A Grammar of the Polish Language
Previous part – Back to the first page
Consonants – general remarks
First of all it is useful to divide the Polish consonants into plain ones and palatalized ones (pronounced with some additional raising of the dorsal part of the tongue towards the palate like when pronouncing the i). Many linguists use here the terms “hard” and “soft”, what can be deceiving as the same terms are used for “stops” and “spirants” respectively in the Anglo-Saxon countries. To understand morphological phenomena you should also know that that division took place quite somewhere else in the Old Polish language (on a certain optionally separated stage of the evolution of the language). We could divide all the consonants up into 4 groups:
- Ordinary, non-palatalized now or in the past. These are b, ch, d, f, g, h, k, ł, m, n, p, r, s, t, w, z.
- Palatalized in the past, non-palatalized now. These are c, cz, dz, dż, l, rz, sz, ż.
- Palatalized now, non-palatalized in Old Polish. These are gi, ki as well as the varieties of the ch, h standing before the i.
- Palatalized now as well as in the past. These are bi, ci/ć, dzi/dź, fi, j, mi, ni/ń, pi, si/ś, wi, zi/ź.
The alternations of the consonants non-palatalized – palatalized are of great importance in the morphology. We have 3 cases here:
- Both elements of the pair have not changed their features: b – bi, d – dzi/dź, f – fi, m – mi, n – ni/ń, p – pi, s – si/ś, t – ci/ć, w – wi, z – zi/ź.
- Both elements are “hard” today: ł – l, r – rz.
- The velars behave themselves particularly, because they have been submitted to different palatalizations several times; we can find here the following alternations:
- ch – sz, ch+i;
- g – ż, dz, gi;
- h – ż, dz;
- k – cz, c, ki.
More details on palatalization could be found here.
Another division lets us distinguish sonorants (together with semivowels) and proper consonants (obstruents), and voiceless and voiced consonants among the last ones. Sonorants are marked with j, l, ł, m, n, ń, r, sometimes with i, u; the “ogonek” (“tail”) – the second element of the diphthongs ą, ę also belongs to this class. From the morphological view the rz (not ż!) and w are close to sonorants because of their origin. Sonorants are usually pronounced as voiced with exception when between voiceless consonants (piosnka, kosmki, krtań), word-initially before a voiceless (rtęć, mknąć) or word-finally after a consonant, especially voiceless (wiatr, myśl, piekł). Even then they become voiceless only in fast speech however.
The Polish sonorants never make a separate syllable as opposed to Czech or English (e.g. rythm). When not adjoining any vowels they enter into the composition of hardly-pronounceable for a foreigner clusters of a number of consonants, e.g. Strwiąż, Brda, lśnić, łza, rytm, rdza (all these words are monosyllabic). The j is not allowed to occur in such a position at all, whereas the ł and u must be differentiated: spiekł – spieku, łzy – uzysk (the final ł after a consonant can be even omitted in fast speech, but not everyone says it is correct).
Here are detailed remarks:
- the j (also written i after a consonant and before a vowel in some words – this way we avoid writing j after certain consonants) resembles the English y, e.g. jajo, kolacja, agresja, Azja; as a palatal sound it is beside the division into palatalized and non-palatalized;
- the l resembles the British “light” l, e.g. lala; it belongs to historically palatalized sounds, thus it cannot be followed by the y practically; today it has a “hard” pronunciation, only before the i it is palatalized to a certain degree, e.g. lina; the group li before a vowel always means lj, e.g. kalia;
- the ł (e.g. lała) is the historically non-palatalized counterpart of the l; today it is pronounced like the English w however, the “scenic” dental pronunciation (resembling the English “dark” l) is obsolete and absolutely unfamiliar for most of Poles, even if already a short time ago it was recommended for actors, speakers etc.; except few loanwords this sound is not followed by the i; in some loanwords it is spelt u (auto, Europa), exceptionally also w (weekend) – of course in these cases the dental pronunciation ł is and was impossible;
