Discussion on the way of presentation of the declension

Previous partBack to the first page

Even if my website is to have popularized scientific character and not scientific one, I have been inclined to write this motive by many reflexions which arose after studying the bibliography of the subject. As in many cases I do not copy the substance and the conceptions included somewhere else, but I rather make my own solutions, perhaps I ought to say these several words to all the interested ones.

There is not one standard method of presenting of diversity of declension types of Polish nouns. To avoid complications, I limit myself only to substantives here.

Old grammarians tried to take Latin authors as their models and they divided substantives into declensions depending on the genitive singular ending. Such a division however, even if it proves correct in a description of Latin, remains little practical in the Polish language for here the set of endings depends in great part on the gender of the substantive. Such a criterion is simply more comfortable. It has been introduced on this website with full consistency.

Jan Tokarski, the author of widely used now conception of division of substantives, distinguished declension groups in each of the genders depending on the final stem sound:

  1. the soft-stemmed substantives (with stems ending in b′, ć, dź, f′, j, l, m′, ń, p′, ś, w′, ź);
  2. the soft-stemmed hardened substantives (with stems ending in c, cz, dz, dż, rz, sz, ż);
  3. the substantives with stems ending in velar consonants (k, g, ch, h);
  4. the other hard-stemmed substantives (with stems ending in b, d, f, ł, m, n, p, r, s, t, w, z).

The additional groups in this classification (V, VI) are determined with substantives with stem expansions and those with special declension (of the type muzeum).

This very accurate, in my opinion, conception contains however some shortcommings. Even a rough analysis of the declension patterns shows for instance that substantives of the ż III group with stems ending in ch, h are declinated after the group IV and not like the others from the group III. Also the groups I and IV can be easily divided into less, distinct subgroups. It sentenced me on necessity to apply my own solution: I applied the modified Tokarski’s conception by distinguishing not four but thirteen classes of substantives. I numbered them contrary to Tokarski, but with agreement with the tradition of Slavists – first hard-, next soft-stemmed substantives. In numbering declension types I have used the alphabetic order of endings as the criterion. Despite of tradition, the neuter gender precedes the feminine gender – it is so because of more morphologic similarity between neuter and masculine.

I also agreed – contrary to Tokarski – with very well motivated, in my opinion, thesis that there exist five, not three genders in Polish. This thesis is now accepted by many present-day grammarians. Evidently it gives birth to nasty consequences in the shape of increase of the quantity of the paradigms. It releases anyone from a necessarity to assume whether a given substantive is personal, animate or inanimate. Unfortunately it is not so obvious (e.g. trup ‘dead body, corpse’ is animate (!) in Polish, whereas anioł ‘angel’ was a person sometime, today it has become an animal – naturally from the point of view of the language).

In my opinion the following can be also numbered among separate groups: feminine substantives ending in a consonant, substantives with a stem expansion, masculine-personal with feminine declension (mixed as a matter of fact – cf. D pl poetów with the ending “prohibited” in the feminine gender), as well as those with adjective declension. This is not an unimportant and exceptional group. It is very hard to delimit it – many adjectives in fact may become substantive in a given statement. There are however nouns having shape of adjectives but used only as substantives (e.g. luty, królowa, myśliwy) or being adjectives but having also a constant and precise substantive meaning (e.g. chory). The distinguishing of so many (not less than 15) groups of substantives may terrify but in my opinion this describes language features the best and this does not push relatively frequent phenomena into the margin of the grammar system.

A very interesting conception of classification of substantives has been presented by Włodzimierz Gruszczyński (see the bibliography). Some of his solutions are controversial however (anyway, he reports many critical remarks of the reviewers by himself). It is uncommon to rely in assuming only on the written form of words. Among others the symbols with two letters like ni, ki, si (describing single sounds in some contexts) are treated by Gruszczyński like simple compositions of two letters, which are separated then between the stem and the ending. Maybe it would be acceptable, however I am surprised that the same author admits reasons for existence of the digraphs like rz, sz! Separating of the stem as the longest possible repeatable sequence of symbols is a very false conception which does not take singularities of the Polish spelling into account and which gives birth to consequences unexpectable and surprising. According to Gruszczyński the form of D lm kowali consists of the stem kowal~ and the ending ~i while another, built in similar way form of D lm śledzi is a stem with the zero ending according to him. I will not continue with presenting more critique of his work but I will only inform of three requirements that suggest themselves to me:

