Talk:Czech phonology

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search
WikiProject Languages (Rated Start-class, Low-importance)
WikiProject icon This article is within the scope of WikiProject Languages, a collaborative effort to improve the coverage of standardized, informative and easy-to-use resources about languages on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 Low  This article has been rated as Low-importance on the project's importance scale.
WikiProject Linguistics / Phonetics  (Rated Start-class, Low-importance)
WikiProject icon This article is within the scope of WikiProject Linguistics, a collaborative effort to improve the coverage of linguistics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 Low  This article has been rated as Low-importance on the project's importance scale.
Taskforce icon
This article is supported by Phonetics Task Force.


I found this hiding in the article. It looks like it might be useful but it needs translation. I don't even know of the article already covers what this does. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 03:36, 29 December 2006 (UTC)

I intend to translate this into English, but I haven't had enough time to do it yet.--Pajast 13:55, 1 January 2007 (UTC)

Alternace samohláskových fonémů[edit]

Délka (kvantita) samohlásek patří k základním distinktivním (rozlišovacím) rysům ve fonologii češtiny. Střídání (alternace) krátkých a dlouhých fonémů jsou v češtině běžné při odvozování slov i při ohýbání. V průběhu vývoje jazyka se však některé fonémy měnily (/oː/ > /uː/, /uː/ > /oʊ̯/), proto při střídání krátkých a dlouhých hlásek nemusí tvořit pár dva kvalitativně stejné/podobné zvuky.

Krátký foném Dlouhý foném Příklad, poznámka
/a/ /aː/ zakladatel – zakládat
/ɛ/ /ɛː/ letadlo – létat
/ɪ/ /iː/ litovat – lítost
vykonat – výkon
/o/ /uː/ koně – kůň
/ʊ/ /uː/ učesat – účes (pouze na začátku slovních základů)
/ʊ/ /oʊ̯/ kup – koupit (v ostatních pozicích)

Pronunciation of ř[edit]

The articles Czech phonology and Caron are conflicting in how they explain the pronounciation of the ř letter.

  • From Caron: Ř/ř (only in Czech: special fricative trill /r̝/, transcribed as /ɼ/ in pre-1989 IPA, pronounced roughly as a compound of trilled /r/ and /ʒ/, e.g. Antonín Dvořák About this sound listen  )
    • This explanation explains the fricativeness very carefully. Nothing of that kind was in the Czech phonology article.
  • From Czech phonology: The phoneme /r̝/ (written as <ř>) is an alveolar fricative trill (see alveolar trill) its rarity makes it difficult to produce for foreign learners of Czech. The basic realization of this phoneme is voiced, but it is voiceless [r̝̊] when surrounded by voiceless consonants.
    • Except for the voiceless/voiced thing, this makes the reader think that ř is the same as r in Finnish. Or at least, it does not explain the difference.

I would like to have some clarification to this issue. Either one of the articles should be fixed. Probably the Czech phonology one. --Bisqwit 02:37, 4 March 2007 (UTC)

It is difficult to describe Czech /ř/. It is neither trilled [r] nor fricative [ʒ] (or [ʃ]), it is something in between. Primarily it is a trill as [r], but the tongue is closer to the hard palate, which gives it a fricative character (partially). The vibrations of the tongue are softer and more frequent than in [r]. In the contemporary IPA version, it is recorded as raised alveolar trill, i.e. r with a special diacritic mark ("up-tack").
I do not know what is the realization of /r/ in Finnish. I cannot explain the difference. The voiceless /ř/ is similar to the retrofex /rs/ [ʂ] in Swedish but it is not the same. If you had some better formulation, feel free to improve this article. --Pajast 16:47, 6 March 2007 (UTC)
Finnish /r/ is definitely different from Czech /ř/. Finnish /r/ is pretty the same as Czech /r/ (it could be a bit more trilling sometimes but I guess, it depends much on a particular speaker). Miraceti (talk) 11:11, 29 April 2008 (UTC)

The article says that /ř/ is pronounced as voiceless when it is surrounded with voiceless consonants. I think that it is enough, if it is preceded with a voiceless consonant. For example in the name About this sound Třinec. However, I have no literature to support or disconfirm it. Jan.Kamenicek 15:20, 11 March 2007 (UTC)

