Curiosities of the Polish spelling

Part one

Previous part – Back to the first page

Even if the Polish spelling is phonetic at high degree, not always a given letter corresponds to one sound. I am presenting here some problems corresponding with this.

1. Three ways of writing of postalveolar sounds
2. Various meanings of the sequences with the letter i
3. Exceptional pronunciation of the digraphs
4. Ways of writing of the consonant [j]
5. Use of the letter i
6. Consequences of the morphological rule and the historical rule
7. A voice of a professional linguist on the Polish spelling

1. The 5 postalveolar-palatal sounds are spelt in three ways, depending on their position:

• with the letters ć, dź, ń, ś, ź before a consonant or word-finally;
• with the letters c, dz, n, s, z before i;
• with di- and trigraphs ci, dzi, ni, si, zi before any other vowels.

Examples:

• ćma, wiedźma, bańka, wieśniak, jeździsz;
• sieć, jedź, koń, wieś, maź;
• cichy, jeździsz, niski, siny, zima;
• ciuchy, dzień, wieśniak, sieć, ziąb.

2. The sequences of letters ci, dzi, ni, si, zi can denote:

1. one sound, i.e. a postalveolar-palatal consonant, e.g. ciocia, dzień, konie, siano, łazienka;
2. two sounds, i.e. a postalveolar-palatal consonant and the following vowel i, e.g. cioci, dziś, niszczyć, kosić, łazik;
3. two sounds, a dental consonant and the following vowel i, e.g. cibalgina, na cito, sinolog, sinus, silos, zin.

The instances a and b are the result of using rules presented in the point 1. The sequences ci, dzi, ni, si, zi compose di- or trigraphs (i.e. they denote one single sound) before a vowel, but they do not compose di- or trigraphs before a consonant and word-finally. In this position however they can denote:

• [ć+i], [dź+i], [ń+i], [ś+i], [ź+i] in native and assimilated words;
• [c+i], [s+i], [z+i] in loanwords; [n+i] is not possible at all (the pronunciation [ń+i] is always obligatory), whereas [dz+i] is extremally rare (e.g. in the proper name Dzibanitum coming from a certain computer game) and it is being replaced with [dź+i] like in Godzilla;
• [zji] or [z’i] (cf. 4; [’] denotes the laryngeal stop here) occurs on the border of a prefix and a stem, e.g. ziścić [z’iścić].

Unfortunately, rules determining when the clusters ci, si, zi (before consonant and word-finally) should be read [ć+i], [ś+i], [ź+i], and when [c+i], [s+i], [z+i], do not exist. The spelling is not precise here and it does not let us foresee the pronunciation. Using intuition when estimating whether the word is considered to be native or foreign, whether the prefix z- does occur or not, can be deceptive, especially for a foreigner who learns Polish.

3. The clusters ch, cz, dz, dź, dż, rz, sz denote one sound as a rule, i.e. they are digraphs, e.g. kochać, czuwać, sadza, dżokej, chodź, marzyć, kosz. It should however be remembered that:

• in some words the rz denotes two sounds, [r+z], e.g. marznąć, zmarzlina, obmierzły, mirza, superzespół, erzjański, Murzasichle (in this place-name the si should be read “normally”, i.e. as [ś+i], while rz does not denote the digraph rz here, but 2 sounds, [r+z]); the spelling likes to be deceitful here, cf. przemarzać [r+z] – przekomarzać [ż];
• the sequence rz can denote two sounds [r+ź] (before i), e.g. mierzić;
• the sequence rzi can denote two sounds [r+ź] (before a vowel), e.g. mirzie;
• sometimes dz, dż denote two sounds ([d+z], [d+ż]), e.g. podzelować, odznaka, nadżerka; dictionaries do not note words with dź meaning [d+ź] but such a sequence should be able to occur, e.g. in the geographic name Przedźródle;
• sometimes dz denotes [d+ź] (before i), e.g. przedzimie;
• sometimes dzi denotes [d+ź], e.g. podziemny.

Non-existence of single letters for the sounds [dz], [dź], [dż] makes impossible to distinguish these africates from the clusters [d+z], [d+ź], [d+ż], which usually occur on the border of a prefix and a stem. However there are not rules making it possible to determine when we really have a prefix in the word. Hence similar words can have different pronunciation, e.g. nadziemny [d+ź] – nadzieja [dź], przedzimie [d+ź] – przedziwny [dź], przedzamcze [d+z] – wyprzedzam [dz], podzwrotnikowy [d+z] – podzwaniać [dz].