- the m occurs as “hard” as well as “soft”; the palatalized variant can occur only before vowels; when before consonants and word-finally, historically palatalized m underwent depalatalization; before the y it is always “hard”, e.g. myła; the palatalized variant occurs before i, e.g. miła, before the other vowels it is spelt with the digraph mi, e.g. miała, what it is not quite out of right considering its asynchronous pronunciation – it means that in fact the mi represents (in accordance with the spelling!) the sequence of two sounds (always belonging to the same syllable): the m and a “palatalizing element” (different than the j); the group mi sometimes means mj in loanwords, where either sound may belong to a separate syllable, e.g. amia [am-ja];
- the n is “hard” while the ń is “soft”; the y can occurs only after the non-palatalized n; the spelling lets write ń as usually only before consonants and word-finally, e.g. koń, końmi; before the i it is sufficient to write n because the non-palatalized variant is not possible here at all, e.g. koni [końi]; before other vowels the digraph ni is used, e.g. konie [końe]; in borrowings the group ni can mark ńj (never just nj) where either sound may belong to a separate syllable, e.g. Dania [dań-ja];
- the r is non-palatalized in principle, its “soft” counterpart was the rz; the r is a sound pronounced like in Russian, Italian, Spanish or Scots; the realization of the r like the English, French or uvular German one may be received as a grating fault of speaking; normally it can be followed by the y, e.g. ryta, in loanwords also by the i (then the r undergoes some palatalization), e.g. Rita; the group ri before a vowel means rj, e.g. aria [ar-ja].
Voiceless consonants, as well as in other Slavic languages, are pronounced without vibration of the vocal cords, they are longer than voiced ones, but practically they are not pronounced with more force, in this connexion they have no aspitation either. The distinction between the Polish t and d is in something else than between the English or German t and d, which differ first of all in force (in Germanic languages the t is stronger than the d and aspirated as a rule).
The principle of assimilation of sounds which is very widespread in Polish introduces a number of restrictions in liberty of occurrence of the voiceless ones and the voiced ones (these limitations do not concern to sonorants):
- A voiced consonant can occur before a vowel (odejść), a sonorant (odłożyć) or another voiced consonant (odbić, odwaga, Afganistan [awgańistan]).
- Before a voiceless consonant and in the absolutely final position (before the pause and in the end of the sentence) only a voiceless consonant may occur (odkładać [otkładać]; hence the three words buk, bóg, Bug are pronounced identically – at the end of a pronouncement like [buk]).
- A voiceless consonant can occur before a vowel (sen), a sonorant (słodki), another voiceless consonant (spać, walizka [waliska]), in the absolutely final position (kasz, karz, każ, all pronounced here as [kasz]), as well as before the rz, w, which become voiceless (these sounds originate from sonorants, hence their particular behaviour: trzeba [tszeba], twój [tfuj]).
- Before a voiced consonant (except the rz, w) only a voiced consonant may occur (zadbać, także [tagże]).
The spelling denotes only assimilation of a rare type as well as occurring within variant forms of morphemes, e.g. dech ‘breath’ – tch-u ‘of breath’, deszcz ‘rain’ – dżdż-ownica ‘earthworm’, Sącz ‘name of a town’ – sądecki ‘of/from Sącz’, while it omits typical assimilations at the point of morpheme junction (in the internal sandhi), e.g. buzia [buźa] – buźka [buśka], wsuwać [fsuwać] – wsuwka [fsufka], kosić [kośić] – kośba [koźba].
In the course of speech assimilations at the point of word junction occur as well (the external sandhi, the liaison), as a rule in accordance with the above-mentioned principles. In this case however the initial rz, w do not become voiced, and on the contrary, like other voiced consonants they cause voicing the final consonant of the preceding word, e.g. podnieś rzeczy [podńeź żeczy], brat wyjechał [brad wyjechał]. When a word starts with a vowel or a sonorant and the preceding word ends in an obstruent (regardless of its spelling: brat ojca, brzeg jeziora), the obstruent is pronounced depending of the region as voiceless (Warszawa – Warsaw: [brat ojca, bżek jeźora]) or as voiced (Poznań, Kraków – Cracow: [brad ojca, bżeg jeźora]). Both styles of the liaison are considered to be correct. The initial vowel is preceded by the glottal (laryngeal) stop. A preposition leaves its final consonant voiced before a vowel in the literary language (the initial laryngeal stop may be very weak or absent in this case): pod okiem.