Oh, no. The assimilation of /ř/ is both progressive and regressive. Therefore, "preceded" is not enough. If it followed by a voiceless consonant, it is voiceless too. What about words like "buřt, stopařka, mateřství" etc.? --Pajast 16:41, 12 March 2007 (UTC)
Yes, you are definitely right. But it is another argument that the word "surrounded" should be replaced, otherwise it seems that the consonant /ř/ has to be both preceded and followed by voiceless consonants at the same time to be voiceless too. And there is even one more possibility, when it is at the very end of a word, for example "Jaroměř". Therefore I suggest to rewrite the sentence as follows:
"The basic realization of this phoneme is voiced, but it is voiceless [r̝̊] when preceded or followed by a voiceless consonant or at the end of a word." Jan.Kamenicek 18:36, 12 March 2007 (UTC)
I don't think at the end of a word is accurate, because when a word is followed by another word whose first phoneme is a voiced obstruent, /ř/ behaves just like the other obstruents - it undergoes voicing, cf. [ˈbiːlaː ˈjaro̞mɲɛr̝̊] vs. [ˈjaro̞mɲɛr̝ ˈbiːlaː ].Pětušek 23:35, 21 March 2015 (UTC)

OK, no problem. --Pajast 15:53, 13 March 2007 (UTC)

Can ř be considered as an apical vowel? (As in Chinese pinyin?)- Hello World! 07:25, 3 November 2008 (UTC)

By no means. It's an obstruent. Pětušek 23:35, 21 March 2015 (UTC)

I hope somebody will be so kind as to enlighten me (I'm Italian, and I'm learning Czech). I'm under the impression (but I may be quite wrong, of course) that the articulation of /r̝/ is the result of two simultaneous articulations: 1) an apico-alveolar trill, [r] (as in Italian etc.), and 2) a laminal fricative, i. e. a [ʒ], but less retracted than the normal [ʒ] of, say, French. So: both the tip and the blade of the tongue are involved, and there are two contemporary articulations. In other words, if I'm right, the frication is not produced by the vibrations of the tongue, but is another, though simultaneous, gesture. Might this be a good description? Do at least some native speakers articulate /r̝/ like that, or roughly like that?

Actually, the friction results from the trill being raised (hence the IPA diactritics), which makes the turbulences stronger as the constriction is tighter. The vibration of the apex thus competes with the friction, so to speak, the latter being easier to maintain. As a result, the trill is usually only realized as a single flap, although in careful speech two or three flaps may be present (and more in exaggerated, expressive or fun contexts). Hence, it really is a single gesture, not a combo of two. Also, as you need the apex to vibrate, I don't think it could be described as laminal, although the mass of the tongue is raised to facilitated the pronunciation. It is also true, however, that it's slightly further back than simply (pre-)alveolar, closer to the post-alveolar region in fact, so little childer often replace it with [ʃ, ʒ] before they learn it (in most Polish varieties, the sound has eventually evolved into retroflex fricatives). So, I (and most of my colleagues) would probably describe the sound as, well, (post)alveolar (apical) fricative (and thus non-sonorant) trill. As you know, the basic realization is voiced, but in assimilative environments (next to voiceless obstruents) it gets devoiced. I hope this helps a bit. :-) If you do want to learn it, most speech therapists teach the sound via regular [r]: learning how to make the apex vibrate is just one step closer to making it vibrate faster. Sometimes, learners do learn <ř> before they have learnt proper /r/, because, auditorily speaking, you mostly only need the right type (colour, frequency) of friction and a lucky single flap, most of the time. ;-)
If you also want to learn the "right" /r/, speech therapists generally recommend replacing all the r's with d's. Once you have managed to weaken the pronunciation of the d's to mere flaps/taps, you've got it! ;-) So, make a list of a few dozen words containing <r>, and try to pronounce those with /d/ first. Later, you can focus on words with syllabic <r>, such as <trnka>, <trpět>, <trdlo>, <drtit>, <pudr> etc. Of course, they will be difficult at first, but, for starters, it is really ok if <trnka> sounds like [tədəŋka] at first, or if <drtit> sounds as [dədəcɪt], or <pudr> as [pudədə]. Hopefully, with enough time and practice, you will achieve [təɾəŋka], [dəɾəcɪt] and [pudəɾə], and later [tɾ̩ŋka] (or [tɾ̩ɾŋka] = [tr̩ŋka]) and so on. ;-) Hope this helps a little. --Pětušek 09:33, 17 September 2017 (UTC)

And one more question, if I may: I suppose that in Czechia too there are some speakers who, like me, are unable to articulate the standard [r] (<r>), and substitute it with some kind of French-style r, such as [ʀ] or [ʁ]. So, how do these speakers pronounce <ř>?