4. To write the consonant [j], we use:

• the letter j word-initially, after a vowel, after c, j, s, z, also after d belonging to a prefix, e.g. jabłko, mijać, kolacja, rubajjat, misja, Azja, nadjechać;
• the letter i after b, ch, cz, d, dż, f, g, h, k, l, m, p, r, sz, t, w, ż, e.g. Arabia, marchia, welwiczia, dieta, lodżia, fiord, Rugia, hiena, makia, kalia, mumia, Pius, awaria, Dżanaszia, kutia, rewia, ażio;
• inconsistently, the spelling j is applied after cz, sz in particular words: Czuwaszja, welwiczja, glediczja (or welwiczia, glediczia);
• the letter i after the n, pronounced always as [ń], e.g. unia, niala (‘nyala’, a genus of antelope);
• occuring of [j] after ć, dź, ś, ź is limited to several foreign words, the spelling j is applied here, e.g. wajśja;
• after dz, ł, [n], rz there cannot be the [j] in Polish.

5. The letter i can denote:

1. the vowel [i], e.g. sinus, cirrus;
2. the vowel [i] and the changed pronunciation of the preceding consonant at the same time, e.g. siny, cisza;
3. the consonant [j], e.g. historia, hiena, kopia; the pronunciation of the preceding consonant usually changes slightly, except the n which is replaced with [ń] instead, e.g. mania pronounced [mańja];
4. only a change of the pronunciation of the preceding consonant, e.g. kiedy, siedzieć;
5. “the palatalizing element” [′], e.g. biały, wiara, łamię; in simplification we can suppose that the palatalization of the preceding labial consonant simply occurs here, just like in the previous sub-point;
6. the sequence of the sounds “juncture” + [i] realized in practice as [ji] or [’i] ([’] denotes here the laryngeal stop while the [′] is the mark of palatalization), e.g. ziścić, szyi, wyimaginowany.

In principle all sounds preceding the i, regardless of what is denoted with this letter, undergo some softening (palatalization). The biggest changes concern c, dz, n, s, z (rzeka) – if the i follows these letters, they usually denote completely different sounds, [ć], [dź], [ń], [ś], [ź]. It always concerns the n. In few loanwords, the c, dz, s, z preserve their pronunciation and undergo only slightly palatalization (like in sinus). The spelling does not give any instructions, which pronunciation occurs in the given word.

The letter [i] denotes a vowel when before a consonant and word-finally. There exists a small group of words however, in which the letter i denotes the sound [j] before a consonant or word-finally, if it is used after a vowel, e.g. mozaika (pronunciation [mozajka]). In the prevailing part of words with such a sequence (a vowel + i + a consonant or the end of the word), the letter i preserves the reading of the vowel [i], or as a matter of fact the sequence “the juncture” (see above) + i, e.g. bonsai, zaimek, ziścić, żmii, statui.

Before a vowel, the i can denote the changed pronunciation of the preceding sound, “the palatalizing element” or the consonant [j]. And so, in such a position the letter i denotes:

• always the changed pronunciation of the preceding c, dz, s, z into [ć], [dź], [ś], [ź] respectively, e.g. ciepły, dzień, siew, zioło;
• often the changed pronunciation of the preceding n into ń, e.g. niania, Mania;
• often the changed pronunciation (palatalization) of the preceding g, k, e.g. giąć, biegiem, kiedy;
• always the consonant [j] after ch, cz, d, dż, h, l, r, sz, t, ż, e.g. ischias, glediczia, Indianie, lodżia, hierarchia, liana, Maria, Dżanaszia, tiul;
• some people read the palatalized [h′] however without the following [j] in certain (not all) instances when chi or hi is spelt, e.g. hierarchia can be read [h′erarhja]; the spelling difference chi – hi has nothing to do with the difference in pronunciation [h′] – [hj]; it seems that the [h′] occurs in pre-stressed syllables (e.g. hieroglif [h′e'roglif]) while the [hj] in stressed (e.g. hiena ['hjena]) and post-stressed ones (e.g. marchia ['marhja]);
• often “the palatalizing element” after b, f, m, p, w; in practice we can tell here about soft varieties of these labial consonants (phonology) in spite of the clearly asynchronous pronunciation giving right to separate two succeeding elements (phonetics), e.g. zarabia, trafię, ramię, piach, trzewia;
• the consonant [j] after b, f, g, k, m, p, w in other words, e.g. Arabia, fiołek, magia, kiur, ramia, opium, rewia;
• the consonant [j] after n in certain words with changing it into ń at the same time, e.g. niala, mania;
• we do not meet the i after ć, dź, j, ł, ń, rz, ś, ź and before a vowel.

