'-vowel, this characteristically Polish vowel, high, midway between front and back, crossed-i, y'know... . So it was by, cy, dy, fy, etc. No means restricted to plosives. If it still is applied I ignore.

    As for the schwas, if there be any, in g-schwa-dańsk and so on, please sir have mercy: we naive Polish speakers simply can't hear them, no matter how obsequiously we defer to the authority of youguys schooled phoneticians. So we have trouble with them not _qua_ bona-fide vowels, but _qua_ any-fide vowels, be she (the fides) bona, mala or what not. We simply don't perceive them. You're like a physicists getting furious with a layman for failing to see atoms. 'But sir, they're are too small.'

    Ad Lipman

    maybe the lexification of plice and medsine is due to the frequency of these words (high), as distinct from the low frequency of 'monotonous'? As is maybe too the frequent schwa-deletion in 'suppose', a fairly frequent word in English, I s'ppose (tho' the Yank says 'I guess'...).

    Ad nul

    So the 'ze' in 'ze sobą' is borrowed as it were from 'ze mną?' Reinforced by the low trust in gemination?

  61. @ nul

    Thanks for pointing out that the /e/ in 'ze mna' isn't just there to simplify clusters. Your example of 'zmniejszyc', however, doesn't quite give us the same phonetic context. I can't find any words in my dictionary that begin with /zmn-/. Any suggestions?

  62. Ad Paul Carley

    I am not sure now but it seems to me 'zmnieść' would be the perfective aspect of 'miąć', to crumple. Maybe I am wrong -- 'zmiąć' is correct(er).

    People say sometimes 'zmnieść' for 'to bear, to stand' or 'zmnienić' for 'zmienić', to change. This is all wrong, but occurs here and there. Also, 'zmniana' for 'zmiana', change.

    There is, too, 'zmrozić', to freeze, 'zmleć' to grind, all perfective aspects. The perfective aspect prefix z- is pretty productive, provides too for many interesting consonant clusters.

  63. @Wojciech

    OK, my usage of "naive" was uncalled for. I'm sorry! But -- as I'm sure you know -- it does exist as a term quite widely used in linguistics texts, possibly in the form of "linguistically naive speakers" etc.

    Also, pointing out that people who are not "instrumental phoneticians" (hobbyist or otherwise) may have trouble with this type of phenomenon (and many others) was in no way meant to be condescending. I was just pointing out the physical evidence. I'm not getting furious with anybody, far from it.

    I have myself learned a lot about Polish -- stuff I'd had no idea about as a "native speaker" -- from a bit of instrumental work I've done. And yes, I do realise you can't see atoms with the naked eye.

    Sorry once again if it struck a nerve.

  64. A lot of Polish people (at least those under 30) spell out words by saying "by", "cy", "dy", "fy", "gy" (, t͡sɨ, , , ɡɨ), etc. They simply add "y" (ɨ) to the standard pronunciation of a letter. It's true that this "y" may sometimes be shorter. If short, then it may even tend towards a more open quality, but it is still central and unrounded.

    This manner of spelling out words is used for all consonantal letters. Even today one of my colleagues was spelling to me a word like this instead of using the standard letter names.

    As for "z czy bez", I agree that the "z" may be pronounced as a letter spelled out in the above manner, with a following short vowel.
    Without the vowel, the "z" word would sound like s, because of word-final devoicing. But in this particular contrastive context, such unvoiced pronunciation would sound unnatural. So it turns out that the vowel after "z" in "z czy bez" is obligatory. This seems very strange to me, but that's my final conclusion.

    The only Polish words with zmn (ex. "rozmnożyć" and cognates) have a syllable boundary there (rɔzˈmnɔ.ʐɨt͡ɕ), so they shouldn't count, I think.

  65. @ahven

    I came across 'rozmnozyc' when I did a quick search of an online dictionary, but discarded it for the reason you suggest.

    I got the impression that it was a close schwa, so maybe we've got something similar in mind. But such a form of 'w' would be pretty distinct from 'wy', though, wouldn't it? That's what my informant seemed to think.

    As you say, maybe pronouncing a fully voiced fricative as a complete word and it being 'word final' was what made it feel awkward for me and made me listen out for how the natives do it.

  66. Duchesse

    My ears have an unfortunate tendency to hear what they expect to hear. When I was living and working in Poland they initially expected to hear gdansk and then, fed with a little more information, they expected to hear gdaɲsk. But this was so obviously not what people were saying that in time I even I could tell that it was something like ɡdaj̃sk.

  67. Ad wjarek

    I was not offended by anything -- so no need for apologies --- and I do consider myself naive in the sense you made explicit, indeed, very naive.

    My point was and is of methodological nature. (I am generally interested in the methodology of various sciences.) In an exchange between a specialist and a layman there comes a point when the layman says: 'I do not understand'. That is because he is simply not schooled enough or too stupid. This happens often with abstract theories---I experience this often when I try to explain something from set theory to the unprepared. OR the layman says: I understand what you're saying but --- I can't confirm your conclusions on the basis of my experience. This happens in more empirical fields of study, for instance phonetics. You guys have your expertise, whose value I of course fully respect, I by contrast have my ears and many years of exposure to various languages.