Or, are there any acceptable, or at least tolerable, alternative articulations for those who, like me, are unable to pronounce [r], and consequently also [r̝], correctly? I mean: not the sequence [rʒ]. Perhaps a retroflex [ʒ]?? I have no idea, and I'm at a loss! It's important for me, but it's also important for this article, of course (so, don't tell me that I'm using this discussion page for my personal doubts ;)). Thank you very much, at any rate. Pio d'Ausonia (talk) 15:40, 7 September 2017 (UTC)

Usually, a wrong pronunciation of "r" (rhotacism) does not mean a wrong pronunciation of "ř" (Czech rhotacism) and vice versa. Of course, there are children who have problems with both of them. It is interesting that rhotacism is much more socially acceptable than the Czhech rhotacims. Adults who cannot pronounce "r" are therefore much more common than adults who cannot pronounce "ř". Miraceti (talk) 12:19, 9 September 2017 (UTC)
Thank you very much, Miraceti. Does that mean that in the articulation of [r̝] there is not also an apico-alveolar gesture? Pio d'Ausonia (talk) 11:12, 13 September 2017 (UTC)
This question is to be answered by somebody else. Sorry. Miraceti (talk) 20:07, 14 September 2017 (UTC)
It varies, really. Some of those who can pronounce the former cannot pronounce the latter, and vice versa. Anything that shares enough phonological features (by means of phonetic information) to make the outcome comprehensible counts. Read my instructions above to see if they are of any help to you. I would definitely recommend retroflex fricatives along the lines of the Polish ones, as they contain friction (albeit different) and also evoke the impression of rhotacity (to me, at least). I still think, however, you're gonna be able to learn both <r> and <ř>. Small steps, remember! ;-) [(ə)d(ə)] → [(ə)ɾ(ə)] → [(ə)r(ə)r] → [rʒ / r̊ʃ] or [rʐ / r̊ʂ] or similar → [r̝ / r̝̊]. Oh, and one more thing: with <ř>, the lips are neither rounded nor protruded; unlike with the other post-alveolars, <š>, <ž>, <č> and <dž>, which are somewhat rounded and could be written as [ʃʷ], [ʒʷ], [t͡ʃʷ] and [d͡ʒʷ], respectively, in a narrower phonetic transcription, the lips are rather spread. --Pětušek --Pětušek 09:22, 20 September 2017 (UTC)
Thank you very much indeed. Your explanations are extremely clear and precise, and very useful to me. I'm studying them carefully, and I'll try to put into practice your advice. The articulation of /r̝/ is now quite clear, in every particular (before, it wasn't!) ;-) Pio d'Ausonia (talk) 22:20, 21 September 2017 (UTC)

consonant assimilation[edit]

Could someone please explain this in more detail? For example, are the z and h in nezhasni voiced or voiceless and is this so according to some "rule"? --Espoo (talk) 18:09, 22 April 2008 (UTC)

According to this rule (reggressive assimilation), this group is voiced, because h is voiced, so the whole group (zh) is voiced. --Pajast (talk) 13:51, 24 April 2008 (UTC)
So is the group sn also voiced? Is it the bad recording here or is Lucia Popp singing all or some of these as voiceless consonants? --Espoo (talk) 18:51, 27 April 2008 (UTC)
Continue reading below the table. Sonorants (/m,n,ɲ,r,l,j/) and /v,r̝/ do not trigger regressive assimilation, so "sn" is still /sn/. — Emil J. 15:40, 23 July 2009 (UTC)


The phonemes /f/ , /g/, and the affricates /d͡z/ and /d͡ʒ/ occur in words of foreign origin only. Phonetically, the affricates can occur at morpheme boundaries (see consonant merging below)

Please compare to [1]: V domácích slovech se znělé afrikáty vyskytují pouze v trsech před znělým konsonantem – potom ovšem můžeme v tomto případě znělé afrikáty chápat jako pouhé znělostní varianty neznělých afrikát (džbán). Translation: Voiced affricates of home origin appear in groups in front of voiced consonant - then one can of course understand these voiced affricates as mere voiced variants of unvoiced affricates (example "džbán").