And so, if the A marks a vowel of your choice:

• the writings ćiA, dźiA, jiA, łiA, ńiA, śiA, źiA are inadmissible, also the words with the rziA group do not exist;
• the writings ciA, dziA, siA, ziA are univocal; the i always denotes only changed pronunciation in them and it does not denote any sound;
• the writings cziA, diA, dżiA, liA, riA, sziA, tiA, żiA also are univocal, and i denotes here the consonant [j];
• the writings biA, chiA, fiA, giA, hiA, kiA, miA, niA, piA, wiA are not univocal; in some words the i denotes the [j] as well as additional (obligatory) palatalization (especially distinct in the instance of the n), in others it is only a sign of palatalization.

There exist pairs of words of a very similar shape, in which one contains the [j], while the other only palatalization, e.g. Arabia – zarabia, Albania – bania, rewia – trzewia, and even pairs of the same spelling but different pronunciation, e.g. mania ‘mania, fixed idea’ – Mania ‘diminutive form of Maria’, ramię ‘accusative of ramia, the plant Boehmeria nivea’ – ramię ‘arm’. Dictionaries usually ignore this problem unfortunately.

It should be noticed that in some form there happen the sequences ii, marking 2 [i] vowels, separated with “the juncture”, e.g. żmii [żmiji ~ żmi‘i], rakii; in other forms of this type the first i denotes the consonant [j], e.g. ramii [ramji]. Hence the presence of the ~ii ending in inflexional forms of the word is an indication concerning its pronunciation (and vice versa) since the spelling permits such an ending only in words with the consonant [j] pronounced (in the case of żmii also written in other forms: żmija).

However there exists a number of words, in which the pronunciation fluctuates or even the form without the [j] dominates in spite of the spelling. E.g. in the speech of the older generation you can still hear [bijologja] (with the juncture, however such a pronunciation one can hardly acknowledge to be correct), one can often hear [bjologja] (with the iota), but more and more often also [b’.ologja] (only with palatalization). Similarly apart from the older [mańjak] (spelt maniak), perhaps it dominates [mańak] today, nevertheless one can still hear mania [mańja]. Apart from older kiosk [kjosk], the [k′osk] becomes common. The iota disappears (or it has disappeared) mainly after groups of consonants, e.g. in such words like armia, relikwia, szałwia, Warmia. The obvious result of it is the pronunciation of their declension forms [armi, relikfi, szałwi, Warmi] instead of the conservative forms with the iota suggested by the spelling armii, relikwii, szałwii, Warmii.

The pronunciation with [j] preserves pretty well in names of countries, foreign lands etc. (e.g. Albania, Arabia, Hiszpania, Kenia, Norwegia, Szwabia), as well as in words less frequently used. Moreover, sometimes the non-etymological [j] is being inserted in rare native words, like konopie (instead of the historically correct pronunciation [konop′e] one can hear [konopje]), hence the false form konopii, suggested by the pronunciation (the spelling permits only konopi), happens (even in scientific works). Maybe with the time it will be laid down the rule saying that after one (internal) consonant [j] can happen, and after two (or more) – palatalization only? Another limitation should be also considered. And so, the palatalized [g′, k′] can occur only before certain vowels, namely before i, e, ę, ą. Because during declension other vowels can occur, preserving of the [j] does not permit occurring the palatalized velar in non-typical environment and all words ending with ~gia, ~kia contain the pronounced iota. In the often used word kiosk, a seemingly exceptional reduction of the [j] before [o], being an invariable part of the stem, has taken place. In fact the presence of the prevelar [k′] immediately before [o] is not so unusual at all, but in giąć [g′ońć] also [o] occurs, in spite of suggestions of the spelling!

Main page – Polish grammar

2008-02-21