Detailed remarks on the pronunciation of consonants:
- The rz and ż are pronounced identically, even if they differ in their behaviour in the internal sandhi; in some dialects they are pronounced differently – in such or another way – it is however not accepted in the literary language.
- The ch and h are also pronounced identically; some authors of dictionaries and other descriptions forget it. I was learning Esperanto at one time: in the first lesson I was told that the seldom used h^ corresponds with the Polish ch, while the h is pronounced like the Polish h. It may have been true in Białystok in Zamenhof’s days – of the inventor of the language. To be sure, till the present day there has still existed a small group of Poles pronouncing the h in another way than the ch – as a laryngeal voiced spirant, but most of Poles cannot even imitate this pronunciation, and when they can hear the h pronounced in this manner, they cannot notice its separate character. There is not any aspiration in Polish – Poles must learn the pronunciation of the English or German h. A similar sound to the ch (but uvular, not velar) can be found in German, Greek, Scots. On the other hand the pharyngeal-laryngeal voiced h occurs in Upper Sorbian, Czech, Slovak, Ukrainian and Byelorussian. Similar sound (the pharyngeal `ayn) is known only in Arabic (e.g. in the word `Arabiyya), however the Slavic h is laryngeal rather than pharyngeal. One can sometimes hear a slightly similar sound in English when the intevocalic h is pronounced, e.g. in behind. The sound spelt h occurs only in borrowings (and in several onomatopeic words). It corresponds to the Common Slavic g.
- The w is pronounced like the English v and like the German w; when it is under assimilation, it resembles sonorants rather, what is a testimony of its origin (once, as long ago as in the pre-literary epoch, it was pronounced like the Eng. w).
- The c, ci/ć, cz, dz, dzi/dź, dż denote affricates, so simple sounds. Beside them there are the clusters of two sounds t+s, t+si/ś, t+sz/rz, d+z, d+zi/ź, d+ż/rz in Polish, which one should distinguish from the affricates (the spelling not always helps here!). In such clusters the t, d are pronounced usually like the adequate affricate which is still followed by the spirant (t+s > c+s etc.). It is usual to find the clusters of two affricates t/c+c, t/ć+ci/ć, t/cz+cz, d/dz+dz, d/dź+dzi/dź, d/dż+dż, e.g. dżdżownica ‘earthworm’. Each of the words czy ‘whether, if; or’ – trzy ‘three’ – czczy ‘empty; vain; futile’ is pronounced differently. Here are more examples: ociekał ‘he was dripping’ – odsiekał (dial.) ‘he was cutting off’ – odciekał ‘it was flowing away’ (ć – t+ś – t+ć), podział ‘division’ – poddział ‘subdivision’ (dź – d+dź), ocynkować ‘zinckify’ – odsypać ‘pour off a granular substance’ – odcyfrować ‘decipher’ (c – t+s – t+c), szczała ‘she was pissing’ – strzała ‘arrow’ (sz+cz – s/sz+t/cz+sz).
- Sometimes, on the border of a prefix and a stem, the groups dz, dż do not denote affricates but separate sounds rather, e.g. przedzimie ‘early winter’, odzew ‘response’, nadżerka ‘erosion’; also the rz rarely denotes r+z, e.g. marznąć ‘freeze, be frozen’, mierzić ‘disgust’ (we have r+ź here). The pronouncing of sz as [zz] occurs in the Polonised name of one of dinosaur genera: muszaur (its scientific name is Mussaurus).
- The Polish and English fricatives can be aligned according to their acoustic features and to their alternating place of articulation: Pol. = Eng. f – Eng. th – Pol. s – Eng. s – Pol. sz – Eng. sh – Pol. ś – Pol. chi – Pol. ch.
- The b, ł, m, p have labial articulation.
- The f, w have labiodental articulation.
- The c, d, dz, n, s, t, z have dental articulation: alveolar pronunciation resembling English is received as incorrect, lisping.
- The cz, dż, l, r, rz/ż, sz have alveolar articulation: the English s is between the Polish s and sz.
- The ci/ć, dzi/dź, ni/ń, si/ś, zi/ź have postdental articulation; the English sh is between the Polish sz and ś.
- The j is a palatal sound.