It looks pretty similar but there is a significant difference: At least /d͡ʒ/ appears also in words of the Czech origin. Rarely but it does. The affricates just do not appear in front of vowels. The word džbán is not a product of consonant mergning, is it? Miraceti (talk) 10:48, 29 April 2008 (UTC)

Well, I think the confusion is in part due to some minor inconsistencies in the orthography. Consider the fact that even if you spelt the word as čbán (phonologically /t͡ʃbaːn/), which is actually how the word is attested in Old Czech, the voicinɡ assimilation would render it as [d͡ʒbaːn] anyway. Moreover, its etymology shows it was /t͡ʃ/ originaly, indeed (Proto-Slavic */t͡ʃĭbanŭ/). I think, however, we could at least exemplify the phenomenon:
[d͡z] occurs as an allophone of [t͡s] due to (regressive) voicing assimilation only:
At morpheme boundaries within a word, e.g. leckdo, leckdy,leckde [lɛd͡zɡdo, lɛd͡zɡdɪ, lɛd͡zkdɛ] (the prefix /lec-/ meaning "every, any", and "kdo", "kdy" and "kde" meaning "who", "when" and "where", respectively)
Word-finally before voiced obstruents, e.g.moc by chtěl [ˈmod͡zbɪ ˈxcɛl] ("He would like [that] a lot."), víc doleva [ˈviːd͡z ˈdolɛva] ("further to the left")
[d͡ʒ] occurs as an allophone of [t͡ʃ] due to (regressive) voicing assimilation only:
At morpheme boundaries within a word, e.g. tlučhuba ("gasbag") ['tlud͡ʒɦuba]
Within voiced consonant clusters, e.g. džbán ("jug"), džber ("bucket") [d͡ʒbaːn, d͡ʒbɛr] (from original /č/ in both cases)
Word-finally before voiced obstruents, e.g. proč by chtěl ['prod͡ʒbɪ ˈxcɛl] ("Why would he want/like?"), rýč do země [ˈriːd͡ʒ ˈdozemɲɛ] ("a spade into the ground")
It's true, however, that, thanks to the loanwords, /d͡ʒ/ is slowly becoming an independent phoneme, entering opposition with its voiceless counterpart, cf. džíny "jeans" vs. Číny "China" (gen./ or džin "jin" vs. čin "deed, act" or džem "jam" vs. čem "what" ( etc.--Pětušek 09:28, 20 September 2017 (UTC)

/f/ and /g/ occuring "only in foreign words"[edit]

(related to same section). Call me ignorant but does /g/ not occur in the word "kdo" as a result of voicing the cluster? I'm pretty sure that's not of foreign origin. - filelakeshoe 21:05, 26 January 2009 (UTC)

Keep in mind the difference between /g/ (with slashes) and [g] (with brackets). As a contrastive phoneme (with minimal pairs and all that), /g/ only occurs in loanwords but [g] may also occur as a result of voicing assimilation of /k/, as the spelling suggests. At least, that's how I'd describe it. — Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 05:31, 27 January 2009 (UTC)
That's right. In the native lexicon, [g] is always an allophone of /k/ only.
As for /f/, however, it's not quite the same: It is still pretty marginal, but it does occur in two native words: doufat ("to hope") and zoufat ("to despair"), both derived from Proto-Czech *upvati (*/pv/ → /f/) "to believe strongly".--Pětušek 09:27, 20 September 2017 (UTC)

Velarized lateral in Czech[edit]

What is the exact IPA value of Czech <l>. It clearly sounds different than Polish or French <l>. May it be velarized? —Preceding unsigned comment added by (talk) 20:00, 24 May 2008 (UTC)

No, it should not be velarized in the standard variety. It's true, however, that especially the western varieties (especially Common Czech) do tend to velarize it, and the situation can be even more complex in some estern Moravian/Silesian varieties, where the velarization is conditioned positionally. Hence, orthoepically, it should be the normal alveolar lateral /l/, with the velarized variant /ɫ/ occurinɡ reɡionally and sometimes individually.Pětušek 10:58, 22 March 2015 (UTC)


One Confused user keeps reverting the transcription to one which distinguishes the height of the u-like vowels [ʊ,uː], instead of [u,uː]. The last time he finally gave something resembling an argument in his edit summary: "rv. It's sourced, so se the source." Apart from the fact that he kan't spel, this apparently tells us to look in the references mentioned in this article as well as cs:Fonologie češtiny for support of his theory. Well, unlike him, I did. It cannot hurt to know what the sources say on the matter, so I looked up all I could find in a library, and these are the results. Note first that none of the sources use IPA regularly, they employ the traditional Czech transcription based on letters of the Czech alphabet, so short vowels are denoted by [a,e,i,o,u], and long vowels by the same letter with a colon or an acute accent [a:,á]. Translation is mine.