  1. The written shape and the spoken shape are tied strictly (more than in English, that’s for sure!) and all attempts of such a servile following of the spelling only, ignoring the function of characters (especially of the character i), will be always unnatural and unhandy.
  2. Even when completely ignoring the pronunciation all digraphs – like rz as well as like ni – should be treated in the same manner – i.e. should be all acknowledged or all not acknowledged. In the Polish language both those types are not univocal. I will remind here e.g. the words marznąć and odzew where the rz, dz are not digraphs, the same like e.g. the ni in the form słoni is not a digraph (but it is a digraph in the form słonia).
  3. Seeing that in the language consciousness the ~i serves the same function in the form kowali as well as in the form śledzi, there is no sense – and it is with detriment to the system – to treat them differently. Perhaps then we can have less variant forms of the stem (only in writing!) but more diversity of endings instead.

For these reasons, even if I consider some elements of the Gruszczyński’s work to be worthy of notice, I do not foretell the unearthly future to certain of his concepts. This author postulates besides to give in dictionaries a symbolic marking of the type of declension which each single substantive belongs to. His criteria of the division into groups are pretty “exotic” – in the first row it is a presence of identical forms of different cases. For me, on the other hand, the existence of formal equalizations remains secondary towards more natural criteria like the gender and the sound ending the stem. I am convinced that using of the great number of not enough intelligible symbols is inattractive for the reader. Many of forms of a substantive can be easily foreseen from its shape and gender without any additional symbols. Even if I also use here an extended number of symbols, I try to bind it in the shape of words rather. All the same the Gruszczyński’s symbols seem to be interesting from the theoretical point of view and that is why I have decided to present them in my tables of declension.

Even if I have ambition to present the declension of the Polish substantives exhaustively, some types given by Gruszczyński are not present in my tables. The reason is that his work is based on the monumental dictionary “Słownik języka polskiego” edited by Doroszewski which contains a large amount of dialectal and out of date words and forms, completely not used today (or used in other forms), and even unknown to me or whose meaning might be only assumed by me, like pałąg, reż, szczebrzuch or żomp (D lp żompia – perhaps I met the last one in regional spoken language – with meaning rzęch, grat, rupieć ‘bone-shaker, piece of junk’, but its spelling is astounding me – I would rather write rząp! – such a form was found by me in another source, as well as żąp). On a scale of common nouns I have limited myself in general to words present in the 3-volumed dictionary edited by Szymczak (see the bibliography) after regarding modern dictionaries of the Polish orthoepy and my own feeling of the language.

And here is the meaning of the symbols (endings and marks using here are the same as on other pages of this website):

