Polish being “ergative”? The riddle of “X NOT BE AT Y”-Constructions


Joanna Błaszczak (Potsdam)


I. In this talk I am going to make an unorthodox claim that Polish in some sense behaves like an ergative language. Moreover, I will argue that applying a kind of analysis proposed for ergative languages to Polish existential-locative constructions offers an elegant and simple solution to otherwise puzzling data.


II. The relevant piece of data is given in (1)-(2). In Polish there is one type of construction which bears a strong resemblance to the constructions found in split-ergative languages like Hindi or Georgian. In the latter languages a special Case marking, i.e., the ergative marking is triggered by a particular tense or aspect; cf. (1). Similarly in Polish, depending on the aspectual properties of the verb być ‘to be’ the “subject NP” [the term “subject” is used here in a purely descriptive pre-theoretical sense] is marked either for NOM or GEN; cf. (2).


III. The proposed analysis will proceed in a few steps. First, I will show that the GEN marking of the “subject NP” in (2a) cannot be simply subsumed under the general rule of Genitive of Negation in Polish (GoN) since (i) only direct objects of transitive verbs are subject to the rule of GoN in Polish (cf. e.g.Witkoś 1998), and (ii) unlike in (2), the aspect of the verb does not seem to have any influence on the Case marking of the direct object (irrespective of the aspectual properties of the verb, the object is marked for GEN in negated sentences); cf. (3). Neither can (2a) be explained by claiming that GEN in (2a) is GoN of the Russian sort. GoN in Russian shows a broader distribution, applying among other to subjects of “existential predicates”; cf. 4 (Babby 1980). However, as far as the influence of aspect on the Case marking is concerned, also in Russian there seems to be no direct connection between the aspect of the verb and the Case of its object (Pereltsvaig 1999 contra Timberlake 1975). On the contrary, Pereltsvaig claims that GoN can be used in Russian when the verb is used generically (‘habitual use’). However, this goes contra what we observe in (2).


Next, it will be shown that the facts in (2) cannot be explained by appealing to some special properties of existential constructions. As pointed out by Grzegorek (1984:107), in Polish other “notional verbs” can replace być ‘to be’ in existential constructions; cf. (5). However, unlike negated być no other “notional verb” requires its subject to be marked for GEN; cf. (6). (This contrasts with the situation found in Russian, as evidenced by (4)).


IV. To solve the puzzle posed by the data in (2) I will assume that (2a) displays an ergative structure known form (split)-ergative languages. Mahajan (1994, 2000) and Hoekstra (2000) among others take transitivity to be a derived property. The “ergative pattern”, normally restricted to perfective sentences, is claimed to be basically unaccusative. Ergative Case marking results from the the non-incorporated preposition (of the possessor/locative phrase). (In Accusative languages this preposition is incorporated into BE resulting in HAVE and NOM marking of the possessor; cf. Kayne 1993).


I will show that something along these lines is also going on in the Polish example (3a): BE is a non-Case-assigning predicate taking (a small clause consisting of) a NP as an internal argument and a PP as an external one. P(reposition) does not incorporated into BE. In affirmative sentences a NP can be assigned Case only by T (hence the NOM marking) since there is no other Case assigner in the clause; in negative sentences, however, there is a closer Case assigner than T, namely NEG(egation) (hence the GEN marking).


For this analysis to go through it must be shown that (i) there is an aspectual difference between być and bywać; desirably in terms of być being perfective (to explain why być but not bywać triggers an ergative pattern); and (ii) there is a structural difference between być and bywać: być being an unaccusative predicate and bywać being unergative; the “subject” NP would be accordingly in the former case an internal argument, and in the latter case an external argument. Given that GoN-rule applies only to internal arguments, the lack of the GEN marking in (2b) could be attributed to this fact. I will present evidence for these claims respectively.

(1)         a.        siitaa         ne      vah    ghar                 khariidaa                      (thaa)               Hindi

                        Sita-FEM-ERG  that   house-MASC  buy-PERF-MASC      be-PAST-MASC

                        ‘Sita had bought that house.’

             b.        siitaa                            vah       ghar     khariidegii                                           

                        Sita-FEM-NOM         that      house   buy-FUT-FEM

                        ‘Sita will buy that house.’


(2)         a.        Jana                 nie        było                                         na        przyjęciu.      Polish

                        John-GEN       NEG    BE-3.SG.NEUT.PAST           at         school

                        Lit.: ‘There was no John at the party.’ / ‘John was not at the party.’

             b.        Jan                   nie        bywał                                                  na        przyjęciach.

                        John-NOM     NEG    BE-3.SG.MASC.PAST.HABIT        at         parties

                        Lit.: ‘John was not at parties.’ / ‘John didn’t use to come to parties.’


(3)         a.        Nie      czytałam                                              tej gazety.

                        NEG    read-1.SG.FEM.PAST.IMPERF        [this newspaper]-GEN

                        ‘I didn’t read this newspaper

             b.        Nie      przeczytałam                                                   tej gazety

                        NEG    read-through-1.SG.FEM.PAST.PERF            [this newspaper]-GEN

                        ‘I didn’t read (completely) this newspaper.”

             c.        W młodości     nie        czytywałam                                         gazet.              

                        in  youth           NEG    read-1.SG.FEM.PAST.HABIT           newspapers-GEN

                        ‘In my youth I didn’t use to read newspapers.’


(4)         Zdes’  ne        voditsja            losej.

             here    NEG    roam-3.SG.PRES        elks-GEN

             ‘No elks roam here.’


(5)         Wzgórze         porastała                                 trawa.

             hill-ACC         grow-3.SG.FEM.PAST          grass-NOM.SG.FEM 

             ‘There was grass on the hill.’             


(6)         a.        Wzgórze          nie        porastała                                 trawa/*trawy.

                        hill-ACC          NEG    grow-3.SG.FEM.PAST          grass-NOM/*GEN    

                        ‘There was no grass on the hill.’                      

             a.’       *Wzgórze        nie        porastało                                 trawy.

                          hill-ACC        NEG    grow-3.SG.NEUTR.PAST      grass-GEN



Babby, L. (1980). Existential Sentences and Negation in Russian. Ann Arbor, MI: Karoma.

Grzegorek, M. (1984). Thematization in English and Polish. A Study in Word Order [= Uniwersytet im. Adama Mickiewicza w Poznaniu. Seria Filologia Angielska NR 18]. Poznań.

Hoekstra, T. (2000). The Nature of Verbs and Burzio’s Generalization. In: Reuland, E. (ed.) (2000), 57-78.

Kayne, R. (1993). Towards a modular theory of auxiliary selection. Studia Linguistica 47: 3-31.

Mahajan, A. (1994). The ergativity parameter: have-be alternation, word order and split ergativity. Proceedings of NELS 24: 317-331.

Mahajan, A. (2000). Oblique Subjects and Burzio’s Generalization. In: Reuland, E. (ed.) (2000), 79-102.

Pereltsvaig, A. (1999). The Genitive of Negation and Aspect in Russian. Ms., McGill University.

Timberlake, A. (1975). Hierarchies in the genitive of negation. SEEJ 19: 123-138.

Witkoś, J. (1998). The Syntax of Clitics: Steps towards a Minimalist Account. Poznań: motivex.