  • M. Krčmová, Úvod do fonetiky a fonologie pro bohemisty. These are lecture notes (skripta) of the University of Ostrava, my library does not hold it. However, I checked other work by the same author:
  • M. Krčmová, Fonetika a fonologie. Zvuková stavba současné češtiny, Masarykova Universita, Brno, 1996, ISBN 80-210-1376-1. On pp. 91–92, she writes: Long vowels usually have formant structure very similar to the structure of a short vowel; the largest difference is for [í], where formants are raised by up to 10%.
  • M. Krčmová also maintains on-line lecture notes for her course on phonetics: [2]. Remarkably, she has a nice IPA chart for Czech there, which, as you can see, promotes IPA [u,uː] for Czech u-like vowels.
  • Z. Šiška, Fonetika a fonologie. These are lecture notes of the Palacký University in Olomouc, my library does not hold it.
  • T. Duběda, Jazyky a jejich zvuky. Univerzálie a typologie ve fonetice a fonologii. Currently on loan, I'll check it later.
  • P. Karlík, M. Nekula, J. Pleskalová (eds.), Encyklopedický slovník češtiny. On pp. 492–493, they propose the following transcription of Czech vowels in IPA (italics are theirs): [a,o,u,ɛ,ɪ,a:,ɛ:,i:,o:,u:,ou,au,eu] or [ou̯,au̯,eu̯].
  • P. Karlík, M. Nekula, Z. Rusínová (eds.), Příruční mluvnice češtiny. Only has a list of Czech vowel phonemes accompanied by a simple table (with no differences between short and long vowels shown other than length) on pp. 28–29, there is no discussion of their phonetic realization.
  • F. Čermák, Jazyk a jazykověda. Currently on loan, I'll check it later. However, this is a textbook of general linguistics with no emphasis on either phonetics or the Czech language; there's unlikely to be anything useful there wrt our problem. There's nothing relevant there, as I suspected. Updated: 09:40, 24 July 2009 (UTC)

I also looked at two other important relevant publications:

  • M. Dokulil, K. Horálek, J. Hůrková, M. Knappová (eds.), Mluvnice češtiny I, Academia, Praha, 1986. On p. 31: It follows from experimental work that the difference of values of F1 and F2 between short and long vowels makes about 5 %. However, influence of phonetic context on formant frequencies can be as much as 40 %. [...] In the case of long vowels, it is essentially a more perfect realization of the articulation position characteristic for the short vowel.
  • Z. Palková, Fonetika a fonologie češtiny, Karolinum, Praha, 1994, ISBN 80-7066-843-1. On p. 171, she writes, The difference between short and long vowels, which tends to be connected to a difference in vowel quality in [various] languages, is in Czech primarily a difference in duration, the colour of the vowels is essentially the same. The i-like vowels are an exception: short i is somewhat more open than long close í.

I conclude that user Confused monk makes baseless claims, the sources do not support the idea of there being any significant height difference between Czech long and short vowels, save for [ɪ, iː]. — Emil J. 15:28, 23 July 2009 (UTC)

Oh, and let us not forget about recommendation of the IPA itself:

  • International Phonetic Association, Handbook of the International Phonetic Association, Cambridge University Press, 1999, ISBN 9780521652360. Its chapter on Czech (pp. 70–73, written by Jana Dankovičová) uses /ɪ, ɛ, a, o, u, iː, ɛː, aː, oː, uː, ou, au, ɛu/ for Czech vowels.

— Emil J. 10:44, 24 July 2009 (UTC)

Pojď sem[edit]

"Consonant merging is perceived as careless at word boundaries, e.g. pojď sem (come here) realized as [pot͡sɛm]. It is necessary to pronounce all phonemes clearly and separately: [pojc.sɛm]."

But then, I have never heard a native speaker say "[pojc.sɛm]". To me the normal realisation is at best [pocsɛm], if not [pot͡sɛm] (though I find [c] and [t] quite difficult to distinguish in fast speech). It may not be super duper Czech but in reality, some kind of merging usually occurs.. - filelakeshoe 18:48, 26 May 2010 (UTC)

That's a good point, indeed. I think Pojď sem. is a poor choice, because its pronunciation as [pot͡sɛm], though not orthoepical is pretty lexicalized in everyday spoken Czech. We need to find a better example (I'm unable to come up with one now, though).--Pětušek 09:24, 20 September 2017 (UTC)