symbol meaning and examples
A masculine inanimate; M lp = B lp (ul), M lm = B lm (ule), Ms lp = W lp (ulu)
A' masculine-personal; D lp = B lp (studenta), D lm = B lm (studentów), Ms lp = W lp (studencie)
A'' masculine animate; D lp = B lp (słonia), M lm = B lm (słonie), Ms lp = W lp (słoniu)
Au like A, but D lp = Ms lp = W lp (budyń, metalu, rodzaju)
Av, A'v, A''v like A, A', A'', but Ms lp <> W lp (głupcu, głupcze)
A'c like A', but M lm = Mx lm (ci goście, te goście)
A'd like A', but Ms lp = W lp = M lm (cyganie)
A'e like A', but C lp = Ms lp (panu) <> W lp (panie)
B feminine (also masculine animate); C lp = Ms lp (kobiecie), D lp = M lm = B lm (kobiety)
B' masculine-personal; C lp = Ms lp (poecie), D lp = Mx lm (poety) <> M lm (poeci) <> D lm = B lm (poetów)
Ba like B, but D lp (szansy) <> M lm = B lm (szanse)
B'a like B', but M lp = W lp (tato)
Bb like B, but C lp = M lm = B lm (ręce)
C feminine; D lp = C lp = Ms lp = D lm (stacji)3 <> M lm = B lm (stacje)
C' masculine-personal; D lp = C lp = Ms lp (bacy)
Ca like C, but D lp = C lp = Ms lp (pracy) <> D lm (prac)
C'a like C', but D lp = M lm (zdrajcy)
Cb like C, but do not build the extra form of D lm3
C'b like C', but M lm = Mx lm (woźnice, kaznodzieje)
Cc like C, but M lp = D lp = C lp = Ms lp = W lp (gospodyni)
like Cc, plus B lp = N lp (panią)
D feminine; M lp = B lp (skroń), D lp = C lp = Ms lp = W lp = D lm (skroni), M lm = B lm (skronie)
D' masculine-personal (old); M lp = B lp = W lp (waszmość), D lp = C lp = Ms lp (waszmości)
Da like D, but D lp = C lp = Ms lp = W lp = M lm = D lm = B lm (wszy)
E neuter; M lp = B lp = W lp (pole, udko), C lp = Ms lp (polu, udku), D lp = M lm = B lm (pola, udka)
E' masculine-personal5; D lm = B lm (chłopisków)
Ea, E'a like E, but C lp (bagnu, mikadowi) <> Ms lp (bagnie, mikadzie)
Eb, E'b like E, but D lp (kurczęcia, imienia, księcia) <> M lm = B lm (kurczęta, imiona, książęta)
Ec like E, but M lm = D lp (dzieci)
G neuter; they have the same form in all the cases of lp (muzeum)
1, 2, … particular patterns of declension
A1 MsW lp ~u, Mx lm ~i, D lm ~ów (storczyk, elektryk, strzyżyk)
A2 MsW lp ~e, Mx lm ~y, D lm ~ów (młot, emeryt, kogut)
A3 MsW lp ~u, Mx lm ~e, D lm ~ów (kloc, kibic, kuc)
A4 MsW lp ~u, Mx lm ~e, D lm ~i (hacel, kowal, daniel)1
A5 MsW lp ~u, Mx lm ~e, D lm ~i (według autora: ~0; staj, mazgaj, cygaj)1
A6 MsW lp ~u, Mx lm ~e, D lm ~i (według autora: ~0; cierń, gamoń, śledź)1
B1 D lp = Mx lm ~i, M lm m-os ~y (lektyka, kolega)
B2 D lp = Mx lm ~y, M lm m-os ~i (według autora: ~0; kobieta, poeta)
C1 D lp = C lp = Ms lp = D lm ż ~i (funkcja, Aria)4
C2 D lp = C lp = Ms lp = D lm ż ~y (łosza, baca)
C3 D lp = C lp = Ms lp = D lm ż ~i (idylla)4
C4 D lp = C lp = Ms lp = D lm ż ~i (według autora: ~0; kopalnia)4
C5 D lp = C lp = Ms lp = D lm ż ~i (według autora: ~0; tuja)4
D3 D lp = C lp = Ms lp = W lp = D lm ż ~y (noc)
D4 D lp = C lp = Ms lp = W lp = D lm ż ~i (gardziel)1
D5 D lp = C lp = Ms lp = W lp = D lm ż ~y (według autora: ~0; kolej)1
D6 D lp = C lp = Ms lp = W lp = D lm ż ~y (według autora: ~0; skroń, waszmość)1
E1 D lm ~0 (skupisk)6
E2 D lm ~0 (serc)6
E3 D lm ~y (naręczy)
E4 D lm ~i (osiedli)
E6 D lm ~0 (mień)6
N with the suffix ~in~ (dominikanin : dominikanie) or ~en / on~ (plemię : plemienia : plemion)
T with the suffix ~ęć / ęt / ąt~ (kurczę : kurczęcia : kurczęta : kurcząt)
* floating (moving) e (kotek : kotki)
ó the alternation ó : o (słój : słoje)
ą the alternation ą : ę (wąż : węże)
the alternation ą : ę; ę only before ~y and ~mi (miesiąc : miesięcy, pieniądz : pieniędzmi)
^ floating o (in the original marked with the capital delta; kocioł : kotły, kozioł : kozły, osioł : osły)
+e the alternation a : e (kwiat : kwiecie, las : lesie) or o : e (anioł : aniele)
,Y M lm ~y with the r : rz alternation in the stem (aktor : aktorzy)
,I M lm ~i with the ł : l alternation in the stem (Hucuł : Huculi)
,& M lm ~owie (mistrzowie, szejkowie)
,{&} only Mx lm, not building the masculine-personal M lm form, potentially ~owie (karzeł – te karły)
,u D lp ~u (only A2: konfliktu, zjazdu, zbioru; the others – Au)
C lp ~u (chłopu, kotu, światu)
,H MsW lp ~u (when D lp ~a; brzuchu, mnichu, molochu)
,$ M lm ~e without stem changes (kwadranse, preferanse)
,a M lm ~a (in the original marked with the small alpha; cuda)
,% D lm ~y / ~i or ~ów (symbol not used here)
,y D lm ~y (only A3: kaloszy, kolejarzy, kosarzy)
,ów D lm ~ów (only A6: chlebusiów, kiciusiów, uczniów)
,MI N lm ~mi (liśćmi, pieniędzmi)
,! non-typical forma W lp ~u (ciociu, matulu)
,# irregular lm (bracia, księża, przyjaciele)
F8:-i non-typical ending, in this case of M lm; I give endings in my interpretation2
() alternative forms