- The gi, ki as well as the ch/h before the i are prevelar, more front than their “hard” counterparts.
- The ch/h, g, k have velar articulation.
- If a vowel commences a syllable, the glottal (laryngeal) stop is often pronounced before it. In English similar phenomenon is rare, in some varieties of the language this sound is pronounced for the t when it ends a syllable. Such a syllable-final laryngeal stop happens in Polish in some words, e.g. in nie! ‘no’ when pronounced with firm tone.
- Polish is famous for existing in it other various complicated groups of consonants. Some examples of tongue-twisters: ssak, brzdąc, brzmieć, rżnąć, prztyczek (p+sz+t), drwal, drżeć (d+r+ż), grdyka, Gwda, zstąpić (s+s+t), ckliwy, tkliwy, trzcina (t/cz+sz+ć), wzdłuż, źdźbło, bezkształtny (s+k+sz+t), bezskrzydły (s+s+k+sz), bezwstydny (s+f+s+t), podstrzesze (t+s+t+sz or cz+sz+cz+sz), ćwierćton, społeczeństwo (ń+s+t+f), wewnętrzny (n+t/cz+sz+n), zmartwychwstanie (ch+f+s+t), embrion (m+b+r+j), barszcz, mistrz (s/sz+t/cz+sz). I recommend also the distich: W Szczebrzeszynie chrząszcz brzmi w trzcinie. More recorded examples can be found on this page. Nice tongue-twisting for all learning Polish by the author of this webpage. Share your impressions! I will add by way of consolation that there are languages with clusters containing a greater number of consonants, e.g. in Georgian words vpckvni ‘I peel’ lub gvprckvnis ‘he has exploited us (to be out of money)’.
- When we include the above remarks, we could say that the consonants b, d, f, g, k, l, m, n, p, s, t are pronounced quite like in English, moreover the j, ł, w resemble the English y, w, v respectively.
- The c, ch, cz, dz, h, j, r, rz, si, sz, w, zi have different pronunciation than in English.
Here you can find tables presenting a number of problems of the Polish phonetics with more details.
Another interpretation of the Polish phonetics may be found here.
Stress is called in Polish akcent ‘refrain’ and this is the right name for this phenomenon, because the Polish stress is first of all tonic (melodic), not only dynamic (which was suggested by the authors of many older works). The stressed syllable is articulated basically at a higher tone of speech (even if it is true that also with a little more intensity). Because of mainly tonic nature of stress there is no reductions in the unstressed syllable in Polish (compare to English, German and Russian).
As a rule the penultimate (second from the end) syllable is stressed in Polish. Here are some exceptions ("`" means stress on the following syllable):
- o`kolica, `reguła, `ryzyko, `opera, `czterysta, `osiemset and some others CAN be proparoxytona (stressed on the third syllable from the end), but paroxytonesis (stressing the penultimate) here is normal and rather often;
- `osiemkroć, w `ogóle, `muzyka, mate`matyka, `technika, `fizyka, `Korsyka etc. are more often proparoxytona, but you can stress the penultimate as well (it sounds more colloquial: `muzyka can mean ‘classical music’, mu`zyka can mean ‘pop music’ – however it is not a strict rule);
- there is a number of oxytona (stressing the last syllable), too, for example composites in the forms with the zero-ending – in nominative singular: eks`mąż, wice`mistrz, and in genitive plural: tych eks`żon, abbreviations PeKa`eS (for Państwowa Komunikacja Samochodowa), few others aku`rat (pronounced with expression, normally a`kurat);
- verbs in 1st and 2nd person of plural of the Past Tense and in the conditional are stressed as if the particle by and personal endings were separate words: `nieśliśmy, `nieślibyśmy, pi`saliśmy, pi`salibyśmy, also `byleby; this rule does not concern the other forms of the Past Tense: akcento`wałem, akcento`wałam, but akcen`towałbym, akcento`wałabym;
- a complex of a preposition with the following personal pronoun builds the whole, stressed as if it was one word: na`de mną, `bez niej;
- the negative particle nie makes the whole with the following verb, e.g. `nie ma, `nie rób `tego.
If you can understand Polish, you can compare my observations with the orthoepic norm given in Polish publications – go here.
Main page – Polish grammar