    I once knew a chap who claimed to literally see the movement of the sun along the ecliptic. Well, lucky him; I only know the sun moves (or seems to) along the ecliptic, but I can't see the apparent movement (other than by comparing the positions of the sun now and in a month's time, say). Here: they say, there is a schwa between the 'g' and the 'd' in 'gdańsk'. OK, I believe they that you have all kinds of sophisticated reasons to say so, but I simply can't hear it---no madder what they say, as Sting once sang. A good empiricist must be a bit stubborn, hard-headed, needn't he?

    A discussion I have recently had on the existence of long and short vowels in Italian is a different matter from the methodological point of view. Germanophone and Anglophone scholars seem to sordda believe in vowel quality in Italian, I firmy hold this is a mistake, resulting from a projection of some of their native tongue's properties on the language of Dario Fo, plus --- er, I daresay ---a kind of too theoretic, too scholarly knowledge of same, what the German calls 'Verkopftheit', too little first-hand everyday familiarity with Italian, as it is really spoken at home and at large (now you recognise I must have been a diligent reader of Dr. M. Luther's famous 'Letter on Interpretation'). We--those scholars and I---hear the same, i.e. Italians lengthening sometimes their stressed vowels, but to them (the scholars), it's vowel quantity, to me it's an articulatory habit (or sumping of this sort) situated on a different level of language than that on which it is situated in languages which really have vowel quantity, such as Czech, or Finnish, or German, or British English, for that matter. Here not the testimony of the senses is decisive but a theory in the background. And of course theirs is sophisticated while mine is naive, inchoate, intuitive... . And yet, I can't help feeling they are wrong and I am right. In the case of the Polish vocoid between initial 'g's and 'd's I don't have this kind of strong opinions; maybe there is one, some persons seem to hear one, OK, all the better. Vive le vocoid! Not without reason did Rimbaud write his famous poem 'Les Voyelles'.

  68. Ad Wojciech

    I'd rather say that Proto-Slavic preposition *sъ (with) has evolved into two different forms (a feature shared by many other prepositions in the most of the Slavic languages): z and ze; moreover, it has merged with another preposition *jьz (from, out of). Both forms can be sometimes used more or less interchangeably (like ze sobą and z sobą, but in some certain phrases, one form sounds just plainly wrong (like z mną), even if you don’t happen to be interested in obscure sound changes that occured a thousand years ago.

    Nevertheless, apart from those certain traditional phrases, the most important factor now is, in fact, the ease of articulation (as we say ze związkiem rather than z związkiem, which even for Polish speakers sounds awkward – we tolerate geminated consonants in intervocalic position only); prosody also matters (and for poets it’s especially handy to have an optional syllable in a phrase).

    Note also that in casual speech we can clamp together much more complex consonant clusters like Trakt Gdański, where you have ktgd cluster pronounced with four consecutive stops. And if someone asked me if there’s any vowel between them, I’d say of course not – but, if it helps – then yes; but for me it’s just a sequence of non-overlapping plosives, so if I were to advice, just say them trippingly on the tongue :-)

    Ad Paul Carley

    You're right, that was unfortunate example, just the first one that came to my mind. Much better examples are similar prepositional phrases like z mnichem, z mnóstwem. Alternate forms: ze mnichem, ze mnóstwem are much more rarely encountered, as consonant cluster breaking is simply unnecessary there.

  69. @nul

    I once saw a copy of a Polish book on Polish phonotactics on I wish I'd bought it now! It's a bit tricky, though, to pay for things and get them delivered to the UK.

    By the way, your 'z mnichem' example suffers from the same problem as your earlier example - the two nasals are bilabial and palatal, not bilabial and dental, as in 'mna'.

    "we tolerate geminated consonants in intervocalic position only" - an interesting point. How about 'ssac, ssanie, ssak' with the geminate in word-initial position. Is there any tendency to simplify to a single /s/?

  70. Ad Paul Carley

    Thanks for pointing out my mistakes. I was just thinking lexically and I should be more aware of the difference between n and ɲ (I think that would be more accurate).

    I was also clearly wrong to write that gemination occurs only in intervocalic position as you've provided the examples of geminated s in initial position. What I meant was that we avoid preconsonantal geminates when we choose the form of a prepositional phrase. Hence w wojsku, w Warszawie, but we wtorek, we Wrocławiu

  71. dżdżu, e.g. krople dżdżu, raindrops. 'dżdżu' seems to be an archaic genitive of 'deszcz', but as kid I seriously thought the nominative case of 'dżdżu' was 'dżdż'.

    Today I caught myself using a schwah after the 't' in the German word 'vielleicht' (possibly, perhaps) when uttered in hesitation. 'vielleichty'. The word, by the nature of things, is in fact quite often said in hesitation.

    Part of the problem, methinks, is that you linguists try to codify too much, to class various things as parts of the system of a language, while they are just part of the rough-and-ready, quick-and-dirty use that people make of that system. Adding a schwah here and there, drawling a vowel, 'singing' a word and great many other things. Doing things TO words, to paraphrase the famous title by J. L. Austin. You guys would straight off like to see it as part of the system, codify it in your API and what not...

  72. 'Gedynia I love you' is what foreign artists tend to say here. I heard it from Snoop Dogg first (and as a result moved to Gdynia) and last month it was nonpareil Prince who uttered 'gedynia'. I'd say we Poles are used to it and even surprised when a native speaker of English can manage a schwa-less 'gd' cluster. Unless they're phoneticians of course.