  1. The author distinguishes types A4, A5 and A6 as well as D4, D5 and D6 basing on his interpretation of the terms “stem” and “ending”. The differences are explained by the following examples: kowal~0 : kowal~a : kowal~i (A4), mazgaj~0 : mazgaj~a : mazgai~0 (A5), śledź~0 : śledzi~a : śledzi~0 (A6); gardziel~0 : gardziel~i : gardziel~e (D4), kolej : kolei~0 : kolej~e (D5), skroń : skroni~0 : skroni~e (D6). As one can see, there is one stem shape in the types A4, D4, two shapes in the types A5, D5 and A6, D6 with different distribution (according to the author). When taking my (classic) interpretation of the stem and the ending into account, it would be enough one type A4 and D4 together with a set of rules of stem changes, depending on its termination in a simple way. In my opinion it proves the unpractical sense of his solution, being not in conformity to (good) tradition.
  2. The numbering of the case forms used in symbols: F2 – D lp, F7 – W lp, F8 – M lm, F10 – D lm.
  3. The author considers here furthermore special, additional forms of genitive plural of substantives of that type (e.g. stacyj). In my opinion such forms are flagrantly archaic today and that is why I omit describing them as living forms in principle on this website. In this connexion I believe that the distinguishing of the additional type Cb is needless.
  4. The distinguishing of the types C1, C3, C4 and C5 as well Ca3, Ca4, Ca5 is linked to the different interpretation of the terms “stem” and “ending” and to considering of alternative forms of genitive plural of feminine nouns. The differences are explained by the following examples: funkcj~a : funkcj~i : funkcj~i / funkcyj~0 (C1), idyll~a : idyll~i : idyll~i / idyll~0 (C3), bil~a : bil~i : bil~0 (Ca3), kopalni~a : kopalni~0 : kopalni~0 / kopalń~0 (C4), bani~a : bani~0 : bań~0 (Ca4), tuj~a : tui~0 : tui~0 / tuj~0 (C5), szyj~a : szyi~0 : szyj~0 (Ca5). Accepting my interpretation, all these examples could be reduced to the types C1 and Ca1; the form funkcyj is obsolete today (only one D lm funkcji, equal to D lp, remains – the type C1). I would not use the form idyll (idylla – C1), then the D lm kopalni instead of kopalń would be less natural for me (kopalnia – Ca1). Regarding the alternation tuj / tui (I would like the first form personally), I would write tuja – C(a)1.
  5. In fact substantives of the type E' are bi-gendered (m-os., n) or even tri-gendered (m-os., m-żyw., n: widzę to chłopisko – widzę te chłopiska, widzę tych chłopisków).
  6. The type E5 does not exist; the distinguishing of the types E1, E2 and E6 is connected with differences in building the forms of N lp and D lm Accepting the interpretation of the terms “stem” and “ending” like the author, we have skupisk~o : skupisk~iem : skupisk~0 (E1), serc~e : serc~em : serc~0 (E2), mieni~e : mieni~em : mień~0 (E6).


Main pagePolish grammar